Datasets:

medarc

/

clinical_pile_v1_minhash_deduped

like 1

in_situ_formation_of_photoactive_b-ring_reduced_chlorophyll_isomer_in_photosynthetic_protein_lh2

## Abstract: Natural chlorophylls have a D-ring reduced chlorin π-system; however, no naturally occurring photosynthetically active B-ring reduced chlorins have been reported. Here we report a B-ring reduced chlorin, 17,18-didehydro-bacteriochlorophyll (BChl) a, produced by in situ oxidation of B800 bacteriochlorophyll (BChl) a in a light-harvesting protein LH2 from a purple photosynthetic bacterium Phaeospirillum molischianum. The regioselective oxidation of the B-ring of B800 BChl a is rationalized by its molecular orientation in the protein matrix. The formation of 17,18-didehydro-BChl a produced no change in the local structures and circular arrangement of the LH2 protein. The B-ring reduced 17,18-didehydro-BChl a functions as an energy donor in the LH2 protein. The photoactive B-ring reduced Chl isomer in LH2 will be helpful for understanding the photofunction and evolution of photosynthetic cyclic tetrapyrrole pigments. Cyclic tetrapyrroles with modified skeletons and peripheral groups have essential roles in various biofunctional proteins . Chlorophyll (Chl) molecules, involved in the solar-energy conversion processes of oxygenic photosynthesis, typically contain an unsymmetrical conjugated tetrapyrrole π-system, in which the C17-C18 bond in the D-ring is hydrogenated 1, . The D-ring reduced chlorin (17,18-dihydroporphyrin) skeleton is responsible for efficient light absorption in the visible portion of the solar spectrum. The in vivo conversion from the porphyrin macrocycle in the precursor of Chls, protochlorophyllide a, to the D-ring reduced chlorin ring is regio-and stereoselectively mediated by protochlorophyllide oxidoreductase (POR) . In photosynthetic bacteria, further hydrogenation of the C7 = C8 double bond in the B-ring of the chlorin macrocycle is mediated by chlorophyllide oxidoreductase (COR) to produce a bacteriochlorin (7,8,17,18-tetrahydroporphyrin), which leads to bacteriochlorophyll (BChl) a (Fig. 1A) . However, it remains unclear why natural phototrophs select D-ring reduced chlorin and both B-and D-rings reduced bacteriochlorin for light-harvesting and charge separation processes. B-ring reduced chlorin (7,8-dihydroporphyrin), in which the position of the hydrogenation in the cyclic tetrapyrrole is diagonally opposite to that in naturally occurring D-ring reduced chlorins, is a fascinating Chl isomer 11,12 . However, no photosynthetically active B-ring reduced chlorins have been discovered in nature, likely because the order of the ring reductions in the (B)Chl biosynthetic pathway is determined by the substrate specificities of POR and COR 5,13 . Substitution of D-ring reduced natural Chls with corresponding artificial B-ring reduced chlorin isomers in photosynthetic proteins provides new insights into the photofunctional mechanisms of Chl-polypeptide complexes. Additionally, this substitution might indicate the reason for selection of the D-ring reduced chlorin π-system in photosynthetic evolution. Herein, we report the in situ formation of a B-ring reduced chlorin pigment, 17,18-didehydro-BChl a (1, Fig. 1B), from BChl a in light-harvesting protein LH2 from a purple photosynthetic bacterium Phaeospirillum molischianum (denoted as molischianum-LH2) by treatment of this protein with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ). LH2 proteins are light-harvesting pigment-polypeptide complexes in purple photosynthetic bacteria. The BChl a pigments are circularly arranged by support of transmembrane α-and β-polypeptides in LH2 proteins . BChl a pigments in LH2 are classified into two types, B800 and B850, on the basis of the peak positions of their lowest energy absorption bands called Q y bands. Intracomplex excitation energy transfer (EET) occurs efficiently from B800 to B850 BChl a 17,18 . Here, we focus on B800 BChl a, which is sandwiched between the outer β-polypeptides in LH2 proteins. In molischianum-LH2, the D-ring of B800 BChl a is oriented toward the surface of the protein matrix (Fig. S1A) 14 . This molecular orientation markedly differs from that in the other structural-determined LH2 protein from Rhodoblastus acidophilus (denoted as acidophilus-LH2), where the D-ring in B800 BChl a is embedded in the protein matrix (Fig. S1B) 15,16 . The B800 geometry in molischianum-LH2 shows promise for producing a photoactive B-ring reduced chlorin by selective in situ oxidation of the D-ring in B800 BChl a. ## Results and discussion Figure 2 compares the electronic absorption spectrum of molischianum-LH2 after incubation with DDQ (denoted as oxidized molischianum-LH2) with that of native LH2. A Q y absorption band at 799 nm of B800 BChl a was absent from the oxidized molischianum-LH2, and a new absorption band appeared at 700 nm. We attribute this feature to the Q y band of the oxidized pigment, produced from B800 BChl a. Difference spectrum of the oxidized/ native molischianum-LH2 revealed a Soret band for the oxidized pigment at 449 nm (Fig. S2). In the oxidation of B800 BChl a, the peak position and the bandwidth of the Q y band of B850 BChl a remained unchanged. The absorption spectrum of oxidized molischianum-LH2 barely changed by incubation with sodium ascorbate (100 mM), indicating the oxidation of B800 BChl a is irreversible (Fig. S3). Spectral changes of oxidized molischianum-LH2 at 40 °C are compared with those of native LH2 to check the stability of oxidized LH2. Both the LH2 samples exhibited no change in their spectra (Fig. S4). These results suggest that oxidized molischianum-LH2 is stable in a similar level to native LH2. The stability of oxidized molischianum-LH2 is also confirmed by no degradation of oxidized molischianum-LH2 during its purification and various measurements. The Q y peak position of the new pigment in oxidized molischianum-LH2 shifted to a longer wavelength than that of D-ring reduced 3-acetyl-Chl (AcChl) a (Fig. 1C) in the B800 site of molischianum-LH2 at 690 nm 19 . This result suggests that the newly formed pigment in the oxidized molischianum-LH2 was not AcChl a. This assumption was confirmed by high-performance liquid chromatography (HPLC) analysis and electronic absorption spectroscopy of the pigments from oxidized molischianum-LH2. Oxidized molischianum-LH2 had two major chlorophyllous pigments (Fig. 3A,B). The former pigment (#1 in Fig. 3A) was BChl a, which we attribute to residual B850 BChl a (Fig. 3C). We assigned the latter pigment (#2 in Fig. 3B) to the oxidized pigment. This oxidized pigment eluted more slowly in the HPLC analysis than AcChl a (Fig. 3D). In addition, the Q y peak position of the oxidized pigment at 692 nm in methanol was red-shifted compared with that of AcChl a at 685 nm (Fig. S5). The oxidized pigment had a signal at m/z 909.5378 in high-resolution mass-spectrometry (HRMS) measurements. This value corresponds to that of the calculated value (MH + , 909.5375) of a didehydrogenated pigment derived from BChl a. These results suggest that the pigment formed through the DDQ oxidation of molischianum-LH2 is an isomer of AcChl a. To assign the oxidized pigment formed in molischianum-LH2, the pigment was demetallated and compared with a structurally determined 17,18-didehydro-bacteriopheophytin (BPhe) a that was synthesized from BPhe a by oxidation with FeCl 3 11,12 (see "Materials and methods" as well as Fig. S6). Note that the removal of Mg from BChl a is necessary for synthesis of the B-ring reduced chlorin possessing the 13 2 -methoxycarbonyl group. HPLC analysis revealed that the retention time of the free-base derived from the oxidized pigment in molischianum-LH2 was identical to that of authentic 17,18-didehydro-BPhe a (Fig. 4A,B), but differed from that of the D-ring reduced 3-acetyl-pheophytin a (Fig. 4C). The electronic absorption spectrum of the oxidized pigment in methanol was identical to that of 17,18-didehydro-BPhe a (Fig. 4D). These results indicate that the newly formed pigment in molischianum-LH2 by the DDQ oxidation is 17,18-didehydro-BChl a (hereafter denoted as 1). No porphyrin-type pigment was detected in the pigments that were extracted from oxidized molischianum-LH2, indicating that the treatment of molischianum-LH2 with DDQ did not oxidize the B-ring of B800 BChl a. No conversion from B800 BChl a to the corresponding porphyrin-type pigment, 3-acetyl-protochlorophyll a, has been reported in the DDQ oxidation of acidophilus-LH2 20 . The regioselective oxidation of B800 BChl a observed here suggests that the polypeptides sterically protect the pyrrole ring of BChl a inside the protein matrix. We confirmed the effects of the DDQ oxidation on lycopene, a major carotenoid in molischianum-LH2, by HPLC analysis. Lycopene from oxidized molischianum-LH2 was detected at the same retention time as that of native LH2 and a standard sample (Fig. S7), indicating that the lycopene in molischianum-LH2 was unaffected by the DDQ oxidation. The stable peak positions of lycopene in the absorption spectrum of oxidized molischianum-LH2 (Figs. 2 and S2) confirmed that DDQ oxidation had no effect on lycopene. The effects of the B800 oxidation on the protein structure of molischianum-LH2 were examined by a frequency modulation atomic force microscopy (FM-AFM), size-exclusion chromatography (SEC), and circular dichroism (CD) spectroscopy. The crystallographic structure of LH3 protein 21 indicates that the circular arrangement is conserved even if local interactions of BChl a with polypeptides; namely substitution of amino acid residues that participate in the formation of the BChl-binding pocket is proved to produce no effect on the overall structures of light-harvesting proteins from purple bacteria. In contrast, information on the effect of chemical modification of BChl a bound to LH2 and related proteins on their protein structures is less available. The structural analysis demonstrated here will provide useful information in this regard. The ring structure of oxidized molischianum-LH2 was clearly visualized by FM-AFM (Fig. 5). This circular arrangement is quite similar to that of native molischianum-LH2 22 . The height profiles in the AFM results (Fig. S8) indicated the averaged top-to-top distance of oxidized molischianum-LH2 to be 4.7 ± 0.2 nm (average and standard deviation of 16 samples). This value is almost identical to that of native molischianum-LH2 (4.5 ± 0.5 nm) 22 . www.nature.com/scientificreports/ Therefore, no deformation of the LH2 ring structure was induced by the in situ oxidation of B800 BChl a in molischianum-LH2. The elution volume of oxidized molischianum-LH2 in the SEC chromatogram was the same as that of native molischianum-LH2 (Fig. S9), indicating that the protein size was unchanged by the oxidation of B800 BChl a. In the CD spectrum of oxidized molischianum-LH2, a negative CD signal of B850 BChl a was observed at around 855 nm (Fig. S10C). This signal was similar to that of native molischianum-LH2 (Fig. S10A) 19,22 , indicating that the orientation and electronic structures of B850 BChl a were not influenced by the B800 oxidation. On oxidation of B800 BChl a, a reverse S-shaped signal for B800 BChl a at around 800 nm disappeared and a new negative CD signal appeared at around 700 nm (Fig. S10C). This negative signal is assigned to 1 in B800 sites. Note that AcChl a reconstituted into the B800 site of molischianum-LH2 barely exhibited CD signal in the Q y region 19 . This difference suggests some disarrangements of reconstituted AcChl a in the B800 site of molischianum-LH2 19 , although the interactions of AcChl a with the surrounding amino acid residues are essentially the same as those of 1 and native B800 BChl a. The negative CD signal in the spectrum of oxidized molischianum-LH2 at around 220 nm (Fig. S10D) closely resembles that of native LH2 (Fig. S10B), indicating that the content of α-helices was unchanged by the B800 oxidation. Excitation of oxidized molischianum-LH2 at 700 nm produced an emission from B850 BChl a at 860 nm (Fig. S11B, black curve). This emission is in line with the B850 emission by excitation of B800 BChl a in native LH2 (Fig. S11A, black curve). In the excitation spectra of oxidized and native molischianum-LH2, the bands at around 700 and 800 nm were detected, respectively (Fig. S11, red curves). These results indicate that 1 and B800 BChl a function as energy donors in oxidized and native molischianum-LH2, respectively. We used femtosecond transient absorption (TA) spectroscopy to examine the EET dynamics from 1 to B850 BChl a in the oxidized molischianum-LH2 (Fig. 6). Excitation of oxidized molischianum-LH2 at 700 nm produced a negative band with a minimum at 700 nm, which was assigned to superposition of the ground state bleach (GSB) and stimulated emission (SE) of the energy-donating 1 in the B800 site. This negative band subsequently decayed, accompanied by new positive and negative bands at 825 and 855 nm, respectively, indicating EET from 1 to B850 BChl a 19,23,24 . We obtained two decay-associated spectra (DAS) (Fig. 7) by global analysis of the time dependence of the differential absorbance (ΔAbs) at various wavelengths (Fig. S12). This global analysis is in line with our previous analysis of native molischianum-LH2 and AcChl a-reconstituted molischianum-LH2 19 . The DAS component A with a shorter lifetime (6.8 ps) has a negative band at 700 nm, which corresponds to the decrease of the mixed GSB/SE band of 1, with a pair of 825-nm negative and 855-nm positive bands due to the increase of the positive excited-state absorption band and the negative GSB/SE band of B850 BChl a, respectively. The DAS component B with a longer lifetime (1.2 ns) represents the decay of the excited state of B850 BChl a. The TA spectroscopy results reveal that 1 can function as an energy donor in oxidized molischianum-LH2. The intracomplex EET dynamics from 1 to B850 BChl a is homogeneous, suggesting that the molecular orientation of 1 in the eight B800 pockets is not affected by the DDQ oxidation. The lifetime of the DAS component A, which originates from the combination of the decay of 1 with the rise of B850 BChl a, indicates that intracomplex EET in oxidized molischianum-LH2 occurs approximately sevenfold slowly relative to that in native LH2 (990 fs) 19 . This slow intracomplex EET is attributed to low spectral overlap between energy-donating 1 and energy-accepting B850 BChl a. The intracomplex EET in oxidized molischianum-LH2 was slightly slower than that in AcChl a-reconstituted LH2 (5.0 ps) 19 despite a slight red-shift of the Q y band of 1 in the B800 site (700 nm) compared with that of AcChl a (690 nm). One possible reason for this result is the difference in the electron delocalization between B-ring reduced and D-ring reduced chlorins. Density functional theory (DFT) calculations indicated that electron density was delocalized over the β-position of the D-ring in the LUMO of derivatives of 1, although no delocalization was found at the β-positions of both the B-and D-rings in the LUMO of AcChl a derivatives 11 . Such a difference in the distribution of the electron densities might lead to slightly slow intracomplex EET in oxidized molischianum-LH2. To summarize, a regioselective isomer of natural Chls, namely B-ring reduced chlorin, is produced by selective in situ oxidation of B800 BChl a in molischianum-LH2. The B-ring reduced chlorin functions as an antenna pigment in the LH2 protein. Regioselective dehydrogenation of the D-ring of B800 BChl a occurs because of its 20 . This different regioselectivity can be rationalized by the difference in the B800 orientations between the two types of LH2 proteins (Fig. S1), and indicates an importance of the polypeptides, which protect the pyrrole rings in BChl a in the protein matrix. Additionally, the regioselectivity we observe in the protein matrix contrasts with that in preferential dehydrogenation of the B-ring of BChl a derivatives in organic solvents 11,25 . This is the first example of a photoactive B-ring reduced chlorin pigment in photosynthetic proteins. We hope that these findings will be helpful for understanding the roles of photosynthetic cyclic tetrapyrrole pigments on the light-harvesting proteins and wrestling with an open question why nature selects D-reduced Chl pigments. ## Materials and methods Apparatus. Electronic absorption and CD spectra were measured with a spectrophotometer (UV-2600, Shimadzu) and a spectropolarimeter (J-820, JASCO), respectively. Fluorescence emission and excitation spectra were measured with a spectrophotometer (F-7100, Hitachi). HPLC was performed with a pump (LC-20AT, Shimadzu) and detectors (SPD-M20A and SPD-20AV, Shimadzu). SEC was performed with an ÄKTAprime plus system (GE Healthcare). 1 H NMR spectra were measured with an NMR spectrometer (ECA-600, JEOL); chemical shifts were expressed (in ppm) relative to CHCl 3 (7.26) as an internal reference. HRMS measurements were conducted with a spectrometer (micrOTOF II, Bruker) by atmospheric pressure chemical ionization (APCI). Materials. LH2 protein was isolated from the cultured cells of a purple photosynthetic bacterium Phaeospirillum molischianum DSM120 26 . BChl a was isolated from a purple photosynthetic bacterium Rhodobacter sphaeroides, and was converted to AcChl a and BPhe a by DDQ oxidation 27 and demetallation under acidic conditions 28 , respectively. Lycopene and DDQ were purchased from Wako Pure Chemical Industries. A detergent n-dodecyl-β-D-maltoside (DDM) was purchased from Dojindo Laboratories. DDQ treatment of LH2. Molischianum-LH2 was treated with DDQ basically according to a previous report 20 . A solution of native molischianum-LH2 in 20 mM Tris buffer containing 0.1% DDM (pH 8.0) was mixed with 1/10 volume of an acetone solution of DDQ. The final concentration of DDQ was 0.2 mM. The mixed solution was incubated at 35 °C in the dark. After disappearance of the Q y band of B800 BChl a, DDQ was removed by ultrafiltration using Amicon centricon concentrators (50 kDa cutoff, Merck Millipore). The oxidized LH2 was purified by anion-exchange column chromatography using Whatman DE52 resin (GE Healthcare). The protein was desalted by ultrafiltration using Amicon centricon concentrators (50 kDa cutoff, Merck Millipore). ## HPLC analysis of chlorophyllous pigments in LH2. Chlorophyllous pigments in molischianum-LH2 were analyzed by reverse-phase HPLC reported elsewhere 20 . Molischianum-LH2 proteins were concentrated by ultracentrifugation with Amicon centricon concentrators (50 kDa cutoff, Merck Millipore), followed by evaporation with a smart evaporator (BioChromato) in the dark. Then, chlorophyllous pigments were extracted from the samples with methanol, and were eluted on a reverse-phase column 5C 18 -AR-II (6 mm i.d. × 250 mm, Nacalai Tesque) with a guard column 5C 18 -AR-II (4.6 mm i.d. × 10 mm, Nacalai Tesque) with methanol at the flow rate of 1.0 mL/min. ## Isolation of oxidized pigment 1. Oxidized molischianum-LH2 was concentrated by ultracentrifugation using Amicon centricon concentrators (50 kDa cutoff, Merck Millipore), followed by evaporation with a smart evaporator (BioChromato) in the dark. The pigment 1 was extracted from the resulting sample with methanol Synthesis of 17,18-didehydro-BPhe a. 17,18-Didehydro-BPhe a was synthesized from BPhe a by oxidation with FeCl 3 according to previous reports 11,12 . A nitromethane solution of FeCl 3 ⋅6H 2 O (4.0 eq.) was added to a dichloromethane solution of BPhe a, and the mixed solution was stirred at room temperature in the dark for 5 min. The reaction was quenched by addition of methanol and washed with distilled water. The organic layer was dried over anhydrous Na 2 SO 4 , followed by evaporation under reduced pressure. HPLC analysis of oxidized pigment 1 after demetallation. The central magnesium was removed from 1, which was purified described above, by addition of an aliquot of hydrochloric acid in methanol. The demetallated pigment was quickly extracted with diethyl ether and dried over anhydrous Na 2 SO 4 , followed by dryness under stream of nitrogen gas. The residue was dissolved with methanol and eluted on a reverse-phase column 5C 18 -AR-II (6 mm i.d. × 250 mm, Nacalai Tesque) with methanol at the flow rate of 1.0 mL/min. The elution profile of this pigment was compared with that of the synthesized 17,18-didehydro-BPhe a under the same HPLC conditions. HPLC analysis of lycopene. LH2 proteins were concentrated by ultracentrifugation with Amicon centricon concentrators (50 kDa cutoff, Merck Millipore), followed by evaporation with a smart evaporator (BioChromato) in the dark. Lycopene was extracted from the samples with a mixture of methanol and dichloromethane (1/1, vol/vol), followed by evaporation under reduced pressure. The residues were dissolved with an HPLC eluent (hexane/ acetone = 99/1, vol/vol) and eluted on a normal-phase column 5SL-II (6 mm i.d. × 250 mm, Nacalai Tesque) with hexane/acetone (99/1, vol/vol) at the flow rate of 0.5 mL/min. ## AFM measurements. A 100 μL aliquot of oxidized molischianum-LH2 solution in 20 mM Tris buffer containing 0.02% DDM and 150 mM NaCl (pH 8.0) was placed onto a cleaved mica (SPI Supplies). After standing for 30 min at room temperature in the dark, the mica surface was rigorously rinsed with 20 mM Tris buffer containing 150 mM NaCl (pH 8.0). The oxidized molischianum-LH2 was then observed in 20 mM Tris buffer containing 150 mM NaCl (pH 8.0) by a laboratory-built frequency modulation AFM controlled by a commercially available AFM controller (ARC2, Asylum Research) 29 with a silicon cantilever with gold coating of deflection side (160 AC-NG, MikroMasch), which had a nominal spring constant of 26 N/m. The typical resonance frequency and Q factor in an aqueous buffer solution are 120 kHz and 7, respectively. Size exclusion chromatography. SEC analysis was performed basically according to previous reports 21,30,31 . Molischianum-LH2 proteins were eluted on a HiPrep 16/60 Sephacryl S-300 HR column (GE Healthcare) with 20 mM Tris buffer containing 0.02% DDM and 150 mM NaCl (pH 8.0) at the flow rate of 0.4 mL/min. ## Transient absorption spectroscopy. Femtosecond time-resolved TA spectroscopy was done according to a previous report 19 with a pair of noncollinear optical parametric amplifiers (NOPA) (TOPAS-white, Light-Conversion), pumped by a regeneratively amplified Ti:sapphire laser (Solstice, Spectra-Physics), as light sources. Output of one of the NOPAs was set at 700 nm for excitation of the Q y band of the oxidized pigment in oxidized molischianum-LH2. A prism pair was used to pre-compress the pulses and the pulse duration at the sample position was 16 fs (fwhm), which was measured by the self-diffraction frequency-resolved optical gating (SD-FROG) method. The excitation intensity at the sample position was 20 μW (20 nJ), and the diameter of the focused laser beam was ca. 0.15 mm. The polarization between the pump and probe pulses was set at the magic angle by rotating the polarization of the pump pulse by a Berek compensator (Model 5540, New Focus). White-light supercontinuum (410-930 nm) was generated by focusing the output of another NOPA centered at 1100 nm into a rotating CaF 2 window (thickness: 2 mm) and it was divided into probe and reference pulses. The probe pulse was focused into a rotating sample cell excited by the pump pulse, and the transmitted light was guided into a multichrometer (MSP1000-V, Unisoku). The reference pulse was directly guided into another multichrometer of the same type and the differential absorbance (ΔAbs) of the sample was calculated. The heterodyne-detected optical Kerr effect (HD-OKE) signal between the pump and the probe pulses was obtained by replacing the sample solution in the rotating cell with neat carbon tetrachloride, and the electronic response signal was used to compensate the group velocity dispersion of the TA signal. Oxidized molischianum-LH2 was solubilized in

chemsum

{"title": "In situ formation of photoactive B-ring reduced chlorophyll isomer in photosynthetic protein LH2", "journal": "Scientific Reports - Nature"}

modulation_of_protein_fouling_and_interfacial_properties_at_carbon_surfaces_via_immobilization_of_gl

## Abstract: Carbon materials and nanomaterials are of great interest for biological applications such as implantable devices and nanoparticle vectors, however, to realize their potential it is critical to control formation and composition of the protein corona in biological media. In this work, protein adsorption studies were carried out at carbon surfaces functionalized with aryldiazonium layers bearing mono-and disaccharide glycosides. Surface IR reflectance absorption spectroscopy and quartz crystal microbalance were used to study adsorption of albumin, lysozyme and fibrinogen. Protein adsorption was found to decrease by 30-90% with respect to bare carbon surfaces; notably, enhanced rejection was observed in the case of the tested di-saccharide vs. simple mono-saccharides for near-physiological protein concentration values. ζ-potential measurements revealed that aryldiazonium chemistry results in the immobilization of phenylglycosides without a change in surface charge density, which is known to be important for protein adsorption. Multisolvent contact angle measurements were used to calculate surface free energy and acid-base polar components of bare and modified surfaces based on the van Oss-Chaudhury-Good model: results indicate that protein resistance in these phenylglycoside layers correlates positively with wetting behavior and Lewis basicity.Much effort towards the design and fabrication of biomaterials and medical devices is dedicated to the attainment of desirable surface chemistry and surface physical properties, as these can often determine the biological response to materials in vivo 1 . There is therefore a strong interest in investigating surface modification strategies that enable a degree of control over interfacial biointeractions. Protein-surface interactions are thought to be of particular importance due to the abundance of these molecules in tissues and biological fluids and due to the central role of peptides and proteins in cell adhesion and signalling. Depending on the specific biomaterial and its application (e.g. biosensor, implant) it might be desirable to either promote protein adsorption or repel protein build-up in order to modulate performance 2-5 . Therefore, much effort has been devoted to developing surface modification strategies to modulate protein-surface interactions.Various forms of carbon find multiple applications as biomaterials; coatings such as pyrocarbon and amorphous carbons (e.g. a-C, a-C:Si, a-C:H, ta-C) 6,7 , are promising for biomedical applications because of their frictional and mechanical properties, their corrosion resistance and chemical inertness, and their bio-and hemocompatibility. Carbon nanomaterials, such as nanotubes and nanodiamonds, have also received much attention as delivery agents for in vivo imaging and sensing 8,9 . Finally, materials such as diamond electrodes, carbon coatings and carbon nanofibers are routinely used for in vivo and in vitro bioanalytical chemistry 10,11 . For all of these applications it is critical to achieve control over interfacial interactions of the carbon solid surface with proteins in solution, to avoid unspecific adsorption that might result in undesirable cell-surface events, or in blocking of sensing/binding sites [12][13][14][15] . Several surface modification methods have been investigated in order to control and minimize protein fouling at surfaces: cationic polymers, enzymes or peptides are effective but costly and often present problems of leaching and durability 16 . Poly and oligo (ethylene glycol) (PEG, OEG) coatings have been shown to successfully minimize protein adsorption 12,17 ; however, PEG/OEGs can easily oxidize, losing their antifouling properties 16 . This problem has prompted a search for alternative antifouling coatings with enhanced chemical stability. In an effort to mimic biological antifouling strategies, work has focused on the use of immobilized carbohydrates, given the presence of these molecules in the antiadhesive glycocalyx that surrounds certain cells 18,19 . Research shows, in fact, that oligo-and polysaccharide coatings can control fouling and protein adsorption, while being extremely stable to oxidation . The use of aryldiazonium salt chemistry for the immobilization of simple carbohydrates on carbon surfaces was recently reported by our group 27 . Aryldiazonium chemistry offers a versatile route for surface immobilization with key advantages for carbon applications: (a) functionalization can be carried out from solution, (b) it occurs under mild conditions without the use of multistep reactions, and (c) it leads to the formation of robust functional layers via formation of strong C-C covalent bonds between R-Ph groups and carbon substrates 28 . This is a desirable property that imparts chemical and thermal stability to carbohydrate adlayers under a variety of conditions thus preventing interfacial exchange between the layer and biomolecules in solution. The ability to solution process surfaces also makes it intrinsically scalable and thus relevant for widespread applications. We have recently shown that immobilized phenylglycosides bearing mono-saccharide groups obtained via aryldiazonium chemistry can reduce the unspecific adsorption of Bovine Serum Albumin (BSA) at carbon surfaces 27 . However, it remains unclear whether antifouling properties can be observed with other proteins and whether specific carbohydrate structural properties are responsible for the antifouling behavior. Interestingly, we have also identified that phenyl-lactosides are more effective than mono-saccharide glycosides at preventing adsorption on polymer surfaces 20 . Herein, we report a study of protein adsorption at phenylglycoside-modified and bare amorphous carbon surfaces using five different glycosides, four bearing mono-saccharide moieties and one being a phenyl-lactoside. We use three proteins with different levels of structural complexity and isoelectric points to understand the generality of protein adsorption trends. Importantly, we investigate the relationship between protein adsorption at phenylglycoside layers and surface free energy, charge and glycoside structure with the aim of improving our current understanding of key properties that result in antifouling activity of aryldiazonium carbohydrate layers. ## Results Protein adsorption studies. Amorphous carbon (a-C) films used in our experiments were deposited via magnetron sputtering. These films had previously been characterized via a combination of spectroscopic methods 29 . Briefly, they consist of approximately 80% trigonally bonded carbon (sp 2 centers), as estimated via X-ray photoelectron spectroscopy (XPS) and Raman spectroscopy. The films also contain oxidized groups resulting in a 9% O/C atomic ratio as determined via XPS. Modification of a-C with aryldiazonium salts was carried out as in our previous work (Fig. 1), via diazoniation of 4-aminophenyl glycoside precursors in situ. Precursor glycosides bearing glucose (Glc), galactose (Gal), mannose (Man), rhamnose (Rha) and lactose (Lac) groups (compounds 1-5, Fig. 2), yielded surfaces from here onwards referred to as Glc-C, Gal-C, Man-C, Rha-C and Lac-C, respectively. Figure 3 shows examples of IR reflectance absorption spectroscopy (IRRAS) of Gal-C, a monosaccharide-modified surface, and of Lac-C, a disaccharide-modified surface, obtained from precursors 2 and 5, respectively. Both IRRAS spectra show the characteristic infrared absorbances of glycosides in the region 1290-950 cm −1 due to C-O stretching modes arising from the carbohydrate ring 27,30 . Peaks in the region 1550-1500 cm −1 arise from C-C skeletal vibrations of phenyl rings 30 ; in particular, it was possible to observe in all spectra the presence of a peak at 1508 cm −1 which can be attributed to the strong 19a stretching mode of phenyl rings 27 . Similar IRRAS spectra were obtained for Glc-C, Man-C and Rha-C surfaces. The thickness of phenylglycoside layers was characterized via atomic force microscopy (AFM) using previously reported methods 31,32 . Briefly, phenylglycoside-modified surfaces were first imaged in tapping mode; subsequently, a section of the film was removed by scratching the sample with the AFM tip in contact mode; finally, the step created in the organic film was imaged in tapping mode. Figure 4 shows an example of a Lac-C surface imaged after the scratching process and of a height profile across the step. Height profiles were used to obtain an average thickness which was found to be 0.8 ± 0.1 nm in the case of both Gal-C and Lac-C layers. These thickness values are slightly lower than estimates of molecular length of 1.0 nm and 1.5 nm for phenyl-β -galactoside and benzyl-β -lactoside conformers, respectively, obtained from optimized glycoside geometries 33,34 . Thickness results therefore indicate that layers prepared via aryldiazonium chemistry using both mono-and di-saccharide groups reach a surface coverage of at most 1 monolayer, as expected based on the presence of bulky terminal groups such as Lac and Gal glycans 35 . In order to evaluate the antifouling properties of glycosylated interfaces, both unmodified and modified a-C substrates were incubated in buffered protein solutions for 1 h and analyzed by IRRAS ex situ. Three proteins with different properties were chosen for our studies: BSA, lysozyme (Lyz) and fibrinogen (Fib); a summary of their main physical properties is reported in Table 1 . Figure 5 shows IRRAS spectra in the region 1900-1300 cm −1 of bare a-C, Gal-C and Lac-C surfaces after incubation in BSA, Lyz and Fib solutions at two different concentrations; dotted lines in the central and right hand panel show the IRRAS spectra of Gal-C and Lac-C surfaces prior to protein adsorption in the same spectral region. After adsorption, spectra exhibit the characteristic bands of amide groups in polypeptides: the two strong, broad peaks at ~1675 cm −1 and ~1540 cm −1 are assigned to the amide I and II modes, respectively 30 . Noticeably, the sharp peak at ~1510 cm −1 assigned to the aromatic ring appears in all of the spectra, thus confirming that the phenyl groups used for surface modification are strongly bound to the surface and are not displaced by adsorbed proteins. Similar results were obtained in the case of Man-C, Glc-C and Rha-C surfaces. The relative amounts of protein adsorbed at bare and saccharide-modified surfaces can be estimated from the net absorbance of amide bands in IRRAS spectra, under the assumption of no preferential orientation of peptide bonds at the carbon surface. Net absorbance values of amide I peaks at bare a-C, and phenylglycoside-modified carbon are reported in Fig. 6, where the inset shows the same results as percentage adsorption with respect to the bare surface. Values in Fig. 6 were obtained from adsorption experiments carried out at two different molar concentrations: 7 μM, equivalent to 0.5, 0.1 and 2.5 g L −1 for BSA, Lyz and Fib, respectively, and 0.30 mM, equivalent to 20 and 4.3 g L −1 for BSA and Lyz, respectively. These two concentrations are relevant for understanding the behavior of surfaces in physiological conditions since molar concentrations of 7 μM are in the normal range for Fib in plasma, while a 0.30 mM concentration is close to that of albumin in serum 39 . Fib could not be studied at the higher concentration because it falls beyond its solubility limit 40 . IRRAS results indicate that at bare a-C surfaces, adsorption increases with increasing molar concentration for the same protein. Fib solutions yielded the strongest adsorption among all protein solutions tested. These observed trends are in general agreement with previous reports of adsorption isotherms of human albumin and fibrinogen at isotropic carbon surfaces by Feng and Andrade 41 . Adsorption values on monosaccharide-modified surfaces were significantly lower than at bare a-C for all three proteins at all concentrations studied. Similar results were obtained for surfaces modified with Glc, Man and Rha units: only small differences were observed in protein resistance among the four monosaccharides used in our studies. The amount of protein adsorbed at Lac-C was however found to be significantly lower than at either bare a-C, or monosaccharide-modified surfaces, thus indicating that Lac-C surfaces are better at resisting protein adsorption. In order to obtain quantitative estimates of protein adsorption at mono-and disaccharide modified surfaces, Quartz Crystal Microbalance (QCM) measurements of protein mass were also carried out ex situ. Upon incubation in 7 μM BSA for 1 h, bare a-C surfaces reported a mass increase of 1.02 ± 0.27 μg cm −2 , whereas Gal-C and Lac-C surfaces yielded increases of only (0.35 ± 0.22) and (0.10 ± 0.11) μg cm −2 (C.I. 95%), respectively. The above estimates likely constitute upper boundaries for BSA adsorption at the three surfaces, given that ex situ QCM also measures contributions from the mass of water trapped within the BSA layer. Table 2 summarizes BSA mass densities and relative adsorption mass values measured via ex situ QCM, together with the corresponding adsorption estimates obtained from amide I peak absorptions in IRRAS spectra. The comparison between the spectroscopic and gravimetric determination of protein adsorption was found to be satisfactory, thus indicating that amide I peak intensities are proportional to surface mass density of proteins on these surfaces. Also, gravimetric analysis confirms that Lac-C layers perform better than Gal-C layers in terms of protein rejection. Surface contact angle and surface free energy studies. Surface free energy (SFE) and wettability play an important role in defining the extent to which a surface can resist biofouling. The SFE of unmodified and modified a-C substrates was determined via contact angle (CA) measurements of multiple solvents using the sessile drop method. In order to obtain the SFE, we used the model of van Oss, Chaudhury and Good (vOCG) 42,43 . This model assumes that the total surface tension results from additive contributions of apolar, or Lifshitz-van der Waals (γ LW ), and polar forces (γ AB ): LW AB where γ AB includes contributions γ − and γ + from electron donor-acceptor interactions, respectively, also called Lewis base-acid interactions. The model assumes that the work of adhesion at the solid-liquid interface, W SL , can be approximated by geometric means as below: where the subscripts "L" and "S" indicate components of the liquid and solid, respectively. vOCG is considered to be a suitable model for describing the asymmetric nature of polar interactions when hydrogen bonding contributions are present 42,43 : electron donating and accepting groups must interact "reciprocally" to contribute to surface tension, as reflected by mixed donating/accepting products in Equation (2). Equation ( 2), in combination with the Young-Dupre equation results in: which can be used to obtain γ S LW , γ − S and γ + S by measuring the CA of three liquids with known surface tension components γ L LW , γ − L and γ + L . Carbon films used for CA measurements were deposited on Si wafers and were found to display low rms roughness before and after modification (see Supporting Information). Surface tension components of the three test liquids at 20 °C are taken from van Oss's data compilation 43 and are reported in Table 3; the choice of liquids was based on the analysis of solvent triplets by Della Volpe et al 44 . γ S LW was first calculated using eq. ( 3) and the CA of diiodomethane, a liquid with γ . CAs of water and glycerol were then used to set a system of two linear equations that were solved for γ + S and γ − S 45 ; positive values were obtained from our calculations thus confirming that all surfaces yield physical solutions for γ + S and γ − S . CA values and surface tension components for all surfaces examined in this work are reported in Table 4. Bare a-C displayed a water CA of 35.3°, total SFE γ S = 63.7 mJ m −2 and components γ = . 49 7 S LW mJ m −2 and γ = . 14 0 S AB mJ m −2 . These values are in good agreement with those reported by Leezenberg et al. 46 for sputtered a-C films, but the polar component and total surface energy are higher than those obtained by Zebda et al. 45 via vOCG analysis. Differences in components and total SFE could arise due to variations in material properties (e.g. sp 2 /sp 3 or O-content) or film history 46 . Surface modification with saccharides leads to a significant decrease in water CA for all saccharide units tested, with the lowest CA observed for Lac-C surfaces. The total SFEs of phenylglycoside layers are slightly higher than that of bare a-C (<3% difference), with negligible differences observed among saccharides. Similarly, the apolar γ LW contribution does not change significantly with surface treatment, remaining approximately constant across all surfaces (<3% difference). The most striking differences among surface modifications were observed in the acid-base components. The vOCG model does not permit to directly compare the solid acid/base contributions of a solid surface 44 ; however, as discussed by Della Volpe et al. 44 , using the same solvent triplet it is possible to examine relative changes in acid and basic character brought upon by the surface modifications studied. Bare a-C displays the minimum γ − S value; modification with monosaccharides increases surface basicity by 30-40%, and a further and significant increase is observed when the disaccharide Lac is used. This result is surprising as carbohydrate units are typically classified as hydrogen bond donors and thus would not be expected to increase the Lewis basicity of a surface; possible explanations for these findings are included in the Discussion section. Surface charge density at bare and modified carbon surfaces. Electrostatic interactions can play an important role in protein adsorption phenomena given that proteins and most surfaces possess ionizable groups whose charge is dependent on pH. To investigate whether electrostatic interactions could contribute to observed changes in protein adsorption upon carbon modification, we carried out ζ -potential measurements using standard solutions of tracer particles. Table 4 summarizes ζ -potential results obtained for a-C, Gal-C and Lac-C surfaces in 1 mM NaCl solutions at pH 9.2. The ζ -potential of a-C was found to be − 55 ± 3 mV, whereas surface modification with phenylglycosides led to ζ -potential values for Gal-C and Lac-C of − 56.3 ± 1.9 mV and − 58.0 ± 2.6 mV, respectively. These results indicate that surface modification via aryldiazonium chemistry results in negligible changes in ζ -potential with respect to that of the bare a-C substrate. This indicates that that functionalisation with phenylglycosides via this methodology offers a route for increasing the wettability of carbon surfaces without the introduction of significant changes in electrostatic charge, as is often the case with other modifications (e.g. oxidation). The implications of these results for understanding the origin of protein antifouling properties of aryldiazonium carbohydrate layers and for the design of carbohydrate coatings with enhanced antifouling properties will be discussed in the following section. ## Discussion Protein adsorption studies on phenylglycoside layers obtained via aryldiazonium chemistry show that this functionalisation strategy leads to the formation of glycoside adlayers that impart resistance to protein adsorption. Spectroscopic and gravimetric studies carried out ex situ, all indicate that coated surfaces adsorb less protein than the unmodified carbon, with phenyl-lactoside groups appearing to be particularly effective at reducing unspecific adsorption. Solvation/hydration forces have been identified as important for determining protein adsorption trends, given that solvation and desolvation processes play a key role in protein adsorption 47 . Many studies 1,7,21, have in fact concluded that highly hydrophilic surfaces tend to prevent unspecific protein adsorption, whereas hydrophobic surfaces are more likely to favor protein adsorption because they are easier to dehydrate and because they can maximize their interactions with protein hydrophobic groups through changes in protein secondary structure upon adsorption 51 . In the case of aryldiazonium carbohydrate layers, CA measurements indicate that modification results in greater hydrophilicity; this correlates well with the reduction in protein adsorption that was observed in general, for all the three proteins at both concentration ranges examined. Lac-C surfaces were found to be the most effective carbohydrate-modified surfaces in terms of repelling protein fouling, and the ones with the lowest water CA in agreement with trends that positively correlate wettability with protein resistance. The contributions of polar and dispersive interactions resulting in the observed wettability were obtained from a multisolvent determination and analysis of Surface Free Energies (SFE). Carbohydrate surfaces obtained via aryldiazonium chemistry possess SFEs that are <3% higher than that of a-C. However the analysis based on the vOCG model suggests that large differences are introduced in the polar contributions to the total SFE, via modification of carbon with phenylglycosides. The solid-water interfacial SFE can be estimated from the data in Table 4, according to S L , which yields values of 4.3, − 0.2 and − 6.6 mJ m −2 for a-C, Gal-C and Lac-C surfaces, respectively. The observation of decreasing fouling in the order a-C > Gal-C > Lac-C is therefore consistent with expectations based on values of γ SL calculated from CA results. Analysis of SFE components also indicates that surface modification via aryldiazonium phenyl-glycosides increases the Lewis basicity of the carbon surface: Glc-C, Man-C, Gal-C and Rha-C have 30-40% greater γ − S values than that of bare a-C, while phenyl-lactoside immobilization leads to a 60% increase. This is somewhat surprising as carbohydrate units are typically classified as hydrogen bond donors and, thus, would not be expected to increase the Lewis basicity of a surface. Evidence from studies on alkylthiols indicates that the presence of groups that are polar, neutral and hydrogen-bond acceptors promotes fouling resistance 21,52 . Carbohydrates have been identified as exceptions to the hydrogen-acceptor requirement, however vOCG results suggest that this might not be the case and that once carbohydrates are immobilized they can actually enhance the hydrogen-acceptor character of surfaces. We speculate that saccharide-saccharide and saccharide-water intermolecular bonding within a dense glycan layer, might result in the basicity displayed by phenylglycoside layers. It is likely that engagement of hydroxyl groups in intra-layer hydrogen bonding modulates the hydrogen bonding properties displayed by the saccharide layer at the interface. Çarçabal et al. 33 carried out experimental and computational work on Man, Gal and Glc phenylglycosides and on benzyl-β -lactoside in the gas phase, showing that hydration leads to the formation of extended intra-and intermolecular hydrogen bond networks. The effect of hydration was greater in the case of benzyl-β -lactoside which was found to effectively lock into conformation through cooperative hydrogen bonding. It appears therefore likely that the water shroud associated with saccharide units would create a barrier to dehydration, and contributes to the protein resistance of carbohydrate aryldiazonium coatings. Further studies that directly probe hydrogen bonding within aryldiazonium layers would be desirable, to determine whether trends observed for phenylglycosides in the gas phase also translate to thin films of surface-immobilized groups. Finally, the surface-blocking effect and the steric hindrance of the saccharide moiety in phenylglycoside layers is likely to also contribute to preventing adsorption of proteins, given that coatings displaying bulky groups can screen protein-substrate interactions. Molecular density however might play a role beyond blocking access to the carbon surface, by also regulating the observed basicity of saccharide layers through intermolecular interactions within the adlayer. Thus it would be important in future studies to identify whether the observed basicity and protein resistance behavior vary significantly with molecular surface density, given the same carbohydrate motif. Conversely, carbohydrate structure might be leveraged to enhance or reduce hydrogen bonding by selecting units with different propensity to engage in inter/intra molecular hydrogen bonding. Studies of layers prepared with oligosaccharide moieties that display predominantly inter-or intra-chain bonding might reveal more about the role of inter and intra-chain interactions in determining basicity and protein fouling resistance in phenylglycoside layers. ## Conclusions We have investigated the adsorption of three proteins at carbon surfaces modified with phenylglycoside layers prepared via aryldiazonium chemistry; layers bearing both monosaccharides and a di-saccharide, lactose, were prepared and compared in their properties and protein resistance to bare carbon surfaces. Results indicate that these coatings display good protein resistance and that judicious choice of synthetic phenylglycosides can be used to optimize resistance. This is an important finding from a practical standpoint because aryldiazonium covalent immobilization is a versatile method for the functionalization of carbons and nanocarbons. Furthermore, it is known to work with a wide range of substrate materials beyond carbon and it is applicable under mild conditions from dip, spray and contact deposition methods. Thus, the methodology offers a versatile route to imparting antifouling properties onto surfaces of complex, mixed material devices, e.g. for biosensing, implantation, blood contacting applications. A study of interfacial physical properties revealed that the protein resistance of these layers correlates well with their hydrophilic character when compared to the bare carbon material. An increase in wettability with respect to bare carbon is achieved without a significant change in surface charge density. Interestingly, we notice that mono and di-saccharides increase the Lewis basicity of the surface, contrary to expectations from typical reactivity patterns of carbohydrates in solution. This finding is consistent with empirical rules on the type of properties that lead to protein fouling resistance of thin-organic layers. We propose that the observed basicity might arise from inter-and intra-molecular hydrogen bonding networks, which could alter the acid-base properties of units exposed at the surface. Further studies would be desirable for understanding the correlation between Lewis basicity and inter-and intra-molecular hydrogen bonding in the phenylglycoside layer. The vast number of existing carbohydrate structural motifs offers an exciting landscape for exploring the potential of these layers to leverage structural variability and achieve tunable fouling resistance. ## Experimental Methods Chemicals and Materials. Diiodomethane (99%), glycerol (≥99.5%), sulfuric acid (95-97%), hydrochloric acid (37%), hydrogen peroxide (30%), fluoroboric acid (48 wt.% in H 2 O), sodium nitrite (≥99.0%), acetonitrile (HPLC grade) and methanol (semiconductor grade) were purchased from Sigma and used without further purification. B-doped Si wafers were purchased from MicroChemicals and 10 MHz quartz crystals were purchased from International Crystal Manufacturing. Bovine Serum Albumin (BSA, ≥96%), Lysozyme from chicken egg white (Lyz), Fibrinogen from bovine plasma (Fib, 65-85% protein) and phosphate saline buffer tablets (PBS, 0.01 M, 0.0027 KCl and 0.137 NaCl pH 7.4) were purchased from Sigma. Millipore water was used for all experiments. Precursors 4-aminophenol-β -D-glucoopyranose (1), 4-aminophenol-β -D-galactopyranose (2), 4-aminophenol-α -D-mannopyranose (3), 4-aminophenol-α -L-rhamnopyranose (4) and 4-aminophenol-β -D-lactopyranose (5) (see Fig. 2) were synthesized as previously reported 20,27 . ## Substrate preparation. Amorphous carbon films (a-C) with thickness 73.6 ± 0.6 nm (C.I. 95%) were prepared via DC magnetron sputtering (Torr International, Inc.) at a base pressure ≤ 2 × 10 −6 mbar and a deposition Ar pressure of 7 × 10 −3 mbar, as previously described 29 . Silicon wafers were cleaned in piranha solution prior to deposition (H 2 SO 4 : H 2 O 2 in a 3:1 ratio -WARNING: Piranha solution is a strong oxidant and reacts violently with organic materials and presents an explosion danger; all work should be performed under a fume hood). For infrared reflectance absorbance spectroscopy (IRRAS) measurements, Si wafers were coated prior to a-C deposition, with an optically thick (449 ± 29) nm (C.I. 95%) Ti layer via DC magnetron sputtering. Surface modification with carbohydrate moieties was carried out as previously reported 27 , and following a protocol summarized in Fig. 1. Briefly, 4-aminophenyl glycosides were dissolved in acid; while keeping the solution in an ice bath, NaNO 2 was added yielding the corresponding aryldiazonium salt in situ at a final concentration of 1.0 mM. Carbon samples were immersed in the aryldiazonium salt solution for 1 h, rinsed in acetonitrile and methanol and dried under argon prior to further use. Characterization Methods. Static contact angles (CA) were measured on a CA analyzer (FTA) under ambient conditions of temperature and humidity; samples were rinsed in methanol immediately prior to CA characterization 45 and a minimum of three CA measurements were obtained for each surface. Spectroscopic Ellipsometry (SE) was carried out using an alpha-SETM ellipsometer (J.A. Woolam Co.). a-C films were deposited on clean Si wafers and measured at 65°, 70°, 75° incidence angle over the 370-900 nm range; SE data was then fitted using the CompleteEASE ® software package using a three layer model to account for Si, a-C and air phases (see Supporting Information). ζ -potential measurements were carried out using a Malvern Zetasizer Nano-ZS equipped with a surface ζ -potential cell; standard 300 nm latex tracer particle suspensions, NaCl 1 mM, at pH 9.2 (Malvern, DTS1235) were used in all experiments. IRRAS was carried out on a Fourier Transform Infrared (FTIR) spectrometer (Tensor 27, Bruker) equipped with a Mercury Cadmium Telluride (MCT) detector, a specular reflectance accessory (VeeMax II), and a ZnSe polarizer. Spectra were taken at 80° incidence using p-polarized light; 100 spectra were collected at 4 cm −1 resolution using a bare substrate as background. All spectra reported in this work were baseline corrected using commercial FTIR software (WinFIRST). Quartz Crystal Microbalance (QCM) measurements were carried out ex situ following a previously reported procedure 27 . The resonant frequency of a carbon coated QCM crystal was measured in air before and after protein adsorption, and the difference was used to calculate the mass change at the crystal via the Sauerbrey equation 53 . Measurements were carried out in a home-built chamber at the same temperature before and after modification; in the case of lactose-modified surfaces it was necessary to introduce a dessicant (Drierite ® ) in the measurement chamber in order to achieve frequency stability, likely due to water adsorption by surface-bound disaccharide units. Thickness and surface roughness measurements were carried out via Atomic Force Microscopy (AFM, Asylum Research) using silicon catilevers. Protein adsorption experiments. BSA, Lyz and Fib were dissolved in 0.01 M PBS buffer (pH 7.4) at different concentrations for each protein: 0.5 and 20 mg/mL for BSA, 0.1 and 4.3 mg/mL for Lyz and 2.5 mg/mL for Fib. Carbohydrate-coated and bare a-C surfaces were incubated in buffered protein solutions for 1 h at ambient temperature (20 °C). Substrates were rinsed, immersed for 10 min in water, and finally dried under argon prior to characterization.

chemsum

{"title": "Modulation of Protein Fouling and Interfacial Properties at Carbon Surfaces via Immobilization of Glycans Using Aryldiazonium Chemistry", "journal": "Scientific Reports - Nature"}

extraction_of_single_serve_coffee_capsules:_linking_properties_of_ground_coffee_to_extraction_dynami

## Abstract: The objective of this paper is to elucidate the variables that govern coffee extraction from single serve coffee capsules. The study was conducted on 43 Nespresso and Nespresso-compatible capsules of the same geometry, from all of which the coffee was extracted on the same machine. This allowed the link between a range of coffee and capsule (input) parameters with coffee brew (output) variables to be studied. It was demonstrated that the most efficient way to increase total dissolved solids in the brew is to use more coffee for extraction, and/or to grind the coffee more finely. However, grinding too finely can lead to excessive flow restriction. The most significant new insight from this study is the importance of the proportion of fines (particles smaller than 100 µm) regarding the capsule extraction dynamics. Capsules with a higher share of fines, for similar median particle size of the ground coffee, led to longer extraction times. General rules applicable for capsule coffee product development were established, although fine-tuning of parameters for successful capsule coffee extraction remains specific to production line and type of coffee. The single serve coffee sector for ground coffee is divided into high pressure extraction from capsules and low-pressure extraction from soft pods. It is the fastest growing segment within the larger category of coffee products. The global coffee pod and capsule market combined is expected to grow from USD 15 billion in 2017 to USD 29 billion by 2025 at a compound annular growth rate (CAGR) of 8.5%. The high-pressure extraction capsules segment alone, which is the focus in this study, currently leads the coffee pod & capsule market and was valued at around USD 9.61 Billion in 2017 (63% of the whole single serve coffee segment). The domestic consumption of single serve coffee is currently expected to be increasing due to the lockdowns imposed during the 2020 Coronavirus pandemic. This growing consumer interest and economic importance of single serve coffee products has led to growing scientific interest but remains a relatively little studied topic. Single serve coffee has been used to study the extraction process, the composition of brews and in the preservation of freshness of the ground coffee in hermetically sealed packaging 11,12 . The single serve coffee sector encompasses a broad definition of packaging options, including pods and/or capsules, and is used to prepare a wide range of coffees from filter to espresso. Different systems vary with shapes and sizes of capsules, with the most common espresso capsule system being the Nespresso system. Recently, a surge of Nespresso-compatible products entered the market. The compatibles have the same shape as Nespresso capsules, but may differ in the material they are made of as well as in the origin and quality of the coffee itself. The advantage of studying the extraction of Nespresso and Nespresso-compatible coffee capsule is that the human factor is essentially eliminated from the brewing process. Furthermore, the brewing unit, and coffee bed geometry are very similar for all capsules. When these parameters are handled manually in preparing coffee (e.g. semi-automatic espresso), this results in a higher variability in extraction and cup quality 13 . Therefore, the wide spectrum of Nespresso and Nespresso-compatible coffee capsule products available on the market provides an excellent platform to systematically study the impact of specific parameters in high-pressure extraction from capsules, while brewing unit parameters are essentially kept constant. Ground coffee analysis. The capsules were opened and thoroughly emptied to measure the weight of the coffee per capsule (mean of 3 capsules). The roast level (colorimetric) was determined using a Colorette 3b meter (Probat, Emmerich, Germany). Particle size analysis was performed using an imaging particle size and shape analyser with a dual camera system Camsizer X2 (Retsch Technology GmbH, Haan, Germany). At least 10 g of coffee was used for each particle size analysis, and three measurements were performed per capsule type. Median particle size (X50) was determined as the volume weighted median particle size, based on the particle area projected on the images. The share of fines (Q 100µm ) was defined as the volume share of particles smaller than 100 µm in size. Three particle shape parameters were determined as volume weighted averages, sphericity (SPHT), symmetry (Symm) and aspect ratio (b/l) values. ## Capsule coffee extraction. The coffees were extracted from the capsules using a Krups Inissia XN1005 capsule machine. The weight of the brews was measured on-line during extraction whereby the aim was to reach 40 g by stopping the extraction manually. The resulting brew weight was 40.1 ± 0.3 g (mean ± s.d), with all brews between 39.6 and 40.7 g. The extractions were timed manually from starting the pump to cessation of the pump operation. ## Brew analysis. The concentration of coffee in the filtered brew (0.45 µm filter) was analysed using a VST LAB Coffee III refractometer (VST inc., Tulsa, US), calibrated using a soluble coffee sample prepared to a known concentration as according to the instrument instruction manual to display total dissolved solids (TDS) in coffee brews. The extraction percentage was calculated based on the mean weight of the coffee in each capsule type, brew weight and refractometric TDS measurement. The sensory test of coffees was performed by three Q-graders (certified Q Arabica Grader, by the Coffee Quality Institute). The sensory quality attributes that were determined were flavour, acidity, body and balance on a scale from 5 to 10. The quality of each attributes was assessed by tasting three replicates of the brew by each of the Q-grader. The final values were based on a subjective consensus of the tasters on the quality score that they give to each samples, rounded to the nearest 0.25 units. The reference of the attribute range was the quality that could potentially be achieved by an espresso coffee extracted as a semi-automatic espresso, where a score of 6 is average quality, and 10 exceptionally high quality. ## Statistical analysis. All data analysis and statistical analysis was performed using R statistical computing (R Foundation for Statistical Computing, Vienna, Austria) using integrated packages. All measurements of brews and capsules (Table 1) were conducted in 3 replicates per each sample. ## Results and discussion Defining espresso coffee and its quality is subject to discussion. Some link it to the concentration of solids found in the espresso brew 15 -the TDS. Others to the amount of water used per gram of coffee 16 -the brew ratio. Regardless of the perspective taken on the definition of what an espresso coffee is, the resulting low volume beverage should have a pleasant, distinct aroma, a complementary balanced profile between acid, bitter and sweetness as well as a pleasant tactile thickness, known as the body 15,16 . As TDS will be heavily influenced by beverage volume, espresso brews prepared in this study were kept at a fixed beverage weight of 40 g. While brew weight was kept constant, the duration of extraction was allowed to vary and recorded to determine the average extraction flow rate. In the following, linear and multivariant correlations of extraction variables will be discussed first. Second, quality and sensory aspects of the extraction process will be addressed. Characterizing the coffee grounds. Grinding coffee generates particles with a bimodal distribution 17 . One local maximum occurs at greater particle sizes 18 and its location varies according to the adjustment (spacing) of the grinder burrs. For the sample in this study, this maximum is in the range 250-550 µm. The second local maximum occurs at smaller particle sizes, usually at around 50 µm. The particularity of this maximum is that its position seems to not be affected by the spacing of the grinding burrs. Median particle size rather than average was used in this study. The reason is that median particle size is less impacted by the amounts of fines, www.nature.com/scientificreports/ when such similar bimodal distributions are compared where the two peaks stay at the same position but only change in height. In Fig. 1, the share of fine particles, Q 100µm is plotted as a function of the median particle size for the capsules measured in this study. The capsule coffee grinds show a wide spread in the amount (proportion) of fines, over the range from 6 to 16% of volume. Generally, there is a trend; as the coffee is ground more coarsely (larger Table 1. Properties of the capsules analysed, ground coffees (X50-volume median mean particle size; Q 100µm -share of fine particles; Roast -roast level in Colorette Pt unit), brews (TDS -total dissolved solids; Extr. -extraction percentage), and the sensory scores of brews, sorted by ascending median particle size. Each line represents a capsule coffee product. The values with errors represent an average value and standard deviation (n = 3). median particle size), the shares of fines become smaller. In this study, a large variation in the share of fines for a given median particle size was found, which is most likely due to different capsule product lines. Coffees can have the same 10% share of fines while the median particle size can vary from 300 µm up to 550 µm. When only a single product line is considered (e.g. capsules from a single manufacturer; red dots, Fig. 1), a distinct grouping is observed. This result possibly reflects the grind technology used by this manufacturer, since the share of fines is clearly negatively associated with the median grind size, with a narrow spread of the data, compared to all the capsules tested in this study. For a specific median particle size, the capsules from this manufacturer systematically show the lowest proportion of fines. Furthermore, the proportion of fines increases in a linear fashion with decreasing median particle size. Influence of particle size distribution on extraction dynamics. Extraction times were used as an indirect reflection of the extraction dynamics. The extraction times were found to be highly variable between different types of capsules, covering a range between 10 and 41 s (Fig. 2), which corresponds to average flow www.nature.com/scientificreports/ rates in the 1-4 g/s range. When relating extraction times to particle size distribution, a highly significant positive association was found with Q 100µm (p < 0.001). Extraction time increased with increasing proportion of fines (Fig. 2b). More specifically, over the sample population an increase in the proportion of fines from 10 to 15% corresponded to an increase in the extraction time of 10 s. In contrast, the volume median particle size was not found to be correlated with the extraction time (Fig. 2a) as otherwise seems to be the case in the extraction with semi-automatic espresso machines 19,20 . Studies on the impact of the grind size for espresso extraction have mostly focused on the median or mean particle size 7,13,20 . The key variable that influences the hydrodynamics of espresso extraction is coffee bed permeability, which is a function of the particle size distribution of solids in the coffee bed and how the particles pack together to form the bed 19 . Hence, it is not only the median or mean particle size that plays a role in the permeability, but the overall particle size distribution also including the share of fine particles 19,21 . Due to the large range of capsules and manufacturers included in this study, results are spread over a wide variety of Q 100µm ; X50 combinations (Fig. 1). Even though both parameters are related (Fig. 1), and that probably a range of different grinding technologies are used by different manufacturers, this study indicates that coffee bed permeability is impacted more by the share of fines in the capsule coffees than by the particle size itself. When grinding coffee on a single grinder type, both Q 100µm and X50 are changing simultaneously (for example Fig. 1, red dots), therefore extraction times will be related closely to the median particle size, as has been previously reported 19,20 . However, as Fig. 2 illustrates, the extraction time seems to be much more significantly impacted by the shares of fines than the median particle size. In addition to searching for linear correlations, multiple regression analysis was applied to the data, in order to unveil multivariant correlations between capsule variables and extraction time (Fig. 3). Multiple regression analysis confirms what is observed with linear correlations. Extraction time as a function of coffee weight in the capsule, volume median particles size, roast level, share of fines and capsule pressure was modelled using multiple regression analysis. With such a model, 42% of the extraction time variance can be predicted, and the two parameters Q 100µm (p < 0.0001) and X50 (p < 0.01) have significant importance for the regression. The extraction time can be predicted based on the model (Fig. 3a), and the prediction works well for coffee extracting until 30 s extraction time, whereas for those that extracted longer than 35 s, the predicted values were smaller than the measured ones. The standardised model coefficient is double in value for Q 100µm than for X50 (Q 100µm 0.88, X50 0.45), and furthermore the expected coefficient for X50 would be negative, since smaller particle size could in theory only be associated with a longer extraction time. This result shows that there is some multicollinearity between share of fines and particle size, and confirms the higher importance of share of fines for controlling capsule extraction time, rather than particle size. We suspected that higher capsule pressure would lead to longer extraction time, due to a larger amount of trapped carbon dioxide in the coffee bed, a residue from the degassing process 22 . No correlation of the capsule pressure with the extraction dynamics was observed. Influence of capsule coffee properties on the cup properties. Espresso quality may be viewed as a complex extraction puzzle where the proper ratio of desirable components must be extracted simultaneously, in a matter of seconds, in order to produce the most desirable product. One of the characteristics of espresso coffee is its viscous, thick appearance, which is caused by high concentration of coffee solubles (TDS) 15 . High TDS is usually a highly desired quality of the espresso brew and is related to the sensory perception of "body". TDS is a www.nature.com/scientificreports/ non-specific analysis technique used routinely to gauge how much of the coffee was extracted into the cup, also called extraction yield. Extraction time, coffee weight and proportion of fines were found to positively correlate with the TDS of the resulting brew (Fig. 4b-d). Negative correlation of the particle size with the TDS was found (Fig. 4a), which is expected and was reported before 23 . The extraction percentage-shows essentially the same relations with the variables mentioned above. This is because extraction percentage is calculated by multiplying TDS by brew weight and dividing by the coffee weight, and the brew weight was kept constant for all extractions in this study. Unsurprisingly, the most direct relation and strongest correlation among those observed in this study is the weight of the coffee in the capsule with the TDS of the brew. Packing more coffee in the same extraction system will lead to a more concentrated brew when brewing to a fixed beverage weight. The linear relation observed follows the theoretical increase of TDS that should be observed when assuming the same extraction percentage from all capsule coffees (Fig. 4c). Indeed, over the whole range of capsules tested here, no relation of extraction percentage (in %) with the weight of coffee in the capsule was found (variable relation not shown). Longer extraction times and hence slower flow rates are related to brews with higher TDS. The effect is surprisingly high, even though as reported by Melrose et al. the bed extraction efficiency theoretically does not increase significantly with longer extraction times 21 . The data from this study indicate a possible "saturation" effect at extraction times above 25 s (Fig. 4d). The extraction time and coffee weight show no significant correlation. The inverse correlation of TDS (and extraction percentage) with particle size is known and is explained by increased extraction efficiency from smaller particle size that enable the solubles to diffuse from the particles more quickly 21 . The complex correlation of variables observed here could also be impacted by the product specifications. Since consumers expect an espresso to be a thick and a strong brew, finer grind sizes with a higher share of fines are used to maximize the extraction percentage. Hence, long extraction times might be acceptable, and the pressure needed to extract the capsule pushed to the limit of the capsule machine by producers to achieve desired brew quality. We see some weak correlation (p < 0.05, data not shown) of TDS with roast level. This result may be attributed to the fact that the recipes for the capsules vary greatly, and darker roasted coffees could be preferred by consumer, as this leads to coffee with a stronger body. Hence, the correlation of roast level with TDS could not be attributed directly to better extractability of darker roasted coffee, since the grind properties play a more important role in the extraction behaviour. Multiple regression analysis of TDS as a function of coffee weight, X50, Q 100µm , roast level and capsule pressure (52.7% of TDS variance predicted) revealed that decreasing the particle size of the grounds, X50 and increasing the coffee weight, had the most prominent impact on TDS (standardized model coefficients -0.43 and 0.32, p < 0.01 and p < 0.01). The prediction of TDS based on the multiple regression model is shown in Fig. 3b. The authors are aware that the indicators of importance of variables (coefficient and p-value) in models are not completely robust, since the independent variables have clearly some multicollinearity. While building the multiple regression models it was found that the results are consistent, even when subsets of samples were modelled. This was the case when independent variables that were not important were excluded from the model, or other variables were included (for example, including time as an independent variable for the TDS model). Hence, it seems reasonable to conclude that the results from the multiple regression analysis revealed with good statistical confidence the importance of the variables predicting the TDS in the brew and the extraction time. ## Sensory analysis. As 43 different commercial capsule coffees were tested (Table 1) in this study, it was anticipated that flavour attributes would vary between products due to the potential use of different origins, varietals, processing methods, roast levels, grinds and blends. Nevertheless, body is particular as it represents a tactile sensation attributed to solids extracted from the coffee during brewing and becoming suspended in the beverage, making them quantifiable by measuring the TDS 24 . Consequently, the sensory results have not shown correlations with technical variables apart from a suspected trend no significance of body correlated with TDS (p = 0.21). The Q-grading sensory evaluation was focused on the quality of the brew. The sensory results were mainly of values close to 6, since the coffee brews in this study did not reproduce the range of Q-grader results that could be expected for coffees prepared using a semi-automatic espresso machine and highest quality coffees. Some of the outliers in the sensory results could be explained by the parameters of the capsules. Two samples were found to have higher oxygen content in the capsule (samples 27 and 42) and scored lower in the sensory. Another two samples that were found to be higher than average in the sensory results (samples 19 and 40) could be explained to be of higher green coffee quality, since those samples were single origin Colombian and Ethiopia coffee. These sensory results and the lack correlation to technical parameters are not surprising since they show that most of the capsules have been optimised and produce average quality results. Only outliers in the sensory space were found as coffees of poorer production (oxygen in capsule) or coffees of above average green coffee quality. The orthogonality of sensory analysis results with capsule parameters is also shown by the principle component analysis (PCA), as displayed in Fig. 5. The differentiation of the capsule samples by PCA is caused by ground coffee properties in PC1 axis, together with the sensory descriptor body, which has the same direction as coffee weight along the PC1 axis. As expected, based on results discussed earlier, the loadings of particle size and roast level have the opposite value in PC1 direction to other loadings, since lower values of those have been www.nature.com/scientificreports/ found associated with higher values of other ground coffee variables and TDS. The remaining sensory attributes are oriented orthogonal to the technical variables, parallel to the PC2 axis, further confirming that the sensory quality properties of the brew within the capsule landscape of studied samples are not significantly impacted by the physical properties of the ground coffee and by variations in extraction variables. The samples mentioned in the previous paragraph (19, 27, 40 and 42) are also outliers on the PC2 axis and are likely the cause of the orthogonality of the sensory to technical parameters on the PCA plot. ## Conclusions Despite the wide variety of different capsule coffee formulations used for production of capsules subjected to this study, we can draw some general conclusions with respect to the extraction of coffee from capsules and espresso brew properties as a function of ground coffee parameters. The capsule extraction dynamics are highly influenced by the particle size distribution of coffee grounds. Only mean or median particle size information is insufficient to understand capsule extraction, therefore in this study we propose the parameter "share of fines; Q 100µm " as the second parameter for a simplified description of coffee grinds. The proportion of fines was highly correlated with increasing extraction time. The best way to increase TDS is to use more coffee for extraction, or to grind the coffee more finely to increase extraction efficiency. Grinding finer to increase extraction comes with a technical problem; finer grind sizes tend to form more finer particles in the grounds, which cause a decreased permeability of the coffee bed. This study shows that while there are certain general rules that can be followed for capsule coffee product development, the fine-tuning of parameters for a successful capsule coffee extraction is a problem specific to capsule production line (e.g. different grinder technology produces different share of fines in the coffee grinds). The coffee sensory quality in the cup was not correlated to the technical variables of the grounds. It is suspected that because most of the samples used in this study are optimised for the extraction, the cup quality can be only further improved by using higher quality green coffee. To better understand in what range the optimisation of technical parameters and green coffee quality affect brew quality, future studies with capsules produced using a controlled range of conditions need to be conducted.

chemsum

{"title": "Extraction of single serve coffee capsules: linking properties of ground coffee to extraction dynamics and cup quality", "journal": "Scientific Reports - Nature"}

ultrastable,_cationic_three-dimensional_lead_bromide_frameworks_that_intrinsically_emit_broadband_wh

## Abstract: Herein, we report the unusual broadband white-light emission as an intrinsic property from two cationic lead bromide frameworks. This is the first time that the metal halide materials adopting a purely inorganic positively-charged three-dimensional (3D) topology have been synthesized, thus affording highly distorted Pb II centers. The single-component white-light emitters achieve an external quantum efficiency of up to 5.6% and a correlated color temperature of 5727 K, producing typical white-light close to that of fluorescent light sources. Unlike the air/moisture-sensitive 3D organolead halide perovskites, our cationic materials are chemically "inert" over a wide range of pH as well as aqueous boiling condition. Importantly, these long-sought ultrastable lead halide materials exhibit undiminished photoluminescence upon continuous UV-irradiation for 30 days under atmospheric condition ($60% relative humidity, 1 bar). Our mechanistic studies indicate the broadband emission have contributions from the self-trapped excited states through electron-vibrational coupling in the highly deformable and anharmonic lattice, as demonstrated by variable-temperature photoluminescence/absorption spectra as well as X-ray crystallography studies. The chemical robustness and structural tunability of the 3D cationic bromoplumbates open new paths for the rational design of hybrid bulk emitters with high photostability. ## Introduction Solid-state lighting, an energy-saving alternative technology to conventional lighting sources, has attracted increasing attention in recent years. 1 Among them, realizing white-light luminescence from light-emitting diodes (LEDs) is of particular interest for general illumination applications. A typical whitelight LED device includes a blend of red, green and blue (RGB) LEDs or coating a blue (or near UV) LED with a yellow phosphor (or a mixture of RGB phosphors). These multi-color and/or multi-component strategies suffer from a variety of inevitable drawbacks, such as efficiency losses arising from selfabsorption and long-term instability due to different degradation rates of the phosphors. 8,9 Thus, a single-component broadband white-light emitter covering the entire visible spectrum is an attractive and challenging target in white-light LED research. However, very few known examples of single-source phosphors intrinsically emit white-light luminescence, thus hindering rational synthetic design. 3,5,13 Hybrid inorganic-organic lead halide perovskites containing vertex-sharing metal halide octahedra are an emerging class of photoactive materials, which have promising applications in photovoltaics and light-emitting devices. 14,15 Among them, twodimensional (2D) organolead halide perovskites usually exhibit narrow-band photoluminescence owing to their flat (100)oriented layers and large exciton binding energy. 16 Recently, a few instances of (110)-oriented 2D perovskites showed broadband photoemission as an intrinsic property, presumably ascribed to the formation of self-trapped excited states (e.g. Pb 2 3+ , Pb 3+ , X 2 ## À ) (X ¼ Cl, Br, I). Lowering the dimensionality of lead halide facilitates the self-trapping process and enhances the photoluminescence quantum efficiency (PLQE), albeit with a sacrifce in structural stability and photoluminescence tunability. The air/moisture-sensitive nature of organolead halide perovskites results in a gradual decrease of photoemission intensity upon UV-irradiation in air, thus hindering their industrial applications in LED technology. Therefore, the development of chemically ultrastable lead halide materials with efficient, tunable and broadband white-light emission is crucial to extend their photoactive applications. Inorganic extended frameworks usually adopt a neutral or anionic inorganic host, including zeolites, aluminophosphates and metal halide perovskites. Purely inorganic structures bearing an overall positive charge are very rare, presenting merely <0.1% of over 150 000 crystal structures in the inorganic crystal structure database (ICSD). Layered double hydroxides are a widely studied class of cationic 2D inorganic materials, possessing trivalent-ion-substituted brucite layers intercalated with charge-balancing anions. Other examples include francisite minerals (Cu 2 BiSe 2 O 8 X, X ¼ F, Cl, Br, I) and their derivatives, 36,37 as well as layered heavy p-block hydroxides and fluorides. Until now, only two synthetic examples of purely inorganic cationic three-dimensional (3D) frameworks have been reported, namely a thorium borate and an ytterbium oxyhydroxide, respectively. 43,44 Among the few examples of 3D lead halide inorganic frameworks, none of them bear a positively charge. 45,46 Very recently, we reported a class of 2D (layered) cationic lead halide materials with high chemical resistance and high photoluminescence quantum efficiency. 47 Herein, we report the synthesis, crystal structures and broadband photoluminescence of the frst two 3D cationic metal halide frameworks, [Pb ] 2+ , TJU ¼ Tongji University). The unique positively charged 3D lead bromide networks defne the arrays of unidimensional channels, in which reside the bridging organic anions. Intriguingly, both materials are intrinsically bulk white-light emitters spanning the entire visible-light spectrum. In contrast to lead halide perovskite and other hybrid bulk emitters, our cationic materials are essentially unaffected upon near-UV-irradiation (365 nm) for 30 days under ambient condition ($60% relative humidity, 1 bar). The temperature-dependent photophysical studies (e.g. UV-vis absorption spectra and photoluminescence spectra) and X-ray crystallography studies attribute the broad emission to the self-trapped states from electron-vibrational coupling in the strongly deformable and anharmonic lattice. ## Results and discussion Hydrothermal reaction of PbBr 2 , adipic acid disodium salt (NaO 2 C(CH 2 ) 4 CO 2 Na), and perchloric acid afforded colorless block-shaped crystals of TJU-6 (Fig. S1a †). Specifcally, slowcooling of the autoclaves at the rate of 10 C h 1 after the static heating is necessary to obtain high phase purity of TJU-6. In addition, perchloric acid was discovered to tune the pH as well as act as a stabilizer, like the role of fluoride ion in zeolite synthesis. 48 X-Ray crystallography reveals that TJU-6 is crystallized in the highly symmetric tetrahedral P4 1 2 1 2 space group (Table S1 †). TJU-6 consists of edge-and vertex-sharing PbBr 3 units extending in three dimensions (Fig. 1a and S2 †). Adipates covalently bridge and crosslink the inorganic extended connectivity, further enhancing the structural stability that will be discussed later. The crystallographically independent Pb II atom occupies a highly distorted octahedral geometry in TJU-6, having three bridging bromines and three oxygens from two adipate anions (Fig. S3 † inset). In metal halide materials (e.g. organolead halide perovskites), there is a strong correlation between the distortion of Pb II centers and the self-trapped excitons from short-range electron-lattice interactions. 14,18,19,23 Noting the Pb-O bond length is obviously shorter than Pb-Br, we sought to use the octahedral angle variance s oct 2 to quantitatively evaluate the deformation of O h symmetry in each PbBr 3 O 3 octehedra: 49 where a i are the X-Pb-X (X ¼ O/Br) angles. The PbBr 3 O 3 units in TJU-6 have a large structural distortion (s oct 2 ¼ 735.0), which probably result from the inert pair of s 2 electrons as well as the different electronegativity of oxygen and bromine atoms. The edge-sharing PbBr 3 O 3 octahedra in TJU-6 are connected in both of the crystallographically (100) and (010) planes, defning the honeycomb arrays of 6-membered ring channels along the a-and b-axis, respectively (Fig. 1c and S3 †). Adipates covalently bridge the lead centers within the hexagon-shaped channels, and the Pb-O bond lengths (2.505-2.664 ) are well within the accepted Pb-O covalent range. The two crystallographically independent Br atoms are either vertex-bridging (m 2 -Br) or quadruply bridging (m 4 -Br) in a highly distorted tetrahedral geometry. Overall, the high-coordinated bridging Br atoms and the low-coordinated Pb centers collectively attribute to the cationic feature of a rare 3D metal halide topology. Synthesis of the cationic 3D bromoplumbate framework was successfully extended to a more compact-packing inorganic topology via employing a shorter a,u-alkanecarboxylate as the anionic structure-directing agent. Colorless block-shaped crystals of TJU-7 (Fig. S1b †) were prepared using succinate in place of adipate under otherwise identical synthetic conditions. TJU-7 crystallizes in the orthorhombic crystal system with Pbcn space group (Table S2 † Non-perovskite metal-halide hybrid materials usually have a low-dimensional inorganic extended structure (1D or 2D). In addition, all of the previously reported 3D lead halide examples occupy an anionic inorganic network hosting organoammonium cations. 45,46,50,51 Overall, TJU-6 and TJU-7 are the frst two lead halide materials with a cationic 3D M-X-M (M ¼ metal, X ¼ halogen) connectivity that have been unambiguously identifed by single-crystal X-ray crystallography. The high yield and phase purity of TJU-6 and TJU-7 was evidenced by Fourier-transform infrared spectroscopy (FTIR), elemental analysis and powder X-ray diffraction (PXRD), which matches well with the theoretical patterns simulated from single-crystal data (Fig. 2, S6 and S7 †). Unlike organoammonium cations in perovskites, the anionic structuredirecting agents (e.g. a,u-alkanecarboxylate) covalently crosslink the metal centers within the pore channels of the cationic inorganic host. This structural feature plays a signifcant role to afford the chemical "inertness" of the resulting lead bromide materials. Stability tests were performed by treating the materials in water, ethanol, HCl solution (pH ¼ 2), and NaOH solution (pH ¼ 12) for 24 h. PXRD of the post-treated TJU-6 and TJU-7 remained intact, confrming the well-retained cationic topology (Fig. 2). In addition, no apparent loss in mass was observed during the chemical treatment, further proving the chemical inertness of TJU-6 and TJU-7. Moreover, thermogravimetric analysis (TGA) and ex situ thermodiffraction indicate that TJU-6 and TJU-7 are thermally stable up to 250 C under air (Fig. 2, S8 and S9 †). Overall, our cationic 3D bromoplumbate materials strikingly push forward the chemical-resistance and related photoactive applications of organolead halide hybrid materials. Given the high robustness and lattice deformation of TJU-6 and TJU-7, we sought to investigate their photoluminescent properties (Table 1). Both materials show a typical semiconductive properties with an optical bandgap of 3.70 eV (335 nm) for TJU-6 and 3.50 eV (354 nm) for TJU-7, respectively, evidenced by UV-vis absorption spectra (Fig. 3a and b). The large band-gaps probably result from the open inorganic framework of [PbBr] + and [Pb 3 Br 4 ] 3+ , which leads to an increase in the bandgap and the energy level of conduction band. 52 In addition, excitonic features (the shoulder peak at 3.99 eV for TJU-6 and 3.90 eV for TJU-7) are observed in the optical absorption spectra. The less-defned excitonic peak is largely ascribed to the a l abs is the wavelength at absorbance maximum; l ex is the excitation wavelength; l em is the wavelength at the emission maxima; 4; is the external photoluminescence quantum efficiency; s av is the PL lifetime. moderate exciton binding energy (290 meV) from the absorption spectra of TJU-6 at 103 K (Fig. S10 †). Interestingly, colorless block-shaped single crystals of TJU-6 show a pronounced white-light emission upon irradiation with a 4 W, 365 nm UV-lamp (Fig. 3c). The photoemission spectra (under 360 nm excitation) of TJU-6 demonstrates the unusual broadband photoluminescence spanning the entire visible spectrum (400-700 nm). The strongly Stokes-shifted broadband emission of TJU-6 has a maximum at 530 nm with a large FWHM up to 124 nm (0.56 eV), signifcantly overcoming the self-absorption problems in the near-UV region. In addition, mm-sized particles of TJU-6 were prepared via manual grinding, and showed nearly identical photoemission spectra (Fig. 3 and S11a †). These results confrm that the broadband emission arises from the bulk materials instead of from surface defect sites (which often contribute to the photoluminescence of lead halide perovskite nanocrystals 53,54 ). The high-energy shoulder at 420 nm is more evident in microscale crystals, likely due to the strong self-adsorption in cm-sized single crystals. 25 Since the two-band emission profles are similar to those of lead halide perovskites, 17,18 it is reasonable to attribute the high-energy shoulder to free excitons and the low-energy broadband emission to self-trapped excitons. The detailed photoemission mechanism will be discussed later. Time-resolved photoluminescence decay experiments indicate the lifetime of the emission measured at 537 nm is 1.7 ns for TJU-6 (Fig. 4a). The ns-ranged lifetimes are characteristic of fluorescence emission, while the longer lifetime of TJU-6 over the previously reported 2-D lead-based perovskites (s av ¼ 0.23-1.39 ns) is presumably attributed to the populated self-trapped excitons. Moreover, the cm-sized single crystals and mm-sized particles of TJU-6 show similar lifetimes, again suggesting the broadband emission as an intrinsic effect. The strongly Stokes-shifted broadband photoemission covering the entire visible-light spectrum has also been observed in TJU-7. A more evident high-energy photoemission shoulder at 415 nm blue-shifts the maximum emission wavelength to 480 nm with a large FWHM up to 166 nm (0.88 eV). Importantly, no apparent change to the color properties of the broadband emission is observed in TJU-6 when adjusting the excitation from 320 nm to 360 nm, confrming its nature as a broadband monochromatic emitter (Fig. S12 †). The Commission International de l'Eclairage (CIE) chromaticity coordinates of the overall TJU-6 emission is determined to be (0.33, 0.48), corresponding to a color temperature of 5727 K (Fig. 3d). The emission is so-called "cold" white-light, and very close to that of conventional fluorescent light sources ($5100 K). Meanwhile, the mm-sized crystals of TJU-6 exhibit a CIE coordinate of (0.30, 0.45) with a color temperature of 6464 K (Fig. 3d). The external quantum efficiency of the TJU-6 is measured to be 5.6%, which is signifcantly higher than most white-light emitters based on 2D organolead halide perovskites as well as a 3D anionic lead chloride framework (Table S3 †). In addition, the photoemission band and intensity was monitored upon irradiation with a 4 W, 365 UV lamp under atmospheric condition ($60% relative intensity, 1 bar, room temperature). Intriguingly, the broadband emission is remarkably stable after UV-irradiation in air for 30 days, achieving a substantial advance in hybrid lead halide white-light emitters (Fig. 4b). The TJU-7 light emission was determined to have CIE chromaticity coordinates of (0.25, 0.32) for bulk cm-sized crystals and (0.25, 0.33) for mm-sized crystal particles, which are all closer to the coordinates of the sunlight source (0.33, 0.33) (Fig. 3d). The bluish white-light emission affords the correlated color temperature of 11 967 K for cm-sized single crystals and 10 640 K for mm-sized particles, respectively. The external PLQE of the TJU-7 was measured to be 1.8% with an expected lower efficiency from the dense [Pb 3 Br 4 ] 2+ framework. In order to verify the broadband emission of TJU-6 arises from the self-trapped excited states attributable to the lattice deformation, we performed variable-temperature experiments of photoluminescence and UV-vis absorption spectra (Fig. 4d-f). In contrast to intrinsic white-light emitters of organolead halide perovskites, 18,55,56 an obvious and gradual blue-shift of the maximum emission wavelength was noticed when the temperature was decreased from 473 K to 77 K (Fig. 4d). The overall photoluminescent peak position was shifted to higher energy by 27 nm (117 meV), while the integrated intensity of the photoemission was increased by 40 times (Fig. 4e). In order to investigate the origin of the photoluminescence blue-shift, we performed variable-temperature UV-vis absorption spectra of TJU-6 from 133 K to 298 K. In agreement with the photoluminescence experiments, a blue-shift of the absorption edge was clearly observed with decreasing temperature (Fig. 4c). Based on these variable-temperature photophysical studies, it is concluded that TJU-6 has a temperature-dependent (133-293 K) optical bandgap. The temperature dependence of the semiconductor bandgap, determined from the UV-vis absorption edge, is usually well described using Varshni's equation: 57 where E g (T) is the energy gap at T K; b is physically associated with the Debye temperature, and g is the temperature coefficient of the band gap. The best ft is obtained with E g (0) ¼ 3.815 (5) eV, g ¼ 0.55(5) meV K 1 , and b ¼ 135(35) K (Fig. S13 †). The value of b ¼ 135(35) K for TJU-6 is within the reasonable range of Debye temperature for PbBr 2 between 100 and 150 K. The temperature coefficient of the band gap (g ¼ 0.55(5) meV K 1 ) is higher than some other typical semiconductors (e.g. InN), 61 suggesting the possible influence of thermal expansion and/or electron-phonon interaction on the band gap in our cationic bromoplumbate materials. A small thermal expansion of the unit cell volume from 1160.5(3) 3 (150 K) to 1166.5(3) 3 (298 K) was observed using single-crystal X-ray crystallography (Tables S1 and S4 †). In addition, the anisotropic thermal expansion was noticed with stronger elongation (0.25%) along the c-axis than that of the other two directions (0.13% for the a-and b-axis). Despite the contribution of the anharmonic lattice, the high temperature coefficient of the band gap (g ¼ 0.55(5) meV K 1 ) is largely attributed to the strong electron-phonon coupling. In addition, broadening of the emission bandwidth was observed from 118 nm to 129 nm as the temperature was increased from 77 K to 493 K, further confrming the electronphonon interactions (Fig. 4d and e). The temperature dependence of the FWHM in TJU-6 can be described using the following model: 62 Here, G 0 represents the emission FWHM at T ¼ 0 K, E LO is the energy of the longitudinal optical phonon energy, and E b represents the average binding energy of the defect states. G LO and G inh give the relative contributions of exciton-phonon coupling and inhomogeneous broadening (induced by trap states), respectively. A ft to the data gives G 0 ¼ 219 AE 2 meV, G LO ¼ 131 AE 7 meV, G inh ¼ 287 AE 3 meV, E LO ¼ 13 AE 5 meV, and E b ¼ 12 AE 1 meV (Fig. 4e and S15 †). The LO phonon energy obtained from the ftting corresponds to a frequency of $110 cm 1 , which lies well within the range of Pb-Br stretching vibration frequencies from Raman spectroscopy (Fig. S16 †). Based on our photophysical mechanistic studies, we confrm that the broadband emission and the self-trapped states of TJU-6 are mainly attributed to the electron-phonon coupling in a strongly deformable 3D lead halide lattice (Fig. 4f). 63 Recent frst-principle calculations of lead halide materials indicate the electron-phonon coupling in a deformable lead halide lattice forms self-trapped excitons (e.g. Pb 2 3+ , Pb 3+ , and Br 2 species), which act as radiative color centers. 64,65 ## Conclusions We have discovered two hybrid bromoplumbate bulk emitters with an unusual broadband white-light emission spanning the entire visible-light region. These are the frst two 3D lead halide frameworks possessing an overall positive charge, thus exhibiting high distortion of Pb II centers and generated self-trapped states. Upon 360 nm excitation, the [PbBr] + exhibit a moderately high PLQE of 5.6%, exceeding most of the perovskite-type whitelight emitters (<1%). Importantly, the ultrastable nature of our materials affords undiminished photoemission throughout UVirradiation for 30 days under air, largely overcoming the chemically labile problems of 3D lead perovskites. Our mechanistic studies confrm that the broadband emission arises from electron-phonon coupling in a strongly deformable and anharmonic lattice, which contributes to the formation of selftrapped states. The long-sought "inert" lead halide materials are amenable to synthetic design for enhancing PLQEs, thus serving as potential alternatives to commercial white-light LED phosphors. ## Conflicts of interest There are no conflicts to declare.

chemsum

{"title": "Ultrastable, cationic three-dimensional lead bromide frameworks that intrinsically emit broadband white-light", "journal": "Royal Society of Chemistry (RSC)"}

correlation_between_structural,_physical_and_chemical_properties_of_three_new_tetranuclear_ni<sup>ii

## Abstract: Based on a new family of Ni II cubane-like clusters, this work addresses current challenges in the synthesis, analysis and dynamics of single-molecule-magnet (SMM) systems. Investigation of various synthetic routes and desorption-sorption processes yielded a series of isomorphous compounds: [Ni 4 L 4 (CH 3 OH) 4 ]Áxsolv, [Ni 4 L 4 (CH 3 OH) 4 ] and [Ni 4 L 4 ]. In order to analyze these deceivingly simple materials, several analytical and quantum mechanical methods had to be used. This revealed materials with mixed lattice solvents, statistical disorder of the solvent, and disordered [Ni 4 L 4 ] cores which offered an insight into the risks of the self-assembly process and interconversion dynamics of the investigated Ni II family.These findings also allowed structural-magnetic relationships to be established, and the outcomes were exploited in two ways: first, the effect of the coordinated and lattice solvent on the magnetic properties was examined, and second, magnetic properties were used to facilitate crystal structure determination. ## Introduction Single-molecule magnets (SMM) are a specific class of metalloorganic complexes whose magnetic behavior depends on the intrinsic magnetic properties of the individual molecules, i.e. it is a direct consequence of inter-and intramolecular interactions in the system. In general, the choice of ligands will govern the topology of the system, as well as the number and type of intra-and intermolecular interactions. This, in turn, depends on the choice of the synthetic route, and synthetic parameters such as temperature, pH and metal-to-ligand ratio play an important role in the synthesis of the targeted product. As these systems found their potential application in data storage, memory devices, switches and sensors, the targeted synthesis of SMMs became increasingly important. 4 Significant progress in synthesis was made by implementation of various ''bottom-up'' approaches based on the sensible choice of ligands and metal ions. These facilitated the formation of clusters with well-defined geometries and related magnetic properties. In the last two decades, a series of compounds based on polynuclear complexes of Mn, Fe, Ni, V and Co, or mixed metals have been reported, of which cubane-like magnetic clusters with a [Ni 4 (m 3 -O) 4 ] core are especially a wellstudied class. Particular interest in this class of compounds arises from their specific core shape and magnetic exchange pathways, governed by m 3 -atoms. Consequently, extensive investigations of the [Ni 4 (m 3 -O) 4 ] core provided an insight into the delicate relationship between magnetic and structural properties. The importance of the exact geometry of the [Ni 4 O 4 ] core is emphasized in a series of studies, as the symmetry and the corresponding Ni-m 3 -O-Ni angles play a crucial role in the intramolecular magnetic interactions. Moreover, fine details of the cluster structure, such as the type and the amount of solvent, are extremely important, as these factors can determine the magnetic properties of SMM systems. This solvent-effect was first reported in 1999 by Khan and Larinova, who introduced the concept of a ''magnetic sponge'' to describe magnetic materials that can reversibly release and reabsorb both coordinated and non-coordinated solvent molecules. 28 Nowadays, solvent effects are becoming increasingly interesting. In the latest investigation of [Ni 4 L 4 (solv) 4 ] type of complexes it was shown that at ambient temperature, these compounds show reversible exchange of the coordinated solvent molecules (MeOH vs. H 2 O), causing the spin to switch the ground state from S = 4 to S = 0. 29 Even more interesting is the recent research of desolvation of SMM materials (Mn 12 , 30 Fe III 4 , 31 metal-organic frameworks 32 ) reporting a serious effect of lattice (non-coordinated) solvents and experimental conditions on magnetic properties. All these findings inspired us to study the dynamics of exchange of coordinated and noncoordinated solvents in a new family of Ni II compounds. Investigation of synthetic paths, desorption-sorption studies, stability and interconversion processes of these closely related compounds offered a more detailed insight into the outcome of the self-assembly processes. Structural analyses of the obtained compounds, coupled with quantum mechanical studies, were correlated with magnetic measurements. The findings were exploited in two ways. First, the effect of the type and the amount of the solvent in the structure on the magnetic properties was examined, and second, magnetic properties were used to facilitate crystal structure determination. The change in temperature from 110 1C to room temperature yielded green hexagonal prisms of compound 2. Furthermore, adding water to the mixture (2.5 MeOH : 1 H 2 O, volume ratio) resulted in green plates 1, while adding acetic anhydride to the reaction mixture yielded compound 2. In both cases, it was noticed that the synthesis path is not sensitive to the molar ratio of the precursors. Under the same synthetic conditions, both 1 : 1 and 1 : 2 precursor ratios yielded the same products. ## Synthesis Although these experiments showed that 1 and 2 products are kinetically and thermodynamically favored, it also became apparent that water molecules compete with methanol for a place in the crystal lattice, and that the outcome can depend on the temperature and solvent mixture. This flexibility of the systems was further supported by stability and sorption studies. When exposed to room conditions, both materials 1 and 2 changed the colour from green to orange-red. While these conversions resulted in similar products 1a and 2a, their respective kinetics differed. The products 1a and 2a have the same molecular formula, [Ni 4 L 4 (CH 3 OH) 4 ], but their structures are slightly different. Conversion 1 -1a took one day, and 2 -2a only a few hours, indicating different types of intermolecular interactions (in 1 and 2) between the Ni II complex and the noncoordinated solvent. Considering that the reactions resulted in single crystal species, a subtle structural change upon the loss of the non-coordinated solvent could be proposed. In contrast to the room temperature observations described above, heating of clusters 1 and 2 up to 180 1C resulted in the red polycrystalline sample 3 (Scheme 2). When exposed to methanol, the colour of the material changed back to green, indicating that sample 3 is most likely to have the formula Ni 4 L 4 . Additionally, sorption experiments conducted on compounds 1 and 2 with water vapors showed that water molecules can be incorporated as solvents in the crystal lattice (Fig. S1-S18, ESI †). These processes (sorption/ desorption, ageing, and heating) yielded a variety of materials, with different types and amounts of coordinated and crystalline solvents, which required a multitude of analytical techniques to be used: single crystal X-ray diffraction (SCXRD), X-ray powder diffraction (XRPD), thermogravimetry (TG), infrared spectroscopy, chromatography and magnetic measurements. A combination of these techniques with quantum mechanical calculations allowed for characterization of the obtained materials, and thus structure-magnetic relationships to be proposed. ## Thermogravimetric analysis (TG) In order to examine the solvent content and thermal stability of clusters 1, 2 and 3 the samples were heated in an oxygen atmosphere from 25 to 600 1C. Clusters 1 and 2 displayed one-step weight loss, occurring between 25 and 180 1C, and 40 and 140 1C, respectively. Such behaviour was ascribed to the combined loss of lattice and coordinated solvent molecules. However, in the case of compound 1, the relative amounts of water and methanol could not be discerned. The further weight loss occurring between 340 and 430 1C (for 1), and 300 and 430 1C (for 2) was associated with cluster degradation (Fig. S19 and S20, ESI †). Compound 3 was found to undergo a one-step weight loss in the temperature range 302-540 1C, which confirmed the absence of both lattice and coordinated solvent molecules. Thermal stability of the samples obtained by sorption/ desorption or by ageing of materials 1 and 2 showed that the combined loss of lattice and coordinated solvent molecules occurred in one step up to 160 1C and 130 1C. This is very similar to that described for compounds 1 and 2, and the small differences could be a consequence of different contents of solvent molecules. ## Structural studies Single-crystal X-ray structure analysis revealed that compounds 1, 1a, occupy alternate vertices of the cube, resulting in interpenetrating two concentric tetrahedra, one made by the nickel centres and another by the asymmetrically spanning m 3 -O oxygen atoms. Each Ni II ion is six-coordinated by five oxygen atoms and one imino nitrogen atom. Three of five oxygen atoms are oxygen atoms of the methylphenolato moiety which span three nickel centres, one is from the neutral MeOH monodentate ligand and the fifth is the oxygen atom from the salicylideneimine part of the ligand (Fig. 1). In the case of compound 1, methanol and water were present as lattice solvents, while in compound 2, only methanol solvent was found, resulting in general formulae [Ni 4 L 4 (CH 3 OH) 4 ]Á0.32CH 3 OHÁ0.32H 2 O (1) and [Ni 4 L 4 (CH 3 OH) 4 ]Á 0.63CH 3 OH (2). The solvent molecules, MeOH/H 2 O and MeOH, in the unit cell of compounds were highly disordered and their respective amounts could not be derived from SCXRD data only. These values were obtained by combining chromatographic and SCXRD data. Coulometric titration and head-space gas chromatography were used for the determination of water and methanol in 1 and 2, respectively. The unit cell volumes of 1 and 2 differ by 32 3 which is 1% of the unit cell volume. As expected, the crystal structures of compounds 1a and 2a showed a similar Ni II core, with no solvent incorporated into the lattice (Fig. S21, ESI †). The difference in the unit cell volumes of 1a and 2a is not negligible (less than 1% of unit cell volume), and it is an indication of possible small structural differences between 1a and 2a. All four crystal structures (1, 1a, 2, and 2a) exhibit intramolecular hydrogen bonds between hydroxyl groups of coordinated MeOH molecules and phenolate oxygen atoms. However, no intermolecular classical hydrogen bonds were observed. In all investigated structures (1, 1a, 2, and 2a), the tetrameric unit is cross-linked by C-HÁ Á ÁO/N intramolecular hydrogen bonds of expected geometries. The basic supramolecular motif that is common for all structures at either room or low temperature is the dimerization of cluster units via the C-HÁ Á ÁO type of the weak hydrogen bond (Fig. 2). CÁ Á ÁO and HÁ Á ÁO separations are approximately 3.4 and 2.5 , respectively, in the low-temperature structures 1 and 2 and 3.5 and 2.6 in the room-temperature structures 1a and 2a. The HÁ Á ÁO separations fall in the long distance range for HÁ Á ÁA contacts, according to T. Steiner. 33 The pertinent structural parameters are listed in Tables S1-S6 in the ESI † in addition to a detailed description of the molecular and crystal structures (Scheme S1 and Fig. S22-S24, ESI †). Compounds 1 and 1a, and on the other side, 2 and 2a, differ in C-HÁ Á ÁO intermolecular hydrogen bonds formed with noncoordinated MeOH/H 2 O molecules and -C 6 H 5 phenyl groups of the ligand. Their geometries belong to the long distance range of approx. 3.6 (for CÁ Á ÁO) indicating that the solvent molecules are weakly bound to the tetranuclear cluster unit. A detailed inspection of the metrical parameters of these hydrogen bonds reveals that HÁ Á ÁO distances in 1 are generally shorter (2.30-2.75 ) than in 2 (2.86 ) (Table S3, see hydrogen bonds with the O1W and O5 atoms acting as proton-acceptors in 1 vs. one hydrogen bond with the O5A atom as a proton-acceptor in 2, ESI †). Structural analysis of cluster 3, a red polycrystalline material obtained by thermal treatment of cluster 1 (or 2), was carried out using XRPD data, as described in detail in the Experimental section. At this point, it is important to differentiate between the ideal structure that corresponds to the energetically ideal case, as found by DFT-optimizations, and the average structure of the investigated material, which was found to be disordered (XRPD pattern) (Fig. 3). As expected, the ideal structure of compound 3 featured two Ni 4 L 4 species with distorted cubane-type geometry, where each Ni atom exhibits square planar coordination: three oxygens from methylphenolato moieties and one imino nitrogen atom. All corresponding Ni-O-Ni bond angles are found to be larger S3 (ESI †). than 901 (Table S7, ESI †). Considering that Bertrand and co-workers already suggested in 1971 that the deviation of AE141 from 901 angles can be tolerated before the direct ferromagnetic exchange terms cease to be dominant, this structure suggested diamagnetic behaviour of the molecule. 35 Each Ni 4 L 4 cluster exhibits two intramolecular C-HÁ Á ÁO bonds, formed between the oxygen and carbon from methylphenolato moieties. One C-HÁ Á ÁO bond is formed between two parallelly aligned ligands and the other connects two neighbouring perpendicular ligands. The structure of cluster 3 features a series of intra-and intermolecular C-HÁ Á Áp hydrogen bonds, established between the methyl group and the aromatic chelate six-membered ring of the coordinated ligand. In particular, the intramolecular C-HÁ Á Áp bond is formed between one set of parallelly aligned ligands. Intermolecular bonds of the same type showed a much richer network: C-HÁ Á Áp interactions were established between parallelly and perpendicularly oriented neighbouring ligands, forming a rich network across the crystal. This description of the ideal structure of cluster 3 allowed diamagnetic behaviour to be proposed. However, examination of the peak widths in the XRPD pattern (Fig. 4) suggested that the material exhibits both crystalline and disordered domains, while a high background suggested that species with low-range order could also be present in the material. Therefore, the average structure of the material and its influence on magnetic properties required further analysis. Considering the facts that (i) cluster 3 was obtained by thermal treatment of cluster 2a (1a), (ii) the molecular structure is stabilized by C-HÁ Á ÁO and C-HÁ Á Áp interactions and (iii) the crystal structure of cluster 3 is governed mostly by weak C-HÁ Á Áp interactions, the following explanation can be proposed. In order to remove coordinated methanol from material 2a (1a), the molecule needs to adopt a suitable geometry. Four ligands in the Ni 4 L 4 cluster need to be slightly shifted and rotated, which results in increased Ni-O-Ni angles. While this increase generates square-planar geometry of the Ni 4 L 4 cluster, the resulting Ni-O-Ni angles do not need to be increased isotropically. Due to the rich inter-and intramolecular interactions of C-HÁ Á ÁO and C-HÁ Á Áp type, Ni-O-Ni angles can adopt a wide range of values. Consequently, such a range of molecular geometries can exhibit slightly different magnetic properties. Moreover, ambiguities in unit cell determination and the possible presence of species with a local structure also need to be taken into account. First, a limited number of peaks and anisotropic peak broadening allowed only average (most probable) values of unit cell parameters to be determined: existence of other, slightly different cells or supercells cannot be excluded. Second, the possible local structure will also contribute to the magnetic behaviour. Thus, deviations from ideal magnetic values can be expected. However, although the material features disordered domains, it appears to be thermodynamically stable. Comparison of the initial XRPD pattern with the pattern collected after two months, using the same setup, revealed no differences. ## Sorption and stability studies In a series of sorption and ageing studies, the stability of synthesised compounds, desolvation-solvation processes, and possibilities of interconversions of non-coordinated solvents were investigated. Water sorption studies on 1 and 2 clusters. Water sorption/ desorption isotherms were measured using an accurate humidityand temperature-controlled microbalance system, at 26 1C. The RH was increased in steps from 0 to 90%, where each step corresponds to an increase of relative humidity of 10% (1 cycle). Similarly, desorption (90% to 0% RH) was carried out cycle-wise, where each cycle corresponds to a decrease of relative humidity of 10% (Tables S8 and S9; Fig. S25 and S26, ESI †). At 90% RH, both complexes 1 and 2 showed that water sorption resulted in an increase of mass of less than 0.1%. The corresponding desorption curve did not match the absorption curve, and showed a decrease in mass of about 2.5% and 4% for compounds 1 and 2, respectively. Considering that both structures 1 and 2 feature dynamical disorder (0.32H 2 O and 0.32CH 3 OH for structure 1, and 0.63CH 3 OH in compound 2) these values indicate that during sorption, water molecules are incorporated into the structure until voids in the crystal are fully occupied by water, that is, until every unit cell is occupied by solvent molecules. The morphology of both the clusters was held but strong surface erosion was observed in cluster 2 (Fig. S27, ESI †). Interconversions under non-ambient conditions. Ageing of compounds 1 and 2 was tested by exposing them to solvent vapors in three cycles. In the first cycle, compound 1 (2) was exposed to 75% RH for seven days. In the second, the obtained material was exposed to MeOH vapors, and in the third, again to water. In both cases (1 and 2), comparison of the XRPD patterns obtained at each cycle showed differences in peak positions and their relative intensities (Fig. 5a). These differences showed that H 2 O/MeOH recycling does not lead to a unique product and indicated that both lattice and coordinated solvents can be exchanged (for details, see Fig. S1-S18, ESI †). In recycling by annealing, compound 1 (2) was annealed at 180 1C and then exposed to MeOH vapors in three cycles. Comparison of XRPD patterns collected for samples exposed to MeOH vapors revealed differences in relative peak intensities (green in Fig. 5b), showing once again that the process is not fully reversible. If these differences are correlated with TG-studies, they can be explained as variations in the amount of lattice solvent present in the material. Similar differences observed between XRPD patterns obtained by annealing (red in Fig. 5b) suggest that the Ni 4 L 4 core can adopt a slightly different structure. ## Magnetic studies In order to establish magneto-structural correlations, two series of magnetic investigations were carried out for compounds 1, 2 and 3 (Fig. S28-S33, ESI †). The first series (heat-MeOH cycle) temperature dependence of the molar magnetic susceptibility of (a) clusters 3 -3 3 obtained after annealing of 1 -1 3 and (b) clusters 1 1 -1 3 obtained by exposing clusters 3 -3 3 to MeOH vapors is shown in Fig. 6 (for 2 Fig. S32, ESI †). The second series (water-MeOH cycle) changes of magnetic behavior are induced by exposure of 1 (or 2) to water and methanol vapors alternatively (Fig. 7). The return to the initial state is almost equally effective in both cycles after exposure to methanol vapors. Small differences in the efficacy are connected with small differences in the dynamics of solvent release observed in other experiments. The Curie constant C of the first series clusters 1 -3 - 3 , in the heat-MeOH cycle, has a mean value of 5.1(3) K emu mol 1 Oe 1 . This value confirms that in the starting green cluster 1 and recycled compounds (also green clusters 1 1 , 1 2 and 1 3 ) the four spin 1 nickel ions per formulae unit are magnetically uncoupled at the high temperatures. The obtained g-factors (2.26, 2.21, 2.19 and 2.19) are in agreement with the expected values for Ni ions in octahedral coordination, within the 2% error of the measurement. 12,15,25 Below B100 K, the susceptibility starts to deviate from the Curie-Weiss law, and the wT product increases, reaching max. at 3.8 K, suggesting the ferromagnetic coupling within the cubane. The red cluster 3 (and also red clusters temperature behaviour of the susceptibility with the Curie-Weiss law gives the C values in the range from 1.4 to 3.4 K emu mol 1 Oe 1 . This decrease indicates the partial transition, between 24 and 64% (sample dependent), of Ni II ions from spin 1 to the diamagnetic state. The explanation of this effective magnetic moment reduction is possible only with the transition of some Ni II ions from spin 1 to spin 0 state. Namely, the obtained C constant is too small to originate from the four spins 1 per cubane unit, and because of the structural improbability of the large antiferromagnetic coupling surviving in the whole temperature range, reduction of the number of magnetic units is the obvious prerequisite. It is known that spin 1 to spin 0 transition for nickel ions occurs with the change of coordination when the electron levels of t 2g and e g orbitals change their order when compared to the octahedral coordination in the original complexes. The green-red-green interconversions observed in the series of conducted experiments (1 -3 -1 1 . . .) were assigned to the reversible release and reabsorption of the solvent and to reversible change of coordination around Ni II from six to four. However, not all Ni II ions come to the diamagnetic state, possibly due to the remaining disorder around some nickel ions preventing against the ideal surrounding needed for transition to low spin. The observed downturn of wT with the lowering of the temperature is similar in all compounds. This decrease is coming from the zero-field splitting. 29 Susceptibility of a series of samples obtained after successive exposure to water vapors and recycling in methanol vapors alternately is shown in Fig. 7 (for 2 Fig. S33, ESI †). In the case of the water-methanol cycle, 1 -3 w -1 1 -3 w1 -1 2 -3 w2 -1 3 -3 w3 , the decrease of Curie constant and Weiss parameter values after exposure to water vapors is less expressed than in the case of the heat-methanol cycle. Exposure of 1 to water vapors leads to some more complex mixed state in the sense that it is not saturated, but magnetization is still higher than expected for independent ions. Quantitative analysis of these data is not practical since there are many possible combinations of partial transitions within cubanes, making the fitting procedure overparameterized and impossible to perform. However, the observations are informative enough to help in explaining the magnetic changes within the light of solvatomagnetic effects. Effective exchange interaction parameters J, g-factors, and percentage p of the remaining S = 1 ions for cluster 1 (for 2 Table S10, ESI †) in both cycles are given in Table 1. The obtained values of J are in accordance with similar structures, 29 where similar corresponding angles within the Ni4 core were reported. Since the maximum of susceptibility is reached at 3.8 K resulting from the exchange coupling of cubane, and the M(H) dependence is reversible and linear for small fields down to 2 K, there is no possibility to observe SMM phenomena under investigated conditions, as was the case in other Ni4. 29 More details about modelling of magnetic behaviour as well as the corresponding results of the equivalent series of compound 2 are presented in the ESI. † ## Quantum chemical calculations Geometry optimizations for the [Ni 4 L 4 (CH 3 OH) 4 ] cluster and its modeled derivative were performed using the density functional theory with two hybrid functionals B3LYP and MN12SX. Calculations started from the experimentally determined crystal structures of compounds 1a and 2a (C 2 point group of symmetry). In each case, all calculated geometrical parameters were very similar to the experimental ones, thus confirming the validity of the methods chosen. In order to investigate how the lattice solvent affects the magnetic properties of the compound, models of the [Ni 4 L 4 (CH 3 OH) 4 ] cluster with a toluene molecule placed near the phenyl ring of the ligand (L 2 = C 6 H 5 (O)-CQN-(O)C 6 H 4 CH 3 ) were constructed (Fig. 8). Since the only experimentally observable difference between unit cells of 1 and 1a in comparison with 2 and 2a was essentially a slight change in the C-H-p distance (Fig. 2), a toluene molecule was placed in two different positions, mimicking the crystal structures of 1 and 2 (Fig. 8). After a series of geometry optimizations, no significant geometrical change in the (Ni II ) 4 core could be observed in either of the models. Natural bond orbital (NBO) analyses performed on each structure revealed that the Fig. 7 Temperature dependence of molar magnetic susceptibility measured in the field of 0.1 T multiplied with temperature, for cluster 1 series, measured successively after exposure to water vapor and recycling in methanol vapor alternately. Lines are fitting curves. Inset: Field dependence of magnetization per cubane unit measured at the temperature of 2 K. Table 1 Effective exchange interaction parameters J, g-factors, and percentage p of the remaining S = 1 ions for 1 in both cycles: heatmethanol (1 -3 -1 1 -3 1 -1 2 -3 2 -1 3 -3 3 ) and water-methanol (1 -3 w -1 1 -3 w1 -1 2 -3 w2 -1 3 -3 w3 ) Step Heat-methanol cycle of cluster 1 Water-methanol cycle of cluster 1 change in the Ni-solvent distance has a negligible effect on the NBO occupancies of nickel atoms (Tables S11-S14, ESI †). Consequently, the observed differences in magnetic properties are most probably not caused by minor changes within the unit cells of 1 and 1a in comparison with 2 and 2a, but are due to the different solvent disorder acting as an outer electrostatic potential for the (Ni II ) 4 core. Loss of the coordinated solvent caused the colour to be changed to red, and a partial transition of the spin of Ni II ions from the ground state from S = 1 to S = 0. This partial transition could be assigned to temperature-induced disorder of the Ni 4 L 4 core in the material. Several cycles of annealing-MeOH sorption experiments showed that the process is not fully reversible. Namely, the geometry of the N 4 L 4 core was shown to vary, and materials obtained by sorption processes were also found to differ, probably in the amount of non-coordinated solvent. Similarly, a series of experiments where compounds 1 and 2 were consecutively exposed to H 2 O and CH 3 OH vapors showed differences in their powder diffraction patterns. This indicated that the number and amount of coordinated and non-coordinated solvents can vary, and are not easy to predict. Moreover, according to the quantum-mechanical calculations, it is the disorder that plays an important role in determining magnetic properties of the material. Existence of mixed solvates, and statistical and dynamic disorder of solvent molecules presented here indicate that there are multiple risks in the design of materials with desired magnetic properties: (i) fully controlled synthesis, (ii) establishing the exact magneto-structural relationship, (ii) water/methanol ambiguity, and (iii) stability of synthesised products. Detailed interpretation of the structural and magnetic correlation remains a challenge for future studies of these systems. ## Synthesis and materials All chemicals obtained from commercial sources were analytically pure and were used without further purification. The Schiff base N-(2-hydroxy-5-methylphenyl)salicylideneimine was prepared according to the literature data. (d) The same crystalline product 2 was obtained using a similar procedure as in (a) in the mixture of dry methanol (20 mL) and acetic anhydride (1 mL). (e) Cluster 3, as dark red prisms or plates, was obtained after removal of the solvent molecules (coordinated and solvated) of ] with toluene (B3LYP/ 6-31G(d) level of theory); interatomic distances frozen during optimizations are given using dashed lines and the solvent accessible surface is displayed. 36 The other complex with slightly shifted toluene molecules is not presented here. modules in sealed aluminium pans (40 mL), by heating in a flow of nitrogen or oxygen (200 mL min 1 ) at 5 or 10 1C min 1 . The data collection and analyses were performed using the program package STARe Software 9.01. 38 Chromatography. The content of methanol and water was determined using Headspace-GC system Agilent Technologies 7890 and Metrohm 831 KF coulometer, respectively. Calcd DVS experiments. The water sorption isotherms were measured using an accurate humidity and temperature-controlled microbalance system, Dynamic Vapour Sorption (DVS 1, Surface Management Systems, UK). The relative humidity (RH) was increased in steps from 0 to 90% and back to 0% at 26 1C. The equilibration condition at each step of the RH for the rate of change in mass with time (dm/dt) was selected (Fig. S25 and S26, Tables S8 and S9, ESI †). Magnetic studies. All compounds in the form of crushed single crystalline samples were measured using a commercial MPMS5 SQUID magnetometer. Contribution of ampoule was subtracted properly and the susceptibility was corrected with respect to the temperature independent diamagnetic and paramagnetic contributions. Temperature dependence of magnetic susceptibility was calculated from measurements in 0.1 T magnetic field. Field dependence of magnetization was measured at several temperatures in order to complement the susceptibility analysis. Single crystal X-ray crystallography. The selected geometries including valence bonds of each Ni II coordination sphere, NiÁ Á ÁNi distances, the comprehensive hydrogen bond geometry, C-HÁ Á Áp interactions and some dihedral angles are given (Tables S1-S5, ESI †). The data collection was performed by applying the CrysAlis Software system, Version 1.171.32.24, 39 on the Oxford Xcalibur diffractometer equipped with a CCD detector and using graphite-monochromated MoKa radiation (l = 0.71703 ) at 296 K (1a and 2a) and 150 K (1 and 2) and o-scans. The programs CrysAlis CCD and CrysAlis RED (Version 1.171.33.41) 39 were employed for data collection, cell refinement and data reduction. The Lorentz-polarization effect was corrected and the diffraction data were scaled for absorption effects by the multiscanning method. Structures were solved by direct methods and refined on F2 by weighted full-matrix least-squares. The programs SHELXL-2014 40 and the WinGX v. 1.80.05 41 software system were used to solve and refine structures. All non-hydrogen atoms were refined using the anisotropic displacement parameters. The solvent content in 1 and 2 was crystallographically modelled. The solvent molecules are found to be present in the lowtemperature crystal structures with the partial occupancy factors of atoms (H 2 O/MeOH = 0.32 : 0.32 in 1 and 0.63 in 2), which have been refined as free variables, but at the final least-square refinement cycle, in order to obtain the final molecular formulae, the occupancies were fixed. By using SUMP instructions the sum of the occupancies of the two possible oxygen positions O5A and O5B in 2 was constrained to 0.63. The hydrogen atoms belonging to the stereochemically different carbon atoms and due to the data collection temperature were placed in the geometrically idealized positions with assigned isotropic displacement parameters and they were constrained to ride on their parent atoms by using the appropriate SHELXL-2014 HFIX instructions. The hydrogen atoms belonging to the coordinated methanol molecules were firstly found in the difference Fourier maps at the final stages of the refinement procedure and then refined by SHELXL-2014 DFIX instructions at 0.82(1) and 0.84(1) (for 1a and 2a compounds at 296 K and for 1 and 2 at 150 K, respectively) with assigned isotropic displacement parameters being 1.2 times larger than the equivalent isotropic displacement parameters of the parent oxygen atoms. The hydrogen atoms which belong to the methyl group of the solvent methanol molecule (or water) are not modelled. The molecular geometry calculations and graphics were done using ORTEP-III 42 integrated in the WinGX software system, PLATON 43 programme and Mercury. 44 Supplementary crystallographic data (atomic coordinates, thermal parameters, all intramolecular distances and angles for all structures) are given in the ESI † (in the CIF format). X-ray powder diffraction (XRPD). All X-ray powder diffraction (XRPD) experiments of the samples were performed on a PHILIPS PW 1840 X-ray diffractometer with CuKa 1 (1.54056 ) radiation at 40 mA and 40 kV. The scattered intensities were measured using a scintillation counter. The angular range (2y) was from 3 to 501 with steps of 0.021, and the measuring time was 0.5 s per step. The data collection and analysis were performed using the program package Philips X'Pert. 45,46 Structural analysis of compound 3 involved several XRPD patterns. Initially, two materials, one obtained by heating compound 1 and the other obtained by heating compound 2, were measured. Although comparison of these two patterns revealed a poorly crystalline substance, these two patterns were not the same. This suggested that the molecular moieties in these two materials have different arrangements. For the purpose of structure determination of cluster 3, a fresh sample of cluster 1a was produced, and then heated up to 180 1C. This sample was first examined under the microscope, which revealed small crystals with cracked surfaces, resembling a deposit of many thin platelets closely stacked together (Fig. S40, ESI †). Grinding the sample in the mortar resulted in paste-like material that was filled in a 1 mm capillary. X-ray powder diffraction (XRPD) data were measured using the MYTHEN detector 47 installed at the Swiss Light Source, 48 using an X-ray energy of 17.7 keV (0.7012 ). Any radiation damage was avoided by collecting 1201 of data within ten seconds (Table S15, ESI †). The measured XRPD pattern showed only a few sharp peaks and several broad ones, resulting in low data resolution (d min = 3 ). The XRPD pattern was compared to the calculated patterns of 1a (2a), and observed similarities led to the conclusion that these structures are correlated. Therefore, unit cell parameters of 1a (2a) were used to facilitate the indexing of the pattern of cluster 3, and the starting model of the [Ni 4 L 4 ] cluster could be built using the structural information of 1a (2a). Considering that magnetic measurements of cluster 3 revealed antiferromagnetic properties, a planar-square geometry of the [Ni 4 L 4 ] cluster was expected. In order to construct such a molecular structure, coordinated methanol from the 1a structure was removed (Fig. 3a, (a)), and this model was subjected to a series of energy optimizations 49 (for details, see Computational methods). As expected, all calculations resulted in clusters of four square-planar coordinated Ni-centers, with Ni-O-Ni angles larger than 991, suggesting the antiferromagnetic properties of the molecule (Fig. 3b, (b)). 24,29 This molecular structure and unit cell parameters calculated from XRPD data were used as input for the crystal structure determination program FOX, based on direct-space optimization. 50 Although all investigated samples (1, 2, 1a, and 2a) exhibited P% 1 symmetry, direct-space optimization processes were carried out in P1 symmetry without any restraints, in order to minimize model bias. However, this approach resulted in chemically non-reasonable structures. The reason for this failure was sought in the preferred orientation of the crystals, resulting in incorrect assignment of the relative intensities of the peaks. This obstacle was overcome by employing soft geometrical restraints. This way, solutions with chemically reasonable intermolecular distances could be favored. The obtained crystal structure of cluster 3 was subjected to Rietveld refinement, using the XRS suite of programs. 51 Considering that insufficient data resolution did not allow for refinement of the atomic coordinates, only non-structural parameters (peak shapes and positions, zero shift, sample displacement, and scale) were refined (Fig. 4 and Table S15, ESI †). Subsequently, the geometry of the resulting crystal structure was optimized using the quantum mechanical approach (Fig. S41a, ESI †). Comparison between the optimized (ideal) molecular structure and the molecular structure obtained from FOX (average) is given in Fig. S41b (ESI †). All discrepancies between two models are considered to be a consequence of disorder in the material. Computational methods. Quantum chemical calculations (compounds 1 and 2, toluene model) were carried out using the Gaussian 09 program package. 52 Geometry optimization of ground states was performed using hybrid functionals B3LYP 53,54 and MN12SX 55 in combination with the 6-31G(d) basis set. For all optimized structures harmonic frequencies were calculated to ensure that the obtained geometries correspond to the minimum on the potential energy surface. To simulate the influence of packing within the unit cell, models containing the [Ni 4 L 4 -(CH 3 OH) 4 ] complex and a toluene molecule were built. In each of the two separate models the toluene molecule was placed in a different position near the phenyl ring of L and interatomic distances between the carbon atom of the toluene's methyl group and carbon atoms of the phenyl ring were fixed during the optimization. Natural bond orbital (NBO) analysis 52 on all optimized structures was conducted using the Gaussian 09. Ab initio calculations for structure determination of compound 3 were performed within the framework of density functional theory (DFT) as implemented in the VASP code 56 using the projector augmented wave (PAW) method. 57 We have used the self-consistently implemented vdW-DF 58,59 functional for correlation in combination with optB88 exchange. In all calculations the expansion in plane waves was done using the cut-off energy of 750 eV. The Brillouin zone was sampled by the Monkhorst-Pack choice of k-points, namely 2 1 1. The atoms were allowed to relax until the forces on them were below 1 meV 1 .

chemsum

{"title": "Correlation between structural, physical and chemical properties of three new tetranuclear Ni^II clusters", "journal": "Royal Society of Chemistry (RSC)"}

responsive_hetero-organelle_partition_conferred_fluorogenic_sensing_of_mitochondrial_depolarization

## Abstract: Malfunctioning organelles are often difficult to probe with classical organelle-homing sensors owing to disruption of physiological organelle-probe affinity. We herein report the use of a responsive heteroorganelle partition and signal activable probe (RC-TPP) for detecting mitochondrial depolarization, a pathologically relevant event featuring loss of the electrical potentials across the mitochondrial membrane (DJ m ). Partitioned in mitochondria to give blue fluorescence, RC-TPP relocates into lysosomes upon mitochondrial depolarization and exhibits red fluorescence triggered by lysosomal acidity, enabling determination of autophagy relevant mitochondrial depolarization and the chronological sequence of mitochondrial depolarization and lysosomal neutralization in distinct cell death signalling pathways. As an alternative to classic homo-organelle specific molecular systems, this hetero-organelle responsive approach provides a new perspective from which to study dysfunctional organelles. ## Introduction Defning the parameters of dysfunctional organelles is valuable for probing their functionality in diseases. Mitochondria are dynamic organelles hallmarked by electrochemical gradients of protons and negative electrical potentials across the mitochondrial membrane (DJ m ), which are critical for diverse biological activities such as ATP biosynthesis. 1 By contrast, mitochondrial depolarization, owing to loss of DJ m , triggers distinct cellular and pathological events, e.g. autophagy and cell death. 2 Mitochondria are routinely imaged with cationic dyes, e.g. rhodamine 123, which accumulate in mitochondria, driven by DJ m . 3 Albeit widely used, these DJ m -responsive dyes quickly leak from mitochondria upon loss of DJ m . Alternatively, cationic dyes with a reactive moiety have been employed to form covalent bioconjugates with intramitochondrial protein sulfhydryls once sequestered inside mitochondria. 4 Although the covalent linkage between dyes and proteins prevents loss of intramitochondrial fluorescence on dissipation of DJ m , these probes are incapable of discerning depolarized mitochondria. Recently, detection of mitochondrial depolarization was achieved using a bipolar probe featuring a positively charged hydrophilic group and an environment sensitive fluorophore which exhibits an altered fluorescence lifetime related to membrane polarization. 5 Herein we report responsive heteroorganelle partition mediated detection of mitochondrial depolarization using RC-TPP, which preferentially accumulates in mitochondria to give blue fluorescence. RC-TPP relocates from depolarized mitochondria into acidic lysosomes to give intense rhodamine fluorescence, allowing a signal-on report of mitochondrial depolarization (Fig. 1). ## Results and discussion Given the myriad roles of altered DJ m in cell homeostasis and death, 2 it is imperative to detect mitochondrial depolarization. As escaping of DJ m -responsive dyes from depolarized mitochondria is concomitant with loss of mitochondrion-specifc fluorescence, we envisioned that conjugation of mitochondriatargetable agents with a profluorophore that is activable by extra-mitochondrial factors could potentially yield a signal-on assay for mitochondrial depolarization. Rhodamine-lactams, a group of nonfluorescent rhodamine derivatives featuring intramolecular spiro-lactam under alkaline conditions, are prone to proton triggered opening of the lactams to give bright rhodamine fluorescence. 6 We previously used rhodamine-lactams for imaging of lysosomes. 7 As such, RC-TPP was constructed to combine a coumarin fluorophore to indicate the sensor's intracellular distribution, a DJ m responsive TPP moiety widely used to ferry various cargoes into mitochondria, 8 and a rhodamine-lactam profluorophore activable by acidic lysosomes, which are the ubiquitous digestive organelles of mammalian cells (Fig. 2A). Acidic pH mediated switched-on fluorescence of RC-TPP pH titration of RC-TPP shows "always-on" coumarin fluorescence whereas rhodamine fluorescence, negligible at pH 7.0-9.0, occurred in acidic media and intensifed as pH decreased (pH 6.5-4.5) (Fig. 2B and C), which is consistent with proton mediated opening of the intramolecular lactam (Fig. 2A). Lysosomes are of acidic pH whereas mitochondria contain alkaline lumen with the pH maintained by the transmembrane proton gradients. 9 The ratiometric titration shows distinct blue to red fluorescence patterns of RC-TPP respectively matching pH windows of lysosomes (pH 4.5-6.5) and mitochondria ($pH 8.0) (Fig. 2D). Accumulation of RC-TPP in polarized mitochondria RC-TPP was evaluated for its ability to target mitochondria. HeLa cells stained with RC-TPP (1 mM) display negligible red fluorescence and strong blue fluorescence colocalized with Mitrotracker Green specifc for mitochondria (Fig. 3A; ESI, Fig. S1 †). Cells that are stained with an elevated concentration of RC-TPP (3 mM) exhibit sparsely punctate rhodamine signals eclipsing lysosomes bearing Lamp2-GFP (GFP-tagged lysosome associated membrane protein 2) (Fig. 3B). These results show that RC-TPP has a high tendency to be partitioned in mitochondria over lysosomes, and is prone to lysosome activation to give red fluorescence. We next probed the structural factor that leads to accumulation of RC-TPP in mitochondria. Rhodamine- To ascertain the dependence of rhodamine fluorescence of RC-TPP on lysosomal acidity, HeLa cells were probed with RC-TPP in the presence of baflomycin A1 (BFA), which inhibits vacuolar ATPase-H1 pump and neutralizes lysosomes. 11 The lack of rhodamine fluorescence in BFA-treated cells proves lysosomal acidity triggered rhodamine fluorescence of RC-TPP in live cells (Fig. 4, ESI, Fig. S4 †). The demonstrated translocation of RC-TPP into lysosomes is ascribed to loss of affinity of TPP to mitochondria upon depolarization, and protonation of the rhodamine-lactam by lysosomal acidity and possibly the contribution of the tendency of high molecular weight dyes to accumulate in lysosomes. 12 Taken together, these data clearly evidence DJ m mediated intra-partition of RC-TPP between mitochondria and lysosomes, and lysosomal acidity mediated activation of RC-TPP upon mitochondrial depolarization. Fluorogenic sensing of autophagy-associated mitochondrial depolarization with RC-TPP Mitochondrial depolarization correlates with autophagy, a catabolic mechanism mediating cell degradation of unnecessary or dysfunctional cellular components in response to diverse cues such as starvation. 13 Cells cultivated in Hanks' balanced salt solution (HBSS) free of amino acids display dramatically enhanced rhodamine fluorescence (Fig. 5), which is consistent with nutrient starvation induced mitochondrial depolarization. 13 Compared with red signal-free control cells or cells treated with apoptosis-inducing staurosporine (STS), there is moderate rhodamine fluorescence induced by the autophagy inducers rotenone or rapamycin (Fig. 4), 14 demonstrating the utility of RC-TPP in distinguishing the efficacy of autophagy inducers. Flow cytometry analysis confrms enhanced rhodamine fluorescence in CCCP-or HBSS-treated cells, and reveals the synergistic effects of CCCP and HBSS on mitochondrial depolarization (Fig. 6), proving the usefulness of RC-TPP for fluorogenic sensing of mitochondrial depolarization induced by distinct biochemical cues. Historically, distinct methods have been developed to detect autophagy using GFP-tagged LC-3, 15 pH-reporting proteins 16 or dyes covalently immobilized to mitochondria. 17 Complementing these approaches, our method focuses on the analysis of mitochondrial depolarization, a vital event preceding autophagy, using a small molecular probe prone to DJ m mediated partition between mitochondria and lysosomes. ## Analysis of mitochondrial depolarization during apoptosis and necrosis Dying cells undergo mitochondrial depolarization and loss of lysosomal acidity owing to membrane permeabilization. 18 As distinct cell death modalities, apoptosis occurs under physiological conditions and is critical for cell number homeostasis whereas necrosis triggered by external factors leads to inflammatory responses. 19 To date, the chronological order of mitochondrial depolarization and lysosomal neutralization during cell death has been undefned. Tom20-GFP + HeLa cells stained with RC-TPP were treated with human Tumor Necrosis Factor-a (TNF) and Smac to trigger apoptosis. HeLa cells lack the endogenous receptor-interacting protein 3 (RIP3), a mediator which is critical for necrosis. Hence RIP3 + /Tom20-GFP + HeLa cells, used to reconstitute necrosis, were loaded with RC-TPP and then treated with TNF, Smac and Z-VAD to trigger necroptosis, which is a programmed form of necrosis. 20 Time course imaging reveals subsided localization of RC-TPP in mitochondria in both cell populations, whereas punctate rhodamine signals occurred in apoptotic cells over necroptotic cells (Fig. 7 and 8). Given the dependence of rhodamine fluorescence on lysosomal acidity and translocation of RC-TPP from mitochondria into lysosomes, these results suggest that mitochondrial depolarization precedes lysosome membrane permeabilization in apoptosis, but the succession is swapped during necroptosis. ## Conclusions RC-TPP was developed for signal-on sensing of mitochondrial depolarization via lysosomal acidity activated fluorescence. RC-TPP, accumulated in mitochondria to exhibit blue fluorescence, relocates to acidic lysosomes to give rhodamine fluorescence upon mitochondrial depolarization, enabling View Article detection of mitochondrial depolarization during autophagy and determination of the chronological sequence of mitochondrial depolarization and lysosomal neutralization during distinct cell death pathways. Abnormal organelles are manifested in diverse pathological conditions. 21 Defning dysfunctional organelles with classic sensors is challenging owing to frequent loss of physiological organelle-probe affinity. Complementing mono-organelle targeting approaches, 22 the heteroorganelle responsive approach offers a new tool to study malfunctioning organelles. ## Materials and methods Lysotracker green DND-26 and Mitotracker green were purchased from Invitrogen. Baflomycin A1 was purchased from Selleck. Rotenone and carbonyl cyanide m-chlorophenylhydrazone (CCCP), human tumor necrosis factor-a, Smac mimetic, and benzyloxycarbonyl-Val-Ala-DL-Asp(O-methyl)fluoromethyl-ketone (Z-VAD-FMK) were obtained from Sigma. All other chemicals were obtained from Alfa-Aesar unless specifed. Cell culturing and overexpression. HeLa, HEK293T, and L929 cells were obtained from American Type Culture Collection (ATCC). Transient transfection of 293T cells was performed using a calcium phosphate method. Lentiviral infection was used for stable expression. Recombinant lentiviruses were packaged in 293T cells in the presence of helper plasmids (pMDLg, pRSV-REWV and pVSV-G) using a calcium phosphate precipitation method. The transfected cells were cultured for 48 h and the viruses were collected for infection. All cells were maintained in Dulbecco's modifed Eagle's medium (DMEM), supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 IU penicillin, and 100 mg ml 1 streptomycin at 37 C in a humidifed incubator containing 5% CO 2 . Full-length cDNA of Tom20 was cloned into BamHI and XhoI sites of the lentiviral vector pBOB-GFP using the Exo III-assisted ligase-free cloning method as described. 23 All plasmids were verifed by DNA sequencing. For lentivirus production, HEK293T cells were transfected by the calcium phosphate precipitation method. The virus-containing medium was harvested 36-48 h later and was added to HeLa and RIP3 + HeLa cells. Fluorescence spectra and confocal fluorescence imaging. The fluorescence spectra of RC-TPP were performed on a Spectra Max M5. Confocal microscopic imaging was performed on a Zeiss LSM 780 using the following flters: l ex@405 nm and l em@415-491 nm for coumarin, l ex@488 nm and l em@490-553 nm for Lysotracker green, Mitotracker green and GFP, and l ex@565 nm and l em@593-735 nm for rhodamine. Images of the fluorescence of RC-TPP and Lysotracker green in cells were merged using Photoshop CS6. Quantitative imaging analysis was carried out on unprocessed images using ImageJ software. Graphs were generated using GraphPad Prism5 and Origin 8.0 software. Synthesis of RC-TPP. ROX-EDA (2.13 g, 4 mmol) 6a and (2-bromoethyl)-triphenylphosphonium bromide (1.79 g, 4 mmol) were dissolved in dichloromethane (DCM) (40 ml) followed by addition of aqueous potassium carbonate solution (20 ml). The reaction solution was stirred for 2 h, diluted with DCM (40 ml), and then washed with brine (40 ml). The organic solution was dried over anhydrous Na 2 SO 4 . After evaporation of the solvent, ROX-TPP was obtained as an off-white solid in quantitative yield (3.28 g). 1 H-NMR (500 MHz, DMSO) d: 7.85 (dt, J ¼ 7.4, 5.5 Hz, 3H), 7.78-7.68 (m, 13H), 7.49-7.43 (m, 2H), 6.97-6.92 (m, 1H), 5.88 (s, 2H), 3.57 (dt, J ¼ 12.9, 6.5 Hz, 2H), 3.09 (t, J ¼ 5.4 Hz, 4H), 3.03 (t, J ¼ 5.3 Hz, 4H), 2.86 (t, J ¼ 6.7 Hz, 2H), 2.83-2.72 (m, 4H), 2.63-2.53 (m, 2H), 2.44-2.30 (m, 4H), 2.04 (t, 2H), 1.97-1.85 (m, 4H), and 1.69 (pd, J ¼ 12.8, 6.4 Hz, 4H). 13 7-Diethylamino-4-carboxy-coumarin (2.6 g, 10.0 mmol) was dissolved in anhydrous DCM (40 ml) followed by addition of oxalyl chloride (2.52 g, 20.00 mmol) and a drop of N,N-dimethylformamide. The mixture was stirred at RT for 2 h and then evaporated to remove excess oxalyl chloride. The residue was dissolved in anhydrous DCM, and then added dropwise to the solution of ROX-TPP (0.821 g, 1 mmol) in anhydrous DCM spiked with triethylamine (1 ml). The mixture was stirred at RT for 1 h and then evaporated to remove the solvent. The residue was purifed by silica gel column chromatography using DCM/MeOH/TEA (20 : 1 : 1, v/v/v) as the eluent. RC-TPP was obtained as an off-white solid (0.64 g, 60%). 1 H-NMR (500 MHz, CD 3 OD) d: 7.97-7.63 (m, 18H), 7.51-7.48 (m, 2H), 7.34 (d, J ¼ 8.9 Hz, 1H), 6.92 (d, J ¼ 5.8 Hz, 1H), 6.83 (d, J ¼ 8.8 Hz, 1H), 6.61 (s, 1H), 5.82 (s, 2H), 3.56 (dd, J ¼ 14.0, 6.9 Hz, 8H), 3.37 (t, J ¼ 6.0 Hz, 2H), 3.16-3.02 (m, 10H), 2.78-2.69 (m, 2H), 2.66-2.56 (m, 2H), 2.41 (m, J ¼ 14.3, 7.5 Hz, 2H), 2.38-2.30 (m, 2H), 1.97-1.86 (m, 2H), 1.81 (dt, J ¼ 11.9, 5.8 Hz, 4H), and 1.28 (t, J ¼ 6.9 Hz, 6H); 13 Synthesis of ROX-coumarin (RC). To a solution of rhodamine 101 (ROX, 4.92 g, 10 mmol) in methanol (20 ml) was added diethylenetriamine (10.3 g, 100 mmol). The reaction mixture was maintained with stirring in an oil bath at 100 C for 2 h. After evaporation of methanol the residue was extracted with DCM (100 ml) and saturated aqueous NaHCO 3 solution (100 ml) twice. The combined organic phases were dried over anhydrous Na 2 SO 4 and then concentrated to remove the solvent. The residue was purifed by silica gel column chromatography using DCM/MeOH/TEA (20 : 1 : 1, v/v/v) as the eluent to give ROX-ETA as an off-white solid (4.88 g, 85%). 1 H-NMR (500 MHz, CDCl 3 ) d: 7.91-7.86 (1H, m), 7.45-7.40 (2H, m), 7.07 (1H, dt, J ¼ 4.6, 3.3 Hz), 5.98 (2H, s), 3.24 (2H, t, J ¼ 6.4 Hz), 3.15-3.11 (4H, t), 3.10-3.06 (4H, t), 2.89 (4H, dt, J ¼ 10.0, 4.8 Hz), 2.60 (2H, t, J ¼ 5.9 Hz), 2.47 (4H, dd, J ¼ 11.1 HZ, 6.1 Hz), 2.41 (4H, t, J ¼ 6.1 Hz), 2.06-1.99 (4H, m), and 1.88-1.82 (4H, m); 13 To a solution of 7-diethylamino-4-carboxy-coumarin (2.61 g, 10.00 mmol) in anhydrous DCM (40 ml) was added N-hydroxysuccinimide (NHS, 1.73 g, 15.0 mmol) and 1-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride (2.9 g, 15.0 mmol). The mixture was maintained at RT for 2 h with stirring. After removal of the solvent, the residue was purifed by silica gel column chromatography using DCM as the eluent to yield the NHS ester of 7-diethylamino-coumarin-3-acid as a pale yellow solid (1.88 g, 50%). To a solution of ROX-ETA (2.4 g, 4.2 mmol) in anhydrous dichloromethane was added the NHS ester of 7-diethylaminocoumarin-carboxylic acid (1.8 g, 5 mmol) with TEA (1 ml). The reaction mixture was stirred at room temperature for 12 h. After rotary evaporation of the solvent, the residue was purifed by silica gel flash column chromatography using DCM/PE/TEA (20 : 20 : 1, v/v/v) as the eluent to give RC as a pale gray solid (2.4 g, 70%). 1 H-NMR (500 MHz, CDCl 3 ) d: 8.81 (t, J ¼ 5.3 Hz, 1H), 8.64 (s, 1H), 7.86 (dd, J ¼ 5.8, 2.7 Hz, 1H), 7.39 (dd, J ¼ 8.7, Hz, 3H), 7.05 (dd, J ¼ 5.7, 2.6 Hz, 1H), 6.62 (dd, J ¼ 8.9, 2.3 Hz, 1H), 6.47 (d, J ¼ 2.1 Hz, 1H), 5.99 (s, 2H), 3.43 (q, J ¼ 7.1 Hz, 2H), 3.36 (dd, J ¼ 12.2, 6.2 Hz, 4H), 3.23 (t, J ¼ 6.5 Hz, 2H), 3.11 (d, J ¼ 6.4 Hz, 2H), 3.06 (dt, J ¼ 16.3, 5.5 Hz, 4H), 2.93-2.84 (m, 4H), 2.60 (t, J ¼ 6.4 Hz, 2H), 2.55-2.40 (m, 6H), 2.02 (td, J ¼ 12.2, 6.3 Hz, 4H), 1.89-1.80 (td, 4H), and 1.22 (t, J ¼ 7.1 Hz, 6H). 13 pH titration of RC-TPP. Aliquots of the stock solution of RC-TPP in DMSO were added to Britton-Robinson buffer (containing 10% DMSO) of various pH values (4.5, 4.71, 4.8, 5.0, 5.19, 5.30, 5.40, 5.50, 5.64, 5.80, 5.91, 6.09, 6.17, 6.22, 6.59, 6.90, 7.50, 8.00, 8.50, and 9.00) to a fnal concentration of 20 mM. The fluorescence emission spectra were recorded as a function of pH using l ex@425 nm for coumarin and l ex@585 nm for rhodamine. Staining of lysosomes and mitochondria in live cells. HeLa and L929 cells were cultured in DMEM spiked with Lysotracker green (2 mM) or Mitotracker green (1 mM) and RC-TPP (1 mM) for 30 min, washed with cold PBS, and then maintained in fresh DMEM for confocal fluorescence microscopic analysis. Cellular distribution of RC-TPP and RC. HeLa cells were cultured in DMEM containing RC-TPP (1 mM) or RC (1 mM) for 30 min, and then analyzed with a confocal fluorescence microscope. Acidity mediated reversal staining of lysosomes. Tom20-GFP + HeLa cells were cultured in DMEM spiked with RC-TPP (3 mM) for 30 min and then further incubated in DMEM spiked with or without BFA (200 nM) for 1 h. The cells were analyzed by confocal fluorescence microscopy. Imaging of CCCP-mediated RC-TPP relocation. Tom20-GFP + HeLa cells were incubated with RC-TPP (2 mM) in the absence or presence of CCCP (0, 10, 100 mM) for 1 h. The cells were analyzed by confocal fluorescence microscopy to pinpoint the locations of RC-TPP and Tom20-GFP. Imaging of lysosome mediated fluorescence-on of RC-TPP. Tom 20-GFP + HeLa cells were cultured in DMEM supplemented with RC-TPP (2 mM) for 30 min and then cultivated in DMEM spiked without or with CCCP (20 mM) in the presence of BFA (200 nM) for 2 h. The cells were analyzed by confocal fluorescence microscopy for GFP and RC-TPP fluorescence. Flow cytometric analysis of CCCP induced activation of rhodamine fluorescence. HeLa cells stably expressing Tom 20-GFP were frst cultivated in DMEM containing RC-TPP (2 mM) for 30 min, and then maintained in HBSS or DMEM spiked with or without CCCP for 4 h. The cells were analyzed by flow cytometry for intracellular rhodamine fluorescence. Cell death related fluorescence activation of RC-TPP. Tom20-GFP + HeLa cells were cultured in DMEM containing RC-TPP (2 mM) for 30 min and then further cultured for 4 h without or with human TNF-a (20 ng ml 1 )/Smac mimetic (100 nM) to initiate apoptosis. At fxed time points, a portion of cells were analyzed by confocal fluorescence microscopy. RIP3 + /Tom 20-GFP + HeLa cells prestained with RC-TPP (2 mM, 30 min) were cultured for 6 h in DMEM containing no addition or human TNF-a (20 ng m 1 )/Smac mimetic (100 nM)/ Z-VAD (20 mM) to initiate necroptosis. The cells were analyzed by confocal fluorescence microscopy. Cytotoxicity of RC-TPP. HeLa cells were cultured with a medium containing RC-TPP (0, 1, 2, 5, 10, or 20 mM) for 30 min and washed with PBS 3 times then incubated with fresh medium for 6, 12, 24, or 48 h. The cell number and cell viability were determined by an MTT assay.

chemsum

{"title": "Responsive hetero-organelle partition conferred fluorogenic sensing of mitochondrial depolarization", "journal": "Royal Society of Chemistry (RSC)"}

parallel_implementation_of_density_functional_theory_methods_in_the_quantum_interaction_computationa

## Abstract: We present the details of a GPU capable exchange correlation (XC) scheme integrated into the open source QUantum Interaction Computational Kernel (QUICK) program. Our implementation features an octree based numerical grid point partitioning scheme, GPU enabled grid pruning and basis/primitive function prescreening and fully GPU capable XC energy and gradient algorithms. Benchmarking against the CPU version demonstrated that the GPU implementation is capable of delivering an impressive performance while retaining excellent accuracy. For small to medium size protein/organic molecular systems, the realized speedups in double precision XC energy and gradient computation on a NVIDIA V100 GPU were 60 to 80-fold and 140 to 780fold respectively as compared to the serial CPU implementation. The acceleration gained in density functional theory calculations from a single V100 GPU significantly exceeds that of a modern CPU with 40 cores running in parallel. ## Introduction Although graphics processing units (GPUs) were originally introduced for rendering computer graphics, they have become essential devices to enhance the performance of scientific applications over the past decade. Examples of GPU accelerated applications span from bioinformatics software that help solving genetic mysteries in biology to data analysis tools aiding gravitational wave detection in astrophysics. 1,2 The availability of powerful general purpose GPUs at a reasonable cost, convenient computing and programming environments such as Compute Unified Device Architecture (CUDA) 3 and especially, the fact that GPUs can perform trillions of floating point operations per second in combination with a high memory bandwidth, outperforming desktop central processing units (CPUs), are the main reasons for this trend. GPUs are also known to deliver outstanding performance in traditional computational chemistry applications, particularly in classical molecular dynamics (MD) simulations and ab initio quantum chemical calculations. In the latter context, Hartree-Fock (HF) 23, 32 and post-HF energy and gradient implementations 18,20,21,27,31 on GPUs have displayed multifold speedups for molecular systems containing a few to large number of atoms. Nevertheless, a majority of the computational chemistry community is unable to enjoy such performance benefits due to the unavailability of an open-source, user friendly, GPU enabled quantum chemistry software. Towards this end, we have been developing a quantum chemical code named the QUantum Interaction Computational Kernel (QUICK) to fill this void. 23,28 As reported previously, QUICK is capable of efficiently computing HF energies and gradients. For instance, the speedup realized for moderate size molecular systems on a Kepler type GPU was about 10-20 times in comparison to a single CPU core while retaining an excellent accuracy. With high angular momentum basis functions (d and f functions), the realized speedup remained 10-18 fold. Such performance gain was primarily due to GPU accelerated electron repulsion integral (ERI) calculations. In our ERI engine, integrals are calculated using vertical and horizontal recurrence relations algorithms 33,34 reported by Obara, Saika and Head-Gordon and Pople. The integrals are calculated on the fly and the Fock matrix is assembled on the GPU using an efficient direct self-consistent field (SCF) scheme. However, the accuracy of the HF method is insufficient, if not totally unsuitable, to study many chemical problems; but having a GPU enabled HF code paves the way towards an efficient post-HF or density functional theory (DFT) package. In the present work, we have undertaken the task of implementing the latter type of methods in QUICK. In fact, given the vast number of research articles published using DFT methods over the past decades, 35 incorporation of such methods into our package was essential. In the context of GPU parallelization of Gaussian based DFT calculations, a few publications have appeared in the literature. 22,36,37 Among these, Yasuda's work 22 is the earliest. His exchange correlation (XC) quadrature scheme consisted of several special features aimed at maximizing performance on GPU hardware. These include partitioning numerical grid points in 3-dimensional space, basis function prescreening and preparation of basis function lists for grid point partitions. Only the electron densities and their gradients and matrix elements of the XC potential were calculated on the GPU. The hardware available at the time limited his algorithm to single precision calculations; but the reported accuracy of benchmark tests was about 10 -5 a.u. A similar algorithm has been implemented by Martinez and coworkers in their GPU capable Terachem software. 36 In addition to grid point partitioning, this algorithm performs prescreening at the level of primitive functions and excludes certain partitions from the calculation based on a sorting procedure. In both of the above implementations, the value of the density functionals at the grid points are calculated on the CPU. We note another DFT package where the XC scheme is GPU enabled, however, the ERI calculations are performed on the CPU. 37 The features of XC quadrature scheme reported in the current work include grid point partitioning using an octree algorithm, prescreening and grid pruning based on the value of primitive Gaussian functions and fully GPU enabled XC energy and gradient calculations where not only the electron density and its derivatives, but also the XC functional values are computed on the GPU. The next sections of this paper are organized as follows. In section 2, we give an overview of the underlying theory of our XC scheme, which was originally documented by Pople and coworkers. 38 The details of the computational implementation are then presented in section 3. Here we first discuss important aspects of data parallel GPU programming. The GPU version of the XC scheme is then implemented following these considerations. Information regarding the parallel CPU implementation using the message passing interface (MPI) 39 is also presented. Section 4 is devoted to benchmark tests and discussion. The tests include performance comparisons between QUICK and the GAMESS GPU version 29,32,40 and accuracy and performance comparisons between the QUICK GPU vs CPU versions. Finally, we conclude our discussion by exploring future directions for further improvement. ## Theory The Kohn-Sham formulation of DFT differs from the HF method by its treatment of the exchange and correlation contributions to the electronic energy. The total electronic energy (𝐸) is given by, 38 ## 𝐸 = 𝐸 where 𝐸 ! and 𝐸 " stand for kinetic and electron-nuclear interaction energies, 𝐸 # is the Coulomb self-interaction of the electron density and 𝐸 $% is the remaining electronic exchange and correlation energies. where na and nb are the number of occupied orbitals. 𝐸 ! , 𝐸 " and 𝐸 # can be written as, 𝐸 $% , within the generalized gradient approximation (GGA), may be given by a functional 𝑓 that depends on electron densities and their gradient invariants, In practical computational implementations, one expresses molecular orbitals as a linear combination of atomic orbitals and 𝜌 & in equation ( 2) becomes, Here 𝜙 0 (𝜇 = 1, . . , 𝑁) are the atomic orbitals. Furthermore, the gradient of 𝜌 & can be written as, A similar expression can be written for 𝜌 ' . Substitution of equations ( 8), ( 9) into ( 2) -( 7) and into the energy equation ( 1) and minimizing with respect to coefficients 𝐶 0( & , 𝐶 The XC contribution to the Kohn-Sham matrix (𝐹 01 67 ! ) is given by, and a similar expression can be written for 𝐹 01 67 " . The gradient with respect to the position of nucleus A is, Where the energy weighted density matrix 𝑊 01 is given by, and the matrix element 𝑋 01 is, Note that for hybrid functionals, the energy in equation (1) will also include a contribution from the HF exchange energy and equations ( 10) and (13) will be adjusted accordingly. Due to the complex form of the XC functionals f, the analytical calculation of the integrals required for the XC energy, XC potential, and its gradients in equations ( 6), (12), and (13), is impossible; hence, this is usually achieved through a numerical procedure. The key steps of such a procedure involve the formation of a numerical grid (also called XC quadrature grid) with quadrature weights assigned to each grid point, calculation of the electron density and the gradients of the density at each grid point, calculation of the value of the density functional and the derivatives of the functional, calculation of the XC energy and the contribution to Kohn-Sham matrix (called matrix elements of the XC potential hereafter). In order to compute the nuclear gradients of the XC energy, one must compute second derivatives of the basis functions, values of the two integral terms in equation ( 13) and add them to the total gradient vector. Finally, due to the involvement of a quadrature weighing scheme in the numerical procedure, the derivatives of the quadrature weights with respect to nuclear displacements must be computed and added to the total gradient. GPUs are ideal devices for data parallel computations, i.e. computations that can be performed on numerous data elements simultaneously. They allow massive parallelization in comparison to traditional CPU platforms, however, at the expense of programming complexity and flexibility. Therefore, a proper understanding of the GPU architecture and the memory hierarchy is essential for writing an efficient code that exploits the full power of this hardware class. GPUs perform tasks according to a single instruction multiple data (SIMD) model, which simply means that they execute the same instructions for a given chunk of data and the same amount of work is expected to be performed for each piece of data. Hence, a programmer should organize the data and assign work to threads that are then processed by the GPU in batches known as thread blocks. The number of threads in a block is up to the programmer, however, there exists a maximum limit allowed by the GPU architecture. For instance, the NVIDIA Volta architecture permits a maximum of 1024 threads per block. NVIDIA GPUs execute threads on streaming multiprocessors (SMs) in warps whose size, 32 for recent architectures, is solely determined by the architecture. Each SM (for example the V100 GPU has 80 SMs, each with 64 CUDA cores for a total of 5120 cores 41 ) executes the same set of instructions for all threads in a warp during a given clock cycle. Therefore, it is essential to minimize the branching in GPU codes (device kernels) to avoid instruction divergence. ## Key considerations in GPU programming A GPU possesses its own physical memory that is distinct from the host (CPU accessible) memory. The main memory called global memory or dynamic random-access memory (DRAM) is relatively large (for example, 32 GB or 16 GB for the V100 and 12 GB for Titan V) and accessible by all threads in streaming multiprocessors. However, global memory transactions suffer from relatively high memory latency. A small secondary type of GPU memory called shared memory is also available on each SM. This type of memory transaction is faster but local, meaning that shared memory of a given SM is only accessible by threads within a thread block that is currently executing on this SM. In addition to these two types of memory, threads in a warp have access to a certain number of registers and this is useful to facilitate the communication between threads of the same warp. GPUs also contain constant and texture memories, which are read-only and capable of delivering a higher performance than global memory in specific applications. If threads in a warp read from the same memory location, constant memory is useful. Texture memory is beneficial when the threads read from physically adjacent memory locations. To maximize the performance of device kernels careful handling of memory is essential. The key considerations for engineering an efficient GPU code include minimizing the warp divergence, minimizing frequent global memory transactions, maintaining a coherent global memory access pattern by adjacent threads in a block (coalesced memory access), minimizing random memory access patterns or simultaneously accessing a certain memory location by multiple threads. Furthermore, constant and texture memories should be employed where applicable. Our existing ERI and direct SCF scheme in QUICK was developed in adherence to this philosophy. As detailed below, we implement our XC scheme following the same practices. ## Grid generation and pruning The selected grid system for our work is the Standard Grid-1 (SG-1) reported by Pople and coworkers. 42 This grid consists of 50 radial grid points and 194 angular grid points for each atom of a molecule. The radial grid point generation is performed using the Euler-Maclaurin scheme and as for the angular grid points, Lebedev grids are used. Following the grid generation, we compute weights for each grid point based on the scheme reported by Frisch et al. 43 We then perform grid pruning in two stages. First, all points with weight less than a threshold (usually 10 -10 ) are eliminated. The remaining points are pruned at a second stage based on the value of atom centered basis/primitive functions at each grid point. As described in section 3.3, we make use of an octree algorithm for this purpose. ## Octree based grid pruning, preparation of basis and contracted function lists In our octree algorithm, the grid points in space are partitioned as follows. First, the lower and upper bounds of the grid is determined and a single cell (node) containing all grid points is formed in 3-dimensional space (see Fig. 1A). This node is then divided into eight child nodes. The child nodes whose grid point count is greater than a user specified threshold are recursively subdivided into octets until the point count of each child node falls below the threshold. The nodes which are not further divisible (leaf nodes, also termed bins below) are considered to be fully partitioned. We then go through each grid point of the leaf nodes and compute the values of the basis and primitive functions at their positions. If a basis/primitive function satisfies the condition for a given grid point, it is considered to be significant for that particular point. We use t = 10 -10 as default threshold, which leads to numerical results that are indistinguishable from the reference without pruning. For basis function-based prescreening, j in equation ( 16) becomes 𝜇 and 𝜁 (@ stands for the value of basis function 𝜙 0 at grid point gi. Similarly, for primitive function based prescreening, 𝜁 (@ represents the value of 𝑐 0A 𝜒 A at grid point gi, where 𝜒 A is the p th primitive function of 𝜙 0 and 𝑐 0A is the corresponding contraction coefficient. Once the basis function values at each grid point are computed, the points that do not have at least one significant basis function are omitted from the corresponding bin (see Fig. 1B). Furthermore, bins without any grid points are also discarded. At this stage, lists of significant basis and primitive function IDs are prepared for each remaining bin of grid points, significantly reducing the number of basis function values and derivatives that have to be evaluated at each grid point during the SCF and gradient computations. Using this algorithm, the number of primitive Gaussian basis function evaluations is reduced from 117400 to 6484 for H2O with a cc-pVDZ basis set (Fig. 1B). ## Grid point packing and kernel launch Following the two-stage grid pruning and the preparation of basis and primitive function ID lists, the grid points are prepared (hereafter grid point packing) to be uploaded to the GPU. Here we add dummy points into each bin and set the total grid point count in each bin to a threshold value used in the octree run. It is worth noting that the threshold we choose is a multiple of the warp size 32, usually 256. By doing so, we are able to pack true grid points into one corner of a block of an array (see Fig. S1A) and this helps us to minimize the warp divergence when we launch GPU threads using the same value for the block size. More specifically, for subsequent calculations (i.e. density, XC energy and XC gradients), we will launch a total of 256*nbin threads (where nbin = number of significant bins) where each thread block contains at least one true grid point and a set of dummy points. The threads launched for dummy points should not perform any computation and these are differentiated from true points by an as-signed integer flag. As mentioned previously, the launched thread blocks are submitted to streaming multiprocessors as warps of 32 by the warp scheduler during runtime. This ensures that at most one warp per thread block suffers from warp divergence and threads of the remaining warps would perform the same task. Once the grid points are packed and basis and primitive function lists have been prepared, these are uploaded to the global memory of the GPU. The corresponding array pointers are stored in constant memory and these are used in launching the kernel and the execution of device kernels during density, SCF and gradient calculations. ## Fig. 2. Flowchart depicting the workflow of a DFT geometry optimization calculation. Brown, blue and mixed color boxes indicate steps performed on CPU, GPU and mixed CPU/GPU respectively. OPT stands for geometry optimization. ## Computing electron densities on the GPU As apparent from eq. 8 and 9, the calculation of electron densities and their gradients require looping over basis and primitive functions at each grid point. This task is performed on the GPU and each thread working on a single grid point only loops through basis and primitive function lists assigned to the corresponding thread block. The retrieval of correct lists from global memory is achieved by using two locator arrays (see Fig. S1B). Note that we use the term "ID" when referring to basis/primitive functions and "array index" for array locations. The first, called basis function locator array, holds the array index ranges of the basis function ID array. Each thread accesses the former based on their block index, obtains the corresponding array index range of the latter and picks the basis function IDs. Second, the primitive function locator array, holds the array index ranges for accessing the primitive func-tion ID array. Each thread picks elements from the primitive function locator array using basis function array indices, obtains the relevant array index range of the latter and takes the primitive function IDs. This retrieval strategy allows us to maintain a coalesced memory access pattern. ## Computing the XC energy and the matrix elements of the XC potential As reported elsewhere, 23,28 our existing SCF implementation assembles the Fock matrix and computes the energy on the GPU. Therefore, our goal is to calculate the XC energy and associated derivatives on the GPU, which highlights the necessity to have density functionals implemented as device kernels. Needless to say coding the numerous available functionals is a cumbersome process and the best solution is to make use of an existing density functional library. Neverthe- In order to do so, we made several minor modifications to the interface and intermediate files. The obtained information is then uploaded to the GPU. We also implemented device kernels for LDA and GGA workers to call functionals during device kernel execution. The corresponding functional source files were also modified with compiler preprocessor directives and during the compilation time, these are included in the CUDA source files and compiled as device kernels. As reported previously, use of compiler directives for porting complex scientific applications to GPUs is rewarding in terms of application performance and implementation effort. 47 Note that our current XC implementation is not capable of computing kinetic energy densities and for this reason, we have not made any changes to the MGGA workers and related source files in LIBXC. During the SCF procedure, the device kernel versions of LIBXC workers are called with pre-stored functional IDs and other parameters. The worker then calls the appropriate functional kernel. Here we have used C function pointers rather than conditional statements due to potential performance penalties introduced by the latter. Following the computation of XC energy density on grid points, we compute the matrix elements of the potential and update the Kohn-Sham matrix using the CUDA atomic add function. In the past, we employed atomic add in our direct SCF scheme and noted that it is not a critical performance bottleneck. ## Computing XC energy nuclear gradients Most of the implementation strategy described above for XC energy and potential holds for our gradient algorithm. More specifically, this is a two-step process consisting of calculating the XC energy gradients and grid weight gradients. While computing the former involves a majority of steps discussed for the energy calculation (i.e. computing values of the basis functions and derivatives and functional values and associated derivatives), additionally, we compute second derivatives of basis functions. The device kernel performing this task is similar to the one that calculates basis function values and their gradients at a given grid point in the sense of accessing basis and primitive function IDs. The grid weight gradient calculation is only required for grid points whose weight is not equal to unity. Therefore, we prune our grid for the third time removing all points that do not meet this criterion. The resulting grid points are packed without dummy points, uploaded to the GPU and the appropriate device kernel is called to compute the gradients. The calculated gradient contribution from each thread is added to the total gradient vector using the CUDA atomic add function, consistent with our existing ERI gradient scheme. a GPU energy column shows the deviation of the energy with respect to the corresponding CPU calculation. ## CPU analog of the GPU implementation In order to perform a fair comparison with our GPU capable DFT implementation, we implemented a parallel CPU version using MPI. In this version, we make use of the standard LIBXC code base and omit the LIBXC dry run from our workflow (step 2 in Fig. 2). Furthermore, the grid generation and octree run are performed in serial but the weight computation, prescreening and preparation of basis and primitive func-tions lists are all performed on in parallel. Moreover, the packing of grid points (step 3g in Fig. 2) is omitted. Computation of electron densities, XC energy, potential and gradients (steps 4-6 in Fig. 2) are performed in parallel. More specifically, once the grid operations are completed, partitioned bins are distributed among slave CPU tasks. The corresponding grid weights, basis and primitive function lists are also broadcast. All CPU tasks retrieve basis and primitive function IDs using locators as discussed in section 3.5, but with the block index now replaced by the bin index. When computing the XC energy and matrix elements of the potential, each CPU task initializes LIBXC through a standard Fortran 90 interface and computes the required functional values. The slave CPU tasks then send the computed energy and Kohn-Sham matrix contributions to the master CPU task. The implementation of the XC gradient scheme is very similar to the above. ## Benchmarking the GPU implementation We now present the benchmarking results of our DFT implementation. First, the performance of the QUICK GPU version is compared against the GPU version of GAMESS. 29,32,40 Then a similar comparison between the QUICK CPU and GPU versions is performed. For the former test, we obtained a precompiled copy of GAMESS (the 17.09-r2-libcchem version) in a singularity image from the NVIDIA GPU cloud webpage. In order to make a fair comparison, the QUICK code was compiled using the comparable GNU and CUDA compilers with optimization level 2 (-O2). Note that performance tradeoff between running a CPU/GPU application native or through a container is reported to be negligible 48,49 and we assume that the container overhead has no significant impact on our GAMESS timings. The selected test cases include morphine, (glycine)12 and valinomycin BLYP gradient calculations using the 6-31G, 6-31G* and 6-31G** basis sets. In GAMESS input files the SG-1 grid system and direct SCF were requested and the density matrix cutoff value was set to 10 -8 . Default values were used for all the other options. All tests were carried out on a NVIDIA Volta V100-SXM2 GPU (32 GB) accompanied by Intel Xeon (R) Gold 6138 CPU (2.10 GHz) with 190 GB memory. The performance comparison between QUICK and GAMESS suggests that the former is significantly faster than the latter (see Table 1). For ERI gradients, the observed speedup is roughly 2-5 fold suggesting that our ERI engine is more efficient, in particular for basis sets with higher angular momentum quantum numbers. Furthermore, QUICK displays higher speedups (~60 or 80-fold in some cases) for the XC gradients, however, this result has to be interpreted carefully since the XC implementation of GAMESS appears to be not GPU capable. Note that in larger test cases, the GAMESS XC gradient time surpasses the ERI gradient time. Based on the times reported for last SCF iteration, the QUICK SCF scheme also outperforms GAMESS. Again, slow performance of the latter may be partially attributed to its CPU based XC implementation. A comparison of separate timings for ERI, XC energy and potential calculations during each SCF iteration was not possible since timings for these individual tasks were not reported by GAMESS. Nevertheless, HF single point calculations carried out for the same systems (see Table S1) show that QUICK ERI calculations are much faster, consistent with the gradient results documented above. We now compare the accuracy and performance between the QUICK CPU and GPU versions using a series of B3LYP 53 energy calculations. As is apparent from Table 2, energies computed by the QUICK GPU version agree with the CPU version up to 10 -6 au or better. Furthermore, the GPU version of the XC implementation delivers at least a 60-fold speedup over the serial CPU version. In both versions, less than 30% of the SCF step time is spent on computing the XC energy and contribution to Kohn-Sham matrix. This suggests that ERI calculation remains the performance bottleneck of QUICK DFT energy calculations. As a result, the speedup observed for the SCF iteration in most cases is somewhat lower than the XC speedup. In Table 3, we report a comparison of B3LYP/6-31G and B3LYP/6-31G** gradient calculation times between the QUICK CPU and GPU versions. The selected test cases are a subset of the molecular systems chosen for the SCF tests. The key observations from Table 3 include: 1) the GPU version delivers significant speedups for both ERI and XC gradient calculations, 2) the speedup observed for ERI gradients resemble the ones realized for energy calculation (~50 times), 3) the speedup delivered for XC gradients is in few hundreds (~100-800 times). Similar speedups observed for ERI energy and gradient calculations can be explained by the fact that their implementations are similar. Indeed, computing the gradients of a given basis function involves the cost of computing the value of higher angular momentum basis function. Understanding the details of the observed speedups for XC energy and gradient computations requires further investigation. As mentioned in section 3.7, we implemented XC energy gradient and grid weight gradient calculations in two separate kernels. Careful examination of kernel run times suggests that the former consumes more than 90 % of the gradient time and gains a better speedup on a GPU. However, the structure of this kernel is similar to that of the XC energy with the exception of computing second derivatives. Therefore, the observed higher speedup must be due to the inefficiency of computing the second derivatives in our CPU implementation. ## Performance of LIBXC functionals As mentioned above, we have integrated the original and modified LIBXC library versions into QUICK CPU and GPU codes. It is important to document the performance of these functionals within our XC scheme. In Table 4, we report a series of taxol energy calculations at the DFT/6-31G level of theory using various functionals. The selected functionals include hand-coded (native) BLYP, B3LYP and their LIBXC versions and additionally, representative LDA, GGA and hybrid-GGA functionals. Comparison of native BLYP and B3LYP total energies against the corresponding LIBXC versions suggests that the latter are as accurate as the former on both CPU and GPU platforms. Furthermore, the reported CPU and GPU XC times remain very similar and the acceleration delivered by the GPU remains substantially the same. In fact, the ca. 20-fold speedup realized on the GPU platform is common for all LDA and GGA functionals. Note that computing the value of density functionals is relatively inexpensive with respect to other operations performed in XC energy or gradient calculations and such a similar speedup is expected. Finally, the fact that we maintain a similar speedup among native and LIBXC functionals suggests that minor modifications performed to density functional source code in the GPU version has not introduced performance penalties. ## Comparison of GPU vs parallel CPU implementations We now compare the performance between GPU and parallel CPU (MPI) implementations using taxol gradient calculations at the B3LYP/6-31G** level of theory. The selected platform for testing is a single computer node equipped with a NVIDIA V100-SXM2 GPU ( In Fig. 3B, we report the total time spent by the ERI and XC calculations during 19 SCF iterations. The maximum speedup observed for the former and latter in the MPI version are ca. 23 and 33-fold respectively. The speedups from the GPU version for the same tasks are ca. 65 and 62-fold. For ERI gradient calculations (see Fig. 3C), the speedup gained from the MPI version is similar to that of the SCF (ca. 22-fold) but significantly less than the corresponding GPU speedup (ca. 50fold). Furthermore, the MPI version delivers about 59-fold speedup for the XC gradient calculations, however, this is below the acceleration achieved by a V100 GPU (ca. 142fold). Finally, as evident from Fig. 3D, the total speedup gained from the GPU is two times over using 40 cores. It is also important to comment on the linearity of our MPI plots in Fig. 3B and 3C. Both ERI SCF and gradient calculations scale well with the number of cores and almost achieve the ideal MPI speedup since we distribute atomic shells among MPI tasks (CPU cores). However, in the XC implementation, we distribute bins containing varying number of grid points. Therefore, the workload is not optimally balanced and plots of the XC MPI timings display a non-regular speedup (not linear) unlike in the ERI case. ## Performance of QUICK on different GPU architectures For all the aforementioned benchmarks, we have used a NVIDIA V100 data center GPU. It is also necessary to document how the current QUICK GPU implementation performs on other available devices. In Table 5, we report the performance of several important kernels on different workstation GPUs (V100, Titan V and P100) and a gaming GPU (RTX2080Ti). The selected test case for this purpose is the taxol gradient calculation at the B3LYP/6-31G level of theory. As is apparent from Table 5, all kernels display their best performance on the V100 GPU. Surprisingly, ERI kernels show their slowest times on the P100 rather than the gaming GPU, suggesting that their performance is not limited by double precision (FP64) operations. Note that the P100 GPU has a higher FP64 capability in comparison to the RTX2080Ti. 60,61 A detailed examination of the two kernels revealed that their performance is bound by high register usage and memory bandwidth. In contrast, XC kernels displayed their highest timings on the RTX2080Ti GPU and therefore, these are bound by FP64 operations. Further examination of the XC energy gradient kernel indicated that its performance is also limited by atomic operations. Overall, excellent performance is achieved both on data center GPUs with Pascal and Volta architectures and gaming GPUs with Turing architecture. a Reported time is the summation over 22 SCF cycles. ## Conclusions We have reported the details of the MPI parallel CPU and the GPU enabled DFT implementation of the QUICK quantum chemical package. Our implementation consists of features such as octree based grid point partitioning, GPU assisted grid pruning and basis/primitive function prescreening and fully GPU enabled XC energy and gradient computations. Performance comparison with the GAMESS GPU version demonstrates that DFT calculations with QUICK are significantly faster. The accelerations observed for the XC energy and gradient computation in the QUICK GPU version with respect to the serial CPU version are impressive. The speedups realized on a V100 GPU for the former and latter are approximately 60 to 80-fold and 140 to 780-fold respectively. Such speedups are out of reach with the MPI parallel CPU version even if one uses 40 cores in parallel. The recommended device for the latest QUICK version (v20.03) is the NVIDIA V100 data center GPU but the code runs very well also on gaming GPUs. The profiling of ERI and XC kernels has shown that there exists room for further performance improvement. In the former context, reimplementing large ERI kernels into smaller kernels may be a viable strategy. This is expected to reduce the register pressure and enhance the performance of the ERI engine. The performance of XC kernels on gaming GPUs may be improved by implementing a mixed precision scheme. Additionally, strategies such as storing the gradient vector in shared memory and latency hiding may be helpful to reduce the computational cost associated with atomic operations in the XC energy gradient kernel. Currently, we are integrating QUICK as a library into the AMBER molecular dynamics package 62 to enable fully GPU enabled quantum mechanics/ molecular mechanics (QM/MM) simulations. Finally, QUICK version 20.03 can be freely downloaded from http://www.merzgroup.org/quick.html under the Mozilla public license. ## ASSOCIATED CONTENT Supporting text and Cartesian coordinates of the test molecules. This material is available free of charge via the Internet at http://pubs.acs.org.

chemsum

{"title": "Parallel Implementation of Density Functional Theory Methods in the Quantum Interaction Computational Kernel Program", "journal": "ChemRxiv"}

transparent,_flexible_supercapacitors_from_nano-engineered_carbon_films

## Abstract: Here we construct mechanically flexible and optically transparent thin film solid state supercapacitors by assembling nano-engineered carbon electrodes, prepared in porous templates, with morphology of interconnected arrays of complex shapes and porosity. The highly textured graphitic films act as electrode and current collector and integrated with solid polymer electrolyte, function as thin film supercapacitors. The nanostructured electrode morphology and the conformal electrolyte packaging provide enough energy and power density for the devices in addition to excellent mechanical flexibility and optical transparency, making it a unique design in various power delivery applications. ## R ecently there has been significant interest in using carbon based nanomaterials as supercapacitor electrodes due to several advantages of carbon such as light weight, high electrical conductivity and electrochemical surface area . There is a large number of interesting possibilities in creating new designs for such energy storage devices if the carbon electrodes can be tailored and engineered to fit new functionalities. Here, we demonstrate the design and fabrication of flexible and transparent supercapacitors using a highly structured carbon thin film, structured inside porous templates by chemical vapor deposition. These carbon films consist of arrays of periodic and interconnected nano-cup morphologies of few layer graphitic structure (from now, termed as carbon nanocups -CNC) and are used as thin film electrodes for the supercapacitor devices. CNCs are architectures precisely engineered from graphitic carbon, within porous templates, having up to 10 5 times smaller length/diameter (L/D) ratios compared to conventional nanotubes, and have unique nanoscale cup morphology. Our CNC filmspolymer electrolyte composites have three remarkable features for the use as a solid state, thin-film supercapacitor device. First, a CNC film has the high surface area offered by arrays of controlled nanoscale cup structures and highly disordered graphitic layers that are keys for the effective permeation of the polymer electrolyte in supercapacitors. Second, unique nanoscale structural and morphological features of CNC films enable the easy access and faster transport of ions at the electrode/electrolyte interface resulting in higher power capability. Finally, high current carrying capability, substantial mechanical strength, and small effective electrode thickness (10 nm) allow us to build multifunctional (optically transparent and mechanically flexible) reliable thin-film energy storage devices. ## Results Fabrications of nano-engineered carbon films. Fig. 1a and 1b-c shows the scanning electron microscopy (SEM) images of top (concave) and bottom (convex) parts of as-synthesized CNC films (with channels of 80610 nm in diameter and 140610 nm in length). By connecting the highly dense and ordered arrays of nanocups with continuous graphitic layer that, large area porous nanostructured films ideal for energy storage electrodes is achieved. To further increase the surface area of CNC films, multi-branched nanocup architecture, with large number of short nanotubes (25 nm in diameter and 330610 nm in length) grown and attached to the bottom of the CNC films, are fabricated by using multistep anodization process followed by a chemical vapor deposition (CVD) method 21 (Fig. 1d-f). Through the control of nanopore dimension in anodic aluminum oxide (AAO) templates and CVD conditions, we can precisely tailor the geometry and structure of the nanocup such as length, diameter, L/D aspect ratio, and their wall thickness, which are important factors to determine capacitor behavior, mechanical stability, and optical property of devices. Our calculation reveals that a branched convex CNC films possess 2.3 times higher surface area exposed to the electrolyte than that of a normal convex CNC film (see the method section for the surface area calculation). The length of branched carbon nanotubes is optimized to obtain optical transparency as well as maximized electrochemically active interface between electrodes and polymer electrolyte. Also it is noted that the innermost layer of the concave nanocup (Fig. 1a) has relatively well-ordered graphitic structure, while the outermost layer which is the surface of the convex part (Fig. 1b-f), is very defective 22 . When used as a supercapacitor electrode, these defects can act as reactive sites providing an effective charge transfer and good electrode-electrolyte interfaces that minimize the internal resistance of supercapacitor devices. Measured surface electrical conductivity of the CNC film is 117 S/m, which is higher than typical activated carbon electrodes (,30 S/m) 23 for supercapacitors. Fabrications of transparent and flexible supercapacitor devices. We have fabricated a flexible and transparent thin-film supercapacitor device by impregnating highly porous CNC electrodes (anode and cathode) in transparent polymer electrolyte films. Fig. 2 shows schematics of the fabrication process of branched CNC-polymer electrolyte thin films and optical images of supercapacitor devices. As shown in the schematics of Fig. 2a, the nanocup films are transferred to polydimethylsiloxane (PDMS) and released by dissolving AAO templates in copper chloride and hydrochloric acid mixture solution to produce transparent and flexible graphitic carbon electrodes. The CNC films are then utilized as a dual function layers in supercapacitor devices where an inner graphitic layer exposed to the electrolyte acts as active electrodes and outer graphitic layer works as current collectors. The fact that ultra-thin and organized arrays of nanostructured graphitic film can be used for both electrodes and current collectors enables us to design and create mechanically flexible and optically transparent supercapacitor film in a scalable and simple manner. For the ionic electrolyte/separator, the polyvinyl alcohol-phosphoric acid (PVA-H 3 PO 4 ) gel electrolyte is then sandwiched between two separated CNC electrode films. For this, PVA-H 3 PO 4 polymer solution is poured over the CNC film and spin-coated with 500 rpm to obtain the effective electrolyte thickness (12 mm). CNC films are transparent with transmittance of 71% at 550 nm wavelength (supplemental Fig. S1) and so the fabricated solid state thin film CNC supercapacitor devices are optically transparent (Fig. 2b) and mechanically flexible (Fig. 2c). Characterizations of the supercapacitors. Cyclic voltammetry (CV) is performed to evaluate the capacitance of all three different types of CNC (concave, convex, and branched convex) electrode based solid state supercapacitors. All CV curves show a very rapid current response on the voltage reversal at each end potential and straight rectangular shapes representing a very small equivalent series resistance of electrodes and faster ionic diffusion in the electrolyte film. The CV curves of CNC devices are measured with various scan rates in the ranges of 10 -500 mVs 21 . Especially when a branched CNC electrode (branched convex) is used (Fig. 3a), nearly rectangular shaped CV curves are obtained even at very high scan rates demonstrating high performance capacitor devices (for concave and convex CNC devices, see supplemental Fig. S2a and S2d). Galvanostatic charge/discharge (CD) is also conducted to evaluate the normalized capacitance and internal resistance of the branched convex CNC supercapacitor device (for concave and convex CNC devices, see supplemental Fig. S2b and S2e). The E-t responses of the charge process show a triangular shape and mirror image with corresponding discharge counterparts confirming the formation of an efficient capacitor and excellent charge propagation across two electrodes. The capacitances by the geometrical area calculated from CD curves are 78, 132, 409 mF cm 22 for concave, convex, and branched CNC supercapacitors, respectively (see method section for the capacitance calculation). Note that we have chosen to use the areal capacitances, and not the gravimetric capacitance, due to the ambiguities in mass determination of CNC films. Though the specific surface area of the convex CNC is 1.25 times higher than that of the concave CNC, the convex CNC device shows 1.7 times higher capacitance than a concave CNC device because of its more open surface morphology and electrochemically active sites formed on the surface of AAO templates 22 as described above. For the branched convex type supercapacitor device, where electrochemically active surface is maximized, measured specific capacity is 3-5 times higher than regular CNC capacitor devices and 6 times higher than what has been reported for a single-layered graphene device 18 . Also the volumetric capacitance of the branched convex type supercapacitor (0.33 F/ cm 23 ) is only 1.36 times smaller while providing high transparency than what has been reported for laser-scribed graphene electrochemical capacitors using the same electrolyte 20 . From the voltage versus time profile, we are able to calculate the coulombic efficiency, g using a ratio of the times for galvanostatic discharging and charging. An ideal capacitor gives 100% efficiency and has mirror inverse V shape from the galvanostatic CD curve. The columbic efficiency for the CNC supercapacitor is 86%. The temperature effects on capacitance and charge-discharge behaviors in CNC devices are also explored. As shown in Fig. 3c, about three times higher capacitance (1220 mF cm 22 ) is observed at 80uC compared to the capacitance measured at room temperature. The increase in capacitance may be partly due to the molecular alignment of PVA-H 3 PO 4 chains and the excitation of charge carriers present on the imperfect sites of the CNC electrode surface with a moderate temperature increase 16 . Ragone chart of the supercapacitors. To evaluate energy storage performance of the branched CNC device, energy density is plotted versus power density (Ragone chart) as shown in Fig. 3d . Using the internal resistance values and capacitances, energy and peak power densities for CNC supercapacitors are calculated (see the methods section). The results are compared with different thin film energy storage devices designed for flexible electronic applications. The volumetric (this includes two CNC films and the polymer electrolyte) peak power and energy densities of the branched CNCbased supercapacitors are 19 mW/cm 3 and 47 mWh/cm 3 respectively. This value is similar to that of the single layer graphene-based solid state supercapacitor 18 and the laser-scribed graphene electrochemical capacitors 20 using the same electrolyte in energy density while offering high mechanical flexibility and optical transparency. In addition, the branched CNC-based supercapacitor exhibits an increase of energy densities of up to 3.1 times higher at 80uC and has similar value to the reduced multilayer graphene oxide 18 and the hydrated graphitic oxide 19 in energy density. This remarkable thin film capacitor behavior can be attributed to the well-textured nanoscale features on the electrode, the significantly increased surface area in the more complex branched CNC films, the excellent conformal filling of polymer electrolyte, and the maximized active electrochemical surface area, respectively. We also speculate that the all connected graphitic structures of CNC films facilitate the charge transfer during charge/ discharge processes leading to the higher power density. 18 , RMGO: reduced multilayer graphene oxide 18 , HGO: hydrated graphitic oxide 19 , LSG-EC: laser-scribed graphene electrochemical capacitor 20 ). Performance of the supercapacitors under mechanical deformations and their long cycle life. In order to evaluate the potential of CNC-based supercapacitors for the use as flexible energy storage, a device was placed under various mechanical deformations and its performance was analyzed. The capacitance and other electrochemical properties changes according to the number of helical wrapping of CNC capacitor film (3.5 cm by 0.5 cm) around glass tube (0.5 cm diameter) are shown in Fig. 4a and Fig. S3. CNC capacitor devices retain their superior capacitance, CV and CD properties even after helically rolled up to 720u. This excellent mechanical flexibility and integrity could be due to the effective conformal filling of polymer electrolyte into organized graphitic nanostructured film. We also measured the long cycle life (number of charge-discharge cycles at constant current) of CNC supercapacitors w/wo mechanical deformation. The normalized capacitance as a function of cyclenumber is shown in the Fig. 4b. The supercapacitor devices even under mechanical stress (45u bending) show long life cycle stability: . 84% of the initial capacitance after 10,000 cycles, indicating that the performance is not limited by parasitic chemical reactions and mechanical breakdown due to swelling of the electrode or mechanical strain during charging-discharging. ## Discussion Finally the performance of our supercapacitor as flexible and transparent devices is demonstrated in Fig. 4c and 4d. For this, a prototype of a large area supercapacitor film (3 cm by 1.5 cm) was fabricated and a light-emitting diode (LED, the working potential is 1.5V) was successfully turned on for 20 min after being charged for 15 min at 2.5 V. We show that such devices could also be integrated into unique applications where power delivery along with transparency or mechanical flexibility could be advantageous, for example as thin coatings on windows, screens, structures with different geometries etc. As a demonstration we placed our transparent supercapacitor on a smart phone screen, and during operation of the device one could clearly see the concurrent transparent nature of the devices (Fig. 4c and supplemental movie S1). In another demonstration, the devices were helically wrapped around differently sized glass tubes in high curvature yet showing excellent performance during operation (Fig. 4d and supplemental movie S2 and S3). In summary, we demonstrated the design of transparent, flexible thin film supercapacitor devices with high performance in energy delivery. The devices were assembled using complex nano-engineered thin graphitic film electrodes with significantly increased surface area and three-dimensional structure. Unique morphological and structural features of the films enable an excellent conformal filling of polymer electrolyte and maximize active electrochemical surface area leading to high energy density. The design of the devices allows for mechanically flexible energy storage devices that could be integrated into unique applications that require high form factor and optical transparency, for example rollup displays, wearable device, and organic solar cell platforms. ## Methods Fabrications of CNC films. A CNC film with an interconnected thin graphitic texture, high specific surface area, and uniform nanopore is obtained using highly engineered AAO nanochannels as templates. Nanoporous alumina template is prepared using a standard electrochemical anodization process 26 . Three differently engineered CNCs, which have concave, convex, and branched convex shapes, are made for the supercapacitor electrodes. For concave and convex (b) Normalized capacitance as a function of cycle-number (10,000) and w/wo the mechanical deformation (45u bending). (c, d) Optical pictures demonstrating optical transparency and mechanical flexibility of a large scale CNC supercapacitor films. Note that a LED is turned on and remain stable even after helically wrapped into the differently sized glass tubes. The large area CNC supercapacitor film (3 cm by 1.5 cm) over the smartphone screen demonstrates optical transparency and the CNC capacitor films (3 cm by 1.5 cm and 3.5 cm by 0.5 cm) helically wrapped around two differently sized glass tubes (1.6 cm and 0.5 cm diameter) show excellent performance during power delivery. nanocups, a two-step anodization process is performed at 45 V in 3% oxalic acid (C 2 H 4 O 2 ) solution for 20 seconds to fabricate short nanochannels. Then, AAO templates are soaked in a 5% phosphoric acid solution for 1 hour at room temperature to widen nanopores. In order to produce branched convex nanocups, third anodization is carried out for 5 minutes at 25 V in 3% oxalic acid solution. Low aspect-ratio CNCs are synthesized by using a chemical vapor deposition (CVD) process at 630uC using 10% acetylene gas as a carbon source. Then convex type nanocups are transferred to PDMS and released by dissolving AAO templates in 3.4 g copper chloride and 100 mL hydrochloric acid mixture solution with 100 mL deionized water to produce transparent and flexible graphitic carbon electrodes. For concave type nanocup devices, first the gel electrolyte is placed between two CNC with AAO templates followed by dissolving the templates. Then both sides of the film are coated by PDMS in order to facilitate the handling. Specific surface area and mass of CNCs. The calculations are based on the following hypotheses. (1) Surface area of cap of a nanocup is consistent with that of mouth of a nanocup pore and curvature of the cap is ignored, and thus both the hexagonal cell and cylinder area in a nanocup geometry are involved into the calculation of the mass, (2) only the surface of carbon layer exposed to the electrolyte is taken into account for specific surface area, (3) the length of the C-C bonds in the curved graphene sheets is the same as in the planar sheet (d c-c 5 0.1421 nm), (4) the nanocups are composed of concentric shells and the inter-shell distance is d s-s 5 0.34 nm, (5) the total carbon layers are about 30 walls in 10 nm thickness, (6) the inter-pore distance is about 105 nm and the total nanocups in 131 cm 2 geometrical area are 1310 10 , (7) the specific surface area of one side of a graphene sheet is 1315 m 2 /g 27 . The surface of a carbon layer exposed to the electrolyte is: S e ~pLd e zS hc The surface of the entire graphene sheet in one nanocup geometry is: The total mass of one nanocup geometry is: W c ~Sc =1315 and thus the specific surface area of one nanocup geometry is: SSA CNC ~Se =W c where L is length of nanocup, d is diameter, d e means inner diameter for concave type and outer diameter for convex type nanocup, S hc is surface of the hexagonal cell, and n is the number of graphene shells. The calculated SSA is 39, 49 and 63 m 2 /g for nanocups of the concave, convex and branched convex type, respectively. The fact that each shell addition does not produce a strong increase of the surface area of the CNCs is due to the much larger increase of its mass. However, the surface of a carbon layer exposed to the electrolyte for branched CNCs is 2.9 times and 2.3 times larger than that for concave and convex CNCs. Fabrications of CNC-based supercapacitors. The gel electrolyte is prepared by mixing PVA powder with water (1 g of PVA/ 10 mL of H 2 O) and concentrated H 3 PO 4 (0.8 g). PVA acts as a host for ionic conduction. The ion source comes from H 3 PO 4 acid as proton donor materials. The gel electrolyte is placed between two CNC electrodes. Upon evaporation of water, the electrolyte solidifies. The polymer electrolyte shows very stable performance in the operating range of 1 V in the CNCbased supercapacitors since the voltage range of water used is closer to an aqueous electrolyte. Characterizations of transparent and flexible supercapacitor devices. Electrochemical properties of CNCs-based supercapacitors are analyzed using cyclic voltammetry (CV), galvanostatic charge-discharge (CD) and cyclic stability. The CV curves of CNC-based devices are measured between 0 and 1 V with various scan rates in the range of 10-500 mVs 21 . CV curves display nearly rectangular shape even at very high scan rates. The CD curves are obtained at a constant current density of 4.2 mAcm 22 for the concave CNCs and 5 mAcm 22 for convex CNCs and the branched convex CNCs. The capacitance and internal resistance values are determined from the slope and the initial voltage drop of the galvanostatic CD curves, respectively. The capacitances C are calculated from the galvanostatic discharge curves using C 5 i/ 2[DV/Dt]A 5 i/2slope 3 A, where C is the capacitance, i is the discharge current, the slope is the slope of the discharge curve after the iR drop, and A is the geometrical area of CNCs on the electrode. The cyclic stability is obtained by performing chargedischarge of the flat and bent CNC supercapacitor over 10,000 cycles. It is apparent that the materials retain good stability over large number of charging-discharging cycles. The efficiency (g), the power density (P) and energy density (E) of the CNCbased supercapacitors are calculated using g 5 (t discharging /t charging ) 3 100, P 5 V 2 / [4RV vol ] and E50.5CV 2 /V vol , respectively where t discharging is the discharging time, t charging is the charging time, V vol is the volume including area and thickness of two CNC films and a polymer electrolyte, C is the measured device capacitance, and R is the internal resistance, respectively.

chemsum

{"title": "Transparent, flexible supercapacitors from nano-engineered carbon films", "journal": "Scientific Reports - Nature"}

volatiles_from_the_tropical_ascomycete_<i>daldinia_clavata</i>_(hypoxylaceae,_xylariales)

## Abstract: The volatiles from the fungus Daldinia clavata were collected by use of a closed-loop stripping apparatus and analysed by GC-MS. A few compounds were readily identified by comparison of measured to library mass spectra and of retention indices to published data, while for other compounds a synthesis of references was required. For one of the main compounds, 5-hydroxy-4,6-dimethyloctan-3-one, the relative and absolute configuration was determined by synthesis of all eight stereoisomers and gas chromatographic analysis using a homochiral stationary phase. Another identified new natural product is 6-nonyl-2H-pyran-2-one. The antimicrobial and cytotoxic effects of the synthetic volatiles are also reported. ## Introduction A large variety of volatile organic compounds from different compound classes including fatty acid derivatives and polyketides, aromatic compounds, terpenes, sulfur and nitrogen compounds, and halogenated compounds is produced by ascomycete fungi . Possibly the most widespread volatile secondary metabolite from fungi is (R)-oct-1-en-3-ol (1, Scheme 1), a compound that was first isolated from Tricholoma matsutake and named "matsutake alcohol" . This odourous volatile is responsible for the typical mushroom smell of many fungi and also contributes to the pleasant aroma of edible mush-rooms such as the button mushroom, Agaricus bisporus . Another widespread fungal volatile is 6-pentyl-2H-pyran-2-one (2) that was first isolated from Trichoderma and exhibits a strong coconut aroma . For fungi producing 2 a plant-growth promoting effect and an induction of systemic resistance in plants has been observed which makes these fungi interesting as biocontrol agents . On the contrary, fungi can also produce mycotoxins, which must be excluded for their safe usage in agricultural biocontrol. Some volatiles, especially terpenes, point to the production of certain toxins in fungi, e.g., aris-Scheme 1: A selection of widespread fungal volatiles. tolochene (3) is the precursor of PR toxin in Penicillium roqueforti , trichodiene (4) is the parent hydrocarbon of the trichothecene family of mycotoxins in various Trichothecium and Fusarium strains , and the diterpene ent-kaurene (5) is the precursor of gibberellins, a class of plant hormones that are produced in large amounts by the rice pathogen Fusarium fujikuroi and other fusaria . The potential beneficial bioactivity and role in the intra-or interspecies communication as well as the possible function as markers for toxin production recently resulted in an increasing interest in volatile secondary metabolites in the scientific community. The Xylariales (class Sordariomycetes) is one of the largest orders of Ascomycota and comprises several thousands of microscopic fungi, as well as numerous "macromycetes" that may produce conspicuous fruiting bodies (stromata) . The Xylariaceae remain the largest family of this order, even though it was recently further divided, based on a multi-gene phylogeny that widely agreed with important chemotaxonomical and morphological traits . A comprehensive overview of the current taxonomy of these families has been published by Daranagama et al. . However, still only little is known about volatile secondary metabolites from Xylariales. Most respective studies have been dedicated to some endophytic strains that can be assigned to the Xylariales based on preliminary molecular phylogenetic data and are being referred to the suggested genus Muscodor. However, this genus was recently rejected, because its erection did not follow good taxonomic standards . The only comparative study available on the production of volatiles that used taxonomically well-characterised Xylariales relied on a panel of strains of the genera Daldinia and Hypoxylon and some allied genera that were previously included in the Xylariaceae, but have recently been reassigned to the Hypoxylaceae . Since many of the compounds observed during GC-MS analyses in the volatile profiles of these fungi could not be identified with confidence in the latter study, we have selected some of these strains for intensified evaluation. Here, we present the identification, synthesis and bioactivities of volatiles emitted by the rare tropical hypoxylaceous ascomycete Daldinia clavata, which has hitherto been only infrequently reported from Africa and Latin America. ## Headspace analysis The volatiles emitted by agar plate cultures of Daldinia clavata MUCL 47436 grown on YMG medium were collected on charcoal filter traps by application of a closed-loop stripping apparatus (CLSA) . Dichloromethane extracts of the charged filters were analysed by GC-EIMS, followed by identification of the captured volatiles by comparison of the recorded mass spectra to data base spectra (NIST and Adams ) and of measured retention indices to reported data. For unknown compounds a structural proposal was developed by interpretation of the mass spectra, followed by the synthesis of reference compounds for unambiguous verification of the suggested structures. ## Volatiles from Daldinia clavata identified by GC-MS A representative total ion chromatogram of a CLSA headspace extract from D. clavata is shown in Figure 1. Several of the emitted volatiles were readily identified from their mass spectra and retention indices, including 4-methylhexan-3-one (6), oct-1en-3-ol (1), octan-3-one (7), 1-phenylethanol (8), and pogostol (16), which was further confirmed by comparison to authentic standards for 1, 7 and 8 (Table 1 and Scheme 2). Furthermore, the two structurally and biosynthetically related compounds 2-methyl-4-chromanone (12) and 5-hydroxy-2-methyl-4-chromanone (13) were tentatively identified from their mass spectra. Compound 13, which exhibits antimicrobial activity, was previously isolated from various species of Daldinia [16, and several endophytic fungi, the latter of which have only been tentatively characterised at the genus level . The compound actually consitutes one of several chemotaxonomic marker metabolites for the clade in the Hypoxylaceae comprising Daldinia and allied genera . The sesquiterpene alcohol 16 was first isolated from the plant Pogostemon cablin (patchouli) , but is also known from fungal sources . We show here the corrected structure as reported by Amand et al. , while for the compound from patchouli oil the opposite absolute configuration has been assigned . The 4-methylhexan-3-one (6) 845 842 ms, ri 2.0% oct-1-en-3-ol (1) 981 974 ms, ri, std 0.4% octan-3-one (7) 988 979 ms, ri, std 0.4% 1-phenylethanol (8) 1060 1057 ms, ri, std 0.2% 6-methyl-5,6-dihydro-2H-pyran-2-one (9) 1072 -syn 0.9% manicone (10) 1136 -ms, syn 0.4% (4R,5R,6S)-5-hydroxy-4,6-dimethyloctan-3-one (11a) 1228 -syn 35.2% 2-methyl-4-chromanone (12) 1366 -ms 0.2% 5-hydroxy-2-methyl-4-chromanone (13) 1467 -ms 2.0% 1,3-dichloro-2,4-dimethoxybenzene (14) 1480 1487 ms, ri, std 0.03% 1,2,4-trichloro-3-methoxybenzene (15) 1504 -ms, std 0.1% pogostol (16) 1653 1651 ms, ri 0.3% 6-nonyl-2H-pyran-2-one (17) 1875 -syn 9.7% a Identification based on ms: identical mass spectrum, ri: identical retention index (standardised GC retention based on comparison to n-alkanes; for C n H 2n+2 the retention index is defined as I = 100•n), std: comparison to a commercially available standard compound, syn: comparison to a synthetic standard. b Peak area in % of total peak area. The sum is less than 100%, because compounds originating from the medium, unidentified compounds and contaminants such as plasticisers are not mentioned. absolute configuration of 16 from Daldinia clavata remains unknown. The headspace extracts from D. clavata also contained two chlorinated compounds as indicated by the isotope pattern of the molecular ions in the respective EI mass spectra. The first of these compounds was readily identified as 1,3-dichloro-2,4dimethoxybenzene ( 14) by comparison to an authentic standard and to synthetic standards of all possible positional isomers. These isomers have been made accessible during our previous work that resulted in the identification of 1,5-dichloro-2,3-dimethoxybenzene as a headspace constituent of an endophytic Geniculosporium sp. . The mass spectrum of the second compound pointed to the structure of 1,2,4-trichloro-3methoxybenzene (15) which was confirmed by comparison to a commercially available reference. Chlorinated anisoles are well known as drinking water contaminants that can be sensed by humans with extremely low detection limits . Their origin by biomethylation of the corresponding phenols is frequently discussed, but the de novo formation of compounds 14 and 15 without administration of the corresponding phenols has not been reported before. ## Volatiles from Daldinia clavata identified by synthesis Several other compounds released by D. clavata could not unambiguously be identified based on GC-MS data only. In these cases structural proposals were delineated from the recorded EI mass spectra and the suggested structures were proven by synthesis of a reference compound. The first compound showed a mass spectrum (Figure 2A) that was similar to a data base spectrum of manicone ((E)-4,6-dimethyloct-4-en-3-one, 10), but no retention index for this compound was available from the literature. For unambiguous structural verification compound 10 was synthesised starting from 2-methylbutanal (18, Scheme 3). A Horner-Wadsworth-Emmons reaction with triethyl 2-phosphonopropionate (19) yielded ethyl (E)-2,4-dimethylhex-2enoate (20) as a separable mixture of E and Z stereoisomers (E/Z = 10:1). The purified E diastereomer was reduced with DIBAl-H to the corresponding alcohol 21. A PCC oxidation and addition of ethylmagnesium bromide gave 22 that was subsequently oxidised with PCC to the target compound 10. Comparison of the natural product to synthetic 10 established their identity by same retention time and mass spectrum. The mass spectrum of one of the two main compounds (11a) in the headspace extracts (Figure 2B) showed some fragment ions that were also observed for 10, suggesting that the two volatiles may be structurally related. The fragment ions at m/z = 57 and m/z = 86 supported the structure of a 4-methyl-3-ketone (these ions would arise by α-cleavage and McLafferty rearrangement). Biosynthetically, 10 is a tetraketide, and if the dehydration step to install the C=C double bond in 10 would be omitted, this would lead to the hydroxy-ketone 11a for which a higher retention time than for 10 would be expected. The structure of such an alcohol was further supported by the fragment ion at m/z = 115 that may result from an α-cleavage next to the alcohol function. To verify this structural proposal for 11a, racemic 2-methylbutanal (18) was reacted in an aldol addition with the enolate anion of pentan-3-one (23) which produced a racemic mixture of all four diastereomers 11a-d (Scheme 4). All eight stereoisomers of 11 were separable by GC on a homochiral stationary phase, one of which matched the natural product in terms of same retention times and mass spectra (Figure 3). To clarify the relative and absolute configuration of the natural stereoisomer of 11 an enantioselective synthesis was performed (Scheme 5). The alcohol (S)-2-methylbutan-1-ol (24) was converted into the corresponding aldehyde by Swern oxidation, followed by the addition of 25 to yield the esters (2E,4S)-and (2Z,4S)-20 as a mixture of diastereomers (ca. 15:1) that was readily separated by column chromatography. The E isomer was reduced with DIBAl-H to 26 that was converted into the epoxide 27a by Sharpless epoxidation with (+)-L-DET. Treatment with TBSOTf and Hünig's base resulted in opening of the epoxide with concomitant hydride migration to yield 28a. The stereochemical course for this reaction has been reported by Jung and D'Amico and proceeds with inversion of configuration at C-2. Grignard reaction with ethylmagnesium bromide to 29a, PCC oxidation to 30a and deprotection with HF-pyri- Scheme 3: Synthesis of manicone (10). dine gave access to (4R,5S,6S)-11c with an overall yield of 7% via seven steps. The 13 C NMR data in CDCl 3 (cf. Supporting Information File 1) were identical to previously reported data . The allyl alcohol 26 was also used in a Sharpless epoxidation with (−)-D-DET to give 27b that was converted into (4S,5R,6S)-11d via the same sequence of steps with a total yield of 6% via seven steps. Starting from the alcohol 24, two more stereoisomers, (4R,5R,6S)-11a and (4S,5S,6S)-11b, would be accessible via the same route from (2Z,4S)-20, but this compound was not obtained in sufficient quantity for a practical approach towards 11a and 11b. Therefore, these two stereoisomers were obtained by epimerisation at the α-carbon (C-4) under mildly basic conditions (Scheme 6). Following this approach, the epimerisation of (4R,5S,6S)-11c yielded (4S,5S,6S)-11b, while epimerisation of (4S,5R,6S)-11d gave (4R,5R,6S)-11a. The epimerisation products were added to the synthetic mixture of all eight stereoisomers, showing by enantioselective GC analysis that the natural product was identical to (4R,5R,6S)-11a (Figure 3). For the chemical characterisation of all eight stereoisomers of 11 the mixture obtained by the reaction shown in Scheme 4 was separated by extensive chromatographic purification. Compound 11c was readily separated from the mixture of 11a, 11b and 11d by simple column chromatography on silica gel. Preparative reversed-phase HPLC allowed for a separation of all three compounds 11a, 11b and 11d. Subsequent preparative HPLC using a homochiral stationary phase gave access to the pure enantiomers of these three compounds, but unfortunately not of 11c. However, as described above, enantiomerically pure (4R,5S,6S)-11c was obtained by enantioselective synthesis, and Scheme 6: Epimerisations of (4R,5S,6S)-11c and (4S,5R,6S)-11d under basic conditions. the peaks for the enantiomers of 11c in the GC analysis on a homochiral stationary phase could be readily assigned by comparison to synthetic (4R,5S,6S)-11c. Finally, a structure could be assigned to each of the eight peaks observed in this analysis, again confirming the structure of (4R,5R,6S)-11a for the natural product from D. clavata (Figure 4). The 13 C NMR data of all four stereoisomers 11a-d are summarised in Table 2. A proposed biosynthetic pathway to (4R,5R,6S)-11a is shown in Scheme 7 that is likely performed by a typical fungal iterative polyketide synthase (PKS). Starting from acyl-carrier-protein (ACP) bound acetate a first elongation step with malonyl-SCoA (Mal-SCoA) catalysed by an acyl transferase (AT) and a ketosynthase (KS) domain yields acetoacetyl-SACP. This may be followed by SAM-dependent C-methylation by a methyl transferase domain (MT). The stereochemical course for this reaction can be inferred from the 4R-configuration of the final product 11a, if indeed an iterative PKS is involved that should have the same stereochemical course for the corresponding reactions in each chain extension step. A keto-reductase (KR) installs the 3-hydroxy group with the same stereochemistry as observed at C-5 in 11a. This is followed by elimination of water by a dehydratase domain (DH) and reduction of the C=C double bond by an enoyl reductase (ER) with installation of the stereocentre in intermediate A corresponding to C-6 of 11a. The next chain extension with malonyl-SCoA and methylation proceeds with the same stereochemical courses as discussed above, but stops after action of the KR to yield intermediate B in which all the stereocentres that occur in 11a are already defined. A third extension with malonyl-SCoA and methylation gives rise to intermediate C that can be released, e.g., by a thioesterase to the β-keto acid D, followed by spontaneous decarboxylation to 11a. Two structurally related molecules to 11a have been reported from endophytic Nodulisporium spp. (shown in the box in Scheme 7) that may be formed by a similar PKS. Further investigations are required to identify the PKSs for this family of metabolites and to confirm the hypothetical biosynthesis as shown in Scheme 7. The volatile 9, a compound emitted in small amounts by D. clavata, showed a mass spectrum that was not included in our database (Figure 5A). However, the mass spectrum of 6-propyl-5,6-dihydro-2H-pyran-2-one (Figure 5B) is very similar, and the mass difference for the molecular ion of 28 Da suggested the structure of 6-methyl-5,6-dihydro-2H-pyran-2-one for compound 9. This compound was synthesised by esterification of pent-4-en-2-ol (31) with acryloyl chloride (32) to 33, followed by ring-closing metathesis using the Hoveyda-Grubbs catalyst of the second generation (Scheme 8). The synthetic material proved to be identical to the volatile of D. clavata, thereby establishing its identity. Finally, major amounts of a compound were emitted by D. clavata whose mass spectrum showed strong similarities to the mass spectrum of 2 (Figure 5C and 5D). The mass difference of 56 Da for the molecular ion pointed to four additional methylene units, suggesting the structure of 6-nonyl-2H-pyran-2-one (17). For comparison a synthetic reference compound was prepared from undec-1-yne (34) and propiolic acid (35) in a gold-catalysed reaction developed by Schreiber and co-workers (Scheme 8). Synthetic 17 and the volatile from D. clavata showed the same mass spectrum and retention time, confirming their identity. The newly identified natural product 17 further extends the number of known 6-alkyl-2H-pyran-2-ones: While the mixture of 6-propyl-, 6-pentyl-and 6-heptyl-2H-pyran-2one occurs in Trichoderma viride , higher homologs have not been reported as natural products so far. ## Biological characterisation of compounds Three of the volatile metabolites identified from D. clavata (compounds 9, 14 and 17) were evaluated for antimicrobial and cytotoxic activities using a standard panel of fungi and bacteria, as well as the murine cell line L929 (Table 3). The 2-pyrone 17 exhibited a moderate antifungal and weak antibacterial effect, while the lactone 9 was moderately cytotoxic, but devoid of significant antimicrobial activity. The chlorinated aromatic com-pound 14 only weakly inhibited the sensitive test organisms Chromobacterium violaceum and Mucor hiemalis. ## Conclusion The current study provides evidence that manifold new volatiles may be encountered in the future through a systematic study of Xylariales and other predominant fungal endophytes. As many of these fungi show interesting activities in screening approaches for biocontrol agents, and some of them are even under development for commercial applications, such studies not only complement the knowledge on the metabolic capabilities, but may even become mandatory in order to provide final proof for the safety of the organisms. There have been some studies on VOC-producing xylarialean endophytes with significant activities against competing microbes, but the volatile profiles of these biocontrol candidates were only evaluated using databases like NIST, which can only serve to detect and identify known compounds. The current study demonstrates the need of chemical synthesis for rigorous identification of new compounds. Some of these metabolites were tested for biological effects and found to display only weak activities in biological systems, providing evidence for their safety. In addition, none of the volatiles detected here represents a metabolite that is biosynthetically linked to a known class of hazardous mycotoxins. Similar studies on other fungal cultures with proven initial antagonistic activities that can be related to VOCs will probably be rewarding. ## Supporting Information Supporting Information File 1 Experimental details, synthetic procedures, and spectroscopic data synthetic [http://www.beilstein-journals.org/bjoc/content/ supplementary/1860-5397-14-9-S1.pdf]

chemsum

{"title": "Volatiles from the tropical ascomycete <i>Daldinia clavata</i> (Hypoxylaceae, Xylariales)", "journal": "Beilstein"}

tautomer_database:_a_comprehensive_resource_for_tautomerism_analyses

## Abstract: We report a database of tautomeric structures that contains 2,819 tautomeric tuples extracted from 171 publications. Each tautomeric entry has been annotated with experimental conditions reported in the respective publication, plus bibliographic details, structural identifiers (e.g. NCI CADD identifiers FICTS, FICuS, uuuuu, and Standard InChI), and chemical information (e.g. SMILES, molecular weight). The majority of tautomeric tuples found were pairs, the remaining 10% were triples, quadruples, or quintuples, amounting to a total number of structures of 5,977.The types of tautomerism were mainly prototropic tautomerism (79%), followed by ring-chain (13%) and valence tautomerism (8%). The experimental conditions reported in the publications included about 50 pure solvents and 9 solvent-mixtures with 26 unique spectroscopic or nonspectroscopic methods. 1 H and 13 C NMR were the most frequently used methods. A total of 77 different tautomeric transform rules (SMIRKS) are covered by at least one example tuple in the database. This database is available as a spreadsheet for free download from https://cactus.nci.nih.gov/download/tautomer/. ## INTRODUCTION Tautomerism is a phenomenon in which a set of molecules can interconvert by movement of a hydrogen or group of atoms and/or molecular rearrangement. The movement of hydrogen atoms along with the migration of pi-bonds is called prototropic tautomerism. The intermolecular arrangements leading to isomerization due to ring opening or cyclization are known as ring-chain tautomerism. (We have recently compiled 11 sets of rules for ring-chain tautomerism in SMIRKS notation. 1 ) Another type of isomerization, which involves rapid reorganization of single and double bonds without migration of any atom or group is termed valence tautomerism. At its core, tautomeric interconversion is a chemical reaction although one that is seen as happening "by itself" given the right conditions, not forced by a chemist's intervention. The boundary between the two is, however, to a certain degree a matter of definition. In this vein, the interconversion of tautomers is affected by solvent, pH, light, temperature, pressure, etc. Tautomers may be in rapid equilibrium or may take days to years to interconvert under experimental, storage, or transport conditions. It is well documented that a tautomer that is dominant in solution may not necessarily be favored in the solid state. 2,3 There has been significant interest in the pharmaceutical industry in the tautomeric polymorphism of drug-related molecules. Tautomerism is one of the causes of polymorphism of drug molecules. These polymorphs may have important, differing, properties such as solubility, bioavailability, physical/chemical stability etc. and may affect the patent life of the molecule. 4,5 Well-known examples of drug molecules that show tautomeric polymorphism are omeprazole, ranitidine, sulfasalazine. Polymorphs of these drugs have been isolated as different tautomeric forms or sometime mixtures of their tautomers. In prototropic tautomers, hydrogen transfer may take place by either intra-or inter-molecular processes. Intramolecular transfer may involve an isolated molecule in the gas phase or in dilute solution in aprotic solvents while intermolecular hydrogen transfer may be observed in the selfassociated aggregates of tautomeric forms, in highly concentrated solution in aprotic solvent, and in the solid state. Protic solvents may facilitate transfer of hydrogen by means of their own proton(s) after forming a complex with the solute's tautomer. 6,7 In most cases, migration of a single hydrogen is involved, however molecules with more polar groups may transfer more than one proton by an either concerted or stepwise mechanism. 8,9 Tautomers usually have different physicochemical properties such as logP, hydrophobicity, pKa, solubility, electrostatic potential, similarity index, etc. with concomitantly computation of such properties and molecular descriptors yielding different results from one tautomer to another. 10 This leads to the question: how to select the tautomer(s) of a molecule that allow one to most accurately predict its properties? One of the issues in this context has been the lack of a publicly accessible database providing a significant number of quantitative ratios or qualitative data of tautomeric forms in different solvents. The difference in the hydrogen bond donors and acceptor patterns between different tautomers of a screening or lead compound may affect interaction of these tautomers with its target protein and so may have profound influence on the drug discovery process. Therefore, consideration of tautomers has been of high interest to the drug design community for decades. A study suggested that using all possible (predicted) tautomers including unstable and stable ones yielded poorer docking results when compared to including only the stable one(s) in water. 12 The oftentimes small structural changes between members of a congeneric series nevertheless can lead to significant changes in the relative abundance of different tautomers. It has also been quite difficult to identify the tautomer that binds to a target protein as it may be any one from the pool of low-to high-energy tautomers found in a different environment. Martin 10 discussed several examples where a minor or less stable tautomer was recognized as binding to the macromolecule. X-ray crystallography, unless solving macromolecular structures at ultra-high resolution (say, below 0.8 ), typically suffers from insufficient precision in the position of heavy atoms and negligible electron density of hydrogen atoms. This lack of experimental evidence of hydrogen positions leads to placement of hydrogens based on chemical assumptions if not default settings of software. Neutron crystallography is more reliable in the determination of the tautomeric state of molecule, unfortunately there are very few structures in the Protein Data Bank (PDB) 15 or Cambridge Structural Database (CSD) 16 solved by neutron diffraction. Computational approaches, many of which need suitable training sets of sufficient size of experimentally determined tautomeric prevalences, may be an, if not in many cases the only, alternative. Borbulevych et al. 17 suggested refinement of X-ray structures based on a quantum-mechanical force field could aid in the determination of the protein-bound tautomer. Different studies have given widely varying estimates of the tautomeric potential of molecules in small-molecule databases. Trepalin et al. 14 indicated the presence of 0.78% of tautomeric moleculesin seven commercial database of bioscreening compounds using 23 pairs of tautomeric fragments. Cruz-Cabeza and Groom 15 suggested that 10% of the molecules in the CSD have tautomeric potential. Milletti et al. 20 estimated that an average of 29% molecules in four databases are tautomeric among about 0.7 million records. Chemoinformatics analysis by our group on 103.5 million records aggregated from 150 small molecule databases suggested that more than 66% of the molecules are susceptible to tautomerism based on about 20 tautomerism transform rules applied. 21 Recently, we found more than 30,000 cases where two or more products from the Aldrich Market Select (AMS) database of (then) ~ 6 million chemicals were just different tautomeric forms of the same compound -with the interesting corollary that these different tautomeric forms of a compound were offered at different unit prices by the same vendor. 22 In another study, we reported numbers in the same range for typical small synthetic organic molecules from the AMS and natural products, yielding an average of 76% and 56%, respectively, of structures susceptible to prototropic tautomerism. In contrast hereto, the occurrence rate of ring-chain tautomerism was found to be twice in natural products compared to AMS (16% vs. 8%). 1 Organizations oftentimes develop their own protocols for compound registration, which may lead to incompatible ways of handling tautomerism during, e.g., database and company mergers or switch to other compound registration and management software. Warr 23 reviewed different approaches used by 27 software vendors and database producers for compound registration including how tautomers were handled. While thus a significant body of work exists of experimental, theoretical (quantum chemical), and -to a limited extent -chemoinformatics (rule-based), studies of tautomerism, very few systematic collections of experimental results in this field have been undertaken so far. A set of 785 transformations belonging to 11 types of tautomeric reactions with equilibrium constants measured in different solvents and at different temperatures was recently used in an effort to build QSPR models of equilibrium constants of tautomeric molecules. 24 To the best of our knowledge, there is currently no database publicly available that provides details of a significant number of molecules and their experimentally investigated tautomers under specific experimental conditions along with a detailed chemoinformatics analysis. Here, we report on a tautomer database we have created from literature in an attempt to compile experimental, quantitative, tautomeric preferences together with chemical and bibliographic information as well as an analysis along a set of more than 70 tautomeric transforms. We hope this resource, which we have made freely downloadable at https://cactus.nci.nih.gov/download/tautomer/ will allow the scientific community to more easily explore the phenomenon of tautomerism by finding several thousand such molecules in one place. ## Dataset of database The current tautomer database consists of 2,819 entries, each comprising an n-tuple of tautomers (n = 2…5) studied in a particular set of experimental conditions (pH, solvent, solvent mixture, temperature, experimental technique used). All these tuples together comprise a total of 5,977 records. The data were extracted from 171 publications, which included a number of reviews (see full list in the Supporting Information Table S1). The initial extraction from these literature sources was done by a contract mechanism (Parthys Reverse Informatics, http://www.reverseinformatics.com) whereas the significant work-up and curation of the inital data was performed by the authors. For each entry for all n-tuples in the tautomer database the corresponding NCI/CADD Chemical Structure Identifiers 25 were calculated using the chemoinformatics toolkit CACTVS 26 (in which they have been implemented as standard molecular properties). The nature of these identifiers, which are based on the standard CACTVS molecular hashcodes, is based on turning off or on sensitivity to the following five chemical features: fragments, isotopes, charges, tautomers, and stereochemistry. In this database, we used the FICTS, FICuS and uuuuu identifiers out of possible 32 possible set of variants (see the original publication for explanation of the nomenclature). 27 The FICTS identifier, in which all five features are turned on, represents the original input structure as is. It is sensitive to fragments (such as counterions), isotopes, charges, and stereochemistry in the input structure as well as to the specific tautomer drawn. The FICuS identifier is tautomer invariant (but sensitive to all four other features), meaning that different tautomers have the same FICuS hashcode. In the case of the uuuuu identifier, all five features are turned off, implying that molecules differing only by fragment, isotope, charge, tautomer and stereochemistry have the same uuuuu (which can thus be regarded as a sort of parent structure identifier). Based on its design, the FICuS identifier is thus conceptually similar to the InChIKey, though the latter handles tautomerism less comprehensively than FICuS due to an only limited range of tautomerism transforms implemented in its current version (v. 1.05). Additionally, it is not currently possible to add entirely new types of tautomerism to the InChI[Key] calculation. It should be noted that this and other shortcomings of the current InChI in the handling of tautomerism has led to an IUPAC-sanctioned project of redesigning the handling of tautomerism for an InChI V2 28 , for which the tautomer database described in this paper forms an experimental backdrop of sorts, and whose authors are involved in the IUPAC project. The standard InChI and InChIKey were also calculated with CACTVS and are included in our database. In order to describe the tautomeric transformation(s) between the members of each of the tautomeric n-tuples in the database, we used a total of 77 rules. This set is closely related to, and essentially a major subset of, the 86 rules described in the context of redesigning of the handling of tautomerism for InChI V2 in the accompanying paper. 29 All these rules were encoded in SMIRKS line notation developed by Daylight Chemical Information Systems, Inc. 30 They were all processed in the CACTVS chemoinformatics toolkit, which comes with a default set of 20 prototropic rules covering a wide range of common and some rarer types of tautomerism, of which 12 have a representative in our database. To these, we added a subset of 8 SMIRKS from our recently published 11 types of ring⇌chain rules (encoded in a total of 38 SMIRKS) 1 , plus 57 out of 61 heretofore unpublished rules, which are detailed in the accompanying publication. 29 We use the following nomenclature (again aligned with the accompanying paper 29 ) for the three types of rules discussed in this paper: 1. Prototropic tautomerism rules are called PT_nn_mm, where nn and mm are the number of the rule, and a possible variant, respectively. The names of most rules end with an 00 indicator, indicating that there is only one variant. 2. Ring-chain tautomerism rules are named RC_nn_mm, where nn and mm have the same meaning as described above. 3. Valence tautomerism rules are termed VT_nn_mm according to the same scheme. To determine the single transform or sequence of transforms connecting the tuple members with each other, we applied the following procedure: In a first step, we enumerated all possible tautomers from each tautomeric tuple; in the second step, we generated a tautomer network among those enumerated tautomers. In such a network, we typically have several pathways that connect one tautomer to the other by different tautomeric transforms. As the final step, we searched for the shortest pathway, defined by the smallest number of transformation steps within the tautomeric pair. If two different paths had the same number of steps, we used a notation of the type of: {PT_03_00/PT_06_00} > PT_09_00. This means the pathway can either use PT_03_00 or PT_06_00 in the first step, followed by PT_09_00 in the second step. ## Database description To recapitulate, our database of tautomers contains 2,819 entries, extracted from 171 publications, including reviews and research articles. It is provided to the user as a spreadsheet in Excel format. Each entry consists of three major segments: conditions, tautomer and publication. Each segment has several fields as listed in Table 1. For each additional (second, third,…) tautomer of a compound, fields in the second (and third etc.) instance of the tautomer-specific columns are populated with data, otherwise left empty. The following provides a brief explanation of some key columns in the spreadsheet. Others should be self-explanatory. A Legends worksheet is also available in the spreadsheet providing explanations for all columns. A. Size. Number of tautomers reported in the publication as being in equilibrium. In a few publications, only the main tautomer of the compound was described, in such cases we entered a second entry based on a possible (calculated) tautomer. B. Solvent. Solvent in which the tautomers were observed. This can be a mixture of solvents. If their concentration is indicated, then it is also mentioned in the solvent column. C. Solvent_proportion. Fraction of solvents or their mixtures (percentage, ratio) used. D. Solvent_Mixture. Indicates whether a single solvent or mixture was used. This column has a "yes" if the "Solvent" column indicates a solvent mixture, otherwise "no". E. Temperature. Temperature (K) at which the tautomers were observed or the experiment carried out. In case of mass spectroscopy experiments, the temperature of the injector was used as the experimental temperature. F. pH. pH of the medium at which the tautomers were observed or experiment was conducted. G. Experimental_Method. This describes the spectroscopic or physical methods that were used in the experimental determination of the tautomers. It may be a single method or a combination of several methods, by which the tautomers were established in the experiments. If the experimental details were not available in the review then those were extracted from the original references. To give an example, data extraction from ref. 9 in the spreadsheet had yielded five entries for five different solvents used in the experiments (acetonitrile, CCl 4 , chloroform, dichloromethane, and THF), each with (the same) two tautomers reported. The following columns therefore carry a suffix ("1" or "_1" for the first tautomer, "2" or "_2" for the second). Explanations are only given for columns pertaining to the first tautomer since columns for subsequent tautomer(s) are entirely equivalent. a Entries with the value "nul" in any column indicate that it was not possible to extract sufficiently specific information from the publication. H. Entry_ID1. Unique ID composed from the publication reference (journal name, year, volume, page numbers) along with the tautomer ID in that publication (if given) and the nature of the tautomerism (e.g., "Keto⇌enol"). I. Type_1. The chemical nature of the tautomer, e.g. keto, hydroxy, imine, enamine, etc. An entry with "nul" in this column indicates that it was difficult to assign any specific name from the molecule's common name or based on similar structures in the database. ## J. Transf_1_2. The rule(s) (prototropic [PT], ring⇌chain [RC], or valence tautomerism [VT]) which transform(s) tautomer_1 into tautomer_2 (single or multiple steps). A forward slash "/" is used to indicate alternative rules for any step. Curly braces "{}" are used to group together alternative rules if these appear in multi-step transforms. The greater than sign ">" is used to separate steps in multi-step transformations. An entry with "no_transform" in this column indicates that these pairs are not covered by our rules because these examples are releated to zwitterionic and complex protonated structures, hence we did not develop any rules for them. If there were 3 or more tautomers reported, there would be corresponding columns in the spreadsheet with "3" or "_3", e.g. Transf_1_3, etc. ## DATABASE ANALYSIS Provenance and relationship of tuples We did not identify any direct tuples' duplicates in terms of both chemical structure and experimental conditions. Purely chemical duplicates were found for 479 tautomeric tuples in the database but they differ in conditions such as temperature, solvent, pH, or spectroscopy method. All tuples came from the same original source, i.e. typically a specific publication. We did not look for, or include, sets of tautomers of the same molecule for which different tautomers in a large database such as PubChem or large vendor catalogs come from different original sources (source set providers, original sample vendors etc.). Such cases -which we have termed "tautomic conflicts" 22 -are numerous within, and even more numerous across, large smallmolecule databases but are not the topic of this paper. They will be more thoroughly discussed in a future publication. ## Size distribution of the tuples The database contains tautomeric tuples ranging in size from 2 to 5. The majority of tautomeric equilibria (2,530 cases) in the database are for 2 tautomers although we also found equilibria with 3 (250), 4 (27), or 5 (11) tautomers, respectively (Figure 1). The tautomeric equilibria for more than 2 tautomers comprise around 11% of the cases. ## Solvent The database contains tautomeric equilibrium studies performed not only in solvents as the medium but also in the solid state, neat liquid, gas matrix, and vapor phase. Still, the majority of experiments were conducted in some kind of solvent or solvent mixture. About 50 different types of solvents were reported for the determination of tautomeric contents of the molecules under study. The frequency distribution of the top 20 solvents is given in Figure 2a. Chloroform (normal or deuterated) was the most frequently used solvent, with dimethyl sulfoxide (DMSO) in second place both together accounting for around 45% of all reported solvents. If we add methanol (~6%) and acetonitrile (~5%), the top 4 solvents cover about 57% of the cases. Looking at the tail end of the distribution shows that polar solvents such as water were used infrequently (~2%), at least as a pure solvent. Turning to the non solvent-based experiments, we note that 6.8% and 3.8% of the cases were determined in gas phase and solid state, respectively. In ~1% of the cases, tautomerism was determined in the neat liquid form of the molecules under investigation. The database has 12 solvent mixtures, in which 12 different types of solvents were used. Around 5% molecules of the database were studied in such mixtures. The majority of these cases were water+ethanol (34), water+DMSO (21) and DMSO+TFA (16), respectively. Their quantitative distribution is given in Figure 2b. In these mixtures, water was one of the common co-solvents mixed with ethanol, DMSO, dioxane, or acetone, respectively, in ~66% of the cases of mixtures. The database has experimental pH details for 100 entries. Of those, 63% of the records were reported to have used acidic medium. The distribution of entries by pH range is shown in Figure 4a. Except for the pH value range 10.01-12.0, most other ranges have at least 10% of the studies each. Most of the acidic and basic condition studies fell in the range of 4.01-6.0 and 8.01-10.0, respectively. We did not find any tautomeric tuples containing four or more tautomers having all been studied within one particular pH range. In most of the cases (91entries), a set of 2 tautomers was observed at the given pH range and in a few cases (9 entries) a set of 3 tautomers was reported in pH based studies. In these studies, medium polar to polar solvents (98%) or their mixtures (2%) were used. The most commonly used solvent for such studies was methanol (Figure 4b). These studies used the following spectroscopy methods: 1 H NMR, Flash photolysis, Raman, UV and UV/VIS. Of these, UV/VIS spectroscopy was used in 79% of the cases with methanol, acetonitrile and DMSO-water. ## Experimental methods In the most of the studies (85%), a single spectroscopy or physical method was used while in remainder of the studies 2 to 3 methods were used, often by way of an additional method used as support of the primary method. In the multiple method studies, spectroscopic methods from among 1 H, 13 C, 14 N, 15 N, 17 O, and/or 31 P NMR spectroscopy were the most common ones and these NMR methods accounts for ~75% cases in multiple method. Out of the total 29 unique methods, 1 H NMR (1014), 13 C NMR (340), UV (253), IR (172) and UV/VIS (139) are the top 5 spectroscopy methods which were used as the major experimental technique (Figure 5a). In the multiple method studies, 1 H NMR and 13 C NMR were frequently used together (131). In addition, 1 H NMR was commonly used together with other methods such as 31 13 C NMR (182 and 84 cases, respectively) (see the Solvents with Spectroscopy Methods in the Supporting Information, Table S2. In IR chloroform (59) and nujol (38), in UV/VIS methanol (89) and acetonitrile (8), and in UV ethanol (76) and water (31) were used extensively. ## Analysis by tautomeric transform rules As already mentioned, we used as the starting point for the tautomeric rule compilation (a) 20 standard prototropic rules (comprised by the 20 default CACTVS rules PT_02_00 -PT_21_00); (b) 11 ring⇌chain (RC_01_00 -RC_11_00) rules that have been published by our group recently. 1 We point out that most of these 11 rules have more than one variant, yielding in fact a total of 38 SMIRKS strings. In addition, we have compiled 29 61 new tautomeric rules derived from various literature sources. These new rules consist of 34 prototropic rules (PT_22_00 -PT_49_00) plus two variants with nn > 00 and variants of PT_11_nn for long range hydrogen migration, where nn ranges from 01-04), 16 ring⇌chain rules (RC_03_03, RC_03_04, RC_04_04, and RC_12_00 -RC_24_00), and 11 valence rules (VT_01_00, VT_01_01 -VT_10_00). (See footnotes of Table 2 for rule naming and numbering nomenclature.) Table 2 shows the frequency of the applicability of all these rules to the entries in our database showing both the cases where the transformation between the experimental tautomers only required the application of a single rule as well as of cases that needed additional, or allowed alternative, rules in the single/multistep transformation between observed tautomers. It needs to be pointed out that there are cases for all "New" rules in our database simply because we added to the database all the cases from the literature source(s) that actually gave rise to each of the new rules as described in the accompanying paper. 29 In contrast hereto, the original (contractor-based) extraction from literature had not been rule-aware. Therefore some of the standard rules, de facto being quite rare even though labeled as "Standard," are not represented in the current version of the database (PT_13_00 -PT_15_00, PT_17_00 -PT_21_00). a Different classes of tautomerism are defined by prefixing each rule with PT, RC, or VT for prototropic tautomerism, ring⇌chain tautomerism and valence tautomerism, respectively. The second placeholder in rule name between underscore indicates rule number in that category (i.e. "02" in PT_02_00) and the last number in name indicates a variant of that rule (i.e."01" in VT_01_01, "03" in RC_03_03). Rule ending with "_00" occurs only in one variant for that rule. This naming scheme allows us to add more variants in that rule if it is required in the future. b We also have four variants of PT_11_nn for long range hydrogen migration, where nn ranges from 01-04. b SMIRKS of these tautomeric rules are given in Table S3 of Supporting Information. The most commonly encountered prototropic, ring⇌chain and valence rules are shown in Figure 6. The tautomeric transformation of one tautomer to another can take place in one or more steps, where each step corresponds to one rule, however in some cases these tautomeric transformation can be achieved by other rules. The majority of transformations from our database occur in a single step (60%) while the others involve the use of additional rules to complete the transformation. About 35% of transformations are achieved by the application of PT_06_00 and PT_07_00 in a single step. There are some rules (PT_02_00 -PT_05_00 , PT_08_00, PT_10_00 -PT_16_00 ) which do not appear in any single-step transformation, but appear several times in multistep transformations or as alternative rules to others. For example, 353 cases need additional one-step (for a total of two steps) and 27 other cases require two or more steps (for a total of three or more steps) to complete the observed tautomeric transformations. Generally, a hydrogen atom migrates in a tautomeric transformation from its initial position in the molecule to an odd numbered (relative) position (such as 3, 5, 7, 9, or 11), designated as "1,3H shift," "1,5H shift" etc. In contrast hereto, migration to an even position (such as 2, 4, or 6) is rare. Analysis of the database showed that in most cases hydrogen migrates via 1,3H shift (1,120), followed by 1,5H (707) and 1,7H (91) shifts, respectively, in the single step transformations. One notes that this distance travelled by the hydrogen is well correlated with the respective H shifts frequency of observation. We also have observed cases where the 1,3H shift can be achieved alternatively via long distance migration using 1,5H (30) or 1,7H shift (118), respectively. Likewise, 1,5H shift based transformations can be in competition with 1,7H and 1,9H shifts in a single step equilibria. For 2-step transformations, we observed the order by frequency of occurrence shown in Table 3. We note that as in single-step transformations, shorter distances hydrogen migrations are more prevalent than longer ones for the 2-step transformations too. The database contains significantly fewer cases (388) of ring⇌chain tautomerism than of prototropic tautomerism, generally belonging to cyclization to 4, 5 and 6 membered ring systems. This ring cyclization can occur either via endocyclic or exocyclic process where the double bond becomes part of the ring or of the side chain, respectively. Examples for both types of ring cyclization occur in our database. In 180 cases of endocyclic ring⇌chain transformations, ring closure happens at digonal (sp), trigonal (sp 2 ), or tetrahedral (sp 3 ) centers. In our recent NMR experimental study, we found 5 or 6 membered endocyclization at a trigonal center (represented by rules RC_09_00 and RC_10_00) to be the most "reliable" ring⇌chain rules in the sense that they predicted truly occurring tautomeric interconversion. 22 As per our new rule RC_13_00, 6-membered ring closure can also occur at a digonal center (sp) through an endocyclization process. Moreover, according to new rules RC_12_00, RC_18_00 and RC_24_00, 5-membered ring cyclization can also occur at a tetrahedral center. These three rules RC_12_00, RC_18_00, and RC_24_00 do not follow the concept of ring closing/opening according to Baldwin's rules. In contrast to other rules, RC_24_00 involves tautomerization between trivalent (chain) and pentavalent (ring) tautomers. It is worth pointing out that during ring closure the hydrogen atom usually travels by 1,3 H shift as noted in the case of all of old, and some of the new, ring⇌chain rules. However, there are some rules that involve 1,2 H shift (i.e. RC_24_00), 1,4 H shift (i.e. RC_15_00, RC_20_00 and RC_23_00) and 1,5 H shift (RC_22_00) during ring closure. We thus note that experimental evidence can introduce the need for new rules (at the chemoinformatics level) that are not covered by previous work aimed at creating comprehensive rule sets. In 193 cases of exocyclic ring⇌chain transformations, the ring closing/opening takes place at trigonal or tetrahedral centers. 5-or 6-membered cyclization occurs only at trigonal centers while 5-membered cyclization via RC_18_00 takes place only at tetrahedral centers. There are some instances of ring⇌chain tautomerism in the thiadiazoline (RC_14_00), boryl/borate (RC_16_00 and RC_17_00), and λ 5 /λ 3 -phosphane (RC_24_00) systems that do not involve any unsaturated electrophilic center (or endo-or exo-cyclic bonds) during interconversions but rather involve saturated sulfur, boron, and phosphorus centers, respectively. The ring-chain rules do not occur in combination with any prototropic or ring⇌chain rule, i.e. in all cases transformation proceeds in a single step. Therefore, ring⇌chain rules appear to be more selective and specific than prototropic rules. Generally, ring⇌chain tautomerism showed a high prevalence for the chain form over of the ring form. There are 20 cases of ring⇌chain tautomerism where three tautomers are in equilibrium with each other in solution, the two ring tautomers existing as cis and trans isomers, respectively. This provides interesting systems in which cyclic stereoisomers are in equilibrium with each other via the chain form; however, the chain form dominates over the two ring forms. There are 228 cases of valence tautomerism in the database. They all involve ring opening/closing in 4, 5 or 6 membered ring systems without migration of any hydrogen atom. The ring-opened tautomers of four rules (VT_02_00, VT_04_00, VT_05_00 and VT_08_00) have a charge-separated moiety in their structures and this charge disappears in the ring-closed tautomers. In contrast hereto, a charge-separated moiety is present in the ring-closed tautomer of both VT_07_00 and VT_10_00. The tautomeric equilibrium via VT_06_00 involves ringcontraction (6 membered) and -expansion (7 membered) in the tautomers. Out of the 11 rules, VT_09_00 is the only one rule that involves valency change during tautomerization:, between trivalent phosphinoimine and pentavalent diazaphosphazole tautomers. Among these 11 valence tautomerism rules, our database contains significant counts only for the tetrazole⇌azide tautomerism (VT_02_00). The dataset allows us to conclude that for the tetrazole⇌azide tautomerism, the tetrazole tautomer is more favored in polar aprotic solvent while the azide tautomer is dominant in nonpolar solvent. The available dataset for benzoxathiete derivatives is quite small, thus does not allow one to deduce any specific preference for one or the other tautomer, both being described in the available literature references. ## Type of tautomerism Many of the transforms listed in Table 2 align quite closely with chemotypes the way the organic chemist would usually perceive them. However, others among these transforms, as they are expressed as general SMIRKS patterns {Formatting Citation} , cover a broader range of compound types. For example, transform PT_06_00 (1,3 heteroatom H shift), recognizes C, O, N, S, P, Se, and Te in its SMIRKS pattern, thus covering quite diverse types of compounds and tautomerism based on those. Conversely, the interconversion between the hydrazine and the azo species of a compound can be effected at the transform level by 1,3H, 1,5H and 1,7H shifts, which are encoded in different transforms. We therefore felt it to be useful to analyze the distribution of the records in our database along these more chemical types of tautomerism. The database in this sense comprises of more than 50 types of tautomeric pairs. The most commonly encountered tautomer types are listed in Table 4 in the order of their occurrence (see molecular examples in the Supporting Information Table S4). We see that azo⇌hydrazone tautomerism is one of the most studied types of prototropic tautomerism, followed by keto⇌enol and oxo-enamine⇌oxo-imine forms, respectively. The number of records for azo⇌hydrazone is more than twice compared to the simple keto⇌enol form. Table 5 shows commonly identified sets of three tautomers with their occurrences. There are 51 cases in which a set of geometrical tautomers of enethiol (cis and trans or Z and E) is found in equilibrium with the thioketo form. On the other hand, in 42 cases of enol-imine⇌oxo-enamine⇌oxo-imine tautomerism, three chemically distinct tautomers are in equilibrium. Moreover, a set of 2 similar thioimidol, enamine, enol, and keto tautomers are found to exist in equilibria with thioamide, imine, keto, and enol form, respectively. Table 6 shows the distribution of some of the common tautomers across the 5 different prevalence categories described above (0-4). To remind the reader, category 0 indicates that the tautomer was not observed or was not in the detectable range; 4 indicates that the tautomer was exclusively observed; and categories 1-3 form the class that comprises various degrees of predominance between the two extremes. There are cases where the numbers of exclusively observed tautomers are higher than the numbers in the predominant class. For example, enethiol, oxo-imine, imine (counter tautomer of enamine) and phenol-imine are exclusively identified in 16, 9, 19, and 16 cases, respectively. In the case of keto-enethiol⇌thioketo-enol type tautomerism, all the tautomers fall in 3 categories (1-3). Ketoenethiol, oxo-enamine, pyridol, isoindole, phosphine and thioketo are some of the tautomers for which there is no single example in prevalence category 4 in the database. This does not exclude the possibility of their observation in other experiments, however, our database does not have examples from them. As per the data in Table 6, azo, keto, oxo-enamine, keto-enol, and thioketoenol are examples of tautomers usually preferred over their corresponding other tautomer(s). It is interesting to note that oxo-enamine can be in equilibrium with oxo-imine, enol-imine, and phenol-imine tautomers. It is usually preferred over both oxo-imine and enol-imine, however phenol-imine always dominates the oxo-enamine. Imine can be in equilibrium with enamine and amine tautomers, being typically preferred over enamine while it is less preferred over amine. c "nul" indicates cases of tautomeric equilibria for which no name for one or the other or both tautomer was given in the references and we were not able to assign any specific name. a See Table 4. Closely following the generic ring⇌chain case count of 388 (Table 4 and 5), the database contains 333 records for the hydrazone⇌azo type of tautomerism. Additionally, there are 3 cases where a zwitterionic species competes with both of the above-mentioned tautomers. Most of the studies were carried out in methanol (82) or DMSO (68), mainly using 1 H NMR (109) and UV/VIS (88) spectroscopic methods. This tautomerism can involve 1,3H, 1,5H, or 1,7H shift. In the simplest case, the hydrogen is hopping between carbon and nitrogen atoms via 1,3H shift as shown in Figure 8. If the X position in that figure is O, then the azo form is termed hydroxy-azo or enol. Its hydrazone counter tautomer is found with several different names in the literature such as quinone-hydrazone, keto-hydrazone, oxo-hydazone, or keto. If X is NH, then this pair is described in the literature by the terms azo and quinoimine-hydrazone tautomers, respectively. Of these, 145 cases included a single-step migration described by transforms PT_07_00 (99), PT_09_00 (79), or PT_06_00 (17) while others involved two or more steps of migration. The numbers in Table 6 show that the azo form was more favored in 122 cases whereas the hydrazone form was found to be more dominant in 72 cases as defined by prevalence category 3 and 4 collectively. The database contains 257 cases where any two tautomers from the following three tautomers were in equilibrium with each other: enol-imine, oxo-enamine and oxo-imine (Figure 9a). For example, oxo-enamine⇌oxo-imine, enol-imine⇌oxo-enamine, and enol-imine⇌oxo-imine type equilibria are present in the database in 113, 104, and 40 cases, respectively. These cases were mainly observed using 1 H NMR, UV/VIS, and mass spectroscopic methods in chloroform, ethanol, acetonitrile, and gas phase. They can involve 1,3H, 1,5H, or 1,9H shifts in a single-step or multi-step transformation with rules PT_02_00, PT_03_00, PT_04_00, PT_05_00, PT_06_00, PT_07_00, PT_09_00, or PT_10_00. Based on prevalence category, we observe that oxoenamine is found to be predominant or exclusively present in 65 cases, whereas oxo-imine is observed as predominant or exclusively present in 11 cases. The equilibrium between enol-imine and oxo-enamine was quite competitive, either state being observed predominantly or exclusively in 45 and 57 cases, respectively. The enol-imine is generally preferred in molecules in which the tautomeric skeleton is part of aromatic (benzene) or heteroaromatic (pyridine) ring systems. However, both tautomers are found with almost equal preference in molecules containing a hydroxyl group on the naphthalene ring, with only a few exceptions. Finally, in the third pair, enol-imine and oxo-imine were observed in prevalence category 3 and 4 in 12 and 19 examples, respectively. Simple keto⇌enol is the second most commonly observed type of tautomerism (137 pairs) after azo⇌hydrazone tautomerism. They were mostly studied in chloroform (53) and gas phase (25), and the rest in about 10 different solvents. We find equilibria for 133 and 3 cases based on PT_06_00 and PT_07_00, respectively. Among these pairs, the keto form was exclusively observed, or more stable, in 73 cases, whereas the enol form was found to be prevalent in 19 cases. β-dicarbonyl compounds can exist in the form of diketo and keto-enol type tautomers (Figure 9b). The acidity of such compounds is higher compared to monocarbonyl compounds hence are more prone to tautomerism. The database contains 108 pairs of diketo⇌keto-enol type tautomerism, which were mainly reported in chloroform, gas phase, and DMSO using 1 H NMR, 13 C NMR and gas chromatography/mass spectroscopy methods. The keto-enol form was observed predominantly or exclusively in 43 cases, while the diketo form was favored in 31 cases. This higher preference for the keto-enol tautomer may be due to possibility of hydrogen bond stabilization of the keto-enol form and the presence of resonance-stabilized conjugated πsystems. In addition, in these molecules, the position of the tautomeric equilibrium is also affected by the presence of electron withdrawing and donating substituents at the one β-and the two α-carbons. The database contains 33 pairs of keto-enol⇌keto-enol type tautomers that exist in equilibrium with other tautomer(s) (mainly observed in chloroform) via 1,5H shift as per rule PT_07_00. Such pairs could have been in equilibrium with a third diketo tautomer via 1,3H shift; however, we have only 2 cases where such two keto-enol tautomers were found in equilibrium with the diketo form. The keto-enethiol⇌thioketo-enol type of tautomerism appears in monothio analogues of βdicarbonyl compounds (Figure 9c). Analysis of 82 cases shows that the thioketo-enol tautomer is predominant mainly in acetonitrile (14), cyclohexane (43) and gas phase (1), and its counter tautomer is favorable in acetonitrile (3) and in the gas phase (1). Out of 26 cases of the enethiol⇌thioketo type tautomerism in simple thioketo molecules, the enethiol was predominant in 17 cases. It is interesting to note that in thio analogues of β-dicarbonyl compounds the thioketo-enol tautomer is favorable while in simple thioketo molecules the enethiol is favored. Pyridone⇌pyridol type tautomerism (58 pairs in 2-pyridone, 4-pyridone and 4-pyrimidinone derivatives) (Figure 10a) was studied at least in 10 different media including solvents, vapor phase, solid phase, liquid phase, N 2 /Ar matrix, KBr, and others. In these molecules, the pyridone type tautomer was found to be predominant in 29 cases, mainly in cyclohexane, chloroform and acetonitrile solvents while its counter tautomer pyridol was found to be favorable only in 8 cases in polar and nonpolar solvents. The amine⇌imine type of tautomerism (83) involves hopping of hydrogen between nitrogen atoms (Figure 10b) where one of the nitrogen atoms usually is present in an aromatic or alicyclic ring and the other in a side chain of the molecule. These molecules were studied using mainly dipole moment measurement, IR spectra and 15 N NMR spectra in at least 7 media. The transformation between tautomers mostly occurs via 1,3H shift in a single step and in a few cases in two steps. Molecules with the aromatic ring involving such tautomerism usually show a strong predominance for the amine form. On the other hand, molecules with alicyclic rings were found to exist predominantly in either the imine or the amine form. Overall, the database shows that the amine is favorable in 36 cases while the imine is predominant or exclusive in 27 cases irrespective of molecule types. The database contains 72 tautomeric pairs of enamine⇌imine type tautomerism (Figure 10c), which were mainly reported in TFA, DMSO and chloroform by 1 H and/or 13 C NMR spectroscopy. The imine form was predominant or exclusively observed in 27 cases (mainly in TFA and chloroform) while enamine form was favorable in 14 cases (mainly in chloroform and DMSO). Further, in a subset of 26 molecules out of the 72 pairs both tautomers were observed as major species. A dataset of 65 pairs of oxo-enamine⇌phenol-imine type tautomerism (Figure 10d) was mainly studied with 1 H and/or 13 C NMR spectroscopy or combination of those with 17 O and 15 N NMR spectroscopy in 5 different solvents. Chloroform (32) and acetone (11) were the main solvents for such tautomerism. In the literature, phenol-imines are often called by names such as hydroxyimine or enol-imine, which have by other authors been given to counter tautomers of other types of tautomers such as oxo-imines. The phenol-imine tautomer was observed predominantly or exclusively in 31 cases, while its counterpart oxo-enamine was not found to be predominant in even a single case. There are some dihydropyrimidines derivatives that show equilibrium between 1,4-and 1,6dihydro forms, and 1,2-and 2,5-dihydro forms (Figure 11). The insertion of one more nitrogen In our database, 1,3H shift of the amide bond in nonaromatic heterocyclic ring is considered in the lactam⇌lactim interconversion (Figure 12). This tautomerism was mainly investigated using electron ionization/mass spectra and UV spectra in gas phase and water. Out of 31 pairs, lactam tautomers were found to be predominant in 16 cases, while lactim tautomers were found predominantly or exclusively in 10 cases. The tautomers of the amide and thioamide bond in acyclic compounds are described by amide⇌imidol and thioamide⇌thioimidol tautomerism, respectively. In phenol-quinone⇌phenol-quinone type tautomerism, the hydrogen migrates from one oxygen to another oxygen, so these tautomers cannot be differentiated for their preferences; still, we have some examples in the database where both identical tautomers were reported to be in an equilibrium. Similarly, in tropolone-based molecules, hydrogen migrates between nitrogen atoms or oxygen atoms via a long conjugation path generating similar tautomers, i.e. without a change of chemotype. In amino-substituted tropolone molecules, the hydrogen shift occurs between oxygen and nitrogen atoms and in all cases, nitrogen in amino form and oxygen in keto form were exclusively observed in the examples in our database. The database contains 388 examples of ring⇌chain tautomerism covered by 24 rules, where 368 and 20 cases are from tuples of 2 and 3 tautomers, respectively. Of these, 85 pairs belong to 5_endo_trig (RC_09_00, RC_20_00, and RC_22_00), followed by 6_exo_trig (80 pairs from RC_04_01/RC_04_02, RC_04_04, and RC_19_00) and 5_endo_tet (39 pairs for RC_12_00) types, respectively. Their ring⇌chain equilibria were explored using about 30 media. Chloroform (113) was the most frequently used medium followed by DMSO, acetone, and carbon tetrachloride. There are 22 examples that were studied in their neat liquid phase, i.e. without using any solvent. In addition, water mixed with DMSO, dioxane, ethanol, or acetone was used as the solvent in 51 cases. ## Scoring for canonical tautomers We analyzed two-tautomer tuples (2,530) to check how likely CACTVS determines the canonical form of a given experimentally determined tautomers via the scoring scheme 21 developed for the NCI/CADD calculable chemical structure identifiers. 25 The current scoring scheme was able to consistently predict the right tautomer in 26% of the cases while in 36% it consistently failed to predict the correct tautomer. The most important aspect is to determine always the same tautomer regardless of which tautomer was used as the starting input to algorithm. In contradiction to this, in the remaining 38% of the cases the current scoring scheme predicted two different canonical tautomers for a given tautomer tuple. We mention this brief experiment here to illustrate that this database may open up a significant tool to implement, or improve currently implemented, tautomer scoring schemes to empirically predict a canonical tautomer. ## SUMMARY AND CONCLUSIONS A significant variety of structures, chemotypes, analytical procedures, and experimental conditions including solvents has been compiled to form the Tautomer Database. It contains, at the time of this writing, a set of 2,819 molecules extracted from experimental literature as tautomeric pairs or higher-order tuples. We have associated them with prototropic, ring⇌chain, or valence tautomeric transform rules, found applicable in 79%, 13%, and 8% of the cases, respectively. As expected, 1 H and 13 C NMR methods were most frequently used in experimental tautomer equilibrium studies. Out of more than 40 solvents reported in the publications, chloroform, DMSO and ethanol were the most commonly used solvents. In solvent mixtures, water with ethanol or DMSO was commonly used. The temperature of the experiment was reported for 1,389 cases, out of these 50% and 32% studies were carried out in the range of 251-300K and 301-350K, respectively. No more than 100 entries had pH details reported, with most of those entries coming from a pH range of 0-7 mainly in methanol and water solvents. We hope that this database of experimental data and its included analysis by chemoinformatics methods (by way of annotation with tautomeric transform rules) may provide a set of data useful for future work in the field of tautomerism. This would include tools such as software and chemical identifiers that could be used to avoid tautomeric duplication in chemical databases and compound registration systems. We also hope it may help in developing approaches to predict the most "medicinally" relevant and "reasonable" tautomer forms. This dataset could be a useful training set for machine learning models based on quantum mechanics 32,33 to rapidly identify the lowest energy tautomer.

chemsum

{"title": "Tautomer Database: A Comprehensive Resource for Tautomerism Analyses", "journal": "ChemRxiv"}

data-efficient_machine_learning_for_molecular_crystal_structure_prediction

## Abstract: The combination of modern machine learning (ML) approaches with high-quality data from quantum mechanical (QM) calculations can yield models with an unrivalled accuracy/cost ratio. However, such methods are ultimately limited by the computational effort required to produce the reference data. In particular, reference calculations for periodic systems with many atoms can become prohibitively expensive for higher levels of theory. This trade-off is critical in the context of organic crystal structure prediction (CSP). Here, a data-efficient ML approach would be highly desirable, since screening a huge space of possible polymorphs in a narrow energy range requires the assessment of a large number of trial structures with high accuracy. In this contribution, we present tailored D-ML models that allow screening a wide range of crystal candidates while adequately describing the subtle interplay between intermolecular interactions such as H-bonding and many-body dispersion effects. This is achieved by enhancing a physics-based description of long-range interactions at the density functional tight binding (DFTB) level-for which an efficient implementation is available-with a short-range ML model trained on high-quality first-principles reference data. The presented workflow is broadly applicable to different molecular materials, without the need for a single periodic calculation at the reference level of theory.We show that this even allows the use of wavefunction methods in CSP. ## Introduction The capability to reliably predict the structure of molecular crystals is considered one of the holy grails of molecular modeling. 1,2 Applications for such crystal structure prediction (CSP) methods range from fnding new drugs with improved dissolution properties (and thus bioavailability) to organic semiconductors with novel optoelectronic properties. 3,4 CSP for these molecular materials is so elusive because both their properties and stabilities are critically determined by the interactions of their molecular building blocks in the condensed phase. Indeed, the competition of different interaction types (e.g. dispersion and hydrogen bonding) within molecular crystals often leads to the coexistence of multiple similarly stable crystal structures-so-called polymorphs-each exhibiting different physical properties. 5,6 The ability to predict these polymorphs from simulations would therefore allow the efficient exploitation of the great technological potential inherent in this structural diversity, but requires an unparalleled CSP accuracy/efficiency ratio to explore the vast confguration spaces with highest energetic precision. In practice, this search requires the reliable assessment of the relative stability of different structures, as measured by the lattice energy: where E crys is the total energy of the crystal per unit cell, N is the number of molecules in the unit cell and E iso is the energy of an isolated molecule in its most stable conformation. Here, the main challenge lies in the large number of possible polymorphs and the small energy differences between them. 5,7,8 In practice, there is thus a trade-off between the ability to screen a wide range of candidates (which requires a fast evaluation of free energy or other stability measures) and applying higher levels of theory that adequately describe the subtle interplay between different intermolecular interactions such as H-bonding, electrostatic, induction and dispersion effects. Many CSP approaches are therefore structured hierarchically using a computationally less demanding stability assessment for screening a large set of candidates, while more advanced methods (typically based on density-functional theory, DFT) are used for the fnal ranking of the most promising structures. 9,10 In recent years, a range of methods have been developed for the approximate stability assessment in the initial screening step. Li et al. 11 for instance evaluate stabilities of trial confgurations by applying the Harris approximation to DFT, with crystal electron densities constructed from the superposition of frozen single molecule densities. Tailor-made empirical potentials have also been successfully used for the screening step, as demonstrated for instance by Neumann et al. 12 in the blind tests of organic crystal structure prediction organized by the Cambridge Crystallographic Data Center 9 (CCDC). Finally, semiempirical electronic structure methods like densityfunctional tight-binding (DFTB) have also emerged as promising tools to efficiently rank the stabilities of molecular crystal structures. 13,14 Note that the initial screening can itself be hierarchical, so that the overall CSP workflow often resembles a funnel of increasingly narrow and accurate selection schemes. Nevertheless, regardless of how the most promising candidates are selected, the fnal step of a hierarchical CSP workflow requires an accurate frst-principles method that allows resolving the subtle stability differences between competing polymorphs, presently typically semi-local or hybrid DFT with a many-body dispersion correction (DFT+MBD). 10 There are essentially two sources of error in such hierarchical CSP schemes. First, the initial screening may either not consider the true lowest-energy structure in the frst place or discard it erroneously. Second, the high-level method in the fnal layer may not produce the correct ranking of the remaining candidates. Unfortunately, the obvious solutions to these issues preclude each other: on the one hand, the selection issue can be mitigated by starting with a larger set of candidates and less severe fltering. On the other hand, better ranking can be achieved with more elaborate methods, at a higher computational cost per evaluation. For a fxed computational budget one cannot do both of these things. What is worse, in general it is not clear at the outset which of the two is more critical. A potential way out of this conundrum is offered by modern machine-learning (ML) techniques, which have been found to combine the accuracy required in many chemical applications with affordable computational costs (most of which is associated with the generation of training data rather than the actual application of the potential). In particular, much progress has recently been made in the development of ML-models for high-dimensional potential energy surfaces such as Neural Network Potentials (NNPs) via the Generalized Neural-Network Representation of Behler and Parrinello 18 or the Gaussian Approximation Potentials (GAP) framework developed by Bartók et al. 19 A more comprehensive overview of ML techniques for the generation of interatomic potentials can be found elsewhere. The high flexibility of ML models-which can be considered the reason of their success-can also lead to unphysical results, however, if the model is forced to extrapolate beyond its training set. Consequently, robust and accurate ML potentials are often trained on tens of thousands of confgurations, for which accurate reference data is required. 23 Fortunately, interatomic potentials need not necessarily be created from scratch. Instead, ML models have also been used to improve the description of an underlying baseline. 24,25 Ramakrishnan et al. 26 coined the expression D-ML for this approach and showed that one needs signifcantly fewer training examples in this case, compared to learning a complete interatomic potential. In the context of CSP, there is a further strong argument for D-ML: most ML potentials are inherently local, meaning that the energy is composed of atomic contributions that only depend on the immediate environment of each atom. Yet, intermolecular interactions like electrostatics and (many-body) dispersion can be quite long ranged. A local ML potential will neglect those contributions, whereas a D-ML approach can incorporate them in the baseline model without altering the ML framework. In this paper we therefore develop a D-ML approach to CSP, yielding accurate models for the description of individual molecules and the corresponding molecular crystals. The approach is characterized by high data efficiency, meaning that the workflow is designed to keep the computational effort for training data generation as low as possible. This is achieved by using a robust and computationally efficient baseline method, a diversity-driven selection of training points and the complete avoidance of periodic calculations at the target level of theory (here full-potential DFT with a many-body dispersion correction or spin-component-scaled second order perturbation theory). ## Levels of theory Baseline method. We begin by defning an appropriate baseline method for our approach. Most importantly, this method should be computationally efficient (to allow application to a large set of test structures) and adequately describe the relevant intra-and intermolecular interactions (so as to minimize the required D-ML correction). In particular, it should provide a reasonable description of long-range interactions that are outside the range of the ML model. In our experience dispersion-corrected DFTB methods, in particular using the 3ob parameterization, 27 fulfll these criteria. 3ob is based on the expansion of the DFT total energy up to third-order in density-fluctuations (DFTB3), which provides a sophisticated description of electrostatics, charge transfer and polarization. 28 This leads to marked improvements in the description of organic and biomolecular systems and hydrogen bonding, compared to earlier variants. Since DFTB uses a minimal basis set and tabulated matrix elements, it provides speedups up to three orders of magnitude compared with semilocal DFT. We further apply the Tkatchenko-Scheffler (TS) correction, 14,29 which allows for an accurate incorporation of dispersion interactions at virtually no additional computational cost. Our baseline method is thus defned as DFTB3(3ob)+TS (called DFTB+TS in the following). Target method. The primary high-level target method in this study will be semi-local DFT (using the PBE functional 30 ) with a many-body dispersion correction. 31,32 This method (DFT+MBD in the following) is known to generate lattice energies in good agreement with experiment for the targeted molecular crystals. This can, e.g., be seen by its excellent performance for the X23 database, which contains the experimental lattice energies of 23 crystals (obtained by back-correcting experimental enthalpies of sublimation). 33 Since X23 covers van der Waals (vdW)-bonded, hydrogen-bonded and mixed molecular crystals, this shows that DFT+MBD offers a balanced description of all interactions relevant for CSP. Moreover, relative stabilities of different polymorphs are also described well, as recently demonstrated by Shtukenberg et al. 34 for the rich polymorphism of coumarin. For comparison, the presented scheme is fnally also applied to spin-component-scaled second-order Møller-Plesset theory (SCS-MP2) in one case. 35 D-ML method. We now aim to learn a correction that fxes the shortcomings of our baseline method relative to the target method. This entails, among other things, multi-center contributions to the Hamiltonian, many-body dispersion effects and exchange-correlation contributions inadequately described by the two-center repulsive potential of DFTB. 36,37 To this end, we use Gaussian Process Regression via the Gaussian Approximation Potential (GAP) framework. 19 Kernel methods like GAP use a similarity measure between atomic confgurations (the kernel) to infer the interatomic potential. Here, we use the smooth overlap of atomic positions (SOAP), 38 which is an inherently many-body representation of atomic environments, in line with the types of contributions we want to describe. As noted above, SOAP and related methods use a local representation, meaning that in the fnal D-ML model, all long-range physics are still described at the baseline level of theory. Full details about the ftting procedure are provided in the ESI. † With the above defnitions of the target (DFT+MBD) and baseline (DFTB+TS) methods and the D-ML approach (GAP) used to connect the two, the lattice energy as measure of crystal stability is written as: where DE GAP is the learned D-ML correction. In the following, we further separate this D-ML contribution into intra-(DE GAP(intra) ) and intermolecular (DE GAP(inter) ) contributions. This has both theoretical and practical reasons. Firstly, the energetic contribution of, e.g., stretching a covalent bond is orders of magnitude larger than the contribution of changing the distance between two molecules in a crystal by the same amount. Nonetheless, the intermolecular contributions are arguably much more important for CSP and fnal polymorph ranking, as evidenced by the wide application of CSP protocols with completely rigid molecules. 11,39,40 By ftting separate models, the intermolecular contributions are not overshadowed by the intramolecular ones. Secondly, data generation for an intramolecular correction is very cheap, as it only requires calculations on the gas-phase molecule. It is therefore practical to separate the two training processes. Using this separation, we can rewrite eqn (2) as where the sum runs over all molecules i in the unit cell, and only intramolecular corrections DE GAP(intra) iso appear, of course, for the isolated molecule. ## Training data The generation of training data is a crucial part of constructing any ML model. This data represents all knowledge about the target function that will be integrated into the ft. The required calculations at the target level of theory, however, typically also make this the most expensive part of any ML workflow. It is therefore essential to strike a balance between covering a wide range of confgurations and requiring a manageable number of calculations. To address this issue, we generate a large pool of trial confgurations and subsequently select a maximally diverse subset using the farthest point sampling (FPS) method. 21,41 This entails the iterative selection of confgurations so that each new datapoint is maximally dissimilar to the previously selected structures. In this context, the similarity between confgurations is measured using the averaged SOAP kernel. 42 Clearly, the most straightforward training data for the D-ML correction would be obtained from periodic calculations on the FPS crystals at the target level of theory (DFT+MBD in this case). However, these are precisely the kinds of expensive calculations that we would like to avoid by ftting a D-ML model. Furthermore, it would in principle be interesting to use even higher levels of theory (e.g. Coupled Cluster or Symmetry Adapted Perturbation Theory) as the target method, for which periodic calculations are either impossible or extremely demanding. Fortunately, we found that it is possible to ft accurate D-ML models without using periodic calculations at the target level of theory at all. Specifcally, we use crystal structures as templates to generate molecular clusters (called X-mers in the following), which reflect the diverse relative orientations of the molecules in a crystal, in addition to providing realistic monomer confgurations (see Fig. 1). The idea of using X-mer training data is reminiscent of a many-body expansion (MBE) of the lattice energy. 43 This is, however, notoriously difficult to converge for (polar) organic crystals and liquids, both in terms of length-scale and bodyorder. For this reason, highly accurate MBE-based water models separate the description of long-range electrostatics from short-range interactions. 47 It is therefore highly benefcial to work in a D-ML framework herein, where long-range interactions are covered by the baseline method. Indeed, a ML correction for force-feld lattice energies based solely on two- body terms was recently reported by Day and coworkers. 48 In our work, we found that a pure two-body correction still displays signifcant errors in predicted lattice energies, and thus opted for the X-mer approach. To this end, an initial pool of crystals is generated via the Genarris package. 11 Subsequently, we apply FPS to select 500 maximally diverse structures from this pool. These structures are then relaxed at the baseline level of theory, with fxed unit cells. Afterwards, a second FPS selection is performed on the relaxed crystals to obtain 250 training structures, while the rest are used for testing. Further details about training and test sets are given in the ESI. † Note that the training data for the intramolecular model is, inter alia, further supplemented with monomer confgurations obtained from gas-phase MD simulations (see ESI † for details). ## Model tting Using the above defned training data, we can now train separate GAP models for the intra-and intermolecular corrections. Specifcally, we train the intramolecular correction on energy and force differences: The intermolecular correction is trained on differences in Xmer interaction energies: The index i runs over all X molecules that constitute a cluster. Details about the underlying concepts of SOAP and GAP are provided in the original literature. 19,38,49 A detailed listing of all hyperparameters and computational settings used in this work can be found in the ESI. † ## Results and discussions To illustrate the accuracy and efficiency of our D-ML approach, we will frst separately discuss the accuracy reached for the intra-and intermolecular corrections, relative to their training targets. We then consider the accuracy of predicted lattice energies. For this we employ a representative set of four molecules and their molecular crystals, namely water (H 2 O), pyrazine (C 4 N 2 ), oxalic acid (C 2 O 4 H 2 ) and tetrolic acid (C 4 O 2 H 4 ). ## Model performance: intramolecular D-ML The accuracy of the intramolecular correction is assessed on monomer confgurations extracted from the test and training crystals. Fig. 2 (top) shows the mean absolute error (MAE) of relative energies, compared to the high-level target method (DFT+MBD). For the DFTB+TS baseline, this MAE can be as high as 150 meV (for oxalic acid). This is a serious liability for CSP, where energy differences between polymorphs are often only tens of meV. In contrast, after the D-ML correction, the MAEs are reduced by orders of magnitude. Even in the most challenging case (oxalic acid) the corrected MAE is below 2 meV. Moreover, the good agreement between training and test errors shows that the models are not overftted. For the analysis on the accuracy of forces the reader is referred to the ESI. † As a case in point, the excellent performance of the D-ML correction is confrmed when analyzing the seven predicted conformers of oxalic acid in detail. Indeed, conformer searches are themselves an integral part of molecular CSP studies, as gasphase geometries are typically used as building blocks for the generation of trial crystals. Furthermore, the globally most stable gas-phase conformer is of special interest as the lattice energy is measured relative to it. Fig. 3 compiles the ranking of these seven conformers obtained at the different levels of theory, where we follow the nomenclature proposed in the literature 50 and refer to the conformers with a capital C (cis) or T (trans) depending on the relative orientation of the carboxylic acid groups, framed by lowercase c or t indicating whether the hydrogen atoms point to the inside or the outside. For the twisted conformer, where this nomenclature is not applicable, we use the symbol X. For this highly sensitive test case, the D-ML method fully reproduces the energetic ordering of the target DFT+MBD method-which in turn is in agreement with the literature. 50,51 In contrast, the baseline DFTB+TS energies differ signifcantly and not even the lowest-energy conformer is correctly identifed (reflected by the negative relative energy). In particular, DFTB+TS erroneously predicts most conformers to be rather close in energy, which could have severe consequences for an intended use as an initial screening method. It is further revealing to consider the quality of the predicted geometries (see Fig. 3, bottom). For each conformer, the differences between geometries optimized with the low-cost methods (DFTB+TS or DFTB+TS+GAP) and the respective DFT+MBD reference is measured in terms of their root-mean-square deviation (RMSD). Similarly to the energies, the GAPcorrection strongly improves the RMSD of all conformers-in most cases by more than an order of magnitude. At the same time it can be seen that DFTB+TS alone already provides quite accurate geometries in most cases. Here, the GAP correction cures only some subtle structural differences with respect to the DFT+MBD reference, as can be seen from the cTc overlay in Fig. 3, where the C-O-H angle in DFTB+TS is slightly too large. The exception to this is the tCt conformer. Here, DFTB+TS predicts a considerably different structure, which is brought into excellent agreement with the reference by the GAPcorrection. This is again illustrated by the overlayed geometries, where DFT+MBD and DFTB+TS+GAP are almost indistinguishable. ## Model performance: intermolecular D-ML To evaluate the accuracy of the intermolecular D-ML contribution, we consider the intermolecular energies of X-mers, which are the training targets of this correction (see ESI † for a corresponding analysis of crystals). To this end, we consider X-mers of various sizes, again obtained from the training and test crystals. Interestingly, tetrolic and oxalic acid show slightly larger MAEs, compared to pyrazine and water. We speculate that this is due to the higher flexibility of these molecules (see e.g. the oxalic acid conformers of Fig. 3), which causes a more diverse range of intermolecular arrangements. Overall, the GAP correction nevertheless improves the MAE per molecule by an order of magnitude (except for the tetrolic acid case, where the pure DFTB+TS description already yields a low MAE of around 20 meV per molecule). ## Lattice energies So far, we have analysed the accuracy of the intra-and intermolecular corrections on their respective training targets, and found large improvements relative to the baseline. However, the goal of the proposed method is to improve the description of crystal lattice energies. To evaluate this, we now benchmark the baseline and D-ML methods against the DFT+MBD target method for the lattice energies of molecular crystals. We again consider the crystals used to generate training and test sets separately. Note however, that even for the "training" crystals, the lattice energies were not used to ft the models. In this sense, all predictions in this section can be considered a validation of the D-ML model. Note that the lattice energies are referenced to the global gas-phase minimum of the molecule, calculated with the respective method. In the case of oxalic acid, the DFTB+TS lattice energies are therefore given with respect to a different gas-phase geometry. The results are summarized in Fig. 2 (bottom). This fgure shows that the improved description of intra-and intermolecular interactions also translates to an improved description of lattice energies, as expected. Specifcally, the MAEs of the D-ML model lie between 12 and 24 meV per molecule, which in most cases corresponds to about an order of magnitude improvement. The exception is again tetrolic acid, which is already well described at the DFTB+TS level (but still improved by the GAP correction). These small MAEs also confrm our initial assumption, namely that the DFTB+TS baseline we employ adequately describes long-range interactions. This is further substantiated by considering the intermolecular contributions to the lattice energy separately, as shown in the ESI. † From a CSP perspective, the lattice energies are arguably less important than the energetic ordering of the crystal structures, since we are more interested in which is the most stable crystal, rather than how stable it is in absolute terms. Fig. 4 therefore also includes the coefficients of determination (R 2 ) for the ranking order of the structures, which maps the correlation between reference and predicted data in a range between 0 (no correlation) and 1 (perfect correlation). Again, these are significantly improved by the GAP correction, with values between 0.967 and 0.995 indicating an excellent correlation between the energetic orderings of our D-ML model and the DFT+MBD target. Importantly, errors for test crystals and the ones that (implicitly) enter the training are also in excellent agreement. This indicates a good generalization of the D-ML models beyond their training sets, also for the application to periodic systems. It is further notable that the MAEs for the baseline method are consistently larger for the training than the test set. This confrms that the workflow for training data selection leads to a set of particularly challenging and diverse systems. This can also be seen from the lattice energy correlation plots in Fig. 4, where the training structures cover the full range of lattice energies. In this context, it should be noted that the sampled range covers both negative and positive lattice energies. Although the focus of CSP is obviously on the systems with the most negative lattice energies, there are many trial crystals that need to be evaluated in the process. As these are not necessarily stable, creating a model that covers both ranges is actually desired, not least to be able to confdently discard unstable structures. Fig. 4 provides more detailed insight into the performance of the baseline and D-ML models for the individual systems. As mentioned above, the baseline already provides a reasonable description of tetrolic acid. Nonetheless, there is signifcant scatter in the DFTB+TS correlation plot, which is also reflected in the energy ranking. Here, the GAP correction accounts for the subtle differences between baseline and target, leading to signifcant improvement. In contrast, the lattice energy correlation plot for pyrazine displays a large systematic error, reflected in an erroneous slope (and consequently a large MAE). This deviation can be traced back to the fact that, for this system, unfavourable intermolecular interactions are less repulsive at the baseline level, compared to DFT+MBD (see ESI †). These systematic errors do not affect the ranking, however, which is in good agreement with DFT+MBD (R 2 ¼ 0.944). The GAP correction is able to correct the systematic error in the lattice energies, leading to a strongly improved MAE. Importantly, however, the correction also further improves the energy ranking (R 2 ¼ 0.989). For water and oxalic acid, we observe both systematic errors and signifcant scatter in the predictions of the baseline method. Here, the GAP corrections need to account for a mixture of different effects simultaneously. The lattice energy correlation plots indicate different types of systematic deviations for these systems. While the slope for the water lattice energies is too small, oxalic acid additionally shows an offset of roughly 200 meV with respect to the DFT+MBD values. As with pyrazine, the erroneous slopes are explained by a systematic underestimation of repulsive intermolecular interactions (see ESI †). Meanwhile, the offset for oxalic acid is due to differences in intramolecular interactions at the baseline and target levels (compare Fig. 3). Here, the different predicted global minimum conformers result in a discrepancy of the intramolecular contributions to the lattice energy. As shown in Section 3.1 the GAP correction is very well suited to account for this situation. More generally, the GAP corrections lead to strongly improved lattice energies and ranking orders for both systems. To quantify the error introduced by the X-mer approach, we further created an alternative set of D-ML models (see ESI †). Here, the intermolecular corrections were trained on FPSselected crystals instead of the X-mers. Compared to the Xmer approach, these models display slightly improved lattice energies for most cases (by 4-6 meV per molecule) and are slightly worse in one case. The error incurred by the X-mer approach is thus small or non-existent for the systems considered herein. ## Crystal structure prediction To allow for a pointwise comparison of interaction potentials, the lattice energies in the previous section were computed via single point energy evaluations for frozen geometries (relaxed at the baseline level). Indeed, this strategy has also been employed in 'real' CSP applications. 14 However, the results in Section 3.1 show that the DFTB+TS baseline used herein can yield signifcantly erroneous geometries. This is an uncontrolled source of error, which will propagate through the entire CSP workflow. Fortunately, GAP models are differentiable, so that geometry relaxations at the D-ML corrected DFTB+TS+GAP level are also possible, at essentially no added cost. In this section, we will illustrate the beneft of this feature. For this purpose, we consider target XXII of the most recent blind test of organic CSP. 9 It corresponds to the crystallized form of the tricyano-1,4-dithiino[c]-isothiazole (C 8 N 4 S) molecule. Notably, the six-membered ring in this molecule can be hinged, which induces a chiral-like character to the molecule and, thus, affects the number of space groups allowed in the solid state. A D-ML model for target XXII was generated following the method detailed in Section 2. All results discussed in the following are for randomly generated trial crystal structures not included in the training process. Additionally, the known experimental crystal structure of the molecule is included, 52 to test whether it would have been correctly identifed. Unlike in the previous section, all trial structures are relaxed at the baseline DFTB+TS and D-ML corrected DFTB+TS+GAP levels of theory, and validated with single point calculations at the target DFT+MBD level (see ESI † for an analysis as in Section 3.3). Fig. 5 shows the corresponding lattice energy correlation plot, as well as the ranking order. The most striking feature of the lattice energy plot is a large offset between the baseline and target predictions. Similar to the oxalic acid case, this is-at least partly-explained by deviations in the intramolecular descriptions. DFT+MBD favours the two symmetry-equivalent conformations that exhibit a kink in the six-membered ring. Fig. 6 shows the DFT+MBD minimum energy path for the interconversion of these structures, obtained from a nudged elastic band (NEB) calculation. Here, the flat conformation of the molecule is found to be a saddle point, in agreement with previous reports. 9 This profle changes dramatically when the minimum energy path is reevaluated with the baseline DFTB+TS method: the barrier turns into a broad valley. In fact, the gas-phase optimum found with DFTB+TS corresponds to the flat conformer, as can be seen from the overlay on the right-hand of Fig. 6. In combination with additional geometric deviations (e.g. a more acute C-S-N angle of the fve-membered ring), this causes an energy difference of 670 meV between the gas-phase minima of the baseline and target methods (when evaluated at the DFT+MBD level). As can be seen in Fig. 5 and 6, the D-ML correction cures these discrepancies and largely eliminates the offset. More importantly, the correction also strongly improves the correlation in the energy ranking and correctly identifes the experimental structure to be the most stable. In the CSP context, the most pertinent comparison of the two methods is provided by the ranking order plot in Fig. 5. Here, the baseline method displays a large scatter, with some structures that are deemed among the most stable by DFT+MBD being assigned high ranks (and vice versa). This results in a low coefficient of determination of 0.483. In contrast, the energetic ordering predicted by the D-ML model correlates very well with the DFT+MBD reference (R 2 ¼ 0.907). This good agreement makes DFTB+TS+GAP a very promising method for CSP, particularly as a pre-screening method in hierarchical schemes. In this context, the most stable structures from the prescreening would be further investigated with highly accurate (and expensive) methods, e.g. including vibrational contributions to the lattice free energy at the DFT+MBD level. To illustrate the benefts of the GAP correction for this purpose, the ranking plot in Fig. 5 As mentioned above, the experimentally determined crystal structure is indeed found to be the most stable structure at the D-ML level. Furthermore, the corresponding D-ML geometry is also found to be the most stable at the DFT+MBD level. In contrast, the baseline method predicts several other structures to be more stable than the experimental one. Critically, the experimental structure is not even the lowest energy one when DFT+MBD single point calculations are performed on DFTB+TS geometries. This is again due to signifcant deviations in the predicted geometries of DFTB+TS. Meanwhile there is excellent agreement between the predicted DFTB+TS+GAP crystal structure, and the one relaxed at the DFT+MBD level (see ESI †). Finally, we return to the question of computational efficiency. As stated above, the main motivation for the presented D-ML approach is to avoid the large computational effort of calculations at the target level of theory. Most importantly, the savings of the D-ML model at prediction time should signifcantly outweigh the cost of generating the training data. To this end, the computational effort for generating the D-ML models and performing 10 000 crystal relaxations (a reasonable number for a CSP application) 9 is shown in Fig. 7. It can be seen that the cost of the training procedure is almost exclusively determined by reference calculations at the target level of theory (in particular for the X-mers). For comparison, a D-ML model that exclusively uses the underlying crystals instead of X-mers requires ca. 5000 CPU hours for performing DFT+MBD reference calculations. At this level of theory, the cost for training with periodic crystal data is thus actually somewhat lower than with the X-mer approach. Note, however, that the accuracy of this model is actually slightly inferior to the X-mer approach (see ESI †). Furthermore, the growth in computational costs when including more training data will be steep, especially when considering higher reference levels of theory, as shown below. the D-ML models are roughly equivalent to the cost of explicitly relaxing just seven crystals at the DFT+MBD target level-an insignifcant number compared to the requirements of a fullblown CSP study. ## Crystal structure prediction beyond density functional theory Dispersion corrected semi-local DFT is known to be quite accurate for noncovalent interactions, but it nevertheless displays some pathologies that can be problematic for CSP. 53 Most prominently, the self-interaction error in most functionals causes the over-delocalization of electrons, which leads to errors in the description of electrostatic potentials and charge transfer. 54 In contrast, correlated wavefunction (WF) methods do not suffer from this problem. Furthermore, with these methods convergence to the exact result is, at least in principle, possible. Consequently, there has been much interest in applying WF theory to molecular crystals. This has been prohibitively expensive until recently, but new algorithms and hardware have made some benchmark calculations possible. In this context, highly accurate (sub-kJ mol 1 ) lattice energy predictions have been demonstrated, e.g. by Yang et al. 43 via a fragment strategy and by Zen et al. via diffusion quantum Monte Carlo. 58 While this highlights their potential for CSP, applying such methods to periodic systems is still far from routine and will not be feasible in a highthroughput context for the foreseeable future. The X-mer approach presented herein does not require periodic reference calculations, however, and thus opens the door to WFbased CSP. To illustrate this, a modifed version of the model from the previous section was developed, for which the intermolecular GAP was trained using spin-component-scaled second-order Møller-Plesset theory (SCS-MP2). 35,59 This highlights an additional feature of the presented approach, namely that different reference methods can be used for the intra-and intermolecular models. This can be particularly useful for flexible molecules, where an accurate prediction of torsional barriers, e.g. at the CCSD(T) level, may be required. 53 To evaluate the new intermolecular model, the interaction energies for a test set of X-mers was considered. This reveals a MAE of 7 meV, slightly lower than the one obtained with the DFT+MBD reference (see ESI † for details). The corresponding full model was then used to relax the 251 trial crystals used in Section 3.4. While no periodic MP2 data is available for benchmarking in this case (for the reasons outlined above), the model correctly identifes the experimental geometry to be the most stable (see ESI † for details). The possibility of crystal relaxations with the ML model is particularly attractive in the context of WF methods, where gradients are much more expensive than single-point energy evaluations. 60 As a fnal note, it should be mentioned that SCS-MP2 is not necessarily more accurate than DFT+MBD for this application. While the former offers a better description of electrostatics and Pauli repulsion (because the method is self-interaction free), the latter offers a true many-body description of dispersion, which is lacking at the (SCS-) MP2 level. 61 Nonetheless, this example demonstrates that the presented scheme can be used to apply correlated wavefunction methods in a CSP context. The computational costs to produce the SCS-MP2 X-mer training data lies at 190 000 CPU hours, while the direct application of SCS-MP2 for crystal relaxations in a molecular CSP study is simply not feasible. ## Conclusions In this work, we have presented a computationally efficient and accurate D-ML approach to CSP, using a low-cost baseline (DFTB+TS) that adequately describes long-range interactions. The method is characterized by addressing intra-and intermolecular corrections separately and features a high efficiency in terms of training costs. In particular, this is achieved by selecting diverse training confgurations and completely avoiding periodic calculation for training data generation. The overall accuracy of lattice energies and relative stability rankings has been demonstrated on a representative set of test systems. Importantly, the approach yields models that allow for reliable structure relaxations, with a computational effort that is orders of magnitude smaller than the high-level target method (PBE+MBD or SCS-MP2), even taking training costs into account. To the best of our knowledge, this is the frst generally applicable ML approach that allows structure relaxations in the context of CSP. This opens the door to a CSP workflow that allows screening large candidate pools with unprecedented accuracy. We further note that the accuracy of the D-ML can, in principle, be further refned by including more data. Beyond this, the fact that no periodic calculations are required means that higher levels of theory, such as hybrid DFT or (correlated) wavefunction methods, can be used as the target method. Finally, having a differentiable model also allows the calculation of vibrational zero-point and free energy contributions to the crystal stability. This will be explored in future work. ## Computational details All DFT calculations were performed with FHI-aims, 62 using the PBE functional, 30 tier2 basis sets, tight integration grids and the MBD dispersion correction. DFTB3 calculations were performed using DFTB+ 63 together with the 3ob parametrization 27 and TS dispersion correction. 14,29 For periodic calculations at both levels of theory, the k-grids were converged to obtain energetic accuracies of 1.5 meV per atom. SCS-MP2 (ref. 35 and 59) calculations were performed with ORCA 64,65 using the resolution of identity approximation. 66 GAP potentials were trained and evaluated with the QUIP package. 49 Candidate crystal structures were obtained with the Genarris package. 11 Additional tasks such as FPS and hyperparameter optimization were performed with the MLtools package available at https://github.com/ simonwengert/mltools.git.

chemsum

{"title": "Data-efficient machine learning for molecular crystal structure prediction", "journal": "Royal Society of Chemistry (RSC)"}

secondary_batteries_with_multivalent_ions_for_energy_storage

## Abstract: The use of electricity generated from clean and renewable sources, such as water, wind, or sunlight, requires efficiently distributed electrical energy storage by high-power and high-energy secondary batteries using abundant, low-cost materials in sustainable processes. American Science Policy Reports state that the next-generation "beyond-lithium" battery chemistry is one feasible solution for such goals. Here we discover new "multivalent ion" battery chemistry beyond lithium battery chemistry. Through theoretic calculation and experiment confirmation, stable thermodynamics and fast kinetics are presented during the storage of multivalent ions (Ni 2+ , Zn 2+ , Mg 2+ , Ca 2+ , Ba 2+ , or La 3+ ions) in alpha type manganese dioxide. Apart from zinc ion battery, we further use multivalent Ni 2+ ion to invent another rechargeable battery, named as nickel ion battery for the first time. The nickel ion battery generally uses an alpha type manganese dioxide cathode, an electrolyte containing Ni 2+ ions, and Ni anode. The nickel ion battery delivers a high energy density (340 Wh kg −1 , close to lithium ion batteries), fast charge ability (1 minute), and long cycle life (over 2200 times).The use of electricity generated from clean and renewable sources, such as water, wind, or sunlight, requires efficient distributed electrical energy storage by high-power and high-energy secondary batteries using abundant, low-cost materials in sustainable processes 1 . The secondary batteries capable of storing enormous electric energy at a very large power are of importance for our society. Battery, whose chemistry is based on cathodic and anodic reactions occurring at the interface between the electrodes and electrolyte, generally composes of a cathode, an anode, an electrolyte and a separator 2 . Secondary battery is rare in battery industries because it is difficult to gather two electrochemically reversible cathodic and anodic reactions in one electrolyte as the battery chemistry. There are only several kinds of secondary (rechargeable) batteries in the world: lithium, lithium ion (LIB), sodium ion, nickel cadmium (Ni-Cd), lead-acid, magnesium, calcium and aluminum batteries 1,[3][4][5][6][7][8][9][10][11][12] . Most of the current batteries, for example lithium ion batteries, utilize univalent ions (i.e. H + , Li + , Na + or K + ) as media to store energy. Moreover, most of them are incapable of fast charge 8,13 . Here, we show "how to discover the secondary battery chemistry with the multivalent ions for energy storage" and report a new rechargeable nickel ion battery with fast charge rate. There are three steps for the fabrication of a battery. Firstly, we need choose two reversible reactions in one electrolyte as the cathodic and anodic reactions, respectively. Secondly, suitable cathode and anode materials are required to carry out these reactions. Finally the rechargeable battery is fabricated with certain cathode, anode, electrolyte, and a separator.The multivalent ions, for example Mg 2+ or Al 3+ ion, are used for energy storage to fabricate magnesium or aluminum battery [10][11][12][14][15][16][17] . The investigation on the reversible intercalation of Mg 2+ ions into Chevrel phase such as Mo 3 S 4 indicated an extremely slow intercalation kinetics and low charge capacity 17 . It is generally believed that the kinetics of insertion of multivalent ions into host solid state electrode is much slower than that of univalent ions. While in MnO 2 -based supercapacitors the charge rate and capacitance of the host electrode material (for example MnO 2 ) can be doubled by using multivalent Ca 2+ ion compared with univalent Na + ion . By using trivalent La 3+ ion, the charge rate and capacitance of MnO 2 can be further improved 22 . The research in the supercapacitor application inspired us that we can set up suitable systems for multivalent ions to obtain high intercalation capacity and a fast charge rate. We absorbed the idea and use it in battery field to invent rechargeable batteries with high energy density and fast charge rate. We realized this idea by using the insertion of multivalent Zn 2+ or Ni 2+ ion into alpha type manganese dioxide to invent two rechargeable batteries with a very fast charge rate 23 . In this manuscript, we report the energetic nickel ion chemistry and nickel ion battery for the first time. The nickel ion battery generally uses an alpha type manganese dioxide cathode, an aqueous NiSO 4 electrolyte, and Ni anode. The nickel ion battery displays a high energy density (340 Wh kg −1 , close to that of lithium ion batteries), fast charge ability (1 minute) and long cycle life (over 2200 times). ## Results The common view that the multivalent ion is unsuitable for energy storage at a fast rate is not correct. Below we show that the storage of multivalent ions in certain host material with a large tunnel structure is feasible both thermodynamically and kinetically. Table 1 shows that the ion dynamic diameters of the multivalent ions are close to that of the univalent ions. It enables the probability of multivalent ions to be stored in alpha type manganese dioxide (α -MnO 2 ) with a tunnel diameter of 0.28 nm. We simulated the insertion of one ion in several possible positions (such as 2a, 2b, 4e, 8 h' and 8 h) of one α -MnO 2 tunnel by first principle calculation (Fig. 1) 9,25 . The summary of Li + , Na + , K + , Ni 2+ , Zn 2+ , Mg 2+ , Ca 2+ , Ba 2+ , or La 3+ ion inserting in various positions can be seen in Table S1. The Li + , Na + , Zn 2+ , Mg 2+ , and Ca 2+ ions favour inserting in the 8 h position of the tunnel, while Ni 2+ ion favour the 2a position and K + , Ba 2+ and La 3+ ions favour the 2b position. The bond length of A-O ranges from 1.83 to 2.89 for different ion species (Table S1). When an ion inserts into the tunnel of α -MnO 2 , the tunnel slightly changes and the electrons are stored and shared by adjacent Mn and O atoms. The lowest binding energy (△ E) after ion-insertion for each ion is listed in Table 1. The lower binding energy of multivalent "ion-insertion" indicates that insertion of multivalent ions is thermodynamically easier and more stable than insertion of univalent ions 9,25 Moreover, the kinetics of charge rate by using multivalent ion is faster than using univalent ion, which is directly seen as nD 0 in Table 1. D 0 represents the diffusion coefficient and n is the valence state of these cations. Most importantly, compared with univalent ion, the multivalent ions have the advantage that each multivalent ion inserting into α -MnO 2 host results in the charge storage of over one electron (Figure S6). The lower required energy and faster charge rate of multivalent ions inserted into α -MnO 2 enable the high capacity and fast charge ability for energy storage, which is also consistent with the experimental results in literature 18, . In the first time, we experimentally found that each multivalent Ni 2+ , Zn 2+ , Mg 2+ , Ca 2+ , and Ba 2 ion can electrochemically insert/extract into/from the tunnels of α -MnO 2 as cathodic reaction. where A n+ represents Ni 2+ , Zn 2+ , Mg 2+ , Ca 2+ , Ba 2+ , or La 3+ ion and n is the charge number 18, . The electrolyte refers to the aqueous solution of each multivalent ion with pH value ranging from 4.0 to 7.0, for example 1 mole per liter (mol L −1 ) Ba(NO 3 ) 2 or NiSO 4 aqueous solution. Figure 2 shows the storage of each multivalent ion in α -MnO 2 . It can be seen that Ni 2+ ion can insert/extract into/from α -MnO 2 within a potential window ranging from 0.2 to 1.3 V (vs. NHE) reversibly. Insert of Fig. 2 shows the charge/discharge curves of reversible insertion/extraction of Ni 2+ ion into/from α -MnO 2 . The results regarded to Zn 2+ , Mg 2+ , Ca 2+ , Ba 2+ , or La 3+ ion have already been published and showed that the capacitance and charge rate of α -MnO 2 could significantly double when replacing univalent Li + , Na + , or K + ion with them . In this paper, the fast reversible insertion/extraction of Ni 2+ ion into/from α -MnO 2 is firstly investigated. Most importantly, we use multivalent Ni 2+ ion to invent a new rechargeable battery, named as nickel ion battery. Apart from α -MnO 2 , MnO 2 samples with other typical tunnel structures, for instance, β -MnO 2 , γ -MnO 2 and δ -MnO 2 , are also explored to store Ni 2+ (See Part 1 in SI). However, a reversible intercalation process of Ni 2+ ions only occurs in α -MnO 2 due to its large and unique tunnel structure. We further confirmed the storage of Ni 2+ ions into α -MnO 2 by X-ray photoelectron spectroscopy (XPS) and element mapping measurements. In order to explore the storage of Ni 2+ ions into α -MnO 2 , individual α -MnO 2 electrode is discharged from 1.25 to 1.1, 0.6, 0.3 V (denoted as M1, M2, and M3) by XPS and element mapping measurements, respectively. XPS analysis on the M1, M2, and M3 electrodes is employed to monitor the valence change of Mn and the molar ratio between nickel and manganese element after discharging (See part 2 in SI). The XPS survey of M1, M2 and M3 electrodes is shown in Fig. 3a. Figure 3b exhibits the Mn 3 s core levels for as-prepared, M1, M2, and M3 electrodes. The difference values of peak energies of the Mn 3 s are 5.08, 5.10, 5.32 and 5.43 eV for as-prepared, M1, M2, and M3 electrodes, respectively. These values are compared to 5.79, 5.50, 5.41, and 4.78 eV for reference samples of MnO, Mn 3 O 4 , Mn 2 O 3 , and MnO 2 , respectively 9,23 . A clear change of peak splitting of Mn 3 s was obtained during discharging, which indicates that the oxidation state of manganese changed from IV in oxidized state to III in reduced state. This plot shows that the average valence state of Mn decreased when the MnO 2 electrode discharged, which means that a part of Mn(IV) ions is reduced to Mn(III) ion and the energy (electrons) is stored. Moreover, the oxidation state of alpha-MnO sample synthesized by the co-precipitation technique is rod-like shape with tens of nanometers in diameter. During discharging process, the electrode Ni ions are stored in MnO 2 . Therefore, results from XPS and TEM element mapping measurement directly demonstrated that Ni 2+ ions are stored in α -MnO 2 . Furthermore, X-ray diffusion (XRD) measurement is used to monitor the structure change of MnO2 during the storage process of Ni 2+ ions. It suggests that the main infrastructure of MnO 2 has not been changed dramatically with the insertion of Ni 2+ ions (Figure S7). ## Discussion Generally, a battery composes of a cathode, an anode, an electrolyte and a separator. And battery chemistry is based on cathodic and anodic reactions occurring at the interface between electrode and electrolyte. We firstly discover the fast reversible insertion/extraction of Ni 2+ ion into/from α -MnO 2 in 1 mol L −1 NiSO 4 aqueous solution as shown in equation 2. Therefore, the α -MnO 2 electrode and 1 mol L −1 NiSO 4 aqueous solution can be used as the cathode and the electrolyte, respectively. It also indicates that the reaction of equation 1 can be the cathodic reaction. Then reliable anode as well as anodic reaction has to be found for each multivalent ion to fabricate a full battery. Metal anode is the typical anode used in secondary battery chemistry. In the aqueous electrolyte, Ni metal is suitable for the anode. Deposition/ dissolution of Ni 2+ /Ni occurs at ca. − 0.23 V vs. NHE in 1 mol L −1 NiSO 4 aqueous electrolyte (red line in Fig. 6a). The deposition of Ni 2+ ion on Ni metal is found to be spherical shape with tens of nanometers in diameter (Figure S8). With the α -MnO 2 as cathode, Ni metal as anode, and 1 mol L −1 NiSO 4 aqueous solution as electrolyte, we can invent a rechargeable battery. The battery chemistry is based on two electrochemically reversible cathodic and anodic reactions (equation 2 and equation 3). The potential ranges of cathodic reaction and anodic reaction are showed in Fig. 6a. Subsequently, the energetic nickel ion chemistry as shown in Fig. 6b is proposed by using Ni 2+ ion as the energy storage medium. Nickel ion battery composes of an α -MnO 2 cathode, a nickel metal anode, and a NiSO 4 mild aqueous electrolyte. During discharging process, anodic nickel is dissolved in the form of Ni 2+ ion, which then inserts into α -MnO 2 cathode, generating an electron current flow in the electrical loop and vice versa. Because the storage/release of energy is based on the migration of Ni 2+ ion between cathode and anode, we name this battery as nickel ion battery (NIB). In addition, Fig. 6 also shows the information of zinc ion battery (ZIB) to compare with NIB. ZIB is assembled by using α -MnO 2 cathode, Zn anode and ZnSO 4 aqueous electrolyte 23 . It can be seen from Fig. 6 that NIB and ZIB are similar in cathode due to the origin of the multivalent ion storage mechanism of α -MnO 2 . However, the electrolyte, anode and the most important battery chemistry are different. We assembled the prototype NIB (See Part 4 in SI) and ZIB (See Part 5 in SI), whose charge/discharge curves are shown in Fig. 6c. The NIB has a maximum operating voltage value of 1.5 V and delivers a capacity up to 298 milliampere hour per gram (mAh g −1 ) calculated based on the mass of MnO 2 . The long cycle life test has been performed on NIBs by the continuous galvanostatic charge/discharge at current densities of 200 milliampere per gram as shown in Fig. 7a. The continuous charge/discharge cycle curves of NIB are showed in the insert of Fig. 7a. After 2200 cycles NIB still shows a stable capacity and good columbic efficiency (Figure S13). The Ragone plots of NIB and ZIB are shown in Fig. 7b. It exhibits that these energy storage devices with multivalent Zn 2+ or Ni 2+ ions for energy storage cover a very wide range from batteries to supercapacitors and fill the gap between them. Batteries with the multivalent ions for energy storage, for example NIB and ZIB, are capable of fast charge/discharge (1 minute) as shown in the insert of Fig. 7a and Figure S15. The energy densities of ZIB and NIB are estimated to be ca. 320 and 340 watt hours per kilogram (Wh kg −1 ) based on the weight of the total active mass of both cathode and anode. In the battery industry, the weight of the active mass of both cathode and anode is 30-50% of the total battery 13 . Therefore, the energy densities of total battery of ZIB and NIB should be up to 170 Wh Kg −1 and ca. 120 Wh Kg −1 , which are higher than lead-acid, or Ni-Cd batteries (ca. 40 ∼ 80 Wh kg −1 ), and close to LIB (ca. 150 ∼ 400 Wh kg −1 ) 2,9,13,26-29 . In addition, NIB shows a longer cycle life over 2200 cycles than current aqueous batteries (ca. 1000 cycles) 2,30 . Furthermore, we have performed the nailing experiment on these batteries at a fully charged state. No sign of flash or smoke has been detected, which indicates the excellent safety. Therefore, these secondary batteries have great advantages in terms of safety, cycle life and energy density over the existing rechargeable batteries. These significant advantages enable the exploration of new industrial applications or to rival lead-acid or nickel cadmium batteries. For example, these secondary batteries are advanced candidates for hybrid electric vehicles or electric bicycles, or for storage of electricity generated from clean and renewable sources, such as water, wind or sunlight. This is just the beginning of searching for new batteries with multivalent ions. With the further exploration we anticipate that more rechargeable batteries with multivalent ions as energy storage will emerge. In summary, we show that storage of multivalent Ni 2+ or Zn 2+ ion in alpha type manganese dioxide presents a stable thermodynamics and fast kinetics, which is confirmed by theoretic calculation and experiment confirmation. We further use storage process of the multivalent Ni 2+ ions to discover a new battery chemistry and invent the rechargeable nickel ion battery. The secondary battery with multivalent Ni 2+ ions for energy storage is advantageous in energy density (340 Wh kg −1 ), fast charge ability (1 minute), and long cycle life (over 2200 times). As multivalent ions are rich in quantity, we believe that the utilization of them may trigger a renaissance of new battery chemistry in the future. ## Methods The MnO 2 powder has been synthesized by a one step process or co-precipitation technique 19 (See Methods in SI). Electrodes were prepared by mixing 70 wt% of MnO 2 or MnO 2 /graphene as active material with 20 wt% of acetylene black (conductive additive) and 10 wt% of binder polytetrafluoroethylene (PTFE). 70 mg of MnO 2 powder and 20 mg of acetylene black were firstly mixed and dispersed in ethanol by ultrasound for 30 min. Then the ink was dried at 80 °C for 4 h to get dark mixed powder and 10 mg of PTFE was added to get a paste with a few of ethanol. Then the paste was dried at 80 °C and a few of 1-methy-2-pyrrolidinone (NMP) were added to get a syrup. The syrup was cold rolled into thick films. Pieces of film with 1 ∼ 5 mg weight, typically 1 cm 2 in size, were then hot-pressed at 80 °C under 100 MPa on a stainless steel mesh. The prototype nickel ion battery with α -MnO 2 cathode, 1 mol L −1 NiSO 4 aqueous electrolyte, a fiber paper separator and a nickel foam anode as shown in Figure S10. The nickel foam is used due to its larger surface area than plate. Electrochemical tests were performed with Solartron 1480 electrochemical station and Land CT2001A equipment. The discharge capacity of the cell is calculated according to the formula: Where C is specific capacity (milliampere•hour per gram, mA•h g −1 ), I is the applied current (milliampere, mA), t is discharge time (hour, h), and m is the mass of the active material (gram). The energy density is calculated by the following equation: Where C is specific capacity (mA•h g −1 ) and V is the average voltage of battery. The average voltage of NIB and ZIB is 0.85 and 1.45 V. The energy density of zinc ion battery (ZIB) and nickel ion battery (NIB) are listed in Table S2.

chemsum

{"title": "Secondary batteries with multivalent ions for energy storage", "journal": "Scientific Reports - Nature"}

sterol_uptake_by_an_alkali-b-cyclodextrin_metal-organic_framework

## Abstract: b-Cyclodextrin is well known in cellular biology for its ability to moderate cholesterol levels in lipid bilayer membranes. Its use in extended network solids remains elusive due to the low symmetry of this macrocyclic system. Self-assembly of two different b-cyclodextrin MOFs with extended nanotube structures is achieved by crystallization with excess potassium hydroxide, one in the presence of cholesterol. We then further demonstrate the proclivity of one of these MOFs to absorb cholesterol and two other sterols from solution using NMR and confocal microscopy techniques. This work demonstrates that these network solids show great potential in both substrate delivery and/or extraction.Cyclodextrins (CDs) are a unique class of material with uses spanning the biological, chemical and materials sciences. The three most common forms of CD are comprised of six (a-), seven (b-), and eight (g-) 1,4-linked pyranose units giving rise to cylindrical-shaped structures with hydrophilic exterior and a hydrophobic core. Their unique three-dimensional shape offer materials scientists several chemical handles for functionalization, and predictable behaviour with the primary and secondary faces of the toroid pointed with equatorially disposed glycosidic 1,3-and 1,2-diols, respectively (Figure 1). [1][2][3][4] Ideally positioned for meta-ligand chelation, the development of coordination networks that incorporate the toroidal motif in a manner that gives rise to extended ordered porosity has received notable interest in recent years. 5 While there is still substantial untapped promise in the use of these sugar-centric network solids (also referred to as metal-organic frameworks; in this case CD-MOFs), to date they have demonstrated limited success beyond carbon dioxide uptake, 6,7 gold ion extraction, 8,9 separating small chiral / aromatic compounds, 10,11 and mediated drug release. 12,13 This is in contrast to myriad of other applications that now employ multi-topic carboxylate-linked MOFs including but not limited to gas-sorption/separation, 14,15 water sorption, 16 catalysis, 17 sensing, 18,19 and drug delivery. 20,21 Few of these systems, however, are derived entirely from renewable or naturally available components, making the pursuit of CD-based MOFs with demonstrable utility particularly important. While the components of CDbased MOFs would be considered naturally occurring and/or derived from renewable resources, very few have demonstrated usefulness within -or interacting with components of-the biological arena. 12,27 We found this particularly intriguing considering the important role that b-CD plays in biochemical research. It has been well documented that b-CD and by extension, methyl-b-CD (MBCD) have become quintessential tools in the mediation of intralamellar cholesterol levels from outside the membrane environment in order to influence cholesterol-dependent cellular processes. 28 For example, MBCD has been used to treat tissue culture cells to control cholesteroldependent budding of influenza viruses, 29 and was separately demonstrated to modulate cholesterol interaction with the in-membrane oxytocin receptor protein. 30 In parallel and of particular importance in medical research, is the study of b-CDs as potential lipoprotein mimics by moderating in vivo cholesterol metabolism for combating atherosclerosis. Through the formation of a [n]pseudorotaxane-style host-guest (HG) inclusion complex, hydrophilic b-CD solubilizes the highly hydrophobic sterol (and others) in aqueous environments with a remarkable association constant of Ka = 1.7 x 10 4 M -1 , as determined by the spectral displacement method. 34 In fact, this is such an effective solubilizing system that a b-CD-cholesterol HG-complex is commercially available from a number of suppliers as 'Cholesterol Water Soluble'. In this account, we report the self-assembly of two new b-cyclodextrin-centered MOFs with apertures that align to form extended nanotubular arrays; one of which includes full characterization due to its broader HG applications (b-CDMOF-1) comprised of b-CD and K + . Separately, crystalline b-CDMOF-2•Chol was grown in the presence of cholesterol and structure confirmed by single crystal XRD, a first for this particular HG complex as an extended network and only the second time as a discrete HG complex. 35 This is surprising considering the ubiquitous use of the b-CD-cholesterol complex in biology and across multiple divisions of chemistry. 36,37 Second, we demonstrate that (b-CDMOF-1 is capable of extracting cholesterol (along with other sterols) from solution into the extended network pores of the sugar-based nanotube structures, and further examine the crystal sponge behaviour by BODIPY-labelled cholesterol and separately resorufin uptake by fluorescence microscopy. Self-assembly of b-CDMOF-1 and -2•Chol was achieved by slow solvent diffusion (either vapour or layering as noted) of methanol into combined aqueous solutions of b-CD (1.0 eq.) and potassium hydroxide (KOH; 20 eq.) in the presence or absence of desired the guest species. Crystal growth of described topologies was highly reproducible, and achieved within a 5-day timeframe. Volatility of the methanol supernatant resulted in rapid solvent loss of the crystals and disintegration of the crystal lattice. Removal of the supernatant and soaking in ethanol for 24 hours followed by two 24-hour dichloromethane soaks afforded starburst crystals stable enough to be easily handled for TGA, elemental analysis, and X-ray powder diffraction (see supporting information). In fact, the DCM soaking protocol was demonstrated by TGA to slightly increase the thermal stability of the coordination assembly (See SI, Fig. S3), though any solvent loss adversely affected the extended crystallinity of the material as seen in the X-ray powder diffraction. Porosimetry experiments were unsuccessful, as the material was not stable enough to withstand the activating conditions required. Nonetheless, surface area was estimated using low level Connolly Surface calculation to be 1250 m 2 g -1 , using the single crystal diffraction data. When the material remained solvated, however, the crystals remained intact and could be easily handled and transferred between vessels. b-CDMOF-2•Chol crystallized as colorless cuboid crystals in a Monoclinic P21 space group with b-CD units aligned to form parallel one-dimensional nanotubes, with primary and secondary faces of the CD toroids again assembled in a head-to-head / tail-to-tail arrangement, stabilized by several complimentary H-bonds at each interfacial junction. Three potassium ions (one of which is partially occupied) participate in inter-nanotube coordination, forming a network of parallel nanotubes along the unit cell's a-axis. The pores of these tubes contain guest cholesterol molecules (1/3 occupancy for each pair of CD host molecules; thus 1:6 cholesterol/b-CD ratio), which have been crystallographically characterized in-situ, as shown in Figure 3. The observation of cholesterol within the pores suggests that these networks may be capable of cholesterol uptake in the form of a crystal sponge. 38 Considering that the conditions for assembly of b-CDMOF-2•Chol mirrored that of b-CDMOF-1, we posit that the presence of cholesterol contributed in the templating of the parallel one-dimensional porous network, an attribute we are currently exploring. Investigation of the structure reveals no significant intermolecular interactions between any cholesterol functionality and the interior walls of the CD channels. Specifically to this structure, we see no hydrogen bond contacts to the free secondary alcohol of cholesterol by any b-CD oxygen. As this is a coordination network and not a solvated intermolecular HG system, the two environments are not comparable but this observation supports conclusions that a driving force for the solvated HG assembly is predominantly by solvophobic Van der Waals' attraction. S5). Single crystal diffraction analysis of these hybrid crystals revealed a single crystal pattern indistinguishable from that of b-CD, superimposed with a powder diffraction pattern of cholesterol originating in the crystallites that had grown on the surface. This indicates that cholesterol does not penetrate into b-CD crystals, rather interacting with the surface only. However, soaking of b-CDMOF-1 under the same conditions resulted in unchanged crystal morphology, presumably because the network solid is being loaded with cholesterol instead of nucleating on the surface. 1 H NMR analysis of the digested solids followed to assess this behaviour. Again, crystals of b-CDMOF-1 were soaked in an ethanolic solution of cholesterol for 24 hours, the supernatant removed, and crystals rinsed twice to dissolve away any surface adsorbed cholesterol with remaining solvent being removed in-vacuo. The b-CDMOF-1 solids were then digested in the chosen NMR solvents to assess host-guest ratios in identifying degree of cholesterol uptake. This analysis revealed an approximate ratio of 2:1 b-CD to cholesterol (Figure 4a) by integration of the respective 1 H NMR signals. We surmise that this ratio is too large to be merely surface adsorption of cholesterol to the crystal surface (particularly after a rinsing protocol), nor do we feel any accessible crystal fracture planes would accommodate cholesterol due to the highly ionic nature of the adjoining space between the nanotubular arrays. This study was also extended to include deoxycholic acid, b-estradiol, and a size-comparable dye molecule named resorufin (Figure 4b). The two related sterols showed similar uptake capacities (3:1 b-CD to deoxycholic acid and 14:1 b-CD to b-estradiol; see SI, Section 6) establishing that b-CDMOF-1 does indeed demonstrate crystal sponge behaviour. The root cause of this host-guest interaction is largely driven by solvophobic effects due to high polarity of the ethanol solvent and low polarity of the b-CD nanotubes. This is comparable to the complexation of free b-CD or MBCD with cholesterol in water (a well-documented system). 34 In contrast, we see no uptake of resorufin by NMR analysis, for which we attribute this to its smaller size (fewer Van der Waal's interactions) and much higher localized polarity. Since 1 H NMR is not direct evidence of cholesterol uptake, and SCXRD on cholesterol-soaked crystals afforded pore contents of intractable disorder, further evidence of this phenomenon was collected using fluorescence confocal microscopy, by employing guest molecules containing strongly emitting fluorophores (Figure 5). To accomplish this, we chose to employ commercial Bodipy-cholesterol (BO-C), which is as the name suggests, a Bodipy-conjugated cholesterol commonly used in cell imaging, 39 and separately, the aforementioned resorufin dye. The distinction between the two emitters is wavelength of emission maximum (507 nm and 586 nm, respectively) and molecular size (Bodipy is a large pendant group, while resorufin is small and initially thought to permeate through the pores). Samples of b-CDMOF-1 were loaded with each respective substrate in accordance with the above NMR analysis procedures, followed by crystal selection for microscopy. The samples were thoroughly rinsed to limit background fluorescence of free substrate and analyzed under a blanket of ethanol to prevent desolvation. Quantitative analysis of b-CDMOF-1•BO-C reveals that emission properties of BO-C are more prominent on the crystal edges (a-axis) due to cumulative intensity of the higher fluorescence signal along the crystal edges (Figure 5a). Here, BO-C can bind on the surface by inclusion of the cholesterol portion, but not totally enter caused by the restrictive size of the Bodipy moiety. Intensity mapping illustrates this phenomenon with some emission intensity along the b-axis (crystal face), but with lower intensity, indication of a single layer (or lower cumulative concentration) of fluorophore. This image visually illustrates the surface inclusion of BO-C. Analysis of b-CDMOF-1•resorufin revealed quite different behaviour (Figure 5b). As noted in the NMR analysis, it was initially posited that the resorufin dye was of sufficient size to be included in the extended b-CD pores, however, this was not observed upon digestion of the MOF material. Confocal analysis revealed that that resorufin appears to permeate into the crystal fracture planes, which would presume to be areas of high polarity (opposite of what would be found inside the b-CD pores. Fluorescence intensity mapping revealed quite regular peak intensities across the crystal surface, indicative of dye being situated within the orthogonal crystal grain dislocations where the sugar oxides and potassium ions reside, rather than being restricted to the surface. This result nicely contrasts that of the inclusion of BO-C and further demonstrates the proclivity of cholesterol uptake in this unique system. In conclusion, we have presented two new cyclodextrin-based MOFs employing b-CD as the structural building unit, one of which containing cholesterol within its pores. We also demonstrate that b-CDMOF-1 is capable of sterol uptake within its non-polar pores, and are able to contrast this be behaviour with similarly-sized dye molecule that exhibited no uptake tendency. This work lays the foundation for our group to develop new MOF technologies related to extraction therapeutics, in this case, towards the combating of Atherosclerosis, and the potential for delivery of steroidal drugs.

chemsum

{"title": "Sterol Uptake by an Alkali-b-Cyclodextrin Metal-Organic Framework", "journal": "ChemRxiv"}

discovery_of_a_new_function_of_curcumin_which_enhances_its_anticancer_therapeutic_potency

## Abstract: Curcumin has received immense attention over the past decades because of its diverse biological activities and recognized as a promising drug candidate in a large number of diseases. However, its clinical application has been hindered due to extremely low aqueous solubility, chemical stability, and cellular uptake. In this study, we discovered quite a new function of curcumin, i.e. pH-responsive endosomal disrupting activity, derived from curcumin's self-assembly. We selected anticancer activity as an example of biological activities of curcumin, and investigated the contribution of pH-responsive property to its anticancer activity. As a result, we demonstrated that the pH-responsive property significantly enhances the anticancer activity of curcumin. Furthermore, we demonstrated a utility of the pH-responsive property of curcumin as delivery nanocarriers for doxorubicin toward combination cancer therapy. These results clearly indicate that the smart curcumin assemblies act as promising nanoplatform for development of curcumin-based therapeutics.Curcumin (CCM), a naturally-occurring polyphenol derived from the turmeric plant, has been received immense attention over the past decades because of its diverse biological activities, including anticancer, antioxidant, anti-amyloid, anti-inflammatory, antidiabetic, antibiotic, and antiviral activities 1-6 . Hence, CCM is recognized as a promising drug candidate in a large number of diseases such as cancer, neurodegenerative diseases, infectious diseases, and diabetes. However, the application of CCM in the therapeutic treatment has been hindered due to three obstacles. The first obstacle is extremely low aqueous solubility of CCM. CCM is hydrophobic molecule, and thus the maximum water solubility is about 30 nM, whereas the required concentration to exhibit various bioactivities is micro molar. Therefore, it is necessary to dissolve CCM in appropriate organic solvent for the use 7 . The second obstacle is chemical instability in aqueous condition. CCM quickly hydrolyze under physiological pH 7.4 in phosphate buffer with a half-life (t 1/2 ) of only 20 min 8,9 . The third obstacle is low cellular uptake. It is demonstrated that CCM tends to deeply insert into the cell membrane through hydrophobic interaction with fatty acyl chains of lipids and hydrogen bonding with the phosphate group in a manner similar to cholesterol 10,11 , and thus limited amounts of CCM diffuse into the cytoplasm, while the main target action site of CCM for the most of its bioactivities is the cytoplasm. Therefore, overcoming these obstacles are the key challenges to achieve clinical application of CCM.To date, several types of approaches have been performed to overcome these obstacles. For example, various kinds of chemical modification to CCM including polymer-conjugation have been performed [12][13][14][15][16][17] . Moreover, various types of nanomaterials including liposomes, polymeric micelles, and silica nanoparticles have been used as carrier materials for CCM [18][19][20][21][22] . In this context, we have adopted a supramolecular self-assembly approach to overcome these obstacles. We have recently reported the synthesis of amphiphilic CCM with several types of molecular architectures through conjugation of CCM with ethylene glycol oligomers and the fabrication of CCM nanovesicles (curcumisome) through supramolecular self-assembly of the CCM amphiphiles via hydrogen-bonding, π -π stacking, and hydrophobic interactions generated among the CCM segments (Fig. 1a). The CCM nanoassemblies exhibited several hundreds of thousands-fold higher aqueous dispersibility than CCM and significantly higher resistance against hydrolysis (t 1/2 : 168 h) than CCM, allowing intravascular administration of CCM at a high dose with keeping CCM's structure intact for several days 23 . Most recently, we serendipitously discovered a new function of the CCM nanoassemblies, i.e. pH-responsive endosomal disruption activity, when we examined cellular uptakes of these CCM nanostructures. In this study, we therefore selected anticancer activity as an example of biological activities of CCM, and investigated the impact of the pH-responsive endosomal disruption activity of CCM on its anticancer activity both in vitro and in vivo. Moreover, we investigated the molecular mechanism of the pH-responsive properties. ## Synthesis and characterization of CCM nanoassemblies. CCM amphiphiles with different numbers of CCM per molecule and hydrophilic/hydrophobic balance, PEG-CCM (PC), PEG-CCM-PEG (PCP), CCM-PEG-CCM (CPC), were synthesized by the same methods with our previous reports (Supplementary Fig. 1) 23 . In this study, we newly synthesized a type of CCM amphiphile with branched architecture, 4-arm PEG-CCM4 (PC 4 ) (Fig. 1b). The characterizations of the CCM amphiphiles estimated by 1 H-NMR (Supplementary Figs 2-5) and GPC analyses were summarized in Supplementary Table 1. Aqueous dispersions of the CCM amphiphiles were analyzed by DLS and TEM analyses (Supplementary Fig. 6). These dispersions showed a unimodal peak, and the averaged diameter was 178 ± 27 nm for PC, 195 ± 21 nm for PCP, and 205 ± 36 nm for CPC, and 79 ± 17 nm for PC 4 , respectively. PC, PCP and CPC nanoassemblies showed spherical shape, while PC 4 nanoassemblies was square-shaped nanoparticles with multiple small CCM domains observed as dark spots in the TEM image. To investigate the architecture of nanoassemblies in molecular level, their 1 H-NMR spectra were measured in D 2 O. As reported in our previous report 23 , the characteristic peaks of CCM segment (5.8-7.6 ppm) and the methylene peaks in PEG segment (3.7 ppm) were clearly observed in CDCl 3 which is good solvent for CCM and PEG segments, while these peaks were completely disappeared in D 2 O and the methylene proton signal of PEG segment was detected as a sharp peak (Supplementary Figs 2-5). These results indicate that hydrophobic domains of CCM locates inside of the nanoassemblies and PEG segments extend out into aqueous environment at the surface of nanoassembly, as illustrated in Fig. 1a. The critical assembly concentration (CAC) values of PC, PCP, CPC, and PC 4 , estimated from the eosin Y assay 24 , was 1.14 ± 10 −6 M, 2.60 ± 10 −6 M, 4.80 ± 10 −7 M, and 3.54 ± 10 −6 M, respectively. As reported in our previous report 23 , CCM amphiphiles with higher CCM contents were prone to show lower CAC, indicating that main driving force to form nanoassemblies is a couple of interactions between CCM segments. These interactions were characterized by FTIR analyses to be intermolecular π -π stacking among aromatic groups and π -conjugated linker in CCM segments, hydrogen bonding among ketone groups in CCM segments, and hydrophobic interaction of CCM segments. ## Cellular uptakes of CCM nanoassemblies. Efficient cytoplasmic delivery of CCM is crucial challenge to enhance the biological activity. We investigated cellular uptakes of CCM nanoassemblies by human cancer cells (PC-3). CCM exhibits green fluorescence in aqueous condition, allowing the monitoring of cellular uptake and intracellular distribution of CCM-based nanoassemblies by fluorescence microscopy. Figure 2a (inhibitor − ) shows CLSM images of PC-3 cells treated with four types of CCM nanoassemblies at equivalent CCM concentration. As a negative control sample, cellular uptake of free CCM was also examined. Although very weak green fluorescence was detected in PC-3 cells after 6 h incubation with free CCM (Supplementary Fig. 7), much strong fluorescence was observed in cells treated with CCM-based nanoassemblies. To characterize the pathway of cell uptakes, endocytosis inhibitors were used to perturb entry. In this study, inhibitors which affect endocytic pathways in a more general manner were used. In particular, clathrin-mediated endocytosis, caveolae/raft-mediated endocytosis, and micropinocytosis were inhibited by chlorpromazine, genistein, and cytochalasin D, respectively. As shown in Fig. 2, cellular uptakes of all types CCM nanoassemblies was drastically reduced by chlorpromazine, while the cellular uptakes was also slightly reduced by cytochalasin D. This result indicates that the CCM nanoassemblies were mainly entered into cancer cell through clathrin-mediated endocytic pathway. Clathrin-mediated endocytosis is the most understood pathway. In this pathway, nanoparticles were transport into cells within vesicles, referred as early endosomes, and the endosomes either were recycled and exocytose out of the cell or maturated within cells. During endosome maturation process, early endosomes are fused with cytoplasmic vesicles and matured into late endosomes and lysosomes 25 . The entrapment of CCM nanoassemblies within endosomes is undesirable because CCM is rapidly degraded in the matured endosomes/lysosomes. Therefore, intracellular fate of CCM nanoassemblies after clathrin-mediated endocytosis was investigated with focusing on endolysosomal network. Endosomes and lysosomes were stained with Lysotracker Red giving a red color, and the nuclei were counterstained with Hoechst33342 giving blue color. Figure 2c shows CLSM overlap images of PC-3 cells treated with CCM nanoassemblies at equivalent CCM concentration. All types of CCM nanoassemblies rapidly entered into the cancer cells and yellow fluorescent spots were observed at early stage (1 h), indicating colocalization of nanoassemblies with endosomes. Importantly, green fluorescence arising from CCM nanoassemblies was also observed in cytoplasm even after early stage, indicating cytoplasmic distribution of nanoassemblies. An obvious increase in the green fluorescence intensity in cytoplasm was also seen over time for all types of nanoassemblies. Moreover, the decrease in the numbers of red and yellow spots was clearly observed over time. These results suggest that the maturation process of early endosomes to late endosomes and lysosomes is interrupted in the presence of CCM nanoassemblies. Additionally, the amount of nanoassemblies in cytoplasm clearly increased over time, indicating release of CCM nanoassemblies from endosomes to cytoplasm. pH-responsive properties of CCM nanoassemblies. The maturation of endocytic vesicles into late endosome and lysosome is characterized by acidification from pH 7.4 to 4.5 within the vesicle 26 . Considering the rapid cytosolic dispersion of CCM nanoassemblies, we hypothesized that the CCM nanoassemblies would have pH-responsive property under endosomal acidic pH condition, and the pH-responsive property could contributes to the cytosolic dispersion. Therefore, we investigated pH-responsive properties of CCM nanoassemblies. We first examined the effect of pH decrease on the zeta potential of nanoassemblies. The average zeta potential of nanoassemblies at different pH values are shown in Fig. 3a. At pH 7.4, PC, PCP, and CPC nanoassemblies had a negative zeta potential ranging from − 35 to − 43 mV, while PC 4 nanoassembly had a positive zeta potential (+ 5 mV). When pH values decreased from 7.4 to 5.5, the gradual increase in zeta potential was observed for all types of nanoassemblies. In particular, PC and PCP nanoassemblies showed more increase in the zeta potential than CPC and PC 4 nanoassemblies. These results suggest that the CCM nanoassemblies have an ability to adsorb H + ions. Considering the significant increase in zeta potential for PC and PCP and the slight increase for CPC and PC 4 nanoassemblies, it is suggested that PC and PCP nanoassemblies adsorb H + ions at the near surface, while CPC and PC 4 nanoassemblies adsorb H + ions inside of nanoassemblies. We next examined pH buffering capacity of CCM nanoassemblies derived from the H + ion adsorption. The acid-base titration curves of the CCM nanoassemblies are shown in Fig. 4b. It is noteworthy that, all types of CCM nanoassemblies had an obvious pH buffering capacity between pH 7.2 to 5.6, corresponding to pH value within endosomes during the maturation. The buffering capacity of PC and PCP was almost same. The buffering capacity of CPC and PC 4 nanoassemblies were 1.5-fold and 1.4-fold higher than PC and PCP nanoassemblies, respectively. The pK a values were estimated to be pH 6.4 for all types of nanoassemblies, indicating that the origin of buffering capacity is the same. It is well known that PEG does not have pH buffering capacity because nonionic PEG can not act as proton acceptor 27 . Considering the report, the proton acceptor of CCM nanoassemblies can be assigned to the CCM segments. CCM displays typical keto-enol tautomerism as illustrated in Supplementary Fig. 8. CCM has three proton acceptable sites per a molecule, two phenolic hydroxyl groups and a diketone group (keto form) or an enolic hydroxyl group (enol form) 3 . PCP molecule, which both two phenolic hydroxyl groups were used to the conjugation with PEG, also showed buffering capacity and the capacity was similar with PC having a free phenolic hydroxyl group. Considering this, it is plausible that the diketone group or enolic hydroxyl group of CCM segments is the possible proton acceptor. In contrast, free CCM in aqueous solution with equivalent concentration at CCM level did not show any buffering capacity (Fig. 3b). These results suggest that the buffering capacity is attributed to the nanoassembly formations of CCM. To assess the thought, we examined pH buffering capacity of CCM-PEG conjugates in water/methanol (70/30, v/v) mixed solution in which CCM-PEG conjugates do not form nanoassemblies. Supplementary Fig. 9 shows the acid-base titration curves of the CCM-PEG conjugates with equivalent concentration at CCM level. As we expected, CCM-PEG conjugates did not show any pH buffering capacity and the profiles were as same with free CCM. Importantly, it is thus concluded that the nanoassembly formations of CCM is an essential for exhibition of the pH buffering capacity. The equilibrium between keto form and enol form of CCM is dependent on the molecular environment, including pH in the solution and its states of matter (solid or liquid) 28 . As described above, the diketone group and enolic hydroxyl group of the CCM segments in nanoassemblies is potential proton acceptor related to the pH buffering capacity. It was reported that the keto form dominates at acidic or neutral pH, while the enol form is favored at alkaline pH. The pK a value of enol group and two phenolic hydroxyl groups in aqueous solution is 8.38, 9.88, and 10.51, respectively 29 . The facts indicate that most of enolic hydroxyl groups is protonated at neutral pH and thus cannot act as proton acceptor at acidic pH. On the other hand, Akulov and co-workers have reported that diketo oxygen atoms in CCM adsorb a H + ion in weak acidic condition and the protonated forms are a stable six-membered ring as shown in Supplementary Fig. 8 30 . Moreover, Basnet and co-workers have reported that the keto form dominates in solid state, and the keto form acts as a potent proton acceptor in the solid state 31 . Consequently, the proton acceptor for the pH buffering capacity of CCM-based nanoassemblies can be assigned to diketone group of the CCM segments. Thus, it was demonstrated that the formation of supramolecular nanoassembly allows CCM molecules to maintain keto form through the formation of solid state CCM domains in nanoassemblies, and the pH buffering capacity was successfully achieved. This is the first discovery on pH buffering capacity of CCM. Based on the above results, the CCM nanoassemblies probably exhibit the "proton sponge effect" in endosomes/lysosomes. Then, we investigated the effects of proton adsorption of the CCM segments under endosomal acidic pH on the size and morphology of CCM nanoassemblies by DLS and TEM analyses. Figure 3c shows the impacts of pH change on the size of CCM nanoassemblies based on DLS analyses. The size of CPC and PC 4 nanoassemblies obviously increased with the decrease in pH from 7.4 to 5.5, while no changes were observed for PC and PCP nanoassemblies. Figure 3d shows TEM images of CPC and PC 4 nanoassemblies incubated under pH 7.4 and pH 5.5, respectively. CPC and PC 4 nanoassemblies at pH 7.4 had spherical and square-shaped morphology with average diameter of about 205 nm (CPC) and 79 nm (PC 4 ), respectively. However, after incubation under pH 5.5, the diameter significantly increased up to 310 nm (CPC) and 450 nm (PC 4 ), respectively. Moreover, the remarkable morphology change (swelling) of both CPC and PC 4 nanoassemblies were observed, and uniformity in the size and the morphology was disordered. As mentioned above, CPC and PC 4 nanoassemblies adsorb H + ions to the CCM domains located inside of nanoassemblies, while PC and PCP nanoassemblies adsorb H + ions at the near surface. Under the physiological environment (pH 7.4), most diketone groups of CCM segments in all types of nanoassemblies were deprotonated, and the hydrophobic interaction, π -π stacking, and hydrogen bonding between the CCM segments were the dominant forces that form condensed CCM domains to maintain nanoassemblies stable. When incubated at weak acidic pH, the hydrophobicity of CCM segments can be reduced due to the adsorption of H + ions, leading to destabilization of the CCM domains. Furthermore, the cationic six-membered ring on protonated diketone groups in the CCM domains caused sensitive electrostatic repulsive force. These pH-responsive synergistic phenomena probably drove the swelling of nanoassemblies. It is reasonably understood that the destabilization of protonated CCM domains located inside of nanoassemblies (CPC and PC 4 ) have great impact on the pH-responsive size and morphology change than that of CCM domains located at the near surface of nanoassemblies (PC and PCP). Interestingly, average size of PC 4 nanoassembly gradually decreased when pH was reversely increased from 5.5 to 7.4, and the resultant nanostructures exhibited inherent square-shaped morphology with average diameter of about 115 nm and a narrow size distribution at pH 7.4 (Supplementary Fig. 10), indicating the reversible pH-responsiveness of PC 4 nanoassemblies. It is well known that early endosome and matured endosomes (late endosomes and lysosomes) typically displayed diameters ranging from 100 to 200 nm and from 200 to 400 nm, respectively 25 . Thus, the size of CPC and PC 4 nanoassemblies swelled at pH 5.5-6.5, corresponding to endosomal pH under the maturation process, was similar or larger than the size of endosomes, implying that the expansion force generated by the swelling of CPC and PC 4 nanoassemblies can enhance their endosomal escaping activity. ## Endosomal escaping activity of CCM nanoassemblies. To assess the endosomal escaping activity of CCM nanoassemblies arising from the pH-responsive properties, damage of endosomal membrane was examined using acridine orange relocation assay 32 . Acridine orange is a pH-sensitive dye which diffuses into cells and accumulates in the acidic endosomes/lysosomes by proton trapping. The accumulation of acridine orange in acidic endosomes/lysosomes gives red-orange fluorescence, while acridine orange shows yellow-green fluorescence in neutral pH environment. Therefore, the rupture of endosomal/lysosomal membrane can be indicated by a shift of red-orange fluorescence to a yellow-green fluorescence in the presence of acridine orange. Figure 4a shows CLSM images of PC-3 cells treated with CCM nanoassemblies and free CCM in the presence of acridine orange, and Fig. 4b shows the time course of the average numbers of endosomes/lysosomes observed in a PC-3 cell based on the CLSM images. PC-3 cells treated with free CCM showed red-orange granules in the endosomes/ lysosomes and a diffuse weak green fluorescence in the cytoplasm after 0.5 h. The numbers of endosomes/lysosomes slightly increased over time in the presence of free CCM, meaning that free CCM does not have endosome disrupting activity. PC-3 cells treated with CCM nanoassemblies revealed similar red-orange granules after 0.5 h, however the red-orange granules gradually reduced over time, indicating the expansion of endosomal/ lysosomal membrane damage in the presence of CCM nanoassemblies. Thus, it was found that all types of CCM nanoassemblies have obvious endosomal/lysosomal membrane disrupting activity. The noticeable decrease in the red-orange granules is attributed to the burst release of acridine orange from endosomes/lysosomes to cytoplasm. Moreover, the following obvious increase in cytoplasmic green fluorescence intensity can be attributed to the accumulation of acridine orange as well as CCM nanoassemblies in the cytoplasm. Thus, it was clearly demonstrated that CCM nanoassemblies have effective endosomal escaping activity and sequential cytoplasmic self-delivering ability. Among them, CPC and PC 4 nanoassemblies showed continuous decrease in the numbers of endosomes/lysosomes within 5 h, while increase of the numbers was started after 3 h for PC and PCP nanoassemblies. As mentioned above, only CPC and PC 4 nanoassemblies exhibited remarkable pH-responsive size increase up to 350-450 nm which is similar or larger than the diameter of endosomes/lysosomes. Considering this, the pH-responsive size increase of CPC and PC 4 nanoassemblies occurred within endosomes/lysosomes probably contribute to the endosomal escaping activity. In particular, only PC 4 nanoassembly showed great decrease in the numbers of endosomes/lysosomes within 1.5 h, implying that PC 4 nanoassembly have further additional mechanism for enhancement of the endosomal escaping activity. To obtain further information about pH-responsive properties of PC 4 nanoassembly at molecular level, 1 H-NMR spectra of PC 4 nanoassembly were measured in D 2 O at different pH. As shown in Fig. 5a, the proton signal of PEG segment (3.4-3.8 ppm) was detected as a sharp peak, while the characteristic peaks of CCM segments (5.8-7.6 ppm) were completely disappeared at pH 7.4 because CCM segments form hydrophobic domains inside of the nanoassemblies and the major part of PEG segments extend out into aqueous environment at the surface of nanoassemblies. The peak intensity of CCM segments obviously increased with the decrease in pH from 7.4 to 5.5, indicating that the CCM domains were partially disrupted and the some parts of protonated CCM segments released from their domains were displayed at near the surface. This result is well consistent with the disordered CCM domains and swelled morphology of PC 4 nanoassemblies at pH 5.5 detected by TEM observation. Importantly, such noticeable increase in the peak intensity of CCM segments was observed for only PC 4 nanoassembly, as shown in Fig. 5b. Thus, it is plausible that the surface display of protonated CCM segments in response to weak acidic pH can be an additional source to enhance the endosomal escaping activity of PC 4 nanoassembly. Barry and co-workers reported that CCM molecules anchored in cell membrane with high density caused disruption of the membrane structure 10 . Inspired by this report, pH-responsive membrane-lytic property of CCM-based nanoassemblies was further studied with erythrocytes as a model for the endosomal membrane 33 . As shown in Fig. 5c, membrane-lysis was not induced by treatments of PC, PCP, and CPC nanoassemblies under any pH conditions. By contrast, PC 4 nanoassembly clearly showed membrane-lytic activity and the activity was enhanced at weak acidic pH condition (pH 6.5 and 5.5). Taken together, it was found that the membrane-lytic activity of PC 4 nanoassembly is caused by the protonated CCM segments, displayed at the surface, which are capable of biding to the erythrocyte membrane through electrostatic interaction and subsequent insertion into the membrane, resulting in the disruption of cell membrane. ## In vitro cytotoxicity of CCM nanoassemblies. To evaluate the potential of CCM nanoassemblies as anticancer nanodrugs, in vitro cytotoxicity was evaluated using cancer cell lines (PC-3 and HepG2 cells). As shown in Fig. 6, cancer cells treated with all types of CCM nanoassemblies (with the same concentration at CCM level) showed a typical dose-dependence sigmoidal curve. This result indicates that the cytotoxicity is derived from the CCM nanoassemblies, thus CCM nanoassemblies can act as anticancer nanodrugs. The half maximum inhibitory concentration (IC 50 ) after 24 h were calculated from the obtained sigmoidal curves and the values were summarized in Supplementary Table 2. All types of CCM nanoassemblies showed lower IC 50 values than free CCM for both PC-3 and HepG2 cells. Importantly, PC 4 nanoassemblies showed the lowest IC 50 value for PC-3 cell, and CPC and PC 4 nanoassemblies showed the lowest IC 50 values for HepG2 cell, indicating that the in vitro cytotoxicity depends on the endosomal escaping activity to deliver themselves into cytoplasm as the site of action of CCM for cytotoxicity. In vivo studies of anticancer CCM nanodrugs. Generally, nanoparticles with a suitable size (< 250 nm) show a longer blood retention time as compared to free small-molecule drugs 34 . To evaluate the influences of supramolecular nanoassembly of CCM on the blood circulation profiles, tumor-bearing mice were treated with single intravenous injection of CCM nanodrugs or free CCM, collected plasma at different time intervals, and then estimated the plasma concentration at CCM level by UV-Vis measurements. As shown in Fig. 7a, the plasma concentration of free CCM sharply decreased to approximately 35% of the initial maximum dose within 0.5 h, indicating rapid clearance of free CCM from the circulation system. By contrast, all types of CCM nanodrugs showed much prolonged blood circulation time with significantly higher CCM concentration over the free CCM. To evaluate the biodistribution profiles, tumor-bearing mice treated with single intravenous injection of CCM nanodrugs or free CCM were sacrificed, and the amounts of CCM accumulated in major organs were estimated by UV-Vis measurements at 0.5 h, 4 h, 12 h, and 24 h post-injection. As shown in Fig. 7b, biased accumulation in specific organs was not observed for all types of nanodrugs. The amounts of CCM nanodrugs accumulated in reticuloendothelial systems, such as liver and spleen, and kidney which are responsible for active clearance of CCM from circulation, was almost similar with that of free CCM. This result indicates that the prolonged blood circulation time of CCM nanodrugs over free CCM would be mainly due to the improved hydrolysis resistance, as described in our previous report 23 . The amount of CCM nanodrugs accumulated in tumor tissues increased over time at least within 24 h. After 24 h, the amounts accumulated in tumor were reached to 316 ng (PC), 345 ng (PCP), 322 ng (CPC), and 334 ng (PC 4 ), while that of free CCM was 238 ng, indicating higher tumor accumulation of CCM nanodrugs than free CCM owing to possible passive targeting, EPR effect. To elucidate whether pH-responsive endosomal escaping activity of anticancer CCM nanodrugs results in the enhancement of anticancer therapeutic potency, PC-3 tumor-bearing mice were intravenously injected with CCM nanodrugs dispersed in PBS, free CCM dissolved in glycerol formal, or PBS only as negative control via the tail vein with volume at 2 mL/kg and an equivalent dose of CCM in nanodrugs at 10 mg/kg, respectively. Tumor volume and body weight of tumor-bearing mice were monitored every day for 30 days. Tumor growth was clearly inhibited after the treatment with free CCM and all types of CCM nanodrugs as compared with PBS treatment (Fig. 7c). The tumor inhibitory rate was calculated from tumor volume. Compared with tumor inhibitory rate of PBS group (100%) after 30 days, the inhibitory rate was 42.5% (free CCM), 51.1% (PC), 54.3% (PCP), 22.3% (CPC), and 12.8% (PC 4 ), meaning that CPC and PC 4 nanodrugs showed significantly better in vivo anticancer efficacy than free CCM and even PC and PCP nanodrugs. It is noteworthy that PC 4 nanodrugs completely inhibited the tumor growth (the tumor volume was not changed during the treatment). The in vivo undesired toxicity arising from nano-sized materials which can cause harmful side effects has usually been one of major concerns in the development of nanomedicines 35 . Therefore, the body weight change of tumor-bearing mice treated with CCM nanodrugs or free CCM with the same dose at CCM level was examined. As shown in Fig. 7d, there is about 7% loss of body weight for mice treated with PC and PCP nanodrugs and free CCM after 5 days. After that, the weight loss was stopped and gradual increase was observed for free CCM-treated mice, while the weight loss was continuously caused and reached to about 15% after 30 days for PC and PCP Formation of tumor-associated neovascular networks has usually been one of major concerns in tumor metastasis. Indeed, it has been reported that CCM inhibits angiogenesis in vitro through down-regulation of the expression of proangiogenic genes, such as VEGF, angiopoietin 1 and 2 36 . To elucidate the inhibition of tumor-associated angiogenesis by CCM nanodrug treatments, mice were sacrificed and tumors were separated at the end of experiments. Although obvious neovascular networks were generated at the tumor surface for PC, PCP, and CPC nanodrugs and free CCM treatments, PC 4 nanodrugs completely inhibited tumor-associated neovascularization (Fig. 7e), suggesting a potential inhibitory effect on tumor metastasis of PC 4 nanodrug. All these results from in vivo experiments demonstrated that CPC and PC 4 nanodrugs have obvious anticancer potency without harmful side effects available for anticancer therapy. Hybrid anticancer nanodrugs fabrication. Combination therapy which uses two or more kinds of drugs with different mechanism of action has recently gained much attention as new trend in the field of nanomedicines. Therefore, we evaluated the potential utility of CPC and PC 4 nanodrugs as delivery carriers for combination therapy. As a combination drug with CCM, we selected doxorubicin (DOX), a common anthracycline antibiotic. DOX interacts with DNA in cell nucleus through the intercalation into DNA double helix, leading to inhibition of the DNA synthesis or poisoning of topoisomerase II 37 . DOX is used for treatment of various cancers, especially breast, ovarian, prostate, brain, lung, and leukemia 38,39 . However, the anticancer efficacy of DOX is often compromised by multidrug resistance mechanisms involving P-gp proteins. DOX is a substrate for drug transporters such as of Multidrug Resistance 1 (MDR1) and Multidrug Resistance-Related Protein 1 (MRP1) 38 . Furthermore, DOX is a weak base with a pK a near neutrality; it is prone to accumulating in acidic cytoplasmic vesicles of cancer cells by active transport mechanism, thereby forming sink conditions in the cytosol of cancer cells 39 . These drawbacks lead to decreased sensitivity of DOX to cancer cells 40 . To overcome the drawbacks of DOX, high cellular uptake and subsequent cytoplasmic delivery via pH-responsive endosomal escaping of CPC and PC 4 nanodrugs can be helpful. Moreover, synergistic anticancer effects can be expected because CCM and DOX have different anticancer mechanism. Therefore, we evaluated a potential utility of CPC and PC 4 nanodrugs as DOX delivery carriers. From the viewpoint of molecular structure, DOX has π -conjugated framework and multiple hydroxyl groups (Supplementary Fig. 11c) capable of generating π -π stacking interaction and hydrogen bonding with CCM molecule. Considering this, we attempted to fabricate supramolecular co-assembled hybrid nanodrugs consisting of CCM and DOX. In this study, molar ratio of CCM amphiphile (100 μ M at CCM level) and DOX (50 μ M) was fixed at 2:1. CPC or PC 4 molecules were dissolved in aqueous DOX solution and then sonicated to give co-assembled hybrid nanodrugs. Supplementary Fig. 12 shows DLS data of CPC/DOX and PC 4 /DOX hybrid nanodrugs, and free DOX. Both CPC/DOX and PC 4 /DOX hybrid nanodrugs showed a unimodal peak with narrow size distribution, and free DOX peak was not detected in the presence of CPC and PC 4 , revealing the formation of co-assembled hybrid nanodrugs with well-defined size. The average diameter was 182 nm (CPC/ DOX hybrid) and 73 nm (PC 4 /DOX hybrid), respectively. The architecture of CCM, DOX, and PEG in the hybrid nanodrugs at molecular level was analyzed by 1 H-NMR measurement in D 2 O. Supplementary Fig. 11a,b shows 1 H-NMR spectra of CPC/DOX and PC 4 /DOX hybrid nanodrugs. The proton signals of CCM segments and DOX were not detected but the proton signal of PEG segment was detected as a sharp peak. This result indicates that hybrid domains consisting of CCM segments and DOX were formed inside of hybrid nanodrugs and the major part of PEG segments were extended out into aqueous environment at the surface of hybrid nanodrugs. Cellular uptakes and intracellular trafficking of the CCM/DOX hybrid nanodrugs were investigated using PC-3 cancer cells. As shown in Fig. 8, both green (CCM-based nanodrugs) and red (DOX) fluorescence intensity in cytoplasm of PC-3 cells treated with CPC/DOX and PC 4 /DOX hybrid nanodrugs significantly increased over time, and the red fluorescence intensity in cytoplasm was much higher than that with free DOX treatment, revealing that CPC and PC 4 nanodrugs have a potential as cytoplasmic delivery carriers for DOX. Importantly, the red fluorescence intensity in cells treated with PC 4 /DOX hybrid nanodrugs was much higher than CPC/DOX hybrid nanodrugs, indicating that the superior endosomal escaping activity of PC 4 nanodrugs was still functioning in PC 4 /DOX hybrid nanodrugs. Note that, yellow fluorescence and green fluorescence signals were detected in cytoplasm and red fluorescence signal was detected in nucleus according to the merged images, indicating that the hybrid nanodrugs were effectively delivered to cytoplasm, and then DOX was released into cytoplasm. Most importantly, CPC/PC 4 nanodrugs stayed in the cytoplasm as the site of action of CCM, while released DOX entered into the nucleus as their action site. Thus, the CCM/DOX hybrid systems can overcome the drawbacks of DOX, and thus both CCM and DOX could exhibit inherent anticancer activity. ## Discussion Based on the experimental results described above and the facts reported previously, mechanism of pH-responsive endosomal escape of CCM nanoassemblies was proposed. It is well known that polymer-based carrier nanomaterials with high pH buffering capacity, such as polyethylenimine, polyhistidine, and polyamidoamine dendrimer, perform endosomal escaping activity via proton sponge effect . Since CCM nanoassemblies also have effective pH buffering capacity, proton sponge effect can be one of potential mechanisms for the endosomal escaping activity. CCM nanoassemblies are entered into cancer cell through clathrin-mediated endocytosis and trapped in acidic endosomes. The CCM segments in nanoassemblies become protonated by H + ion adsorption under endosomal acidic pH condition. H + ions are further supplied by the V-ATPase (proton pump) during the endosomes maturation. This process keeps the pump functioning and leads to the influx of one Cl − ion and one water molecule per proton 46 . The influx of Cl − ions and water molecules into the endosomes/lysosomes caused an increase in osmotic pressure and tension on the endosomal/lysosomal membrane, which subsequently induces the endosomes/lysosomes swelling and possible disruption of the membrane, resulting in the release of CCM nanoassemblies to cytoplasm. CPC and PC 4 nanoassemblies with higher buffering capacity induce more influx of Cl − ions and water molecules into the endosomes/lysosomes, generating higher osmotic pressure which facilitates the endosomal escape. For PC and PCP nanoassemblies, the increased osmotic pressure and the related tension arising from proton sponge effect within endosomes/lysosomes is possible mechanism for their endosomal escaping activity (Supplementary Fig. 13). In addition to the osmotic pressure and internal tension, expansion force arising from the pH-responsive size increase (swelling) generated within endosomes/lysosomes are possible endosomal escaping mechanism for CPC nanoassemblies (Supplementary Fig. 14). Thus, the synergistic mechanism facilitates the endosomal escaping activity of CPC nanoassemblies. In case of PC 4 nanoassemblies, the increased osmotic pressure and internal tension, expansion force arising from the pH-responsive size increase, as well as pH-responsive membrane-lytic activity are possible endosomal escaping mechanism (Fig. 9). Thus, the synergistic and sequential multiple mechanism significantly enhances the endosomal escaping activity of PC 4 nanoassemblies. Impotrantly, the in vivo anticancer activity of CCM nanodrugs is in accordance with the results of in vitro cytotoxicity and endosomal escaping activity. As mentioned above, there was no significant difference in the tumor accumulation between four types of CCM nanodrugs. Considering the results, the superior in vivo anticancer activity of PC 4 nanodrugs can be attributed to the highly effective pH-responsive endosomal escaping activity and subsequent cytoplasmic self-delivering ability. Thus, it is demonstrated that the pH-responsive endosomal escaping activity of CCM, newly discovered in this study, directly results in enhancement of their anticancer potency. In summary, we have developed CCM nanoassemblies by adopting supramolecular self-assembly approach. Due to the pH-responsive endosomal disrupting activity of CCM nanoassemblies newly discovered in this study, the CPC and PC 4 nanoassemblies performed endosomal escape via synergistic and sequential mechanisms. Importantly, systemic administration of CCM nanoassemblies with high endosomal escaping activity showed superior in vivo anticancer efficacy than nanoassemblies with low endosomal escaping activity and CCM alone. Thus, we demonstrated an important contribution of the pH-responsive endosome disrupting activity of CCM to enhance its anticancer activity. Furthermore, we demonstrated an additional usable ability of CCM nanodrugs as cytoplasmic delivery carriers for DOX toward combination cancer nanomedicine. Overall, this study can powerfully move forward with a clinical application of CCM and open new opportunity of CCM as biomaterials in nanomedicines. ## Methods Materials. Curcumin was purchased from Tokyo Chemical Industry (Tokyo, Japan). Poly(ethylene glycol) methyl ether (MePEG750, M w : 750 Da), poly(ethylene glycol)-diol (PEG1500, M w : 1500 Da) and 4-arm branched poly(ethylene glycol) (4-arm PEG, M w : 10,000 Da) were purchased from Sigma-Aldrich Japan (Tokyo, Japan). 4-Dimethylaminopyridine (DMAP) and N,N′-carbonyldiimidazole (CDI), dehydrated tetrahydrofuran (THF), dehydrated dichloromethane, diethyl ether, dimethyl sulfoxide (DMSO), methanol, deuterated chloroform (containing TMS), and deuterated water were purchased from Wako Pure Chemical Industries (Osaka, Japan). Dialysis membrane Spectra/Por 7 (MWCO: 3500) was purchased from Spectrum Laboratories. HepG2 human hepatocellular liver carcinoma cell and PC-3 human prostate cancer cell were purchased from the American Type Culture Collection (ATCC). 3-(4,5-dimethylthial-2-yl)-2,5-diphenyltetrazalium bromide (MTT) and Hoechst 33342 were purchased from Dojindo (Kumamoto, Japan). Synthesis of CCM amphiphiles. MePEG750 (PC: 308 mg, 0.41 mmol, PCP: 615 mg, 0.82 mmol), PEG1500 (CPC: 308 mg, 0.21 mmol), or (4-arm PEG: 200 mg, 0.020 mmol) were dissolved in 3 mL of THF (for PC, PCP, and CPC) or dichloromethane (for PC 4 ) at room temperature. The obtained PEG solution was added into 4 mL of THF solution containing CDI (PC: 100 mg, 0.62 mmol; PCP and CPC: 200 mg, 1.23 mol; and PC 4 : 20 mg, 0.12 mmol) and DMAP (PC: 50 mg, 0.41 mmol; PCP: 100 mg, 0.82 mmol; CPC: 50 mg, 0.42 mmol; and PC 4 : 10 mg, 0.080 mmol), and then stirred at room temperature for 6 h. After that, the resultant solution was added dropwise into 5 mL of THF solution containing CCM (PC, PCP, and CPC: 150 mg, 0.41 mmol; and PC 4 : 30 mg, 0.08 mmol) and DMAP (PC and CPC: 100 mg, 0.82 mmol; PCP: 150 mg, 1.23 mmol, and PC 4 : 80 mg, 0.64 mmol), and stirred at room temperature for 40 h. The reaction solution was added dropwise into 250 mL of cold diethyl ether and stirred for 120 min to give precipitate of CCM-PEG conjugates. Diethyl ether dissolving non-reacted CCM and reaction byproducts was removed by centrifugation (3000 rpm for 3 min) and the precipitate was washed by cold diethyl ether twice, and then the precipitate was dried under vacuum for 48 h to give yellow flake of CCM-PEG conjugates. Characterization of the obtained CCM amphiphiles were carried out by 1 H-NMR measurement (JEOL, ECA-500, solvent: CDCl 3 ) and gel permeation chromatography (GPC) (JASCO, HPLC LC-2000Plus, column: TSKgel ® , detector: RI, standard: PEG, eluent: DMSO dissolving 10 mM LiBr). Preparation and characterization of CCM nanoassemblies. The CCM amphiphiles were directly dissolved in pure water at room temperature and the solution was sonicated for 10 min to give their self-assembled nanostructures. The average diameters of nanoassemblies in pure water were measured by dynamic light scattering (DLS, ZETASIZER NanoSeries ZEN-3600, Malvern). The critical assembly concentration (CAC) of CCM-PEG conjugates in pure water was analyzed by UV-Vis spectroscopy using eosin Y as probe molecule at 20 °C. Eosin Y solution (final concentration: 20 μ M) was added to CCM-PEG conjugate solution and incubated for 2 h, and then the absorbance of eosin Y at 518 nm was measured by UV-Vis spectrophotometer (JASCO, V-630). The shape of nanoassemblies was visualized by transmission electron microscope (TEM) observation (JEOL, JEM-1400). One drop of aqueous nanoassembly solution (concentration: 100 μ M) was carefully placed on a copper grid, air-dried under vacuum prior to TEM observation. The morphology of CCM nanoassemblies was analyzed by 1 H-NMR measurement in D 2 O. Cellular uptake. PC-3 cell were cultured in DMEM supplemented with 10% heat-inactivated FBS, 0.15% NaHCO 3 , 2 mM L-glutamine, 100 U/mL penicillin, 100 μ g/mL streptomycin and the culture was maintained in a humidified incubator at 37 °C with 5% CO 2 . PC-3 cells (5.0 × 10 4 cells) were seeded on 24-well cell culture plate and incubated for 24 h. Cellular transport inhibitors (chlorpromazine hydrochloride: 40 μ M, genistein: 40 μ M, cytochalasin D: 4 μ M) were added to medium and incubated for 2 h, and medium was freshly changed prior to addition of CCM nanoassemblies (200 μ M). After 1, 2, 4, and 6 h incubation with the nanoassemblies, 100 μ L of medium containing nanoassemblies was taken and analyzed by UV-Vis spectroscopy at 430 nm, corresponding to λ max of CCM, to estimate the amount of CCM nanoassemblies in medium. PC-3 cells treated with nanoassemblies for 6 h were observed by fluorescence microscopy measurements. The nuclei were stained with Hoechst 33342. PC-3 cells (2.5 × 10 5 cells) were seeded on glass bottom cell culture dish (35 mm) and incubated for 24 h prior to addition of CCM nanoassemblies (100 μ M). After 1, 4, 8, and 12 h incubation with the nanoassemblies, supernatant was removed and cells were washed gently with PBS twice, and then fresh medium was added to the dish. Intracellular distribution of CCM nanoassemblies was observed by fluorescence microscopy measurements. Late endosomes and lysosomes were stained with LysoTracker Red and the nuclei were stained with Hoechst 33342. Acid-base titration. The buffering capacity of CCM nanoassemblies was analyzed by an acid-base titration method. CCM nanoassemblies were directly prepared in 10 mM NaOH solution and the solution pH was adjusted to pH 9.0 by addition of 10 mM NaOH solution to obtain the nanoassembly solution of 200 μ M in CCM level. Thus, the CCM concentration of PC, PCP, CPC, and PC 4 nanoassembly solution was the same. Fifty μ L of 2 mM HCl was added to the solution and the pH value was measured. This titration procedure was continued at pH 3.5. As a control, buffering capacity of CCM was also analyzed by the same method. ## pH-responsive size, morphology, and zeta potential changes of CCM nanoassemblies. CCM-PEG conjugates were directly prepared in pH-controlled water (pH 5.5, 6.0, 6.5, and 7.4) at 100 μ M concentration, and pH was finally adjusted to 5.5, 6.0 6.5, or 7.4 by addition of 2 mM HCl or 10 mM NaOH solution. The size and zeta potential of the obtained CCM nanoassembly solutions were analyzed by DLS with capability to measure surface zeta potential. The morphology of nanoassemblies at each pH was analyzed by TEM observation. One drop of aqueous nanoassembly solution at various pH was carefully placed on a copper grid, air-dried under vacuum prior to TEM observation. The morphology of CCM nanoassemblies at molecular level in D 2 O (200 μ M) at different pH (5.5, 6.5, and 7.4) was analyzed by 1 H-NMR measurement. Endosomal escape of CCM nanoassemblies. Endosomal escaping activity of CCM nanoassemblies was assessed using acridine orange method. PC-3 cells (3.0 × 10 4 cells) were seeded on glass bottom 8-well plate and incubated for 24 h prior to addition of CCM nanoassemblies (100 μ M). After 0.5, 1.5, 3.0, and 5.0 h incubation with the nanoassemblies, acridine orange (3.0 μ g/mL) solution in PBS was added to medium and incubated for 20 min. After that, medium was removed and cells were washed gently with PBS twice, followed by fluorescence microscope observation. pH-dependent hemolytic activity of CCM nanoassemblies. The pH-responsive membrane-lytic activity of CCM nanoassemblies was assessed using erythrocytes as a model of the endosomal membrane. EDTA-treated sheep whole blood (7 mL) was centrifuged for 5 min at 1600 rpm and the resultant pellet was washed several time with PBS (pH 7.4) until the supernatant was clear and colorless. Finally, 10 mL of erythrocyte suspension was prepared by PBS (pH 7.4). Erythrocyte suspension (180 μ L) was seeded into a 96-well plate followed by centrifugation (1600 rpm, 5 min) to obtain pellets of erythrocyte in each well. The aqueous CCM nanoassembly solution (200 μ M) at pH 5.5, 6.5, and 7.4 was added to each well, and the pellet was gently resuspended with the sample solutions. After 2 h incubation with the nanostructures at 37 °C, 96-well plate was centrifuged at 4000 rpm for 5 min. The supernatant (100 μ L) was transferred to new 96-well plate and the absorbance at 540 nm was measured by microplate reader (Bio-Rad, iMark) to determine the released hemoglobin. Two controls were prepared by resuspending erythrocyte either in PBS alone with pH 5.5, 6.5, and 7.4 (negative control) or in pure water (positive control). The percentage of hemolysis for CCM nanoassemblies was estimated by comparing the absorbance of samples with that of positive control. In vitro cytotoxicity of CCM nanoassemblies. Cytotoxicity induced by CCM nanoassemblies was investigated by conventional MTT assay. 200 μ L of cell suspension (PC-3 and HepG2) was seeded to 96-well plate (1.0 × 10 4 cells) and incubated for 24 h in a humidified incubator at 37 °C with 5% CO 2 . PBS solutions of nanoassemblies with defined concentrations were added to each well and incubated for 24 h. Then, 10 μ L of MTT solution was added to each well and incubated for 4 h. The supernatant was removed and MTT-formazan crystals formed in cells were dissolved by addition of 200 μ L of isopropyl alcohol containing 0.04 M HCl and 10% Triton-X, and the absorbance at 570 nm was measured by microplate reader. In vivo distribution of CCM nanoassemblies. All animal experiments and all experimental protocols were approved by Konan University (protocol No.: K-13-06), and conformed to the Guidelines for the Care and Use of Laboratory Animals published by the National Institutes of Health. Female nude mice (BALB/cSlc-nu/nu) at 6 weeks of age were purchased from Charles River. For surgery, the animals were anesthetized using isoflurane and the surgical area was cleaned. PC-3 cells (2.0 × 10 6 cells) in 200 μ L of Matrigel (BD Bioscience) were injected subcutaneously into the back bilaterally using a disposable syringe and a 26-gauge needle. When the tumor size reached 200 mm 3 , animals were randomly divided into 6 groups (PBS only, free CCM, PC, PCP, CPC, and PC 4 , n = 3). CCM nanoassemblies dispersed in PBS or free CCM dissolved in glycerol formal were injected to tumor-bearing mice intravenously through the tail vein with volume at 2 mL/kg and an equivalent dose of CCM in each nanoassembly at 10 mg/kg, respectively. Then, 0.5, 4, 12, and 24 h after administration, the mice was sacrificed and tumor tissue, blood, and main organs (brain, spinal cord, heart, lung, liver, spleen, kidney, stomach, and intestine) were excised carefully for quantitative analysis of the accumulated CCM. These tissues and organs were homogenized and 500 μ L of mixed solution (DMSO/MeOH = 1/4, v/v) was added followed by centrifugation at 4000 rpm for 30 min. Then, the supernatant was measured by UV-Vis spectroscopy at 430 nm, corresponding to λ max of CCM, to estimate the amount of CCM nanoassemblies in these tissues and organs. In vivo antitumor activity of CCM nanoassemblies. Female nude mice (BALB/cSlc-nu/nu) at 6 weeks of age were anesthetized using isoflurane and the surgical area was cleaned. PC-3 cells (2.0 × 10 6 cells) in 200 μ L of Matrigel (BD Bioscience) were injected subcutaneously into the back bilaterally using a disposable syringe and a 26-gauge needle. When the tumor size reached 200 mm 3 , animals were randomly divided into 6 groups (PBS only, free CCM, PC, PCP, CPC, and PC 4 , n = 3). CCM nanoassemblies dispersed in PBS or free CCM dissolved in glycerol formal were injected to tumor-bearing mice intravenously through the tail vein with volume at 2 mL/kg and an equivalent dose of CCM in each nanoassembly at 10 mg/kg, respectively. These samples were administrated thrice a week for 4 weeks. The tumor volume was measured everyday using caliper and calculated according to the formula: tumor volume = (shorter diameter) 2 × (longer diameter)/2. Body weight of the mice was recorded every day. At the end of the study, tumor tissues were carefully excised and weighted. Co-assembly of CCM amphiphiles with DOX. CCM amphiphiles (CCM concentration: 100 μ M) were dissolved in aqueous DOX solution (50 μ M) at room temperature and the solution was sonicated for 10 min to give co-assembled hybrid nanostructures composed of CCM amphiphiles and doxorubicin. The average diameters of the hybrid nanoassemblies were measured by DLS. The morphology of the hybrid nanoassemblies was analyzed by 1 H-NMR measurement in D 2 O. Cellular uptake of CCM/DOX hybrid nanoassemblies. PC-3 cells (3.0 × 10 4 cells) were seeded on glass bottom 8-well plate and incubated for 24 h prior to addition of CCM/DOX hybrid nanoassemblies (100 μ M/50 μ M). After 1 h and 4 h incubation with the hybrid nanoassemblies, medium was removed and washed gently with PBS twice, and then fresh medium was added to each well. Free DOX (50 μ M) was used as negative control sample. Cell uptake and intracellular distribution of hybrid nanoassemblies were observed by fluorescence microscopy measurements.

chemsum

{"title": "Discovery of a new function of curcumin which enhances its anticancer therapeutic potency", "journal": "Scientific Reports - Nature"}

exhaustive_reduction_of_esters_enabled_by_nickel_catalysis

## Abstract: We report a one-step procedure to directly reduce unactivated aryl esters into their corresponding tolyl-derivatives. This is achieved by the action of a Ni/NHC catalyst and an organosilane reducing agent that is activated in situ by stoichiometric KO t Bu. The resulting conditions provide a direct and efficient alternative to multi-step procedures for this transformation that often require use of hazardous metal hydrides. Applications in the synthesis of -CD3 containing products, derivatization of bioactive molecules, and chemoselective reduction in the presence of other C-O bonds is demonstrated. ## Introduction The reduction of carboxylic acids, as well as their derivatives such as esters and amides, is of fundamental importance in organic synthesis. 1 This is most commonly achieved by the action of aggressive metal hydride reducing agents, enabling the formation of the corresponding alcohol product. 2 Exhaustive reduction -that is, reduction all the way to the methyl oxidation state -is less readily achieved. Reasons why one may seek to perform this reaction include the profound effects of methyl groups on bioactivity, 3a-b the ability to exploit the electronic effects of an ester group prior to reduction, 3c or its ability to create deuterium-labelled tolyl derivatives. 3d-e In rare cases, single step transformation of an ester into a methyl group can be realized, 4 for example by the use of excess lithium aluminum hydride in refluxing ethereal solvent (Scheme 1a). 5 More often, this transformation is carried out by a three-step sequence via initial reduction to the alcohol, functional group interconversion to an alkyl halide, and further reduction to the methyl product (Scheme 1b). 6 Alternatively, catalytic hydrogenolysis can be used to reduce an alcohol to the methyl oxidation state (Scheme 1c). 7 Scheme 1. Contemporary methods to achieve the ester-tomethyl transformation Our group and others have recently demonstrated that Ni catalysts are capable of catalytic activation of methyl esters, presumably initiated by oxidative addition into the C(acyl)-O bond. 8 In the course of these studies, we hypothesized that useful catalytic reductions of these esters may be achievable using a similar strategy. Related work from the Martin lab for reducing anisoles to arenes (Scheme 2a), 9 the Rueping lab on the reduction of phenyl esters to arenes (Scheme 2b), 10 and the Garg lab on transforming amides to amines (Scheme 2c), 11 among others, 12 served as inspiration that silane reagents could act as mild hydride donors in Ni-catalyzed reactions. With this precedent in mind, we initiated a study on the reaction of silanes with methyl esters in the presence of Ni. To our surprise, when using 1,1,3,3tetramethyldisiloxane (TMDSO) 13 as a reducing agent in the presence of Ni(cod)2, 1,3-dicyclohexylimidazolium tetrafluoroborate salt (ICy•HBF4), and potassium tert-butoxide, Scheme 2. Ni-catalyzed reductions mediated by organosilane reducing agents. the major product was the corresponding tolyl derivative (Scheme 2d). Given the absence of general, functional group tolerant methods to achieve this transformation in a single step, we further explored this reaction and report the results herein. The Ni-catalyzed exhaustive reduction of methyl esters to the corresponding tolyl derivative was optimized using N-Me indole 1. Heating the substrate at 110 °C in toluene in the presence of Ni(cod)2 (10 mol%), ICy•HBF4 (20 mol%), TMDSO (2 equiv), and KO t Bu (1 equiv) afforded methylbearing indole 2 in 84% yield (Table 1, entry 1). Running the reaction in the absence of either KO t Bu or siloxane resulted in no conversion (entry 2). However, removing just the Ni(cod)2 led to complete consumption of 1, with the corresponding benzyl alcohol as the only major identifiable product after working the reaction up with TBAF (entry 3). Using this alcohol as a starting material in the presence of the Ni catalyst resulted in efficient formation of 2, suggesting that the reaction proceeds through a non-catalyzed reduction to a silylated alcohol 14 followed by catalytic reduction of this species to form the alkane (entry 4). Alternative ligands (entry 5, 6), bases (entry 7, 8) and organosilane reducing agents (entry 9, 10) were all less efficient. Despite being highly reducing conditions, Ni(II) precatalysts such as NiCl2 (entry 11) or NiBr2•glyme (entry 12) gave poor yields. The addition of 0.2 equiv Mn gave a moderate improvement to 35% (entry 13), while 1 equivalent of Mn gave 78% yield (entry 14), providing a viable alternative set of conditions that can be performed without the use of a glovebox. Applying these conditions in the absence of TMDSO resulted in no formation of 2 (entry 15), indicating that TMDSO is ultimately responsible for the substrate reduction. Notably, the reaction was found to be similarly efficient when starting from the corresponding t Bu, Bn or Ph ester (entry 16). With optimized conditions, we turned our attention towards the reaction scope (Scheme 3). Electron-neutral and rich methyl arenes 3-7 were formed from their corresponding esters in 73-86% yield. Subjecting a styrene-bearing methyl ester to the conditions led to partial reduction of the olefin; however, slightly milder conditions enabled 8 to be selectivity formed in 74% yield. Slightly milder conditions also allowed recovery of stilbene 9 in similar yields. In contrast with prior work on Ni-catalyzed hydrogenolysis of anisoles and related compounds, 9a,12c-f no C(aryl)-O cleavage was observed when using ethereal or phenol-containing substrates (10-15). In many classical heterocycle-forming reactions, ester-containing starting materials like malonates or propiolates are used to facilitate cyclizations, thus making the ester-containing products more accessible than the corresponding methyl analogs. With this in mind, we were excited to see that many heterocycles were tolerated, including pyridines (16-20), morpholines (21, 22), piperidines (23, 24), indoles (25-29), carbazoles (30-32), an indazole (33), quinoline (34), benzofuran (35), and dibenzofuran (36). Derivatization of commercial pharmaceuticals bifendatatum and probenecid successfully formed reduction products 37 and 38, respectively. Lastly, reduction of a lactone proceeded smoothly to form alcohol-bearing tolylderivative 39 upon quenching with TBAF. a Reactions ran on a 0.20 mmol scale. Yields determined by NMR using 1,3,5-trimethoxybenzene as internal standard. b Alcohol (1-methyl-1H-indol-6-yl)methanol observed as major product after work-up with TBAF. Alkyl-substituted esters, carbon-halogen containing substrates, free N-H bonds, and several other functional groups were unfortunately not tolerated (Scheme S3, Supporting Information). Glovebox-free conditions using NiBr2•glyme were also studied with a range of substrates -these conditions were observed to give synthetically viable yields, albeit with slightly lower efficiency than when using Ni(cod)2. Lastly, a gram scale reaction was performed in a round-bottom flask, providing 7 in 71% yield. Given the importance of deuterated molecules in the study of reaction mechanisms, pharmaceuticals, and as tools for understanding metabolic pathways, 15 we next explored the potential of our exhaustive reduction to synthetically introduce -CD3 groups. While the development of methods to perform this transformation is of contemporary interest, 16 this is commonly achieved by using LiAlD4 or LiEt3BD as a reducing agent in the three-step procedure described in Scheme 1. 6f-k For instance, Salvadori and co-workers used a three-step, three-day procedure to synthesize CD3containing 2,5-dimethoxy-d3-toluene (40) towards the preparation of d3-δ-tocopherol. 17 Reduction of the corresponding ester to alcohol with LiAlD4, halogenation with PBr3, and a second LiAlD4 reduction afforded product in an overall yield of 69%. Using d2-TMDSO 18 alongside Ni(cod)2 gave incomplete deuterium incorporation. In contrast, the use of NiBr2•glyme/TMDSO/Mn in toluene-d8 resulted in an efficient 73% yield with >95%D incorporation(Scheme 4a). This procedure was found to be general, enabling the synthesis of several -CD3-containing molecules in good yield and high deuterium incorporation (41-43). Experiments in non-deuterated toluene gave ~5-10% reduced deuterium incorporation. We were intrigued by the similarity between our conditions to reduce methyl esters and those which Martin and coworkers used to reduce anisole derivatives. 9a Notably, these methods differ primarily in the choice of ligand and the presence of base. To directly probe the selectivity, esters 44-47 were subjected to both sets of reaction conditions (Scheme 4b). Using a Ni/ICy catalyst in the presence of KO t Bu resulted in chemoselective formation of the ester reduction products 48-51 in 54-82% yield Using base-free conditions with a Ni/PCy3 catalyst resulted in a complete switch, cleaving of the ethereal bond while leaving the methyl ester untouched in products 52-54. The Ni/PCy3 system has been proposed to act via a Ni(I)-SiR3 active cata-lyst; 9b the contrasting chemoselectivity in our system suggests that both a different active catalyst and different mechanistic pathway are operative. ## Scheme 4. Applications of exhaustive reduction of esters To gain further mechanistic information, the method of variable time normalization analysis was applied to the reduction of methyl naphthoate 53 (Scheme 5). 19 Positive apparent first order kinetics were observed for all studied species -the Ni catalyst, substrate, siloxane, and base. While the exact nature of reducing agent is not clear, the rate dependency on the base and literature precedent both support a hypervalent O t Bu-siloxane species being involved in the key hydrogen transfer. 20 We tentatively propose that this hypervalent siloxane reduces the ester into a silylated alcohol, which reacts with a Ni(0) catalyst by oxidative addition to form a benzylic Ni(II) intermediate that is reduced in a ratedetermining hydride transfer. Scheme 5. Apparent rate law determination by variable time normalization kinetic analysis In summary, we have developed a new method to convert methyl esters into the corresponding tolyl derivative. In contrast to established multi-step sequences for this transformation, the reaction takes place under a single set of conditions, comprised of a rapid, non-catalyzed siloxane-mediated reduction to a silylated alcohol followed by subsequent Ni-catalyzed reduction to the corresponding -CH3 group. We believe this reaction will be particularly useful in the synthesis of methyl-bearing heterocycles, where the corresponding esters are more readily abundant, as well as in reductive deuteration as a method to install -CD3 groups. This transformation sharply contrasts previous reports of Nicatalyzed reductions with siloxanes of, e.g., amides, ethers, and phenyl esters. The inclusion of a KO t Bu is proposed to be a key feature, which generates a more aggressive siloxane reducing agent in situ. ## AUTHOR INFORMATION Corresponding Author *stephen.newman@uottawa.ca ## Author Contributions The manuscript was written through contributions of all authors. ## Supporting Information Information regarding experimental procedures, optimization tables, troubleshooting, characterization of organic molecules, mechanistic studies is available free of charge via the Internet at http://pubs.acs.org.

chemsum

{"title": "Exhaustive Reduction of Esters Enabled by Nickel Catalysis", "journal": "ChemRxiv"}

measuring_the_coefficient_of_friction_of_a_small_floating_liquid_marble

## Abstract: This paper investigates the friction coefficient of a moving liquid marble, a small liquid droplet coated with hydrophobic powder and floating on another liquid surface. A floating marble can easily move across water surface due to the low friction, allowing for the transport of aqueous solutions with minimal energy input. However, the motion of a floating marble has yet to be systematically characterised due to the lack of insight into key parameters such as the coefficient of friction between the floating marble and the carrier liquid. We measured the coefficient of friction of a small floating marble using a novel experimental setup that exploits the non-wetting properties of a liquid marble. A floating liquid marble pair containing a minute amount magnetite particles were immobilised and then released in a controlled manner using permanent magnets. The capillarity-driven motion was analysed to determine the coefficient of friction of the liquid marbles. The "capillary charge" model was used to fit the experimental results. We varied the marble content and carrier liquid to establish a relationship between the friction correction factor and the meniscus angle. A liquid marble is a liquid droplet coated with hydrophobic powder, which can roll on a solid surface and float on a liquid surface due to its non-wetting coating . A liquid marble can be driven across a liquid surface using thermocapillarity 17,18 . Liquid marbles served as micro bioreactors for culturing cells . A number of excellent review papers provided a comprehensive summary on properties and applications of liquid marbles 1,14, . Recently, Bormashenko et al. demonstrated that a floating liquid marble can propel itself across the water surface due to the Marangoni solutocapillary effect 16 . Self-propulsion provides a means of transport for a small volume of an aqueous solution. Our previous study on the autonomous motion of a floating liquid marble focused on operation parameters such as the geometric constraint and the concentration of a volatile compound within the marble 13 . The dynamics of the motion was beyond the scope of the previous study as critical parameters such as the coefficient of friction of a floating marble were unknown. The marble motion is driven by surface tension gradient and resisted by friction. The knowledge of both the friction force and the resultant motion can lead to the determination of the surface tension gradient. This gradient can then be correlated to the amount of volatile compound released to the surface of the carrier liquid. Insights into the autonomous motion of a self-propelled floating liquid marble allows for better prediction of its velocity, carrying capacity and lifetime. The coefficient of friction of a solid spherical particle straddling between two fluids have been studied extensively . Some previous studies analysed the motion of small spherical particles under the action of known capillary forces 30,31 . The parameters were chosen such that the particles experienced Stokes drag. The Newtonian equation of motion was then applied to solve for the friction term, which has a correction factor. This model requires the measurement of several key parameters. First, the radius of the three phase contact line (TPCL) must be known to calculate the Stokes drag. Second, the meniscus angle created by the particle needs to be measured as it is related to the capillary force. The measurement of these two parameters for a solid spherical particle is relatively straightforward. However, a floating liquid marble deforms, and its meniscus angle does not necessarily depend on the apparent contact angle 12 . Furthermore, the radius of the TPCL varies with the marble volume. For simplicity, the liquid marbles investigated in this paper are sufficiently small to be assumed as a sphere. To date, the measurement of the coefficient of Stokes friction for a floating liquid marble has not been reported. Although we have previously estimated the correction factor of the ideal Stokes friction coefficient based on indirect measurements and calculations 13 . In the present work we seek to measure this correction factor by applying known theories and a novel experimental setup, which exploits our recently reported magnetic actuation concept for liquid marbles 32 . A liquid marble can be controlled magnetically by filling it with a small amount of magnetite without significantly changing its density or surface tension. Two floating marbles containing magnetite were immobilised with permanent magnets positioned below. If the magnets are dropped vertically, the marbles attract each other through the capillary force of the deformed liquid surface. Using the calculated capillary force, we determined the correction factor for the coefficient of friction of the floating marbles. We then correlate this correction factor with the meniscus angle of the carrier liquid, thus creating a convenient method of estimating this factor. ## Theory Two floating spherical particles, placed close to each other, experience a resultant force, F R . In the horizontal (or x) direction, the component of the resultant force F R(x) acting on a particle is: where m is the mass of the particle, ẍ is the horizontal acceleration of the particle; and F C(x) and F f are the capillary and friction force, respectively. The friction force can be calculated using Stokes' law 30,31 : where r is the particle radius, μ is the dynamic viscosity of the carrier liquid,  x is the horizontal velocity of the particle, and β is the correction factor for the coefficient of friction. The seminal work by Kralchevsky et al. 33 concluded that each particle behaves like a "capillary charge", which is analogous to electrically charged particles. The capillary forces can be expressed as: where L is the distance between the particles and the first order of the modified Bessel function of the second kind K 1 that can be approximated as: where z is a dimensionless quantity (z = qL in our case), γ the surface tension between the fluids, q the inverse capillary length of the carrier fluid: w where ρ w is the density of the carrier liquid. Q 1 and Q 2 are the capillary charges of individual particles that can be calculated as: 0 where r 0 is the contact radius of the TPCL, which is typically much smaller than the distance between the two marbles L ref. 33. For example, in our experiments r 0 is on the order of 5 × 10 −4 m, whilst L is set to be 1.5 × 10 −2 m. In Equation (6), ψ is the meniscus angle of the carrier liquid relative to a reference horizontal level, Fig. 1. Assuming the marble remains spherical, the contact radius can be estimated as: 0 Substituting ( 7) into (6) results in: For two liquid marbles of the same composition and radius, Q = Q 1 = Q 2 ; L = 2x given that x is measured from the midpoint between the two marbles. Substituting m = (4π r 3 /3)ρ m , where ρ m is the effective density of the liquid marbles, and making the appropriate arrangement in (1) gives: (2 ) 0 In (9), the physical properties (μ, r, ρ m and ρ w ) are known. From the experiments, x is measured as a function of time, and the resultant data can be differentiated with respect to time to obtain both velocity,  x, and acceleration, ẍ . Consequently, the measurements can be used in combination with (9) to obtain the values of β for a known meniscus angle ψ. The meniscus angle ψ is determined by application of a vertical force balance to the liquid marble (at equilibrium): where F g is the weight of the marble, F b is the buoyancy force, and F C(y) is the vertical component of the capillary force. Making the appropriate substitutions for F g , F b and F C(y) gives: where h(r, ψ) is the height of the meniscus, measured vertically from the TPCL to the free liquid surface sufficiently far away from the marble; h is a function of r and ψ, and can be solved using methods described in our previous work 12 . Equation ( 11) can be solved numerically to obtain an equilibrium value for ψ for a given set of values for r, γ, ρ m and ρ w . With the value of meniscus angle ψ then, β can subsequently be obtained from (9). Note that the analysis given above is only valid for small particles, i.e. (qr) 2 ≪ 1, and small meniscus slopes, i.e. ψ  sin 1 2 34,35 . In our case, both (qr) 2 and ψ sin 2 were on the order of 10 −1 . The correction factor, β is known to be dependent on the amount of perturbation of the liquid interface, which in turn depends on the meniscus angle ψ and the submerged depth of the liquid marble. Since this depth is a function of ψ for a known volume, β should be a function of ψ as well. For a small floating liquid marble, the buoyancy force F b is small compared to the surface tension force F C(y) 15 . Neglecting F b in (10), Equation ( 11) can be simplified as: As β should be a function of ψ for a given volume, it should also be a function of the liquid surface tension, γ and ρ m . Note that the liquid density and marble surface tension terms are missing, because buoyancy force and marble deformation are neglected. This assumption gives us a convenient way to estimate β based on reasonably measurable quantities. In this paper, we vary the effective marble density ρ m and the carrier liquid surface tension γ to determine the relationship between the correction factor β and sin ψ. The marble radius r is kept constant, whereas ρ m is measured for marbles containing different NaCl concentrations. The surface tension γ and the density ρ m were changed in our experiments, independently. With an aspect ratio of ε ≈ 0.9 for a 5-μ L marble, the liquid marbles were assumed to be spherical in our experiments 12 . We also assumed that the marbles maintain the spherical shape during the motion. Although the marble is spherical in shape, its physical properties differ from that of a solid spherical particle. For a solid spherical particle, the contact angle is constant regardless of floating position due to its rigidity. The meniscus angle can hence be calculated if the contact angle and surface inclination at the TPCL are known. For a floating marble, the apparent contact angle changes due to the deformation of the marble itself 15 . In this paper, the deformation of the floating marble is negligible due to the small volume used. Therefore, the model is similar to that of a solid spherical particle. For larger floating marbles (> 10 μ L), this deformation needs to be taken into account. Capillary force arising from the edge of the container is neglected, as the meniscus created by the container is miniscule and far away from the floating marbles. The marble is assumed to maintain its initial volume throughout the experiment as the time scale of the experiments (~30 seconds from sample preparation until the end of collision) is much smaller than that of the evaporation process of the marble of approximately one hour. ## Materials and Methods A 5 m (mole/kg) NaCl stock solution was prepared and diluted to desired concentrations (NaCl acquired from Chem-supply, dissolved in DI water). NaCl solutions of 0 to 5 m with 1 m interval were used as carrier liquids and liquids of the marble. Magnetite (Sigma-Aldrich ® iron oxide (II, III) 5 μ m nominal diameter) was mixed with the NaCl solutions at 0.25 wt%. Weight measurements were conducted using an electronic balance (RadWag ® AS82/220.R2 analytical balance). The mixture was shaken thoroughly to increase particle suspension before being dispensed onto a bed of polytetrafluoroethylene (PTFE) powder (Sigma-Aldrich ® 1 μ m nominal diameter, ρ = 2.2 g/cm 3 ), as reported previously 32 . A 5-μ L droplet was dispensed using a micropipette (Thermo Scientific Finnpipette 4500 0.5-10 μ L) to ensure high accuracy. The micropipette has an uncertainty of ± 4.3%. The liquid marble was formed by rolling the NaCl solution and magnetite mixture droplet in the powder bed, followed by rolling it around on a clean stainless steel spoon to dislodge excess powder from the marble surface. The weight of a liquid marble was measured by averaging over 10 marbles 36 , whereas the volume of the liquid marble was calculated based on visual measurements of its diameter. The effective marble density can be found by dividing the average mass by the volume. As the floating marble has a small submerged volume and most of its weight is supported by the surface tension of the carrying liquid 15 , the change in the density of the carrier liquid has a negligible effect on the floating position. Five experimental runs were conducted with DI water as carrier liquid and marbles containing NaCl solution (1-5 m); five runs were conducted with NaCl solution as carrier liquid (1-5 m) and marbles containing DI water; one run was conducted with DI water as both carrier liquid and liquid marble content. Each run was repeated three times. All experimental works were carried out at a temperature of 293.5 ± 0.5 K and atmospheric pressure. A 140-mm-diameter Petri dish with an opaque white base was centred on a stationary platform with two square-shaped through holes spaced 15 mm apart. The Petri dish diameter is 140 mm and made of polystyrene, which has a contact angle of about 90° with water 37 . Two identical, small, cubic permanent magnets (4.5 mm for all edges) were placed inside the through holes with the same poles facing up. The characterisation data of these magnets are provided in the supplementary information. These magnets are resting on a separate sliding platform mounted on a linear stage motor (Zaber Technologies T-LS28M). All platforms were laser-machined poly(methyl methacrylate) (PMMA) slabs. The petri dish was filled with water up to a depth of about 5 mm. A USB camera (Ximea MQ013CG-ON with 0.3X EO telecentric lens) was mounted directly above the centre of the Petri dish, as shown in Fig. 2. The two marbles were placed on the liquid surface in the vicinity of the permanent magnets, one after another. The floating marbles were attracted to the permanent magnets and were then immobilised. Once the marbles were in place, the linear stage was triggered to withdraw the platform supporting the permanent magnets at a speed of 4 mm/s. When fully withdrawn, the permanent magnets were dropped and video recording was initiated immediately. As the magnets were about 4.5 mm away from the floating liquid marbles, the effective magnetic flux density is about 0.02 T, too small to magnetise the magnetite particles. Thus, magnetic force is negligible in the collision process. The entire path of collision of the marbles was recorded within less than 5 seconds. Video recording was conducted at a resolution of 638 × 508 pixels and a fixed frame rate of 50 frames per second (fps). The video was then processed frame-by-frame using MATLAB (MathWorks) to extract the centroid position of individual marbles. The data points were then fitted with a curve and the best fit generated the value of β. ## Results and Discussion When the marbles were floating above the magnets, they experienced vertical forces from the magnets. Hence the floating position would be lower with a larger contact radius. Once the magnets were dropped, the liquid marbles rapidly recover their stable positions without any influence of the magnetic field. Therefore neither the magnets nor the magnetite play any role in the subsequent marble motion generated by the capillary force. Video recordings show that marble pairs collided and bounced off each other elastically. Some runs with 5 m NaCl marbles collided with enough impact to rupture one of the marbles. A sample video recording is provided in the supplementary information. Table 1 shows effective densities ρ m of the marble needed to solve for Equation (9). By solving this equation, the position and velocity for different NaCl concentrations can be determined, Figs 3 and 4. Note that the curves do not overlap for the same concentration due to the slightly different initial positions and velocities. Based on the measured initial values, a curve (solid line) is fitted for each experimental run using different values of β. Although NaCl can be dissolved up to 6 m, solid precipitation was observed after extended period of time. Therefore, 5 m was the maximum concentration attempted. NaCl solution was used because: (i) NaCl salt is cheap and readily available; (ii) Properties of NaCl solution are well understood and widely reported; (iii) NaCl has high solubility in water; (iv) NaCl is one of the few additives which increases the viscosity and surface tension of water, hence reducing errors generated by ambient air currents. At 5 m, the surface tension, the density and the dynamic viscosity of NaCl solution are 80.0 mN/m 38 , 1.169 g/cm 3 39 , and 1.71 mPa•s 40 , respectively. All the intermediate values were acquired from the respective references. Although NaCl solution is significantly more viscous than water, the increases in surface tension and density are minor. Figures 3 and 4 show the experimental data of marble position x. A smoothing spline was fitted to the data points to reduce data spread caused by centroid calculations in the image analysis process. The smoothed curve was differentiated with respect to time to yield the velocity data; and differentiated twice to yield the acceleration data. As the smoothing magnitude was relatively small, minute spreads in the position data were amplified in the acceleration data. Figures 5 and 6 show the calculated β values from each data point, plotted against the marble position. β values are relatively constant for a large distance (x > 3 mm) but increases rapidly as the distance shortens. Thus, a two-term exponential fitting function was fitted onto the data points to obtain an averaged β: where a and b are fitting parameters. Ideally, the correction factor should be independent of the position, assuming that the floating body is a rigid sphere. However, this assumption might not be true as the correction factor is a function of the position, Figs 5 and 6. The submerged portion of the floating marble might deform, especially at higher speeds or at lower position values, affecting the value of the correction factor. The proposed exponential fitting equation was only used as an empirical description of the obtained data. This function was chosen because it takes into account the two distinct regions of β values. The data spread increases with increasing distance. This uncertainty is likely caused by the fact that the marbles dwelled at the vicinity of the starting position (≈ 7.5 mm) with low velocities (< 1 mm/s). Thus, they were very susceptible to micro currents and minute vibration sources. Since β is relatively linear in the region of (3 mm < x < 7.5 mm), we took the average value of β at these two position as the correction factor for the coefficient of friction β β β = . + . [ (0 003) (0 0075)]/2. Figure 7 shows the averaged value of β versus the sine function of the meniscus angle ψ (sin ). If the marble floats with its equator intersecting the interface, its meniscus angle ψ is zero. Consequently, the value of β should be 0.5. In our case, the liquid marble floats relatively high on the free surface with a small meniscus angle ψ. Therefore, we expect a smaller correction factor of β < 0.5. Our results disagrees with this hypothesis to a certain extent, as experiments of water marble floating on water have an average β of 0.52, higher than the expected value. This discrepancy is likely contributed by the perturbation of the water surface due to the weight of the marble and rough contact area caused by the particle coating. Nevertheless this value is very close to our estimated value reported earlier 13 . By increasing the surface tension of the carrier liquid, the marble is lifted higher with a smaller meniscus angle ψ. Hence, the correction factor β is reduced (Fig. 7, region to the left of the dash-dot line). Conversely, increasing the density of the marble generates a larger ψ, as β is increased (Fig. 7, region to the right of blue dash-dot line). The value of β may exceed 0.5 because increasing both the submerged area and meniscus angle contribute to a larger flow resistance. The carrier liquid surface tensions and marble effective densities were adjusted to reflect the changes of the meniscus angle, with reference to Equation (12). Within our experimental dataset, β can be approximated to have a linear relationship with sin ψ where β ψ = . − . 3 64 sin 0 78, r 2 = 0.73. The fitting function is shown as a solid line in Fig. 7. With this fitting function, we can estimate β of a small floating marble based on the marble and carrier liquid properties. Substituting Equation ( 12) into our fitted line results in the approximated relationship: where C 1 and C 2 are constants (C 1 = 3.64 and C 2 = − 0.78 in our case). Equation ( 14) provides a convenient way to estimate β for small floating marbles without the need to run an experiment for every parameter change. This model relating the correction factor β to the meniscus angle ψ is limited for small, floating and non-wetting spherical objects because the buoyancy forces are relatively small. Since this model assumes spherical marbles, it is valid for small Bond numbers where γ m is the marble effective surface tension. For future works, using compatible liquids of larger differences in surface tension and density can further expand the range of the meniscus angle ψ . ## Conclusion This paper reports the experimental determination of the Stokes friction correction factor of a small floating liquid marble for a small range of meniscus angles. The experimental setup was designed to simplify the calculations of the various force terms involved in the marble motion generated by capillary force. We exploited the non-wetting properties of a liquid marble and its magnetic actuation to design the experiments. The correction factor of a 5-μ L water marble was found to be about 0.52. We found that the correction factor could be estimated to vary linearly with the sine function of the meniscus angle, which itself could be expressed in terms of marble size, effective density, and carrier liquid surface tension. The method reported here provides a convenient way to estimate the correction factor of a small floating liquid marble without the need of characterising its motion. The known friction factor enables proper designing and modelling liquid marbles as a digital microfluidics platform for various applications.

chemsum

{"title": "Measuring the Coefficient of Friction of a Small Floating Liquid Marble", "journal": "Scientific Reports - Nature"}

direct_methylation_and_carbonylation_of_<i>in_situ</i>_generated_arynes_<i>via</i>_hdda-wittig_coupl

## Abstract: A highly efficient HDDA-Wittig coupling strategy for the synthesis of fully functionalized benzenes, such as ethyl 2-methylbenzoates and o-tolylethanones, is reported. The formation of four new C-C bonds via a one-pot, multicomponent cascade proceeded through the formation of a benzyne intermediate by self-cyclization, which then reacted with a phosphorus ylide. The target bicyclic aromatic compounds were prepared by the reaction of tetraynes with (acetylmethylene)triphenylphosphorane/(carbomethoxymethylene)triphenylphosphorane, and trace water allowed direct methylation and played a pivotal role in the construction of the natural carbonylated 2,3-dihydro-1H-indene cores, which were highly substituted.This report describes a robust method for the production of fused polyfunctional aromatic hydrocarbons. Scheme 1 Target carbonylated 2,3-dihydro-1H-indene structures. † Electronic supplementary information (ESI) available: Preparation of substrates, characterization data, 1 H and 13 C NMR, MS and IR spectra. CCDC 1888961 and 1903151. For ESI and crystallographic data in CIF or other electronic format see Wittig reagents or phosphorus ylides (P-ylides) are suitable tools for CvC bond formation and are frequently used in organic synthesis. 1 The aryne intermediate is highly reactive and offers good regioselectivity. 2 This intermediate has unique advantages in the construction of cyclic compounds, especially certain complex PAH compounds. 3 The hexadehydro-Diels-Alder (HDDA) reaction is widely used in the conversion of functional groups to tune molecular functions. 4 Yang et al. used a magnesium-catalyzed hemiacetal and bench-stable P-ylide for a step-economic asymmetric reaction. 5 Hoye used benzynes with BF 3 as a Lewis acid in cascade reactions to promote carbene-like reactivity. 6 Stuart et al. presented the C-H deprotonation activation to avoid the necessity for difunctionalized aryne precursors. 7 Ohmori et al. established dual benzyne monomer coupling methods for the natural product synthesis of Actinorhodin. 8 Zhu reported the heteroannulation of arynes as precursors in the natural total synthesis of (+)-hinckdentine A. 9 However, it is difficult to introduce methyl and carbonyl groups into benzene rings. We tested a novel cyclization method involving a benzyne intermediate with a Wittig reagent to prepare rare, fully substituted benzene cores. 10 The synthesis of an aryne precursor system that is both methylated and carbonylated that does not utilize a stepwise bimolecular nucleophilic substitution reaction (Scheme 1) is very challenging for organic chemists. 11,12 Surprisingly, the benzannulation of triynes and acetyl-methylenetriphenylphosphorane/carbomethoxymethylenetriphenylphosphorane and trace water in toluene typically yields fused poly-functional aromatic hydrocarbon derivatives as the major products via waste-free, green transformations that also exhibit high atom economy. 13 The fully substituted benzene derivatives were compared with general aromatic hydrocarbon derivatives, and the former, with multiple rings and sophisticated and diverse structures, were prepared by the present reaction method, high-lighting its great potential for chemical production and pharmaceutical synthesis. 14 ## Results and discussion Herein, we developed a mild, catalyst-free procedure for the synthesis of polyfunctional fused aromatics in good to excellent yields. This method was simple and met the requirements of atom economy for green chemistry. It was verified that H 2 O played a crucial role in this reaction. With this novel method, the simultaneously methylation and carbonylation of the aro-matic ring was achieved. The polyfunctionalized fused aromatic derivatives were generated by cascade HDDA reactions of different P-ylide and tetrayne substrates. We proposed a possible mechanism for the reaction: the tetrayne substrates form benzyne intermediates by self-cyclization and then react with the P-ylides to yield the fused aromatics compounds via an HDDA reaction. This method, which does not require a directing group, exhibits precise regioselectivity and can generate fused methyl 4-methyl-2,3-dihydro-1H-indene-5-carboxylates and ethyl 4-methyl-2,3-dihydro-1H-indene-5-carboxylates via a one-pot, multiterminal cycloaddition reaction (1a-1q) with excellent yields (Table 1). Consequently, this reaction provides an economical, efficient, and direct method for the synthesis of highly substituted ethyl 2-methylbenzoate and o-tolylethanone compounds. A reaction scheme for accessing carbon-bridged tetraynes with ylides in a catalyst-free manner was designed. The optimum reaction conditions in terms of water addition, temperature, solvent, and reaction time are discussed. Initially, with dry toluene as the solvent, substrate 1a was reacted with (carbomethoxymethylene)triphenylphosphorane at 100 °C. TLC was performed to monitor the reaction, and the starting material disappeared after 9 h. After separation and purification of the products, the reaction with a toluene : water ratio of 100 : 1 using (carbomethoxymethylene) triphenylphosphorane for 9 h at 100 °C yielded 10% of the desired product, and when the temperature was 90 °C (with a reaction time of 9 h), the yield was 80%. Because TLC did not fully reflect raw materials, the actual yield was 72%. Then, the temperature was increased to 110 °C and 120 °C with the same 9 h reaction time. TLC-based monitoring showed that the raw materials had been consumed, and the reaction had reached completion. We investigated the reactions in acetonitrile, toluene and cyclohexane and found that toluene was the most effective. The optimum reaction conditions for the tetrayne substrate (1 equiv.) with (carbomethoxymethylene)triphenylphosphorane (1.05 equiv.) and water (2 equiv.) were as follows: toluene 1.5 mL, 100 °C, 9 h. First, the effect of the structure of the tetrayne substrate on the yield was investigated. As shown in Table 1, a series of products (3a-3p) were obtained from the reactions of tetraynes with (carbomethoxymethylene)triphenylphosphorane or ethyl 2-(triphenylphosphoranylidene)acetate, and the yields ranged from 75% to 85%. The effect of different tetrayne substrate on the product yield was examined. When O i Pr, OEt and OMe were connected to the carbonyl carbon in the alkyne substrates, the yields were almost the same (3a (80%) and 3d (80%), 3i (79%), 3n (77%) and 3p (77%)). While the yields of 3c and 3h were 78% and 75%, respectively. Compounds containing benzene rings with fluorine, chlorine or other electronwithdrawing substituents exhibited lower yields than those with unsubstituted rings, for example, the yield of chloridecontaining substrate 3c was 78%, while that of the molecule with an unsubstituted benzene ring was 80% (3a); the yield of the substrate with a benzene ring bearing a fluoride (3o) was 73%, while the molecule with an unsubstituted benzene ring has offered a yield of 77% (3n). a Reaction scale: tetraynes 1 (1.0 equiv.), (carbomethoxymethyl ene)triphenylphosphorane 2 (1.05 equiv.), water (2.0 equiv.), toluene 1.5 mL, 100 °C. b Isolated yield. As shown in Table 2, yields of 76-87% obtained from the reactions of tetraynes with (acetylmethylene)triphenylphosphorane. When O i Pr, OEt and OMe were connected to carbonyl carbon of the tetrayne substrates, the yields were similar, such as those of 3q (80%), 3x (81%), 3r (82%) and 3v (81%). The yields with benzene ring-containing tetraynes with alkyl groups as electron donors were slightly higher than those of tetraynes directly bearing alkyl groups as electron donors. For example, the yields of 3r and 3s were 82% and 85%, respectively (Table 2), while the yields of 3t and 3u were 80% and 79%, respectively. The yields of benzene rings with alkyl substituents were higher than those of benzene rings without substituents. For example, the yield of 3s, with an n-propyl group on the benzene ring, was 87%, and the yield of 3r, with a methyl on the benzene ring, was 82%, while the yield of the unsubstituted benzene ring was 80% (3a). The structures of 3h and 3y were confirmed by X-ray diffraction. 16 Scheme 2 shows a possible mechanism of the reaction. Tetrayne substrate 1 formed a benzyne intermediate by selfcyclization and then reacted with a Wittig reagent to yield fused aromatic compounds B via nucleophilic addition reaction. The tetrayne substrate first undergoes an HDDA reaction to form benzyne intermediate A, and then, the carbanion in the phosphorus ylide attacks benzyne intermediate A to form four-membered ring intermediate B. 17 Triphenyl oxyphosphorous C is produced after the self-cyclization performance of intermediate B, affording the intermediate D via a 4π-electrocyclic ring opening process. 18 With the elimination of phosphine oxide, 19 E, which is the different resonance structure of D then transformed to product 3f. Deuterated benzene was initially used as the solvent to explore the source of the methyl hydrogen in the product. 1 H NMR spectroscopy confirmed that deuterated benzene was not the source of the hydrogen in the product, demonstrating the importance of D 2 O in this process. ## Conclusions In summary, we describe the first method for the synthesis of fully substituted benzenes, such as ethyl 2-methylbenzoates a Reaction scale: tetraynes 1 (1.0 equiv.), (acetylmethylene) triphenylphosphorane 2 (1.05 equiv.), water (2.0 equiv.), toluene 1.5 mL, 100 °C. b Isolated yield. and o-tolylethanones, via the reaction of acetylmethylenetriphenylphosphorane/carbomethoxymethylenetriphenylphosphorane and trace water polyynes. The formation of four new C-C bonds via a one-pot, multicomponent cascade led to the formation of benzyne intermediates by self-cyclization, and these intermediates then reacted with phosphorus ylides. The reactions produced highly substituted targets with excellent regioselectivity. Rare, fully substituted benzene derivatives were obtained in high yields under aerobic conditions. Future studies will be focused on the construction of directly methylated and carbonylated aryne precursors and on the development of a highly efficient pathway for the production of the ubiquitous carbonylated 2,3-dihydro-1H-indene and polyfunctional aromatic hydrocarbons.

chemsum

{"title": "Direct methylation and carbonylation of <i>in situ</i> generated arynes <i>via</i> HDDA-Wittig coupling", "journal": "Royal Society of Chemistry (RSC)"}

simultaneously_boosting_the_conjugation,_brightness_and_solubility_of_organic_fluorophores_by_using_

## Abstract: Organic near-infrared (NIR) emitters hold great promise for biomedical applications. Yet, most organic NIR fluorophores face the limitations of short emission wavelengths, low brightness, unsatisfactory processability, and the aggregation-caused quenching effect. Therefore, development of effective molecular design strategies to improve these important properties at the same time is a highly pursued topic, but very challenging. Herein, aggregation-induced emission luminogens (AIEgens) are employed as substituents to simultaneously extend the conjugation length, boost the fluorescence quantum yield, and increase the solubility of organic NIR fluorophores, being favourable for biological applications. A series of donor-acceptor type compounds with different substituent groups (i.e., hydrogen, phenyl, and tetraphenylethene (TPE)) are synthesized and investigated. Compared to the other two analogs, MTPE-TP3 with TPE substituents exhibits the reddest fluorescence, highest brightness, and best solubility. Both the conjugated structure and twisted conformation of TPE groups endow the resulting compounds with improved fluorescence properties and processability for biomedical applications. The in vitro and in vivo applications reveal that the NIR nanoparticles function as a potent probe for tumour imaging. This study would provide new insights into the development of efficient building blocks for improving the performance of organic NIR emitters. ## Introduction Near-infrared (NIR, >650 nm) fluorescence imaging has attracted considerable attention from both fundamental researchers and the clinical community, as it has lower autofluorescence interference, causes less photodamage, and has better penetration capability as compared with visible light, making it a preferable option for biological imaging. A lot of NIR fluorophores have been developed, for example, quantum dots, carbon nanomaterials, rare earth-doped nanoparticles (NPs), and organic emitters. Among them, organic materials hold great promise for clinical translation due to the salient merits of good biocompatibility, a well-defned chemical structure, facile modifcation, large-scale production, and potential degradation. Therefore, the development of efficient organic NIR emitters is highly important for advancing biomedical applications. 11,12 At present, organic NIR molecules face several limitations, such as short emission wavelengths, low brightness, unsatisfactory processability, and the aggregation-caused quenching (ACQ) effect. Moreover, it is very difficult to optimize all these important performance at the same time. Organic NIR chromophores are usually developed by extending the conjugation length and using a donor-acceptor (D-A) approach, however, both would cause strong intermolecular p-p interactions, and give rise to more pronounced processability/solubility and ACQ issues. The most popularly used strategy is the introduction of a lot of alkyl chains to increase the solubility and disrupt the intermolecular interactions. 19,20 Although effective, the bulky insulating groups would impact the photophysical properties adversely. 21,22 As a result, the introduction of a large amount of alkyl chains is a compromised or temporary selection, which calls for more efficient methods. After frst being coined by Tang and co-workers in 2001, aggregation-induced emission (AIE) has been considered as an effective solution to the notorious ACQ phenomenon. 23,24 AIE luminogens (AIEgens) are weak or non-luminescent in dilute solution but become highly emissive in the aggregate or solid state due to the restriction of intramolecular motion (RIM) mechanism. Freely rotating molecular rotors have been employed to endow the molecules with AIE signature, for example, hexphenylsilole (HPS) and tetraphenylethene (TPE). Many AIEgens with emission colours covering the entire visible spectral region and even the NIR range have been developed, and some of them exhibit great promise for biomedical applications. Nevertheless, the exploration of NIR AIEgens is suboptimal, and more studies are needed to optimize the emission wavelength and brightness simultaneously. One of the most serious obstacles for developing organic NIR luminogens is the low photoluminescence quantum yield (PLQY). According to the "energy gap law", the brightness of organic molecules usually decreases as the emission wavelength red shifts, especially in the NIR region, because the large vibronic coupling between the ground and excited states, and the non-radiative deactivation pathways become dominant when the electronic bandgap decreases. Red-shifting the emission wavelength and increasing the brightness at the same time are ideal for organic NIR bioprobes, but this is scarcely reported as it is indeed a challenging task. In this contribution, we report the simultaneous bathochromic emission, boosted PLQY, and increased solubility of organic NIR fluorophores by simply introducing AIEgens as the substituent groups, which is benefcial for biological imaging. By employing the typical AIE building block, TPE, as the substituent unit, the absorption/emission wavelength red shifts with the extension of the conjugation length, the brightness increases greatly because of the suppressed non-radiative processes, and the solubility also increases since the molecular rotors disrupt the intermolecular interaction. Both in vitro and in vivo experiments suggest that the bioimaging performance is enhanced for the probe with AIEgen-based molecular rotors. This work for the frst time demonstrates that AIEgens could signifcantly increase the conjugation, brightness and solubility of organic NIR luminogens, being better than the typically used alkyl chains, and represents a new strategy for developing highperformance fluorophores. ## Results and discussion The D-A structured molecules with methoxy-substituent TPE (MTPE) as the donor and thieno [3,4-b]pyrazine (TP) as the acceptor were synthesized, in which the electronic bandgaps could be tuned by changing the conjugation. TP is an electronwithdrawing group based on planar thiophene fused with pyrazine, which facilitates the intramolecular charge transfer (ICT) in D-A molecules. The MTPE unit exhibits stronger electron-donating properties than the widely used building block of AIEgen, TPE, and thus leading to a smaller bandgap and longer emission wavelength. Based on the molecular architecture of "MTPE-TP-MTPE", a series of compounds with different substituents (i.e., hydrogen, phenyl, and TPE) in the TP core (Scheme 1) were synthesized to study their influence on the molecular conjugation and conformation, and photophysical properties. The synthetic route of MTPE-TP1-3 is presented in Scheme 1. The Suzuki coupling reaction was carried out between 2-(4-(2,2-bis(4-methoxyphenyl)-1-phenylvinyl)phenyl)-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (4) and 2,5-dibromo-3,4-dinitrothiophene (5) to yield the dinitro molecule (6), followed by a reduction reaction to afford the diamine compound (7), which subsequently underwent cyclization with benzyl derivatives to obtain the TP structure. For MTPE-TP3, bromobenzene was frst attached to the TP core, and then reacted with 4,4,5,5-tetramethyl-2-(1,2,2-triphenylvinyl)-1,3,2-dioxaborolane (12) to construct the TPE group. The intermediates and fnal compounds have been characterized using 1 H NMR, 13 C NMR and high-resolution mass spectra (HRMS). Detailed synthesis processes and characterization are presented in the ESI (Fig. S1-S22 †). To gain deeper insight into the molecular conformation, density functional theory (DFT) calculations were performed. As depicted in Fig. 1, MTPE-TP1-3 possess rather twisted molecular geometries, in which the intramolecular rotors would dissipate the excited-state energy through the free highfrequency rotation of phenyl rings in solution, and the ACQ effect would be signifcantly inhibited in the aggregation/solid form, being favourable for realizing the AIE signature. 40,41 Interestingly, it is obvious that these three molecules have different conjugations. For instance, the dihedral angles between TP and adjacent phenyl rings are 24.7 /26.1 , 36.5 / 29.5 , and 40.4 /40.1 for MTPE-TP1-3, respectively. The increased backbone distortion is probably due to the steric resistance between MTPE and substituents in the pyrazine ring, which would impact the photophysical properties. 42,43 The highest occupied molecular orbital (HOMO) is distributed in both MTPE and TP units (Fig. 1), suggesting good intramolecular conjugation. While the electron cloud of the lowest unoccupied molecular orbital (LUMO) is mainly located in the TP core, indicating an obvious intramolecular D-A interaction and ICT effect from MTPE to TP. 44,45 It is noted that the electron density of the LUMOs of MTPE-TP2,3 is also distributed in the phenyl rings attached to pyrazine, which is an indicator of better conjugation, and thus smaller bandgaps can be expected. We frst evaluated the solubility, since processability is of vital importance for application of organic molecules. The solubilities of MTPE-TP1-3 in dimethyl sulfoxide (DMSO) are 0.58, 0.36, and 1.22 mg mL 1 , respectively. The lower solubility of MTPE-TP2 as compared with MTPE-TP1 can be ascribed to the extended conjugation, while the highest solubility of MTPE-TP3 is likely due to the introduction of violently rotated TPE substituents. The absorption spectra of MTPE-TP1-3 in DMSO (Fig. 2a) show the maximum absorption in sequence at 518 nm, 538 nm, and 543 nm, which are in good agreement with the calculation results that the conjugations of MTPE-TP2,3 are better than that of MTPE-TP1, and the slightly longer absorption of MTPE-TP3 is probably attributed to the contribution of conjugated TPE side groups. As shown in Fig. S23, † the photoluminescence (PL) spectra exhibit a similar trend to the absorption change, and MTPE-TP3 shows the reddest emission. Of note, although the planarity/conjugation between MTPE and TP decreases from MTPE-TP1 to MTPE-TP3 for the increased dihedral angles, the introduction of phenyl and TPE substituents on TP contributes more to the molecular conjugation. We next investigated the fluorescence properties by adding water (poor solvent) into DMSO solution (good solvent) of the compounds. The PL intensity decreases in low water fractions and then intensifes (Fig. 2b, c and S24 †). The decrease of PL intensity and concurrent redshift of the emission wavelength can be explained by the solvatochromic effect in polar solvents (Fig. S25 †), which is usually observed in D-A type compounds. 46,47 The intensifed PL intensity is due to the formation of aggregates, representing a typical AIE signature. The AIE amplitude increases from MTPE-TP1 to MTPE-TP3, especially there is a notable enhancement for MTPE-TP3, which is likely due to the more twisted molecular geometry and inhibition of intermolecular interactions in the aggregate state. We further studied the PL properties in different viscosity environments, as molecular motions would be restricted in high viscosity. 48,49 The PL spectra and corresponding intensity changes in dimethylformamide (DMF)/glycerol mixtures with various glycerol fractions are presented in Fig. 2d, e and S26. † When increasing the glycerol fractions, the PL intensity frstly decreases and then rises. This phenomenon could also be explained by the solvatochromism and AIE effect for the high polarity and viscosity of glycerol. In high glycerol fractions, the viscosity becomes rather high, and the molecular motions are restricted greatly, which could effectively suppress the energy consumption via non-radiative deactivation. 50,51 The PL spectra of solid powders (Fig. 2f) suggest a bathochromic shift compared to solution states and very bright emission in the NIR region beyond 650 nm. To endow the hydrophobic compounds with good water solubility and biocompatibility, the nanoprecipitation method (Fig. 3a) was adopted to formulate the self-assembled AIE NPs. With the assistance of an amphiphilic polymer surfactant, Pluronic F-127, uniform and stable organic NPs were obtained. The morphology and size of the NPs were characterized by dynamic light scattering (DLS) and transmission electron microscopy (TEM). DLS measurements reveal that the NPs formed by MTPE-TP1-3 (NPs1-3) exhibit similar average diameters of about 110 nm (Fig. 3b, S27, † and Table 1), while TEM images suggest a kind of spherical morphology with an average diameter of about 90 nm. As compared to the solution state, both the absorption and emission spectra of NPs redshift (Fig. 3c and d). The PL maxima of NPs1-3 are 660 nm, 678 nm, and 685 nm, respectively, which are located in the NIR tissue transparent window and favourable for biological imaging. The large Stokes shift of about 130 nm suggests low self-absorption, a quality that is highly desirable for fluorescence imaging. PL excitation (PLE) mapping of NPs3 (Fig. 3e) indicates that it can be excited efficiently by the available excitation light (e.g., 535 nm) of an in vivo imaging system (IVIS). The photographs of the NP solutions under 365 nm UV light irradiation reveal very bright emission. As shown in Fig. 3f and Table 1, the PLQYs of NPs1-3 are measured to be 9.6%, 12.1%, and 20.7%, respectively. The comparison of some organic/polymer NIR NPs with similar emission wavelengths is displayed in Table S1, † suggesting high brightness of NPs3. The PL intensity obtained from the IVIS and corresponding images (Fig. 3g) also manifests that the fluorescence brightness follows the order: NPs3 > NPs2 > NPs1. It is worth noting that from MTPE-TP1 to MTPE-TP3, the PL maximum red shifts for 25 nm, and the PLQY exhibits a more than two-fold increase. This phenomenon is different from most previous results which show that the non-radiative decay usually increases when the bandgap is reduced. In this work, the introduction of AIE blocks into the TP unit and therefore the hindrance of molecular motions in the aggregate form is considered to be the main reason for the high brightness. In order to uncover the photophysical processes of the simultaneously red-shifted emission and enhanced brightness, we measured the fluorescence lifetime of encapsulated NPs. As depicted in Fig. 3h, the lifetime increases from NPs1 to NPs3, which is in the same trend as PLQY. The fluorescence properties are closely linked to radiative and non-radiative decay rates (k r and k nr ) from the excited state to ground state. Their relationships can be expressed as k r ¼ F F /s, and k nr ¼ 1/s k r , where F F is the PLQY and s is the fluorescence lifetime. Accordingly, the radiative/non-radiative decay rate constants of NPs1-3 are calculated to be 0.49/4.61 10 8 s 1 , 0.54/3.88 10 8 s 1 , and 0.64/2.43 10 8 s 1 , respectively (Table 1). Interestingly, the radiative decay rate increases slightly, while the non-radiative decay rate decreases a lot. The twisted molecular structure is benefcial for enhanced radiative decay, and more importantly, signifcantly inhibits the non-radiative process. 52,53 This result clearly demonstrates that the non-radiative deactivation pathway is greatly suppressed by the restriction of molecular motions, which is effective for realizing highly bright luminogens. 54 After studying the photophysical properties, we next exploited the application of the water-soluble NIR NPs in biological imaging both in vitro and in vivo. First, their potential in cellular imaging was studied with confocal laser scanning microscopy (CLSM). 4T1 breast cancer cells were respectively incubated with NPs1-3 in the same concentration (20 mM) for 4 h, and then the cells were fxed and the cell nuclei were stained with 4 0 ,6-diamidino-2-phenylindole (DAPI). As depicted in Fig. 4, an intense red signal is obviously observed in the cell cytoplasm, demonstrating successful uptake of the NPs. It is noted that relatively weak red fluorescence is observed for the NPs1-treated cells, whereas the cells incubated with NPs3 at the same concentration exhibit robust red fluorescence signals. This result is in line with the PL brightness of the NPs. Moreover, little side effect in cell viability is observed as more than 90% of cells remain alive after treating with a high concentration (50 mM) of NPs (Fig. S28 †). These results demonstrate that the bright NIR AIEgens, especially NPs3, offer great promise for fluorescence cell imaging. To further assess the in vivo imaging performance of AIE NPs, non-invasive whole-body fluorescence imaging was conducted on 4T1 tumour-bearing mice. In this study, the same dose of NPs1-3 (200 mL, 500 mM based on AIEgens) was intravenously injected into tumour-bearing mice through the tail vein, respectively. As displayed in Fig. 5a, the tumour tissues signifcantly light up by the NPs, which clearly delineates the tumour from the surrounding normal tissues. In particular, the NPs3-treated mice exhibit much stronger tumour fluorescence than that of NPs1 and NPs2, which has also been confrmed by the ex vivo imaging results of resected tumours (Fig. S29 †). The quantitative analysis (Fig. 5b) verifes that the fluorescence intensity of NPs3 in the tumour site is about 1.9-fold higher than that of NPs2, and 3.6 times greater than that of NPs1, which is consistent with the in vitro cellular imaging results and PLQY data. The best tumour imaging performance of NPs3 is mainly ascribed to the longer emission wavelength and higher PLQY. This result manifests that MTPE-TP3 can be used as a highly efficient NIR probe for realizing high-contrast biomedical imaging in vivo. After verifying the greater potency of NPs3 for bioimaging than the other two NPs, we next explored the detailed timedependent tumour accumulation and imaging performance in vivo. The fluorescence images of mice were recorded at different S2. time points after the administration of NPs3. As shown in the representative images in Fig. 5c, and the corresponding fluorescence intensity of the tumour site in Fig. 5d, the NPs exhibit a tendency to accumulate into the tumour with time elapsing, suggesting a good tumour preferential profle. The maximal tumour signal is observed at about 8 h post tail vein injection, which then gradually decreases. However, even after 24 hours, the tumour site is still highly visualized, and the strong fluorescent signal makes the tumour clearly distinguishable from normal tissues. The efficient accumulation of AIE NPs at the tumour site is mainly due to the enhanced permeability and retention (EPR) effect. The results show that the bright NIR NPs obtained through rational molecular design hold great promise for precise non-invasive tumour diagnosis in a high contrast manner. ## Conclusions In summary, we demonstrate for the frst time that AIE building blocks (e.g., TPE) could efficiently extend the conjugation, boost the brightness, and enhance the solubility of organic NIR fluorophores at the same time. After introducing two TPE groups, the absorption/emission wavelength red shifts with the extension of the conjugation length, the brightness increases greatly because of the suppressed non-radiative processes, and the solubility also increases since the molecular rotors disrupt the intermolecular interaction, suggesting that the AIE unit works better than the typical solubility enhancer, bulky alkyl chains. Both in vitro cellular imaging and in vivo tumour imaging results verify that NIR AIE NPs are efficient for biomedical imaging, and the reddest and brightest MTPE-TP3 NPs give the best tumour diagnostic outcome. This work highlights that AIEgens could serve as a potent regulator to obtain an optimal emission wavelength, brightness, and solubility of organic emitters, being useful for designing molecular probes. ## General methods All the chemicals and reagents were purchased from chemical sources and were used as received. Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker AV 400 spectrometer. High-resolution mass spectra (HRMS) were measured with a GCT premier CAB048 mass spectrometer in matrix assisted laser desorption ionization-time of flight (MALDI-TOF) mode. The geometry optimization was performed at the level of B3LYP/ 6-31G* using the density functional theory (DFT) method with the Gaussian 09 program package (Cartesian coordinates see Tables S3-S5 †). The UV-vis absorption spectra were recorded using a Shimadzu 2550 UV-vis scanning spectrophotometer. Steady-state photoluminescence (PL) spectra were recorded on a Horiba Fluorolog-3 spectrofluorometer. The PLQY was measured using a Hamamatsu absolute PL quantum yield spectrometer C11347 Quantaurus-QY. Transient PL at room temperature was measured using an Edinburgh FLSP980 fluorescence spectrophotometer. Dynamic light scattering (DLS) was measured on a 90 plus particle size analyser. Transmission electron microscope (TEM) images were acquired from a JEM-2010F transmission electron microscope with an accelerating voltage of 200 kV. ## Synthesis of MTPE-TP1 2,5-Bis(4-(2,2-bis(4-methoxyphenyl)-1-phenylvinyl)phenyl)thiophene-3,4-diamine (0.54 g, 0.6 mmol) and 1,4-dioxane-2,3-diol (0.12 g, 1 mmol) were dissolved in a mixture of chloroform (20 mL) and acetic acid (20 mL) in a 100 mL flask. The reaction mixture was heated to 60 C, and stirred for 12 h. Then water was added, and the mixture was extracted with dichloromethane three times. The organic phase was combined and dried with anhydrous MgSO 4 . After the removal of the solvent under reduced pressure, the residue was purifed by column chromatography on silica gel using dichloromethane/hexane (v/ v 1 : 2) as the eluent to afford 5 ## Synthesis of MTPE-TP2 2,5-Bis(4-(2,2-bis(4-methoxyphenyl)-1-phenylvinyl)phenyl)thiophene-3,4-diamine (0.54 g, 0.6 mmol) and benzil (0.21 g, 1 mmol) were dissolved in a mixture of chloroform (20 mL) and acetic acid (20 mL) in a 100 mL flask. The reaction mixture was heated to 60 C, and stirred for 12 h. Then water was added, and the mixture was extracted with dichloromethane three times. The organic phase was combined, and dried with anhydrous MgSO 4 . After the removal of the solvent under reduced pressure, the residue was purifed by column chromatography on silica gel using dichloromethane/hexane (v/v 1 : 2) as the eluent to afford 5,7-bis(4-(2,2-bis(4-methoxyphenyl)-1-phenylvinyl) phenyl)-2,3-diphenylthieno [3,4-b]pyrazine (MTPE-TP2) as a dark red solid (72% yield). 1 Preparation of the NPs 1 mM of AIEgens and 4 mg of Pluronic F-127 were dissolved in 1 mL of tetrahydrofuran (THF). The obtained THF solution was poured into 10 mL of deionized water under sonication with a microtip probe sonicator (XL2000, Misonix Incorporated, NY). Subsequently, the mixture was sonicated for another 1 min and violently stirred in a fume hood overnight at room temperature to evaporate residue THF, and the NP solution was used directly. The concentration of the NPs was estimated based on the initial feeding AIEgen in the nanoprecipitation process, i.e., the concentration of AIEgens in the solution. ## Cell culture 4T1 breast cancer cells were cultured in Dulbecco's Modifed Eagle's Medium (DMEM) containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37 C in a humidifed environment containing 5% CO 2 . Before experiments, the cells were precultured until confluence was reached. ## Cellular imaging 4T1 breast cancer cells were seeded and grown on a 35 mm Petri dish with a cover slip at a density of about 2 10 5 cells per well in 2 mL of culture medium. The cells were incubated respectively with NPs1-3 (fnal concentration: 20 mM based on AIEgens) in the medium for 4 h. Then the cells were washed with 1 PBS three times, fxed in 4% paraformaldehyde for 20 min, washed three times with 1 PBS, and incubated with 4 0 ,6-diamidino-2-phenylindole (DAPI) for 10 min. The cells were then washed three times with 1 PBS. Fresh PBS was added to the confocal chambers, and laser scanning confocal microscopy (Zeiss LSM 710, Jena, Germany) was performed (l ex ¼ 405 nm for DAPI, l ex ¼ 543 nm for NPs1-3; fluorescent signals were collected at 630-760 nm for NPs1-3 and 430-475 nm for DAPI, respectively). ## Cytotoxicity study 3-(4,5-Dimethyl-2-thiazolyl)-2,5-diphenyl-tetrazolium bromide (MTT) assay was used to evaluate the cytotoxicity of the NPs. 4T1 breast cancer cells were harvested in a logarithmic growth phase and seeded in 96-well plates at a density of about 4 10 4 cells per well for 24 h. Then different concentrations (0, 2.5, 5, 10, 20, and 50 mM based on AIEgens) of NPs1-3 were added into the cell culture medium separately. After incubating for 24 h, the culture medium was replaced with a fresh medium containing 0.5 mg mL 1 of MTT, and incubated for 4 h. The culture medium was discarded and replaced with 100 mL of DMSO with gentle shaking. Afterwards, the absorbance of MTT at 490 nm was measured using a microplate reader (GENios Tecan). Cell viability was expressed by the ratio of the absorbance of cells incubated with NPs to that of the cells incubated with the culture medium only. ## Animal experiments All animal studies were conducted under the guidelines set by the Tianjin Committee of Use and Care of Laboratory Animals, and the overall project protocols were approved by the Animal Ethics Committee of Nankai University. Tumour-bearing mice 6 week-old female BALB/c mice were obtained from the Laboratory Animal Centre of the Academy of Military Medical Sciences (Beijing, China). To establish the xenograft 4T1 tumour-bearing mouse model, 4T1 breast cancer cells (1 10 6 ) suspended in 30 mL of RPMI-1640 medium were injected subcutaneously into the right axillary space of the BALB/c mouse. After 9 days, the mice with tumour volumes of about 80-120 mm 3 were used subsequently. ## In vivo uorescence imaging The xenograft 4T1 tumour-bearing mice were randomly selected for the following fluorescence imaging experiments. The tumour-bearing mice were anesthetized using 2% isoflurane in oxygen, and the NPs (200 mL, 500 mM based on AIEgens) were intravenously injected into the tumour-bearing mice using a microsyringe, respectively (n ¼ 3 mice for each group). In vivo fluorescence imaging was performed on a Maestro EX fluorescence imaging system (CRi, Inc.) with 535 nm excitation and signal collection in the spectral region of 600-850 nm.

chemsum

{"title": "Simultaneously boosting the conjugation, brightness and solubility of organic fluorophores by using AIEgens", "journal": "Royal Society of Chemistry (RSC)"}

how_important_are_dispersion_interactions_to_the_strength_of_aromatic_stacking_interactions_in_solut

## Abstract: In this study, the contributions of London dispersion forces to the strength of aromatic stacking interactions in solution were experimentally assessed using a small molecule model system. A series of molecular torsion balances were designed to measure an intramolecular stacking interaction via a conformational equilibrium. To probe the importance of the dispersion term, the size and polarizability of one of the aromatic surfaces were systematically increased (benzene, naphthalene, phenanthrene, biphenyl, diphenylethene, and diphenylacetylene). After correcting for solvophobic, linker, and electrostatic substituent effects, the variations due to polarizability were found to be an order of magnitude smaller in solution than in comparison to analogous computational studies in vacuo. These results suggest that in solution the dispersion term is a small component of the aromatic stacking interaction in contrast to their dominant role in vacuo. ## Introduction Aromatic stacking interactions play a key role in determining the stability, activity, and utility of many supramolecular processes such as the structure of biopolymers, 1-3 host-guest complex stability, and the selectivity of asymmetric catalysts. The importance and utility of aromatic stacking interactions have provided the motivation to study the fundamental nature of the interaction and to develop models that can accurately predict their stability trends. 11,12 Thus, the influence of variables to the strength of stacking interactions such as charge, 13,14 substituent effects, 15,16 and solvent effects 17,18 has been an active area of research. The goal of this study was to experimentally assess the role that dispersion interactions play in the aromatic stacking interaction in solution using a dynamic small molecule model system (Scheme 1). Our approach was to systematically vary the size of one of the aromatic surfaces involved in an aromatic stacking interaction. This strategy mirrors that of a computational study by Zeinalipour-Yazdi and Pullman, 22 which predicted a dramatic strengthening of the aromatic stacking interaction with increasing size of an aromatic surface due to an increase in the dispersion term. Attractive London dispersion interactions are weak attractive interactions that can form between both polar and non-polar molecular surfaces. 12,23 In the gas-phase, dispersion interactions have been identifed as the dominant contributing term for aromatic stacking interactions. For example, Sherrill's component analysis estimated that the dispersion term comprises 61% of the overall stacking energy for the benzene dimer. 27 However, the role of dispersion forces in solution has been much more controversial. Specifcally, the dispersion contributions in solution have been proposed to be much smaller. 28 The rationale is that there are roughly an equal number of dispersion interactions on either side of the stacking equilibrium (Scheme 1). The aromatic surfaces still form attractive dispersion interactions in the stacking complex. Scheme 1 Representation of the intramolecular aromatic stacking interaction in the folded conformer of a molecular torsion balance model system and the influence of solvent molecules (red spheres) on the stability of the folded-unfolded conformational equilibrium. However, solvent molecules form additional dispersion interactions on either side of the binding equilibrium with the uncomplexed aromatic surfaces and with each other. 28 Thus, the question is whether the net dispersion interactions on the right-hand side of the equilibrium are stronger than those on the left-hand side of the equilibrium. The measurement of dispersion interaction in solution has posed a number of experimental challenges. 29,30 First, the dispersion contributions are expected to be small, and thus a very sensitive method with sub kcal mol 1 accuracy is required. Second, in contrast to electrostatic or solvent trends, it is difficult to systematically vary the dispersion term of a non-covalent interaction. Third, dispersion interactions are very difficult to differentiate from solvophobic interactions because both scale with increasing size of the aromatic surfaces. 31,32 Thus, studies that have observed a correlation between the size of the aromatic surface and the strength of the stacking interaction could be attributed to solvophobic or dispersion effects. In this study, a small molecule model system was designed to specifcally address the above challenges. 36,37 First, the model system is an example of a "molecular torsion balance", which has been demonstrated to provide a very accurate and sensitive measure of non-covalent interactions. 37 Variations in the strength of the intramolecular interaction as small as AE0.03 kcal mol 1 can be measured by monitoring their influence on the folded-unfolded equilibrium (Scheme 1). Second, the dispersion term was systematically varied by increasing the conjugation length and polarizability of one of the interacting surfaces. Dispersion interactions are known to increase with increasing molecular polarizability because dispersion interactions are the result of the electrostatic attraction between polarizable molecular surfaces. 22,42,43 Third, the dispersion effects were differentiated from the solvophobic effects by keeping the contact area between the two stacking aromatic surfaces constant. The rigid bicyclic framework of the molecular balances fxes the geometry and contact area of the aromatic surfaces in 1a-f (Fig. 2a). Only the frst benzene ring of the aromatic arm, regardless of its size, was in contact with the phenanthrene shelf. Thus, the extended aromatic surfaces in 1b-f did not form any additional stacking or solvophobic interactions. An additional advantage of these molecular systems was that the results could be directly compared with computational studies. 22,40,44 These computational studies provided theoretical in vacuo benchmarks to compare the magnitudes of our experimentally measured trends in solution. The most similar computational studies were by Zeinalipour-Yazdi and Pullman, which measured the stacking energy of a benzene with aromatic surfaces of varying size. 22 The molecular balances performed an analogous comparison as the outer most benzene ring of the phenanthrene shelf forms stacking interactions with the aromatic arms of varying size. In the computational studies, the stacking energies of the benzene unit were found to systematically increase with the increasing size and polarizability of the opposing aromatic surface. A steep linear correlation was predicted between the size of the opposing aromatic surfaces and the stacking energies. This trend is consistent with the dispersion term representing a signifcant portion of the stacking energy in vacuo. For example, the stacking interaction energy of the benzene-naphthalene complex was 2.3 times larger than the benzene-benzene complex. Similarly, the stacking interaction energy of the benzene-anthracene complex was 3.9 times larger than the benzene-benzene complex. ## Molecular balance design The rigid bicyclic N-arylimide framework of the molecular balances utilized in this study had been previously employed to study a range of non-covalent interactions such as aromatic stacking, CH-p, and cation-p interactions. In particular, substituted derivatives of balances 1a were able to accurately measure small differences in the aromatic stacking interactions (<0.05 kcal mol 1 ) due to substituent effects. 45 For the measurement of weak non-covalent interactions, this molecular balance framework has a number of attractive features. First, restricted rotation of the N-arylimide rotor leads to the formation of distinct folded and unfolded conformers that are in equilibrium at room temperature. Second, the bicyclic framework holds the aromatic surfaces of the arm and shelf at different distances in the two conformers. In the folded conformer, the arm and shelf surfaces are in close proximity allowing formation of an intramolecular off-set stacking interaction. In the unfolded conformer, the arm and shelf surfaces are held apart and cannot form a stacking interaction. Thus, the folded/unfolded equilibrium ratio provides a very sensitive measure of variations in the strength of the intramolecular interactions. A strengthening of the intramolecular stacking interactions is evident by a shift in the folded/unfolded ratio towards the folded conformer. We have previously demonstrated that the parent balance 1a with a phenyl arm forms a well-defned off-set stacking interaction in the folded conformer (Fig. 2a). 40,41 X-ray and NMR analyses of 1a found that the phenyl arm and phenanthrene shelf adopt a parallel stacking geometry in the folded conformer. The phenyl ring of the arm is in contact with outer most ring of the phenanthrene shelf with an arm centroid-toshelf plane of 3.75 . Furthermore, the proximity of the arm and shelve surfaces and the rigidity of the bicyclic framework do not provide sufficient freedom and space to form the alternative arene-arene geometries such as the perpendicular edge-to-face and T-shaped geometries. The rigidity of the balance framework also ensures that the extended aromatic surfaces of the arms in balances 1b-f (represented as dotted lines in Fig. 2a) should not form additional stacking interactions with the phenanthrene shelf. In addition, the contact area between the arm and shelf surfaces should remain constant despite the variations in the size of the aromatic arms. We have also confrmed that the parent control balance 2a with the phenyl arm is unable to form an intramolecular stacking interaction in the folded conformer due to its shorter benzene shelf (Fig. 2b). 41 The absence of stacking interactions in 2a was confrmed by NMR and modeling studies. Thus, control balances 2a-f should provide a measure of the other factors that influence the folded-unfolded equilibria such as solvent, dipole, linker, and secondary interaction effects. The subtraction of the folding energy of 2 from the folding energy of 1 should isolate the intramolecular stacking energy. For this study, six balances (1a-f) and six control balances (2a-f) were prepared via previously described synthetic routes. 40,41 The six aromatic arms (a-f) include aromatic surfaces of varying size, conjugation length, and polarizability. These aromatic surfaces fell into three groups (Fig. 1). The frst was the unsubstituted phenyl arm (a) that had the smallest common aromatic surface. The next were the fused aromatic surfaces with naphthyl and phenanthryl arms (b and c). The last group was the non-fused aromatic surfaces (d-f). These include the biphenyl (d), stilbene (e), and diphenylethynyl (f) arms, which extend the conjugation of the parent phenyl ring from a single substitution point at the para-position. ## Verication and measurement of the intramolecular stacking interactions The formation of the expected intramolecular stacking interactions within the new balances 1b-f was established by comparison of their 1 H NMR spectra (CDCl 3 , 298 K) with those of the parent balance 1a and control balances 2a-f. The NMR analyses were facilitated by a separate set of peaks for the folded and unfolded conformers due to slow exchange on the NMR time scale. The frst indication of stacking interactions in 1b-f was the observation of the expected upfeld shifts of the aromatic arm and shelf protons. Due to the proximity of the arm and shelf aromatic surfaces in the stacked structure, upfeld shifts of up to 1.0 ppm were observed in the folded versus the unfolded conformers. The direction and magnitude of these peak shifts were identical to those observed in the parent stacking phenyl balance 1a. 41 By comparison, these same aromatic protons did not display upfeld shifts in the folded versus unfolded conformers of control balances 2a-f, which cannot form intramolecular stacking interactions. The formation of stacking interaction was also evident from a comparison of their folding energies. The folded/unfolded ratios of 1a-f and 2a-f and their corresponding folding energies were measured from their peak areas in the 1 H NMR spectra (Table 1). The folding energies of the stacking balances 1a-f were consistently stronger (more negative) than the folding energies of the corresponding control balances 2a-f. This was consistent with the stabilization of the folded conformers of 1a-f by the formation of attractive stacking interactions. The intramolecular stacking interactions were estimated from the difference in the folding energies of 1a-f and 2a-f (DDG 1-2 ). The stacking energies ranged from 0.92 to 1.33 kcal mol 1 , which were comparable with previous measurements of stacking interactions of benzene surfaces in organic solution. 15 Next, the influence of the different sized aromatic arms on the stacking energies was examined. This analysis suggested that the dispersion contributions to the stacking interaction are not dominant in solution. These conclusions were based on two observations. First, the stacking energies in 1a-f were very similar despite the large variations in size of the arm surfaces. The DDG 1-2 values spanned a relatively narrow range from Table 1 1 H NMR measured folding energies of balances 1 and 2 (DG 1 and DG 2 ), the aromatic stacking energies (DDG 1-2 ), and characteristics of the arm aromatic surfaces (polarizability, ESE) 0.92 to 1.33 kcal mol 1 . Overall, these variations were an order of magnitude smaller than those predicted by the computational studies. 22 For example, the stacking interaction energy of the naphthyl arm in 1b was only 26% greater than the phenyl arm in 1a (1.16 versus 0.92 kcal mol 1 ). By comparison, the computational studies predicted a 230% increase in the stacking energies of naphthalene versus benzene surfaces. 22 Second, no clear correlation was observed between the stacking energies and the polarizabilities of the respective arms. The polarizabilities of the aromatic surfaces in the arms were estimated using computational methods (B3LYP, 6-31G*) (Table 1). 47 The most polarizable arms such as the diphenyl acetylene and stilbene did not show the strongest stacking energies. More convincingly, a plot of the calculated polarizabilities versus the measured stacking energies (DDG 1-2 ) did not show a clear correlation (Fig. 3). The stacking energies of the different sized arms appeared to strengthen (more negative) with increasing polarizability of the fused arms (b and c). However, an inverse correlation was observed between polarizability and stacking energy for the non-fused arm (d-f). ## Explanations for the inability to observe the dispersion contributions to the stacking energy trends The inability to observe the dispersion of the stacking interaction energies in solution was consistent with the hypothesis that the dispersion contributions would be smaller in solution because of the counter-balancing dispersion interactions of the solvent molecules. However, alternative explanations were also explored. First, the possibility was that the variations in the stacking energies were within the error for the measurement. The standard deviation of the stacking energies was 0.15 kcal mol 1 . While this value is small, it is greater than the error for the analysis which was estimated to be AE0.04 kcal mol 1 . The second possible explanation was that the variations in the stacking energies were due to electrostatic substituent effects (ESEs). Substituents on aromatic rings have been shown to stabilize and destabilize the aromatic stacking interaction in computational and experimental studies. 15,16, Along these lines, we have previously characterized the electrostatic substituent effects (ESE) of this specifc stacking model system. 45 This allowed us to estimate the influence of the substituent effects and to test whether the substituent effects can explain the observed minor variations in stacking interaction energies. To assess the ESEs in this system, the extended conjugation of arms b-f was classifed as meta-and/or para-substituents on the core phenyl arm a (Table 2). Arms d-f were treated as monosubstituted phenyl rings. For example, the biphenyl arm in balance 1d was categorized as a phenyl ring with a paraphenyl substituent. The fused naphthyl and phenanthryl arms b and c were treated as disubstituted phenyl rings with one metaand one para-substituent. The expected stabilizing or destabilizing ESEs were calculated based on the Hammett s meta parameters for respective substituents in Table 2 and the slopes of the previously measured Hammett plots for this balance system. 51 The details of this calculation are provided in the ESI. † For the disubstituted arms (b and c), the ESEs were calculated as the sum of the individual substituent effects. This analysis is based on the recent fnding that the substituent effects for stacking interactions are additive. 45 The estimated ESEs were able to explain half of the variation in the DDG 1-2 values. The ESEs for arms b-f were mostly stabilizing, which was consistent with the observed stronger stacking energies for the extended arms b-f. The vinyl, phenyl, styrene, and phenylacetylene substituents in the arms are all weak electron withdrawing groups with small positive Hammett s meta values (0.03 to 0.14). Electron withdrawing substituents have been shown to stabilize stacking interactions due to the formation of attractive electrostatic interactions. 52 To assess the importance of the substituent effects, the predicted ESE values were subtracted from the measured DDG 1-2 values to give a substituent corrected stacking energy (DDG 1-2 ESE). The corrected stacking energies (0.92 to 1.11 kcal mol 1 ) had approximately half the variation than the uncorrected stacking energies (0.92 to 1.33 kcal mol 1 ). The third alternative explanation that was examined was that the substituent effects had been obscuring the smaller dispersion effects. To test this possibility, the correlation between the substituent corrected stacking energies and polarizability was examined (Fig. 4). The plot of DDG 1-2 ESE versus polarizability was relatively flat, as the majority of variance had been removed. However, the remaining variance did not show a correlation with the polarizabilities of the aromatic surfaces. The fnal explanation for the inability to observe the dispersion effects was that the experimental design has not properly isolated the stacking energy from the folded-unfolded equilibrium energies. Of particular concern were solvophobic effects, which have been cited as the dominant term for the weak non-covalent interactions of non-polar surfaces in solution. 53,54 However, there were two observations that suggested that solvent and solvophobic effects had been effectively isolated in this study. First, relatively small differences in the stacking energies in 1a-f were observed despite the large variation in the size of the aromatic arms. Thus, the geometric constraints in the balance framework appear to have been effective in keeping the surface area contact and solvophobic effects constant for the series. Second, the folding energies of balances 1a-f and 2a-f were measured in two additional solvent systems. The folding energies of balances 1a-f and control balances 2a-f were measured in a more polar, acetone-D6, and a less polar, bromobenzene-D5, solvent (see ESI †). The overall trends and conclusions were analogous to those observed in CDCl 3 , suggesting that the solvent effects were not the reason for the inability to observe dispersion effects. The uncorrected (DDG 1-2 ) and corrected (DDG 1-2 ESE) stacking energies in these two additional solvents had relatively small variations and did not show any clear correlation with the polarizabilities of the aromatic surfaces (ESI Fig. S22 and 23 †). ## Conclusions In this study, we designed a series of molecular torsion balances 1a-f to assess the importance of dispersion interactions to the aromatic stacking interactions in solution. These model systems measured the strength of an intramolecular stacking interaction via changes in a folded-unfolded conformational equilibrium. The contribution of the dispersion term was assessed by systematically varying the size and polarizability of one of the aromatic surfaces and measuring the effect on the stacking energies. Through the use of control systems 2a-f, geometrical constraints, and studies in multiple solvents, the stacking interaction energies were separated from other factors that influence the conformational equilibrium such as solvophobic, dipole, linker, and steric effects. No correlation was observed between the polarizabilities of the aromatic surfaces and the stacking energies. There was relatively little variance in the strengths of the stacking energies despite the wide range in the sizes and conjugation lengths of the aromatic surfaces. These results suggest that the dispersion contributions to the aromatic stacking interaction in solution are relatively minor. The approach and conclusions of this study nicely complement studies of the origins of alkyl-alkyl interactions in solution. 53 Cockroft and co-workers used a perpendicular approach of assessing the contributions of dispersion interactions in solution. Instead of varying the size and polarizability of the interacting surfaces, they systematically varied the solvent environment, which allowed them to measure and subtract out the solvent effects from the overall interaction energy. Although the approach was different, the conclusions were similar to those in this study as the dispersion term of the interaction could not be observed in solution. The majority of the interaction energy was attributed to the solvent and solvophobic effects. A more recent study by Cockroft was able to measure the dispersion contributions to the interaction energy of non-polar surfaces in organic solution. However, the dispersion terms in non-polar organic solvents were small and were on the same order as the solvophobic interactions. 55 The conclusion that dispersion interactions do not play a dominant role in stacking interactions in solution is in contrast to computational studies and gas-phase studies, 56 which have found a dramatic correlation between the stacking energy versus the size and polarizability of an aromatic surface. 22 It is important to note that the smaller influences of dispersion interactions in solution is not due to an absence of dispersion interactions. Aromatic surfaces still form attractive dispersion interactions in solution just as they do in vacuo. However, the aromatic surfaces also form dispersion interactions with solvent molecules, which attenuates the overall magnitude of the dispersion term. ## Experimental section Synthesis of balances 1a-f and control balances 2a-f Balance 1a-f and control balance 2a-f were synthesized as previously described. 41 The balances and control balances were prepared by the following general route. First, the phenol corresponding to the arm was reacted with 2-fluoronitrobenzene in an S N Ar reaction that formed the diphenyl ether. The nitro-ether was reduced to the amino-ether and then reacted with the endo-bicyclic anhydride to yield the balance or control balance. Specifc procedures and characterization data for 1a-f and 2a-f are provided in the ESI. † ## Measurement of the folding energies The folded/unfolded ratios were measured by integration of the 1 H NMR spectra at 25 C. The peak areas of the singlets corresponding to the succinimide methine protons were measured by line ftting analysis. The folding energies were calculated from the equation DG ¼ RT ln([folded]/[unfolded]). The error in the folding energies was estimated to be AE0.03 kcal mol 1 based on a conservative estimate of the NMR measured folded/ unfolded ratio of AE5%. 57

chemsum

{"title": "How important are dispersion interactions to the strength of aromatic stacking interactions in solution?", "journal": "Royal Society of Chemistry (RSC)"}

the_construction_of_a_two-dimensional_organic–inorganic_hybrid_double_perovskite_ferroelastic_with_a

## Abstract: Two-dimensional (2D) hybrid double perovskites have attracted extensive research interest for their fascinating physical properties, such as ferroelectricity, X-ray detection, light response and so on. In addition, ferroelastics, as an important branch of ferroic materials, exhibits wide prospects in mechanical switches, shape memory and templating electronic nanostructures. Here, we designed a 2D phasetransition double perovskite ferroelastic through a structurally progressive strategy. This evolution is core to our construction process from 0D to 1D and AgBi-based 2D. In this way, we successfully synthesized 2D lead-free ferroelastic (DPA) 4 AgBiBr 8 (DPA ¼ 2,2-dimethylpropan-1-aminium) with a high Curie temperature (T c ), which shows a narrower band gap than 0D (DPA) 4 Bi 2 Br 10 and 1D (DPA) 5 Pb 2 Br 9 .Moreover, the mechanism of structural phase transition and molecular motion are fully characterized by temperature dependent solid-state NMR and single crystal XRD. (DPA) 4 AgBiBr 8 injects power into the discovery of new ferroelastics or the construction and dimensional adjustment in new hybrid double perovskites. ## Introduction In recent years, lead-based perovskites represented by MAPbI 3 have swept across numerous scientifc research areas including light emitting diodes, 1-5 ferroelectrics, solar cells, gas sensors, 16 lasers, 17 catalyst, photodetectors, 18 and so on. The excellent features stem from their comparative advantages including high absorption coefficient, high charge-carrier mobility, narrow and tunable bandgaps, and other physical and chemical characteristics in perovskites. However, toxicity and long-term instability of lead-based materials restrict their further development. 24,25 Its potential negative impacts on animals, plants and the environment are our concerns and need to be overcome. In this research context, lead-free and lead-replacement ones have naturally become new explorations. This is the most direct method for the research and development of lead-free materials. Therefore, a lot of attention is focused here in order to make new scientifc breakthroughs. As a feasible method, homo-valent replacement using Ge and Sn was proposed to construct lead-free perovskite with superior optical and electronic properties. Nevertheless, Ge/ Sn-based perovskites have been criticized for their instability, motivating us to try hetero-valent replacement. To maintain charge neutrality, hetero-valent replacement can be divided into two subcategories, namely ion-splitting and ordered vacancies. In the ion-splitting subcategory, mixed cation materials at the B site, with a chemical formula of A 2 B I B III X 6 , are featured with appropriate electronic dimensionality and rich chemistry, besides their stability, compared to ordered vacancy (A 3 ,B III X 9 and A 2 ,B IV X 6 , , is vacancy). Here, the B I -site cation mainly includes alkali metal and group IB elements, and the B III -site cation are abundant elements that can locate at group B and group A, and the X at corner can contain halogen, CN and NO 3 . As a member of double perovskites, two-dimensional double perovskites with a formula of A 4 B I B III X 8 show multiple fascinating properties, such as ferroelectricity, piezoelectricity, 39 X-ray detection, light response, 35,36,43 broad photoluminescence, 44 phase transition 38 and so on, 45 which inspire us to grope new 2D double perovskites with desired properties in lead-free exploration. According to the strategy in Scheme 1: First, we synthesized a zero-dimensional organic-inorganic hybrid compound (DPA) 4 Bi 2 Br 10 , which disappointingly does not show the phase transition or other properties we expected. Then a double-row 1D compound (DPA) 5 Pb 2 Br 9 with room-temperature phase transition was synthesized by the introduction of Pb 2+ . Based on this construction strategy, fnally, the lead-free double perovskite ferroelastic (DPA) 4 AgBiBr 8 with a high T c and narrower band gap was successfully constructed. This is exactly what we expected. In this work, we deeply explored the structure-activity relationship between structural phase transition and metal substitution. In addition, the single crystal XRD and solid-state NMR were employed to fully characterize the order-disorder characteristics of structural phase transition. As a ferroelastic phase transition material (DPA) 4 AgBiBr 8 with an Aizu notation of mmmF 1, the transformation of ferroelastic domain structures was clearly observed by using a variable-temperature polarizing microscope. And it is confrmed by UV-vis absorption measurements and density functional theory (DFT) that the band gap of (DPA) 4 AgBiBr 8 (2.44 eV) is lower than that of (DPA) 4 Bi 2 Br 10 (2.80 eV) and (DPA) 5 Pb 2 Br 9 (2.96 eV). In a word, the current report is helpful to the exploration and excavation of more similar lead-free ferroelastics with high temperature phase transformation and hybrid double perovskites. ## Basic crystal structure analysis The crystal structures of (DPA) 4 Bi 2 Br 10 , (DPA) 5 Pb 2 Br 9 and (DPA) 4 AgBiBr 8 were determined by single crystal X-ray diffraction at low temperature. The structure of (DPA) 4 Bi 2 Br 10 is characterized by structural analysis and crystallizes in the P 1 (no. 2) space group of a triclinic system (Table S1, ESI †). It adopts zero-dimensional coordination packing, in which octahedrons are connected by edge sharing, and the N in the cation is oriented towards Br in the adjacent octahedron (Fig. 1a). Subsequently, the lead-replacement (DPA) 5 Pb 2 Br 9 located in P 1 (Table S1, ESI †) was successfully constructed. Unlike the traditional 1D face-sharing PbBr-based perovskites, (DPA) 5 Pb 2 Br 9 adopts a unique double column corner-sharing connection to form a 1D chain. Due to the H-bond interaction, DPA cations in (DPA) 5 Pb 2 Br 9 are interspersed orderly by N atoms facing the adjacent inorganic skeletons (Fig. 1b). Then two-dimensional (DPA) 4 AgBiBr 8 was successfully assembled by a lead-free Ag/ BiBr scheme, which also crystallizes in P 1 (Table S1, ESI †). It is a corner sharing Ruddlesden-Popper perovskite structure. The N atoms in the upper and lower layers face the inorganic octahedron, showing an upward and downward posture respectively (Fig. 1c). In order to better present the structural dimension in the three compounds, an inorganic skeleton stacking as shown in Fig. 1d-f is drawn. The achievement is helpful to design and regulate the structural transformation with expected physicochemical features. All the structural information including hydrogen bonds, bond lengths, bond angles and hydrogen-bond geometry in the three compounds are listed in Fig. S1 and Tables S2-S7. † ## Analysis of phase transition behaviors Due to their potential application in temperature sensors and solid-to-solid phase transition substances, phase-transition features in the hybrids have attracted much attention. Therefore, various design strategies and construction schemes have been implemented in order to make new progress and improvement. The structural order feature provides a fxed stacking arrangement for the inorganic skeleton, so the chargebalance cations are also arranged regularly in the gap of the inorganic skeleton, and sufficient space is provided to realize possible thermal movement. Then the order-disorder structural phase transition is triggered, and a series of physical and chemical characteristics are induced. So, differential scanning calorimetry (DSC) and temperature-dependent dielectric measurements were carried out to prove the occurrence of phase transition. As speculated, (DPA) 5 Pb 2 Br 9 and (DPA) 4 -AgBiBr 8 exhibit phase transition behavior at 300.6 K and 375 K in Fig. 2a, respectively. However, due to the different intermolecular forces, (DPA) 4 Bi 2 Br 10 is not a phase transition one. And the temperature-dependent dielectric constant 3 0 was obtained. The corresponding conductivity was calculated by using the formula 3 00 ¼ 3 0 tan q and sa.c. ¼ u3 00 3 0 , where 3 0 is the permittivity of vacuum. The curves of the dielectric constant of (DPA) 4 Bi 2 Br 10 (Fig. 2b) are nearly linear, and the dielectric constants for (DPA) 5 Pb 2 Br 9 (Fig. 2c) and (DPA) 4 AgBiBr 8 (Fig. 2d) are abnormal with the temperature heating/cooling. This is consistent with DSC analysis. ## Variable-temperature crystal structure analysis For solid-to-solid phase transition compounds (DPA) 5 Pb 2 Br 9 and (DPA) 4 AgBiBr 8 , it is necessary to comprehend the relationship between the structures and physical properties. Therefore, single crystal X-ray diffraction of (DPA) 5 Pb 2 Br 9 and (DPA) 4 AgBiBr 8 in the high temperature phase was performed to determine the structures after phase transition. The space group of (DPA) 5 Pb 2 Br 9 is also P 1, and the bond lengths/angles of the inorganic skeleton change slightly (Tables S3 and S8, ESI †) at 310 K. Besides, as shown in Fig. 3a and S2a of the ESI, † half of the cations in (DPA) 5 Pb 2 Br 9 undergo order-disorder transition in the high temperature phase, which is the main contributor for the phase transition. In contrast, the space group of (DPA) 4 AgBiBr 8 changes from P 1 to Cmmm (no. 65), which can be classifed as a ferroelastic phase transition with an Aizu notation of mmmF 1. At 375 K, the DPA cation in (DPA) 4 AgBiBr 8 is located at a special symmetry site of 2mm and undergoes molecular thermal vibration, which leads to its multi-oriented disordered state similar to the state of rotational motion (Fig. 3b and S2b, ESI †). In addition, the inorganic skeleton also changes signifcantly, and the frontal (156.29 ) and lateral (24.79 , inset) torsion angles of the Bibased octahedron with a Ag-based octahedron as the reference at 150 K (Fig. 3c) change to 179.5 and 0.48 at 375 K (Fig. 3d), indicating that the structure gradually changes from the distorted form to an inorganic perovskite-like structure. The change of the inorganic part including torsion of the octahedron and the shift of metal atoms is also observed from Fig. and f, and the conclusion is consistent with the results discussed above. As another piece of evidence, variabletemperature powder X-ray diffraction of (DPA) 4 AgBiBr 8 changes signifcantly (Fig. S3, ESI †), indicating that the phase transition occurred near 375 K. ## Variable-temperature solid-state NMR analysis In order to prove the rotational motion of cations after phase transition, variable-temperature solid-state NMR was performed (solid-state NMR measurements, ESI †). Fig. 4 shows the experimental and simulated 2 H NMR spectra of DPA cations at different temperatures. It can be observed that both the experimental spectra acquired at 273 K and 380 K exhibit axially symmetric powder Pake patterns, indicating that the DPA cations undergo some restricted reorientation processes. By simulating the spectra, we have obtained detailed information on the reorientation processes. When the temperature is just above the phase transition temperature (e.g., 380 K), the Pake pattern exhibits a sudden narrowing and the n QS is motionally averaged to about 6 kHz, indicating that DPA cations undergo an overall cation reorientation (Fig. 4b). We introduce a motion model in which all the DPA cations perform a 4-site jump pattern along the C n axis (close to the C-C axis, see rotation mode). The rotation angle, namely the included angle between the C-N bond (R C axis) and C n axis, is represented as s. In addition, the internal rotational motion about the C-N bond (the C 2 axis) also remains activated. Deuterium spectrum simulation is implemented based on this motion model and the result found that the spectrum simulated by using s ¼ 61.5 agrees very well with the experimental spectrum acquired at 380 K. It is worth noting that the bond angle of C-C-N is about 67 , very close to the rotation angle s, indicating that the C n axis is almost parallel to the C-C bond adjacent to the C-N bond in the DPA cation. In general, combined with variable-temperature single crystal X-ray diffraction and variable-temperature solid-state NMR, we conclude that the phase transition of (DPA) 4 AgBiBr 8 is due to the rotation of organic cations and the torsion of an octahedron in an inorganic skeleton. ## Ferroelasticity Based on the Aizu notation of mmmF 1, (DPA) 4 AgBiBr 8 can be classifed as a typical ferroelastic. And the ferroelastic domains can be easily observed by using a variable-temperature polarizing microscope. Compared with inorganic materials, organicinorganic hybrid ones can be characterized by both the bulk and thin flms. Of course, the observation of thin flms is the best choice to show clearer domain characteristics. Here, the thin flm was prepared by dropping 20 mL of solution containing HBr acid (AR $ 40%, 500 mL) and (DPA) 4 AgBiBr 8 (25 mg) on the ozone treated (20 minutes) ITO glass and then heating at 323 K for 45 minutes, as shown in Fig. 5a. By using a polarizing microscope at 330 K, the ferroelastic domain structures can be clearly observed (Fig. 5b). With the continuous increase of temperature from 330 K to 380 K, the domain structures gradually disappear near 375 K, suggesting the presence of the paraferroelastic phase in higher temperature, and then they gradually recover by cooling to the ferroelastic phase at 350 K. The disappearance and recovery of domain structures in the heating/cooling circles are completely consistent with the DSC, dielectric and temperature-dependence single structures. The spontaneous strain tensor can be calculated by using the following matrix (1) according to its Aizu notation of mmmF 1 from the high-symmetry phase (orthorhombic) to the lowsymmetry phase (triclinic): The formula corresponding to the elements in this matrix can be obtained in the ESI, † by substituting unit cell parameters at 150 K and 375 K in formula (1) and (2). In this way, a reasonable calculation is successfully completed, in which the total spontaneous strain 3 ss is 0.15593. In addition, the domain transformation Video S1 † was recorded, showing the readers a beautiful reversible ferroelastic transition process. This is very helpful to analyze and explore the domain structure and the micro mechanism. ## Semiconducting behavior The UV-vis absorption and density functional theory (DFT) calculations of (DPA) 4 Bi 2 Br 10 , (DPA) 5 Pb 2 Br 9 and (DPA) 4 AgBiBr 8 were implemented to consider and discuss the change of their band gaps in Fig. 6. As shown in the Tauc plot, (DPA) 4 AgBiBr 8 exhibits an indirect band gap of 2.44 eV, which is less than that of (DPA) 4 32 which is lower than that of BA-based (BA ¼ butylammonium) double perovskite 54,55 and higher than that of Cs + double perovskite. 32 By comparing with AgBiI-based ones, the band gap of AgBiBrbased double perovskite is signifcantly higher (Table S9, ESI †). 30,56 Therefore, it can be found that the influence of the inorganic part on the band gap is greater than that of the organic part. With spin-orbital coupling (SOC), the band gap, the valence band maximum (VBM) and the conduction band minimum (CBM) in the three compounds are predicted in Fig. 6b, d and f. In addition, the energy band structure of (DPA) 4 AgBiBr 8 without SOC was obtained (Fig. S4, ESI †).The differences between SOC and non-SOC are mainly derived from a split-off conduction band within Bi-6p states according to the partial density of states, and a precise prediction implies the necessity of SOC in Bi-based and Pb-based organic-inorganic perovskite. 45, For these three compounds, the valence band maximum (VBM) and conduction band minimum (CBM) are mainly contributed by the inorganic part. Besides, in (DPA) 4 Bi 2 Br 10 (Fig. 6b), (DPA) 5 -Pb 2 Br 9 (Fig. 6d) or (DPA) 4 AgBiBr 8 (Fig. 6f), the organic part, and the H-s, N-p and C-p states overlap widely, indicating a strong interaction. ## Conclusions In this work, three target compounds were synthesized through the metal substitution strategy as shown in Scheme 1. In particular, an AgBi-based hybrid double perovskite was successfully constructed. The (DPA) 5 Pb 2 Br 9 and (DPA) 4 AgBiBr 8 show solid-to-solid phase transition. It is worth mentioning that (DPA) 4 AgBiBr 8 is a high temperature ferroelastic material clas-sifed as mmmF 1, and its phase transition comes from the rotational cation motion and the torsion of the octahedron in the inorganic skeleton. Theconstruction of (DPA) 4 AgBiBr 8 broadens the potential applications of two-dimensional double perovskite. At the same time, (DPA) 4 AgBiBr 8 also provides power for the explosion of new ferroelastics or hybrid double perovskites. ## Synthesis All the reagents and solvents mentioned in this work were purchased from commercial suppliers and were not further purifed. (2,2-Dimethylpropan-1-aminium) 4 Bi 2 Br 10 (DPA) 4 Bi 2 Br 10 Stoichiometric amounts of 2,2-dimethylpropan-1-amine (0.2 mmol) and bismuth bromide (0.2 mmol) were added into a beaker. After this, hydrobromic acid (AR $ 40%, 20 mL) was added into the beaker. A colorless prism crystal of (DPA) 4 Bi 2 Br 10 was obtained for single-crystal X-ray diffraction study through slow evaporation of the mixed solution at room temperature after several days.

chemsum

{"title": "The construction of a two-dimensional organic\u2013inorganic hybrid double perovskite ferroelastic with a high <i>T</i>_c and narrow band gap", "journal": "Royal Society of Chemistry (RSC)"}

direct_evidence_for_distinct_colour_origins_in_roy_polymorphs

## Abstract: ROY is one of the most well-studied families of crystal structures owing to it being the most polymorphic organic material on record. The various red, orange, and yellow colours of its crystal structures are widelybelieved to originate from molecular conformation, though the orange needle (ON) polymorph is thought to be an exception. We report high-pressure, single-crystal X-ray measurements which provide direct experimental evidence that the colour origin in ON is intermolecular, revealing that the molecule undergoes minimal deformation but still exhibits a pronounced, reversible, pale orange / dark red colour change between ambient pressure and 4.18 GPa. Our experimental data are rationalised with band structures, calculated using an accurate hybrid DFT approach, where we are able to account for the variation in colour for five polymorphs of ROY. We highlight the outlier behaviour of ON which shows marked p/p stacking interactions that are directly modified through application of pressure.Band structure calculations confirm these intermolecular interactions as the origin of the colour change. ## Introduction The 'ROY' family of crystal structures, crystallising from the 5methyl-2-[(2-nitrophenylamino)]-3-thiophenecarbonitrile compound, is numerous, with the most recent literature citing the discovery of the thirteenth form, 4 making it the most polymorphic material in the Cambridge Structural Database (CSD) at the time of writing, ahead of other rivals including galunisertib (eight forms), the structurally-similar flufenamic acid (ten), and aripiprazole (twelve). The ROY molecule exhibits signifcant flexibility about the s SCNC dihedral angle between the crystal structures and, to a lesser extent, the s CNCC angle (shown in Fig. 1); the molecules show markedly different conformations and packing across all the known crystal forms. 5,6 The propensity for ROY to form so many crystal structures spontaneously, from the melt or solution, 7 has drawn signifcant attention from the crystal growth community, in attempts to control the polymorphic outcome. More recent work on crystal growth has exploited the cross-nucleating ability of ROY where synthetic analogues can be used to seed supercooled melts of structurally 'normal' material, leading to yet more polymorphs. 14 The newest forms have been obtained via small-scale crystallisations from droplets of liquid ROY/oil-encapsulated ROY solutions. 4 While ROY has been a fruitful system for experimentalists, with new polymorphs being found relatively frequently, calculating its solid state landscape and polymorph properties has proved more challenging. This is exemplifed by the difficulties in ranking the internal energies of its many crystal structuresthese often differ both from other theoretical studies and also from the experimentally-observed hierarchy. In some studies, some of the known forms of ROY have not ranked within the top one hundred most stable predicted structures at all. 16 A complicating factor is the apparent difficulty in producing accurate potential energy surfaces as a function of Fig. 1 (a) Molecular structure of ROY. Carbon atoms are shown in black, nitrogen in blue, oxygen in red, sulfur in yellow, and hydrogen in white. The two rotatable bonds s SCNC and s CNCC that can influence intramolecular conjugation are indicated. (b) An alternate measure of the extent of ring coplanarity; the angle q between mean planes calculated through the phenyl and thiophene groups. molecular conformation; Thomas and Spackman have explored this in more detail, showing that DFT energies are unreliable for this purpose and are outperformed by MP2 calculations. 17 Subsequently, Nyman et al. found that DFT+D approaches overestimate the stability of coplanar conformations, due to excessive predicted p-electron delocalisation. 18 Though the ROY system is interesting in its own right, it also represents a more general challenge facing the materials chemistry community: how to compute accurate material physical properties when conventional DFT approaches-the 'go-to' method for solidstate calculations-prove too limited, 19 in this case for a relatively simple geometry optimisation procedure. The conformational flexibility of ROY in turn suggests a malleable electronic structure, and it is widely-accepted that its crystal colours arise from the degree of conjugation between the nitrophenyl and thiophene moieties. 6 Each crystal form sits somewhere on the red-orange-yellow region of the visible spectrum (hence the 'ROY' moniker) where, broadly, coplanar and perpendicular molecular arrangements lead to red and yellow colours, respectively, and orange colours are represented by intermediate angles. Previous work by one of us demonstrated that the yellow (Y) polymorph exhibits piezochromic properties, 20 where the crystal became progressively red in colour as the molecule was driven towards planarity, and crystal density increased, on applying pressure. Similarly, the molecular geometry of the orange plate (OP) form also showed susceptibility to pressure, 21 though we were unable to comment on the crystal colour due to the lack of optical access to the pressure device. Given the strong colouration of the ROY family, and its prominence in organic solid-state literature, there is a notable absence of calculated electronic band structures. A recent computational study by Feng et al. 22 advanced this area, providing a set of band gap values for many of the ROY forms. However, the authors highlighted the aforementioned difficulties posed by DFT, which they note led to additional flattening (ca. 10 ) of the s SCNC angle in the red and orange polymorphs, and systematic underestimation of the band gaps. A particularly intriguing observation, arising from their calculated molecular singlet excitation energies, was that the orange needle (ON) form appears to be an outlier; the authors postulated that its colour may actually originate from intermolecular interactions, in contradiction to the decades-held view that molecular conformation is predominantly responsible. We have sought to explore the ROY colour origins in detail, reporting the frst electronic band structures of ROY made possible by making use of a hybrid DFT approach with crystalline orbitals-its proven ability to calculate band gaps with greater accuracy than conventional DFT will likely see it gain further traction in the materials chemistry and physics communities, particularly in assisting band-structure engineering. 26,27 In combination with direct evidence provided by high-pressure, single-crystal, X-ray diffraction measurements we show conclusively that the ON form is indeed an anomaly, where its colour is intermolecular in origin. ## High-pressure crystallography-cell compressibility High-pressure diffraction provides the ideal experiment with which to test whether any observable change in colour occurs in conjunction with either intramolecular or intermolecular modifcations (or both) to the structure. Earlier work on the Y form frst revealed the piezochromic nature of ROY, showing that a reversible yellow / red colour progression was observed with pressure. 20 This was accompanied by both a large deformation in the lattice-expected behaviour for a compressed molecular crystal structure-but also conformational change in the molecule. A similar structural response to pressure was also observed in the OP polymorph. 21 Fig. 2 summarises the behaviour of ON ROY under pressure-there is no transformation in the crystal structure (see panel Fig. 2a); its P2 1 /c symmetry is retained over the pressure range investigated here. The unit cell simply becomes a more compressed version of its ambient-pressure form. Details of the cell compression characteristics are given more comprehensively in the ESI, † but we note here that the most compressible principal axis is approximately aligned with the a-direction, having a compressibility K of 21.6(7) TPa 1 , corresponding to a decrease in length of 13.4%. The other two principal, orthogonal, directions are less compressible (11.5(5) and 3.7(10) TPa 1 ), ultimately leading to a bulk modulus B 0 of 5.9(14) GPa. Table 1 provides crystal structure refnements statistics for select pressure points. ## High-pressure crystallography-colour origin Similar to the Y form, there is a clear pale orange / red colour progression as pressure is increased (Fig. 2b). This is completely reversible, with the crystal returning to a pale orange colour on complete decompression. The deep red colour of the crystal is obscured at 4.18 GPa, due to the apparent crystallisation of the pentane mixture at a lower-than-expected pressure (ordinarily 5.4 GPa); premature crystallisation of the n-pentane component is now known to occur on occasion. 28 Fig. 2c depicts the most conclusive experimental observation that the crystal colour arises from intermolecular interactions; the conformational geometry of the molecule is insensitive to the effects of pressure, showing only negligible change. Though the molecular conformation is usually discussed in terms of the s SCNC angle, which measures 52.9(11) at 4.18 GPa (cf. 52.6(3) at ambient pressure, 29 and 47.7(13) at 0.02 GPa), a more accurate measure of the conformation is the angle made between mean planes q drawn through the phenyl and thiophene groups, shown schematically in Fig. 1b as this does not neglect the (often small) effect of the s CCNC dihedral angle. Between 0.02 GPa and 4.18 GPa, q is effectively unchanged, decreasing from 52.79 to 52.41 , and measures 53.68 for the literature 0 GPa structure (CSD ref. code: QAXMEH). As an aside, we note that the difference between s SCNC and q can actually be quite pronounced in some other ROY forms, e.g. for the R form (QAXMEH02), they measure 21.74 and 45.56 , respectively. Lastly, the relatively compressible a-axis indicates that the most deformable intermolecular interactions are aligned with this direction. These are comprised of two sets of p/p stacking interactions, one between nitrophenyl rings and the other between thiophene groups. The crystal symmetry constrains both of these interactions to be the same length, which decrease from 3.92 at 0.02 GPa to 3.50 at 4.18 GPa. The authors of ref. 22 postulated that intermolecular interactions are most likely to be responsible for the colour origin in ON are the p/p stacksour experimental results show that this is almost certainly the case. ## Electronic band structures In order to more fully understand the colour origin in ROY polymorphs-specifcally, the nature of the orbitals involvedwe performed electronic structure calculations using a hybrid DFT approach that has a proven track record in determining band gaps. Five polymorphs (red R, orange plate OP, orange needle ON, yellow Y, and yellow needle YN) that encompass the full colour spectrum exhibited by ROY were selected for computation-starting coordinates for each were obtained from literature structures in the CSD. First, geometries were optimised, while holding unit cell parameters fxed at experimental values. A strong level of agreement between our calculated interplanar angles q and the experimentally observed angles, shown in Table 2 confrms that the simulations are providing accurate models of the original crystal structures, avoiding the over-stabilisation of the more planar forms seen with conventional DFT. 18 Following optimisation, electronic band structure diagrams and projected density of states (PDOS) were computed, with the outputs shown in Fig. 3 alongside their respective Brillouin zone paths. The computed band gap values are given in Table 2 and appear to correlate well with the known colours of the polymorphs, providing some confrmation that the hybrid DFT approach does not suffer from the same difficulties as conventional DFT, in agreement with observations made in ref. 22. For R, the computed band gap (1.95 eV) is similar to the experimental value reported for a-HgS (2.0 eV), from which the red pigment vermillion is derived. 30 The orange polymorphs ON and OP present near-identical band gaps of 2.32 and 2.37 eV, respectively, which closely matches that of the orange pigment lead(II) chromate (2.3 eV), 31 while the larger band gaps for YN and Y (2.54 and 2.80 eV) are similar to CdS (2.5 eV), which is used to obtain the pigment cadmium yellow. 32 The strong link between observed colour and electronic band gap confrms that the electronic transitions responsible for the colouration of ROY are confned to the frontier orbitals, located either side of the Fermi energy level. From the band structure plots-shown in Fig. 3-we observe two valence and two conduction bands for R and YN, whereas ON, OP, and Y have double this number; this simply reflects the number of molecules in each respective unit cell. The frontier bands are largely k-invariant, indicating that the corresponding crystalline orbitals are localised at the molecular level and are not strongly influenced by the lattice environment. However, as identifed in ref. 22, ON is an exception (Fig. 3) as the valence bands show a degree of energy dispersion with respect to the k-point paths G / D/B and Y / A/E. Superposition of the Brillouin zone path with the real-space crystal lattice (Fig. 3f) shows that this corresponds to the a-direction, suggesting that the pertinent intermolecular interactions must be aligned with this unit cell vector; notably the deformable p/p nitrophenyl and thiophene stacking interactions are coincident with this direction. However, the ring separation distance at ambient pressure is long (3.95 ), so this interaction is likely to be weak. ## Pressure-dependent band dispersion Having demonstrated the reliability of our calculations in reproducing ambient-pressure molecular geometry and band gaps, we applied the same computational strategy to select high-pressure ON structures (1.37, 3.00, and 4.18 GPa), effectively accounting for the change in cell volume, which is known to have computational implications. 22 These geometry optimisations returned q values of 52.0, 51.6, and 51.1 , con-frming that the molecular geometry is largely unaltered with pressure. The resulting electronic band structure diagrams are shown in Fig. 4, along with the ambient pressure band structure for direct comparison. It is immediately apparent the degree of dispersion in the frontier bands increases with applied pressure. The variance in energy is exclusively confned to the valence bands along the k-point paths G / D/B and Y / A/E. This implies that the strength of the intermolecular interactions along the a-direction increase with applied pressure; an observation substantiated by a contraction in the aforementioned p/p stacking of 0.45 , at the highest pressure measured here. Non-covalent interaction (NCI) plots for the optimised structures at ambient pressure and 4.18 GPa, shown in Fig. 5 show the reduced electron density gradient isosurface (s), coloured according to the strength (r) of the interactions. The enhanced blue colouration in the high-pressure structure indicates stronger p/p stacking interactions-approximately double that of the ambient pressure structure. Crucially, the simulations support the presence of piezochromic behaviour as the size of the band gap reduces from 2.32 to 1.89 eV as pressure is applied, corroborating the pale orange / red colour change that is visually evident in Fig. 2. ## Frontier crystalline orbitals Plotting the values of the valence and conduction band energies raises a further interesting point, namely that the pressureinduced decrease in band gap can be attributed, almost entirely, to an increase in energy of the valence band-see Fig. 4. This in turn can be accounted for by the dispersion in these frontier bands, which was explained on the basis of the p/p stacking interactions. Fig. 4 also shows the frontier orbital energy separation for an isolated (gas phase) molecule of ROY with atomic coordinates frozen to those obtained from geometry optimisation of the ambient pressure structure. The influence of the crystalline environment clearly impacts on both valence and conduction band energies, serving to raise the former and reduce the latter. More specifc information on the nature of these frontier bands can be provided by the PDOS, shown in Fig. 3 alongside the ambient-pressure band structure diagrams. From these, we can deduce that the top of the valence band is derived mostly from carbon and nitrogen states, with a small contribution from oxygen, whereas the bottom of the conduction band has near equal weighting across all three atomic states. Sulfur makes a small contribution to both orbital states. This behaviour is mirrored across all fve polymorphs investigated here, suggesting that the frontier orbitals are invariant to crystal packing. Fig. 4 also presents a visualisation of the frontier orbitals at the G-point for the ON form. Both the highest occupied crystalline orbital (HOCO) and lowest unoccupied crystalline orbital (LUCO) are p-type orbitals, with the HOCO delocalised across the whole molecule, while the LUCO is more localised on the nitrophenyl ring. Absorption of light will therefore likely accumulate electron density in the p-orbitals of the nitrophenyl region of the molecule, further enhancing the p/p interactions. ## Conclusions We have provided conclusive experimental and computational evidence that the colour origin in the ON ROY polymorph arises from intermolecular p/p interactions, making it something of an anomaly in the ROY family. These weak, deformable, interactions strengthen on application of pressure and, in doing so, introduce energy dispersion in the valence bands which progressively narrow the band gap to values commensurate with the colours seen experimentally. Though the colours in the other ROY polymorphs investigated here are clearly intramolecular in nature, the extended crystal lattice is still important as it plays a role in stabilising the respective molecular conformations. That ROY can adopt distinct mechanisms in different polymorphs to produce its strong colouration, is owed to its flexibility, and this has some precedent in other flexible polymorphic compounds. 33 This only becomes evident by directly comparing the band structures between polymorphswere the ON band structure calculated in isolation, the extent of inter/intramolecular influence would be less clear. The accuracy in geometry optimisation and band gap calculation of the hybrid DFT approach we have used is highly encouraging, and can be straightforwardly transferred to other solid-state materials. In particular, it has allowed us to ascertain the level of influence the intermolecular interactions have on a material property (in this case colour), through the extent of band dispersion. Our study concerns just fve of the thirteen known ROY polymorphs (and an additional three were also considered by Feng et al.), 22 which leaves open the possibility that the colours of some of the other forms could also be a result of intermolecular excitations. The ROY polymorphs have been loosely categorised based on distinct regions they occupy on their conformational potential energy surface, 5,6 however if additional polymorphs were revealed to show similar electronic behaviour to the ON form, then perhaps grouping the forms by intra/intermolecular colour origins might be an appropriate, alternative, classifcation system. ## High-pressure X-ray diffraction ROY was obtained in powdered form from TCI Chemicals as the OP polymorph. Small crystals of the ON form were visually identifed, and isolated, from other concomitantly-occurring forms following recrystallisation from acetone; these were then used to seed saturated ROY:acetone solutions. A suitable crystal was identifed and loaded in a Merrill-Bassett diamond anvil cell (DAC), 34 equipped with Boehler-Almax anvils with an 85 opening angle and WC backing seats. 35 A 1 : 1 volume mixture of pentane : isopentane was included as a pressuretransmitting medium, 36 and a ruby chip as a pressure calibrant; pressure was determined using the ruby fluorescence method. 37 X-ray diffraction data were collected on a Rigaku Synergy diffractometer, using Mo K a radiation, at pressures of 0.02, 0.62, 1.37, 1.91, 2.41, 3.00, 3.53, and 4.18 GPa as well as a further measurement on complete decompression. The raw diffraction data were integrated and corrected for absorption using CrysAlisPro. Structure refnements were carried out using Crystals. 38 A starting model for the lowest-pressure refnement was obtained from an earlier ambient-pressure dataset (unreported). To avoid any potential bias of the dihedral angles, after atomic coordinates were imported, the entire thiophene moiety was deleted and relocated in a Fourier difference map. Owing to the low completeness of the data, only the sulfur atom was refned anisotropically, and hydrogen atoms were constrained to ride on their host atoms. All covalent bond distances were restrained to values informed by the ambient pressure structure, and 1-2, 1-3 vibration and thermal similarity restraints were also applied. Refned models at each pressure were then used as a starting set of coordinates for the following pressure point. ## Hybrid DFT calculations Solid-state calculations were performed using CRYSTAL17, 39,40 with triple zeta quality all-electron basis sets with valence polarisation used for all atoms, combined with the HSE06 hybrid functional, along with a Grimme D3 dispersion correction. 41,42 This choice of functional can be justifed from its proven track record in accurately calculating electronic band gaps. Ambient-pressure structures deposited in the CSD, and our high-pressure X-ray structure determinations of the ON form were used as input geometries for atom-only optimisation. K-Space was sampled using a Monkhorst-Pack net of 8 8 8 for all structures. 43 Increasing the k-point sampling to a larger Monkhorst-Pack net of 16 16 16 proved convergence with respect to k-points had been achieved to within 1 10 7 a.u. Tolerances for the bielectronic Coulomb and exchange contributions to the Fock matrix are controlled by fve parameters, the frst four of which were set to 1 10 7 , and the ffth to 1 10 14 . 39,40 Convergence criteria were set on the root-meansquare (RMS) and absolute values for both the gradient (i.e. atomic forces) and estimated atomic displacements at 3 10 4 a.u. and 1.2 10 3 a.u., respectively. In addition, the energy convergence threshold between successive cycles was required to be below 10 7 a.u. 39 Following optimisation, the electronic band structures, PDOS and localised crystalline orbitals were obtained, with the latter computed at the Brillouin zone G-point. The non-covalent interaction plots were obtained using the CRITIC 2 code. For the isolated molecule optimisations, the same procedure as documented above was employed, with the exception that the contents of the unit cell were deleted to leave just one ROY molecule inside a non-periodic system, and the dihedral angle was constrained to the same value as observed in the ON crystal structure (53.4 ).

chemsum

{"title": "Direct evidence for distinct colour origins in ROY polymorphs", "journal": "Royal Society of Chemistry (RSC)"}

persistent_nucleation_and_size_dependent_attachment_kinetics_produce_monodisperse_pbs_nanocrystals

## Abstract: Modern syntheses of colloidal nanocrystals yield extraordinarily narrow size distributions that are believed to result from a rapid "burst of nucleation" (La Mer, JACS, 1950, 72(11), 4847-4854) followed by diffusion limited growth and size distribution focusing (Reiss, J. Chem. Phys., 1951, 19, 482). Using a combination of in situ X-ray scattering, optical absorption, and 13 C nuclear magnetic resonance (NMR) spectroscopy, we monitor the kinetics of PbS solute generation, nucleation, and crystal growth from three thiourea precursors whose conversion reactivity spans a 2-fold range. In all three cases, nucleation is found to be slow and continues during >50% of the precipitation. A population balance model based on a size dependent growth law (1/r) fits the data with a single growth rate constant (k G ) across all three precursors. However, the magnitude of the k G and the lack of solvent viscosity dependence indicates that the rate limiting step is not diffusion from solution to the nanoparticle surface. Several surface reaction limited mechanisms and a ligand penetration model that fits data our experiments using a single fit parameter are proposed to explain the results. ## Introduction Nanometer scale colloidal crystals of metal, metal oxide, metal chalcogenide, and metal pnictide materials can be synthesized with extraordinary size and shape uniformity. It is widely assumed that monodisperse particle size distributions result from a short burst of nucleation followed by diffusion controlled growth. However, even in the case of colloidal quantum dots (QDs) where the nanocrystal size, size distribution, and nanocrystal concentration are readily monitored with optical spectroscopy, direct measurements of the nucleation kinetics are challenging. Few experimental studies simultaneously achieve the necessary time resolution and precision to monitor the nanocrystal concentration and size during the nucleation phase. It is therefore uncertain whether the duration of the nucleation period explains the monodispersity. In typical QD syntheses molecular precursors are injected into hot surfactant solution at a temperature that converts them to solutes. 10 These solutes supersaturate leading to crystal nucleation and growth. When the conversion reaction is very rapid, nucleation can occur in a transient region of solution with non-uniform temperature and concentration that is impractical for mechanistic study. Slow and controlled precursor conversion reactivity, on the other hand, allows nucleation and growth to occur after mixing has eliminated concentration or temperature gradients. In these syntheses, the precursor conversion reaction is rate limiting and governs the kinetics of solute supersaturation and nucleation. 10,11 Among QD materials, lead sulfde produced from chalcogenourea precursors and lead oleate form with extraordinarily narrow size distributions. 12 By modifying the chalcogenourea substitution pattern, the kinetics of solute generation can be precisely controlled over a wide range of temperatures. These reagents have recently been used to study the temperature dependence of the crystal nucleation and growth steps, where in situ X-ray pair distribution function analysis demonstrated the formation of molecular solute intermediates. 14 Moreover, the precisely controlled solute supply is orthogonal to the crystal growth reaction and therefore ideally suited to explore the connection between the solute concentration and the kinetics of lead sulfde nucleation and growth. High supersaturation and diffusion limited growth kinetics are widely believed to cause size distribution focusing. 15,16 While numerous studies argue for focusing on the basis of diffusion limited growth kinetics, we are unaware of any direct evidence for diffusion limited growth in colloidal QD formation, nor any measurements of the solute concentration during growth. Moreover, claims that size distribution focusing occurs during the formation of colloidal QDs typically rely on the temporal evolution of the absorption spectral linewidth. 2 That analysis, however, does not properly account for the intrinsic width of a single QD absorber nor its size dependence. 13,17,18 In addition, arguments based on changes to the percent standard deviation rather than changes to the absolute polydispersity are obscured by the evolving nanocrystal size. Thus, the mechanistic origins of size distribution focusing are unclear. To probe the origins of monodispersity, we performed a direct measurement of QD nucleation and growth kinetics using in situ X-ray scattering. Coupled with measurements of the solute supply kinetics using NMR spectroscopy and 13 Clabeled thiourea precursors, we deduce the solute concentration throughout the reaction. These experiments reveal a slow steady increase in the number of nanocrystals that continues throughout a large fraction of the synthesis. Population balance modeling and a weak dependence on the solvent viscosity demonstrate that the canonical diffusion limited growth mechanism cannot explain the size focusing. ## Results and discussion The kinetics of PbS nanocrystal formation were monitored in situ using optical absorption and time resolved small angle and wide angle X-ray scattering (SAXS and WAXS) on the ID02 beamline of the European Synchrotron Radiation Facility (ESRF). 19 Several disubstituted thiourea (Fig. 1) derivatives were selected whose conversion to PbS QDs (d ¼ 3.1-7.5 nm) reaches completion in 1-30 minutes at 110 C. Following injection of the thiourea (t ¼ 0 s), the reaction solution is pumped through a thin X-ray capillary and SAXS and WAXS patterns are recorded every second (Fig. S1 †). In parallel, an optical probe records the absorbance of the solution at l ¼ 400 nm. Fig. 1B shows the evolution of the absorbance at 400 nm, where an induction delay is observed (21-56 seconds), following which the absorbance increases with kinetics that are well described by a single exponential (k UV-Vis obs ¼ 2.1-7.2 10 3 s 1 ). Fig. 2 shows the evolution of the SAXS and WAXS signal during a synthesis. The increasing SAXS intensity at small wave-vectors (q) and the appearance of oscillations at high q are characteristic of an ensemble of monodisperse nanoparticles that grow in size. The SAXS patterns were ftted using a quantitative model consisting of the sum of two terms, each one having the form of a polydisperse distribution of spheres (see ESI † for details). The frst term captures the concentration, mean radius and polydispersity of PbS nanoparticles. The second term captures the concentration of lead oleate and is based on the signal of a lead oleate solution measured separately that is well ft to a spherical scatterer with radius of 0.13 nm. The lead oleate precursor was further probed using small angle neutron scattering (SANS) in tetradecane-d 30 solution. The strong scattering contrast between the hydrogenated surfactant alkyl chains and the solvent produces a SANS signal that can be ft to a spherical scatterer with radius of 1.74 nm and polydispersity of 20%. Both measurements are consistent with a small molecular complex composed of one or two lead oleate units (Fig. S7 †). As the reaction proceeds the scattering from lead oleate decreases while the scattering from the nanocrystals increases (Fig. S20 †). Fits of the nanocrystal component provide the mean radius of the size distribution, its polydispersity and the nanocrystal concentration at each time point. The signal to noise ratio of the data at very early times prevents precise measurement of the polydispersity and concentration (see ESI † discussion on this point) but following this period, oscillations in the SAXS pattern appear at high q and allow a reliable ft. The absolute value of PbS units within nanocrystals can be retrieved from the size distribution and the overall concentration in nanocrystals. A single exponential ft of PbS concentration provides rate constants for PbS precipitation of k SAXS obs ¼ 2.4-4.3 10 3 s 1 depending on the precursor used (Fig. S21 and S22 †). WAXS patterns display fve peaks characteristic of the PbS rock salt crystal structure. By ftting the peak to a Gaussian model with a linear background we independently determine the concentration of scattering material and the size of crystallites. The decreasing full width at half maximum (FWHM) is indicative of an increasing mean nanoparticle size that can be extracted using the Debye-Scherrer relation. This estimate of the crystallite size is systematically slightly smaller than the one estimated from SAXS measurements. A single exponential ft of the WAXS peak area versus time provides a rate constant for the precipitation of PbS which is equal to the one found by SAXS (Fig. S22 †). After 50 s, good agreement between SAXS and WAXS signals indicate that the PbS is crystalline rather than amorphous throughout the reaction. The kinetics of nanocrystal formation were compared to the kinetics of the precursor conversion reaction using 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopy. The disappearance of N-(p-X-phenyl)-N 0 -dodecylthiourea-13 C (X ¼ Cl, H, OMe) and the formation of the N-oleoyl-N(p-X-phenyl)-N 0dodecylurea- 13 C coproduct could be monitored by tracking the integral of the thiocarbonyl and urea carbonyl carbons as shown in Fig. 3. Both measures provide similar rate constants for the conversion of each thiourea respectively (k NMR obs ¼ 8.8-13.3 10 3 s 1 ) (see ESI † and Fig. 3). The kinetics measured by NMR are faster than those measured using UV-Vis absorption spectroscopy, which are both faster than those measured using SAXS and WAXS. The differences between each measurement technique can be explained by the formation of PbS solutes that do not contribute signifcant intensity to the SAXS or WAXS signal, but do influence the absorbance at l ¼ 400 nm. Spectroscopic and X-ray scattering characterization of these solutes suggests a molecular complex with formula [(PbS)(Pb(oleate) 2 )] 2 . 14 Collectively, the NMR, UV-vis, and X-ray scattering kinetics data support a stepwise process where the lead oleate and thiourea are converted to solutes that accumulate and are then consumed by nanocrystal growth. The concentration of solute (C solute ) can be determined from the amount of precursor converted and the amount of PbS in nanocrystals ( As can be seen in Fig. 3c, the C solute for each thiourea peaks following $50% conversion of the precursors with a maximum solute concentration of $4 mM or $45% of the total sulfur. Both nucleation and growth to a larger average size occurs in parallel with the rise in C solute . The concentration of nanocrystals measured with SAXS steadily increases to 0.9-4.4 mM over $200 seconds and then plateaus as the nanocrystals grow to their fnal size (Fig. 4 and S23 †). The production of such a large concentration of crystallites is in itself evidence of a homogeneous nucleation process. No further evolution of the nanocrystal concentration or size is observed after the plateau. We conclude that the nanocrystals do not agglomerate or ripen following precursor conversion as was previously reported. 12,13 The increasing number of nanocrystals can be used as a measure of the nucleation kinetics without considering the nucleus size or structure. The instantaneous nucleation rate is extracted from a polynomial ft to the nanocrystal concentration as described in the ESI † and shown in Fig. 4b. These curves illustrate how the nucleation phase continues across $2 minutes, long after the induction delay observed using optical spectroscopy and spanning more than $10% of total reaction time (>50% of the precipitation). Reactions performed at lower temperatures display even longer nucleation times. 14 Interestingly, the nucleation rate decays as the solute concentration climbs to its maximum value. The inverse correlation between nucleation rate and solute concentration is counterintuitive, however, a variety of factors may directly or indirectly influence the nucleation rate. For example, conversion of the thiourea precursor consumes lead oleate, which may impact the nucleation or growth rates. High lead oleate and lead chloride precursor concentrations are known to increase the extent of PbS nucleation and narrow the size distribution. 23,24 Nanocrystal coalescence could also cause an apparent reduction in the nucleation rate, but we do not observe ripening or reduction of the nanoparticle concentration over long periods of time at the reaction temperature. While it is not yet clear why a decrease in the nucleation rate accompanies the increase in solute concentration, this behavior may be related to the prolonged nucleation period observed. On the other hand, the burst of nucleation postulated by LaMer occurs when the nucleation rate is a stronger function of solute concentration than the growth rate, 25 a condition that is not met in our experiments. Thus, the prolonged nucleation period and the inverse relation between nucleation and solute concentration are perplexing but possibly related observations that warrant further study. The size distributions obtained here cannot be explained by a short nucleation phase that is separate from growth. Instead, a size focusing mechanism is needed to achieve the characteristically narrow size distribution. We use population balance modeling to probe the size dependence of the crystal growth rate and determine the magnitude of the size focusing effect. The particle size distribution n(r, t) during nucleation and growth evolves according to the population balance equation (PBE, eqn ( 1 here, n is the number density of particles, r is the particle radius, t is the time and G h r 0 (t) is the growth rate of particles. B is the net generation of particles through events like aggregation, break-up, and nucleation. The PBE framework can describe different nucleation and growth processes by incorporating different B(n, r, t) and G(r, t) models, respectively. In the absence of agglomeration and secondary nucleation we can omit the B(n, r, t) term and account for nucleation with a boundary condition: n(0, t)G(0, t) ¼ J(t). The nucleation rate expression (J(t)) comes directly from the time derivative of the nanoparticle concentration (Fig. 4). Using J(t) and C solute (t) as experimental inputs, we used the PBE to test different growth rate models of the form G(r, t) where k G is a kinetic prefactor, v m the molar volume (v m ¼ 31.5 10 6 m 3 mol 1 ), and a is a parameter corresponding to different limiting resistances. Numerical methods for solving the PBE are described in the ESI. † PBEs with growth models where a ¼ 0-4 were ft to the radius evolution with time by optimizing the corresponding k G , the only adjustable parameter for each growth model (Fig. S15 †). An example particle size and size distribution predicted by the PBE is shown in Fig. 5. Specifc values of correspond to different limiting resistances in the attachment of PbS. A model with size independent growth kinetics (a ¼ 0) was unable to ft the observed evolution in average radius for any value of k G (Fig. S13 †). Moreover, the standard deviation in diameter obtained from the SAXS data is much smaller than that predicted from the length of the nucleation phase followed by a linear growth rate. Thus, our results cannot be explained by a size independent growth rate, nor by a brief "burst" of nucleation followed by growth. Models with a ¼ 1-1.5 are sufficient to describe the mean particle size and polydispersity as a function of time (Fig. S15 †) and the total amount of solid PbS (Fig. S16 †) across all precursors using a narrow range of k G . An inverse dependence of G on radius (a ¼ 1) is consistent with a diffusion controlled growth process in which k G ¼ D so that G ¼ v m DC solute (t)/r. However, the ftted value of k G corresponds to a solute diffusion constant (D ¼ 3.46 10 11 m 2 s 1 ) on the same order as the diffusivity of a 10 nm nanocrystal. 1 H nuclear magnetic resonance diffusion order spectroscopy (DOSY NMR) of lead oleate in hexadecane at 110 C gives a diffusivity of 3.6 10 10 m 2 s 1 , an order of magnitude faster than the diffusivity from the G ¼ v m DC solute (t)/r model ft. Thus the ftted value of D is too slow to be explained by diffusion of surfactant stabilized monomers through the solvent. To further test whether diffusion through the bulk solution limits the growth kinetics, PbS syntheses were performed in several n-alkane solvents (C8 to C20), whose viscosity varies from 0.24 to 1.41 N m s 2 at 100 C. 29 The viscosity increase causes a small reduction ($25%) in k UV-Vis obs and slightly reduces the fnal nanocrystal size (Fig. S19 †). If the growth is limited by diffusion under these conditions, the seven-fold increase in viscosity should cause an approximately seven-fold decrease in the diffusion coefficient and a corresponding increase in the nanocrystal volume. As seen in Fig. S19, † the increase in solvent viscosity slightly decreases the fnal size, contrary to this prediction. Instead, the increased viscosity correlates with a slight increase in the growth rate. These observations are inconsistent with a growth process that is limited by diffusion through bulk solution. Thus, another mechanism is required to explain the size dependence of the growth rate. Each of the microscopic steps during the surface reaction may depend on the nanoparticle size, including ligand penetration, surface binding, migration, and facet nucleation. Any or all of these steps could be influenced by size dependent structures, such as: (1) strain, 30,31 ligand coverage and binding strength, 32,33 and the ratio of atoms on corners, edges, and facets, (2) transitions between magic sizes with kinetics that are governed by 2D facet nucleation, 34 and (3) attachment kinetics that are limited by penetration of a surfactant ligand layer, among others. Mechanisms ( 2) and ( 3) are amenable to simple phenomenological models based on the nanoparticle size and geometry. (2). A single growth rate constant is found that is consistent across all three thioureas. The central idea in mechanism (2), for example, is that nanocrystal shapes with completed facets constitute highly stable magic clusters. 34 Each facet represents a confned region where 2D-nucleation and growth must occur to make the transition to the next size. Models suggest that the transitions from one magic cluster size to the next become more difficult as the nanoparticle grows larger. 35 It is not yet clear whether the magic cluster mechanisms are important for quasi-spherical metal chalcogenide nanocrystals. Here we elaborate on model (3), i.e. attachment limited by penetration of the surfactant ligand shell. We construct a model based on three key assumptions: (i) that PbS units in the bulk solution diffuse to the outer edge of the ligand shell with negligible resistance, (ii) that the ligand shell has approximately uniform thickness around the nanoparticle, and (iii) those PbS units that penetrate to the inner edge of the shell are immediately incorporated into the PbS nanoparticle. According to this mechanism, the key parameters that control the attachment rate are the thickness of the oleate ligand shell, the partition coefficient of PbS to the oleate ligand shell from the bulk solution, and the effective diffusivity of PbS through the ligand layer. The predicted growth rate is here l is the ligand shell thickness, v is the molar volume of solid PbS, K is the partition coefficient of PbS from bulk solution to the ligand shell, and D shell is the diffusivity of PbS through the ligand shell. The derivation of this equation can be found in the ESI. † Fig. S17 † shows how our assumptions determine boundary conditions on the PbS concentration at the inner and outer edges of the ligand shell. For large radii (where r [ l), the growth rate becomes proportional to the ligand shell permeability (G ¼ v m C solute (t) D shell K/l). At small sizes (l > r) the growth rate approaches v m -D shell KC solute (t)/r. Note that the oleate ligand shell thickness is about 2 nm, which is similar to the diameter of nanocrystals produced during the nucleation period shown in Fig. 3 and 4. In our calculations, v and l are known quantities, and C solute (t) was experimentally determined as described above. Only the distribution of radii (r) is being predicted and D shell K is a ftting parameter. Fig. 5 shows that the model based on eqn (2) accurately predicts the experimental mean radius and polydispersity data across all three precursors using a narrow range of the lumped adjustable parameter D shell K. The ability to ft our results using a single parameter across a factor of two in solute supply kinetics highlights the reproducibility of our reaction conditions and the reliability of our model. The binding strength and coverage of ligands was recently proposed to cause size dependent growth kinetics and size distribution focusing of Pd nanocrystals. 33 Population balance modelling and density functional theory calculations demonstrated that phosphine ligands bind small Pd nanocrystals with lower affinity and can thereby induce size dependent growth kinetics and size distribution focusing. Another computational study on the growth mechanism of indium phosphide nanocrystals 32 demonstrated size dependent surface reactivity including tighter binding of carboxylates and less favorable attachment of phosphide to large crystallites. These fndings were proposed to explain the well-known reluctance of InP nanocrystals to grow beyond a few nanometers in size. 36 Similarly, the binding of lead oleate to PbS is known to be weaker for small crystallites. 37 Moreover, increasing the concentration of lead oleate and lead chloride during the synthesis of PbS nanocrystals, 23,24 or phosphines in the synthesis of Pd nanocrystals 33 decreases the fnal size and size distribution. These effects are consistent with slower, more size dependent growth kinetics at higher ligand coverages. Finally, it has been demonstrated that increasing the chain length of carboxylate surfactants decreases the size and size distribution of colloidal CdSe nanocrystals. 38 These results clearly illustrate how ligand binding, coverage, and structure can induce size dependent growth kinetics. Together they point to a new picture of size distribution focusing that can be addressed using surface coordination chemistry. The slow nucleation observed here clearly demonstrates that the burst of nucleation inherent in LaMer's proposal 39 is not applicable to this canonical colloidal crystalline material. Neither is the diffusion limited size focusing mechanism described by Reiss. 15 Several recent studies on platinum, 40 cadmium selenide, 41 indium phosphide, 36 and palladium 33 nanocrystals reach a similar conclusion. These results require a reinvention of the core principles used to explain the size distributions of colloidal crystals. More detailed investigations of the rate limiting attachment process and its structural origins are important opportunities to advance the rational design of crystalline materials. models of PbS nanocrystals. We thank Vincent Klein and the "ELINSTRU" team of the Laboratoire de Physique des Solides for the conception and building of the remotely controlled injector.

chemsum

{"title": "Persistent nucleation and size dependent attachment kinetics produce monodisperse PbS nanocrystals", "journal": "Royal Society of Chemistry (RSC)"}

fiddle._simultaneous_indexing_and_structure_solution_from_powder_diffraction_data_using_a_genetic_al

## Abstract: The usual process for crystal structure determination from powder diffraction data consists of (1) indexing of the powder pattern, (2) space group determination, (3) structure solution and (4) structure refinement. Despite the success of methods for powder indexing, there are many cases in which the very first step in a structure determination fails or is far from straightforward. Due to a number of fundamental and experimental problems, like peak broadening, the presence of impurity phases, dominant zones and geometrical ambiguities, powder indexing will remain difficult in many cases, thereby hampering the next steps in the structure determination. We present a method for the determination of crystal structures from powder diffraction data that circumvents the difficulties associated with separate indexing. Structure determination from powder diffraction data can be seen as a process of global optimization of all model parameters, including the unit cell parameters. For the simultaneous optimization of the parameters that describe a crystal structure a genetic algorithm is used together with a pattern matching technique based on auto and cross correlation functions. We show that this "onepot" strategy for indexing and structure determination can successfully be used for cases for which indexing is problematic. ## INTRODUCTION Nowadays, single-crystal X-ray diffraction is a standard method for the determination of molecular structures and for the elucidation of intra and intermolecular interactions in various materials. Often, however, crystals suitable for single-crystal diffraction cannot be obtained. Compounds that show polymorphism or solvate (hydrate) formation frequently cause difficulties in the crystallization of one of the modifications. In addition, most of the times recrystallization of unstable compounds that are initially obtained as microcrystalline powders is not feasible. A possible route for crystal structure analysis of "problematic" compounds is to make optimal use of powder diffraction data. These data contain less information than single crystal diffraction data but for a lot of compounds powder diffraction has allowed for a full crystal structure determination. The complexity of crystal structures determined from powder diffraction data has steadily increased through further development of "traditional" methods for structure determination in reciprocal space and application of global optimization algorithms in direct space (Harris et al., 2001;David et al. 2002;Favre-Nicolin et al., 2002;Altomare et al., 2004;Altomare et al., 2004;Tremayne, 2004;David and Shankland, 2008). The published number of crystal structures determined by powder diffraction is still rather limited compared to the enormous number of structures determined by single-crystal diffraction. Clearly, structure determination from powder diffraction data is not a generally applicable method yet. The lack of success in the indexing step has become one of the major bottlenecks in structure determination from powder diffraction data. There are a number of fundamental and experimental problems that can make powder indexing problematic or simply impossible: peak broadening (leading to a loss in resolution), increasing peak density at higher angles, peak shifts, systematic or accidental absences of reflections, the presence of impurity phases, inaccuracies in the experimental measurements, dominant zones and geometrical ambiguities. Problems that originate from the use of laboratory instruments, like peak broadening caused by the instrumental profile and experimental inaccuracies, can be reduced by measuring on synchrotron facilities. However, problems originating from specific powder imperfections (impurities, crystallite size, strain, lattice mistakes) cannot be solved by going to more sophisticated instruments. This means that for powder patterns measured on laboratory as well as synchrotron facilities the problem of limited quality can exist. It is clear that particularly for compounds for which no suitable single crystals can be grown the powder method is an attractive alternative. Often, however, the problematic crystallization behavior of these compounds is also reflected by the moderate quality of their microcrystalline powders and their corresponding powder diffraction data. Despite the success of standard methods for powder indexing, like ITO (Visser, 1969), TREOR (Werner, 1985) and DICVOL91 (Boultif, 1991), there are many cases in which the very first step in structure determination fails or is far from straightforward. New strategies and enhancements will certainly extend the possibilities of powder indexing methods (Coelho, 2003;Neumann, 2003;Altomare et al., 2000). However, the problems associated with the nature of powder diffraction data (absences, increasing peak density etc.) together with the usual strategy of applying a separate indexing step, which may lead to geometrical ambiguities, will be a fundamental limitation for improvement. A method called OCEANA was reported for structure determination from powder diffraction data without prior indexing (Padgett et al., 2007). OCEANA is a grid search based method which uses a genetic algorithm to adjust the cell parameters, potential packing energy and Rwp to locate the global minimum in a given space group. The method is limited to Z'≤1. In this paper a different methodology for the determination of crystal structures from powder diffraction data is presented that circumvents the difficulties associated with separate indexing. The method, implemented in the program FIDDLE, has the possibility of searching through the most common space groups, while varying Z' and without taking into account any energy or packing considerations. By using correlations functions for pattern matching the method is able to optimize cell and structural parameters simultaneously using a genetic algorithm, without any assumptions on cell parameters or density. Results from several tests of the method on various organic compounds are presented to demonstrate the effectiveness of the concept. Moreover, using FIDDLE, three previously unknown crystal structures, cases for which indexing was problematic, could be determined successfully. ## Program description There is no fundamental reason to separate the process of unit cell determination and the process of structure solution. If one is able to describe the main part of a crystal structure by a model consisting of a discrete set of parameters (cell parameters {a, b, c space group number S and -using information on molecular structures -a limited set of structural parameters describing the positional parameters {(x, y, z)j}), then structure determination from powder diffraction data can be seen as a process of (global) optimization of all crystal structure parameters, including the unit cell parameters. This is the strategy implemented in the FIDDLE program. Structure solution then means that in the end the calculated powder diffraction pattern corresponding to the complete set of optimized parameters, including the cell parameters, must match the experimental pattern, assuming that there is one unique structure which has this property. The FIDDLE method is a direct space method. Prior knowledge of the structure of molecular fragments (and the internal degrees of freedom) is needed to define a model of the crystal structure in terms of structural parameters that can be optimized. Like in other direct space methods FIDDLE reduces the set of positional parameters to a set of rotation angles {}, which define the orientation of a molecular fragment in the unit cell, a translation vector {x, y, z}, which defines the position of the molecular fragment in the cell and a set of n torsion angles n}, which describe the intramolecular geometry of the fragment. Part of the parameters, the cell parameters, define the positions of the peaks in the powder pattern. The other parameters, space group and positional parameters, define the intensities of the peaks. Together, they define the total pattern (not regarding zero-point shifts and peak shapes). A suitable measure for similarity between experimental and calculated pattern must be used to be able to optimize both sets of parameters simultaneously. FIDDLE uses weighted auto and cross correlation functions for pattern matching and the mathematics for this fitness function are described below. For space group determination FIDDLE uses a strategy that is also applied in crystal structure prediction: the frequency of the occurrence of space groups in the CSD (Cambridge Structural Database) is used and systematically the most common space groups are explored. About 79% of all organic and organometallic compounds crystallize in only 5 space groups: P21/c, P-1, P212121, C2/c and P21 (Allen, 2002). Chiral molecules (non-superimposable on their mirror image) crystallize in Sohncke space groups and this further reduces the number of possible space groups. FIDDLE explores the space groups P21/c, P-1, P212121, P21, C2/c, Pbca, Pnma, Pbcn, and P1 and varies Z'. Prior knowledge about Z' obtained from methods like solid-state NMR can reduce the problem. Fig. 1 shows the main steps in the FIDDLE procedure. The Genetic Algorithm A Genetic Algorithm (GA) is a global optimisation method based on evolution principles (Harris et al., 1998). Evolutionary operations such as mating, mutation and "natural selection" are applied, through which members with the highest fitness value within a population survive and procreate. Genetic algorithms can be applied to any problem in which the quantity to be optimized can be written as a function of a set of variables (Goldberg, 1989;Cartwright, 1993;Keane, 1996). The settings of the Genetic Algorithm implemented in the FIDDLE program were determined by trial and error and were initially based on experience with the determination of molecular constants from rovibronic spectra with genetic algorithms (Hageman et al., 2000). Table 1 shows the genetic algorithm settings used in FIDDLE. The unit cell parameters (cell edges and angles) and the positional parameters such as orientation angles and the torsion angles are randomly set within a range of values during initialization (see Table 1). This also holds for the translation vector(s) but the boundaries for these parameters are dependent on the space group. ## Comparison of powder diffraction patterns in FIDDLE: the fitness function The RMS (Root Mean Square) criterion is a well known criterion that is based on the sum of the squared differences between n observed and n calculated data values, respectively yi(obs) and yi(calc). A well-known criterion for powder pattern similarity is Rwp (R-weighted pattern). This criterion is used for the refinement of crystal structures on XRPD data and also for the structure solution from XRPD data. Rwp is similar to an RMS criterion but the squared differences are weighted according to the standard deviations in the observed intensities. Rwp can only be used when the peak positions in the observed and calculated powder patterns are the same or at least very close. It is therefore useless for the comparison of powder patterns corresponding to structures with significantly different unit cells. The WCC (Weighted Cross Correlation) criterion was developed to deal with cases where two patterns are different with respect to peak positions. It is based on correlation functions (de Gelder et al., 2001): The weighting function, which extracts information from the correlation functions, can be adapted to influence the sensitivity for shifts in peak positions, in XRPD as a result of lattice parameter variations. This is done by changing the value of width l: The WCC criterion is always normalized and scaling is unnecessary since this is done implicitly. Comparison of deformed patterns, caused by unit cell variations, is possible with the WCC criterion since it recognizes shifted peaks. When we divide the intensities of the powder patterns by the standard deviations in the observed intensities and calculate WCC at r = 0, we obtain the relation between WCC and Rwp, a relation between a sine and cosine function: So, (only) for r = 0 there is a direct and clear relation between Rwp and WCC. EXPERIMENTAL 16 compounds covering a range of structural complexity, number of torsion angles, different Z' and different space groups were selected from the CSD in order to test the methodology implemented in the FIDDLE program (Table 2). The known crystal structure of morphine anhydrate was re-determined from laboratory X-ray powder diffraction data with the FIDDLE program. This re-determination was carried out to test the performance of the FIDDLE method for experimental data. The crystal structures of three compounds (ethinyl estradiol anhydrate, naltrexone monohydrate and creatine anhydrate) were determined from laboratory X-ray powder diffraction data using the FIDDLE program. Indexing the three powder patterns with the most commonly used indexing programs, such as: DICVOL91, ITO and TREOR, was not successful and therefore the determination of these three crystal structures was of particular interest. The X-ray powder diffraction measurements for the structure solutions were performed using a Bruker D8 AXS Advance X-ray Diffractometer. The D8 was equipped with a Johansson type monochromator with a focusing curved Ge 111 crystal. A VNTEC-1 detector was used with an effective angular region of 2. The data were collected in transmission capillary geometry using monochromatic Cu K1 radiation. The most important instrumental and data collection parameters are presented in Table 3. ## TESTS APPLIED TO CALCULATED POWDER DIFFRACTION DATA (A) The crystal structures of 16 compounds selected from the CSD database were determined with FIDDLE to verify the effectiveness of the program and the methodology, for cases with various symmetries and structural complexities (Table 2 and 4). The input molecular models were also obtained from the CSD. The calculations were performed on 6 AMD Dual-Core Opteron 280 machines (two dual core 2.4GHz processors each), all running under the LINUX operating system. The total number of seeds used for the various crystal structure determinations was between 2400 and 4800. The maximum 2θ range used was 40. For each compound the tests were done in the correct space group and for the true value for Z'. On the best FIDDLE solutions having the highest fitness values a full final Rietveld refinement was carried out using Topas 3. Fig. 2 shows a representative overlay between the solution with the highest fitness value obtained from FIDDLE and the known crystal structure in the CSD for one of the studied compounds (FAFWIS02). It is clear that the final result obtained with FIDDLE is close to the true crystal structure of this compound. The compound has one torsion angle and crystallizes in the orthorhombic space group P212121 with unit cell parameters: a=39.700(1) , b=14.129(2) , c=3.835(6) , V=2151.51 3 , Z'=2. From the data presented in Table 4 it can clearly be seen that when the correct space group and Z' is used, the correct unit cell parameters are always found for the top twenty solutions. Always the first solution with the highest fitness value appeared to be the correct crystal structure. ## Target pattern Best pattern in population Target pattern ## Best pattern in population To illustrate that in general the correct crystal structure is the one having the highest fitness value and that the correct unit cell parameters are more often found than a complete solution, a simple test on sucrose data was performed. Sucrose is a compound crystallizing in the monoclinic space group P21, with unit cell parameters: a=7.7235(5) , b=8.6786(7) , c=10.824(1) , =102.982(3)°, V= 706.98 3 , Z'=1 and having five torsion angles. For this compound 20 optimizations were performed. The graph showing the different fitness values vs. the number of generations for sucrose is presented in Fig. 3. The simple test for sucrose showed that the correct unit cell parameters are found for all 20 optimizations and that the solution with the highest fitness value corresponds to the correct crystal structure. ## TESTS APPLIED TO CALCULATED POWDER DIFFRACTION DATA (B) Another experiment was carried out on the sucrose data in order to check the influence of the value of the width of the weighting function. For this test the space group and Z' were again fixed. 2400 Seeds and a maximum 2θ value of 40° were used. Fig. 5 shows the width vs. the number of correct solutions found in the top 80. There is a tendency to a smaller number of correct solutions for smaller values of the width, which indicated again that going to narrow weighting functions, with Rwp as an extreme case, reduces the performance of the genetic algorithm. On the other hand there is a wide range for the value of the width for which good results are obtained. Experience showed that a width of 0.7 is a good compromise between calculation speed and efficiency. Figure 5. The number of correct solutions vs. the width Searching through all ten space groups implemented in FIDDLE while varying Z' from 1 to 3 and using 1.5 for the width, the correct crystal structure of sucrose was again found, the best solution having a fitness value of 0.9690 (see Fig. 6). This experiment mimics the actual situation in a structure solution from powder diffraction data where in principle space group and Z' are unknown. No. correct solutions top 80 0.5 0.6 0.7 0.8 0.9 Figure 6. The overlay between the solution with the highest fitness value obtained from FIDDLE (red), obtained after varying the space group and Z', and the known crystal structure in the CSD (green) for sucrose TEST APPLIED TO EXPERIMENTAL DATA When re-determining the crystal structure of morphine anhydrate, the effectiveness of the program was proved once again. For this test the space group and Z' were set to their correct value. The number of seeds used was 9600. A width of 0.7 and a maximum 2θ of 50° was used. The first 8 solutions having the highest fitness value returned the correct crystal structure, although the correct unit cell parameters were found within the first 37 solutions. The quality of the best solution returned by FIDDLE could be used directly for Rietveld refinement. The full Rietveld refinement proceeded smoothly to reach a minimum characterized by a good fit to the diffraction profiles presented in Fig. 7 Of particular interest were the crystal structures of ethinyl estradiol anhydrate, naloxone monohydrate and creatine anhydrate. After collecting the X-ray powder diffraction data for the three compounds, indexing using DICVOL91, ITO and TREOR failed. The crystal structures of ethinyl estradiol, naloxone monohydrate and creatine anhydrate could successfully be determined using FIDDLE. Prior information regarding the number of molecules per asymmetric unit was obtained from solid-state NMR and used as input in FIDDLE, in order to reduce the computing time. The input molecules were obtained from related structures in the CSD. Searching through all ten space groups implemented in FIDDLE, using the information about Z' obtained from solid-state NMR, and using a maximum 2θ of 40° and 0.7 for the width, the final crystal structures of the three compounds were determined successfully. 4800 Seeds were used in total for all space groups. For ethinyl estradiol the first 40 solutions obtained from FIDDLE indicated the correct unit cell. In the case of naloxone monohydrate and creatine anhydrate the first 5, respectively 4 solutions indicated the correct unit cell. On the best FIDDLE solutions having the highest fitness value a full final Rietveld refinement was carried out using Topas 3. The full Rietveld refinements proceeded smoothly to reach a minimum characterized by a good fit to the diffraction profiles presented in Fig. 9 for all three cases. Comparisons between the best solution obtained from FIDDLE for ethinyl estradiol anhydrate, naloxone monohydrate and creatine anhydrate and the final structures obtained after Rietveld refinement are presented in Fig. 10. Initially, the crystal structure of ethinyl estradiol was determined with the use of four programs: IsoQuest, FIDDLE, DASH and TOPAS 3. The IsoQuest program was used in order to get information about the unit cell parameters from isostructural compounds in the CSD. IsoQuest is a program, which can use as input the X-ray powder diffraction pattern and can search for isostructural compounds in the CSD (de Gelder et al., 2006). As a result an isostructural compound (CSD refcode -EYHENO) was found. IsoQuest has an extremely low computing time and therefore is a good choice for obtaining information about the space group and Z'. Therefore the computing time of FIDDLE can be reduced significantly. The information about the space group obtained from IsoQuest together with the information obtained from solid-state NMR about Z' was used in the FIDDLE program in order to determine the crystal structure of ethinyl estradiol anhydrate. After running 2400 seeds and a width of 1.5 the unit cell of ethinyl estradiol anhydrate was obtained. Using the unit cell from FIDDLE, the DASH program was applied, followed by a full Rietveld refinement (using TOPAS 3). Afterwards, several tests were performed on ethinyl estradiol anhydrate data using the FIDDLE program only. It was shown that ethinyl estradiol anhydrate could also be determined with FIDDLE alone (Rwp=2.15). Therefore, when IsoQuest does not give any useful results, performing a full determination with FIDDLE always leads to a correct result although computing time is higher. Although the quality of the powder diffraction patterns recorded for ethinyl estradiol anhydrate and naloxone monohydrate was high enough for crystal structure determination, standard indexing proved not to be successful. After determining the crystal structure of the two forms, the FIDDLE results pointed at impurity peaks appearing at low 2θ angles, peaks that probably hampered the indexing process. The impurity in the case of ethinyl estradiol corresponds to the hemi-hydrate form and for naloxone monohydrate to the anhydrate form. Eventually we were able to obtain a pure form of ethinyl estradiol anhydrate and therefore the data presented here do not show any impurity corresponding to the hemi-hydrate form. This is not the case for naloxone monohydrate for which the purest form of the monohydrate still had an impurity peak corresponding to the anhydrate form. In the case of creatine anhydrate, the quality of the X-ray powder diffraction pattern was low and therefore the indexing step was impeded by severe peak overlap. Even using the knowledge obtained from FIDDLE about the unit cell parameters, the quality of the recorded data restricted a Pawley refinement and subsequent structure solution with other programs. ## CONCLUSIONS Simultaneous indexing and structure solution is possible using a global optimization approach. FIDDLE was successfully used to determine known and unknown crystal structures and can be applied to compounds for which the indexing step is impeded, for several reasons. Indexing with FIDDLE is computationally cheaper than complete structure determination, meaning that when a structure is not fully solved the program can still deliver the correct unit cell. Any information related to the number of molecules per asymmetric unit obtained from solid-state NMR or knowledge about the chirality of the compounds may be input into the program in order to reduce the computational time. Nevertheless, FIDDLE is capable of solving structures for which unit cell, space group and Z' are unknown.

chemsum

{"title": "FIDDLE. Simultaneous Indexing and Structure Solution from Powder Diffraction Data using a Genetic Algorithm and Correlation Functions", "journal": "ChemRxiv"}

characterization_of_o-acetylation_in_sialoglycans_by_maldi-ms_using_a_combination_of_methylamidation

## Abstract: O-Acetylation of sialic acid in protein N-glycans is an important modification and can occur at either 4-, 7-, 8-or 9-position in various combinations. This modification is usually labile under alkaline reaction conditions. Consequently, a permethylation-based analytical method, which has been widely used in glycomics studies, is not suitable for profiling O-acetylation of sialic acids due to the harsh reaction conditions. Alternatively, methylamidation can be used for N-glycan analysis without affecting the base-labile modification of sialic acid. In this report, we applied both permethylation and methylamidation approaches to the analysis of O-acetylation in sialic acids. It has been demonstrated that methylamidation not only stabilizes sialic acids during MALDI processing but also allow for characterization of their O-acetylation pattern. In addition, LC-MS/MS experiments were carried out to distinguish between the O-acetylated glycans with potential isomeric structures. The repeatability of methylamidation was examined to evaluate the applicability of the approach to profiling of O-acetylation in sialic acids. In conclusion, the combination of methylamidation and permethylation methodology is a powerful MALDI-TOF MS-based tool for profiling O-acetylation in sialic acids applicable to screening of N-glycans.Glycosylation is a universal post-translational modification of proteins in eukaryotic species and plays key roles in protein folding, protein-protein interaction, cell-cell recognition, cancer metastasis, and the immune system 1-4 . In protein N-linked glycans, sialic acids, usually found as terminal monosaccharides, are the most important monosaccharides for human evolution 5 . Sialic acids are a family of 9-carbon carboxylated sugars, in which the most common one is N-acetyl-neuraminic acid (2-keto-5-acetamido-3,5-dideoxy-D-glycero-D-g alactononulopyranos-1-onic acid) (Neu5NAc) 6 . Neu5NAc can be further modified, of which O-acetylation is one of the major modifications that significantly alters biological properties of the parent molecule. O-Acetylation can occur at either 4-, 7-, 8-, or 9-hydroxyl position 7,8 and is regulated in a molecule specific, tissue-specific, and developmentally regulated fashion [9][10][11][12][13][14][15] . O-Acetyl groups may cause conformational change of glycoproteins and reduce the hydrophilic properties of sialic acids 16 . Sialic acid acetylesterase has a strong genetic link to susceptibility in relatively common human autoimmune disorders 17 . N-glycolylneuraminic acid (Neu5Gc) is a sialic acid molecule found in most non-human mammals and closely related to Neu5NAc. Neu5Gc is highly and selectively enriched in red meat, such as beef and pork. It can be metabolically incorporated into human tissues from dietary sources. Moreover, it was found that the interactions of Neu5Gc antigen with circulating anti-Neu5Gc antibodies in human body could potentially induce inflammation 18 .Electrospray ionization (ESI) and matrix-assisted laser desorption/ionization (MALDI) mass spectrometry (MS) have been widely used in glycomics studies [19][20][21][22][23][24][25] . Using capillary electrophoresis-ESI-MS, we investigated the changes of O-acetylation pattern in sialic acids in N-glycans of salmon (Salmo salar) serum 26,27 . The analysis of native N-glycans revealed that the O-acetylation pattern was correlated with long-term handling stress, which was most likely to be a product of immune reaction. With the permethylation process, MALDI-MS is the desired technique for profiling glycans . Methylation of all hydroxyl groups of glycans can significantly increase detection sensitivity because the derivatives are considerably more hydrophobic and stable than native glycans. In addition, permethylated glycans produce more information in tandem mass spectra that is very useful for linkage analysis. Unfortunately, permethylation process removes acetyl esters, which precludes its application to the characterization of O-acetylation of sialic acids. So far, detailed O-acetylation profiles and biological roles of these substitutions have been overlooked because of their lability to conventional purification and detection methods. An alternative strategy for stabilizing a sialic acid is the selective modification of its carboxyl group, including methylamidation , dimethylamidation 19 , methyl esterification 36 and ethyl esterification 37,38 . This study aims at the development of MALDI-TOF MS for glycomics, especially the profiles of O-acetylation in sialic acids. To demonstrate the applicability of the proposed method, we analysed sera from seven different species of carp, including crucian carp, grass carp, silver carp, bighead carp, common carp, bream and black carp from wild fisheries 39 . The enzymatically released and purified fish serum N-glycans were subjected to methylamidation and permethylation derivatization and analysed using MALDI-TOF MS and LC-MS/MS. The tandem mass spectrometry experiments were performed to characterize glycan structures. ## Results and Discussion Crucian carp serum. Permethylation reaction removes O-linked modification, thus the resultant MALDI MS spectrum can be significantly simplified and used to derive the composition of N-glycans. A representative MALDI-MS spectrum of permethylated N-glycans from crucian carp (Carassius carassius) serum sample is presented in Fig. 1a. For simplicity, this serum sample was labelled as Crucian-1. The primary structures of N-glycans isolated from crucian carp serum are similar to that of salmon serum 26,27 . Four major glycoforms, m/z 2431.1, 2792.2, 3241. 4 and m/z 1783.8 (Hex 6 HexNAc 2 ). The non-sialylated biantennary structure was detected at m/z 2069.9 with a composition of Hex 5 HexNAc 4 . The detection of ions at m/z 4051.7 and 4412.9 suggests the presence of tetra-antennary structures with compositions of Neu5NAc 3 Hex 7 HexNAc 6 and Neu5NAc 4 Hex 7 HexNAc 6 , respectively. The MALDI-TOF MS spectrum of methylamidated N-glycans from crucian carp serum sample (Crucian-1) is shown in Fig. 1b. As expected, the methylamidation process completely derivatized all sialic acids and preserved the labile modifications, i.e. O-acetyl groups in sialic acids. The spectrum revealed several clusters of ions, with 42 Da intervals within each cluster. This observation suggests that one sialic acid can be modified with up to three O-acetyl groups. For example, the ions at m/z 1967.7, 2009.7, 2051.7 and 2093.7 may be assigned to monosialylated glycoforms with addition of zero, one, two and three O-acetyl groups, respectively. For the glycans containing two sialic acid residues, the ions at m/z 2271. , respectively. Overall, up to six O-acetyl groups were detected for the N-glycans containing two sialic acid residues. This observation reveals the most extensively O-acetylated N-glycans in any fish species that have been reported so far 26,27 . O-acetylation of sialic acid can occur at four positions, i.e. the 4-, 7-, 8-, or 9-hydroxyl position 7,8 . For any disialylated N-glycan containing six O-acetyl groups, two isomers may exist. Each of two sialic acids has either three O-acetyl groups (3 + 3) or one has four and the other one has two O-acetyl groups (4 + 2). Therefore, it is necessary to perform tandem mass spectrometry to characterize the distribution of O-acetyl groups. We performed LC-MS/MS experiments on the methylamidated sample and the representative MS/MS spectra are shown in Figure S1 in Supporting Information. For non O-acetylated glycans, the prominent fragment ions at m/z 305.1 and m/z 287.1 correspond to the methylamidated Neu5NAc (addition of 13 Da) and its anhydrous form (Figure S1a in Supporting Information). The non O-acetylated parent ions also afforded a Neu5NAc-containing fragment ion at m/z 670.1, corresponding to Neu5NAc 1 Hex 1 HexNAc 1 . When one of two sialic acids was O-acetylated, the tandem MS spectrum showed two fragment ions with a difference of 42 Da at m/z 670.3 and m/z 712.3 (Figure S1b in Supporting Information). The intensities of two fragment ions associated with trisaccharide residues, Neu5NAc 1 Hex 1 HexNAc 1 and OAc 1 Neu5NAc 1 Hex 1 HexNAc 1 , were almost identical. However, the intensities of the fragment ions that corresponded to the monosaccharide residues, Neu5NAc and OAc 1 Neu5NAc, exhibited significant difference (m/z 305.1 and m/z 347.1 in Figure S1b in Supporting Information). For the glycans containing two O-acetyl groups, the predominant fragment ion at m/z 712.3 revealed that each sialic acid contained one O-acetyl group (1 + 1) rather than the combination of two O-acetyl groups and no O-acetyl modification (2 + 0). For the glycans containing three O-acetyl groups, the major species were 2 + 1 (m/z 712.3 and m/z 754.3) distribution, while the combination of 3 + 0 was only a minor species (m/z 796.3 and m/z 670.3) (Figure S1c in Supporting Information). For the glycans with six O-acetyl groups, the tandem MS spectrum demonstrated that each sialic acid was modified with three O-acetyl groups (Figure S1f in Supporting Information), which implied that the number of O-acetyl modification for a sialic acid was not higher than 3. We also noted that the LC-MS/MS experiments on the glycans containing multi sialic acids did not produce any fragments that could be correlated to the presence of disialic acid residues, i.e. Neu5NAc-Neu5NAc. To study the biological variation, N-glycans from two additional individual fish serum samples, labelled as Crucian-2 and Crucian-3, were analysed and the obtained mass spectra of the replicate serum samples are presented in Figure S2 in Supporting Information. Although the relative abundance of O-acetylated species varied, the maximum numbers of O-acetyl groups for disialylated N-glycan remained unchanged among the three individual fish. The detected ions and the proposed compositions are summarized in Table S1 in Supporting Information. ## Repeatability of methylamidation. Methylamidation has been previously proven to be an effective derivatization strategy for sialic acid-containing N-glycans . Complete methylamidation of 2,6-sialyllactose could be achieved within 10 min, whereas the complete reaction for 2,3-sialyllactose took about 30 min 32 . So far, the applicability of methylamidation to profiling O-acetylation patterns of sialic acids has not been investigated. Because no O-acetylated N-glycan standards were available, we used the N-glycans from pooled serum samples of multiple crucian carp for repeatability testing. The glycans Neu5NAc 2 Hex 5 HexNAc 4 with different O-acetyl groups were selected to investigate the stability of O-acetyl groups, by comparing ESI-MS analysis of native glycans and MALDI-MS analysis of methylamidated glycans. The intensities of native glycans from 5 technical replicates in ESI-MS experiments were summed and normalized to the sum of intensities of the highest peak (Fig. 2a). To investigate the stability of O-acetyl groups under methylamidation conditions, 5 samples from serum pool were individually PNGase-F digested and derivatized, followed by MALDI-MS analysis. The intensities of each glycan from 3 spots in MALDI target were summed and normalized to the sum of intensities of the highest peak (i.e. the glycan containing three O-acetyl groups). The average relative intensity values for six glycoforms are presented in Fig. 2b, with the CVs ranging from 1.9% to 25.4%. We then derivatized 5 N-glycan samples from the same serum pool collected on each day for two additional days to evaluate the day-to-day repeatability 38 . The relative intensity values of six glycoforms for three different days are illustrated in Fig. 2c, with the CVs ranging from 0.8 to 27.8%. While the profiles between native glycans (ESI-MS) and their methylamine derivatives (MALDI-MS) were similar, relatively lower intensities in ESI-MS spectra were observed for the glycans containing four, five and six O-acetyl groups, respectively. To further demonstrate the applicability of the approach, we analysed N-glycans from serum samples of additional six fish species. , respectively. Trisialylated glycans were found to be modified with up to six O-acetyl groups, i.e. m/z 3024.9 (2xOAc), 3066.9 (3xOAc), 3108.9 (4xOAc), 3150.9 (5xOAc) and 3192.9 (6xOAc). ## Common carp serum. The MALDI-MS spectra for the N-glycans from serum samples of two additional individual common carp, labelled as Common-2 and Common-3, are presented in Figure S4 in Supporting Information. The O-acetylation pattern in N-glycans from Common-2 (Figure S4b and Figure S4e) was similar to that shown in Common-1 (Figure S4a and Figure S4d). However, O-acetylation pattern in N-glycans from Common-3 showed much lower degree of O-acetylation (Figure S4c and Figure S4f), in which only four and six O-acetyl groups were detected in disialylated glycans and trisialylated glycans, respectively. The detected ions and their corresponding compositions are summarized in Table S2 in Supporting Information. Two additional biological replicates of silver carp serum samples were analysed and significant variations were observed between the three replicates. For example, the glycan with the composition of Neu5NAc 1 Hex 6 HexNAc 5 (m/z 2880.3) was only detected in the MALDI MS spectra of permethylated glycans from Silver-2 (Figure S8b and Figure S8e in Supporting Information) or Silver-3 (Figure S8c and Figure S8f in Supporting Information), and the corresponding methylamidated glycans were detected at m/z 2332.9 (Neu5NAc 1 Hex 6 HexNAc 5 ) and m/z 2374.9 (OAc 1 Neu5NAc 1 Hex 6 HexNAc 5 ). Table S4 in Supporting Information summarizes the detected ions and their corresponding compositions. ## Bream carp serum. The N-glycans from a representative serum sample of bream carp, labelled as Bream-1, were initially analysed using permethylation method (Fig. 6a). The most abundant structures are the triantennary glycoforms with different numbers of sialic acids and Hex terminal motifs. For example, the ion detected at m/z 2723.3 is a triantennary structure with the chemical composition of Hex 7 HexNAc 5 . The ions at m/z 3084.4 correspond to the composition of Neu5NAc 1 Hex 7 HexNAc 5 , with the addition of a sialic acid group. More complex structures were observed as the tetraantennary glycans with ions at m/z 3894.8 and 4255.9, respectively. Methylamidation was used to investigate the O-acetyl modification in bream carp serum N-glycans as well (Fig. 6b). The monosialylated glycans at m/z 2212.7, 2374.7 and 2536.7 can be assigned to the addition of an O-acetyl group to the oligosaccharides with m/z 2170.8, 2332.7 and 2494.7, respectively. The disialylated glycans, with the addition of two O-acetyl groups, were detected at m/z 2720.8 (OAc 2 Neu5NAc 2 Hex 6 HexNAc 5 ) and 2882.8 (OAc 2 Neu5NAc 2 Hex 7 HexNAc 5 ). This observation suggests that the O-acetylation patterns of serum N-glycans from bream carp are similar for silver carp and grass carp. The representative MS/MS spectra are presented in Figure S9 in Supporting Information. Similar glycoforms were detected in two additional biological replicates (Bream-2 and Bream-3) as shown in Figure S10b, Figure S10c, Figure S10e and Figure S10f in Supporting Information, although relative intensities varied among the three replicates. The proposed compositions for most detected ions are presented in Table S5 in Supporting Information. ## Bighead carp serum. The N-glycan profiles of a representative serum sample from bighead carp (Hypophthalmichthys nobilis) were also investigated using both permethylation and methylamidation methods (Fig. 7). For permethylated glycans (Fig. 7a), the most abundant ions corresponded to triantennary glycoforms at S12b, Figure S12c, Figure S12e and Figure S12f in Supporting Information, respectively. The ion at m/z 3037.8 in the spectrum of permethylated glycans corresponded to a composition of Neu5NAc 2 Hex 5 HexNAc 5 . The methylamidated glycans with zero, one or two O-acetyl groups were detected at m/z 2474.9, 2516.9, and 2558.9, respectively. The detected ions and their compositions are summarized in Table S6 in Supporting Information. Black carp serum. MALDI-MS profiling of the permethylated glycans released from a representative serum sample of black carp (Mylopharyngodon piceus), labelled as Black-1, indicated the presence of multi-antennary structures (Fig. 8a), which have been previously found in several different species of fresh-water fish and both unfertilized 40,41 and fertilized fish eggs . The most abundant glycan structures at m/z 2723.3, 3084.5 and 3445.7, correspond to the oligosaccharides with chemical compositions of Hex 7 HexNAc 5 , Neu5NAc 1 Hex 7 HexNAc 5 and Neu5NAc 2 Hex 7 HexNAc 5 , respectively. The glycans from black carp serum sample were mainly found to exhibit complex tri-and tetra-antennary structures. Comparatively, the glycan compositions in crucian carp and common carp sera mainly exhibit biantennary structures and extensive O-acetyl modification. The N-glycans of black carp serum samples were also investigated using the methylamidation method (Fig. 8b). Surprisingly, no O-acetyl modification was detected for this species. The representative MS/MS spectra are shown in Figure S13 in Supporting Information. Two additional biological replicates (i.e. Black-2 and Black-3) confirmed the absence of O-acetylation in sialic acids of N-glycans from black carp serum samples (Figure S14 in Supporting Information). The relative abundance and type of each glycan in black carp serum samples are consistent between the permethylation and methylamidation spectra. The detected ions and their corresponding compositions are summarized in Table S7 in Supporting Information. ## Conclusion In this study, we applied a new derivatization strategy, the combination of permethylation and methylamidation, for the characterization of O-acetylation of sialic acid in N-glycans. Methylamidation allows the stabilization of O-acetylated sialic acid residues; whereas permethylation provides a sensitive technique with simplified MALDI MS spectra that contain unambiguous information on glycan compositions. The serum samples from seven freshwater carp species were analysed, including crucian carp, grass carp, silver carp, common carp, bream and black carp. The results revealed that the N-glycans from sera of different fish species presented significant difference in composition, sialylation pattern and degree of O-acetylation. For example, the crucian carp serum exhibited up to three O-acetyl groups on a single sialic acid residue. The results also suggested that no more than three O-acetyl groups were attached to one sialic acid residue in all glycans studied, even in the ones with the most extensive O-acetyl substitution. While the method was developed for application in aquaculture, it is applicable for the analysis of sialoglycans derived from sera of any other animal species. ## Materials and Methods Chemicals and Materials. 2,5-Dihydroxybenzoic acid (DHB), dimethyl sulfoxide (DMSO), sodium hydroxide, N-methylmorpholine, acetonitrile (ACN), methylamine hydrochloride, (7-azabenzotriazol-1-yloxy) trispyrrolidinophosphonium hexafluorophosphate (PyAOP), trifluoroacetic acid (TFA), 1-butanol, ethanol, porous graphitic carbon (PGC), microcrystalline cellulose (MCC) were obtained from Sigma-Aldrich (St. Louis, MO). Blood was collected from a local fish farm and the serum was separated by centrifugation and stored at − 20 °C until used. N-Glycosidase F (PNGase F) and endoglycosidase buffer pack were purchased from New England Biolabs (Ipswich, MA, USA). Formic acid (FA), methyl iodide, chloroform, empty cartridges and frits were purchased from Aladdin (Shanghai, China). All solutions were prepared using deionized water purified by a Milli-Q purification system (Millipore, MA, USA). Preparation of N-glycans. Fish serum (10 μ L) was dissolved in 50 μ L of sodium phosphate (20 mM, pH 7.5) containing 0.2% SDS and 0.1 M DTT and denatured at 100 °C for 10 min. After cooling, 12 μ L of 10% NP-40 and 38 μ L of water were added. The reaction mixture was incubated with PNGase F (10 units) for 24 h at 37 °C. The sample was then boiled for 5 min to stop the reaction and the released glycans were purified using PGC cartridges. The first serum sample from crucian carp was labelled as Crucian-1 and similar definition was used for the first serum sample of other fish species, i.e. Common-1, Grass-1, Silver-1, Bream-1, Bighead-1 and Black-1. Biological variation was established by analysing two additional replicates of serum samples for each fish species. Accordingly, the additional serum replicates for all species examined were abbreviated as Crucian-2 and Crucian-3, Common-2 and Common-3, Grass-2 and Grass-3, Silver-2 and Silver-3, Bream-2 and Bream-3, Bighead-2 and Bighead-3, Black-2 and Black-3. For comparison, the MALDI-TOF MS results for all three replicates of each species were included in Supporting Information, i.e. Figure S2, Figure S4, Figure S6, Figure S8, Figure S10, Figure S12 and Figure S14. All experiments were performed in accordance with the protocols approved by Ethical Committee of Jianghan University. ## Purification of glycans. PGC cartridges were washed with 3.0 mL of 80% (v/v) ACN containing 0.1% TFA followed by 3.0 mL distilled water. The glycans released by PNGase F were loaded on PGC cartridges and then washed with distilled water (3.0 mL) to remove buffer and salts. Glycans were eluted with 25% ACN in 0.1% TFA (3 × 1 mL). The fraction was collected and dried for further processing. The liquid phase was recovered and dried using a SpeedVac concentrator. ## Permethylation of glycans. Dried oligosaccharide sample was dissolved in an Eppendorf tube using 50 μ L of DMSO. 100 μ L of DMSO-NaOH slurry and 50 μ L of methyl iodide were then added. Tubes were capped tightly and stirred for about 10 min at room temperature. The reaction was stopped by the addition of distilled water (0.5 mL). The permethylated oligosaccharides were extracted into chloroform (0.25 mL) by vortex mixing. The lower organic layer containing the permethylated oligosaccharides was washed with water (3 × 0.5 mL). ## Methylamidation of glycans. Dried glycans were dissolved in 25 μ L of DMSO solution containing 1 M of methylamine hydrochloride and 0.5 M of N-methylmorpholine, followed by addition of 25 μ L PyAOP (50 mM in DMSO) solution. The reaction mixture was vortexed and allowed to proceed at room temperature for 30 min. The glycan derivatives were purified according to a hydrophilic method described previously 32 . Briefly, cellulose particles were first washed with distilled water and then with solvent mixture of 1-butanol/ethanol/H 2 O (4:4:1, v/v/v). The reaction mixture solution was mixed with 5 mg cellulose particles in 1 mL of the organic solvent described above. After gentle shaking for 45 min, the cellulose particles were washed three times with 0.5 mL of the organic solvent. The cellulose particles were incubated with 0.3 mL of a mixture of ethanol/H 2 O (1:2 v/v) for 30 min. The liquid phase was recovered and dried using a SpeedVac concentrator. All samples were redissolved in distilled water prior to MS analysis. ## MALDI-MS analysis. The MALDI-MS spectra were acquired using 4800 MALDI-TOF/TOF (SCIEX, Concord, Canada) equipped with an Nd:YAG laser with 355 nm wavelength of < 500 ps pulse and 200 Hz repetition rate. The spectrometer was operated in the positive reflectron mode. The spectra were accumulated by 1000 laser shots. The MS data were further processed using Dataexplorer 4.0. The samples were loaded onto MALDI target in 0.5 μ L of water and mixed with 0.5 μ L of freshly prepared DHB solution (10 mg/mL in 50% ACN) and allowed to dry in a gentle stream of air. LC-MS/MS analysis. LC-MS/MS experiments were performed using a TripleTOF 5600 System (SCIEX, Canada) and a NanoLC Ultra System (Eksigent, USA), equipped with a trap column (150 μ m i.d. × 1 cm long; PGC, 5 μ m; Proteomics Front, China) and a separation column (75 μ m i.d. × 10 cm long; PGC, 5 μ m; Proteomics Front, China). MS was operated in the positive-ion mode with a mass range of 500-3000 m/z, and MS/MS was acquired in the information dependent acquisition (IDA) mode with a mass range of 100-2000 m/z. The 20 most abundant precursor ions with charge numbers from 2 to 5 were scanned in the IDA mode. Each cycle consisted of a MS acquisition for 0.25 s and a total of 20 MS/MS scans for 2 s. ## Repeatability analysis. Repeatability of methylamidation conditions was investigated by multiple analyses of the same serum samples, including glycan releasing and purification. Using similar experiment design as for testing the repeatability of ethyl esterification conditions 38 , we repeated the entire procedure twice on consecutive days with freshly prepared reagents to establish day-to-day variability. For comparison, the native glycans from the same sample were also analysed using a TripleTOF 5600 System.

chemsum

{"title": "Characterization of O-acetylation in sialoglycans by MALDI-MS using a combination of methylamidation and permethylation", "journal": "Scientific Reports - Nature"}

bifunctional_iminophosphorane_catalysed_enantioselective_sulfa-michael_addition_of_alkyl_thiols_to_a

## Abstract: The first enantioselective sulfa-Michael addition of alkyl thiols to alkenyl benzimidazoles, enabled by a bifunctional iminophosphorane (BIMP) organocatalyst, is described. The iminophosphorane moiety of the catalyst provides the required basicity to deprotonate the thiol nucleophile while the chiral scaffold and H-bond donor control facial selectivity. The reaction is broad in scope with respect to the thiol and benzimidazole reaction partners with the reaction proceeding in up to 98% yield and 96 : 4 er. N-Containing heterocycles are ubiquitous motifs in both biologically active molecules and natural products. Their functionalization, especially when performed in an enantioselective manner, is therefore of particular interest in the feld of organic synthesis. Alkenyl azaarenes have been used extensively as synthetic precursors for the functionalization of N-containing heterocycles. 1 The electron defciency of the aromatic ring, part-activates the conjugated alkene towards Michael-type additions, 2 allowing for the rapid generation of molecular complexity. Most recently, the groups of Harutyunyan, Terada and Meng reported elegant, highly enantioselective Michael additions to alkenyl N-heterocycles employing organocuprates, 3 pyrazoles 4 and B 2 (pin) 2 5 respectively. Our research has focused on developing enantioselective methods utilizing novel bifunctional iminophosphorane (BIMP) organocatalysts, 6 which combine a chiral H-bond donor scaffold 7 with an organo-superbase. 8 More specifcally, BIMP catalysis has been employed in the enantioselective addition of thiols 9,10 to unactivated esters. 6c,g This encouraged us to consider replacing the enoate electrophile with isoelectronic alkenyl benzimidazoles in order to access complex, chiral druglike scaffolds with perfect atom economy and potential applications to medicinal chemistry (Fig. 1). 11 To the best of our knowledge, there have been no reports to date of the enantioselective base catalysed Michael additions to alkenyl benzimidazoles 12 and herein we wish to report our work leading to the frst example, under BIMP catalysis. We chose the readily prepared 4 (E)-2-propenyl-1-tosylbenzimidazole 1 and commercially available 1-propanethiol as model coupling partners to investigate reactivity and selectivity with a selection of bifunctional Brønsted base/H-bond donor catalysts using 3 eq. of thiol at 0.5 M concentration in THF at 22 C for 24 hours (Fig. 2, Table 1). Quinidine derived catalyst A (entry 1) only provided 2 in 12% yield and a negligible 53 : 47 er. We therefore chose to investigate the more basic and more active BIMP catalysts in this reaction and were very pleased to fnd that known BIMP catalyst B 6a bearing one stereocenter provided desired product 2 in 80% yield and 83 : 17 er (entry 2). With signifcant catalyst-enabled reactivity and stereocontrol identifed we then proceeded to investigate second generation catalyst C 6g which provided 2 in improved yield and er at 92% and 86 : 14 respectively (entry 3). Shifting the thiourea moiety further away from the iminophosphorane (D-E) 6d showed no improvement in er over B (entries 4 and 5). We therefore focused on exploring catalysts built around the same chiral scaffold as C. Catalyst F 6g bearing t Bu groups at both stereocenters in the (S,S) confguration afforded 2 in 90% yield and 90 : 10 er (entry 6). Interestingly, a control reaction without any catalyst was found to go to completion (entry 7), indicating that an uncatalysed background reaction 13 pathway was leading to an erosion in the enantiomeric ratio of the product. To suppress this background reactivity, the reaction was diluted to [0.06 M], cooled to 0 C and only 1.2 eq. of thiol were used. The new set of conditions, combined with a solvent switch from THF to Et 2 O, provided 2 in 93% yield and 94 : 6 er using catalyst F (entry 8). Surprisingly a further decrease of the temperature to 40 C led to an erosion of the enantiomeric ratio (entry 9). To further boost the enantiomeric ratio, diastereoisomeric catalyst G 6g was screened. Pleasingly, catalyst G outperformed corresponding diastereomer F affording the desired product in 98% yield and 95 : 5 er (entry 10). With optimal conditions established, we proceeded to explore the scope and limitations of this transformation (Scheme 1). Initially the steric and electronic properties of the thiol nucleophile were varied. Higher order linear, branched and cyclic alkyl substituents on the thiol all provided the corresponding Michael adducts (3-5) with high yields and enantioselectivities. The introduction of a phenyl ring was well tolerated providing 6 in outstanding yield and good er. Appending a silyl group to the thiol nucleophile showed no detrimental effect providing 7 in excellent yield and er. Benzyl thiols provided corresponding Michael adducts 8-10 in high yields in all cases and good enantioselectivity, albeit slightly diminished when compared to simpler alkyl thiols. 15,16 Having investigated the thiol component, we then focused on substituent effects on the benzimidazole core (Scheme 2). Variations to the phenyl backbone did not affect reactivity, disubstitution at C5 and C6 with methyl groups afforded corresponding adduct 11 in 81% yield and 86 : 14 er. Alternating monosubstitution between C5 and C6 did not have a large effect, with bromine containing substrates affording the corresponding Michael adducts (12, 13) in greater than 75% yield and 86 : 14 er allowing for potential further functionalization at both positions. 17 We were pleased to fnd that the high enantioselectivity of the reaction was largely maintained when the nitrogen protecting group was changed from N-tosyl to N-Cbz (14) or N-Boc (15), however in these cases reactivity was found to diminish. This was easily circumvented by running the reaction at 22 C using 3 equivalents of 1-propanethiol. 18 Having varied the substitution pattern on the benzimidazole, we proceeded to investigate the scope with respect to substituents on the alkenyl moiety. The introduction of higher order linear alkyl chains, bearing aromatic, alkene and alkyne substituents, was well-tolerated with all n-propyl thiol Michael additions providing the corresponding products (16-19) in excellent yield and enantioselectivity. When substituting the alkene moiety with an aromatic group, the solvent was switched to THF and reactions were run at 22 C due to decreased solubility and reactivity of the substrates. When a phenyl substituent was introduced on the alkenyl moiety, catalyst G only provided a moderate Michael adduct 20 in 77 : 23 er, however this was boosted to 88 : 12 when using diastereomeric catalyst F. Introducing electron withdrawing groups at either the para or meta positions of the phenyl ring afforded the corresponding products 21-23 in good yield and enantioselectivity; in these cases, however, catalyst G proved superior to F. Finally, when the phenyl ring was exchanged with a 3-pyridyl moiety, it smoothly afforded the corresponding adduct 24 in 74% yield and 90 : 10 er. Increasing the reaction scale 10-fold (1 mmol) afforded 2 in equal yield and er, which upon treatment with HCl (5 M aq.) gave corresponding deprotected product 25 in quantitative yield. Single crystal X-ray analysis of 25 allowed the absolute confguration of sulfa-Michael product 2 to be determined as S when using catalyst G. We were also pleased to fnd that, upon treatment of 2 with m-CPBA, sulfone 26 was obtained in 95% yield with no loss of optical purity (Scheme 3). We used density functional theory (DFT) to investigate the origins of enantioselectivity, performing calculations at the wB97XD/6-31G(d) level of theory (Fig. 3). 19 Calculations considered PPh 3 -derived catalyst G* with the PMP-groups of G modelled by Ph-groups. The most stable conformation of (most enantioselective) catalyst G* has substituents either side of the urea oriented with a hydrogen atom towards sulfur: other rotamers are disfavoured. This creates a pocket with the iminophosphorane positioned above the thiourea (from the perspective of Fig. 3). Two substrate activation modes are possible (A vs. B) and either could in principle lead to the formation of the major observed enantiomer. Computationally, we fnd that the interaction of the thiolate nucleophile with the protonated iminophoshorane and the benzimidazole with the thiourea (mode A) is energetically favored by 4-5 kcal mol 1 over the alternative (mode B) in which the thiourea binds the nucleophile and the benzimidazole to the protonated iminophoshorane. This mode of activation is consistent with the observed sense of enantioselectivity, and with earlier mechanistic proposals of Takemoto. Recent theoretical studies of Grayson and Houk have emphasized the importance of activation mode B in sulfa-Michael reactions promoted by Cinchona-derived catalysts. 20 Our present results suggest that both activation modes may be operative, depending on catalyst and substrate, as originally hypothesized by Soós and I. Pápai. 21 ## Conclusions In summary, the frst enantioselective sulfa-Michael addition of alkyl thiols to alkenyl benzimidazoles has been described. Excellent yields and good enantioselectivities were achieved across a broad range of alkyl thiol and alkenyl benzimidazole reaction partners using a second generation BIMP organocatalyst. This work further demonstrates the versatility and high activity of the BIMP catalyst family, as well as expanding its use in methodology for the synthesis of biologically relevant chiral benzimidazole derivatives. Further investigations into new

chemsum

{"title": "Bifunctional iminophosphorane catalysed enantioselective sulfa-Michael addition of alkyl thiols to alkenyl benzimidazoles", "journal": "Royal Society of Chemistry (RSC)"}

the_substrate_import_mechanism_of_the_human_serotonin_transporter

## Abstract: The serotonin transporter, SERT, initiates the reuptake of extracellular serotonin 1 in the synapse to terminate neurotransmission. Recently, the cryo-EM structures of SERT bound to ibogaine resolved in different states provided a glimpse of functional conformations at atomistic resolution. However, the conformational dynamics and structural transitions to various intermediate states are not fully understood. Furthermore, while experimental SERT structures were complexed with drug molecules and inhibitors, the molecular basis of how the physiological substrate, serotonin, is recognized, bound, and transported remains unclear. In this study, we performed microsecond long simulations of the human SERT to investigate the structural dynamics to various intermediate states and elucidated the complete substrate import pathway.Using Markov state models, we characterized a sequential order of conformational driven ion-coupled substrate binding and transport events and calculated the free energy barriers of conformation transitions associated with the import mechanism. We identified a set of residues that recognize the substrate at the extracellular surface of SERT and our simulations also revealed a third sodium ion binding site coordinated by Glu136 and Glu508 in a buried cavity which helps maintain the conserved fold adjacent to the orthosteric site for transport function. The mutation of these residues results in a complete loss of transport activity. Our study provides novel insights on the molecular basis of dynamics driven ion-substrate recognition and transport of SERT that can serve as a model for other closely related neurotransmitter transporters. ## Introduction The serotonin transporter (SERT) terminates synaptic transmission by catalyzing the reuptake of extracellular serotonin from the synapse. Reuptake is critical for normal serotonergic signaling in the brain with implications on mood, cognition, behavior, and appetite. 1 Consequently, improper SERT function is associated with numerous neurological disorders including depression, autism, and bipolar disorder. 2 Additionally, SERT is expressed in platelet membranes and regulates blood coagulation throughout the circulatory system. 3 Given its medical importance, SERT is a major molecular target for therapeutic drugs and drugs of abuse. 4,5 Similar to other members of the solute carrier 6 (SLC6) neurotransmitter transporter family, SERT mediated serotonin (5-hydroxytryptamine; 5HT) translocation from the synapse and surrounding area is coupled to favorable ion co-transport of one Na + with a Clion dependence, and export of one K + to complete an overall electroneutral cycle. Other conduction states and stoichiometries with unclear physiological significance may occur under different conditions. SERT, and the closely related dopamine transporter (DAT) and norepinephrine transporter (NET), belongs to a class of monoamine transporters in the neurotransmitter:sodium symporter (NSS) family. These members share a distinct 12 transmembrane (TM) helix architecture known as the LeuT fold, which consist of 12 transmembrane (TM) helices, with TM1-5 and TM6-10 forming inverted pentahelical repeats around a pseudo two-fold axis of symmetry. 16,17 Cysteine labeling studies on SERT revealed that the 5HT binding site is accessible from both extracellular and intracellular sides of the membrane, providing the first glimpse of evidence of an alternating access model. 18 Quick and Javitch developed a proteomic approach to characterize the sodium-dependent substrate transport mechanism in the tyrosine transporter Tyt1. 19 These biochemical studies elucidated that the NSS family of transporters function based on the principle of an alternating access mechanism. 20 Crystal structures of the bacterial NSS homolog leucine transporter (LeuT) obtained in three functional states, outward-facing (OF), occluded (OC), and inward-facing (IF) states, have validated the NSS transport process is by an alternating access mechanism, in which the substrate and ions first access their central binding sites via an open extracellular vestibule, and then are released within the cell through the sequential closure of an extracellular gate and opening of an intracellular exit pathway. 17, Historically, LeuT has served as a structural template to study monoamine transporters and based on in-depth studies of bacterial transporters, including electron paramagnetic resonance (EPR) spectroscopy, molecular modeling, 32,33 and single-molecule fluorescence resonance energy transfer (smFRET) experiments, 28,34 substrate permeation through the NSS family transporters is facilitated by reorientation of helices around the central axis, in particular the movement of TM1a away from the helical bundle to open an intracellular vestibule for substrate release. 21,35,36 Despite low sequence similarity with human NSS transporters, these efforts paved the way for rational drug design for treating various psychiatric disorders. Structural investigations into human NSS transporters have further benefited from the more recent resolution of outward-facing conformations of eukaryotic monoamine transporters Drosophila DAT (d DAT) and human SERT (hSERT). The screening and docking studies using these crystal structures provide the structural basis of antidepressant recognition and inhibition. Most recently, cryogenic electron microscopy (cryo-EM) structures of hSERT complexed with the psychedelic non-competitive inhibitor ibogaine reveal the occluded and inward-facing states with similar structural arrangements as seen in LeuT. 17,23,50,51 However, given the structural discrepancies between SERT and other NSS structural models, the molecular basis of transitions between the intermediate states remains unknown. Closure of the extracellular vestibule is coordinated by helix motions of TM1b and TM6a where Arg79 and Glu493 are proposed to serve as extracellular gating residues to stabilize the OC and IF states. The helix orientation of TM1b in the SERT OC conformation is closely aligned to that of OF LeuT. Moreover, among the current SERT OF structures, the distance between the guanidinium group of Arg79 and carboxyl of Glu493 varies from 4.4 to 7.4 , while in the OC and IF states, this distance is 7.2 and 4.6 , respectively. As a result, the role of these gating residues and their interactions during conformational transitions is unclear. The N-terminal loop preceding TM1a and its interactions with TM6 and TM8 regulates the helix motion of TM1a during substrate release and acts as an intracellular gate. 36,52,53 Hydrogen-deuterium exchange (HDX) experiments have provided an alternative approach to understand the conformational dynamics within the NSS family and have shown that ion-substrate binding facilitates changes in dynamics in TM1a, TM6, and TM7. 24, Intricate loop dynamics, specifically motions of extracellular loops (EL) 3 and 4 fluctuates significantly during substrate transport. 24, The combined structural and biochemical studies have provided invaluable insights in the functional mechanism of the NSS family. However, the realistic motions of structural transitions at atomistic resolution are not fully known to understand the conformational driven substrate transport cycle. In this study, we performed unbiased all-atom molecular dynamics (MD) simulations to obtain a comprehensive understanding of the import mechanism for the physiological substrate serotonin in hSERT. Our study shows the key binding and transport events, including substrate interactions at an extracellular allosteric site, neurotransmitter binding within subsite B, coordination of three metal ions, and a single symported sodium ion being displaced into the cytosol by the movement of serotonin into the exit pathway. Using a Markov state model (MSM)-based adaptive sampling approach to explore the conformational landscape, we report a sequential order of the ion-substrate binding and transport processes for any NSS family transporter. The free energy conformational landscape plots reveal that structural isomerization from OC to IF is a rate-limiting step for import that is facilitated by the presence of 5HT in the orthosteric site. We identified a third sodium ion binding site in a buried cavity close to the orthosteric site which helps maintain the fold for substrate transport. We determined the key residues that are involved in 5HT recognition, binding, and transport, and these residues show to have some role in transport using site-directed mutagenesis. Our results provide an in-depth perspective into the molecular recognition and transport of 5HT in SERT and may aid for the development of conformational selective inhibitors for the treatment of psychiatric disorders. ## Results Substrate binding decreases the free energy barrier for SERT conformational transitions to the IF state. To understand the effects of substrate-induced protein dynamics, the entire import process of 5HT was studied using molecular dynamics (MD) simulations. A Markov state model (MSM)-based adaptive sampling approach was used to explore the entire accessible conformational space of the SERT. Simulations were initiated from the OF crystal structure of hSERT (PDB: 5I73) and a total of 130 µs of 5HT-free SERT (referred to as Na + -SERT) was obtained. Na + -bound SERT in an OF conformation obtained from Na + -SERT simulations, with the Na1 and Na2 sites occupied, was used to seed simulations of the 5HT import process (referred to as 5HT-SERT). 100 mM 5HT was added to the simulation box (equivalent to 12 5HT molecules) and a total of 210 µs data were collected. All simulation data were used to construct an MSM, which parses the simulation data into kinetically relevant states and calculates the transition probabilities between the states (See Methods for additional details). MSM-weighted simulation data were projected onto a coordinate system defined by distances between extracellular and intracellular gating residues (Figure 1 and S1). The conformational landscape plots reveal that despite the absence of 5HT binding, Na + -SERT may undergo transitions from the OF state to the IF state (Figure 1A). Extracellular gating residues Arg104 (TM1b) and Glu493 (TM10) can separate to 10 , enlarging the extracellular permeation pathway. The equivalent charged residues in the bacterial transporter LeuT (Arg20 and Asp404) have been previously implicated in the gating mechanism. 17,62 The OF states are stable, with a relative free energy of ∼1.5 kcal/mol. The distance between gating residues Arg104-Glu493 decreases to 3 and is associated with electrostatic interactions (Figure 1 states. An hourglass-like (HG) state, in which both gates are open, was also observed. The SERT structure is represented as cartoon with TM 1, 5, 6, 8, and 10 colored in teal, green, magenta, yellow, and orange respectively. (C) Cross-section through SERT conformational states viewed from the membrane plane, shown as surface representations. The channel pore volume across the transporter is depicted as dark blue spheres and extracellular gates Arg104 and Glu493 are shown as teal and orange sticks, respectively. calculated from MD simulation. 28,29, Distance distribution plots reveal that the extracellular surface remains open while the intracellular vestibule is closed in OF and OC states, and vice versa in the IF state (Figure S3). Closure of the extracellular pathway as SERT isomerizes from the OF to OC state weakens contacts on the intracellular side of the transporter, creating an energetically accessible pathway towards a partial IF state, in which the intracellular pathway measured by the opening of TM1a and TM5 extends to ∼6 . The free energy barrier for transition from the OC-IF state in Na + -SERT is estimated as ∼2 kcal/mol, which is higher compared to the OF-OC transition (∼1 kcal/mol). Biophysical investigations of LeuT also reveal that IF states are less populated and that transitions to IF states are facilitated by the presence of the substrate. 28,29,31,34 Formation of the IF state is associated with the partial unwinding of the cytoplasmic base of TM5, the breakage of electrostatic interactions between Arg79 (N-Term) and Asp452 (TM8) and Glu80 (N-Term) and Lys275 (TM5) at the intracellular gate, and increased dynamics of the flanking loops (Figure S2). Simulated helix rearrangements involved in opening and closing of the transporter agree with the recent cryo-EM structures 50 and other NSS crystal structures 23,44,46 (Figure S4). The substrate-present conformational landscape plot exhibits deviations in the relative free energies of conformational states and reduced free energy barriers between states (Figure 1B). Binding of 5HT in the entrance pathway stabilizes the OF states to a greater extent compared to Na + -SERT simulations (Figure 1, S3C). 28 The gating residues form alternative interactions with Gln332 (TM6) and Lys490 (TM10), thereby widening the extracellular vestibule (Figure S2). Similar to observations seen in human DAT simulations, 40 the diffusion of 5HT to the orthosteric (S1) site via the allosteric (S2) site promotes the inward closure of extracellular gating helices TM1b, TM6a and TM10 to facilitate formation of the OC state (Figure S5). The OF-OC transition has a free energy barrier of ∼1.5kcal/mol, similar to Na + -SERT. The 'downward' movement of 5HT facilitates opening of the intracellular gate and isomerization to the IF state. The free energy barrier for the OC-IF structural transition is estimated as ∼2 kcal/mol. The presence of 5HT in the intracellular pathway stabilizes SERT in a greater IF state, with a relative free energy of ∼1 kcal/mol as compared to ∼2 kcal/mol in Na + -SERT. Compared to Na + -SERT, the intracellular vestibule in the IF state now extends to 7-10 , thus allowing for the substrate to be released and is associated with the breakage of hydrogen bonding networks between intracellular gating helices (Figure S2, S6, S7, S8, S9). Terry et al. has demonstrated that the selective substrate-driven conformational transition to the IF state in LeuT is highly favorable compared to nonselective substrates and ion binding alone. 28 Using Transition path theory, we estimated the mean first passage time (MFPT) for different transitions observed in the simulations. We found that the rate of transition between OC-IF states is rate-limiting as compared to transitions between OF-OC states both in the presence and absence of 5HT (Figure S10). We also observed partial OF-IF like conformations, which we have termed as an hourglass-like (HG) state in which both gates are open but constricted at the center (Figure S11C). Terry ions and substrate molecules through HG states is a more energetically demanding process than the traditional substrate transport via the IF state. In Na + -SERT simulations, we observed uncoupled ion leaks, in contrast, Na + binds tightly to its respective site and no ion leaks were observed in 5HT-SERT (Figure S12). This state has been observed in other membrane transporters including a disease-associated mutant of DAT, 69 tions compared to Na + simulations (Figure 2). Hydrogen deuterium exchange (HDX) mass spectroscopy studies hint that EL2 and EL4 regions are destabilized and show increased deuterium exchange upon ion and substrate binding. 24,54,56,68 Furthermore, EL2 exhibits higher deuterium uptake kinetics in the OF state compared to the IF state in LeuT. 24 In 5HT-SERT simulations, EL4 shows less pronounced fluctuations during OC to IF transitions. The experimental results from HDX show EL4 regions are more stabilized during K + uptake which is hypothesized to stabilize the IF state of SERT based on LeuT studies. 56,70 The increased deuterium uptake of TM1a in the presence of K + agrees with the large fluctuations we observed for transitions to the IF state. The comparison of calculated deuterium exchange fraction of Na + and 5HT-SERT simulation data agree with previous HDX studies (Figure S13). 24 Identification of a new sodium ion binding site in a buried cavity stabilizes the fold for substrate transport. Monoamine transporters utilize an electrochemical gradient to transport substrates across the cellular membrane. 5HT-mediated transport involves the symport of 1 Na + ion and antiport of 1 K + ion. 7,71 Upon the transition of 5HT from the allosteric to the orthosteric binding site, the Na + ion in the Na1 site shifts to a third metal coordination center, which we call the Na3 site to be consistent with prior nomenclature. 72,73 The vacant Na1 site is then filled by two to three water molecules that interact with the charged site until another Na + ion binds to the Na1 site (Figure S14). The calculated electrostatic potential map of SERT OF state cryo-EM structure reveals that the permeation pathway is negatively charged, which allows Na + ions to rapidly diffuse and bind to the Na1 site (Figure S15). At the Na3 site, the Na + ion is coordinated by the carboxylates of Glu136 and Glu508, and the sulfur of Met135 through water molecules (Figure 3). A third metal ion site has not previously been described in SERT, but computational modeling, biochemical analysis, and electrophysiology recordings indicate that equivalent residues of the neuronal GlyT2 transporter also form a third Na + site. 73 The simulation reveals that the presence of Na + in the Na3 site stabilizes TM6 unwinding and the proper orientations of residues in the orthosteric site. There are two pieces of experimental evidence for this in silico discovery. First, a structural alignment of SERT from the MD simulation with the crystal structure shows weak but discernible electron density, comparable to the density of surrounding side chains, near the modeled third Na + (Figure 3). Second, previous experiments have shown a reduction of serotonin transport when Glu136 is mutated, underscoring the role of this residue for appropriate conformational dynamics. 74 We note that the presence of buried glutamates within hydrophobic transmembrane regions is highly unusual; in this case, Glu136 hydrogen bonds to exposed backbone N-H groups to support the unwinding of TM6 near the central substrate binding site. We calculated the pKa of these residues with and without Na3 bound and found that they are not protonated under physiological conditions (Table S1). The calculated pKa values for SERT resolved structures and other NSS structures were also predicted to be in the range of 3-4, given the intracellular neuron pH is 7.17 75 (Table S1). Our simulation shows that the Na3 cavity is well solvated before and after Na3 binding, with an average of 4-5 and 7-8 water molecules with 5 of the Na3 site, respectively (Figure S16A, S16B). We also calculated the protein tunnels and channel-like pores that could facilitate the transport of small molecules and ions for the OF, OC and IF states of SERT. The tunnel plot reveals that the Na3 site is completely buried inside the transporter and is not exposed to the cytoplasmic half (Figure S16C). Therefore, from our simulations, we observed that Na3 is not released to the intracellular side (Figure S17E). Simulations reveal a sequential order of substrate binding and transport in SERT. An aspect of the current NSS transport model that remains unaddressed is the sequential order of substrate binding and transport events. Using transition path theory, the highest flux pathway for conformational change and 5HT import can be determined from the Markov state model and used to predict an ordered sequence of binding events and structural changes. SERT undergoes complete transitions to the IF state in the simulations, with permeation towards the intracellular side upon binding of substrates in the order following Na + , 5HT, and Clions (Figure 4). We describe each step of the import process in detail. The transport process begins with the binding of Na + to the Na1 site, followed by a second Na + binding the Na2 site (Figure S18, S19). These two sites are well-described in the SERT crystal structure and the literature. 23,44,46,76 Na + bound at the Na1 site couples activity between the ion and substrate binding sites, whereas computational studies of related transporters have indicated that Na + coordinated at the Na2 site dissociates during the transporter cycle to become the symported metal ion. 40,77,78 The importance of the Na2 site is underscored by its conservation in distantly related secondary transporters. 79 Na + ions entering the transporter interact with Asp328 and Asn112 at the extracellular surface, then rapidly diffuse into the allosteric site (Figure S18A). Here, the Na + ions interact with Glu493 and Glu494, and a rotameric shift in Glu493 enables the ions to descend past the extracellular gate to their central binding sites (Figure S18B, S18C). Na1 is stabilized by Figure 4: The major flux pathway and mean first passage times for SERT conformational transitions and 5HT import determined from transition path theory. The transport process begins with the binding of 2 Na + ions to the Na1 and Na2 sites in the OF state (2, 3). Substrate diffusion to the orthosteric site shifts a Na + to the Na3 site ( 4). An additional Na + and Clion bind (5), facilitating closure of the extracellular gate to form the OC state (6). Isomerization to the IF state is associated with the release of Na + from the Na2 site and 5HT diffuses out (6-9). Arrow thickness represents relative flux between transitions. Asn101, Ser336, and Asn368, while Na2 is coordinated by backbone carbonyls of Gly94, Val97, Leu434, and side chains of Asp437 and Ser438. Previous simulations of SERT have shown Asp437 as a key residue for maintaining Na2-ion interactions, 78 and mutations of Asp437 and Ser438 have confirmed their role in coordination and dissociation of Na2. 78,80 Similar to other transporters in the NSS family, 29,63,64 the binding of Na + ions to their respective sites stabilizes SERT in the OF state while neutralizing the polar cavity to allow protonated 5HT diffusion. 5HT is recognized by Tyr107, Ile108, Gln111, and Asp328 at the extracellular vestibule to initiate the binding in the OF state. Ile108 forms hydrophobic contacts with the indole ring of 5HT while other residues form polar interactions with the substrate that favors binding. 5HT then diffuses inside the translocation pore and binds to the allosteric site (Figure 5B). The substrate is stabilized by aromatic ring packing against Phe335 and Arg104, and a hydrogen bonding network with Asp328, Gln332, and Glu494 (Figure 5B). Previous mutations of residues in the allosteric site have been shown to alter inhibitor potency. 81 5HT undergoes a 90 • rotation by rapidly exchanging its polar interactions and shifts towards the orthosteric site 5C). The switching of amine group interactions to Glu494 triggers the movement of 5HT from the allosteric site to the orthosteric site. 5HT rotates such that the conformation becomes perpendicular to the membrane which is further favored by polar and hydrophobic interactions by Asp328, Gln332, Leu502, and Ala331. The extracellular gating residues form a salt bridge interaction and enlarges the binding cavity such that substrate can escape to the primary binding site. This substrate binding site in the NSS family contains three well-studies subsites and has served as the basis of designing various trycyclic antidepressant molecules. 82 In the orthosteric site, the protonated amine moiety of 5HT forms charged interactions with Asp98 of subsite A, disrupting the hydrogen bonding interactions between Asp98 and Tyr176 (Figure S20). Also within subsite A, the phenol moiety forms aromatic interactions with Tyr95 and Phe341. Site-directed mutagenesis and computational docking studies have emphasized the significance of Asp98 in 5HT recognition and transport, 83,84 while biochemical studies show that the disruption of aromatic interactions with the substrate leads to a loss of function or decreased potency of antidepressants. 81, The binding of 5HT from the allosteric to the orthosteric binding site promotes Na + in the Na1 site to migrate to the third metal coordination center, the Na3 site. 5HT mediates the binding of an electrogenic Clion. Cland Na + permeate into the extracellular vestibule and bind at the Cland vacated Na1 coordination sites, which leads to the formation of an occluded conformation (Figure S21). Na + is stabilized by Glu493 and Glu494 while Clforms polar interactions with Arg104 and Tyr176. The calculated electrostatic potential map of the SERT-OF cryo-EM structure reveals that the permeation pathway is negatively charged and supports our finding that the binding of Clion will be accompanied by a Na + ion (Figure S15). The additional interaction of Clwith the indole-NH of 5HT further stabilizes the ion in the exposed extracellular recognition site (Figure 5D). The indole ring of 5HT occupies subsite C of the primary binding site, which in turn favors the transition of the Clion to its binding site (Figure S22). As Clenters, it shifts the guanidinium group of Arg104, facilitating the diffusion of both the Cland Na + ions into the central cavity (Figure S20C). The movement of Clthrough the transporter is supported by a network of interactions with Arg104, Tyr176, and the indole-NH of 5HT (Figure 5D). A shift of the Arg104 side chain exposes Gln332 for making contacts with the Clion, facilitating its migration to the Clbinding site. The predicted Na1, Na2, and Clsites in the simulations concur with their respective sites observed in crystal structures of SERT, LeuT, and DAT 23,44,46,76 (Figure S23). Interestingly, we did not observe Clto bind in the absence of the substrate within the given simulation timescales (Figure S24). In order to quantify the binding of Clin the absence of 5HT, we performed umbrella sampling simulations to compute the free energy of Clbinding in Na + -SERT conditions. The free energy barrier for Clbinding without 5HT was estimated on the order of ∼12 kcal/mol as compared to ∼4 kcal/mol in 5HT-SERT simulations (Figure S25). The highest energy barrier observed at approximately 3 away from the Clbinding site is associated with the dissociation of Na + from Clto allow binding to their respective sites suggesting that the substrate in the orthosteric binding site neutralizes the polar cavity to allow for favorable Clbinding. The binding of ions and the substrate to their respective sites leads to the closure of the extracellular gates to obtain the OC state where the permeation pathway is closed at both ends. The decrease of the pore results in shifting of the aromatic ring of 5HT from subsite C to subsite B within the orthosteric binding pocket. The amine moiety of the neurotransmitter remains bound to Asp98, Tyr95, and the C-terminal pole of TM1a within subsite A (Figure 6A,B). The simulated configuration of 5HT in subsite B agrees with the crystal structure of dopamine-bound d DAT 76 (Figure S23F). The indole-NH ring of 5HT interacts with Thr473 and other hydrophobic contacts by Ala169, Ala441, Gly442, and Leu443 residues stabilizing the substrate in subsite B. Additionally, 5HT may enter into subsite B in which the indole-NH is oriented towards Phe341 adopting a binding pose also reported in previous modeling studies. 84,88 Previous experimental studies indicate that mutations of these residues can decrease 5HT transport 81,85 and the binding pose is further supported by several studies showing that interactions in subsite B are critical for inhibitor potency. 82,85,86 Mechanism of 5HT translocation down the exit pathway. The conformational free energy landscape suggests that structural isomerization to the IF state is limited by a large free energy barrier which is decreased in the presence of the substrate. The prolonged binding of 5HT in subsite B weakens Na + interactions in the Na2 site and results in its dissociation. Na + loses its interaction with the backbone carbonyl of Leu434 and enters the intracellular vestibule, thereby initiating structural transitions from the OC to IF state. Koldsø et al. has shown that solvation of the intracellular vestibule allows Asp437 to rotate Na2 towards the intracellular pathway and facilitates transitions to the IF conformation. 78 Similarly, we also observed that the rotameric shift in Asp437 initiates the dissociation of Na + from the Na2 site (Figure S26A, S26B). The Na + ion then interacts with Ser91 and is further stabilized by Phe347 and Phe440 via cation-π interactions (Figure S26C). Finally, the Na + ion engages with Asp87 and diffuses into the intracellular space (Figure S26D). Koldsø et al. also predicted that Glu444 interacts with Na2 in the exit pathway,. 78 yet in our simulations, Glu444 forms an ionic interaction with Arg462 and does not interact with Na2. At this juncture, the intracellular gating residues (Arg79 and Asp452) still hold their hydrogen bond interactions, and Na + can access the intracellular pathway without the breakage of ionic contacts. However, in Na + -SERT simulations, the Na2 ion has a tendency to rebind to the Na2 site after release (Figure S27). Such intracellular binding events were not observed in 5HT-SERT simulations. The coupling of 5HT import to the cytoplasmic release of Na + from the Na2 site explains the 1:1 sodium to neurotransmitter stoichiometry of the transport cycle. The rotameric shift of Tyr95 results in permeation of 5HT to the exit pathway (Figure 6C). The flipping of the indole ring of 5HT displaces the ionic interactions with Asp98 resulting in the aromatic ring of 5HT to be trapped between Tyr95 and Val343. The 'downward' movement of 5HT disrupts contacts between TM1a and TM5, specifically the hydrogen bonding interactions between Asp87 and Trp282 (Figure S9). The distance between Asp87 and Trp282 increases up to 16 , thus drives the dissociation of 5HT to the cytoplasmic half of the transporter. Additionally, the intracellular salt bridge network between Glu80-Lys279 and Arg79-Asp452 weakens as the cytoplasmic base of TM5 unwinds, further resulting in the opening of the intracellular pathway. 5HT shifts to the intracellular vestibule and occupies the Na2 site (Figure 6C). The amine group forms strong polar contacts with residues in the Na2 site, and the indole ring is lodged between Tyr95, Phe347, and Phe440. 5HT further diffuses down, however the amine group of 5HT still forms interactions with Tyr95, and the indole-NH forms additional interactions with Ser91. Finally, 5HT leaves the transporter through a widened intracellular pathway surrounded by TM1a and TM5. Our results show that the rotation of Tyr95 propagates the opening of TM1a and mediates substrate transport to the cytoplasmic half. The equivalent residue in hDAT, Phe76, has been shown to undergo a similar rotameric transition to allow for substrate release. 40 We also observed that the cytoplasmic base of TM5 unwinds and shifts outward by ∼8-10 to facilitate cytoplasmic opening of the exit pathway (Figure 6D), and it is well known that these regions play a crucial role in regulating SERT activity 64 . Simulations identify key residues involved in substrate transport. SERT has been extensively studied by targeted mutagenesis, especially within the orthosteric binding site, confirming key interacting residues with substrate and drugs. 84,89 To provide basic evidence supporting aspects of the simulations, we instead focused mutagenesis to residues predicted to make early interactions with substrate (Figure 7A). Changes in transport activity were assessed based on cellular uptake of a fluorescent 5HT analogue, APP+, which has been shown to have similar transport properties to 5HT. 90 Furthermore, we extracted individual states from the 5HT-SERT simulation data and performed molecular docking experiments to predict the binding mode of APP+, and observed that APP+ binds in a similar fashion and shares similar residue interactions as the native substrate, 5HT (Figure S28). We examined by immunoblot that the decreased transport activity was not a result due to poor SERT expression, with the exception of Ile108Ala (Figure 7B). Alanine substitutions of Tyr107 and Gln111, which are predicted to participate in the early recognition of 5HT as it first diffuses into the extracellular vestibule and substitution of Asp328 in the allosteric site reduce substrate transport (Figure 7B). Both mutations are deleterious for transport activity as they mediate crucial interactions with the substrate in the initial recognition as well as in the allosteric site (Figure 7B and 7C). Leu502 is packed beneath aromatic residues in the allosteric and early recognition sites, and disruption of local structure by its substitution to alanine renders SERT inactive, although we cannot exclude broader effects on the protein fold (Figure 7D). We further tested alanine substitution of Val343, which is packed against Tyr95 at the base of the orthosteric site. Once Tyr95 undergoes a rotameric shift during the OC to IF transition, a void is created between Val343 and Tyr95 that is temporarily occupied by the substrate as it moves into the intracellular vestibule (Figure 7E). SERT Val343Ala has partly reduced activity, which is notable considering the close chemical similarity between alanine and valine side chains. Additionally, we determined the K M -V Max values for a mutant in each interaction site which show a reduction in transport activity (Table S2). Finally, Glu136 and Glu508 that coordinate Na + in the Na3 site were mutated to alanine to cause a complete loss of activity, demonstrating the importance of 392 this region for the transport process. ## Conclusions In this study, we present an atomistic view of the substrate import process in SERT as well as characterizing the thermodynamics of key states involved in substrate transport. By implementing an MSM-based adaptive sampling protocol to sample the conformational landscape, we investigated global transitions from OF to IF for both 5HT-free and 5HTtransporting SERT. MSM-weighted conformational free energy landscapes show the OF and OC states are relatively stable, and transitions to and from OF and OC states are relatively low energy. Transitions from OC to IF are substantially higher, with energy barriers of ∼2 kcal/mol in Na + -SERT; however, the presence of 5HT not only lowers the free energy barrier of OF to OC transitions, but further stabilizes SERT in the IF state as compared to Na + -SERT. The dependence of the IF state on the presence of substrate has experimentally been observed for the bacterial LeuT transporter, where addition of extracellular substrate promotes dynamics at the intracellular gate. 62 The simulated structures show similar helix orientations at the extracellular and intracellular gates with respect to experimental SERT structures. 23,44,46 The comparison of Na + and 5HT-SERT data reveal that the structural transitions from OF to OC involve minimal helix movements in TM1b, TM6a, and TM10, while OC to IF transitions show higher fluctuations of intracellular helix tips of TM1a, TM5, and TM7 to facilitate opening of the intracellular vestibule (Figure S4). Biochemical experiments in conjunction with computational modeling have shown that cholesterol binding to a conserved site comprised of TM1a, TM5, and TM7 inhibits these helix movements and may act as a regulatory mechanism for SERT function. 46,91,92 Deviations were observed in the opening of the intracellular vestibule of the IF state between the cryo-EM and predicted MD structure, specifically the outward motion of TM1a. This might reflect the loss of lateral pressure following detergent-extraction; the membrane is anticipated to constrain the extent to which TM1a can move away from the helical bundle. Nevertheless, our results show that partial opening of the intracellular pathway is sufficient for substrate transport. Experimental studies of LeuT also indicate that a partially-open IF conformation is suitable for substrate transport. 28 Conducting simulations of both serotonin-free and serotonin-present conditions allows us to examine the finer details of the substrate-induced conformational transitions and the cooperative nature between the transporter and the substrate. Motions of TM2 were highly correlated with TM1b in Na + -SERT, whereas the binding of 5HT decreases the cooperativity as TM1b rotates and closes the extracellular cavity to open the exit pathway (Figure S29). The conformation dynamics examined through HDX studies reveal that TM2 undergoes less fluctuations when bound to Na + , in contrast to when the transporter is substrate bound. 54 Similar trends were observed between TM6a and TM2, more specifically, the presence of 5HT destabilizes this region and favors the conformational change required for substrate transport (Figure S29). Gating residues Arg104-Glu493 remain open until the ions and 5HT bind to the primary binding site, and later close the extracellular cavity as the exit pathway opens to allow the release of substrate molecules (Figure S5). At the binding site, Tyr176 has the tendency to form a polar interaction with Asp98 in the OF and OC states. Interactions by the amine moiety of the substrate sever the Tyr176-Asp98 interaction, which is further disrupted when the opening of the intracellular cavity induces a rotameric shift in Tyr176 away from the Asp98 (Figure S19). Most significantly, the closure of the extracellular allosteric cavity is coupled to opening of the intracellular helix TM1a when the substrate is bound (Figure S29C). Experimental studies show that the substrate binding increases dynamics between EL4 and TM12. 24, In simulations, EL4 inserts deep inside the extracellular cavity, which is further facilitated by the movement of TM12 (Figure S30). Furthermore, HDX studies on DAT support the coupled motions between TM7 and TM12 observed in simulations when the substrate is bound, compared to just Na + binding (Figure S29). Our simulations reveal an ordered sequence of binding and transport events that agree with the 1:1 substrate:Na + stoichiometry as previously characterized for the NSS family. In 5HT-SERT simulations when Na1 and Na2 sites are occupied, both Clions and 5HT molecules were present in the solution, but even after extensive sampling, we only observed 5HT binding to the transporter. Similarly, in the Na + -SERT simulations, Na + and Clions were present in the solution, yet we observed rapid diffusion and binding of Na + ions to their binding sites instead of Clion binding. We further determined that the free energy barrier for Clto bind without the substrate in the orthosteric cavity is significantly greater than in 5HT-SERT simulations (Figure S25). Additionally, the calculated electrostatic potential map of the OF SERT cryo-EM structure reveals that the permeation pathway is negatively charged and supports our finding that the binding of Clprior to substrate binding will be associated with a large free energy barrier (Figure S15). We identified a third sodium ion binding site buried deep in the binding pocket that is conserved among various NSS transporters and critical for substrate transport. Similar to other NSS members, the release of Na + from the Na2 site decouples the interactions between TM1a, TM5, and TM8, increasing the solvation of the intracellular permeation pathway and allowing TM1a to open for substrate release. As unbiased molecular dynamics simulations can only sample events with low free energy barriers, we have the highest likelihood of observing the most probable binding sequence of ions and substrate. However, we do not disregard that different sequences of ion and substrate binding events with higher free energy barriers associated with them are possible and may alter the kinetics of the SERT transport cycle. Differences in stoichiometry may arise if the ion unbinding timescales for Na1, Na3, and Clare much longer than the simulated timescales obtained in this study. We investigated the mechanism of 5HT recognition, binding, and translocation in SERT and through mutagenesis provided support for certain residues having potential roles in substrate binding and/or conformational transitions, though we do not exclude that they may have other roles in folding of the transport cycle. In some simulations, we observed 5HT to simultaneously bind both the orthosteric and allosteric site. Substrate bound at the allosteric site hinders the fluctuations of several residues at the extracellular side, hence, allosterically communicates to the primary binding site of the transporter by limiting conformational transitions from the 5HT-bound OF state. The distance distribution plot reveals that the side chain conformation of Arg104 is restricted when 5HT binds to both orthosteric and allosteric sites, whereas when 5HT is solely occupied the orthosteric site, Arg104 may adopt multiple conformations (Figure S31). The binding of the substrate at both sites sterically restricts opening of the exit permeation and substrate release. This study has explored the conformational dynamics and substrate import of a monomeric unit of SERT; however, fluorescent microscopy has shown SERT to form functional oligomers in the membrane. 93,94 The LeuT crystal structure has been resolved as a dimeric unit, and while recent modeling studies have explored possible dimeric SERT units, 95 the exact interface of SERT oligomerization and the effects of coupled dynamics remains unclear. Further investigation is required into understanding how in vivo regulation affects the kinetics and conformational landscape of SERT. ## Methods Molecular dynamics (MD) simulations. The OF SERT crystal structure (PDB: 5I73) was used as the starting model for MD simulation. 44 Thermostable mutations in the crystal structure Ala110, Ala291, and Ser439 were reverted to wild type Tyr110, Ile291, and Thr439, respectively. The protein was embedded in a phosphatidylcholine (POPC) bilayer with CHARMM-GUI 96 and solvated with TIP3P water molecules. 97 150 mM NaCl was added to neutralize the system and mimic physiological conditions. Terminal chains were capped with acetyl and methyl amide groups. Overall, the final Na+-SERT system consisted of ∼70,000 atoms in a period box volume of 77 X 77 X 113 3 . The MD system was built using the tleap module of AMBER14. 98 The MD system was minimized for 20,000 steps using the conjugate gradient method, heated from 0 to 300 K at NVT, and equilibrated for 40 ns under NPT conditions. A Na + bound in the Na1 and Na2 site OF SERT structure, obtained from Na + -SERT simulations, was used as the starting model to capture the mechanism of 5HT import. 100 mM serotonin (equivalent to 12 5HT molecules was randomly added to the simulation box and equilibrated under the same conditions described previously. We note that the simulated 5HT concentration is unphysiological and was implemented to enhance the probability of substrate binding events. During the alternate-access cycle, the transporter only interacts with at most two 5HT molecules. Therefore, the free energy and kinetics of the conformational change processes are independent of the additional 5HT molecules in solution. The ionic charge of 5HT is neutralized with Clions. While the ionic concentration could have some effects on the protein conformation such as altering the structure of the extracellular loops which directly interact with the additional 5HT molecules, we do not observe any shifts in the structure of these loops between the Na + -SERT and 5HT-SERT simulations. All simulations were implemented using Am-ber14 package employing Amber ff14SB 102 force field combined with GAFF force field at constant NPT conditions (300K, 1 atm) and periodic boundary conditions. Temperature and pressure were maintained with Berendsen thermostat and barostat, respectively. 103 Electrostatic interactions were treated with the Partial Mesh Ewald method, 104 and hydrogen bonds were constrained using SHAKE algorithm. 105 Nonbonded distance cutoff was set at 10 , and an integration timestep of 2 fs was used for all simulations. Snapshots were saved every 100 ps during production simulations. ## Adaptive sampling. Obtaining sufficient sampling is a reoccurring challenge in simulating complex biological processes. To overcome this issue, we adopted a Markov state model (MSM)-based adaptive sampling methodology to efficiently explore the conformational landscape. 57,106,107 In each round of adaptive sampling, multiple short MD simulations are conducted in parallel. The simulation data is clustered using the K-means algorithm 108 based on a designated metric and starting structures are chosen from the least populated states to seed the subsequent rounds of simulation. The sampling bias introduced from least-count selection is eliminated during the construction of the MSM by estimating the reverse transition probability matrix for transition between all conformational states. In the limit of long timescales, the sampling errors are expected to be small. We show that there is little difference between the population of states derived from the raw data and the MSM-derived equilibrium population on the conformational landscape (Figure S32). Extracellular and intracellular gating residues were used as adaptive sampling metrics for the conformational sampling of the states and zposition of substrates to capture the import process. A total of ∼130 µs of Na + -and ∼210 µs 5HT transport simulation data was obtained. Multiple substrate binding and transport events were captured as a direct advantage of the MSM-based adaptive sampling approach (Figure S33). The entire MD dataset was used for MSM construction and analysis. ## MSM construction. The CPPTRAJ and pytraj 109 modules in AmberTools and MDTraj Python library 110 were used for post processing the trajectory data. Markov state models (MSMs) were constructed using pyEMMA 2.5.6 Python package. 111 MSMs were constructed for both Na + -and 5HT transport datasets. 15 residue-residue pair distances surrounding the permeation pathway for clustering (Figure S34). Additionally, the z-components of the 2 Na + ions were incorporated to the 15 residue-residue pair distances as featurization metrics for Na + -SERT simulations. The z-components of 5HT, Cl -, and the symported Na + ion were added along with the 15 distances for the 5HT transport process. Time-structure independent component analysis (tICA) was performed on the feature matrix to reduce the dimension space by obtaining the slowest-relaxing degrees of freedom as a linear combination of the features. 112 The optimum number of clusters and time-independent components (tICs) were obtained from the set that yielded the greatest sum of the eigenvalues of the transition matrix, also known as the VAMP1 score. 700 clusters and 8 tICs were used to construct the MSM for Na + -SERT simulations. 700 clusters and 10 tICs were used for 5HT-SERT MSM. (Figure S35A). The lag time of 8 ns was determined for MSM construction from implied timescale plots (Figure S35B). The Chapman-Kolmogorov test, which validates the Markovian behavior of the MSM, 113 was performed on 5 macrostates implemented in pyEMMA (Figure S36, S37). ## Trajectory analysis. MSM-weighted simulation data were plotted on the coordinates of the gating distances, specifically the distances between the closest heavy atom between Arg104 and Glu493 of the extracellular gate, and the closest heavy atom between groups of residues Gly77-Ser91 of the N-terminus and TM1a and Lys297-Trp282 of TM5 for the intracellular gate (Figure S1). MSM states were further clustered into macrostates and visualized with Visual Molecular Dynamics (VMD) 114 and PyMOL (Schrödinger, LLC). The predicted fraction deuteration was calculated according to Adhikary et al. 24 on 1000 OF and IF structures randomly extracted from the Na + -SERT and 5HT-SERT MSM. The 5HT-residue interactions were obtained from python scripts implemented from the GetContacts package (https://getcontacts.github.io/). To calculate the RMSF between transitions, 1000 structures of each OF, OC, and IF states were randomly extracted from the MSM and measured with respect to the cryo-EM structure of the prior conformational state (i.e. OF-OC: 1000 OC structures with respect to OF cryo-EM structure). In-house scripts and matplotlib Python library were used to generate plots. Channel pore radius was calculated using the HOLE program. 115 pKa values were calculated using DelPhiPKa program as it also considers the effects of solute molecules while preforming the calculation in the static structure. 116 Cross correlation analysis of Na + -SERT and 5HT-SERT was performed to identify the allosteric effects of substrate on conformational changes between various states and the substrate transport. The cross-correlation value of 0 represents motions that are not correlated, while a value of 1 shows that the motions are highly coupled. ## Transition path theory. The top flux pathways for conformational changes and 5HT import were determined using transition path theory (TPT) analysis. Given a transition matrix obtained from the Markov state model, TPT examines the transition probabilities and estimates transitions pathways connecting the source (defined as state A) and sink (defined as state B) states and the fluxes associated with the pathway. 117 The flux between state i to j or f ij is defined as where π i is the stationary probability of state i as obtained from the eigenvalue decomposition of the transition matrix, T ij is the probability of transitioning from state i to state j, q − i is the backwards committer probability of transitioning to the source state rather than the sink state from state i, and q + j is the forwards committer probability of transitioning to the sink state rather than the source state from state j. 118 The forward committer probability for a state i, or q + i is computed as the sum of all the transitions from state i the sink state B and the sum of all the transitions from state i to sink state B via an intermediate state I, 118 or The backwards probability committer probability is calculated as TPT analysis was conducted using the pyEMMA 2.5.6 Python package. 111 Umbrella sampling of Clbinding in Na + -SERT Umbrella sampling MD simulations were used to calculate the binding free energy of Cl to the Clsite without the presence of the substrate. The Clbinding reference pathway was obtained from 5HT-SERT simulations. Using the distance of Clto the binding site as the collective variable, 73 equally spaced umbrella sampling windows from 0 to 18 at 0.25 intervals. At each window, the Clion was restrained by a harmonic potential with spring constant 10 kcal/mol-2 . The umbrella sampling simulations were implemented using the Amber18 package. 119 Each window was equilibrated for 10 ns and followed by 25 ns production. The free energy was constructed using the weighted histogram analysis (WHAM) software 120 and averaged over 5 independent runs. ## Plasmid construction. Human SERT was PCR amplified from pcDNA3-hSERT (Addgene 15483) 121 to introduce a consensus Kozak sequence and ligated into the Kpn1-Xho1 sites of pCEP4 (Invitrogen). Alanine substitutions were introduced by overlap extension PCR, and all plasmid sequences were verified. Transport Assay. Expi293F cells (ThermoFisher) were cultured in Expi293F expression medium (ThermoFisher) at 37 • C, 8% CO 2 , 125 rpm, and transfected with 500 ng plasmid per ml at a density of 2 x 10 6 cells/ml using Expifectamine (ThermoFisher). 24-28 hr post-transfection, cells were washed with PBS-BSA (Dulbecco's phosphate-buffered saline supplemented with 0.2% bovine serum albumin), incubated with fluorescent substrate analogue APP+ (Aobious) in PBS-BSA for 2 min at room temperature, washed twice with ice cold PBS-BSA, and analyzed on a BD LSRII cytometer. The main cell population was gated by forward and side scatter properties to exclude debris and doublets, and 10,000 gated events were collected. Cell fluorescence was measured with a 488nm laser, 505 LP dichroic mirror, and 530/30 bandpass filter. Fluorescence corresponding to non-specific uptake by vector-transfected cells was subtracted, and activity of mutants was assessed relative to the wild type construct. ## Immunoblot. Cells were collected 24 hr post-transfection, solubilized in SDS loading dye, and proteins were separated by SDS-polyacrylamide gel electrophoresis. After transferring proteins to polyvinylidene fluoride membrane (Bio-rad) at 4 • C, the membrane was blocked for 1 hr at room temperature with tris-buffered saline containing 0.1% Tween 20 (TBST) and 3% skim milk powder. The membrane was incubated with 1:1,000 anti-SERT rabbit polyclonal (Abcam Cat. No. ab102048) in TBST containing 1% milk overnight at 4 • C, washed 3 times with TBST for 5 min, and incubated with 1:10,000 HRP conjugated goat anti-rabbit IgG (Jackson ImmunoResearch Laboratories) in TBST containing 1% milk for 1 hr at room temperature. The membrane was washed three times with TBST for 5 min, and developed using clarity western ECL substrate (Bio-rad). The signal was visualized using Gel Doc (Bio-rad). ## Molecular Docking. Docking was performed using AutoDock 4.2. 122 The PDBQT format files for the SERT structures and APP+ was obtained using AutoDock python utility modules. The grid center was chosen by selecting the residues within 6 of the ligand and grid maps were generated using AutoGrid. The ligands were docked into the active site using AutoDock. The default Lamarkian Genetic algorithm was used for performing energy evaluation of the ligand conformations. 123 The docked poses were visualized in PyMOL (Schrödinger, LLC). The

chemsum

{"title": "The Substrate Import Mechanism of the Human Serotonin Transporter", "journal": "ChemRxiv"}

foldable_semi-ladder_polymers:_novel_aggregation_behavior_and_high-performance_solution-processed_or

## Abstract: A critical issue in developing high-performance organic light-emitting transistors (OLETs) is to balance the trade-off between charge transport and light emission in a semiconducting material. Although traditional materials for organic light-emitting diodes (OLEDs) or organic field-effect transistors (OFETs) have shown modest performance in OLET devices, design strategies towards high-performance OLET materials and the crucial structure-performance relationship remain unclear. Our research effort in developing cross-conjugated weak acceptor-weak donor copolymers for luminescent properties lead us to an unintentional discovery that these copolymers form coiled foldamers with intramolecular Haggregation, leading to their exceptional OLET properties. An impressive external quantum efficiency (EQE) of 6.9% in solution-processed multi-layer OLET devices was achieved. ## Introduction The past two decades have witnessed great research efforts in the interconversion of light and electricity in the area of organic conjugated polymers. The conversion of photons to electrons takes place in photovoltaic devices and photodetectors, which have been actively pursued. 1,2 The reverse process, converting electrons to photons, occurs in organic light-emitting diodes (OLED), 3,4 which have been commercialized and are now widely used in lighting and display applications. Accompanying the development of OLEDs, organic light-emitting transistors (OLET) emerged as a new class of organic optoelectronic devices that combine both the electrical switching functionality of organic feld-effect transistors (OFETs) and the light-generation capability of OLEDs in a single device. The OLETs, therefore, offer the potential for simplifying circuit design in the electroluminescent displays, electrically pumped organic lasers, and digital displays. However, the requirements of organic semiconductors for OLET based applications are more stringent than those of OLED active materials. They include balanced high ambipolar mobility and high photoluminescent quantum yield (PLQY) simultaneously in the same material, which are usually not compatible and difficult to realize. Current OLET devices are based on the traditional fluorescent semiconductors already used in OLEDs or OFETs. Their performances are relatively poor as they do not satisfy the stringent requirements as mentioned above. 6,7 To address this issue, multi-layer OLET devices that delegate different functions such as charge transport, charge injection, and emission into different materials are being developed. In 2010, Muccini et al. used p-type small-molecule semiconductor, 5,5 000 dihexyl-2,2 0 :5 0 ,2 00 :5 00 ,2 000 -quaterthiophene (DH-4T) and n-type fluorine-substituted DFH-4T as transporting layers, and host tris(8-hydroxyquinolinato)aluminum (Alq 3 ) and guest 4-(dicyanomethylene)-2-methyl-6-(p-dimethylaminostyryl)-4H-pyran (DCM) as emissive layer. 19 This tri-layered OLET device showed a good match of energy levels and balanced charge mobility, which resulted in external quantum efficiency (EQE) as high as 5%, nearly 100 times higher than the corresponding OLED. More recently, Meng et al. used thermally activated delayed fluorescence small-molecule semiconductor and high-k polymer-based dielectric layer to construct a trilayered OLET through vacuum-deposition in which an impressive EQE of 9.01% was obtained. 21 However, the fabrication of devices mentioned above require sophisticated vacuum deposition of multiple layers of smallmolecule materials and is not compatible with polymeric materials and the modern printing electronics industry. Multiple layers formed in solution-processed tri-layered OLETs are obtained by spin-coating, in which materials used must be soluble in orthogonal solvents to avoid re-dissolution. This requirement signifcantly limits the choice of available materials and thus the performance of resulting OLETs is relatively poor (EQE < 1%). 18 Detailed studies lead us to realize that new material systems for high-performance OLETs require not only suitable energy levels, luminescent quantum yields and charge mobility, but also the correct aggregation state. In this paper, we describe a semi-ladder polymer system aimed at addressing these issues and obtaining efficient solutionprocessed multi-layered OLETs. These polymers are designed based on the idea that ring fusion in ladder building block can enhance rigidity in the molecular system, which will minimize the non-radiative decay and thus improve PLQY. Detailed studies demonstrated that semi-ladder polymers forming Haggregated and folded structures can balance the PLQY and charge transport. The H-aggregation exhibit limited p-p interaction between chromophores, yet enough to achieve effective charge transport. The resulting OLETs thus outperform those fabricated from traditional linear conjugated polymers. ## Synthesis, and chemical properties The semi-ladder polymers were synthesized via Suzuki coupling polymerization of electron-accepting monomer 5,11-bis(2butyloctyl)-dihydrothieno[2 0 ,3 0 :4,5]pyrido[2,3-g]thieno[3,2-c] quinoline-4,10-dione (TPTQ) or TPTQF dibromide with electron-donating chromophore carbazole (C) monomer containing bis(pinacolato)di-boron (BPin) moieties. The resulting polymers are cross-conjugated and exhibit excellent fluorescent properties (Fig. 1a). General synthetic procedures of the polymers are summarized in the ESI. † These polymers exhibited sizable molecular weights and generally narrow polydispersity indices (PDI) as summarized in Table S1. † TPTQ-C and TPTQF-C were soluble in common organic solvents such as p-xylene or chlorobenzene. The HOMO and LUMO energy levels of the polymers were calculated from the oxidation onset of cyclic voltammetry (CV) measurements and optical bandgap of thin flms (Table S1 †). The replacement of thiophene on TPTQ with furan (TPTQF) leads to slightly higher HOMO and LUMO energy levels and larger bandgaps. TPTQF-C and TPTQ-C exhibit E HOMO /E LUMO of 5.42/3.04 eV, and 5.44/3.19 eV, respectively. These energy levels are consistent with the energy levels calculated from the density functional theory (DFT) (B3LYP method, 6-31g** basis set) as shown in Fig. S1. † Notably, TPTQ-C exhibited a lower PLQY (30%) than TPTQF-C (50%) which may be attributed to the heavy atom effect (sulfur versus oxygen) and will play an important role on device performance. For comparison, linear semi-ladder polymers, TPTQF-CC and TPTQ-CC were also synthesized as shown in Scheme S1 † and their chemical properties were summarized in Table S1 and Fig. S2. † The optimized geometry obtained from DFT calculations indicated that the cross-conjugated connection makes both TPTQF-C and TPTQ-C coiled with sizes of cross-section around 24.1 /29.5 and 21.1 /26.6 respectively (Fig. 1b). ## Optical properties and aggregations The optical transitions in these cross conjugated polymers, monomers, and model compounds (carbazole-TPTQ(TPTQF)carbazole) were investigated in detail by employing UV-vis spectrometer and the results are shown in Fig. 2a and S3. † The absorption spectra of the TPTQ monomer and model compound exhibited a strong 0-0 transition and weaker 0-1 transition. However, polymers TPTQ-C and TPTQF-C show a signifcant difference in spectral shape, where the 0-0 transition intensity is reduced and 0-1 transition becomes the strongest, indicating the formation of H-aggregates. 22,23,33 Normalized absorption spectra showed almost no change in the spectral shape with decreasing concentration (Fig. 2), indicating that H aggregation exists even at the level of a single polymer chain. This is an evidence for polymer folding as shown in our previous work. 34,35 In comparison, 0-0 transition intensity was reduced gradually and the whole absorption spectrum was blue shifted with increasing concentrations, as shown in linear polymer, TPTQ-CC (Fig. S4 †), which exhibited unfolded H-aggregation. Fluorescence spectra under varied temperatures provided further evidence for polymer aggregation formation. It is known that the ratio of intensities of the I 0-0 peak to the I 0-1 peak will increase as the temperature increases for Haggregation, and decrease for J-aggregation. 23 As shown in Fig. S5 and Table S2, † TPTQF-C exhibited an increased I 0-0 /I 0-1 ratio from 2.03 to 2.18 when the solution temperature increased from 9 to 20 C which is consistent with H-aggregation (due to limitations in our instrument, we could only perform the measurements within this narrow temperature range). Similar trends were observed for polymer TPTQ-C. It was proposed that the formation of H-aggregates in normal semiconducting polymers is due to strong inter-chain interactions. 23,33 The Haggregation in these polymers, however, must be due to intrachain folding as concluded from the spectral analysis above. Due to intrachain H-aggregation, these foldamers exhibit modest PLQY in dilute chloroform solution (0.001 mg mL 1 ). The direct evidence for folded structures came from small angle X-ray scattering (SAXS) measurements using advanced synchrotron light source. 36,37 The SAXS profles of TPTQF-C and TPTQ-C were obtained in THF solutions with a concentration of 5 mg mL 1 , which were used to analyze the structure of the foldamer. As shown in Fig. 3, the two polymers showed strong scattering intensity I(q) at small scattering vector (q < 0.3 1 ). After plotting the characteristic Kratky plots: q 2 I(q) vs. q, folded peaks were observed. Unlike the unfolded samples which have a plateau, the folded structures of TPTQF-C and TPTQ-C could be unambiguously identifed in Fig. 3b. 38 To calculate the particle size for the foldamers, the plots of ln[I(q)] vs. q 2 , were ftted with Guinier relationship: ln where I 0 is proportional to M w and R g is the size of the particle (Fig. 3c). 36,38 The calculated particle size of TPTQF-C (25.3 ) was relatively larger than that of TPTQ-C (24.3 ) which is consistent with the simulated coiled structures (Fig. 1b). The flm photoluminescence spectra for the polymers showed a slight redshift in comparison with corresponding solution spectra. To understand these observations, we measured concentration-dependent photoluminescence spectra (Fig. 2). The range of concentrations used in our study was from 0.001 mg mL 1 to 0.1 mg mL 1 . The polymers showed a gradual redshift of fluorescence upon concentration increase. The shoulder peaks (I 0-1 ) were present even in the most dilute solution for all the polymers and their intensity increased with the increasing concentration. These results seem to be in contradiction with H-aggregates which are known to exhibit a blueshift. However, as demonstrated by Spano et al., polymers containing quadrupole interactions could exhibit a redshift in H-aggregates. 24,39 As shown in Fig. 1a, the D-A -D + resonant structures indeed demonstrate a compound exhibiting quadrupole interactions. Thus, these polymers are special cases with quadrupole interactions that exhibit a redshift in H-aggregates. It is different from typical blue-shifted H-aggregate for small molecules, in which aggregates are mainly influenced by Fig. 3 (a) SAXS scattering intensity, I(q) of TPTQ-C (blue) and TPTQF-C (orange) versus scattering vector q in THF solution (5 mg mL 1 ). (b) Kratky plots, q 2 I(q) q showing the folded peaks in small q-range. (c) Fitting ln[I(q)] q 2 to Guinier relationship to calculate the radius of gyration, R g . (d) GIWAXS profile along q z (out-of-plane) and q y (inplane) in thin-films. This journal is © The Royal Society of Chemistry 2020 Chem. Sci., 2020, 11, 11315-11321 | 11317 intermolecular interaction, while in polymers, the aggregation states are contributed from both interchain and intrachain interaction. 23 To gain deeper insight into photophysical properties, timeresolved fluorescence decay measurements were performed with polymer solutions (Table S3 †). Fluorescence decay curves were ftted with exponential decay equation and fluorescence lifetimes were calculated. It was found that polymer TPTQ-C exhibits the fluorescence lifetime (s) of 2.14 ns with a single exponential decay curve. TPTQ-F showed double exponential decay behavior with s 1 (25%) ¼ 0.77 ns, and s 2 (75%) ¼ 2.44 ns, which may indicate the presence of different relaxation pathways in comparison with TPTQ-C. This value seemed to be consistent with folded H-aggregates in which exciton delocalization elongates the fluorescence lifetimes (Table S3 †). ## Microstructures and charge transporting properties As shown in the two-dimensional (2D) grazing-incidence wideangle X-ray scattering (GIWAXS) images (Fig. S6 †) and profles (Fig. 3d) of the polymer thin flms, it was found that the polymers were almost amorphous, with a slight preference face-on orientation. TPTQ-C and TPTQF-C exhibited similar intermolecular p-p stacking distances of 3.84 and 3.89 respectively. Moreover, TPTQF-C and TPTQ-C showed p-type transport behavior and hole mobilities (m h ) of 5.2 10 6 and 6.9 10 5 cm 2 V 1 s 1 respectively (Tables S4 and S5 †) in bottom gate top contact FET devices with gold as source and drain electrodes (Fig S9 and S10 †). The modest charge mobilities and amorphous characters in the thin flms, may be due to the tight intrachain folding. Expectedly, no electroluminescence was observed in single-layer FET devices, because of the large injection energy barrier (using the same drain-source electrodes) and low charge mobility in pristine flms. ## Fabrication of multi-layered OLET and device performance To address the issue about unbalanced charge injection, multilayers including an electron injection layer, a charge transporting layer, an emissive layer, and a self-assembled monolayer (SAM) were integrated as a device confguration of (Si 3 N 4 / OTS/DPP-DTT/emissive layer/PFN + BIm 4 /Au). SAM (n-octadecyltrichlorosilane, OTS) was vapor-deposited on SiN x as a modifcation layer at 120 C in a vacuum oven to reduce charge trapping and to improve molecular stacking. From the energy level diagram (Fig. 4), the LUMO energy levels of these polymers were aligned too high relative to Au (workfunction, W Au ¼ 5.1 eV) with the electron injection barrier as high as 2.0 eV. Therefore, a thin conjugated polyelectrolyte (CPE) PFN + BIm 4 with a thickness of around 10 nm was inserted between gold (Au) and the emissive layer as an electron injection layer. The ionic effect of PFN + BIm 4 effectively lowered the electron injection energy barrier. 18 Since the thin flm PFN + BIm 4 was spin-coated from a methanol solution, we were able to avoid the dissolution of the emissive layer. The low charge mobility of pristine polymer flms would impede the recombination of electron/hole pairs which dramatically decreases the electroluminescent efficiency. To address this issue, a charge transporting layer was inserted between the gate electrode and the emissive layer. After carefully testing different high mobility FET polymers, we found DPP-DTT suitable for our material (Fig. 4). DPP-DTT does not dissolve in p-xylene and exhibits high hole and electron mobility. 40 Moreover, the HOMO (5.2 eV) and LUMO (3.5 eV) of DPP-DTT matches well with that of the emissive layer which should facilitate hole or electron transport from DPP-DTT back to the emissive polymers (Fig. 4b). The polymer emissive layer was then spin-coated from p-xylene to avoid the dissolution of DPP-DTT and then annealed at 120 C. To simplify device fabrication, we employed symmetric drain/source electrodes. Transfer and output curves were measured at positive and negative source-drain voltages (V DS ) to test for n-channel and pchannel in our device respectively. Fig. 5 shows that our OLET devices exhibit ambipolar behavior with V-shaped transfer curves. It is evident that the charge transport occurs predominantly at the DPP-DTT/dielectric interface. The calculated mobilities of TPTQF-C (m h ¼ 2.5 10 2 cm 2 V 1 s 1 ; m e ¼ 3.2 10 2 cm 2 V 1 s 1 ), and TPTQ-C (m h ¼ 3.5 10 1 cm 2 V 1 s 1 ; m e ¼ 5.1 10 1 cm 2 V 1 s 1 ), from transfer curves and I on/off for tri-layered devices are several orders of magnitude larger than for single-layered devices (Table S4 †). Our OLET devices for TPTQF-C and TPTQ-C exhibited strong yellow-green and yellow emission respectively, although the emission zone was fxed near the electrodes (Fig. 5). A detailed investigation into light emission revealed that the electroluminescence spectra of the polymers were very close to the 0-0 transition band in the flm photoluminescence spectrum (Fig. 4c and d), indicating the identical nature of emissive centers for both PL and EL processes. The transfer curves for the OLET device and the photocurrent for the reverse-biased photodiode were simultaneously measured by placing calibrated photodiode right in front of the device and observing the response. Based on the photocurrent obtained from photodiodes and source-drain current in OLET devices, we can measure the EL intensity and EQE of our OLET devices. 41,42 As shown in Fig. 5c and f, the EL intensity decreased with decreasing gate voltages from negative to positive, and then increased with increasing gate voltages starting from around V G ¼ 40 V. The EL intensity achieved for TPTQ-C (200 nW) and TPTQF-C (216 nW) were comparable to the best trilayered OLET devices reported by Capelli et al. 19 TPTQF-C showed the highest EQE of 3.5% at low applied voltages (V DS ¼ 60 V, V G ¼ 51 V) which was more than three orders of magnitude higher than that of the corresponding tri-layered OLED (Fig. S7 †). In comparison, the intrinsically low PLQY in TPTQ-C, and larger source-drain current due to higher charge mobilities in tri-layered OLET devices (Table S4 †), limited the electroluminescence efficiency, and led to an EQE of only 0.0050%. The same tri-layered OLETs of linear polymers, TPTQF-CC and TPTQ-CC, were fabricated and measured as shown in the ESI (Fig. S8 and Table S5 †) for comparison. EL intensity ($10 1 nW) and EQE obtained in OLET devices of TPTQF-CC (0.0032%) and TPTQ-CC (0.00022%) were much lower than the corresponding cross-conjugated coiled foldamers, TPTQF-C and TPTQ-C. Since the highly fluorescent foldamer, TPTQF-C exhibited good performance in tri-layered OLET devices, further optimization of the device structure was essential. As shown in Fig. 5e, the charge carriers in the DPP-DTT layer recombined with the injection charge carriers from PFN + BIm 4 near the electrodes and displayed a narrow emission zone, which behaved more like an OLED and limited the EQE. It was because the thin PFN + BIm 4 layer was unable to transport charge carriers that led to the exciton quenching on gold. 18 Thus, another charge transporting layer DFH-4T was inserted between the emissive layer and the charge injection layer as shown in Fig. 6 and 4b. 43,44 The The HOMO (6.3 eV), however, was much lower than that of TPTQF-C. Moreover, DFH-4T showed high charge mobility of 0.5 cm 2 V 1 s 1 which is comparable with that of DPP-DTT. As shown in Fig. 6a and S11, † the much stronger yellowgreen emission zone in these new devices extended signifcantly and nearly covered the whole channel. This is in sharp contrast to the device shown in Fig. 5e. This demonstrated that the injection charge carriers in the DFH-4T layer can transport efficiently and recombine with charge carriers from DPP-DTT bottom layer in the middle of the emissive layer. Here, we observed impressive EL intensity (2332 nW) and EQE (6.9%) as shown in Fig. 6d. Notably, an interesting observation is that the emission zone is not a narrow line as observed in those singlelayer devices. 41 This may have complicated reasons. One explanation is likely due to multilayers so that the gating feld sensed by the emission layer was broadened. A more detailed device investigation is in progress to elucidate this point. ## Conclusions Our results reveal that the semi-ladder copolymers TPTQ-C and TPTQF-C exhibiting a foldamer structure show balanced electrical and light-emitting properties. It is known that foldamers are well studied in biological macromolecules or synthetic polymers (oligomers) that adopt highly ordered helical-like selfassembling structures by non-covalent interactions. 45 The investigation of foldamers provides insight into biological systems and is of great importance when developing new selfassembling materials. Artifcial foldamers have shown promising applications in chiral recognition, circularly polarized luminescence, asymmetric catalysis, etc. 45 Though chemists have proposed strategies for dye self-assembly, 46,47 synthetic and design protocols for functional foldamers such as light-emitting materials are rare, if any. This is the frst time anyone has demonstrated that coiled donor-acceptor semi-ladder polymers can form folded structures exhibiting superior device performance. The observed high EQE is remarkable considering the fact that the polymer exhibits only modest PLQY and low mobility. There are three factors that enhanced the EQE of the OLET device. The frst is the unique structures of the foldamer that allows an optimal compromise in light emission and charge transport, leading to high EQEs. The second one is the inserted charge-transporting layers which balanced charge injection and transport so that excitons are formed away from the edge of electrodes, which is evident from the EL image. The third one is that emitted light can be extracted from top side in our bottom gate top contact confguration, and don't need to pass through the highly refractive transparent electrodes like OLED, which can achieve higher outcoupling efficiency. This design strategy may pave the way for the development of even more efficient polymers that can be used in the next generation of high-efficiency OLETs. ## Conflicts of interest There are no conflicts to declare.

chemsum

{"title": "Foldable semi-ladder polymers: novel aggregation behavior and high-performance solution-processed organic light-emitting transistors", "journal": "Royal Society of Chemistry (RSC)"}

a_relay_strategy_actuates_pre-existing_trisubstituted_olefins_in_monoterpenoids_to_form_new_trisubst

## Abstract: A retrosynthetic disconnection-reconnection analysis of epoxypolyenessubstrates that can undergo cyclization to podocarpane-type tricyclesreveals relay-actuated  6,7 -functionalized monoterpenoid alcohols for ruthenium benzylidene catalyzed olefin cross metathesis with homoprenyl benzenes. Successful implementation of this approach provided several epoxypolyenes as expected (E:Z, ca. 2-3:1). The method is further generalized for the cross metathesis of pre-existing trisubstituted olefins in other relay-actuated  6,7 -functionalized monoterpenoid alcohols with various other trisubstituted alkenes to form new trisubstituted olefins.Epoxypolyene cyclization of an enantiomerically pure, but geometrically impure, epoxypolyene substrate provides an enantiomerically pure, trans-fused, podocarpane-type tricycle (from the E-geometrical isomer). ## Introduction Biomimetically inspired polyene cyclizations have emerged as a powerful synthetic strategy for the stereocontrolled construction of complex polycarbocyclic scaffolds of biological significance, 1 where epoxypolyene cyclizations of terminally functionalized geranyl units with nucleophilic aromatic headgroups have provided synthetic access to podocarpane-type tricyclic diterpene skeleta (Figure 1a). 2 Such cyclization substrates are typically constructed in two steps via metal-catalyzed cross-coupling methodology of an electrophilic geranyl species in conjunction with a benzylic organometallic, andeither before or after C-C bond constructionregioselective functionalization of the geranyl alkene at the terminus of the chain (Figure 1a). 3 Each of these steps is subject to a potential disadvantage: the former is subject to competing allylic SN2' substitution, and the latter to non-perfect regioselective oxidation, regardless of the order of implementation. 4 During the course of our studies, we had reason to consider an alternative disconnection of such functionalized linear monoterpenoid derivatives by olefin cross metathesis, but of the two terminal olefin species that are revealed, the epoxide-containing component is synthetically non-simplified (Figure 1b). Nonetheless, a 'reconnection' operation 5 reveals a geraniol derivative with a pre-existing trisubstituted olefin that we expected could be actuated for cross metathesis by the application of Hoye's relay strategy. 6 For reasons outlined below, we also elected to 'reconnect' the terminal alkene component from the initial disconnection as a trisubstituted alkene. The catalyst(s) of choice for the above proposition would be the commercially available well-defined ruthenium benzylidenes as developed by Grubbs. 7 Such catalysts are widely used to accomplish the ringclosing metathesis of disubstituted, trisubstituted and even tetrasubstituted olefins. 8 In contrast, and quite surprisingly, there are only three reports on the formation of unfunctionalized trisubstituted olefins (as required here) by cross metathesis using ruthenium benzylidene pre-catalysts. 9 Grubbs and co-workers initially showed that ruthenium pre-catalyst 1 was competent for the cross metathesis of geminally disubstituted olefins with terminal olefins (Figure 1c). Subsequently, Robinson and co-workers showed that the cross metathesis of sterically challenging allyl branched 1,1-disubstituted olefins performed considerably better using a (terpenoid) prenyl rather than an allyl partner using pre-catalyst 2 (Figure 1d). 12 With this latter literature precedent in mind, we therefore selected trisubstituted olefins as the cross metathesis partners (Figure 1b, reconnection). 13 As envisioned, this overall stratagem not only opens up the possibility of an alternative, modular, synthetic route to such cyclization precursors, but perhaps more significantly could provide a general approach to the functionalization of pre-existing trisubstituted olefins in acyclic monoterpenoid alcohols by cross metathesis (Figure 1e). 14 Herein, we report the success of this unprecedented olefin-olefin combination to form new unfunctionalized trisubstituted olefins by cross metathesis (Figure 1e), where the overall transformation can be classified as a relay cross metathesis. This relay cross metathesis reaction ("ReXM") distinguishes itself from the very limited literature precedent for such reactions by being the first such example to form isolated, unconjugated, trisubstituted alkenes where all previous reports have formed conjugated alkenes. 15,16 4 ## Results and Discussion Scheme 1. Synthesis of relay-modified  6,7 -functionalized monoterpenes (E)-5a and (Z)-5a Scheme 2. Synthesis of various  6,7 -functionalized monoterpenes 5b-g. We commenced our investigations with two main objectives in mind: i) demonstration of proof-of-principle ReXM of monoterpenoid alcohol derivatives with homoprenylbenzenes to prepare representative epoxypolyene cyclization substrates; ii) exemplification of the method as a general approach for the functionalization of pre-existing trisubstituted olefins in acyclic monoterpenoid alcohols. Accordingly, we assembled relay-modified  6,7 -functionalized monoterpenes (E)-5a and (Z)-5a from geraniol [(E)-3] and nerol ## 5 [(Z)-3] via allylation 17 and epoxidation with mCPBA (Scheme 1). We also prepared diols (S)-and (R)-5b via Sharpless dihydroxylation 18 of triene (E)-4 in excellent enantiomeric purityconfirmed by conversion to their respective benzoates 5c and chiral stationary phase HPLC analysis (see SI)and thence acetonides (S)-and (R)-5d (Scheme 2) by ketalization. Relay-free acetonide (S)-5e was prepared from geranyl acetate as a control substrate by the use of Scafato's methods. 19 Control substrate (S)-5f was prepared by the action of Grubbs catalyst 1 on (S)-5d, thereby inherently confirming the ability of the allyl group to function as a relay in this situation. Boronates (S)-and (R)-5g were also prepared from diols (S)-and (R)-5b by direct condensation with phenyl boronic acid in ethyl acetate. These latter substrates, now incorporating UV active chromophores, could be analyzed directly by HPLC for enantiomeric purity and were found to have identical enantiomeric excesses to benzoates (S)-and (R)-5c (see SI). ## Scheme 3. Synthesis of various trisubstituted alkenes Attention now turned to the assembling a collection of suitable trisubstituted alkenes for this study. Trisubstituted alkenes 7a and 7d-7k were prepared by the Wittig reaction of isopropyl phosphonium iodide with aldehydes (Scheme 3). 20 Alternatively, trisubstituted alkene 7c could be prepared by the reaction of the 6 corresponding benzylic Grignard reagent with prenyl bromide under Pd(0) catalysis. 3b The former method is preferred, since non-perfect regioselectivity from competing SN2' attack is possible in the latter. Prenyl acetone 7l was commercially available, as was terminal alkene 7b which was used for control experiments (vide infra). With these substrates in hand we selected relay (E)-5a and trisubstituted alkene 7a as the partner olefin to test in the proposed ReXM reaction. It is well established that trisubstituted olefinsclassified as Type III olefins 21 do not homodimerize, and this prompted us to use trisubstituted alkene 7a in excess with the expectation that this would thereby help facilitate the desired cross metathesis. Although various attempts to mediate the proposed ReXM in toluene or dichloromethane solution failed, neat epoxide (E)-5a underwent smooth ReXM using 10 mol% 1 with trisubstituted alkene 7a (5 equiv.) at 50 ˚C to provide functionalized epoxypolyene 8a in excellent isolated yield (Table 1, entry 1). In stark contrast, the use of terminal olefin 7b under the same conditions (entry 2) with epoxide (E)-5a gave instead direct cross metathesis product 9a and isomerized vinyl ether 10a as the major epoxide containing products, demonstrating that the use of a trisubstituted alkene is critical for these reactions. Control experiments with acetonides (S)-5d-f (entries 4-6) 22 verifies also the vital role of the relay in this ReXM process, and a comparison with the reaction with Z-epoxide (Z)-5a (entry 3) establishes the olefin geometry in the relay substrate as unimportant. Further examples of epoxides (E)-and (Z)-5a with various homoprenyl benzenes 7c-g establishes the generality of the method (entries 7-12). 23 In all successful cases, the ReXM products 8a-g were obtained with moderate E-olefin selectivity (ca. 2-3:1), as inseparable isomers, which is a limitation of the method. 24 However, these selectivities are directly comparable to those previously reported for the formation of trisubstituted olefins by cross metathesis with ruthenium benzylidene pre-catalysts (c.f. Figure 1, c, d). Table 1. ReXM of relay-actuated  6,7 -functionalized monoterpenoids with homoprenyl benzenes using 10 mol% GII catalyst (1) [a] Entry [a] Relay Partner Olefin Product [b-c] 1 (E)-5a 7a [a] 0.25 mmol scale, conditions: olefin (5 equiv.), GII (1) (10 mol%), neat, 50 °C, 1 h; [b] Percentage isolated yields shown after chromatography; [c] A possible catalytic cycle for this ReXM process using representative epoxide (E)-5a with homoprenyl benzene 7a, invokes Diver 15 for the conversion of A to B with loss of dihydrofuran (Figure 2). The regioselective reactions of ruthenium species of type B with trisubstituted olefins have been proposed by Robinson, 12 which would produce the ReXM product 8a, and ruthenium isopropylidene C. In this scenario, the catalytic cycle would be closed by re-initiation of ruthenium isopropylidene C 11 on the terminal olefin of relay epoxide (E)-5a with concomitant loss of isobutylene. 25 This mechanism is consistent also with the results obtained using nerol vs geraniol derived substrates (c.f., Table 1, entries 1 vs 3 & entries 7 vs 8) since the same ruthenium alkylidene of the type B, should be formed after initial relay metathesis. 9 With the ReXM method established for the reaction with homoprenyl benzenes, we explored further reactions with a variety of relay substrates and different trisubstituted alkenes as a general method for the functionalization of pre-existing trisubstituted olefins in acyclic monoterpenoid alcohols (Table 2). Thus, epoxide (E)-5a underwent smooth ReXM with aliphatic trisubstituted alkene 7h to give ReXM product 8h in excellent yield (Table 1, entry 1). -Branching of the alkyl chain as in olefin 7i (entry 2), proved to be detrimental to the process, where ,-dimethylstyrene (7j) and prenylbenzene (7k) (entries 3-4) as partner olefins also failedproducing only truncated alkene 5hpresumably on the basis of increased steric demand in each of these partner olefins. Readily available prenyl acetone 7l gave the ReXM product 8i (entry 5), but diol (S)-5b unexpectedly failed to undergo ReXM (entry 6), resulting in truncated compound 5i and isomerized product 9b (implicating catalyst decomposition to a ruthenium hydride species). 26 Acetonides (S)-& (R)-5d and boronates (S)-& (R)-5g however, participated cleanly in ReXM reactions (entries 7-13) to provide the desired products (S)-& (R)-8b, 8j-l without complication. In these latter instances, these substrates are all derived from highly enantioselective Sharpless dihydroxylations of geranyl allyl ether [(E)-4, vide infra], thereby providing the ReXM adducts in uniformly high enantiomeric excess, which we flag as an advantage of this methodology. Table 2. ReXM of relay-actuated  6,7 -functionalized monoterpenes with various trisubstituted olefins using 10 mol% GII catalyst (1) [a] Entry [a] Relay Partner Olefin Product [b-c] 1 (E)-5a 7h [a] 0.25 mmol scale, conditions: olefin (5 equiv.), GII (1) (10 mol%), neat, 50 °C, 1 h; [b] Percentage isolated yields shown after chromatography; [c] In order to overcome the inherent E/Z mixture limitation of this cross-metathesis method, we elected to demonstrate an epoxypolyene cyclization with the expectation that any resulting products would have more marked polarity differences. Accordingly, we prepared ReXM product (R)-8c from enantiomerically pure epoxide (R,E)-5a and homoprenyl methoxybenzene (7c) in good yield (81%) as an inseparable 2:1 E:Z mixture (Scheme 4). Boron trifluoride promoted epoxypolyene cyclization of this E/Z mixture provided single enantiomer podocarpane-type tricycle 11 (56% yield based on E-8c) as a single diastereoisomer which was readily separated away from the other components in the reaction mixture. 27 To the best of our knowledge, tricycle 11 has not previously been prepared in single enantiomer form, 28 thereby validating the utility of this two-step metathesis-cyclization sequence. 29 Scheme 4. ReXM-epoxypolyene cyclization sequence. ## Conclusion In conclusion, we have designed and demonstrated a novel ruthenium benzylidene catalyzed relay cross metathesis ("ReXM") reaction for the preparation of podocarpane-type epoxypolyene cyclization substrates from relay-actuated  6,7 -functionalized monoterpenoid alcohols with homoprenyl benzenes. It constitutes also a general method for the cross metathesis of pre-existing trisubstituted olefins in other relay-actuated  4, 135.2, 131.8, 124.2, 120.9, 117.1, 71.1, 66.7, 39.8, 26.5, 25.9, 17.8, 16.6; HRMS (EI + ) m/z calcd for C13H22O [M] •+ 194.1671, found 194.1682. 14 31 To a stirred solution of ether (E)-4 (1.41 g, 7.28 mmol, 1.0 equiv.) in CH2Cl2 (30 mL) was added dropwise a solution of mCPBA (1.63 g, 77%, 7.28 mmol, 1.0 equiv.) in CH2Cl2 (30 mL) over ca. 0.5 h at 0 °C. The mixture was allowed to warm to room temperature gradually. After a total reaction time of 18 h, the reaction mixture was concentrated, dissolved in EtOAc (100 mL) and washed with a saturated aqueous NaHCO3 solution (3 × 50 mL), brine (100 mL), dried over Na2SO4, concentrated and chromatographed (20-50% Et2O in pentanes), to give epoxide (E)-5a (1.71 4, 135.1, 121.5, 117.2, 71.3, 66.7, 64.2, 58.5, 36.4, 27.3, 25.0, 18.9, 16.7; HRMS (CI + ) calcd for C13H21O2 [M -H] + 209.1536, found 209.1535. ## (Z)-1-(Allyloxy)-3,7-dimethylocta-2,6-diene [(Z)-4] . 17 Using a modified procedure of Rao and Senthilkumar, to a neat mixture of nerol [(Z)-3] (5.30 mL, 30 mmol, 1.0 equiv), allyl bromide (7.8 mL, 90 mmol, 3.0 equiv) and TBAI (554 mg, 1.50 mmol, 5 mol%) was added crushed KOH pellets (3.37 g, 60.0 mmol, 2.0 equiv) at room temperature and the mixture was stirred for 18 h. The crude reaction mixture was purified by loading directly onto a pad of silica gel and eluting with n-hexane, to give allyl ether (Z)-4 (5.81 g, 29.9 mmol, quant.) as a colourless oil. 5, 135.2, 132.0, 124.0, 122.0, 116.9, 71.1, 66.4, 32.4, 26.8, 25.8, 23.6, 17.7; HRMS (EI + ) m/z calcd for C13H22O [M] •+ 194.1671, found 194.1670. To a stirred solution of ether (Z)-4 (1.41 g, 7.28 mmol, 1.0 equiv.) in CH2Cl2 (30 mL) was added dropwise a solution of mCPBA (1.63 g, 77%, 7.28 mmol, 1.0 equiv.) in CH2Cl2 (30 mL) over ca. 0.5 h at 0 °C, the mixture was allowed to warm to room temperature gradually. After a total reaction time of 18 h, the reaction mixture was concentrated, dissolved in EtOAc (100 mL) and washed with a saturated aqueous NaHCO3 solution (3 × 50 mL), brine (100 mL), dried over Na2SO4, concentrated and chromatographed (20-50% Et2O in pentanes), to give epoxide (Z)-5a (1.60 g, 7.6 mmol, 76%) as a colourless oil. 10.14469/hpc/5742. Rf 0.25 (10% EtOAc in n-hexane); IR (ATR, neat) 3075 cm -1 ; 1 H NMR (400 MHz, CDCl3) δ 5.92 (ddt, J = 17.3, 10.4, 5.7 Hz, 1H), 5.46 -5.38 (m, 1H) 6, 135.0, 122.5, 117.1, 71.3, 66.3, 64.0, 58.5, 29.0, 27.6, 25.0, 23.5, 18.8 (S,E)-8-(Allyloxy) -2,6-dimethyloct-6-ene-2,3-diol [(S)-5b]. 31 Using a modified procedure of Sharpless, 18 to a vigorously stirred solution of AD-mix-α (7.0 g, 1.4 g mol -1 ) in t BuOH (75 mL) and H2O (75 mL) was added MeSO2NH2 (476 mg, 5.0 mmol, 1.0 equiv.) followed by ether (E)-4 (972 mg, 5.0 mmol, 1.0 equiv.). The reaction mixture was allowed to stir for 16 h, the mixture was quenched by the addition of a 20% Na2SO3 3, 135.1, 121.4, 117.3, 78.3, 73.3, 71.4, 66.7, 36.8, 29.7, 26.7, 23.4, 16.7; HRMS (ES + ) calcd for C13H24O3Na [M + Na] + 251.1623, found 251.1631. (R,E)-8-(Allyloxy) -2,6-dimethyloct-6-ene-2,3-diol [(R)-5b] was prepared under identical conditions but using AD-mix-. 31 To a solution of diol (S)-5b (208 mg, 0.91 mmol, 1.0 equiv.) in CH2Cl2 (1.8 mL) at 0 ˚C was added MsCl (0.11 mL, 1.4 mmol, 1.5 equiv.) followed by pyridine (0.59 mL, 7.3 mmol, 8.0 equiv.) dropwise. The reaction mixture was allowed to gradually warm to room temperature and was stirred for 20 h, after which the mixture was poured into a suspension of K2CO3 (1.9 g, 14 mmol, 15 equiv.) in MeOH (9 mL). This suspension was stirred for a further 18 h. The reaction mixture was then concentrated, diluted with H2O (80 mL) and extracted with EtOAc (3 × 40 mL). The combined organics were washed with a saturated aqueous CuSO4 solution (3 × 50 mL), then brine (50 mL), dried over Na2SO4, concentrated and chromatographed (5-10% EtOAc in petrol), to give epoxide (R,E)-5a (171 mg, 0.81 mmol, 89%) as a colourless oil. Rf 0.57 (30% EtOAc in pentanes); Data is otherwise identical to the racemic material (E)-5a. ## (S,E)-8-(Allyloxy)-2-hydroxy-2,6-dimethyloct-6-en-3-yl benzoate [(S)-5c]. To a solution of diol (S)-5b (25.0 mg, 0.11 mmol, 1.0 equiv.) in pyridine (1 mL) was added benzoyl chloride (16 µL, 0.14 mmol, 1.3 equiv.) and the reaction mixture was stirred at room temperature for 18 h. Then was added additional benzoyl chloride (33 µL, 0.28 mmol, 2.6 equiv.) and the mixture was stirred for an additional 5 h. The reaction mixture was diluted with EtOAc (25 mL) and washed with a 1M aqueous HCl solution (3 × 20 mL) and a saturated aqueous NaHCO3 solution (3 × 20 mL). The organics were dried over Na2SO4, concentrated and chromatographed on silica gel (20% EtOAc in n-hexane), to give benzoate (S)-5c (33.0 mg, 0.11 mmol, 99%) as a colourless oil. 8, 139.4, 135.1, 133.3, 130.2, 129.8, 128.6, 121.4, 117.1, 80.4, 72.8, 71.2, 66.6, 36.1, 27.9, 26.7, 25.3, 16.7; HRMS (ES + ) calcd for C20H28O4Na [M + Na] + 355.1885, found 355.1894. HPLC (CHIRALCEL OD; 10% IPA in n-hexane; 0.5 mL/min) tR = 10.1 min (major), 10.9 min (minor) (96:4). (R,E)-8-(Allyloxy)-2-hydroxy-2,6-dimethyloct-6-en-3-yl benzoate [(R)-5c] was prepared under identical conditions from diol (R)-5b. [] D 31 + 12.4 (c 0.8,CHCl3). HPLC (CHIRALCEL OD; 10% IPA in nhexane; 0.5 mL/min) tR = 9.9 min (minor), 10.6 min (major) (4:96). To a solution of diol (S)-5b (1.05 g, 4.59 mmol, 1.0 equiv.) in CH2Cl2 (9 mL) were added dimethoxypropane (5.60 mL, 45.9 mmol, 10.0 equiv.) and pyridinium p-toluenesulfonate (577 mg, 2.30 mmol, 0.5 equiv.) at room temperature and the mixture was stirred for 18 h. The reaction mixture was concentrated and loaded directly onto a column of silica gel and chromatographed (5-10% EtOAc in petrol) to give acetonide (S)-5d (1.11 g, 4.13 6, 135.0, 121.1, 117.0, 106.5, 82.9, 80.1, 71.1, 66.5, 36.7, 28.6, 27.5, 26.9, 26.1, 22.9, 16.6 (S) -2,2,4,4-Tetramethyl-5-(3-methylbut-3-en-1-yl)-1,3-dioxolane [(S)-5f]. To a solution of (S,E)-acetonide (S)-5d (100 mg, 0.37 mmol, 1.0 equiv.) in CH2Cl2 (37 mL) was added ruthenium benzylidene 1 (3.2 mg, 0.0037 mmol, 1 mol%). The mixture was heated to reflux with stirring for 2 h. The reaction mixture was concentrated and loaded directly onto a column of silica gel and chromatographed (5% EtOAc in petrol) to give alkene (S)- 110.1, 106.5, 82.8, 80.1, 34.9, 28.6, 27.4, 26.9, 26.0, 22.9, 22.6; HRMS (EI + ) m/z calcd for C12H22O2 [M] •+ 198.1620, found 198.1612. (S,E)-5- (5-(Allyloxy)-3-methylpent-3-en-1-yl)-4,4-dimethyl-2-phenyl-1,3,2- 135.0, 134.8, 131.3, 127.8, 121.2, 117.1, 85.3, 82.1, 71.2, 66.6, 36.4, 29.9, 28.8, 23.5, 16.7 ] was prepared under identical conditions from diol (R)-5b. []D 25 +15.8 (c 1.0, CHCl3). HPLC (CHIRALCEL OD-H, 1% IPA in n-hexane; 1.0 mL/min) tR = 8.5 min (minor), 10.8 min (major) (3:97). General procedure for preparation of trisubstituted alkenes 7 via Wittig olefination of aldehydes. Using a modified procedure of Pfaltz, 20 to suspension of (CH3)2CHPPh3I 6 (7.8 g, 18 mmol, 1.8 equiv.) in THF ( 20mL) was added n BuLi (11.25 mL, 1.6 M in hexanes, 18 mmol, 1.8 equiv.) dropwise at 0 °C. After 30 min, the required aldehyde (1.0 equiv.) was added dropwise and the mixture was stirred for 18 h. The mixture was quenched by addition of H2O (15 mL) and the THF was removed by concentration. The mixture was diluted with Et2O (100 mL) and filtered through a pad of Celite. The organics were washed with H2O (3 × 20 mL), dried over Na2SO4 and concentrated. The crude material was purified by DCVC, eluting with pentanes, to give the desired trisubstituted alkene. (4-Methylpent-3-en-1-yl)benzene (7a). 32 Following the general procedure for the formation of trisubstituted alkenes using hydrocinnamaldehyde (1.32 mL, 10.0 mmol, 1.0 equiv.) gave alkene 7a (0.99 g, 6. 5, 132.3, 128.6, 128.3, 125.8, 123.9, 36.3, 30.2, 25.8, 17.8; HRMS (CI + ) m/z calcd for C12H17 [M + H] + 161.1330, found 161.1325. ## 1-Methoxy-4-(4-methylpent-3-en-1-yl)benzene (7c). 3b To a solution of p-methoxybenzyl chloride (2.26 mL, 16.6 mmol, 1.0 equiv.) in THF (10 mL) was added dropwise over 20 min to magnesium turnings (0.478 g , 20.0 mmol, 1.2 equiv.) in THF (10 mL) and at 0°C. The reaction was warmed to room temperature and the mixture was stirred for 3 h. To the mixture was added dropwise a solution of tetrakis(triphenylphosphine)palladium(0) (0.226 g, 0.196 mmol, 1.5 mol%) and 1-bromo-3-methylbut-2-ene (2.4 mL, 20.1 mmol, 1.0 equiv.) in THF (10 mL) at -78°C. The reaction mixture turned to green immediately and was stirred for an additional 3 h before warming to room temperature. The reaction mixture was stirred for (m, 5H), 1.28 (s, 3H), 1.24 (s, E-8a, 2.19H), 1.24 (s, Z-8a, 0.81H). The E:Z ratio was determined by integration of the resonances at δ 1.24(4) (major) and 1.23(9) (minor) ppm; 13 C{ 1 H} NMR (101 MHz, CDCl3) δ 142.4, 142.3, 135.0, 135.0, 128.6, 128.4, 128.4, 125.9, 125.9, 125.2, 124.4, 64.3, 64.2, 58.5, 58.5, 36.5, 36.4, 36.2, 30.1, 30.0, 28.7, 27.6, 27.5, 25.1, 25.1, 23.5, 18.9, 18.9, 16.1. The E-isomer was identified as the major isomer on the basis of a characteristic shielded methyl resonance 23 9, 139.1, 133.7, 128.4, 128.3, 127.1, 125.8, 121.5, 70.8, 66.3, 64.0, 58.4, 36.2, 35.5, 34.1, 27.2, 24.9, 18.7, 16.5 2, 145.1, 139.8, 139.7, 128.5, 121.0, 120.4, 101.2, 98.8, 68.2,

chemsum

{"title": "A Relay Strategy Actuates Pre-Existing Trisubstituted Olefins in Monoterpenoids to Form New Trisubstituted Olefins by Cross Metathesis", "journal": "ChemRxiv"}

ligand-mediated_phase_control_in_colloidal_aginse_2_nanocrystals

## Abstract: Synthetic studies of colloidal nanoparticles that crystallize in metastable structures represent an emerging area of interest in the development of novel functional materials, as metastable nanomaterials may exhibit unique properties when compared to their counterparts that crystallize in thermodynamically preferred structures. Herein, we demonstrate how phase control of colloidal AgInSe2 nanocrystals can be achieved by performing reactions in the presence, or absence, of 1-dodecanethiol. The thiol plays a crucial role in formation of metastable AgInSe2 nanocrystals, as it mediates an in-situ topotactic cation exchange from an orthorhombic Ag2Se intermediate to a metastable orthorhombic phase of AgInSe2. We provide a detailed mechanistic description of this cation exchange process to structurally elucidate how the orthorhombic phase of AgInSe2 forms. Density functional theory calculations suggest that the metastable orthorhombic phase of AgInSe2 is metastable by a small margin, at 10 meV/atom above the thermodynamic ground state. In the absence of 1-dodecanethiol, a mixture of Ag2Se nanocrystal intermediates form that convert through kinetically slow, non-topotactic exchange processes to yield the thermodynamically preferred chalcopyrite structure of AgInSe2. Finally, we offer new insight into the prediction of novel metastable multinary nanocrystal phases that do not exist on bulk phase diagrams. Metastability, broadly defined, is the kinetic persistence of a system that exists in a higher free energy state than the thermodynamically most stable state for a given set of conditions. The application of metastable materials are ubiquitous, and include examples from diamond wafers for semiconductor applications to the use of technetium-99m as a radiotracer in gamma ray imaging. 1,2 All nanomaterials are inherently metastable with respect to their bulk material counterparts as a result of their high surface energies and large surface area-to-volume ratios. 3,4 In addition to the useful properties afforded by size effects for colloidal nanocrystal analogs of thermodynamically stable bulk materials of that same crystal phase, the thermodynamic scales of phase equilibria on the nanoscale are often compressed, allowing relatively low-temperature syntheses of crystalline polymorphs that only exist at much higher temperatures and/or pressures in the bulk. Furthermore, entirely new crystal phases can arise on the nanoscale that have no known counterparts in bulk. Because the physical properties of a material are linked to its crystalline structure, the ability to isolate new or difficult-to-access metastable structures on the nanoscale holds promise for the discovery of novel functional materials with properties different from, and possibly superior to, the properties of more thermodynamically stable materials. To synthetically target such materials, it is important to consider that a metastable state is only isolable if, under some set of conditions, that state represents a thermodynamic minimum. 5 In other words, if a state is never the thermodynamically most stable state under any set of conditions, it is not synthesizable. The synthetic chemistry of colloidal nanocrystals that persist in metastable states with respect to their bulk analogs remains a science largely dependent on empirical findings, rather than on bottom-up design principles. This is partially a result of the myriad variables that can contribute to phase determination, such as nanocrystal size, 8 surface area-to-volume ratio, 19 surface functionalization, 14 crystal defects, 9 etc. These confounding variables make it difficult to draw direct analogies between the thermodynamic phase diagrams of bulk materials and corresponding stabilities at the nanoscale, thus making the predictable syntheses of metastable colloidal nanocrystals an outstanding challenge. 23 Diorganyl dichalcogenides (R-E-E-R, where E = S, Se, Te, and R = Ph, Me, Bz, etc.) are proven molecular precursors for the preparation of colloidal metal chalcogenide nanocrystals, and in particular, for the preparation of metastable phases of these nanocrystals, including wurtzite or wurtzite-like phases of CuInS2, Cu2SnSe3, Cu2ZnSnS4, and Cu2-xSe. We were the first to report a previously unknown wurtzite-like phase of CuInSe2 from a synthesis utilizing a diselenide precursor, which was shown to be critical in phase determination of the resulting nanocrystals. We subsequently determined that the functional groups on the diselenide precursor could be leveraged to molecularly program different polymorphs of the resulting colloidal CuInSe2 nanocrystals depending on the C-Se precursor bond strength. 24 Herein, we explore a related ternary chalcogenide, AgInSe2, that is of interest for applications in near-infrared luminescence and as a solar absorber for thin film photovoltaics. Like CuInSe2, AgInSe2 belongs to the family of I-III-VI2 semiconductors that adopt a thermodynamically preferred chalcopyrite structure in bulk. Possessing an A + B 3+ E 2-2 composition, the diamondoid structure of chalcopyrite can be thought of as a supercell of zinc blende, where the A + and B 3+ cations are ordered in the cation sub-lattice and the Se 2sub-lattice adopts a cubic close packed structure. In the case of AgInSe2, a metastable orthorhombic phase is also known to exist only on the nanoscale, where the Se 2sub-lattice adopts a hexagonally close-packed structure. Isostructural with the high-temperature orthorhombic phase of bulk AgInS2, the In 3+ and Ag + cations in this metastable phase of AgInSe2 are ordered and alternate along the crystallographic direction. 34 While dichalcogenides have been utilized to access a wide range of metastable colloidal nanocrystal phases, as previously mentioned, it has also been observed that the presence or absence of coordinating ligands influences phase determination in these reactions. 14, Herein, we elucidate the role of a coordinating ligand in the phase determination of AgInSe2 nanocrystals synthesized using dibenzyl diselenide as the selenium precursor. This mechanism is notably different from previously proposed mechanisms for the formation of metastable orthorhombic AgInSe2 nanocrystals. Finally, we propose a general conceptual framework that explains the isolation of previously empirically discovered metastable polymorphs on the nanoscale and may aid in future rational discoveries of metastable materials that do not exist on bulk phase diagrams. ## RESULTS AND DISCUSSION In a typical reaction, AgNO3 and In(OAc)3 were dissolved together in a mixture of 1-octadecene (ODE), 1-dodecanethiol (DDT), and oleic acid. In a separate flask, the dibenzyl diselenide (Bn2Se2) selenium source was dissolved in DDT and ODE. The metal precursor solution was then heated and the solution containing the diselenide was hot injected into the metal precursors at 200 °C. Under these reaction conditions, we observed the formation of colloidally stable, 10-nm AgInSe2 nanocrystals that crystallize in the orthorhombic Pna21 space group, which is a metastable phase of AgInSe2 known to form only on the nanoscale (Figure 1). 34,41 The powder X-ray diffraction (XRD) pattern of the phase-pure orthorhombic nanocrystals is given in Figure 1a. Rietveld refinement of the XRD pattern using the Pna21 space group returns lattice parameters of a= 7.3151 (2), b = 8.5366(3), and c = 6.9638(1) , with a unit cell volume of V = 434.86(1) 3 . These values are in close agreement with the previously reported experimental values for orthorhombic AgInSe2 (i.e., a = 7.33 , b = 8.52 , and c = 7.02 ; V = 438 3 ). 39 This orthorhombic phase is similar to the wurtzite structure type, with the notable distinction between them being the ordering of Ag + and In 3+ in the orthorhombic structure. Discerning wurtzite from wurtzite-like structures can be difficult and has been a point of interest within studies of metastable ternary chalcogenide materials. 15,33,42 In this case, orthorhombic AgInSe2 in the Pna21 space group exhibits distinct low-angle reflections (at 15-16° 2θ) from the (110) and (011) lattice plane families, which are absent in a higher symmetry wurtzite structure type (space group P63mc, see Figure S1). The Rietveld refinement and the observation of lowangle reflections in Figure S1 leads us to conclude that the metastable AgInSe2 nanocrystals do indeed assume a wurtzite-like structure that maintains Ag + and In 3+ ordering within the crystalline lattice. The diselenide precursor is important for phase determination. When substituting Bn2Se2 for grey selenium in the same solvent mixture, the reaction does not yield product due to the low solubility and reactivity of Se powder under these reaction conditions. However, when Se powder is dissolved in oleylamine and used as the selenium source under otherwise similar conditions, the reaction gives chalcopyrite AgInSe2 nanocrystals. 28 Nonetheless, formation of the metastable orthorhombic phase of AgInSe2 using Bn2Se2 was a surprising result, as it differs from what we observed when employing Bn2Se2 in the synthesis of CuInSe2 nanocrystals; there, diselenide precursors possessing relatively weak C-Se bonds, including Bn2Se2, gave the thermodynamically preferred chalcopyrite phase of CuInSe2. 24 Thus, we anticipated that Bn2Se2 might similarly produce the thermodynamically preferred chalcopyrite phase of AgInSe2, yet this turned out not to be the case. This indicates that the mechanism of formation of this metastable phase when using Bn2Se2 is distinct from that which was previously observed for the formation of CuInSe2. Although Bn2Se2 leads to the metastable orthorhombic phase of AgInSe2, we surmised that increasing the reaction temperature might yield the thermodynamic phase of AgInSe2. The initial reactions with Bn2Se2 to give orthorhombic AgInSe2 nanocrystals were performed at 220 °C. Increasing reaction temperatures to 250 °C still resulted in formation of metastable orthorhombic AgInSe2 with no indication of chalcopyrite formation by XRD (Figure 1a). Annealing powders of the metastable AgInSe2 nanocrystals to 300 °C in the solid state also does not cause the material to thermally relax to the chalcopyrite phase, even after several heating/cooling cycles (Figures S2, S3). Moreover, after leaving the as-prepared orthorhombic AgInSe2 nanocrystals for ~10 months on the lab bench under ambient conditions, they maintain their metastable orthorhombic structure (Figure S2). Heating the as-synthesized orthorhombic AgInSe2 nanocrystals at 300 °C for 1 h as a colloidal suspension in ODE also leaves the metastable phase mostly intact, although some conversion to the chalcopyrite phase was observed, indicating that this metastable phase is more resistant to relaxation as a powder at high temperatures than as a colloid in solution (Figure S2). Empirically, the orthorhombic phase of these AgInSe2 nanocrystals appears to be a local minimum in the energetic landscape of this material system that has a high barrier to reorganization to the thermodynamically preferred phase, and thus the orthorhombic phase remains kinetically persistent. To explore the potential roles of the coordinating species (i.e., DDT and oleic acid) in phase determination, they were systematically omitted from the reactions. When oleic acid is omitted from the reaction by replacing it with an equal volume of DDT, under otherwise identical conditions, the reaction still returns orthorhombic AgInSe2 (Figure S4). This suggests that oleic acid does not play a major role in phase determination. Conversely, when DDT is replaced by an equal volume of oleic acid, we found that the analogous hot-injection reaction with Bn2Se2 performed at 250 °C yields chalcopyrite AgInSe2 with minor Ag2Se impurities (Figure S4). This result illustrates that: (1) Bn2Se2 can give the thermodynamic phase under certain reaction conditions, and (2) DDT plays a critical role in phase determination in this reaction. Formation of Chalcopyrite AgInSe2. To probe the formation of chalcopyrite AgInSe2, a study was performed without DDT in which aliquots were removed at certain time points after the injection of Bn2Se2. Powder XRD patterns of nanocrystal products isolated from each aliquot show a complex mixture of Ag2Se intermediates at early times that, over the span of 15 min, convert into chalcopyrite AgInSe2 upon reaction with In 3+ in solution (Figure 2a). Bulk Ag2Se exhibits two stable polymorphs --namely, a low-temperature orthorhombic phase and a high temperature (T > 130 °C) cubic phase. 9,43 However, an additional metastable tetragonal polymorph is known to form within polycrystalline thin films and for Ag2Se nanocrystals. To the best of our knowledge, the crystal structure of this tetragonal phase of Ag2Se has not yet been unambiguously determined, in large part due to its instability as a bulk material under any known conditions. Even so, Wang et al. conducted a thorough investigation of the phase transitions that occur between the tetragonal, orthorhombic, and cubic phases of Ag2Se nanocrystals by variable-temperature powder XRD measurements. 9 For their system, they reported that the tetragonal phase undergoes a phase transition to the cubic phase at ~110 °C, whereas the low-temperature orthorhombic phase converts to the cubic phase at ~140 °C. Figure 2a illustrates that 1 min after injecting Bn2Se2 into the metal precursor solution in the absence of DDT, all three distinct polymorphs of Ag2Se are present (i.e., orthorhombic, tetragonal, and cubic structures). Phase quantification of each respective polymorph is difficult due to the high degree of overlap of the powder XRD patterns of these three phases. Both the orthorhombic and the tetragonal phases of Ag2Se are likely metastable at the reaction temperature of the aliquot study, and they are both capable of undergoing direct phase transitions to form cubic Ag2Se at elevated temperatures, which led us to believe that perhaps the cubic phase of Ag2Se is the binary intermediate that ultimately gives rise to chalcopyrite AgInSe2. However, a control experiment in which Bn2Se2 was hot-injected into a flask containing only AgNO3 (i.e., with no In(OAc)3 precursor) revealed that these Ag2Se phases do not undergo phase transitions to the cubic phase of Ag2Se after 30 min (Figure S5) under the same conditions used for the aliquot study shown in Figure 2a, suggesting that each of the intermediate Ag2Se phases must be capable of directly converting to chalcopyrite AgInSe2 in the presence of In 3+ cations. This observation is supported by the fact that on the bulk Ag2Se-In2Se3 pseudo-binary phase diagram of AgInSe2, both cubic and orthorhombic Ag2Se can convert to chalcopyrite AgInSe2 with increasing In 3+ content. 41 On the nanoscale, conversion of Ag2Se to AgInSe2 can be thought of as a partial cation exchange in which two equivalents of Ag2Se combine with one equivalent of In 3+ to yield AgInSe2 with the expulsion of three Ag + ions. Neither orthorhombic nor cubic Ag2Se have cubic close-packed Se 2anion sub-lattices (i.e, cubic Ag2Se is body-centered cubic and orthorhombic Ag2Se is nearly hexagonally close-packed, vide infra), whereas the Se 2sub-lattice of chalcopyrite AgInSe2 is cubic close-packed (see Figure S6). Thus, to generate chalcopyrite AgInSe2 nanocrystals from Ag2Se intermediates, a reconstructive transition via non-topotactic cation exchange must occur in which the Se 2sub-lattice reorganizes to a cubic close-packed structure. While this reorganization to the chalcopyrite structure is thermodynamically favored, it is necessarily kinetically slow. For that reason, the hot-injection syntheses without DDT always resulted in products comprised of chalcopyrite AgInSe2 with some binary Ag2Se impurities, even when reactions were carried out in the presence of excess In(OAc)3 and for extended periods of time (Figure S7). To improve the phase purity of the chalcopyrite AgInSe2 products, a heating up procedure can be employed, whereby all reagents were combined in a flask with oleic acid and ODE and heated to the desired reaction temperature. This method proved to be more effective in converting the Ag2Se intermediates to a product containing almost exclusively chalcopyrite AgInSe2 (Figure S7). Formation of Orthorhombic AgInSe2 and the Role of DDT. The formation of orthorhombic AgInSe2 nanocrystals in the presence of DDT suggests that DDT changes the mechanism of formation for the ternary material. To better understand the mechanism of formation of orthorhombic AgInSe2 nanocrystals when DDT is present, we performed additional aliquot studies with a hotinjection of the Bn2Se2 as the selenium precursor. In contrast to the long-lived binary Ag2Se intermediates observed in the aliquot study with no DDT (Figure 2a), the analogous aliquot study in the presence of DDT reveals fast conversion of precursors to the metastable orthorhombic AgInSe2 product (Figure 2b). The amount of DDT was reduced from the 20 equivalents (relative to the metal precursors) used in the original synthesis to 5 equivalents in order to better observe any binary Ag2Se intermediates (Figure 2c). Above this threshold value, conversion happened so quickly that no binary intermediates were observed preceding the formation of orthorhombic AgInSe2. Notably, Figure 2c illustrates how when DDT is present, the predominate intermediate observed is the orthorhombic phase of Ag2Se, and not the complex mixture of silver selenides that was observed in the absence of DDT, indicating that orthorhombic Ag2Se is the intermediate that leads to orthorhombic AgInSe2. This fast conversion of orthorhombic Ag2Se elucidates the role of DDT in the reaction; as a soft base, it is capable of mediating cation exchange from the orthorhombic Ag2Se phase that is otherwise kinetically sluggish to react in the presence of soft In 3+ cations due to the low intrinsic ionic conductivity of orthorhombic Ag2Se (~10 -4 S/cm). 44 While others have observed the presence of orthorhombic Ag2Se prior to the formation of orthorhombic AgInSe2, this transformation is not well understood in the literature. Abazović et al. speculated that the formation of the metastable phase of AgInSe2 is in some way related to how ligands bind to the surfaces of the ternary nanocrystal nuclei, thus directing the phase towards orthorhombic AgInSe2. 38 We propose a more nuanced mechanism of formation for orthorhombic AgInSe2 whereby a DDT-mediated topotactic cation exchange converts orthorhombic Ag2Se to orthorhombic AgInSe2. Structural comparisons of orthorhombic Ag2Se to orthorhombic AgInSe2 reveal similarities between these two crystal structures and elucidate how the process of cation exchange transforms the former into the latter. Upon examining the Se 2sub-lattice of orthorhombic Ag2Se, it is apparent that there exists a nearly hexagonally close-packed network of Se 2anions in the direction (Figure 3a). These hexagonal sheets of Se 2are nearly planar, although the in-plane Se-Se angles are distorted from the 120° inplane angles within the hexagonal lattice of orthorhombic AgInSe2 (Figure 3a, b). The interplanar d-spacing between Se 2sheets along the direction in orthorhombic Ag2Se is 3.56 , whereas the d-spacing along the direction of close packing in AgInSe2 is slightly less, at 3.51 . Moreover, the average Se-Se distance within a hexagonal sheet of Se is 4.53 for orthorhombic Ag2Se and 4.24 for orthorhombic AgInSe2. Topotactic cation exchange from orthorhombic Ag2Se to orthorhombic AgInSe2 should naturally allow for this slight lattice contraction, considering the ionic radius of four-coordinate Ag + is 129 pm and that of four-coordinate In 3+ is 94 pm. Overall, the Se 2sub-lattice of orthorhombic Ag2Se resembles that of AgInSe2, since only slight changes are needed to take the former to the latter. Considering that the Se 2sub-lattices are so similar, the redistribution of cations upon cation exchange with In 3+ comprises a more significant structural transformation in going from orthorhombic Ag2Se to orthorhombic AgInSe2. The asymmetric unit of orthorhombic Ag2Se has one crystallographically unique Se 2site and two unique Ag + sites. 43,45 Of the two Ag + sites, one site resides within a tetrahedral hole. These tetrahedra share edges with two adjacent, symmetrically equivalent tetrahedra along the direction. The other Ag + site exists in a trigonal planar coordination geometry (Figure 3d). The orthorhombic structure of AgInSe2 is a wurtzite-like structure in that the Se 2sub-lattice is hexagonally close-packed and all cations reside in corner-sharing tetrahedral coordination environments. Thus, to form this structure from orthorhombic Ag2Se, cation exchange needs to occur in a manner that disrupts the edgesharing and trigonal planar coordination geometries to yield the requisite corner-sharing tetrahedron motif. To achieve such a transformation, the periodic tetrahedral holes that exist within the structure of orthorhombic Ag2Se (Figure 3c, e) need to be filled by either incoming In 3+ ions or by neighboring Ag + ions that, when migrating, would then leave corner-sharing tetrahedral holes that In 3+ could fill. Figure 3e demonstrates how removing the edge-sharing tetrahedra and trigonal planar coordination environments from the orthorhombic Ag2Se structure, and placing cations within the periodic tetrahedral holes, leads to the wurtzite-like structure of orthorhombic AgInSe2. Occupation of the tetrahedral holes in orthorhombic Ag2Se would lead to unstable, edge-sharing configurations with both the proximal Ag + tetrahedra and trigonal planar sites (Figure 3c). Every In 3+ ion incorporated into the structure necessarily must expel three Ag + ions to maintain charge neutrality. Therefore, it is useful to visualize a transformation wherein each In 3+ atom displaces one Ag + atom from a neighboring tetrahedral coordination site and two Ag + atoms from trigonal planar coordination sites, creating more stable corner-sharing configurations via the displacement of edge-sharing motifs within the structure. While the mechanism described above illustrates how the cornersharing network of tetrahedra in orthorhombic AgInSe2 can be derived from orthorhombic Ag2Se, it does not explicitly explain how or why the specific ordering of cations in orthorhombic AgInSe2 arises through this transformation. In fact, the tetrahedral holes within orthorhombic Ag2Se are periodic such that along the direction, they form a linear channel of vacancies (Figure 4a). If all In 3+ cations were to occupy these vacancies, the resulting ternary structure would contain linear chains of Ag + and In 3+ in the direction, where the cations within each chain would be identical (Figure 4d). However, this arrangement of cations is not present within the orthorhombic structure of AgInSe2, rather, the cationic sites along the direction alternate between Ag + and In 3+ (Fig- ure 4b). This indicates that an ion hopping process is operative during cation exchange such that Ag + ions migrate to accommodate incoming In 3+ . By comparing the calculated electrostatic site potentials and Madelung energy of the orthorhombic AgInSe2 structure to the site potentials and Madelung energy of the structure that would result by simply filling the periodic holes within orthorhombic Ag2Se, we found that there is an electrostatic driving force that causes this shuffling of Ag + during cation exchange; in the theoretical ternary structure, derived directly from orthorhombic Ag2Se with no ion hopping, the calculated In 3+ site potential is greater (-1.54 e/) than that for the In 3+ site in orthorhombic AgInSe2 (-1.67 e/), which is an indication that electrostatic repulsion between neighboring In 3+ ions is more significant in this theoretical arrangement than in the orthorhombic structure of AgInSe2. This finding is also supported by the Madelung energy calculations, which represents the attractive electrostatic component to the lattice energy of an ionic solid. 46 The Madelung energy of the theoretical ternary structure is higher in energy (-7.38 MJ/mol) than that of the experimentally observed orthorhombic AgInSe2 structure (-7.60 MJ/mol; see SI for calculation details). The Materials Project database contains thermodynamic information calculated on six polymorphs on AgInSe2, two of which are experimentally known (!4 # 2%, '3 # )), and four of which are theoretical structures (R3m, I41/amd, Fdd2, P4/mmm) calculated by DFT. Of these, the tetragonal !4 # 2% polymorph is predicted to be stable, with the trigonal '3 # ) polymorph exhibiting a degree of metastability at 0.1 eV/atom above the 0 K convex hull (Figure S8). Typically, materials with a predicted metastability in the range of ~0.1 eV/atom are considered in principle synthesizable under appropriate conditions, although this is highly dependent on chemical composition. 47 To supplement these calculations, an additional calculation was performed on the orthorhombic Pna21 polymorph of AgInSe2. The Pna21 polymorph was found to have a formation energy of -0.412 eV/atom, which is 10 meV/atom above the predicted stable chalcopyrite phase of AgInSe2. Figure S8 combines this result with existing Materials Project data calculated using the phase diagram analysis capability of the pymatgen package. 48 This low lying metastability is not unprecedented; in fact, many metastable metal selenide materials are less than 25 meV/atom above the thermodynamic ground state, and the median energy above the ground state for metastable ternary polymorphs irrespective of composition is 6.9 meV/atom. 5 Thus, while the orthorhombic structure is metastable, it is only higher in energy than the chalcopyrite structure by a small margin, which may explain why it is isolable. Predicting the Syntheses of Novel Metastable Polymorphs on the Nanoscale. Predictable syntheses of metastable materials at large remain a challenge. From this work, and our previous work on phase control of CuInSe2 nanocrystals, 24 we note an interesting pattern emerging. In both cases, the metastable ternary chalcogenide nanocrystals form via cation exchange from low-temperature structures of binary selenides, which are metastable at the relatively high temperatures of their respective nanocrystal syntheses. The highlighted areas indicate the type of Se 2sub-lattice present for each respective phase (blue = pseudo-hcp, red = bcc, white = fcc). Notably, there exist lattice mismatches in going from either phase of Ag2Se to chalcopyrite AgInSe2. Such lattice mismatches can be taken advantage of by leveraging fast cation exchange kinetics on the nanoscale to generate novel metastable ternary structures. (b) Reaction scheme explaining isolation of metastable AgInSe2; the orthorhombic phase of AgInSe2 is only 10 meV/atom in energy higher than the chalcopyrite phase and has a Se 2sublattice analogous to that of Ag2Se, allowing for fast, DDT-mediated conversion to the ternary metastable phase. Notably, for both copper and silver selenides, the low-temperature (Cu3Se2 and orthorhombic Ag2Se) and high-temperature (Cu2-xSe and cubic Ag2Se) phases differ significantly in their Se 2sub-lattices, with the low-temperature phases of each being pseudohexagonal and the high-temperature phases assuming face-centered and body-centered cubic Se 2sub-lattices, respectively. As mentioned above, the chalcopyrite structure type possesses a face-centered cubic Se 2sub-lattice. Therefore, isolating metastable ternaries in these cases relies on the conversion of binary selenides that possess Se 2sub-lattices that do not form in bulk for the ternary materials. In the formation of both metastable polymorphs of AgInSe2 and CuInSe2, kinetically fast topotactic cation exchange mechanisms provide the means of preserving the distinct hexagonal Se 2sub-lattices upon reaction within In 3+ . These mechanisms outcompete processes that would otherwise lead to the thermodynamically preferred crystal structures, and instead lead to metastable ternary structures that do not exist on their respective bulk phase diagrams for the ternary selenides. More generally, a promising area to explore in the rational discovery of new metastable nanomaterials may be within material systems that exhibit sub-lattice mismatches between the binary and ternary anionic sub-lattices, where the binary polymorphs with distinct anionic sub-lattices could generate new metastable ternary structures by reacting with a third element in such a way that the anionic sub-lattice is preserved. In effect, lattice mismatches between anionic sub-lattices can act as effective kinetic barriers that restrict quick access to thermodynamic structures, allowing for the isolation of metastable polymorphs on the nanoscale, exemplified by Figure 5. Inspecting pseudo-binary phase diagrams for ternary material systems, and the phase diagrams of the binaries that could lead to ternary materials, is insightful and can act as a guide when searching for lattice mismatches to exploit for new metastable nanomaterial syntheses. To check that this conceptual framework holds true for more than just CuInSe2 and AgInSe2, we inspected the pseudo-binary phase diagram of the Cu2Se-SnSe2 system; here, with increasing Sn 4+ content, cubic Cu2Se (66.7 % Cu, 33.3 % Se, stable above ~130 °C) converts to a cubic, sphalerite phase of Cu2SnSe3. 49 However, the low-temperature Cu3Se2 phase (60 % Cu, 40 % Se) has a pseudo-hexagonal anionic sub-lattice 24 and it maintains a Cu:Se ratio within the boundaries of the two-phase Cu2Se-SnSe2 region. Therefore, we expect a kinetically fast reaction of Cu3Se2 with Sn 4+ to produce a metastable, hexagonal phase of Cu2SnSe3 since there exists a sub-lattice mismatch in going from Cu3Se2 to the thermodynamically preferred sphalerite phase of Cu2SnSe3. Indeed, such a metastable hexagonal phase exists that was previously unknown in bulk, as we first reported the isolation of wurtzite-like Cu2SnSe3 nanocrystals in 2012. 50 This further illustrates the utility of leveraging sub-lattice mismatches to generate novel metastable materials. We hypothesize that this conceptual framework can also be extended to the predictable isolation of metastable phases not present on bulk phase diagrams for quaternary materials. To support this hypothesis, we turned to the Cu2ZnSnS4 literature. Cu2ZnSnS4 is a quaternary material that possesses a face-centered cubic anionic sub-lattice and crystallizes with a kesterite structure type, 51,52 analogous to the diamondoid chalcopyrite structure type for ternary materials. The quasi-ternary Cu2S-ZnS-SnS2 phase diagram shows that, in bulk, the introduction of ZnS and SnS2 into Cu2S results in the conversion of a Cu2S polymorph (digenite, a high-temperature phase stable up to 1130 °C) with a face-centered cubic anionic sublattice to kesterite Cu2ZnSnS4. 53,54 However, wurtzite-like Cu2ZnSnS4 nanocrystals have been synthesized, 55,56 despite the fact that this phase does not exist in bulk. Phenomenologically, this wurtzite-like phase must be the result of kinetically fast reactions with a low-temperature phase of Cu2-xS (such as djurelite or roxbyite) 23,54 that does not possess an fcc anionic sub-lattice. Thus, the predictive power of this conceptual framework can be proven by using it to explain empirically discovered metastable ternary and quaternary nanomaterials. In summary, coupling the identification of material systems that exhibit lattice mismatches between potential kinetic intermediates and the thermodynamically expected products with computations that reveal the energetics of the predicted metastable phases could provide a useful new methodology for the rational discovery of metastable nanomaterials that have never been observed before on bulk phase diagrams. ## CONCLUSIONS In conclusion, this work sheds light on the mechanism of phase determination in AgInSe2 nanocrystals. More specifically, DDT mediates a fast cation exchange from orthorhombic Ag2Se to form a metastable phase of orthorhombic AgInSe2. Without the use of DDT as an exchange mediator, this orthorhombic Ag2Se intermediate cannot undergo cation exchange due to its low intrinsic ionic mobility. For reactions that occur in the absence of DDT, various silver selenide intermediates form and then convert to the thermodynamic chalcopyrite structure of AgInSe2 via kinetically slow non-topotactic cation exchange processes. In addition to elucidating the mechanism of formation for the metastable orthorhombic phase of AgInSe2, we discovered that its isolation likely also correlates with the fact that it is only marginally metastable at 10 meV/atom above the ground state. Finally, we provide a new conceptual framework to predict metastable polymorphs that do not form in bulk; using phase diagrams, it is possible to identify sublattice mismatches that exist between kinetic intermediates that form quickly in nanocrystal syntheses and the thermodynamically most stable polymorphs for multinary materials. Fast conversion of intermediates with distinct sub-lattices can generate new metastable structures of multinary nanomaterials not present on bulk phase diagrams. In predicting these new phases, convex hull calculations can provide an idea of whether or not such metastable materials should be isolable from a thermodynamic perspective. ## EXPERIMENTAL SECTION Materials and General Procedures. Silver(I) nitrate (AgNO3, Alfa Aesar, 99.9%), indium(III) acetate (In(OAc)3, Alfa Aesar, 99.99%), dibenzyl diselenide (Bn2Se2, Alfa Aesar, 95%), 1-dodecanethiol (DDT, Alfa Aesar, 98%), 1-octadecene (ODE, Sigma-Aldrich, 90%), oleic acid (Alfa Aesar, 90%), and selenium powder ~200 mesh (Alfa Aesar, 99.999%) were all used as received, with no further purification. All solvents were degassed prior to use for 4 h at 105 °C and then overnight at room temperature. Reactions were conducted under a nitrogen atmosphere using standard Schlenk techniques. All reactions employed J-KEM temperature controllers with in-situ thermocouples in order to control and monitor the temperature of the reaction vessel. Synthesis of AgInSe2 Nanocrystals. We adapted the general synthesis of AgInSe2 nanocrystals from Deng et al, 28 but here using a diselenide precursor. In a typical synthesis, AgNO3 (16.9 mg, 0.1 mmol) and In(OAc)3 (29.1 mg, 0.1 mmol) were loaded into a 25 mL three-neck round-bottom flask. Bn2Se2 (34.0 mg, 0.1 mmol) was added to a separate two-neck round-bottom flask. In the syntheses of orthorhombic AgInSe2 nanocrystals, 4 mL of ODE, 0.5 mL of DDT, and 50 µL of oleic acid were added to the three-neck flask and 0.5 mL of ODE and 0.5 mL of DDT were added to the two-neck flask. In the syntheses of chalcopyrite AgInSe2 nanocrystals, all DDT was replaced with an equal volume of oleic acid, keeping the volumes of ODE constant. After adding the solvents, the flasks were degassed at 100 °C for 1 h. The metal precursorcontaining flask was then ramped to 250 °C at 10 °C/min under nitrogen. Upon reaching a high temperature (200 °C for DDTcontaining reactions and 230 °C for reactions not containing DDT), the Bn2Se2 solution was injected into the metal precursor-containing flask, resulting in nucleation of nanocrystals (the relatively low 200 °C injection temperature for DDT-containing reactions was implemented to prevent formation of sulfides prior to injection of the diselenide). Following injection, the three-neck flask recovered to 250 °C and was allowed to heat for a total of 30 min after injection. The three-neck flask was then quenched by placing it in a room temperature water bath. The crude product was then split into two 40 mL centrifuge tubes and filled to volume with ethanol. The centrifuge tubes were bath sonicated for 10 min, and centrifuged for 3 min. The product was redispersed in 5 mL of hexanes in each centrifuge tube and filled to volume with ethanol. This washing procedure was repeated two more times to yield particles for XRD analysis. Aliquot Studies. All aliquot studies were performed using the same experimental protocols described for the synthesis of AgInSe2 nanocrystals; however, to capture the various intermediates that precede AgInSe2 formation, the final reaction temperatures (and amount of DDT in the case of orthorhombic AgInSe2) were reduced from 250 °C for the initial syntheses to 230 °C. In the case of orthorhombic AgInSe2, this reduction in temperature was still not enough to capture the timescale of formation of the ternary phase. To observe the binary orthorhombic Ag2Se intermediate, the amount of DDT for the aliquot study was reduced from 20 equivalents with respect to the metal precursors to 5 equivalents. Under these conditions, we were able to observe the binary intermediate. Characterization. Powder X-ray diffraction (XRD) measurements were performed on a Rigaku Ultima IV powder X-ray diffractometer using Cu Kα radiation (λ = 1.5406 ). Samples were analyzed on a zero-diffraction silicon substrate. Transmission electron microscopy (TEM) micrographs were obtained from dropcast samples supported on holey carbon-coated copper TEM grids (Ted Pella, Inc.). Grids were placed in a vacuum oven overnight at 60 °C for removal of volatile organics. A JEOL JEM-2100 microscope with a Gatan Orius charge-coupled device (CCD) camera was used to take TEM images at an operating voltage of 200 kV. Thermogravimetric analysis (TGA) was performed on a TGA Q50 instrument with a heating rate of 10 °C/min with an approximate sample size of 10 mg in an alumina crucible. Density Functional Theory (DFT). Formation energy calculations were performed on the orthorhombic Pna21 polymorph using the Vienna Ab-initio Simulation Package (VASP), plane-augmented wave pseudopotentials and a k-point density of 64 points per -3 , consistent with Materials Project standard settings to ensure the energies would be directly comparable to existing Materials Project calculations. 57,58 The atomic positions and crystal lattice were allowed to relax, resulting in lattice parameters of a = 7.48 , b = 8.76 and c =7.14 . These calculations were performed using the PBE exchange-correlation functional, and so lattice parameters are expected to be slightly over-estimated compared to experiment.

chemsum

{"title": "Ligand-Mediated Phase Control in Colloidal AgInSe 2 Nanocrystals", "journal": "ChemRxiv"}

carbon_monoxide_and_hydrogen_(syngas)_as_a_c1-building_block_for_selective_catalytic_methylation

## Abstract: A catalytic reaction using syngas (CO/H 2 ) as feedstock for the selective b-methylation of alcohols was developed whereby carbon monoxide acts as a C1 source and hydrogen gas as a reducing agent. The overall transformation occurs through an intricate network of metal-catalyzed and base-mediated reactions. The molecular complex [Mn(CO) 2 Br[HN(C 2 H 4 P i Pr 2 ) 2 ]] 1 comprising earth-abundant manganese acts as the metal component in the catalytic system enabling the generation of formaldehyde from syngas in a synthetically useful reaction. This new syngas conversion opens pathways to install methyl branches at sp 3 carbon centers utilizing renewable feedstocks and energy for the synthesis of biologically active compounds, fine chemicals, and advanced biofuels. ## Background and motivation Synthesis gas (syngas), a mixture of carbon monoxide (CO) and hydrogen (H 2 ), is a crucial relay between the energy and the chemical sector. While produced mainly from fossil resources today, 1,2 it can be obtained also from other carbon feedstocks such as biomass, recycled plastics, 6 or even carbon dioxide (CO 2 ) combined with renewable energy. Therefore, catalytic processes for the chemical conversion of syngas are considered central elements in future sustainable chemical value chains. In particular, conversion of renewable-based syngas via the Fischer-Tropsch process or methanol synthesis is fnding wide-spread interest due to the large product volumes. At the same time, "de-fossilized" syngas may be envisaged also as a C1-building block in later-stages of the chemical value chain. 16,17 Hydroformylation, for example, uses syngas for the production of commodities and fne chemicals. We report here a novel catalytic process using syngas to install methyl branches (H 3 C-) with high selectivity at existing aliphatic carbon chains in the b-position of alcohol substrates (Scheme 1). This transformation combines two distinct features of Fischer-Tropsch chemistry (full deoxygenation of CO, hydrocarbon product) and hydroformylation (chemo-and regioselectivity, C1 building block). The reaction is catalysed by a molecular complex comprising earth abundant and non-toxic manganese as an active metal in the presence of a suitable base. The transformation opens new pathways to introduce renewable carbon into molecular structures with potential applications for the synthesis of fuels, large volume products, fne chemicals, and pharmaceuticals. ## Results and discussion Most recently, catalytic methods using methanol (CH 3 OH) for bmethylation of alcohols have been reported by us and others. These reactions occur via an integrated "borrowing hydrogen" reaction sequence involving metal-catalyzed re- hydrogenation and de-hydrogenation in conjunction with basemediated aldol-condensation/isomerization. Formaldehyde is formed as the C1 building block in situ by de-hydrogenation of methanol in these systems. Starting from syngas as the C1 source would thus require to provide sufficient concentrations of formaldehyde to enter this sequence. Notably, examples for catalytic generation of formaldehyde from syngas using homogeneous catalysis are very scarce. 30 Lately, however, the groups of Prakash as well as Checinski and Beller reported the amine assisted hydrogenation of CO to methanol using homogeneous transition metal catalysts based on ruthenium and manganese, respectively. 31,32 Based on recent progress using Mn-complexes in alkylation 22,29, and CO/CO 2 hydrogenation, 32,37 we explored the potential of the well-established complex [Mn( i Pr-MACHO)(CO) 2 Br] ( i Pr-MACHO ¼ HN(C 2 H 4 P i Pr 2 ) 2 ; 1) as a catalyst precursor for the methylation directly from syngas. At the outset, 2-phenylethanol (2a) was selected as a benchmark substrate to validate the catalytic activity of 1 and to screen a set of parameters for optimization. The initial conditions involved reacting 2a under a mixture of CO (5 bar) and H 2 (15 bar) in the presence of complex 1 (1 mol%) and NaO t Bu as a base (2 equiv. with respect to 2a) in toluene as a solvent. Already under this preliminary set of conditions, analysis of the liquid phase revealed >99% conversion of 2a and the formation of the bmethylated product 3a in a yield of 65% after 24 h at 150 C (Table 1, entry 1). Alkenes and aldol-coupled products were observed in the reaction mixture indicating selectivity as the main optimization target. Using the closely related noble metal complex [RuH(CO)(BH 4 )(HN(C 2 H 4 PPh 2 ) 2 )] as a catalyst resulted in rather unselective conversions with 25% yield of 3a only, highlighting the superior performance of the 3d metal in the diagonal position of the periodic table in this case. The selectivity for b-methylation could be improved signifcantly when the amount of complex 1 was increased to 2 mol%. The desired product 3a was formed with a selectivity of 92% at >99% conversion and isolated in 86% yield by column chromatography (Table 1, entry 2). Lower yields were obtained at lower (120 C) as well as higher (170 C) temperatures (Table 1, entry 3 and 4) reflecting a combined influence of activity and selectivity. Reducing the syngas pressure decreased the yield to the desired methylated product drastically while conversion remained relatively high (Table 1, entry 5). Increasing the syngas pressure (CO: 8 bar, H 2 : 24 bar) gave similar results as under the conditions of entry 2 (Table 1, entry 6). When the amount of NaO t Bu was decreased to 1 equiv. with respect to 2a, the rate of the transformation decreased leading to 3a in 67% yield at 82% conversion (Table 1, entry 7). Replacing the base with KO t Bu led to high conversion but very unselective product formation including polymeric materials, while Cs 2 CO 3 resulted in no signifcant activity (Table 1, entry 8 and 9). Monitoring the pressure over time for the b-methylation of 2phenyl propanol 2a under the conditions of entry 2, Table 1, indicated sigmoidal reaction progress (see the ESI †). In the frst 3 hours, the pressure dropped slowly by 2.5 bar followed by a sharp decrease resulting in a total 12 bar pressure drop over 7 h which continued to fnally reach a stable value of a 15 bar total pressure drop after 16 h. The conversion/time profle obtained by the 1 H NMR analysis of the reaction mixture at different time intervals corroborate this observation (Fig. 1). The reaction starts with a signifcant induction period exhibiting only 5% conversion and 4% yield to the b-methylated alcohol product after 1.5 h. The reaction rate continuously increases reaching a maximum around 50% conversion. Subsequently, the transformation slows down but continues to reach >99% conversion while selectivity catches up to result in 92% yield of the desired product 3a. The 1 H NMR spectra revealed the formation of small amounts phenylacetaldehye ( 1 H NMR ¼ 9.86 ppm, t, J ¼ 4 hz), methanediol ( 1 H NMR ¼ 4.89 ppm, s), and methanol ( 1 H NMR ¼ 3.50 ppm, s) as potential intermediate products in the reaction (see the ESI †). a Reaction conditions: 2-phenylethanol 2a (0.5 mmol), Mn-complex 1 (2 mol%), CO, H 2 , base, and toluene (0.8 mL) were heated in a high-pressure reactor for 24 h. Conversion and yield were calculated using 1 H NMR spectroscopy. b Modifed conditions: 2a (0.5 mmol), 1 (1 mol%), CO, H 2 , base, and toluene (0.8 mL) were heated in a high-pressure reactor for 24 h. Scheme 2 shows a plausible reaction network to rationalize the basic mechanism of the new catalytic process which is consistent with the conversion/time profle of Fig. 1 and supported by the control experiments summarized in Scheme 3. 38 The Mn-pincer complex catalyzed de-and re-hydrogenation steps of the organic substrate and intermediates and the base mediated aldol condensation are well established and the deuterium scrambling over all three carbon centers is fully consistent with this sequence (Scheme 3a). 22,36 The high degree of deuterium incorporation at the aand b-position in product 4 reflects rapid de-and re-hydrogenation at all stages of the catalytic network (Scheme 3a, see the ESI, Section 5.2 and 5.3 †). The manganese-catalyzed generation of formaldehyde from CO/ H 2 is unprecedented, however, and unlocks the overall manifold. It can be reasonably assumed as the limiting factor in the initial phase explaining the observed induction period. While a direct hydrogenation pathway for CO reduction cannot be fully excluded, we favor an indirect conversion similar to previous reports on homogeneously catalyzed methanol formation from syngas. 31,32,53,54 These reports have identifed organic formyl species resulting from base mediated coupling of CO and secondary amines as crucial intermediates for CO hydrogenation. In full analogy, the alcohols used as substrates here can be converted to formate esters in the presence of CO and NaO t Bu. 53,55 This possibility was confrmed for substrate 2a in the absence of hydrogen and a catalyst under otherwise typical reaction conditions (Scheme 3b). In line with this assumption, the reaction of 2-phenethyl formate 5a produced the methylated alcohol 3a in 81% yield under standard conditions (Scheme 3c). Furthermore, the reaction of 2a in the presence of ethyl formate and hydrogen led to the methylated product formation with a yield of 40% (Scheme 3d), clearly demonstrating that formate esters can serve as a formaldehyde source. On the other hand, when syngas was reacted under the standard conditions but in the absence of alcohol, the reaction resulted in only a very small amount of methanol (TON # 2) and no formation of formaldehyde or methanediol was detected under these conditions (Scheme 3e). These control experiments affirmed that the presence of the alcohol substrate is necessary to mediate the hydrogenation of carbon monoxide most likely via formate esters as intermediates. Having established a robust method to construct a methyl group from syngas in the b-position of the aliphatic chain in the benchmark substrate 2a, we set out to explore the synthetic scope of this catalytic reaction. The methyl branch is a highly important structural motif in biologically active products such as pharmaceuticals and agrochemicals. 56,57 Using the reaction conditions of Table 1 entry 2 as standard procedure, various 2- arylethanol derivatives were methylated with CO and H 2 addressing potential building blocks and intermediates for biologically active products (Scheme 4, 3a-3i). Electron donating as well as withdrawing substituents in the phenyl ring were fully tolerated and no dehalogenation was observed (Scheme 4, 3a-3h). Notably, pharmaceutically relevant ibuprofen and naproxen alcohols were prepared by using this new methodology in excellent yields of 92% and 96%, respectively (Scheme 4, 3c and 3f). The sulfur containing heterocyclic 2-(thiophen-2-yl)ethanol was converted effectively and the product 3i could be isolated in 86% yield. Aryl-substituted longer chain aliphatic alcohols reacted also smoothly under standard conditions providing very good to excellent yields of the corresponding methyl-branched products (Scheme 4, 3j-3n). Again, heteroatoms were tolerated and the pharmaceutically important amino alcohol 2-(methyl(phenyl)amino)ethan-1-ol 58 was selectively mono-methylated with a yield of 73% at >99% conversion (Scheme 4, 3o). Secondary alcohols of general structure 7 were used also as substrates for selective b-methylation (Scheme 5). Using the standard reaction conditions, 1-phenylpropan-1-ol was monomethylated at the b-position with 81% selectivity at full conversion and the product was isolated in 77% yield (Scheme 5, 8a). Similarly, 1-(p-tolyl)propan-1-ol was also monomethylated in good yield (Scheme 5, 8b). The reaction also occurred readily when the reactive position was part of a 6membered carbocycle, installing the methyl group with a 2 : 1 preference in the trans-position to the OH group (Scheme 5, 8c). Scheme 4 | Mn I catalyzed b-methylation of aryl substituted alcohols 2 with CO and H 2 . Reaction conditions: 2 (0.5 mmol), Mn-complex 1 (2 mol%), CO (5 bar), H 2 (15 bar), NaO t Bu (1 mmol), and toluene (0.8 mL) were heated at 150 C in a high-pressure reactor for 24 h. Conversion and yield were calculated using 1 H-NMR spectroscopy. Yields in parentheses correspond to isolated yield. Scheme 5 Mn I catalyzed b-methylation of secondary alcohols 7 with CO and H 2 . Reaction conditions: same as Scheme 4. a Modified conditions: 7 (0.5 mmol), Mn-complex 1 (2 mol%), CO (5 bar), H 2 (15 bar), NaO t Bu (1 mmol), and toluene (0.8 mL) were heated in a high pressure reactor for 36 h. b 7 (0.5 mmol), Mn-complex 1 (2 mol%), CO (8 bar), H 2 (24 bar), NaO t Bu (2 mmol), and toluene (0.8 mL) were heated in a high-pressure reactor for 36 h. Di-methylation of 1-arylethanols to generate iso-propyl groups was possible under slightly adjusted conditions (see the ESI †). Increasing the amount of base to 4 equiv. and reacting the substrates under higher pressures of CO (8 bar) and H 2 (24 bar) for 36 h allowed isolation of the di-methlyated products 8a and 8d in 70% and 48%, respectively. Methyl branches in aliphatic carbon chains feature benefcial combustion properties in fuel components. 59,60 As a possible synthetic pathway for the upgrading of biogenic alcohols to fuel components with improved combustion properties, 61,62 we therefore focused next on the selective b-methylation of purely aliphatic alcohols (Scheme 6). Ethanol proved to be a challenging substrate, but a mixture of iso-butanol (46%) from di-methylation and 1propanol (16%) from mono-methylation was obtained under 8 bar of CO and 24 bar of H 2 at an elongated reaction time of 36 h (Scheme 6, 10a). Longer chain aliphatic alcohols that can result in mono-methylation only were almost quantitatively converted under 5 bar of CO and 15 bar of H 2 within 24-36 h providing good selectivity and yields (Scheme 6, 10b-10f). Fatty alcohols including lauryl alcohol, and stearyl alcohol were also converted resulting in high yield to the corresponding b-monomethylated alcohols (Scheme 6, 10h, and 10i). Remarkably, the unsaturated fatty alcohol undec-10-en-1-ol was transformed into the corresponding b-methylated product with 68% yield leaving the C]C double bond in the molecule intact (Scheme 6, 10g). These products may have potential application as fne chemicals for surfactants or in the cosmetic industry. ## Conclusions In conclusion, a catalytic reaction has been developed employing syngas as raw material for the catalytic b-methylation of alcohols enabling the use of carbon monoxide as a renewable C1 source and "green" hydrogen as a reducing agent. The catalyst system comprises the earth-abundant, frst row transition metal manganese in the form of an air-stable pincer complex as the metal component. This new catalytic reaction for the installation of methyl groups at sp 3 C-centers generates water as the sole byproduct. The reaction shows a remarkable broad substrate scope providing very high to excellent yields for primary and secondary alcohols. Potential products include fuel components or commodity chemicals, fne chemicals, and even pharmaceuticals. Even with the resource basis of today's petrochemical industry, the catalytic process described herein opens new retrosynthetic pathways to important target molecules providing potential environmental benefts. Most intriguingly, however, the resulting novel synthetic strategies may help to unlock the potential of waste, biomass, or CO 2 as carbon sources for the chemical value chain in line with the principles of Green Chemistry. 17 The general concept to access formaldehyde from syngas through a catalytic cycle with a molecularly defned organometallic complexes and to intercept this useful building block provides a multitude of further opportunities for chemical synthesis. ## Conflicts of interest There are no conflicts to declare.

chemsum

{"title": "Carbon monoxide and hydrogen (syngas) as a C1-building block for selective catalytic methylation", "journal": "Royal Society of Chemistry (RSC)"}

experiment_and_simulation_of_single_inhibitor_sh110_for_void-free_tsv_copper_filling

## Abstract: Three-dimensional integration with through-silicon vias (TSVs) is a promising microelectronic interconnection technology. Three-component additives are commonly used for void-free TSV filling. However, optimising the additive concentrations is an expensive process. To avoid this, a singlecomponent additive was developed: 3-(2-(4,5-dihydrothiazol-2-yl) disulfanyl) propane-1-sulfonic acid/ sulfonate (SH110). Sodium 3,3′-dithiodipropane sulfonate (SPS) and SH110 were used as additives for TSV electroplating copper filling. SH110 resulted in void-free filling, whereas large keyhole voids were found for SPS. To understand how the additives affect the filling mechanism, linear sweep voltammetry of the plating solutions was carried out. The interactions between the Cu surface and additives were simulated by molecular dynamics (MD) analysis using Materials Studio software, and quantum chemistry calculations were conducted using GAUSSIAN 09W. SH110 adsorbs to the Cu surface by both 4,5-dihydrothiazole (DHT) and 3-mercaptopropane sulfonate (MPS) moieties, while SPS is adsorbed only by MPS moieties. MD simulations indicated that the adsorption of the coplanar MPS moiety is the main factor governing acceleration. Quantum chemistry calculations showed that DHT provides an inhibitory effect for TSV filling, while MPS acts as an accelerator for SH110. SH110 is an excellent additive exhibiting both the acceleration and the suppression necessary for achieving void-free TSV filling.Three-dimensional (3D) integration with through-silicon vias (TSVs) is a promising technology for use in electronic systems, as TSVs can provide extremely short vertical interconnections that can improve performance, increase operating speed, and reduce the volume of devices when compared with conventional integration technologies 1-3 . TSV copper filling is one of the key techniques used for TSV fabrication, as it costs ~ 40% less than conventional integration technologies. However, voids often occur upon filling, which must be overcome for reliable TSV fabrication 4,5 .To accomplish void-free TSV filling, an accelerator, suppressor, and leveller are commonly added to the plating solution. At present, the most commonly used accelerator is sodium 3,3′-dithiodipropane sulfonate (SPS) 6 , where the sulfur S-S bonds and sulfonic acid or sulfonate groups (SO 3 H or SO 3 -) are thought to be the key structures responsible for acceleration effects. The typical leveller employed is an organic monomer containing positively charged nitrogen, such as pyridinium, imidazolium, or ammonium 7,8 . Polyethylene glycol (PEG), polypropylene glycol (PPG), and co-polymers thereof are commonly used as suppressor additives 9,10 . These additives accelerate the via bottom deposition rate and suppress the via mouth deposition rate, to obtain bottomup filling. However, it takes numerous experiments to optimise the concentrations of each component in this complex additive system. Therefore, a single inhibitor that accomplishes void-free TSV filling is needed in order to reduce the time and cost of the optimisation process. Thus far, only one report, by Tang and co-workers, has demonstrated a single-component additive (Janus Green B) that could provide void-free filled-in micro-vias 11 .In this study, a single inhibitor, 3-(2-(4,5-dihydrothiazol-2-yl)disulfanyl)propane-1-sulfonic acid (SH110), was found to possess the ability to fill the TSV without introducing voids. Linear sweep voltammetry (LSV) was performed to determine the electrochemical properties of additives in the deposition process for TSV filling. To understand how SH110 affects the TSV filling mechanism, molecular dynamics (MD) simulations and quantum chemical calculations were used to analyse the configuration and electronic structure of SH110 and its interactions with the copper surface, in comparison with a common accelerator, SPS. sectional scanning electron microscopy (SEM) images of the filled TSVs are shown in Fig. 1a,b. The applied electroplating current density was 1 mA/cm 2 with plating times of 12 h. Using SH110 as the additive, the TSVs were fully filled without any voids, as shown in Fig. 1a. The thin film on top of the silicon indicates that these vias were filled in a bottom-up manner. However, for the sample with SPS as the additive, large keyhole voids were found, as shown in Fig. 1b. Thus, SH110 provides excellent filling behaviour. To better understand how the SH110 additive facilitated void-free TSV filling, TSVs filled at different plating times were obtained, as shown in Fig. 2. (Note that, as we designed five different pitches for the vias in the same die, the pitches of the vias in Fig. 2 vary.) Initially (2 h, Fig. 2a), the TSVs were only plated at the bottom, following the U-shaped model; copper deposition was almost totally suppressed in the top half of the vias. Over time (3 h, Fig. 2b), the thickness of the copper at the bottom half of the vias increased, whereas the top half of the via remained suppressed. Then, after the bottom half of the vias were completely filled (8 h, Fig. 2c), copper began to be deposited in the top half in a bottom-up manner. The entire vias were fully filled after 12 h (Fig. 2d). Therefore, with the single inhibitor SH110, the TSVs were filled according to different models at different stages in the filling process. The first stage is the filling of the bottom half of the vias, which proceeds by the U-shape model; the top half of the vias are filled in the second stage, following the bottom-up model. It should be noted that the thickness of the deposited copper layer at the surface remained unchanged with time, indicative of total suppression of the surface. Figures 1 and 2 is generate by Scanning electron microscope (SEM) (TESCAN VEGA 3 https:// www. tescan. com/). SEM is used to observe the cross section of TSV copper plating filling. The TSV hole was magnified by 2000-3000 times. Electron microscope was used to take photos directly. ## Electrochemical procedures. To understand how the additives affect the filling mechanism for TSVs, LSVs was performed using plating solutions with either SH110 or SPS additives, as shown in Fig. 3 and Table 1. According to the additive theory, a large peak current density (I p ) means acceleration plays a leading role, whereas a large valley (I b ) suggests that suppression is predominant. Thus, a larger ΔI value (I p − I b ) indicates better bottom-up filling ability in TSVs, and a larger ΔE (the potential gap between I p and I b ) indicates that there is a wider potential region accessible for bottom-up filling 12,13 . For the primary solution containing SH110, the peak current density (I p = 0.859 A/cm 2 ) appears at approximately − 0.250 V, with a valley (I b = 0.567 A/cm 2 ) at approximately − 0.521 V. For the plating solution with SPS, the peak current density (I p = 1.03 A/cm 2 ) appears at approximately − 0.314 V, with a valley (I b = 0.91 A/cm 2 ) at approximately − 0.397 V. Moreover, the ΔI and ΔE values of the solution with SH110 are 0.292 A/cm 2 and 0.271 V, respectively, which are both greater than those with SPS (0.12 A/cm 2 and 0.073 V, respectively). Therefore, SH110 provides better bottom-up filling ability than SPS, as it has better suppression effects at high-potential locations, such as the via mouth. The larger ΔE of SH110 also means it has a wider potential operation window than SPS. Employing SH110 as a solitary additive results in a fully filled TSV, while the SPS additive results in keyhole voids (Fig. 1). ## Molecular dynamics simulation. Since the adsorption characteristics of additives are intrinsically linked to their role in TSV filling, MD simulations were performed to study the adsorption behaviour of SPS and SH110 on Cu (001), Cu (101) and Cu (111) surface. Because SPS and SH110 are involved in the filling process of www.nature.com/scientificreports/ electroplated copper, the surfaces most likely to be contacted are the surface of Cu (111), the surface of Cu (111) and the surface of copper during the filling process of electroplated copper, the main consideration in the filling process of TSV is (111) . Figure 4a-d show the initial and equilibrium stages for adsorption of SPS. It can be seen that In the initial state onCu (111) (Fig. 4a), the MPS groups are oriented away from the copper surface, thus giving the SPS molecule a V-shape. SPS is adsorbed coplanar to the copper surface when in equilibrium on Cu (001) (Fig. 4b). SPS www.nature.com/scientificreports/ is vertically adsorbed on the surface of the copper crystal by sulfonic group(SO 3 H) when in equilibrium on Cu (101) (Fig. 4c). SPS is adsorbed coplanar to the copper surface by the two MPS groups when in equilibrium on Cu (111) (Fig. 4d). Since SPS is a commonly employed accelerator, our results suggest that the coplanar adsorption of MPS groups to the copper surface is the main factor contributing to acceleration 10, . Figure 5a-d show the initial and equilibrium stages for adsorption of SH110. Initially (Fig. 5a), SH110 is located far from the copper surface, suggesting only very slight adsorption. It is vertically oriented and positioned in a sideways V-shape. In equilibriumon, Similar to SPS, SH110 is adsorbed coplanar to the Cu(001) surface (Fig. 5b). SH110 is Vertically adsorbed on the Cu(101)surface of the copper crystal by the DHT moiety SH110 is adsorbed coplanar to the copper surface when in equilibrium, but at three points on the molecule: the DHT moiety, the S-S bond, and the -SO 3 − group. The DHT moiety is oriented closest to the surface of the copper, making this group the dominant contributor to this initial (but slight) adsorption. The equilibrium adsorption of S-S and -SO 3 − groups on the copper surface contribute to the acceleration associated with SH110 as an additive. Meanwhile, the DHT group adsorbed to the copper surface contributes to SH110 having an inhibitory effect. Thus, we can conclude that the DHT, S-S, and -SO 3 − groups are crucial to SH110 adsorption behaviour, which results in fully filled TSVs. ## Quantum chemistry calculation. Equilibrium geometry structure. The initial and equilibrium structures of SPS were obtained from MD simulations. To further study how the additives were adsorbed on the copper surface, the initial and adsorbed equilibrium structures of SPS and SH110 were obtained by the DFT B3LYP method, as shown in Figs. 6 and 7. For SPS in Fig. 6, atoms 1, 5, 6, and 10 are sulfur, while 11-16 are oxygen, and 2, 3, 4, 7, 8, and 9 are carbon. For SH110 in Fig. 7, atoms 2, 3, 8, and 11 are sulfur, 12-14 are oxygen, 5 is nitrogen, and 1, 4, 6, 7, 9, and 10 are carbon. The optimised bond lengths of SPS and SH110 in initial and equilibrium conditions on the Cu surface were also obtained. For SPS (Table 2), most bonds became longer in equilibrium compared to the initial state, except for C9-S10. It should be noted that the S-O bonds changed the most (~ 5%); this suggests that the SO 3 − group strongly interacts with the Cu surface. The changes in bond lengths that occur for SH110 from initial www.nature.com/scientificreports/ to equilibrium adsorption are also given in Table 2. For instance, C4-N5 shortens by ~ 13% upon adsorption, suggesting a strong interaction between the DHT moiety and the copper surface. The C10-S11 bond length also shortens by ~ 6% upon adsorption, suggesting that the interaction between the MPS group and copper is also important. Active sites. To reveal how the groups on each additive interact with the copper surface, an active site was established according to two factors: (1) natural atomic charge and (2) distribution of the frontier molecular www.nature.com/scientificreports/ orbitals 17 . This was accomplished by using quantum chemical calculations of the orbital information and the electronic properties. ## Natural atomic charge Table 3 shows the natural atomic charges of SPS and SH110 initially and at equilibrium. For SPS, O11, O12, O14, and O15 carry larger negative charges in the initial state than at equilibrium, and all other atoms exhibit very little change from the initial to equilibrium states. This suggests that O11, O12, O14, and O15 have negative charge centres that could offer electrons to the Cu atoms to form coordinate bonds that are initially strong, but weaken at equilibrium. S1 and S10 carry large positive charges both in the initial state and at equilibrium, indicating that they possess positive charge centres that can accept electrons from the 3d orbital of the Cu atoms to form a feedback bond in both states, although this bond would also be weaker in equilibrium than the initial state. This suggests that SO 3 − is the driving force of the electrostatic interactions and chemical properties of SPS interacting with the Cu surface. This is one reason SO 3 − is often present in compounds used as accelerators. For SH110, N5, O13, and O14 carried larger negative charges in the initial state than at equilibrium. This indicates that N5, O13, and O14 are negative charge centres that can offer electrons to the Cu atoms to form coordinate bonds. Moreover, the bonds with N5 and O13 are stronger at equilibrium than in the initial state. S11 carries a positive charge, which indicates a positive charge centre that can accept electrons from the 3d orbital of the Cu atoms to form coordinate covalent bonds. These S11 bonds are also stronger at equilibrium than in the initial state, further strengthening the interaction of SH110 with the Cu surface. ## Distribution of the frontier molecular orbitals Figure 8 shows the highest occupied and lowest unoccupied molecular orbitals (HOMO and LUMO, respectively) of SPS in both the initial and equilibrium states. We can see that SPS essentially comprises two MPS molecules. The main components of the molecular HOMO and LUMO orbitals are listed in Table 4. In the initial stage of adsorption, S5, S6, O11, O12, and O13 contribute 86.2% to the HOMO, with S6 contributing the most (50.50%). This indicates that the atoms in S-S and SO 3 − play a major role in governing the chemical reactions and can interact strongly with the copper surface. The LUMO is mostly comprised of contributions from C8, C9, S10, O14, O15, and O16, at 9.40%, 29.52%, 13.26%, 10.74%, 13.55%, and 7.79%, respectively. This indicates that SPS can accept electrons from the 3d orbitals of Cu atoms, thus further strengthening the interaction between SPS and the copper surface. After adsorption, the frontier molecular orbitals are redistributed. The HOMO is mainly comprised of O15, C9, and S5. This indicates that O15, C9, and S5 can offer electrons; O15 has the maximum electrophilic electron density for charge transfer (76.64%). Thus, SO 3 − strongly adsorbs to the copper surface. S1, C3, C4, S5, S6, O11, O12, and O13 contribute 90% of the LUMO, indicating that electrons can be accepted from the 3d orbitals of Cu atoms, further strengthening the interaction of SPS with the copper surface. SPS mainly interacts with the copper surface through the S and SO 3 − groups. This may be the reason that SPS is an excellent additive for acceleration. Figure 9 shows the HOMO and LUMO orbitals of the initial and equilibrium states of SH110. The main components of the HOMO and LUMO orbitals are listed in Table 5. In the initial stage of adsorption, N5 and S8 contribute over 87.74% to the HOMO. Thus, N5 and S8 can offer electrons. In addition, the DHT moiety plays www.nature.com/scientificreports/ a major role in governing chemical reactions and can interact strongly with the copper surface. The LUMO is mainly comprised of C1, S2, C9, C10, S11, O12, O13, and O14. This indicates that all these atoms have some ability to accept electrons from the 3d orbitals of Cu atoms, thus further strengthening the interaction of SH110 with the copper surface. DHT is dominant in the initial stage of adsorption (as shown in Fig. 5), while SO 3 − has no ability to donate electrons, which differs considerably from SPS. After adsorption to the copper surface, the frontier molecular orbitals are redistributed. N5 and S8 from DHT contribute over 95.49% to the HOMO in equilibrium, while the atoms associated with the MPS moiety (C9, C10, S11, O12, O13, and O14) do not www.nature.com/scientificreports/ contribute at all to the HOMO. We can therefore conclude that the DHT moiety of SH110 donates electrons to Cu atoms and becomes strongly adsorbed to the copper surface. At equilibrium, N5 and S8 of the DHT moiety can interact more strongly with the copper surface, enabling SH110 to cover the copper surface and prevent further deposition. This may be the reason that SH110 is an excellent additive that provides an inhibitory effect. The MPS moiety (C1, S2, S3, C9, C10, S11, O12, O13, and O14) contributes to the LUMO, indicating that this portion of SH110 can accept electrons from the 3d orbitals of Cu atoms, further strengthening the interaction of SH110 with the copper surface. This is likely the reason that SH110 behaves as an excellent additive for acceleration. Therefore, SH110 is unique in that the DHT moiety provides an inhibitory effect for TSV filling, while the MPS moiety acts as an accelerator. SH110 exhibits an optimal balance between acceleration and suppression that allows for void-free TSV filling. ## Materials and methods All chemical additives of SH110 and SPS were purchased from Jiangsumengde. SH110 and SPS are unstable in their acid forms, whereas the sodium salts are stable. Therefore, in the current study, we utilised the sodium salts of SPS and SH110 in acidic solution to obtain the corresponding acid in situ; the sodium salts were used for MD simulations as well. The acidic and sodium salt forms of SH110 contain S-S bonds, SO 3 − groups (with H + or Na + , respectively), and DHT within their frameworks (Fig. 10a-c). Accelerators commonly contain S-S and SO 3 − 18 , and both are present in the acidic and sodium salt forms of SPS (Fig. 10d,e), which is a valuable accelerator 10, . Notably, DHT is commonly added during TSV filling as a leveller 7,8 , so the SH110 additive is expected to provide both accelerator and leveller properties. The MPS molecule in Fig. 10f corresponds to one half of a SPS molecule. Electroplating. The primary solution prepared for electroplating consisted of 0.78 mol/L CuSO 4 •5H 2 O, 0.2 mol/L H 2 SO 4 , and 0.2 g/L KCl. A single-component additive (SH110 or SPS) was then added to the primary solution without a suppressor or leveller. A silicon chip with TSV (diameter 20 μm and depth 60 μm) was used for electroplating. A blind via was etched by Bosch-type deep reaction ion etching (DRIE). A silicon dioxide layer of approximately 0.3 μm was deposited for sidewall insulation using tetraethoxysilane chemical vapor deposition (TEOS CVD), and then a copper seed layer was deposited through a physical vapor deposition (PVD) process. During electrodeposition, the TSV sample was first immersed in deionized water in a vacuum chamber for 10 min, and then placed in an ultrasonicator for 60 s to remove air and impurities. After this pre-processing, the TSV sample was immersed in the plating solution for 10 min to allow the equilibration state to be reached between the additive and copper in the vias prior to electroplating. TSV filling was carried out with either SH110 or SPS additive in the primary solution, at a current density of 1 mA/cm 2 . After electrodeposition, the cross-section and microstructure of the TSVs were observed using SEM at 2000 × magnification (TESCAN MIRA3LMU). Electrochemical procedures. To study the electrochemical properties of the additives in the deposition process for TSV filling, LSV 19,20 was performed on an electrochemical workstation (Chenhua CHI660E) using a three-electrode cell. The cell consisted of a 5-mm-diameter rotating Pt disk electrode (Pt-RDE) as the working electrode (WE), a Pt counter electrode (CE), and a saturated calomel electrode (SCE) as the reference. The testing potential ranged from 0.6 to 0.7 V versus Pt-RDE with a scan rate of 5 mV/s. The LSV measurements were performed using a potentiostat/galvanostat (PARSTAT 2273) at 25 °C.

chemsum

{"title": "Experiment and simulation of single inhibitor SH110 for void-free TSV copper filling", "journal": "Scientific Reports - Nature"}

supramolecular_reactions_of_metallo-architectures:_ag₂-double-helicate/zn₄-gri

## Abstract: Supramolecular reactions are of importance in many fields. We report herein three examples where complexes of hydrazone-based ligands are involved. A Ag 2 -double-helicate was converted, by treatment with Zn(OTf) 2 , into a Zn 4 -grid (exchange of metal ions and change of the nature of the initial complex). A Pb 4 -grid was converted, upon reaction with ZnCl 2 or ZnBr 2 , into a Zn 4 -grid (exchange of metal ions, but conservation of the nature of the initial complex). The reverse conversions were also achieved. The fusion of a Ag 2 -double-helicate with another Ag 2 -double-helicate was performed (exchange of ligands, but conservation of the nature of the complexes) and resulted in a mixture of three helicates (two homostranded ones and one heterostranded one). ## Introduction Like covalent molecules, supramolecular 1 assemblies may participate in various reactions. The understanding of supramolecular reactions is of much interest because they are involved in many areas such as complex chemical systems and networks, 2 adaptive 3 and stimuli-responsive 4 chemical systems, fabrication of nanodevices and materials, 4 biomolecular processes. Thus, in the complexity and diversity of supramolecular chemistry, the reactivity of supramolecules plays a crucial role. It includes the processes: (a) of (self)assembly (i.e. formation of supramolecular architectures through assembly, but also their participation, as subunits, in more complex assemblies), and correlatively, partial or total disassembly; (b) of partial or total reorganization or exchange (at the supramolecular and, additionally and possibly, at the covalent level), that involves the breaking of several or all of the initial supramolecular connections and formation of new ones; (c) without breaking or formation of new supramolecular connections (e.g. covalent modifcations after self-assembly 5 ). Amongst supramolecular architectures, double helices and helicates, 6 as well as grids 7 arouse much interest and work. For example, DNA 8 and the ion channel generated by gramicidine 9 have a double helical structure, and there are double helical complexes that act as molecular machines 10 or catalysts. 11 Gridlike complexes have been studied for their electrochemical and magnetic properties, 7 for their capacity to encapsulate ions 12 or as starting materials for building more complex architectures (e.g. a Solomon link 13 ), amongst other things. However, supramolecular interconversions of grids and helicates have not, except several examples, 14,15 been much explored. With these ideas in mindand using principles such as the displacement of an equilibrium through precipitation, and the preference of Ag + for tetrahedral and of Zn 2+ for octahedral coordinationwe designed, as reported herein, three supramolecular reactions 16 of reorganization and exchange (Fig. 1) involving grids and double helicates. They are related through the ligands 17 (which are pyrimidine-bis-hydrazones; 18 Fig. 2) that produce the supramolecular complexes, as well as through the nature of complexes, and occur due to the dynamic character of the present metal-ligand connections. These reactions (Fig. 1) can be seen as: (i) a change of the nature of the supramolecular architecture, from a Ag + dinuclear double helicate (DH) into a Zn 2+ tetranuclear grid (G), induced by replacement of Ag + by Zn 2+ (Fig. 3). In this reaction, not only the nature of the complex and that of the metal ion change, but also the conformation of the ligand (helical / unfolded), the charge (2 + / 8 + ) and the nuclearity of the complex (2 / 4) and the number of ligands per complex (2 / 4). In regard to this last change, this process can be compared with the conversion or the equilibrium between supramolecular dimer and tetramer of bioactive proteins, 19 or between other homo-oligomers 20 with influence on the protein functions. (ii) a substitution 21 (metal ion exchange or transmetallation), in a sole operation, of the four Pb 2+ ions of a grid-like 22 complex by Zn 2+ ions (Fig. 4); (iii) a fusion (conproportionation) 23 between two Ag + double helicates 24 (Fig. 5). While in case (ii) the equilibrium is shifted towards the Zn 2+ grid through the precipitation of Pb 2+ as its halides (chloride and bromide), in cases (i) and (iii), the conversions can be done without precipitation. way, Ag + induces the formation of double helicates with ligands 1 and 2. Zn 2+ prefers an octahedral coordination environment that results from 2 three-Nsp 2 -atom tridentate sites of type pyridine-hydrazone-pyrimidine, thus generating a grid. ## Results and discussion Reaction of 1 equiv. of Ag 2 L 2 -DH 15c with 2 equiv. of Zn(OTf) 2 (OTf ¼ CF 3 SO 3 ) produceswithout the need to precipitate Ag + as a halidethe corresponding grid Zn 4 L 4 -G 15b,22a (solvent: CD 3 NO 2 with 6-14% CD 3 CN; ESI, pp. S9-S11 †). Where ZnCl 2 is used in the reaction with Ag 2 1 2 -DH, two equivalents of AgOTf per equiv. of DH are required according to the equation (ESI p. S8 †): On treatment of the double helicate Ag 2 2 2 -DH in CD 3 NO 2 with 2 equiv. of Zn(OTf) 2added as a solution in a small volume of CD 3 CN, or as a solidthe grid Zn 4 2 4 -G was obtained. When the double helicate Ag 2 1 2 -DH in CD 3 NO 2 was treated with 2 equiv. of Zn(OTf) 2 , added as a solution in a small volume of CD 3 CN (about 6-14% of the CD 3 NO 2 volume), the grid Zn 4 1 4 -G was obtained. When Zn(OTf) 2 was added as a solid, without CD 3 CN, was obtained a mixture without the Zn 4 1 4 -G grid; addition of a small volume of CH 3 CN (about 6-14% of the CD 3 NO 2 volume) to this mixture produced the expected grid Zn 4 1 4 -G. A possible explanation could be that, in the case of the reaction Ag 2 1 2 -DH / Zn 4 1 4 -G, the CH 3 CN acts as a coordinating species for the Ag + ions and so contributes to the displacement of the equilibrium from the double helix towards the grid. The grid Zn 4 2 4 -G should bedue to the p-stacking aromatic interaction between a phenyl ring and the two ligands between which that phenyl is located within the gridmore stable than the grid Zn 4 1 4 -G. This stability may be sufficient to make possible the formation of the grid Zn 4 2 4 -G from the corresponding double helicate without, unlike in the case of the grid Zn 4 1 4 -G, the assistance of CH 3 CN. DOSY NMR was also used to study the conversion Ag 2 L 2 -DH / Zn 4 L 4 -G (L ¼ 1, 2). As expected, the volume of the grid species obtained from double helicates on treatment with Zn(OTf) 2 was found in agreement with that of the grid prepared from the free ligands L and Zn(OTf) 2 . The reverse conversion Zn 4 L 4 -G/ Ag 2 L 2 -DH can be done as follows: after treatment of the grid with KOH, the solvent (CD 3 CN or CD 3 NO 2 ) is removed, and the ligand is extracted with CDCl 3 and separated from the solid residue (by centrifugation or fltration); after removal of CDCl 3 , CD 3 NO 2 is added, then AgOTf is added to form the helicate. In order to simplify the procedure, we used ligand 2 and a mixture of CDCl 3 and CD 3 NO 2 where ligand 2, as well as the corresponding grid and double helicate were soluble. After precipitation of Zn 2+ with KOH, the mixture was centrifuged (the ligand 2 being soluble in the mixture of solvents), and to the recovered liquid phase AgOTf was added to produce the Ag 2 2 2 -DH (ESI, p. S13 †). In a pH-dependent system (Fig. 3b), the interconversion between Ag 2 1 2 -DH and Zn 4 1 4 -G was achieved as follows (ESI, p. S10 †): the grid was generated from the double helicate by reaction with Zn 2+ ; then, Zn 2+ was complexed with hexacyclen, and the double helicate was regenerated; partial protonation of hexacyclen with TfOH caused release of Zn 2+ and formation of the grid (incomplete yield); fnally, addition of triethylamine reactivated the hexacyclen that again encapsulated Zn 2+ and resulted in the reformation of the double helicate. (ii) The Pb 4 1 4 -G / Zn 4 1 4 -G conversion (Fig. 4a) can formally be seen as a substitution of Pb 2+ by Zn 2+ ions, although the real mechanism, involving breaking and formation of supramolecular bonds, must be more complex. Reaction of Pb 4 1 4 -G 15b with 4 equiv. of Zn(OTf) 2 produces a mixture which no longer contains the grid-like species Pb 4 1 4 -G or Zn 4 1 4 -G (ESI p. S2 †). This suggests that the affinity of Zn 2+ for the ligand, as well as its preference for octahedral coordination are not sufficient to displace the equilibrium towards Zn 4 1 4 -G. We considered that the involvement of Pb 2+ ions in a weakly dissociating or sparingly soluble compound should displace the equilibrium. Indeed, addition of Br (as Bu 4 P + Br ) to the above mixture, or treatment of Pb Thus, in addition to its self-assembly from Zn 2+ and a ligand, the same Zn 2+ grid, Zn 4 1 4 -G, can be obtained, in reactions (i) and (ii), from a Ag + dinuclear double helicate or from a Pb 2+ tetranuclear grid (exchange of metal ions and reorganization of the architectures). (iii) The fusion (conproportionation) reaction of double helicate Ag 2 1 2 -DH 15c with 1 equiv. of Ag 2 2 2 -DH (Fig. 5) according to the equation conversion, where the exchange of metal ions changes the nature of the metallo-supramolecular architecture, (ii) a Zn 4 1 4 -G grid into Pb 4 1 4 -G grid conversion driven by a halide-induced precipitation and where the nature of the metallo-supramolecular architecture is conserved, and (iii) a double exchange of ligands during the fusion of two double helicates. The grid/grid and double-helicate/grid conversions were made reversible by precipitation of Zn 2+ with KOH and subsequent reaction of the free ligand with Ag + or Pb 2+ , or, for one DH/G interconversion, in a pH-dependent way. In perspective, such ligands could be introduced in larger and more complex, suitably decorated, architectures where such supramolecular reactions can act as actuators of various properties (charge, volume, multivalency).

chemsum

{"title": "Supramolecular reactions of metallo-architectures: Ag₂-double-helicate/Zn₄-grid, Pb₄-grid/Zn₄-grid interconversions, and Ag₂-double-helicate fusion", "journal": "Royal Society of Chemistry (RSC)"}

screening_for_generality_in_asymmetric_catalysis_one-sentence_summary:_a_new_method_for_high-through

## Abstract: Research in the field of asymmetric catalysis over the past half century has resulted in landmark advances, enabling the efficient synthesis of chiral building blocks, pharmaceuticals, and natural products. A small number of asymmetric catalytic reactions have been identified that display high selectivity across a broad scope of substrates; not coincidentally, these are the reactions that have the greatest impact on how enantioenriched compounds are synthesized. We postulate that substrate generality in asymmetric catalysis is rare not simply because it is intrinsically difficult to achieve, but also because of the way chiral catalysts are identified and optimized. Typical discovery campaigns rely on a single model substrate, and thus select for high performance in a narrow region of chemical space. Here, we put forth a practical approach for using multiple model substrates to select simultaneously for both enantioselectivity and generality in asymmetric catalysis from the outset. Multisubstrate screening is achieved by conducting high-throughput chiral analyses via supercritical fluid chromatography-mass spectrometry (SFC-MS) with pooled samples. When applied to Pictet-Spengler reactions, the multi-substrate screening approach revealed a promising and unexpected lead for the general enantioselective catalysis of this important transformation. ## Introduction Since the discovery that chiral phosphine-rhodium(I) complexes catalyze the highly enantioselective hydrogenation of certain dehydroamino acids (1,2), asymmetric synthesis with small-molecule catalysts has been demonstrated in a dazzling variety of contexts (3). It is now widely appreciated that high enantioselectivity, typically defined as >90% enantiomeric excess (ee), is often attainable for specific model substrates in a reaction of interest. While these model substrates may be the targets of a specific synthetic campaign, they are more frequently chosen on the basis of accessibility, ease of chiral analysis, lack of peculiar structural features, or similarity to substrates studied successfully in related reactions. To this day, high enantioselectivity remains the sine qua non of asymmetric catalysis development efforts; it is only after that condition is met with a model substrate that the substrate scope and limitations of the reaction are evaluated (Fig. 1A). Given the truism that "you get what you screen for" (4), an unintended consequence of this paradigm is that optimization in this manner fundamentally selects for success with substrates that are similar to the model. In rare cases, asymmetric transformations with broad substrate scope do emerge, and those are typically the ones that are most impactful as they can be applied predictively in new contexts (5)(6)(7)(8)(9). But such generality is effectively accidental. In the vast majority of cases, the scope is limited, researchers strain to identify enough high ee examples to fill the "Substrate Scope" table requisite for publication, and the methods remain underutilized because synthetic practitioners shy away from trying unproved or unpredictable chemistry. Optimization against multiple, diverse substrates simultaneously rather than against a single model substrate would shift the focus in asymmetric catalysis discovery efforts from identifying circumscribed examples of high enantioselectivity to revealing more general solutions (Fig. 1B). This approach was proposed in 1999 by Kagan et al. (10,11) and articulated compellingly in recent studies by MacMillan et al. (12) and List et al. (13), but its adoption has been largely precluded by the challenges associated with conducting large numbers of chiral analyses on a variety of products. Even though methods for ee determination have advanced substantially over the past several decades, their development and application remain laborious and time-consuming. Accordingly, researchers will only commit to the laborious task of developing and applying ee-determination methods for multiple products after success with a model substrate has been achieved. The most commonly applied analytical methods for ee determination involve high-performance liquid chromatography (HPLC) or supercritical fluid chromatography (SFC) with chiral stationary phases (CSPs) in conjunction with isocratic elution and UV-Vis detection (Fig. 1C). Such methods are inherently limited in throughput because of the requirement for baseline separation and the removal of interfering signals before analysis. While there has been long-standing interest in developing high-throughput chiral analysis methods, the techniques identified to date have been tailored for repeated analyses of specific analyte classes of interest. For instance, the Anslyn and Wolf groups have developed circular-dichroism-based sensors that enable ee determination for compounds containing chelating functional groups (14,15). Other approaches have employed chiral 19 F-NMR shift reagents (16), fluorescent DNA biosensors (17), mass-tagging with pseudo-enantiomers (18), and selective enzymatic oxidation (19,20). Although these methods can provide high accuracy and sample throughput, more general analytical strategies are needed to conduct effective multi-substrate screening across a range of chemical space. We envisioned that combining conventional chromatographic techniques and sample pooling might allow facile and simultaneous ee determination for a diverse panel of products. In particular, we investigated the coupling of SFC, which offers very short analytical runtimes and excellent resolution for most substrate classes, with mass spectrometry (MS) as the detection method (21). By choosing substrate combinations that generate products of differing mass, combining aliquots drawn from multiple independent reactions into a single analysis vial, and subjecting the sample to SFC-MS and extracted ion chromatogram (EIC) analysis, we anticipated that it would be possible to obtain many ee measurements simultaneously (Fig. 1D). This strategy of using SFC-MS to analyze pooled samples has many attractive features. First, SFC already enjoys widespread use and has been applied successfully to the rapid analysis and separation of a wide variety of compounds (21). By applying MS as the detection method, it is possible to use an elution gradient; a single, generic gradient can be applied to analytes possessing a wide range of polarities, greatly simplifying analytical method development (22). Recent advances in immobilization strategies for polysaccharide chiral selectors have given rise to commercially available CSPs with excellent separation properties and improved robustness, increasing the possibility of analyzing crude reaction mixtures of a wide variety of products using only a small number of columns (23). Finally, the EIC analysis would enable ee determinations when the product enantiomers are separated but co-elute with other products or residual reaction components. Overall, we anticipated that the combination of SFC-MS and sample pooling would raise the throughput of ee determination to the point where it may become practical to optimize directly for both enantioselectivity and generality by allowing high-throughput analysis of multiple crude reaction mixtures simultaneously. ## Fig. 1. Approaches to the discovery and analysis of enantioselective reactions. A) The standard approach to discovery of new asymmetric catalytic reactions involves optimization around a single model substrate. The scope of the method is then examined in a separate exercise, often resulting in methods that are only highly effective for substrates similar to the model. B) Optimization of asymmetric transformations via multi-substrate screening improves the chances of identifying more general catalysts and conditions, particularly if screening is performed across a broad cross-section of substrate space. C) Conventional ee determination of single isolated products via chromatography on CSPs and detection by UV-Vis. D) Proposed approach to simultaneous, highthroughput ee determination of multiple products. Crude reaction mixtures are pooled and analyzed by SFC-MS. Signals due to resolved products are detected without interference by other products or other reaction components of different mass. ## Methods The performance of SFC-MS for rapid ee determination of single and pooled samples was first assessed with a set of commercially available compounds using standard chromatographic columns. When pure samples of known enantiomeric composition were analyzed at high concentrations (10 mM), an unacceptably large rootmean-square error (RMSE) of 22% resulted, with the greatest deviations for scalemic mixtures (Fig. S2). A major source of error originates from the nonlinear relationship between detector response and concentration, which is exacerbated at high signal intensities and therefore results in underestimations of the ee (Fig. S2). However, the signal dependence on concentration approaches linearity at lower sample concentrations (<0.1 mM, Fig. S3), and the RMSE in the analyses could thus be reduced to 3%, which is comparable to standard HPLC-UV conditions (24). Consequently, all subsequent samples were analyzed at the lowest concentrations that still afforded good signal-to-noise (typically 0.1 mM, assuming 100% theoretical yield), a practice that was convenient for high-throughput assays of reactions carried out at micromole-scale. Given the intention to carry out analyses of multiple compounds rapidly and simultaneously with the fewest possible injections, we recognized that baseline separation of all enantiomeric pairs would not always be achieved. We therefore sought to develop a general peak-fitting method to extract accurate integrations from partially separated enantiomers. Fitting the experimental data of baseline-resolved peaks produced from scalemic mixtures of known composition to linear combinations of peak functions revealed that the SFC-MS peak shapes were not well reproduced by any single known peak functions. However, a "Frankenstein" model involving a piecewise combination of Gaussian and Voigt functions was found to provide excellent fits (Fig. S1). We combined this peak model with parameter-fitting methods in a convenient web-based application for analyzing chromatographic data. Accurate ee determinations were obtained when the protocol was applied to the analysis of the same scalemic standards on columns that produced only partial separation of the enantiomers, even in the challenging cases of high-ee samples where the first major peak tailed into the second minor peak (Figs. S4-S5). With a rapid and accurate SFC-MS ee-determination method in hand, we evaluated its accuracy for analysis of pooled samples. A mixture of 20 commercially available racemic compounds (mostly pharmaceuticals; Table S3) at low concentration was analyzed with five chromatographic columns, registering ees of -1±7%, with several significant outliers (Fig. 2A). The origin of these outliers was identified upon inspection of the total ion chromatograms displayed as two-dimensional heatmaps (Fig. 2B). The bright horizontal lines represent unavoidable contaminants (e.g., polyethylene glycol) and did not interfere with the analyses. However, the dark vertical bands are the result of ion suppression by strongly ionizing species, which results in reduced intensity for any co-eluting ions (25,26). If one enantiomer falls under such a band and its partner does not, the apparent ee is distorted in favor of the unaffected peak. This effect is illustrated in pairwise experiments with racemates (Fig. 2C). Recognizing the need to minimize ion suppression and increase the capacity of our workflow to handle large panels of products, we introduced an alternative pooling scheme. Rather than combining all compounds into a single vial and injecting the entire pool onto each column ("simple pooling"), we pooled subsets of products into multiple vials and injected each vial only once ("smart pooling", see SI for details). This latter procedure reduces ion suppression, avoids the unnecessary injection of compounds onto columns that do not separate them, and was applied in the screens discussed below. ## Results and Discussion We selected asymmetric catalysis of the Pictet-Spengler reaction as a timely and synthetically relevant platform to illustrate the application of the new high-throughput enantioselectivity determination methodology and the "screening for generality" concept. The condensation of tryptamines with aldehydes or ketones to generate tetrahydro-ꞵ-carbolines (Fig. 3A) is a venerable reaction with crucially important applications in laboratory and biological synthesis (27,28). The reaction has inspired intensive research efforts in search of asymmetric catalytic variants, and so far over a dozen distinct catalytic systems have been described (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43). Each study relied on optimization of catalyst and conditions around very limited numbers of model substrates, and resulted in the identification of highly enantioselective reactions. By traditional standards, this output of new catalysts, high ee examples, and publications in high-impact journals can certainly be viewed as a major success. Yet despite this apparent progress, none of the published methods has found widespread application, and the chemist interested in carrying out an enantioselective catalytic Pictet-Spengler reaction on a never-before-tested substrate combination would be hard-pressed to know which system to try. We sought to establish whether screening across broad stretches of substrate space might prove informative in that regard and possibly enable the identification of general systems. Selection of the panel of model substrates represents a key step in any screening-for-generality exercise, and we sought to maximize the structural and functional diversity of aldehyde and tryptamine combinations. We constructed an in silico library of 340 potential tetrahydro-ꞵ-carboline products, generated molecular fingerprints for each member, and performed UMAP dimensionality reduction to generate a two-dimensional representation of the diversity of potential products (Fig. 3B) (44). By combining insights gained from this representation with practical considerations such as reactivity, ease of separation, and commercial availability, we winnowed the candidates down to a panel of 14 representative products (Fig. 3A). Notably, these products encompass regions of chemical space that are largely unexplored by reported methodologies. Pictet-Spengler reactions can proceed via N-acyl-, N-protio-, and N-alkyl-iminium ion intermediates, and enantioselective catalytic variants have been identified for each of these manifolds (29)(30)(31)(32)(33)(34)(35)(36)(37)(38)(39)(40)(41)(42)(43). We selected reactions between N-benzyl tryptamines and aldehydes as a particularly convenient platform to survey the substrate/catalyst landscape. In particular, reactions with N-benzyl tryptamines were found to remain homogenous, enabling the reactions to be run in a format that is highly amenable to parallel screening with standard equipment (96 well plates, no stirring, 0.01 mmol scale). Acceptable agreement between the enantioselectivity values obtained by single sample SFC-UV and pooled sample SFC-MS was achieved for the resulting N-benzyl-tetrahydro-ꞵ-carboline products (8% RMSE, Fig. S32). ## Fig. 3. High throughput ee-determination of enantioselective catalytic Pictet-Spengler reactions (A) The 14-member panel of products (left) used to study the Pictet-Spengler reaction and a map (right) of potential products (grey) with previously reported products from the literature (blue) vs. our panel (red). (B) Enantioselectivity screen using 14 previously reported organocatalysts against the 14-member panel. Reactions with weakly acidic H-bond-donor catalysts i-ix and xiii-xiv were run with benzoic acid as a co-catalyst. Empty squares represent low-yielding reactions. We evaluated a set of previously reported chiral Brønsted-acid and H-bond-donor catalysts across the selected panel. The resulting ee data (Fig. 3B) reveal that different catalyst classes respond very differently to variations in substrate. In some cases, such as thioureas i-vii, moderate (e.g., 20-40%) enantioselectivity was observed fairly consistently across most substrates. Other catalysts afford higher (e.g., >60%) enantioselectivity, but only for specific subsets of the substrate space; for example, Miller squaramide ix is particularly effective in catalyzing the reaction between neutral indoles (X=Y=H) and aryl aldehydes, whereas the SPINOL-phosphoric acid xii reported by Lin and Wang et al. stood out in reactions of the electron-deficient indole (X=Cl, Y=H) (34,43). Tabulation of the data as a heatmap (Fig. 3B) provides a visually straightforward tool for identifying correlations in behavior between related catalysts. For instance, chiral phosphoric acids xi and xii perform poorly with the electron-rich indole (X=H, Y=OMe), and generally much better with the electron-poor or neutral tryptamine analogs. Similarly, comparison of ee values between rows reveals correlations between substrates. For example, compound 23, a prototypical minimally functionalized model product derived from N-benzyl tryptamine and benzaldehyde, responds similarly to catalyst effects as do products 24 and 28 also possessing the neutral indole. However, product 23 is a very poor model for products possessing electron-rich indoles such as 11, 12, 16, or 19. These findings highlight the risks associated with optimizing around a single model substrate combination and the value of employing multi-substrate screening for the accurate assessment and optimization of a given methodology. To assist in quantifying the level of generality displayed by each catalyst, we constructed a generality metric g that summarizes a collection of enantioselectivity values into a number between 0 and 1, where 1 represents a completely general catalyst that induces 100% ee in every reaction surveyed (see SI Section 3.9 for further discussion). By this analysis, catalyst xii stands out among those surveyed as the most promising from a generality standpoint. Despite the encouraging results obtained with catalyst xii, it is apparent from the data in Fig. 3B that reactions of the electron-rich indole are particularly challenging for that, and indeed all catalysts in the screen. We zeroed in on that substrate class by evaluating several solvent-catalyst combinations for product 16 (Fig. S40). Improved results were obtained with polar aprotic compounds such as 2-methyl THF (2MT) and ethyl acetate (EA), prompting us to evaluate these solvents across the entire substrate panel (Fig. 4A). Notably, the solvent effects were highly catalyst-dependent, with enhanced enantioselectivities observed in reactions performed in 2MT for chiral phosphoric acid xii, but no systematic improvement observed with H-bond donor catalyst i. The substantial benefits of 2MT were possibly missed in the original work that led to the development of catalyst xii (34) because the authors used only a single model product, similar to 23 and related N-protected analogs. While changing from PhMe to 2MT leads to a small decrease in the enantioselectivity of 23, most other substrates benefit substantially, with the increase from 1% to 62% ee for product 11 serving as a particularly notable example. These results illustrate how performing screens on multiple substrates simultaneously can provide valuable and otherwise elusive insights. When xii was used to generate product 6 in a standard, rather than high-throughput, experimental format, a slightly improvement in enantioselectivity from 86 to 91% ee was observed (Fig. 4B), thereby demonstrating the successful translation of the HTE-based screens and analyses to laboratory-scale reactions. Relative to traditional ee-determination methods, the analytical workflow outlined in this study trades a modest decrease in accuracy for the ability to analyze multiple crude reaction mixtures simultaneously with substantial reductions in method development and analysis time. As a result, optimization of catalyst structure and reaction conditions is more readily performed across a variety of substrate combinations, thereby increasing the chances of identifying general protocols. If no broadly general protocols are found for a given reaction of interest, the broad survey of the catalyst/substrate landscape can be used to ensure that highly effective catalystsubstrate combinations are not missed, and to identify "islands" of high enantioselectivity involving specific catalyst structures and subsets of the substrate space. The survey of the substrate/catalyst landscape in the asymmetric Pictet-Spengler reaction reveals how the standard approach of basing optimization campaigns on a single model system has indeed resulted in substrate-specific islands of high enantioselectivity for this reaction. However, an unexpected finding from this study is that there is already a family of catalysts (exemplified by xii) that displays very promising levels of generality across a wide range of substrate combinations. The uniquely interesting features of xii was likely masked by the fact that it was identified through an evaluation of a fairly narrow region of substrate space (34). Similarly, the beneficial effect of 2-methyl THF as a solvent in reactions with catalyst xii was masked by the anomalous behavior of the prototypical model substrates for that reaction. Thus, the use of multi-substrate screening revealed a promising lead for the development of a highly general and enantioselective Pictet-Spengler reaction, and will hopefully prove useful in the identification and discovery of general chiral catalyst systems for other transformations of interest.

chemsum

{"title": "Screening for Generality in Asymmetric Catalysis One-Sentence Summary: A new method for high-throughput enantioselectivity determination on diverse substrates enables an intentional approach for the identification of broadly effective asymmetric catalysts", "journal": "ChemRxiv"}

supramolecular_assemblies_of_a_nitrogen-embedded_buckybowl_dimer_with_c₆₀

## Abstract: A directly connected azabuckybowl dimer was synthesized via a palladium-catalysed C-H/C-Br coupling.The electron-donating nature of the pyrrolic nitrogen atoms of the azabuckybowl enabled a strong complexation with pristine C 60 . In the presence of two equivalents of C 60 , the azabuckybowl dimer formed crystals with a 1 : 2 stoichiometry. Conversely, in diluted solution, complexes with a 1 : 1 stoichiometry of the dimer and C 60 were detected predominantly, and these precipitated upon increasing the concentration of C 60 . Scanning electron microscopy images of the precipitate showed fibre-like aggregates, indicating the formation of supramolecular assemblies with 1D chain structures. A variable-temperature 1 H NMR analysis revealed that the precipitate consists of the dimer and C 60 in a 1 : 1 ratio. ## Introduction Supramolecular assembly is defned as higher-order aggregates constructed from two or more components, in which the molecules interact by non-covalent interactions such as hydrogen bonding, metal coordination, or hydrophobic interactions. 1 As these interactions are much weaker than covalent bonds, cleavage of the aggregates easily occurs by adjusting the temperature or concentration, which regenerates the corresponding monomers. Due to this flexibility, supramolecular polymers are expected to act as stimulus-responsive materials. 2 Among all the research in this feld, supramolecular polymers with fullerenes based on host-guest interactions have been extensively studied in recent years (Fig. 1). Previous studies have often employed molecular tweezers hosts to ensure strong binding with C 60 derivatives. These strategies require a C 60 dimer (type a; Fig. 1) or a functionalized C 60 bearing a binding site (type b; Fig. 1). However, supramolecular polymerization with pristine C 60 remains a challenge because two binding units of the molecular tweezers would be needed to capture one C 60 molecule. 7 Consequently, stronger yet sterically less-demanding host molecules are required. Buckybowls are bowl-shaped p-conjugated molecules, for which sumanenes and corannulenes are representative examples. 8,9 Such curved polycyclic aromatic hydrocarbons have been used for the recognition of fullerenes, given that the concave surface of the former efficiently overlaps with the convex surface of the latter. 10 And it is exactly for this reason that extensive research on assemblies of buckybowls with fullerenes has been carried out. 11 However, due to the poor electron-donating nature of these buckybowls, 12 their binding ability is usually insufficient to construct large supramolecular assemblies. Recently, the group of Nozaki and our own group have independently succeeded in the synthesis of nitrogenembedded buckybowls such as penta-peri-pentabenzoazacorannulene 1 (Chart 1). 13 Due to the electron-donating nature of the pyrrolic nitrogen atom, buckybowl 1 exhibited a large association constant with C 60 in solution. The binding constant of 1 was 3800 M 1 in 1,2-dichlorobenzene, which is a top-class value reported for a bowl-shaped molecule. This result suggests that 1 could be used as a new building block for supramolecular assemblies with C 60 . Herein, we disclose the synthesis of buckybowl dimer 2 as a host molecule for pristine C 60 . Owing to its two binding sites, a Department of Molecular and Macromolecular Chemistry, Graduate School of Engineering, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi 464-8603, Japan. E-mail: hiroto@chembio.nagoya-u.ac.jp; hshino@chembio.nagoya-u.ac.jp dimer 2 was expected to form complexes with C 60 in either a 1 : 1 or 1 : 2 ratio. We discovered that 2 acts as a concentration-dependent fullerene host, showing drastic morphological changes in the solid state depending on the number of C 60 molecules that are contained within the structure. In particular, 1D-chain fbre aggregates consisting of 2 and C 60 were obtained. ## Synthesis and characterization In our previous work, 1 was synthesized via the Pd-mediated C-H/C-Br coupling of tribrominated precursor 3 using an excess of palladium(II) acetate and tricyclohexylphosphonium tetra-fluoroborate. Interestingly, the use of catalytic amounts of these reagents provided the linked azabuckybowl dimer 2 in 31% yield (Scheme 1). The structure of 2 was characterized by NMR spectroscopy and mass spectrometry. The parent mass ion peak of 2 was observed at m/z ¼ 1321.7303, which confrms its dimeric structure. The 1 H NMR spectrum of 2 exhibited fve singlet peaks in the aromatic region, consistent with a symmetric structure for 2 (Fig. S1 †). The downfeld shifts of the H a protons (Scheme 1) in 2 compared to those in 1 indicate a deshielding effect by the second azabuckybowl unit. In addition, the 13 C NMR spectrum exhibited 18 peaks assignable to sp 2 -carbons, which suggests a conformation with C 2v symmetry (Fig. S2 †). ## Optical and electrochemical properties Fig. 2 shows the UV-vis absorption and emission spectra of 1 and 2 in CH 2 Cl 2 . Compared to the spectrum of 1, the lowestenergy band of 2 was red-shifted from 472 nm to 495 nm, which indicates the presence of electronic communication between the two azabuckybowl units through the covalent bond. The emission band of 2 was observed at 517 nm with a quantum yield of 0.17, which is almost identical to that of 1. The electrochemical properties of 2 were investigated by cyclic voltammetry (Fig. S3 †), where 2 exhibited a lower oxidation potential (0.18 V) than 1 (0.20 V), commensurate with higher electron-donating properties for 2. ## Titration experiments between 2 and C 60 In order to examine the binding ability of 2 toward C 60 in solution, we carried out titration experiments under diluted conditions (c ¼ 1.3 10 5 M 1 ) (Fig. 3). The titration was conducted in 1,2-dichlorobenzene and monitored by UV-vis-NIR absorption spectroscopy. Upon addition of a solution of C 60 to a solution of 2, an absorption band around 800 nm appeared, which is similar to the behaviour of 1. The Job's plot for the absorbance at 800 nm indicated the predominant formation of 1 : 1 complexes in solution (Fig. S4 †). A nonlinear curve ftting based on a 1 : 1 binding afforded an association constant of 7.8 10 3 M 1 , which is higher than that of 1 (K a ¼ 3.8 10 3 M 1 ). This result corroborates the superior electron-donating nature of 2 relative to that of 1. It should also be noted that the binding constant reached 1.0 10 5 M 1 in toluene (Fig. S5 and S6 †). Such solvent-dependent association constants should probably be attributed to the different solvophobicity of the fullerene in each solvent. 14 X-ray crystal structure and charge-carrier mobility Fortunately, we obtained single co-crystals of 2 and C 60 that were suitable for an X-ray diffraction analysis, which were prepared by vapour diffusion of acetonitrile into a toluene Chart 1 Azabuckybowl 1 and linked azabuckybowl dimer 2. Scheme 1 Synthesis of directly linked azabuckybowl dimer 2. Fig. 2 UV-vis absorption and emission spectra of 1 and 2 in CH 2 Cl 2 . solution of a mixture of 2 and C 60 (Fig. 4). 15 In these crystals, the two azabuckybowl units face in opposite directions, reflected in the tilt angle (50.1 ) between the two buckybowl units around the central bond. The C 60 molecules are coordinated to the azabuckybowl units in a concave-convex fashion, resulting in a 1 : 2 ratio in the crystal. Between the centroid of the pyrrole ring to the closest surface of the C 60 molecules, distances of 3.28 and 3.29 were measured. Such short distances indicate the existence of strong electronic interactions between 2 and C 60 in the solid state. The packing structure is shown in Fig. 4b. Similar to 1$C 60 , 2 and the C 60 molecules present segregate stacking (Fig. S7 †). The photo-induced transient conductivity of 2 and 2$C 60 was determined by flash-photolysis time-resolved microwave conductivity (FP-TRMC) measurements. 16 The carrier mobility of the 2$C 60 crystals (2.0 10 4 cm 2 V 1 S 1 ) is approximately by one order of magnitude higher than that of 2 (Fig. S8 †). The carrier mobility of the 2$C 60 crystals was similar to that of 1$C 60 , indicating a similar charge-separation state between C 60 and the azabuckybowl unit in the crystal. ## Supramolecular assembly of 2 with C 60 To uncover the different binding ratios of 2 with C 60 in solution and the crystalline state, we carried out titration experiments in toluene at 20 C, which were monitored by 1 H NMR spectroscopy (Fig. 5). With increasing amount of C 60 , the spectral shape of the signal peaks became broader and the peak intensities decreased. Notably, a precipitate was formed in the presence of 1.0 equiv. or less of C 60 . This phenomenon was not observed in the case of 1, for which a clear solution was obtained in the presence of C 60 . To clarify this phenomenon, we performed diffusion-ordered two-dimensional NMR spectroscopy (DOSY) experiments. 17 The diffusion coefficient (D) determined for 2 (6.83 10 10 m 2 s 1 ) decreased by 15% (5.80 10 10 m 2 s 1 ) in the presence of 0.5 equiv. of C 60 . In contrast, a reduction of only 2% was observed in the case of 1 (Fig. S9 †). This drop in the D value of 2 indicates the formation of larger structures. The macroscopic structure of the precipitate formed in the presence of C 60 was investigated by scanning electron microscopy (SEM). For that purpose, samples were prepared by dropcasting toluene solutions onto silicon wafers. Fig. 6 displays the SEM images of 2 and 2 with 1.0 equiv. of C 60 . In the precipitate, fbre-like structures were observed, while a flm-like morphology was observed for 2, similar to the case of 1 with C 60 (Fig. S10 †). These results indicate that 2 and C 60 assemble into The stoichiometric ratio between 2 and C 60 in the precipitate was determined by variable-temperature 1 H NMR measurements in toluene-d 8 (Fig. 7). We conducted experiments at low temperature to accelerate the assembly process. At 40 C, the broad spectrum of 2 in the presence of 0.5 equiv. of C 60 became very similar to the spectrum of 2, which exhibited sharp peaks. Using 1,1,2,2-tetrachloroethane as the internal standard revealed that $50% the original amount of 2 remained in solution (Fig. S11 †). Consequently, we concluded that the precipitate consists of 2 and C 60 in a 1 : 1 ratio. Notably, peaks in the aromatic region appeared upon addition of 2.0 equiv. of C 60 , and these are completely different from those observed for 2 (see also Fig. S12 †). The composition of the fbres was further analysed by MALDI-TOF mass spectrometry (Fig. 8). The spectrum exhibited several intense peaks at regular intervals. The gaps between the peaks correspond to the molecular weight of 2 or C 60 . The largest observable peak was at M w ¼ 13 kDa, corresponding to a 6 : 6 complex of 2 and C 60 . In their entirety, the NMR, SEM, and MS analyses allow the conclusion that the fbres consist of a 1D chain-like assembly of 2 and C 60 in a 1 : 1 ratio. To elucidate more structural details of the fbres, we performed a powder X-ray diffraction (XRD) analysis (Fig. S13 †), which exhibited two broad peaks at 2q ¼ 2.44 (36.2 ) and 5.00 (17.7 ). The spectral pattern of the fbres is thus inconsistent with that of the single crystal of 2$C 60 , suggesting the formation of a different packing structure. On the other hand, a powdered sample of 2 showed weak and broad reflections at 2q ¼ 4.94 (17.9 ) and 6.18 (14.3 ), indicating the lack of structural regularity in 2. Fig. 9 shows the UV-vis-NIR absorption of 2 and its inclusion complexes in the solid state. In contrast to 2, the 2$C 60 crystal exhibits a broad absorption band around 850 nm, which was characterized as a charge transfer (CT) band. The fbre aggregates also exhibit an NIR absorption band, indicating similar concave-convex binding between 2 and C 60 in the structure. However, the intensity of this band was higher for the fbre than for the crystal, suggesting different packing structures for these two samples. Two binding modes are possible for the association of 1 with C 60 . One is a 1 : 1 concave-convex complex and the other one involves the formation of a 1 : 2 sandwich-type complex. We anticipated that these two binding modes should result in different UV-vis-NIR absorption features in the solid state. Fortunately, by changing the solvents used for recrystallization from methanol/toluene to hexane/chloroform, a 2 : 1 complex of 1 and C 60 was obtained. 18 The single-crystal X-ray diffraction analysis of the complex unambiguously revealed a sandwichtype structure, in which two azabuckybowl molecules cooperatively capture a C 60 molecule by concave-convex interactions (Fig. 10). In addition, we recorded the solid-state UV-vis-NIR absorption spectra of the crystals for both 1 : 1 and 2 : 1 binding modes (Fig. 11). Both 1 : 1 and 2 : 1 complexes exhibit CT absorption bands around 850 nm. Notably, the absorption intensity at this wavelength is higher for the 2 : 1 complex than for the 1 : 1 complex. The theoretical calculations by the TD-DFT method also support these experimental results. The simulated absorption bands assigned to the CT transitions are signifcantly large in the 2 : 1 complex as compared to that in the 1 : 1 complex (Fig. S14 †). Such an enhancement of the CT band was also observed in the absorption spectrum of 2 + C 60 . This spectral similarity strongly indicates a sandwich-type binding mode in the 2$C 60 fbres. ## Plausible structures and mechanism of supramolecular assembly Scheme 2 illustrates the plausible association mechanism of 2 with C 60 . In the initial binding step, a 1 : 1 complex between 2 and C 60 should be formed. Binding to the second C 60 molecule would be weaker due to the reduced electron-donating ability of the other azabuckybowl unit after binding the frst electron-defcient C 60 . 19 Consequently, the 1 : 1 complex is obtained predominant in dilution. The 1 : 1 complexes interact with each other under more concentrated conditions to form fbres by sandwich-type binding, which are insoluble in organic solvents. We optimized the structure of the 1D chain-like arrangement using PM6 semi-empirical calculations (Fig. S15 †). The calculated interplanar spacing (17.8 ) is in good agreement with the XRD results (17.7 ). Further addition of C 60 to the fbres induces cleavage of the polymer chain to form soluble fragments, as detected by 1 H NMR spectroscopy. ## Conclusions In summary, we have synthesized a directly linked azabuckybowl dimer from a tribrominated monomeric precursor. The dimer exhibits strong 1 : 1 complexation with C 60 in solution. Segregated stacks of 2 and C 60 were observed in the crystalline state, suggesting efficient photo-excited charge-carrier mobility. Under concentrated conditions, 2 and C 60 form 1D chain supramolecular assemblies with a fbrous structure. The present results demonstrate that an electron-donating bowlshaped p-conjugated molecule can serve as a binding motif for pristine C 60 for the construction of supramolecular assemblies based on strong donor-acceptor interactions.

chemsum

{"title": "Supramolecular assemblies of a nitrogen-embedded buckybowl dimer with C₆₀", "journal": "Royal Society of Chemistry (RSC)"}

cerium–quinone_redox_couples_put_under_scrutiny

## Abstract: Homoleptic cerous complexes Ce[N(SiMe 3 ) 2 ] 3 , [Ce{OSi(OtBu) 3 } 3 ] 2 and [Ce{OSi i Pr 3 } 3 ] 2 were employed as thermally robust, weakly nucleophilic precursors to assess their reactivity towards 1,4-quinones in nonaqueous solution. The strongly oxidizing quinones 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) or tetrachloro-1,4-benzoquinone (Cl 4 BQ) readily form hydroquinolato-bridged ceric complexes of the composition [(Ce IV L 3 ) 2 (m 2 -O 2 C 6 R 4 )]. Less oxidising quinones like 2,5-di-tert-butyl-1,4-benzoquinone (tBu 2 BQ) tend to engage in redox equilibria with the ceric hydroquinolato-bridged form being stable only in the solid state. Even less oxidising quinones such as tetramethyl-1,4-benzoquinone (Me 4 BQ) afford cerous semiquinolates of the type [(Ce III L 2 (thf) 2 )(m 2 -O 2 C 6 Me 4 )] 2 . All complexes were characterised by X-ray diffraction, 1 H, 13 C{ 1 H} and 29 Si NMR spectroscopy, DRIFT spectroscopy, UV-Vis spectroscopy and CV measurements. The species putatively formed during the electrochemical reduction of [Ce IV {N(SiMe 3 ) 2 } 3 ] 2 (m 2 -O 2 C 6 H 4 ) could be mimicked by chemical reduction with Co II Cp 2 yielding [(Ce III {N(SiMe 3 ) 2 } 3 ) 2 (m 2 -O 2 C 6 H 4 )][Co III Cp 2 ] 2 . ## Introduction Quinones are multifunctional organic molecules exhibiting intriguing redox behaviour. 1,2 Of particular note is their importance in biological electron-transfer processes (photosynthesis, respiration) 3 and in industrial catalysis (anthraquinone process for hydrogen peroxide production). 4 Quinones can engage in one or two electron redox processes involving the formation of either semiquinolates or hydroquinolates. 5 Strikingly, the reduction potential of 1,4-benzoquinones (para-benzoquinones) can easily be modifed by introducing electronwithdrawing or donating substituents into the benzene ring. 5,6 As a consequence, tetrachloro-1,4-benzoquinone (chloranil, Cl 4 BQ) and even more so 2,3-dichloro-5,6-dicyano-1,4benzoquinone (DDQ) emerged as efficient oxidants in organic synthesis. 7 DDQ has been further successfully applied in photoredox catalysis. 8 Moreover, anionic h 4 -1,4-benzoquinone manganese tricarbonyl features a quinoid p-complex, broadly used for the fabrication of supramolecular metal-organometallic coordination networks. 9 Relatedly, deprotonated variants of 2,5-dihydroxy-1,4-benzoquinone (DHBQ) were shown to act as rigid ditopic linkers, 10 e.g., to support the formation of pentagonal dodecahedral Ce 2 (H 2 O) 18 cages or in permanently porous aluminium frameworks. 11 DHBQ was also probed as a bridging redox-active ligand in bimetallic [LnCl 2 (thf) 3 ] 2 (mbobq) (Ln ¼ Y, Dy; bobq ¼ 2,5-bisoxide-1,4-benzoquinolato) to build single-molecule magnets. 12 More recently, the related semiquinolato radical-bridged dimeric complexes [LnCl 2 (thf) 3 (m-Me 4 sq) 2 ] 2 (Ln ¼ Y, Gd) were obtained by oxidation of the corresponding in situ formed hydroquinolate complexes with FeCl 3 . 13 Semiquinolato-bridged scandium(III) species were reported to promote self-organised electron transfer from dtransition metals (Ir, Fe) to 1,4-quinones. 14,15 Targeted metal-redox chemistry with quinones has been a recurring issue for the rare-earth-metal couples Ln(II)/Ln(III) 16 and Ce(III)/Ce(IV). 17 Especially in the case of molecular cerium chemistry, 17 its unique single-electron-transfer (SET) pathway has recently been extended beyond the traditional application of ceric ammonium nitrate (CAN; redox potential of 1.61 V vs. NHE) in organic synthesis 18 to photoredox catalysis. 19 On the other hand, redox protocols are known to provide efficient access to metalorganic Ce IV complexes. Typically, such Ce III / Ce IV transformations are promoted by halogenating oxidants (e.g. C 2 Cl 6 , Ph 3 CCl, PhICl 2 , TeCl 4 , FcPF 6 , FcBF 4 , Ph 3 CBF 4 , Ph 3 CPF 6 , I 2 ), 20 silver salts (AgX, X ¼ F, I, BF 4 , OTf) 21 or dioxygen. 20b,22 Archetypical 1,4-benzoquinone (BQ) has been established as a versatile oxidant for the synthesis of homoleptic ceric complexes CeL 4 from cerous ate complexes [CeL 4 M(do) x ] via tandem oxidation-ligand redistribution protocols (L ¼ monoanionic ligand, M ¼ alkali metal and do ¼ donor solvent; separation of an alkali-metal hydro-/semiquinolate). 23 In the presence of sterically demanding ligands L, BQ was also shown to form hydroquinolato (hq)-bridged ceric complexes of the general composition [L 3 Ce-OC 6 H 4 O-CeL 3 ]. 20g,24 This very Ce III / Ce IV transformation was pioneered by Sen et al. in 1992, resulting in the isolation of [(tBu 3 CO) 3 Ce(OC 6 H 4 O)Ce(OCtBu 3 ) 3 ] (Chart 1, I). 24a In the same paper, the oxidation of Ce(OCtBu 3 ) 3 with 2,6-di-tert-butyl-1,4-benzoquinone to the terminal Ce IVsemiquinolate radical (tBu 3 CO) 3 Ce(O 2 C 6 H 2 tBu 2 ) was described as evidenced by 1 H NMR and EPR spectroscopic measurements. 24a More recently, Schelter et al. reported on hq-bridged complex II resulting from the oxidation of cerous Ce(BINOlate) 3 (thf)Li 3 (thf) 4 with 0.5 equivalents of BQ. 24b Similarly, our group synthesized In contrast, the reaction of BQ with [Ce(Me 2 pz) 3 ] x featuring the sterically less demanding and increasingly nucleophilic 3,5-dimethylpyrazolato ligand (Me 2 pz) led in fact to a transient Ce IV hydroquinolate species (as indicated by the characteristic colour change), which, however, at ambient temperature was converted into the isolable trimetallic Ce III complex Ce 3 (pchd) 2 (Me 2 pz) 5 (thf) 2 (pchd ¼ 1,4-bis(3,5-dimethylpyrazol-1-yl)cyclohex-2,5-diene-1,4-diolato). 23c Apparently, the new pchd ligand formed via 1,4-nucleophilic attack at bq by two adjacent Me 2 pz ligands. This nucleophilic reaction pathway could be prevented by using bulky tBu groups on the pz ligand, but homolpetic Ce(tBu 2 pz) 4 was formed as the main ceric product via irreversible ligand rearrangement. 23c As such, cerium-1,4-benzoquinone couples have revealed distinct redox chemistry, we became curious about as to what extent such redox transformations are affected by both the type of 1,4-benzoquinone oxidant and the molecular Ce III precursor. The present study uncovers some unexpected correlation between Ce IV -hydroquinolato stabilisation and quinone oxidant strength, as well as a new path to p-semiquinolatoradical-bridged rare-earth-metal complexes. ## Molecular redox precursors The quinones used in this study comprise 2,3-dichloro-5,6dicyano-1,4-benzoquinone (DDQ), tetrachloro-1,4benzoquinone (Cl 4 BQ), 1,4-benzoquinone (BQ), tetramethyl-1,4-benzoquinone (Me 4 BQ), 2,5-di-tert-butyl-1,4-benzoquinone (tBu 2 BQ), 1,4-naphthoquinone (NQ), and 9,10-anthraquinone (AQ). All are commercially available and were selected according to their reduction potentials spanning a E 0 range of 89 to 887 mV (2e /2H + , vs. NHE, cf., Scheme 1). 5,25 The cerous precursors were chosen according to the criteria solubility, weak nucleophilicity, proven access to the tetravalent state, and a stabilizing effect on the latter. Furthermore, the use of sterically bulky ligands was assumed to minimise the occurrence of ligand redistribution reactions. Accordingly, homoleptic Ce [N(SiMe 3 ) 2 ] 3 (1) appeared to be an ideal benchmark system. 24c After additional investigations into the respective pyrazolate chemistry, the abovementioned [Ce(R 2 pz) 3 ] (R ¼ Me, tBu) were discarded because of persisting alternative reaction pathways like 1,4-nucleophilic attack of BQ by Me 2 pz and ligand redistribution (formation of Ce(tBu 2 pz) 4 ). 23c The new pyrazolate studies clearly confrmed that steric hindrance of both the pyrazolato ligand and the 1,4-benzoquinone can minimise/ counteract such undesired reactions, but the formation of product mixtures seems inevitable. Products crystallised from these reactions include minor amounts of ceric [Ce(tBu 2 pz) 3 (thf)] 2 (Me 4 hq) or a cerous product of partial pyrazolyl-promoted nucleophilic attack Ce 3 (bpad)(pasq)(Me 2 pz) 6 (thf) (bpad ¼ 1,4bis(3,5-dimethylpyrazol-1-yl)anthra-1,4-diolato; pasq ¼ 1-(3,5dimethylpyrazol-1-yl)anthra-1,4-semiquinolato) (84%) mixed with semiquinolate [Ce(Me 2 pz) 2 (thf) 2 (asq)] 2 (asq ¼ anthrasemiquinolato; cf. ESI † for structural details). The use of Ce III halides was discarded mainly for solubility issues. In addition to silylamide 1, the siloxide derivatives [Ce {OSi(OtBu) 3 } 3 ] 2 (2) 21d,26 and [Ce(OSi i Pr 3 ) 3 ] 2 (3) were assessed as suitable cerous precursors. Complexes 2 and 3, with and without intramolecular donor site, respectively, were readily obtained in pure form via protonolysis of 1 with the corresponding silanol. 26 The crystal structure of the new complex 3 revealed a dimeric arrangement with two m 2 -bridging and four terminal siloxy groups (Fig. 1), similar to that found for tris(tertbutoxy)siloxy congener 2 or [Ce(OSiPh 3 ) 3 ] 2 (6 tBu2hq ) (from top down). Ellipsoids are shown at the 50% probability level. Hydrogen atoms, disordered ligands and lattice solvents are omitted for clarity. Selected interatomic distances for 5 hq , 6 Cl4hq , 4 ddq , and 4 Me4hq are given in Tables 1 and 2. Selected bond lengths for 6 tBu2hq []: Ce1-O1 2.109(3), Ce1-O2 2.104(3), Ce1-O3 2.097(3), Ce1-O4 2.107(3), C1-C2 1.391( 6), C2-C3 1.391( 6), C1-C3 0 1.405( 6), C1-O1 1.350 (5). ## Quinone oxidation of Ce Treatment of Ce[N(SiMe 3 ) 2 ] 3 (1) with each 0.5 equivalents of DDQ, Cl 4 BQ, Me 4 BQ, tBu 2 BQ and NQ, in a mixture of toluene and n-hexane, immediately led to a colour change from yellow to dark brown. Upon recrystallisation from toluene/n-hexane mixtures it was possible to isolate the hydroquinolato-bridged complexes ) and [Ce{N(SiMe 3 ) 2 } 3 ] 2 (m 2 -O 2 C 10 H 6 ) (4 nhq ) in very good crystalline yields of 71 to 90% (Scheme 1). The crystal structures of the new complexes 4 xhq are isostructural to the previously reported derivative 4 hq24c and only differ in the bridging hq linker (Fig. 2). The Ce1-O1 distances of 2.084(6) to 2.173(2) (for a full list of interatomic distances see Table 1) are in the same range as found for other hq-bridged cerium complexes (2.086(10)-2.143(5) ). All attempts to isolate putative 4 ahq , derived from the weakest oxidising quinone under study, namely 9,10-anthraquinone (E 0 ¼ 89 mV; 2e /2H + , vs. NHE), 5 were unsuccessful with the reaction mixtures showing no colour change immediately after addition of the anthraquinone. However, a colour change from derived from quinones with a relatively strong oxidising effect were successfully isolated from these reactions (Scheme 1). However, the weakly oxidizing quinones Me 4 BQ and NQ did not lead to tetravalent cerium species, as indicated by the detection of only paramagnetic signals in the 1 H NMR spectra (for an example of such a 1 H NMR spectrum, see Fig. S34 in the ESI; † formation of semiquinolates, vide infra). The accessible complexes 5 and 6 were obtained in moderate to good crystalline yields of 42 to 71% upon recrystallisation from THF or THF/ Et 2 O mixtures. Crystals suitable for XRD analysis were obtained for complexes 5 hq , 5 Cl4hq , 6 Cl4hq , 6 ddhq and 6 tBu2hq , revealing the same structural motif as complexes 4, that is two CeL 3 moieties connected via a hydroquinolato linker (Fig. 2). Strikingly, the 1 H NMR spectrum of 6 tBu2hq indicated the existence of an equilibrium similar to that of ceric 4 tBu2hq (cf. Scheme 2). However, along with the reactants 3 and tBu 2 BQ additional signals assignable to distinct dia-and paramagnetic decomposition products were detected. Further, the crystal structures of complexes 5 and 6 show that the cerium atoms are additionally coordinated by THF donor molecules. The Ce1-O siloxide distances of 2.066(2) to 2.1534(10) (see Table 2 for a complete list of interatomic distances) compare well to other ceric siloxides like Ce{OSi(OtBu ) 3 } 4 (2.089(2)-2.157(2) 26 and 2.084-2.160 21d ) or Ce{OSiPh 3 } 4 (dme) (2.098(1)-2.133(1) ). 29 Also, as seen for the silylamides 4, the Ce1-O hq distances of 2.1244(10) to 2.2325( 16), as well as the C-C and C-O distances underline the formation of an aromatic hq linker. 20g,24 1 H NMR spectroscopic measurements also validate the formation of Ce IV species, showing a sharp singlet for the tert-butyl groups and a doublet plus a septet for the iso-propyl groups depending on the siloxy co-ligand. 1 and 2). The detection of two closely adjacent redox events (E pc values) in some cyclic voltammograms may correspond to a successive reduction/oxidation of the cerium centres. Similar features were also described for the hq-bridged Ce(IV)-BINOLate complex II. 24b All complexes under study display redox processes with a large separation of E pc and E pa (DE z 0.6 V for 4; 1.5 V for 5 and 1.0 V for 6). Representatively, the cyclic voltammograms of the DDQfunctionalized Ce III /Ce IV redox couples are depicted in Fig. 3 (top graphic). The silylamide complexes 4 gave reduction potentials similar to those reported for halogenidofunctionalised ceric complexes Ce[N(SiMe 3 ) 2 ] 3 X (E 1/2 ¼ 0.56 (X ¼ F), 0.30 (X ¼ Cl), 0.31 (X ¼ Br)) with E 1/2 values of 0.46 V for 4 Cl4hq and 0.36 V for 4 ddhq . 30 Only 4 hq with E 1/2 ¼ 0.76 V gave a signifcantly higher stabilisation by 0.20 V. The extra-large separation of the reduction/oxidation events observed for the siloxide complexes 5 and 6 had been noticed previously for rare-earth-metal siloxides and was assigned to oxidation-state-dependent ligand reorganisation processes. 31 Stabilisation of the tetravalent oxidation state of cerium in complexes 4, 5, and 6 increases in the order of N(SiMe 3 ) 2 < OSi(OtBu 3 ) 3 < OSi i Pr 3 as co-ligand (Fig. 3/bottom) which is in accordance with previous fndings. 20f ,26,30,31c Surprisingly, the stabilisation of Ce IV proceeds in reverse order of the oxidation potential of the 1,4-quinones under study giving the most stable complexes for the hydroquinolato-bridged complexes and the least stable compounds for its 2,3-dichloro-5,6-dicyanohydroquinolato congeners. A reason for this trend could be the increasingly electron-defcient nature of the aromatic hydroquinolato linkers due to the large I effect of the substituents. The Ce IV oxidation state can be stabilised by increasing donor strength of the ligands. 32 Based on this, it seems surprising that isolable complexes 4 Me4hq and 4 nhq , derived from weakly oxidizing quinones, are not stable in solution at ambient temperature. This might be a result of another reaction pathway preferred after formation of the hydroquinolato-bridged Ce IV complexes (like following up redox processes and the formation of Ce III semiquinolates, cf. vide infra). ## Reduction of silylamide 4 hq with cobaltocene Having investigated the electrochemical reduction of compounds 4, 5 and 6, the chemical reduction with cobaltocene (CoCp 2 ) (1.31 V vs. Fc/Fc + in DME) 2a was attempted, as it has already been shown to engage in such reductions. 31a,33 Accordingly, treatment of a solution of 4 hq in THF with two equivalents of CoCp 2 resulted in a colour change from dark brown to pale yellow (Scheme 3). The 1 H NMR spectrum of the reaction mixture showed complete consumption of CoCp 2 and only broadened signals indicating the formation of a paramagnetic Ce III species. Crystallisation from a concentrated THF-d 8 solution at 40 C gave light brown crystals of the composition [(Ce {N(SiMe 3 ) 2 } 3 ) 2 (m 2 -O 2 C 6 H 4 )][CoCp 2 ] 2 (7) (Fig. 4). Complex 7 shows the same structural motif as 4 hq but is flanked by two cobaltocenium cations. Compared to 4 hq , the Ce-N and Ce1-O1 distances are elongated by approximately 0.19 as expected for the larger Ce III ion size. 34 On the contrary, the bonding parameters within the bridging hydroquinolato linker did not change, further corroborating a cerium-borne redox chemistry. Reacting 4 hq with one equivalent of CoCp 2 did not lead to a mixed Ce III/IV complex but gave a mixture of 50% of 7 and 50% of unreacted starting material. The reactions of CoCp 2 with other complexes 4 to 6 in THF-d 8 showed immediate decolourisation of the solution while the 1 H NMR spectra of the reaction mixtures displayed only paramagnetic signals (for an example, see Fig. S33 in the ESI †). However, the isolation of additional reduced species similar to 7 was not successful. 391(4), C1-C3 1.392(4), C2-C3 0 1.392(4). ## Cerium semiquinolates A closer look at the reactions of cerous siloxides 2 and 3 with the weakly oxidising quinone Me 4 BQ (which did not produce any tetravalent cerium species; vide supra) revealed another important detail of the cerium-quinone redox system. Treatment of 2 or 3 with 0.5 equivalents of Me 4 BQ led to a colour change from colourless to light blue. Upon recrystallisation from THF dark blue crystals suitable for X-ray diffraction could be grown and were identifed as cerous semiquinolates [CeL 2 (thf) 2 ] 2 (m 2 -O 2 C 6 Me 4 ) 2 (with L ¼ OSi(OtBu) 3 ( 8) or OSi i Pr 3 ( 9)) (Scheme 4). Examining the reaction mixtures by 1 H NMR spectroscopy in THF-d 8 showed, besides paramagnetic signals for 8 and 9, a sharp singlet at 1.39 ppm (for 8) or a doublet plus a septet at 1.13 and 1.04 ppm (for 9), indicating the formation of homoleptic Ce[OSi(OtBu) 3 ] 4 or [Ce(OSi i Pr 3 ) 4 ], respectively, as a result of the one-electron reduction of Me 4 bq followed by ligand redistribution. Crucially, such a reaction pathway seems unfeasible for complexes 4, since putative homoleptic "Ce [N(SiMe 3 ) 2 ] 4 " is unknown. 20a Emergent kinetic constraints in the case of ceric complexes 4 were also suggested by the redox behaviour of [Ce{N(SiHMe 2 ) 2 } 3 ] 2 derived from a less bulky silylamido ligand. Accordingly, the cerous bis(dimethylsilyl)amide complex was treated with one equivalent of both BQ and Me 4 BQ in THF-d 8 and C 6 D 6 (see Fig. S38-S41, ESI †). The 1 H NMR spectra of these reactions suggest the formation of a tetravalent species of the composition "[Ce{N(SiHMe 2 ) 2 } 3 ] 2 (m 2 -O 2 C 6 R 4 )". However, the ceric products appear to be unstable in solution at ambient temperature. In C 6 D 6 , the formation of Ce [N(SiHMe 2 ) 2 ] 4 was observed in the reaction with BQ as well as other insoluble products. In THF-d 8 , the resulting product seemed more stable but after 24 h in solution also traces of decomposition products were found. The Me 4 BQ reaction in C 6 D 6 also indicated successful oxidation, however, after 24 h the 1 H NMR spectrum revealed signals for trivalent decomposition products as well as traces of Ce[N(SiHMe 2 ) 2 ] 4 . In THF-d 8 , the putatively formed hydroquinolate complex was even less stable, showing signals for trivalent by-products directly after addition of Me 4 BQ. In addition, the stability of 4 bq was investigated in THF-d 8 showing small amounts of decomposition products like Ce[N(SiMe 3 ) 2 ] 3 after 24 h (Fig. S42, ESI †). Regrettably, purifcation of complexes 8 and 9 was impeded by co-crystallisation with the ceric by-products CeL 4 . The crystal structures of 8 and 9 revealed two six-coordinate cerium atoms surrounded by two siloxy ligands, two THF donor molecules and two bridging tetramethyl semiquinolato moieties (Fig. 5). The Ce1-O silanolato distances are elongated by about 0.1 compared to the respective tetravalent compounds 5 and 6 and in accordance with other Ce III siloxides like 3, Ce{OSi(OtBu 35 and [Ce{OSiPh 3 } 3 ] 2 (Ce-O term 2.141(7)-2.184(6) ). 27 As expected for semiquinolato ligands the six-membered rings display two shortened C-C and four elongated C-C bonds. Additionally, the six-membered rings are slightly bent in comparison to the flat aromatic hydroquinolato linkers in complexes 4, 5 and 6 resulting in an angle of 170.34 for 8 and 170.29 for 9, respectively (see Fig. 5, bottom). Notwithstanding, the bridging radicals engage in signifcant p-stacking as indicated by close semiquinolato-semiquinolato distances of 3.112 for 8 and 3.156 for 9. Overall, complexes 8 and 9 display the same arrangement of the semiquinolato radical bridges as observed in complexes [LnCl 2 (THF) 3 (m-Me 4 sq) 2 ] 2 (Ln ¼ Y, Gd) (Ct/Ct 3.097 ; Ct ¼ centroid of benzene rings). 13 To investigate the electronic behaviour of the bridging semiquinolates, X-band EPR spectra of compounds 8 and 9 were recorded from a crystal powder sample at 123 K (Fig. 6). For both complexes cw-EPR spectra are composed of two distinct sets of resonances, one that results from the transition within the Kramers doublet corresponding to m j ¼ AE1/2 and one at half-feld, H z 160 mT. The transition for the m j ¼ AE1/2 state associates with an axial g tensor with principal components g II ¼ 2.094 and g t ¼ 2.032 for 8 and g II ¼ 2.088 and g t ¼ 2.032 for 9, respectively. The transition locating at half-feld gives rise to a very broad dispersion line with g II z 4.359 for 8 and g II z 4.351 for 9, respectively. Notably, an identical line pattern derives from a frozen 2-Me-THF solution at 77 K; cf. Fig. S67, ESI, † for pertinent details. The cw-EPR spectra corroborate the presence of Ce 3+ , and agree with early work on mononuclear complexes of Ce 3+ . 36 This indicates a radical- 7), Ce1-O2 2.422(7), Ce1-O3 2.219(9), Ce1-O4 2.237( 6), Ce1-O5 2.648(7), Ce1-O6 2.644 (7), C1-O5 1.306( 5), C4-O6 1.307( 5), C1-C2 1.458( 6), C2-C3 1.401( 6), C3-C4 1.449(8), C4-C5 1.463( 6), C5-C6 1.388( 6), C1-C6 1.463(8). Bottom: side view of [(Ce{OSi i Pr 3 } 2 (thf) 2 ) 2 (m 2 -O 2 C 6 Me 4 ) 2 ] (9). radical p-bonding, as it was recently shown for yttrium and gadolinium semiquinolates [LnCl 2 (thf) 2 ] 2 (m 2 -O 2 C 6 Me 4 ) 2 (Ln ¼ Y or Gd). 13 Additionally and also similar to the yttrium semiquinolate, complex 9 shows a signal with a g value of 1.999, which most likely results from a non-coupled radical impurity. Note that the reactions of cerium siloxides 2 and 3 with 1,4naphthoquinone resulted also in the formation of homoleptic ceric siloxides as well as paramagnetic by-products (cf. Fig. S36, ESI †) indicating a similar reactivity as observed for Me 4 BQ. Unfortunately, any putative semiquinolate complexes could not be isolated.

chemsum

{"title": "Cerium\u2013quinone redox couples put under scrutiny", "journal": "Royal Society of Chemistry (RSC)"}

adhesion-peptide_conjugates_of_thermo-responsive_polymers_co-electrospun_into_fibrous_scaffolds_enab

## Abstract: Human corneal stromal cells (hCSCs) are in high demand for ocular repair following injury or disease but differentiate to an undesirable activated fibroblastic phenotype when cultured by conventional methods. Here we report a new peptide-functionalised thermo-responsive electrospun cell culture scaffold that supports rapid expansion, phenotypic expression and enzyme-free passaging of quiescent hCSCs for potential ocular repair applications.Current methods of adherent mammalian cell expansion for tissue engineering and regenerative medicine applications rely usually on the use of 2D culture platforms, foetal bovine serum containing media, and enzymatic digestion for cell recovery. However, 2D culture of primary cells can result in undesired de-differentiation, 1-3 due to the lack of similarities to the natural extracellular matrix (ECM) environment, and enzyme passaging can significantly reduce cell quality, 4-6 as a result of the destruction of important cellsurface proteins. In recent years, thermo-responsive surfaces 7,8 have been used as supports for mammalian cell culture, as their changes in physicochemical properties at upper or lower critical solution temperatures (UCST, LCST) can facilitate cell detachment without the need for enzymatic passaging. 9-13 However, while the polymer phase transitions result in measurable "hydrophilic-tohydrophobic" behaviour changes, these differences are often only moderate in terms of their effects on cell adhesion. Accordingly, there is a need to generate materials, which provide selective functionality for specific cells to attach, while retaining their ability to release cells via a mild temperature change at the end of the cell culture and expansion period.Here we report the development of a new thermo-responsive adhesion peptide-functionalised (GGG-YIGSR) fibre-scaffold based on co-electrospun poly(DL-lactide) (PLA) and poly(di(ethylene glycol) methyl ether methacrylate)-co-poly(di(ethylene-glycol) carboxylate ester ethyl thiol) (PDEGMA/PDEGSH, Fig. 1A) for enzyme-free 3D mammalian cell culture. We have exemplified this concept via the growth and recovery of human corneal stromal cells (hCSCs) in non-serum containing media. These cells were chosen as hCSCs display a well characterised phenotypic change during conventional in vitro culture from a keratocyte phenotype (quiescent, CD34 + , ALDH + ) to a mesenchymal stem cell/fibroblast/myofibroblast phenotype (proliferative, CD105 + , α-SMA + ). 14 The laminin sequence YIGSR was selected for this study as previous literature has shown that interaction with this peptide sequence supports hCSCs attachment and the maintenance of a quiescent keratocyte phenotype. 15 Co-spinning PLA with thermo-responsive polymers of PDEGMA and a side-chain thiol-functionalised PDEGSH generated fibrous scaffolds with 'click' handles for peptide coupling. The subsequent conjugation chemistries were carried out by reacting norbornene-fuctionalised peptides (Nor-GGG-YIGSR, Fig. 1C) with the side-chain thiols via a UV-mediated thiol-ene reaction (Fig. 1B). Attachment of the peptides on the scaffolds was confirmed by fluorescence labelling, ToF-SIMS and XPS analysis. The resulting bioconjugate scaffolds were assessed for their ability to support hCSC culture by cell adhesion, proliferation and immunocytochemistry assays. The phenotypic profiling of hCSCs after thermo-responsive passaging was evaluated by flow cytometry analysis of key cell surface markers. The combined assay data indicated that the expansion of hCSCs could be conducted with high yield and retention of a therapeutically desirable phenotype without using expensive and potentially cell-damaging enzymatic methods. The ability to culture, expand and recover therapeutically relevant cell types makes these scaffolds a promising new class of materials for application in many areas of regenerative medicine, where cells are destined for the clinic. Polymer synthesisThe co-polymers PDEGMA/PDEGSH in different ratios (100:0, 98:2, 97:3, 96:4, and 90:10, PDEGMA to PDEGSH respectively) were synthesised by a two-step route (scheme. Fig 1A ) and structures were confirmed by 2D-NMR analysis (Fig. disappearance of the CH2OH signal (H 1 -NMR; 3.7-3.8 ppm, t, 2H and C 13 -NMR; 61 ppm) and the appearance of the CH2SH peaks (H 1 -NMR; 2.9 ppm, t, 2H and C 13 -NMR; 34 ppm) in the H 1 and C 13 -NMR of the end product PDEGMA/PDEGSH. Full materials, methods and characterisation can be found in the supplementary section 1, Fig. S1-18 for NMR and Tab. S3 for GPC. Percentages of free thiol in the co-polymers were confirmed by the Ellman's assay (Tab. 1). The phase transition temperatures (Tt) of the polymers were estimated by cloud point measurements (Fig. 1E, S19 and Tab. S4). The norbornene functionalised peptide GGG-YIGSR was synthesised on a rink amide resin, purified by HPLC and its structure verified by electrospray ionisation mass C) the chemical structure of the synthesised Nor-GGG-YIGSR peptide sequence. Nor (blue block) is the norbornene acid photo-reactive group, GGG (green block) is used as a spacer and YIGSR (red block) sequence previously identified to promote hCSCs proliferation. E) Cloud point hysteresis of PDEGMA, PDEGMA/PDEGOH, PDEGMA/PDEGSH and PDEGMA/PDEGS-GGG-YIGSR in water (2 mg/ml). Arrows up are the heating spectra and arrows down are the cooling spectra.Transition temperature (T t) shift down when polymer cools down. This is due to the aggregation of the polymer, which takes time to break up. PDEGOH Tt is slightly higher in temperature due to the additional hydrophilic group OH. PDEGSH polymer shows an earlier Tt change than the PDEGMA and PDEGOH polymers as well as a slow steady slope when cooled. F) FT-IR spectra from pure Nor-GGG-YIGSR (red), PDEGMA/PDEGSH (green) and peptide functionalised PDEGMA/PDEGS-Nor-GG-YIGSR by photo-initiated thiol-ene chemistry (purple). The double bond from the norbornene acid disappeared and the amino g roups of the peptide appeared on the FT-IR of the functionalised polymer-peptide. spectrometry (ESI-MS, Fig. S20). Upon reaction of the peptide with the thiolated polymer, a decrease in percentage of free thiols was observed at increasing UV exposure times (0, 5, 10 and 15 mins), with a conversion of 65% after 15 mins (Tab. 1). Fourier-Transform Infrared (FT-IR) analysis showed the disappearance of the norbornene double bond (1530 cm -1 ) and the appearance of the amino-groups (1655 cm -1 ) of the peptide in the case of the reacted peptide-thiol polymer (Fig. 1D). Cloud point measurements of the PDEGMA/PDEGS-Nor-GGG-YIGSR derivatised polymer showed transition temperatures (22°C when heated and <18°C when cooled, Fig. 1E and Tab. S4), as compared with the PDEGMA/PDEGSH precursor, confirming retention of thermo-responsive behaviour after peptidefunctionalisation. ## Electrospun scaffold fabrication and characterisation The obtained PDEGMA/PDEGSH polymers were successfully blend electrospun with PLA into uniform fibres (19 kV, 17.5 cm, 0.5 hr/ml and 25 w/v% PLA and 10 wt% PDEGMA/PDEGSH polymer in Ac/DMF) with an average fibre diameter range of 800-1200 nm (Fig. S22 and Tab. S5, optimised via previous DoE work 16 ). XPS analysis and fluorescence labelling confirmed the presence of free thiols (S(2p) peak at 168.9 eV and fluorescein-5-maleimide labelling respectively) and peptide (N(1s) peak at 399.8 eV and ATTO NHS ester labelling respectively) on the surface of the scaffolds (Fig. S22 and Tab. S6). The distributions of the free thiols and peptides on the scaffolds surface were investigated further by ToF-SIMS analysis. Mass spectrometry peaks of the ion fragments were identified and confirmed through literature (Tab. S4). The ion fragment CH3Oattributed to the underivatised PDEGMA was present on the surface, and a decrease in peak intensity of CH3Owas observed for all the scaffolds containing PDEGMA/PDEGSH co-polymer (Fig. 2A). ToF-SIMS peak distribution images indicated that SH groups were in regions corresponding to the fibre positions on the substrate (Fig. 2C and Fig. S22), confirming an even distribution of the free thiols on the fibrous surface. A decrease in Sand SHabsorptions were detected for the thiol-ene derivatised scaffolds (Fig. 2A). The CNOpeak of the peptide amino-groups were observed for all scaffolds incubated with peptide (Fig. 2A). Significantly higher intensity was detected for the scaffold 3S-P, with a decrease in CNOintensity being observed for 4 and 10 S-P (Fig. 2A). Nor-GGG-YIGSR peptide positive ion fragment signals were identified and were corroborated to literature 21,22 (see fragment structures of identified signals in Fig. S23). Peptide fragments were detected for all scaffolds, including the 0SH scaffolds (Fig. 2B). This suggested that non-specific electrostatic binding of the peptide occurred with the surface of the scaffolds. The distribution images of the signal, an even distribution of peptide fragment peaks was observed on the 0% thiol scaffold without a defined fibre morphology (Fig. 2C). However, in case of the peptide-derivatised scaffolds, more defined fibre morphologies were found (see Fig. 2C). This observation indicates the occurrence of specific reaction of the fibre surfaces when free thiols are available, as well as some, albeit low-level, non-specific peptide interaction at the surface of the fibres with no thiols present. Cell expansion and phenotype support hCSCs were isolated from corneascleral rims as previously described 23 and cultured on the fabricated scaffold (see protocol sup. Section 1). The media composition of DMEM/F12 with 20% serum replacement, fibroblast growth factor and leukaemia inhibitory factor (SCM) has previously reported to support the keratocyte phenotype of hCSCs 23 and was therefore used in this study. Enhanced proliferation was observed on the scaffolds with an increased content of the laminin mimetic peptide sequence Nor-GGG-YIGSR compared to the PLA/PDEGMA scaffolds (Fig. 3A). Interestingly, significantly higher proliferation was also found on PLA with PDEGMA/PDEGSH in the fibres (Fig. 3A). Morphology staining (actin and vimentin for the cytoskeleton) studies showed clusters of cells on PLA, and PLA with 10 wt% PDEGMA (0SH, Fig. 4 and S26), elongated cell morphologies on PLA with PDEGMA/PDEGSH (90:10, Fig. S26), and an increase in a spindle-like morphology of the cells on the peptide derivatised PLA scaffolds with more PDEGMA/PDEGS-Nor-GGG-YIGSR present (Fig. 4 and S26). The round morphology of the cells on the PLA, and PLA with 10 wt% PDEGMA scaffolds indicated a limited interaction between the cells and the fibrous scaffold. This in turn suggested that protein deposition was not enhanced when the adhesion peptides were merely adsorbed (as detected by ToF-SIMS analysis) rather than being covalently bound to the fibre surface. The elongated morphologies of cells attached to the thiol-functionalised scaffolds was due most probably to direct cell surface-thiol reactions, for example via surface exposed cysteines. This assertion is supported by prior data showing that thiolated surfaces can result in improved ECM protein adsorption and subsequently enhanced cell attachment and proliferation. 24,25,26 However, significantly more cells exhibiting a spindle-like morphology were observed when the laminin mimetic sequence YIGSR was present (see Fig. 4). Uzunalli et al. 15 reported a similar morphology of hCSCs when cultured on YIGSR-containing nanofibres compared to fibronectin-mimetic RGD-peptide nanofibres, as well as increased proliferation, supporting the findings reported here. Immuno-staining of the hCSCs showed the quiescent keratocyte phenotypic markers CD34 and ALDH, and mesenchymal stem cell (MSC) expression of CD105 on all the different scaffolds (see Fig. 4 and Fig. S26). In addition, the expression of myofibroblast marker α-SMA was found when cells were cultured on the PLA and PLA with 10 wt% PDEGMA scaffolds, and a very low expression of α-SMA was also observed when cells were cultured on the PLA fibres with 10 wt% PDEGMA/PDEGSH (90:10, Fig. S26). No α-SMA expression was detected when cells were cultured on any of the PLA fibres functionalised with 10 wt% PDEGMA/PDEGS-Nor-GGG-YIGSR. These data confirmed that the peptide supported the quiescent keratocyte phenotype and suppressed dedifferentiation of the keratocytes to their activated myofibroblast phenotype. ECM expression was investigated to confirm the hypotheses that the free thiol group-containing scaffold supported ECM expression and adsorption and hence cell adhesion and proliferation as observed in Fig. S26. Collagen I expression was observed throughout all the scaffolds (Fig. 4), however, this protein increased in abundance and was more evenly distributed on the peptide containing scaffolds (Fig. 4). These results confirmed that PDEGMA/PDEGS-Nor-GGG-YIGSR peptide functionalisation on the scaffold led to an increased expression of collagen I. Stromal keratocytes in a normal undamaged cornea express the keratan sulphate proteoglycan, lumican, a corneal transparency factor. 27,28 Lumican interacts with collagen fibrils helping to maintain spacing and transparency in a healthy cornea 29 . Production is significantly reduced during wound healing and phenotypic change to a fibroblast morphology, and thus the presence of lumican can be considered a marker of the quiescent phenotype of hCSCs. As apparent from Figure 4, lumican expression was found on the scaffolds with peptide present, including the 0SH scaffold with non-specific bound peptide. This indicated that the non-specifically bound Nor-GGG-YIGSR on the surface detected during ToF-SIMS analysis did support ECM expression. However, myo-fibroblast differentiation and clusters of cells were still observed, suggesting that the non-specifically bound peptide did not support strong cell interactions with the scaffold. It was observed that cells spread out across the surfaces only when the peptide was covalently bound to the fibres. Increased lumican expression was detected for the peptide bound scaffolds with 3, 4 and 10% PDEGSH attached, providing support for assertion (Fig. 4). As the derivatised scaffolds with the Nor-GGG-YIGSR peptide were observed to promote the desired quiescent phenotype, spindle-like morphology and ECM expression, the cellular responses on the surfaces with different percentages of peptides attached were investigated. After 9 days cell culture, significantly higher proliferation was observed for 3 and 10% peptide-derivatised scaffolds (Fig. 3B). The expression of the quiescent keratocyte phenotype CD34 and ALDH markers were detected for all different concentrations of peptide (Fig. 4). A significantly increased expression of ALDH was observed for the 3% peptide scaffold. As ALDH expression plays an important role in the maintenance of corneal transparency, this indicated a desirable response of the cells in terms of their culture on the derivatised fibres. In addition, CD105 MSC marker expression was found on all scaffolds, suggestive of a partial loss of the differentiated quiescent phenotype to a more stem cell-like phenotype. 30 Myo-fibroblastic α-SMA expression was observed for those cells cultured on scaffolds containing 2% peptide and on the 0% control, indicating a threshold of peptide content needed to avoid myofibroblast activation. Peptide concentrations of 2% were too low to support the quiescent phenotype and resulted in a myo-fibroblastic change, while cells cultured on scaffolds containing 3% peptide and above generated no α-SMA expression. The previously rounded and cluster cell morphology was observed on the scaffolds containing 0 and 2% peptide concentrations, while a more fusiform morphology was observed when cultured on scaffolds with peptide-concentrations above 2% (Fig. 4). Moreover, lumican ECM expression was detected when cells were cultured on scaffolds with higher percentages of peptide on the surface (3, 4, and 10%). It can therefore be concluded that scaffold fibres with more than 2% of Nor-GGG-YIGSR covalently bound to their surfaces significantly improved hCSCs attachment, proliferation, ECM expression and supported a quiescent cell phenotype. ## Enzymatic-free passaging The thermo-responsive PDEGMA/PDEGS-Nor-GGG-YIGSR polymer PLA scaffolds were not only designed to improve cell attachment and phenotype support, but also to prevent cell damage, which can occur through enzymatic passaging. Many studies to date have investigated the potential of the thermo-responsive polymer, PNIPAM, for corneal cell type detachment. 8, We have recently shown that blend electrospinning a thermo-responsive polymer poly(PEGMA188) with a bulk polymer of choice will render the fibre thermo-responsive. Mammalian cell adhesion, viability, proliferation and phenotype on this fibrous culture system over numerous thermal enzyme-free passages was achieved. 34 However, no other studies are reported to date utilising such thermo-responsive scaffolds for mammalian cell culture. A preliminary experiment illustrating the ability to culture and detach cells from PDEGMA and PDEGMA/PDEGS-Nor-GGG-YIGSR-containing scaffolds, and the subsequent quantitative assessment of the cell phenotype by flow cytometry analysis (FACS), was performed. hCSCs grown on peptide-derivatised fibrous scaffolds containing different concentrations of PDEGMA/PDEGS-Nor-GGG-YIGSR were passaged after 9 days of cell culture at 37°C by placing the scaffolds in the fridge for 30 mins at 4°C (section 1 supplementary data). Cell detachment was observed for all scaffolds with no significant difference in percentage of population detached (Fig. 5A). However, large standard deviations were observed for the scaffolds containing 2 and 10% PDEGMA/PDEGS-Nor-GGG-YIGSR (Fig. 5A). This may have been due to heterogeneous cell attachment to the lowest percentage peptide derivatives (2% PDEGS-Nor-GGG-YIGSR /PDEGMA) and very strong adhesion to the more extensively peptide-functionalised fibres (10% PDEGS-Nor-GGG-YIGSR /PDEGMA). Nonetheless, more consistent cell detachment was observed for the scaffolds containing 3 and 4% PDEGMA/PDEGS-Nor-GGG-YIGSR. Flow cytometry analysis of cell surface markers of cells detached from the scaffolds was carried out to validate immuno-staining findings. Cells cultured on 3% peptide scaffolds were observed to have an increased sub-population of 16% of cells that only expressed the CD34 + /CD105keratocyte marker (Fig. 5B and Tab. S5-6) compared to those cultured all other scaffolds. However, CD34 -/CD105 + and CD34 + /CD105 + sub-populations combined constituted half of the total cell population (53%, Tab. S5) whereas CD34 + /CD105 + and CD34 + /CD105cells comprised 45% of the total cell population. Therefore, a significant percentage of the population exhibited MSC characteristics besides those of the desired keratocyte CD34 marker. A similar sub-population percentage of CD34 -/CD105 + and CD34 + /CD105 + cells (51%) was detected when cultured on the scaffolds with 4% peptide on the surface, nevertheless there was an increase in the proportion of cells which expressed the CD34 marker (38% of total cell population, Tab. S5). Significantly less CD34 + /CD105 + expression was observed for all scaffolds compared to the 2D TCPS control when comparing the Y-mean (Tab. S6). This observation shows that the 3D culture environment significantly reduced the expression of the MSC marker CD105 but no significant difference was seen for the different percentage content of peptides (Tab. S7-8). In addition, an increased expression of ALDH + /α-SMAwas demonstrated when cells were cultured on 3 and 4% peptide scaffolds (Fig. 5C). In terms of the total cell population, 24, 28, 19 and 32% (ALDH + /α-SMAand ALDH + /α-SMAcombined, Tab. S5) for 2, 3, 4 and 10% peptide-functionalised fibres, respectively, were shown to express the α-SMA activated myofibroblastic marker. This was significantly less than the 93% observed for the 2D control. Hence, the 3D structure of the scaffolds containing covalently bound peptide on the surface reduced the differentiation to the undesired activated myo-fibroblastic phenotype compared to the 2D control. Although no significant difference could be observed between the scaffolds for CD34/CD105 expression, a significant decrease of ALDH -/α-SMA + was seen for the scaffolds with 3 and 4% peptide compared to 2D and the adsorbed peptide scaffold when comparing the X-mean (Tab. S6-8). This gives an indication that the covalently attached peptide reduced the change to the activated myofibroblast phenotype and supported increased ALDH expression. In general terms, these data were in accord with the improved results in 3D fibrous scaffolds for cell culture and tissue engineering compared to 2D cultures reported in recent years. The fibrous structure in these materials was designed to mimic ECM properties, resulting in an increase in desired cell response. However, as the synthetic polymers used in our culture assays were non-native to the cells, we anticipated that peptide or protein incorporation or functionalisation would be required to overcome some key practical challenges. These included poor cell attachment and induction of inflammation or undesired cell differentiation due to the sub-optimal mechanical or chemical properties (such as stiffness and hydrophobicity) of conventional polymer fibres. For example, Rodina et al. 38 investigated migration and proliferation of mesenchymal stromal stem cells (MSSCs) on collagen protein modified PLA fibrous scaffolds. They observed more uniform distribution and cell penetration into these modified scaffolds. In addition to proteins, more specific peptide sequences can be used as ligands for enhanced cell attachment. The Arg-Gly-Asp (RGD) collagen peptide sequence is the most commonly used peptide incorporated in biomaterials to increase their cell adhesion properties. 40 Besides, incorporation of known cell attachment peptide sequences (such as RGD), more cell specific peptide sequences can be incorporated to improve selective cell function or differentiation on the materials. 41 Examples include the laminin-derived peptides IKVAV 42 and YIGSR 43 for cell attachment and anti-cancerous response, specific bone morphogenetic protein-2-related peptide P24 and rhBMP2, 44,45 and enzyme responsive peptides (such as Fmoc-propargyl-GAARGD, which can be cleaved to ARGD by the enzyme elastase expressed by porcine pancreas). 46 In this study, we set out a new approach for 3D mammalian cell culture by functionalising thermo-responsive fibrous scaffolds with a cell responsive peptide sequence. The methodology employed here to functionalise the scaffolds is amenable to the attachment of other peptide sequences as desired. In our case the laminin-1 peptide component, YIGSR, known to promote hCSCs interaction and to support the quiescent keratocyte culture on the scaffolds was chosen as an exemplar. 15 Previous studies have demonstrated that laminin improves growth, organisation and differentiation of many cell types. 15, Nonetheless, the use of laminin has many disadvantages such as degradation in culture to inactive fragments, tumorigenesis and induction of immune-responses, which make the full sequence unsuitable for clinical application. 51 This risk can be minimised by selecting small peptide sequences within the laminin structure, which have the ability to improve adhesion and cell migration. One of these sequences is YIGSR, which has been observed to improve cell adhesion and migration. For example, Uzunalli et al. 15 investigated the response of hCSCs on bioactive self-assembled peptide nanofibres. Two peptide sequences were investigated; 1) YIGSR (laminin derived) and 2) RGD (fibronectin derived). In both these cases, hCSCs were observed to maintain their characteristic morphology, and their proliferation was enhanced when YIGSR was present. Significantly, less cell proliferation and adhesion was observed for the RGD fibres. This study suggested that the YIGSR sequence has a direct impact on hCSCs attachment, proliferation and migration on 3D structures. These observations are in line with the increase in proliferation observed of the hCSCs on our scaffolds with increasing ratio of GGG-YIGSR covalently bound on the surface. Many studies have reported improved attachment and proliferation of human cornea epithelial cells. However, only one paper has described the response of hCSCs on YIGSR incorporated scaffolds, 15 and no phenotypic support study has been performed to our knowledge. The immuno-staining and flow cytometry data reported here suggests that not only does YIGSR enhance cell attachment and proliferation on the fibrous scaffold, but the peptide also supports a higher percentage of the cell population which expresses a quiescent phenotype. Down-regulation of the undesired myofibroblastic marker α-SMA was detected, while up-regulation of keratocyte markers CD34 and ALDH was observed. Phenotypic characteristics were observed to remain constant when cultured and passaged, by the thermo-responsive enzymatic free passaging capabilities of the scaffolds. This gives an indication that such scaffolds can be designed to support the desired cell phenotype during in vitro culture and enzyme-free passaging and therefore this system is suitable for therapeutically relevant cell types. ## Conclusions The peptide functionalised, thermo-responsive electrospun scaffolds described here were designed for human cornea stromal cells (hCSCs) proliferation, phenotype support and enzymatic digestion-free passaging. The initially proposed scaffolds (PLA with 10 wt% PDEGMA) did not facilitate hCSCs adhesion and proliferation. Therefore, a peptide-functionalised scaffold was produced to improve the attachment of hCSCs on the scaffold and support the desired quiescent phenotype. PDEGMA/PDEGSH polymers were accordingly synthesised, their thermo-responsive properties and thiol content confirmed, and the materials were successfully blend-electrospun with PLA into fibrous scaffolds. The presence and homogenous distribution of free thiol on the scaffold surfaces, and the subsequent conjugation of Nor-GGG-YIGSR was confirmed by ToF-SIMS analysis. Non-specific attachment of the peptide was observed on thermo-responsive fibres without free thiol functionality but microscopy studies showed aggregated cells on these scaffolds, while elongated and spread cell morphologies were observed on thiol-containing and covalently-bound peptide derivatised scaffolds. These data indicated that enhanced hCSCs interaction was dependent on covalent peptide-functionalisation of the thermo-responsive fibres. The undesired, activated myofibroblastic phenotype was observed to decrease with increased percentage of Nor-GGG-YIGSR conjugated polymers at the scaffold surface. The highest levels of peptide-functionalisation resulted in a decrease in keratocyte specific markers ALDH and CD34, whereas concentrations of 3 and 4% Nor-GGG-YIGSR resulted in the highest expression of keratocyte markers, while suppressing the active myofibroblastic phenotype. Between 80-100% of cultured cells were observed to detach from the scaffolds at temperatures below the phase transition temperature of the thermoresponsive polymers at the surfaces of the fibres. Despite some variation in detachment of the cells from the scaffolds, cells were able to retain their cell phenotype after thermal cycles of attachment and detachment passaging. Hence, it can be concluded that the designed thermo-responsive peptide conjugate fibrous scaffolds are suitable for the attachment, proliferation, quiescent keratocyte phenotype support and temperature-mediated-passaging of hCSCs. In our studies, we used a sequence to aid in hCSCs cell culture and expansion, but the attached peptide sequence could easily be changed to select for a different cell type, making the fabricated scaffolds a versatile matrix for cell expansion and phenotype support of any cell type including those destined for the clinic.

chemsum

{"title": "Adhesion-peptide conjugates of thermo-responsive polymers co-electrospun into fibrous scaffolds enable growth and enzyme-free recovery of quiescent human Corneal Stromal Cells (hCSCs)", "journal": "ChemRxiv"}

intramolecular_through-space_charge-transfer_emitters_featuring_thermally_activated_delayed_fluoresc

## Abstract: Through-space charge-transfer (TSCT) emitters have been extensively explored for thermally activated delayed fluorescence (TADF), but arranging various donors and acceptors into rigid cofacial conformations for various efficient TSCT TADF emitters has still remained a challenge. Here we report for the first time a "fixing acceptor" design to reach various efficient TSCT TADF emitters. By chemically fixing the acceptor (benzophenone) with a rigid spiro structure and cofacially aligning various donors with the fixed acceptor, a series of efficient TSCT TADF emitters have been developed. Single-crystal structures and theoretical calculations have verified closelypacked cofacial donor/acceptor conformations and favorable TSCT in the emitters. The emitters afford sky-blue to yellow TADF emission in doped films, with high photoluminescent efficiencies of up to 0.92 and reverse intersystem crossing rates of up to 1.0×10 6 s −1 . Organic light-emitting diodes using the emitters afford sky-blue to yellow electroluminescence with high external quantum efficiencies of up to 20.9%.The work opens a new avenue toward a wide variety of efficient TSCT TADF emitters. ## Introduction Charge-transfer emitters with thermally activated delayed fluorescence (TADF) have aroused huge research interest for their wide applications in light-emitting devices , photocatalysis , and biological imaging , et al. They exhibit fast reverse intersystem crossing (RISC) from non-emissive triplet states (T1) to emissive singlet states (S1), due to the small singlet-triplet energy gaps (ΔEST) (< 0.2 eV) produced by spatial separation of the highest occupied molecular orbitals (HOMOs) and the lowest unoccupied molecular orbitals (LUMOs) . With advantages of triplet-harvesting and low-cost, they have demonstrated a bright prospect for the use in electroluminescence (EL) devices . While through-bond charge-transfer TADF emitters have been widely developed by directly linking donor (D) and acceptor (A) units via chemical bonds or conjugated π-linkers and carefully twisting the D/A conformations , through-space chargetransfer (TSCT) TADF emitters have recently attracted fast-growing interest . In the simplest manner, TSCT systems can be developed by physically blending D and A molecules . Such TSCT systems (also known as exciplex systems) can afford tiny ΔEST values (~ meV) due to the almost complete spatial separation of HOMO and LUMO orbitals, which are highly favorable for TADF [6, . However, they usually afford low luminescent efficiencies, due to weak TSCT and pronounced non-radiative decay losses, resulting from loose packing and random orientations of the D/A molecules . To solve this challenge, the D/A molecules have been held close by non-conjugated σ-linkers . Nevertheless, these D-σ-A TSCT systems still show relatively low luminescent efficiencies because of the lack of control over the relative orientations of the D and A units. Recent researches have proved that cofacial D/A alignments are favored for developing efficient TSCT emitters , which can reinforce TSCT and promote radiative decay of charge-transfer states . Very recently, Jiang and Liao et al. have unveiled that fixation of the cofacial D/A conformation is critically important for developing efficient TSCT emitters, which can remarkably suppresses non-radiative decay of charge-transfer states by restricting the rotation or vibration of the D/A units . The fixation has been achieved by chemically fixing the 9,10-dihydroacridine type D units and confining the D/A units into rigid cofacial conformations . By this design, a series of highly efficient TSCT TADF emitters have been developed, which have proved high performances in EL devices . Unfortunately, in this design, the D units are confined to be derivatives of 9,10-dihydroacridine , which imposes some restrictions on donor variation and further molecular engineering. It has still remained challenging to develop various efficient TSCT TADF emitters. Herein, we report for the first time a "fixing acceptor" design to reach various efficient TSCT TADF emitters. As shown in Scheme 1, TSCT emitters 1-5 are developed by chemically fixing the A unit (benzophenone) and confining the D/A units into closely-packed cofacial conformations. The A unit is fixed with a spiro-structure derived from a carbonyl group of the A precursor. The D units are facilely varied with different structures and electron-donating strengths, which are 1,3,6,8-tetramethyl-9Hcarbazole (Me4CAZ), 9,9-dimethyl-9,10-dihydroacridine (DMAC), diphenylamine (DPA), triphenylamine (TPA), and benzo oxazino [2,3,4-kl]phenoxazine (BOP) for 1-5, respectively. Single-crystal X-ray structures and theoretical calculations have verified closely-packed cofacial D/A conformations and favorable TSCT in 1-5. In doped films, 1-5 afford strong sky-blue to yellow TADF emission, with high photoluminescent efficiencies (ΦPL) of up to 0.92 and RISC rates (kRISC) of up to 1.0×10 6 s -1 . Organic light-emitting diodes (OLEDs) using 1-5 as the emitters show skyblue to yellow EL with high external quantum efficiencies (EQEs) of up to 20.9%. Scheme 1. Synthetic routes to TSCT emitters 1-5. ## Results and discussion Scheme 1 outlines the general synthetic route to 1-5. The rigid spiro-structure in a fluorene linker is utilized to fix benzophenone as the A unit. The D-containing compounds M1-M5 firstly reacted with n-butyllithium to form the phenyllithium reagents, which then reacted with one carbonyl group of the A precursor (anthraquinone) to afford the tertiary alcohol intermediates including both D and A units. By intramolecular Friedel-crafts cyclization reactions in the tertiary alcohol intermediates, 1-5 were prepared with total yields of 40−64%. A key design is to direct the cyclization to occur at the ortho-position of the D unit by blocking the other cyclization site with a methyl group. Scheme S1 outlines the synthetic routes to the key intermediate compounds M1-M5. M3 was prepared by a Cu(I) catalyzed Ullman reaction between 2'-bromo-6-methyl-[1,1'-biphenyl]-3-amine (S1) and aniline, while M1, M2, M4 and M5 were synthesized by Pd(II) catalyzed Buchwald-Hartwig or Suzuki reactions between 2'-bromo-5-iodo-2-methyl-1,1'-biphenyl (S2) and D units or boronic acid derivatives of D units. Detailed synthesis procedures to S1, S2, M1-M5 and 1-5 are presented in Supporting Information. It is worth mentioning that the synthetic route reported here can apply to not only various D units but also different A precursors that carry a carbonyl group serving as the spiro-fixing site. S2), respectively, which indicates strong or considerable face-to-face π-π stacking interactions . Theoretical calculations were conducted with density functional theory (DFT) (See ESI). Figure 1b-d show the calculation results of the ground-states (S0). On the optimized S0 states, the closely-packed D/A units exhibit considerable non-covalent van der Waals interactions, as revealed by the reduced density gradient (RDG) isosurfaces (Figure 1b and Figure S2), which agree with the π-π stacking interactions between the D and A units . The HOMOs of 1, 2, 4 and 5 distribute predominantly over the D units, while the HOMO of 3 distributes over the diphenylamine unit, extending to the connected phenyl ring of the fluorene linker (Figure 1c). The LUMOs of 1-5 are all concentrated on the spiro-fixed benzophenone units (Figure 1d), with nearly no distributions on the fluorene linkers, due to the interrupted conjugation by the spirocarbons. It is shown that the HOMOs and LUMOs of 1-5 exhibit nearly no overlaps on the fluorene linkers, which minimizes through-bond charge-transfer. In contrast, the HOMOs and LUMOs largely overlap through the space due to the close cofacial D/A packing, which facilitates TSCT . Figure S3 depicts the natural transition orbital (NTO) analysis based on time-dependent DFT (TDDFT) calculations. For the S0 → S1 and S0 → T1 transitions, the hole and particle orbitals are predominantly delocalized over the D units and the spiro-fixed benzophenone units, respectively, which confirms the TSCT character of the S1 and T1 states. The calculated ΔEST values of 1-5 are as small as 8−27 meV (Table S3), which agrees with the almost complete spatial separation of HOMOs and LUMOs. The S1 and T1 states were further optimized with the TDDFT method. The relaxed S1 and T1 geometries show better cofacial D/A alignments than the S0 geometries (Figure S4), as revealed by the shorter distances and more parallel alignments (except 3) between the D and A units (Table S2), due to the Coulomb attractive interactions on excited-states . Figure S5 shows the absorption and photoluminescence (PL) spectra of 1-5 in toluene solution. Table S4 summarizes the photophysical characteristics. In addition to the strong absorption bands (below 350 nm for 1-4 and below 400 nm for 5) that arise from the D/A-centered 1 π-π * / 1 n-π * transitions, 1-5 exhibit weak charge-transfer absorption bands at around 380 nm for 1-4 and 410 nm for 5 (Figure S5a). The emergence of charge-transfer absorption indicates strong electronic coupling between the D and A units on ground states, due to the marked D/A overlap . In the toluene solution, 1-5 afford green to orange emission centered at 514, 548, 524, 522 and 571 nm (Figure S5b), respectively. The different emission energies for 1-5 result from the different electron-donating strengths of the D units, as shown by electrochemical measurements (vide infra). The gradual red-shift of the emission by increasing the solvent polarity agrees with the charge-transfer character of the emission (Figure S6). In degassed toluene solution, 1-5 afford ΦPL below 0.10 (Table S4), which suggests considerable non-radiative decay losses in the fluid solution, presumably caused by vibrations or rotations of the unfixed D units (especially for 3 and 4 with rotary phenyl rings) . 1-3 show both prompt and delayed components in the transient PL decays (Figure S7 and Table S4), reflecting the TADF feature of the emission, whereas 4 and 5 do not show delayed components, presumably because the T1 states are mostly consumed by non-radiative deactivations before being converted to the S1 states. 1-5 all display aggregation-induced emission phenomena, as revealed by the remarkable enhancement of PL intensity upon increasing the water fraction in the mixed THF/H2O solution (Figure S8), due to suppressed vibrations or rotations of molecular fragments in aggregation states . from further suppression of intramolecular vibrations/rotations in the solid state . Figure S9 shows the low-temperature fluorescence and phosphorescence spectra of the doped films. From the onsets of the fluorescence and phosphorescence spectra, singlet (ES) and triplet (ET) energies were respectively calculated (Table 1). 1-5 afford ΔEST values at 10−70 meV, which are sufficiently small for rapid RISC. Figure 2b shows the transient PL decay curves. 1-5 all show distinct prompt and delayed components in the transient PL decays, with prompt (τp)/delayed (τd) lifetimes at 64 ns/6.8 μs, 457 ns/10.9 μs, 130 ns/5.9 μs, 132 ns/5.3 μs and 99 ns/3.3 μs, respectively. The delayed components are remarkably suppressed at low temperature (Figure S10), which confirms the TADF nature of 1-5. Table S5 summarizes the calculated rate constants of the emission in the doped films. 1-5 show radiative decay rates of singlet state (kr,s) at 0.7−3.3×10 6 s -1 , which are within the normal range reported for TSCT emitters [35,37, 45]. 1-4 show non-radiative decay rates of singlet state (knr,s) at 1.8−7.3×10 5 s -1 , whereas yellow-emitting 5 exhibits a much larger knr,s of 2.0×10 6 s -1 , in accordance with the energy gap law. Notably, 1-5 exhibit high kRISC values at 9.9×10 5 s -1 , 2.3×10 5 s -1 , 3.3×10 5 s -1 , 6.1×10 5 s -1 and 7.1×10 5 s -1 , respectively, indicating their high capability of converting triplets to singlets for emission. Figure S11 shows the cyclic voltammograms measured in solution. 1-5 show similar electrochemical reductions and LUMO levels (around −2.70 eV) but different electrochemical oxidations and HOMO levels (-5.23 ~ -4.99 eV) (Table S6), because they use the same A unit and different D units. The Me4CAZ, DPA and TPA donors have similar electron-donating strengths, which lead to similar HOMO levels (around −5.2 eV) and emission energies for 1, 3 and 4 (Figure 2a). The DMAC and BOP donors have increased electron-donating strengths, which lead to destabilized HOMO levels (−5.14 and −4.99 eV, respectively) and decreased emission energies for 2 and 5 (Figure 2a). As shown by thermal gravity analysis (Figure S12), 1-5 all possess high thermal stability, with 5% weight loss temperatures at 401, 385, 384, 417 and 444 ºC, respectively. To evaluate their application potential in EL devices, 1-5 were employed as emitters in vacuum-evaporated OLEDs. Figure 3a Table 2. Summary of EL performances of OLEDs using 1-5 as the emitters. Figure 3c shows the current-density and luminance versus voltage curves. Table 2 summarizes the device performances. The OLEDs based on 1-5 showed low turn-on voltages below 3.0 V and high peak current efficiencies/EQEs at 50. 2). Efficiency roll-offs in TADF OLEDs are usually ascribed to triplet-triplet/triplet-polaron annihilations in the emissive layers . The high kRISC values of 1-5 indicate that they can efficiently consume triplets for emission, alleviating triplet-triplet/triplet-polaron annihilations . By broadening the recombination zones with optimized device structures, the efficiency roll-offs would be suppressed . The high device efficiencies indicate that 1-5 are promising functional materials for high-performance EL devices. ## Conclusion We reported a "fixing acceptor" design for efficient TSCT TADF emitters. By chemically fixing the A unit (benzophenone) and cofacially aligning various D units with the fixed A unit, TSCT TADF emitters 1-5 were developed. Single-crystal structures and theoretical calculations verified closely-packed cofacial D/A conformations and favorable TSCT in 1-5. In doped films, 1-5 afforded high ΦPL of up to 0.92 and kRISC of up to 1.0×10 6 s −1 . OLEDs using them as emitters showed high EQEs of up to 20.9%. The work has demonstrated that fixing the acceptor can afford efficient TSCT TADF emitters. In principle, the "fixing acceptor" design allows facile variation of donor and feasible variation of acceptor as long as the acceptor precursors carry a carbonyl group that serves as the spiro-fixing site (such as thioxanthen-9-one sulfone and its derivatives). Varying the fixed acceptor is currently ongoing in our lab. The work opens a new avenue toward a wide variety of efficient TSCT TADF emitters, which are promising functional materials for high-performance EL devices.

chemsum

{"title": "Intramolecular Through-Space Charge-Transfer Emitters Featuring Thermally Activated Delayed Fluorescence for High-Efficiency Electroluminescent Devices", "journal": "ChemRxiv"}

copper(<scp>ii</scp>)_ketimides_in_sp³_c–h_amination

## Abstract: Commercially available benzophenone imine (HN]CPh 2 ) reacts with b-diketiminato copper(II) tertbutoxide complexes [Cu II ]-O t Bu to form isolable copper(II) ketimides [Cu II ]-N]CPh 2 . Structural characterization of the three coordinate copper(II) ketimide [Me 3 NN]Cu-N]CPh 2 reveals a short Cu-N ketimide distance (1.700(2) Å) with a nearly linear Cu-N-C linkage (178.9(2) ). Copper(II) ketimides [Cu II ]-N]CPh 2 readily capture alkyl radicals Rc (PhCH(c)Me and Cyc) to form the corresponding R-N]CPh 2 products in a process that competes with N-N coupling of copper(II) ketimides [Cu II ]-N]CPh 2 to form the azine Ph 2 C]N-N]CPh 2 . Copper(II) ketimides [Cu II ]-N]CAr 2 serve as intermediates in catalytic sp 3C-H amination of substrates R-H with ketimines HN]CAr 2 and t BuOO t Bu as oxidant to form N-alkyl ketimines R-N]CAr 2 . This protocol enables the use of unactivated sp 3 C-H bonds to give R-N]CAr 2 products easily converted to primary amines R-NH 2 via simple acidic deprotection. ## Introduction Transition metal-catalysed sp 3 C-H amination protocols have gained immense attention in the synthetic community over the past couple of decades. A majority of these protocols proceed via metal-nitrene 2,5 [M]]NR 0 or metal-amide [M]-NR 0 R 00 intermediates. 1,6 Extensive studies on such intermediates and underlying mechanisms have paved the way towards more efficient sp 3 C-H amination protocols. 1 Related metal-ketimide [M]-N]CR 0 R 00 intermediates, however, have received less attention in C-H amination chemistry. The strong metal-N ketimide interaction makes ketimides effective spectator ligands. For instance, ketimides stabilize high valent homoleptic Mn(IV), 7 Fe(IV) 8 and Co(IV) 9 complexes (Fig. 1a). In some cases, ketimides can also form via nickel and copper arylimido/nitrene intermediates [M]]NAr via C-C coupling at the para-position of the aryl nitrene ligand (Fig. 1b). While this reactivity was initially uncovered with nickel bdiketiminato complexes, 10 reversible C-C bond formation/ cleavage in related copper complexes provides access to terminal copper nitrenes [Cu]]NAr that participate in sp 3 C-H amination. 11,12 Fewer examples of ketimides exist, however, in which the ketimide ligand serves as a reactive functional group in discrete transition metal complexes. 13 Metal ketimide intermediates have been proposed in several Pd-catalysed cross-coupling reactions of aryl (Fig. 1c) 14 and alkyl halides (Fig. 1d) 15 with benzophenone imine. Cu-catalysed photoredox cross-coupling reactions of redox-active alkyl esters (Fig. 1e) 16 and Cu-catalysed benzylic sp 3 C-H amination with benzophenone imine (Fig. 1f) 17 are among other examples that may be mediated by metal-ketimide intermediates. Moreover, Stahl and colleagues have proposed copper(II) ketimides in the N-N oxidative coupling of imines Ar 2 C]NH to azines Ar 2 C]N-N]CAr 2 under aerobic or electrocatalytic conditions (Fig. 1g). 18,19 Herein we describe discrete frst-row transition metal-ketimide complexes intimately involved in C-H amination chemistry. Building upon the Kharasch-Sosnovsky reaction, we previously demonstrated that copper(II) alkyl amides [Cu II ]-NHR 0 , 23 anilides [Cu II ]-NHAr, 6,24 and aryloxides [Cu II ]-OAr 25 serve as key intermediates in a radical relay protocol for sp 3 C-H functionalisation (Fig. 2). Formed via acid-base 6,23,24 or trans-esterifcation 25 reactions between [Cu II ]-O t Bu with H-FG or Ac-FG reagents, these copper(II) complexes [Cu II ]-FG capture sp 3 -C radicals Rc generated via H-atom abstraction from R-H to furnish the functionalized product R-FG. We anticipated that the relatively high acidity of the imine N-H bond 26 coupled with a preference for binding at copper with softer N-donors should enable the formation of [Cu II ]-N]CAr 2 species from [Cu II ]-O t Bu complexes and HN]CPh 2 allow for an examination of copper(II) ketimides in C-H amination catalysis. ## Synthesis and characterization of copper(II) ketimides Monitored by UV-vis spectroscopy, addition of benzophenone imine (1 equiv.) to a solution of [Me 3 NN]Cu-O t Bu (2a) in toluene at 80 C results in decay of the characteristic UV-vis absorption of 2a at 470 nm with growth of a new band at 570 nm (Fig. S2 †). Performed on a preparative scale, this new species [Me 3 NN]Cu-N]CPh 2 (3a) may be isolated as dark purple crystals from pentane at 35 C in 78% yield (Fig. 3a). The X-ray crystal structure of [Me 3 NN]Cu-N]CPh 2 (3a) (Fig. 3a) reveals the Cu-N ketimide distance of 1.700(2) , signifcantly shorter than the Cu-N bond found in the copper(II) amide [Cl 2 NN]Cu-NHAd (1.839(9) ) 23 and copper(II) anilide [Cl 2 NN]Cu-NHAr Cl 3 (1.847(3) ). 6 Copper(II) ketimide 3a possesses a nearly linear Cu-N3-C24 angle of 178.9(2) . The short Cu-N ketimide distance and linear Cu-N3-C24 angle support effective sp hybridization at the ketimide N atom. These values remarkably differ from those in the homoleptic copper(I) ketimide [Cu-N]CPh 2 ] 4 with bridging ketimide ligands that lead to a square-like tetrameric structure with Cu-N distances 1.847(2)-1.861(2) and Cu-N-Cu angles of 94.17(9)-98.25 (9). 27 To outline differences between coordination of anionic ketimide ligands and their neutral ketimine counterparts, we prepared the corresponding benzophenone imine adducts [Me 3 NN]Cu(NH]CPh 2 ) (4a) and [Cl 2 NN]Cu(NH]CPh 2 ) (4b) (Fig. 3b). These copper(I) complexes feature substantially longer Cu-N ketimine distances of 1.8940( 14) and 1.8937( 14) . These ketimine adducts 4a and 4b each exhibit a pronounced bend in the Cu-ketimide linkage with Cu-N-C angles of 132.68 (12) and 130.25 (12) consistent with sp 2 hybridization at N. UV-vis analysis of copper(II) ketimide [Me 3 NN]Cu-N]CPh 2 (3a) reveals the presence of a single low energy absorption band at 570 nm (3 ¼ 1910 M 1 cm 1 ) in toluene at room temperature. The EPR spectrum of 3a in a mixture of toluene and pentane at room temperature shows a signal centred at g iso ¼ 2.081 with very well resolved coupling to 63/65 Cu (A Cu ¼ 298.0 MHz) and additional hyperfne modelled with three equivalent 14 DFT calculations reveal remarkably high unpaired electron density on the ketimide N atom of both 3a (0.58) and 3b (0.61) (Fig. 4 and S23 †). These values are signifcantly higher than values reported for related three coordinate b-diketiminato Cu(II) anilides [Cu II ]-NHAr (0.23-0.25) 6 and a copper(II) amide [Cu II ]-NHAd (0.49). 23 We rationalize this as a result of a 2-center 3-electron p interaction between the highest energy d orbital at the copper(II) center destabilized by the b-diketiminato Ndonors and a p orbital of the sp-hybridized ketimide N atom (Fig. 4a). In addition, the orthogonal orientation of the Cu-N ketimide p-interaction relative to the conjugated ketimide N] CPh 2 p system further limits the delocalization of unpaired electron density away from the ketimide N atom (Fig. 4b and c). ## Copper(II) ketimide reactivity: radical capture and N-N bond formation The ability of many b-diketiminato copper(II) complexes to participate in catalytic sp 3 C-H functionalisation via radical relay (Fig. 2) encouraged us to assess the reactivity of copper(II) ketimides 3 towards alkyl radicals. We fnd that [Cu II ]-N]CPh 2 species 3a and 3b capture alkyl radicals Rc to provide the corresponding R-N]CPh 2 products (Fig. 5a). [Cu I ] is anticipated to form in these radical capture reactions that correspond to step d in the radical relay catalytic cycle (Fig. 2). For instance, reaction of 3a and 3b with (E/Z)-azobis(a-phenylethane) at 90 C that generates the benzylic radical PhCH(c)Me upon heating provides the alkylated imine PhCH(N]CPh 2 )Me in 40% and 74% yields, respectively. Generation of Cyc radicals in the presence of 3a and 3b by heating t BuOO t Bu in cyclohexane (via H-atom abstraction by t BuOc radicals) provides Cy-N]CPh 2 in 58% and 41% yields, respectively. Upon heating to 60 C, copper(II) ketimides 3a and 3b undergo N-N coupling to form benzophenone azine Ph 2 C]N-N]CPh 2 isolated in 66% and 90% yields, respectively (Fig. 5b). This represents a competing reaction for radical capture at copper(II) ketimides 3a and 3b. ## Copper(II) ketimides in sp 3 C-H amination With a fundamental understanding of copper(II) ketimide formation and reactivity, we explored these complexes in catalytic C-H amination via radical relay. Using ethylbenzene as a model R-H substrate, we screened a modest range of copper(I) b-diketiminato catalysts 1 that possess different electronic and steric properties (Table 1). The catalyst [Cl 2 NN]Cu (1b) provides the highest yield compared to more electron-rich (1a and 1c) and electron-poor (1d) catalysts. Increasing the t BuOO t Bu oxidant amount does not signifcantly improve the yield. Lowering the temperature from 90 C reduces the yield drastically (Table S1 †), possibly due to binding of the ketimine HN] CAr 2 to the copper(I) catalyst (Fig. 3b) that inhibits t BuOO t Bu activation. 28 While (1-(tert-butoxy)ethyl)benzene forms in trace amounts via C-H etherifcation, 28 the azine Ph 2 C]N-N]CPh 2 is the main byproduct in these catalytic C-H amination reactions, representing non-productive consumption of H-N]CPh 2 . In a previous study of C-H amination with anilines H 2 NAr employing the [Cl 2 NN]Cu/ t BuOO t Bu catalyst system, electronpoor anilines provided the highest yields in the face of competing diazene ArN]NAr formation. 24 Copper(II) anilido intermediates [Cu II ]-NHAr serve as intermediates in C-H amination with anilines H 2 NAr; those derived from electron-poor anilines H 2 NAr (e.g. Ar ¼ 2,4,6-Cl 3 C 6 H 2 ) proved more resistant to reductive bimolecular N-N bond formation. 6,24 To examine whether similar electronic changes in the ketimine H-N]CAr 2 could similarly promote more efficient catalysis, we explored two electron-poor ketimine derivatives H-N] CAr 2 (Ar ¼ 4-CF 3 C 6 H 4 and 4-FC 6 H 4 ) in C-H amination (Table 2). Although the p-CF 3 substituted imine provides a higher C-H amination yield with cyclohexane (C-H BDE ¼ 97 kcal mol 1 ), 29 the increase in yield is modest with the benzylic substrate ethylbenzene (C-H BDE ¼ 87 kcal mol 1 ). 29 No signifcant differences were observed between benzophenone imine and the p-F substituted analogue. While electron-poor imines can give somewhat higher C-H amination yields, we most broadly examined the commercially available H-N]CPh 2 to survey the scope of R-H substrates in sp 3 C-H amination (Table 3). Ethers such as THF, 1,4-dioxane, or even 12-crown-4 undergo C-H amination at the a-carbon in relatively high yields (6a-6d). Amination of the benzylic secondary C-H bonds in heteroaromatic substrates occurs (6f-6g), though yields may be lower due to the possibility of coordination of these substrates and/or products to the copper(I) centre that can decrease the rate of reoxidation with t BuOO t Bu. 28 Aromatic substrates with benzylic C-H bonds undergo C-H amination in moderate to high yields (6h-6k). Cycloalkanes with stronger, unactivated sp 3 C-H bonds give moderate yields with electron-poor ketimine HN]CAr 0 2 (Ar 0 ¼ 4-CF 3 C 6 H 4 ) (6l-6o). The bicyclic eucalyptol undergoes C-H amination in 32% yield (6e). These aminated products may be isolated either as synthetically versatile protected primary amines R-N]CPh 2 via column chromatography (6a-6g) or as the primary ammonium salts [R-NH 3 ]Cl via deprotection upon simple acidic work up (6h-6o) under mild conditions. The potential to use recovered benzophenone from deprotection of ketimine products and azine byproducts to regenerate the Ph 2 C]NH starting material 30 enhances the overall atom economy of this amination protocol. ## Conclusions The isolation of mononuclear copper(II) ketimides [Cu II ]-N] CPh 2 reveals the role that they play as intermediates in sp 3 3; entries 6l-6o). Nonetheless, facile N-N bond formation also by copper(II) ketimides [Cu II ]-N]CAr 2 underscores the role that they may play in the (electro) catalytic copper(II) promoted oxidative N-N coupling of benzophenone imine to form benzophenone azine (Fig. 1g). 18 ## Experimental section Detailed experimental procedures are provided in the ESI. †

chemsum

{"title": "Copper(<scp>ii</scp>) ketimides in sp³ C\u2013H amination", "journal": "Royal Society of Chemistry (RSC)"}

telomerase_and_poly(adp-ribose)_polymerase-1_activity_sensing_based_on_the_high_fluorescence_selecti

## Abstract: Telomerase and poly(ADP-ribose) polymerase-1 (PARP-1) are two potential cancer biomarkers and are closely related to tumor initiation and malignant progression. TOTO-1 is well-known for differentiating ss-DNA from ds-DNA because it is virtually non-fluorescent without DNA and exhibits very low fluorescence with ss-DNA, while it emits strong fluorescence with ds-DNA. In this paper, for the first time, it was found that TOTO-1 has high fluorescence selectivity and sensitivity towards the G bases in single-stranded DNA and poly(ADP-ribose) (PAR). Poly(dG) was used as the model target to explore its possible mechanism. Molecular dynamics (MD) simulation proved that intramolecular p-p stacking existed in TOTO-1 (in an aqueous solution), while intermolecular p-p stacking formed between TOTO-1 and poly(dG) in a similar way as that observed for dsDNA. Interestingly, telomerase and PARP-1 catalyzed the formation of G-rich DNA and PAR in vivo, respectively. Therefore, TOTO-1 was explored in detecting both of them, obtaining satisfactory results. To the best of our knowledge, no probe has been reported to recognize PAR. It is also the first time where telomerase is detected based on the specific recognition of G bases. Importantly, integrating multiple functions into one probe that can detect not only telomerase but also PARP-1 will significantly raise the specificity of screening cancer and decrease false positive proportion, which make TOTO-1 a promising candidate probe for clinical diagnosis and pharmaceutical screening. Telomerase and poly(ADP-ribose) polymerase-1 activity sensing based on the high fluorescence selectivity and sensitivity of TOTO-1 towards G bases in single-stranded DNA and poly(ADPribose) † Introduction Telomerase and poly(ADP-ribose) polymerase-1 (PARP-1) are two potential cancer biomarkers and are closely related to tumor initiation and malignant progression. 1,2 Integrating multiple functions into one probe that can detect not only telomerase but also PARP-1 will signifcantly raise the specifcity of screening cancer and decrease false positive proportion. To the best of our knowledge, few probes have been reported to detect not only telomerase but also PARP-1. Most telomerase detection methods depend on the unique property of its extended primer, 5 0 -AAT CCG TCG AGC AGA GTT (TTAGGG) n -3 0 . First, the extended primer is a powerful displacement strand that can be used to initiate the signal "turn-on" or "turn-off" based on the hybridization of DNA. Second, the extended G-rich primer forms a G-quadruplex, which has high peroxidase-like activity and can initiate signals closely related to the catalyzed reaction. Third, the extended DNA sequence has ample negative charges that have great influence on the signal depending on the electrostatic interactions. Only few telomerase detection methods have been reported based on the specifc recognition of bases. Most detection methods for PARP-1 are based on its catalyzed production of PAR that has a large number of negative charges, which can adsorb positively charged probes by electrostatic attraction. These methods do not have good selectivity because of the high background signals. Because PAR lacks unique properties as the extended telomerase primer, very few methods have been developed for its detection even though its activity is very important for assessing cancer development. 16,17 Fluorescence analysis has been widely applied in detecting biomolecules because of its inherent advantages of being simple, convenient and sensitive. The fluorescence properties of dyes differ widely when they are in different environments, and they are strongly dependent on the local environment of the macromolecules and their circumstances. 18 DNA carries the genetic information of all known living organisms. The ability to specifcally discriminate DNA sequences is signifcant for disease screening and oncology studies. 19 Some fluorescence probes that specifcally recognize DNA sequences have been developed. Fluorescence nanoparticles such as DNA-silver nanoclusters and small organic dyes such as thioflavin T, ethidium bromide (EtBr), 30 thiazole orange (TO), 31 TOTO, 32 and iridium complex 33 can bind to DNA chains and show various fluorescence properties depending on their manner of binding and binding ability. Some of them can distinguish ss-DNA from ds-DNA or G-quadruplex DNA. However, only a few fluorescence probes are known to effectively distinguish deoxynucleotides 34 even though they are signifcant for DNA sequence discrimination, bio-analysis and clinical diagnosis. TOTO-1, an unsymmetrical cyanine dye dimer, is well-known for differentiating ss-DNA from ds-DNA (Fig. S1 †). 35,36 It is virtually non-fluorescent without DNA and exhibits strong fluorescence when it binds to ds-DNA (Fig. S2 †). 37,38 In this paper, for the frst time, we found that TOTO-1 has unique fluorescence selectivity and high sensitivity towards the G bases in single-stranded DNA and PAR. Its fluorescence intensity was much stronger with poly(dG) than those with poly(dA), poly(dC) and poly(dT) (Fig. S2A †). More interestingly, the fluorescence of TOTO-1 with poly(dG) was far stronger than that with dGTP even when the concentration of dGTP was 4 orders of magnitude higher than that of the G bases in poly(dG). Furthermore, the fluorescence of TOTO-1 with PAR increased sharply compared to those with higher concentrations of ADP or poly(dA). These fndings urged us to develop a simple and sensitive method to assay the activities of both telomerase and PARP-1; they catalyzed the formation of a G-rich elongated TS primer and PAR in vivo, respectively, which were highly sensitive to TOTO-1. The high background signals produced by the telomerase primer and the activated dsDNA were reduced by 9 times and 38 times, respectively, under the influence of digestion by the Exo III enzyme. As a result, the detection sensitivity improved greatly. A linear relationship between the fluorescence intensities and logarithmic concentrations of telomerase was observed between 13 and 4000 cells per mL, and the limit of detection (LOD) was determined to be 13 cells per mL. Also, the fluorescence intensities showed a linear response between 0.02 U and 1.5 U for PARP-1 with LOD of 0.02 U. This simple and sensitive approach is promising for clinical diagnosis and is a powerful tool for pharmaceutical screening. ## Molecular dynamics (MD) simulation The initial single-stranded poly(dG) structure was constructed using the NAB module in the AmberTools16 package. 39 Then, the confguration of TOTO-1 binding to poly(dG) was constructed using AutoDock 4.0 with the Lamarckian genetic algorithm. 40 The ground-state geometry of free TOTO-1 molecule was frst optimized at the M06-2X/6-311++G(d,p) level with the polarizable continuum model (PCM), and the RESP charge of the TOTO-1 molecule was calculated using the GAUSSIAN09 program. 41 The confgurations of TOTO-1 without and with poly(dG) at the center with 8000 water molecules were simulated at 298 K in cubic boxes of 56.61 56.61 56.61 3 and 62.46 62.46 62.46 3 , respectively. The generalized amber force feld (GAFF) and FF99BSC0 force feld were used for the TOTO-1 molecule and poly(dG), respectively. All the water molecules were described by the TIP4PFB force feld, which was found to be able to reproduce the experimental properties of liquid water. 42 Cl and Na + counterions were added to neutralize the system of TOTO-1 without and with poly(dG) in aqueous solutions, respectively. The simulations of the two systems were run on the NPT ensemble at 1.0 bar with a Berendsen barostat. Periodic boundary conditions (PBC) were carried out and the long-range electrostatic interactions were taken into account using the particle mesh Ewald (PME) method. The velocity Verlet algorithm was employed to solve Newton's motion equations with the bonds of hydrogen atoms constrained. The van der Waals cutoff was set as 10 for the two systems. For TOTO-1 without and with poly(dG) in aqueous solutions, two simulation trajectories of 20 ns after 2 ns of equilibrium were carried out and snapshots were collected at the interval of 1 ps with every time step of 1 fs. ## Hybrid quantum mechanical and molecular mechanical (QM/ MM) calculation The electronic absorption spectra of TOTO-1 without and with poly(dG) in aqueous solutions were calculated using the QM/ MM method at the level of time-dependent density functional theory (TDDFT) with the M06-2X functional and the 6-31G(d) basis set. TOTO-1 molecule and TOTO-1 with poly(dG) were treated in the QM part for the two systems, and the remaining molecules including the counterions and water molecules were treated as background point charges in the MM region. ## Procedures of telomerase activity and inhibition efficiency evaluation Cell cultivation, telomerase extraction and primer extension reaction were performed according to previously described methods (ESI †). The extended G-rich sequence was hybridized with cDNA (complementary to the TS primer) at 95 C for 5 min, followed by cooling down slowly to room temperature. Then, Exo III was used to digest the duplex part formed by cDNA and TS primer at 37 C. The resulting solution (30 mL) was diluted to 200 mL with PBS buffer containing 250 nM TOTO-1. Then, the fluorescence spectra were recorded. Various concentrations of BIBR 1532 or curcumin were incubated with 750 cells per mL A549 cells to evaluate their telomerase inhibition efficiency. Heated A549 cells (95 C, 5 min) were used for control experiments. ## Procedures of PARP-1 activity and inhibition efficiency evaluation First, active dsDNA was prepared by hybridizing two complementary sequences ssDNA1 and ssDNA2 in a hybridization buffer at 95 C for 5 min, followed by cooling to room temperature slowly. Then, dsDNA (150 nM) was incubated with PARP-1, NAD + (250 mM) and reaction buffer (R-buffer) for 1 h at room temperature. After PARP-1 catalyzed the polymerization of PAR, dsDNA was digested by Exo III at 37 C to reduce the background signal. The resulting solution was mixed with PBS buffer containing 250 nM TOTO-1. Fluorescence spectra were recorded. For PARP-1 inhibitor assay, the same procedure was carried out except for the treatment of PARP-1 with different concentrations of AG014699. ## Result and discussion High fluorescence selectivity and sensitivity of TOTO-1 towards the G bases in single-stranded DNA and PAR The fluorescence selectivities of TOTO-1 towards 15 mM dGTP, dATP, dCTP, dTTP and ADP were detected (Fig. S3 †). The results showed that TOTO-1 has much stronger fluorescence with dGTP or ADP than that with dATP, dCTP or dTTP. Considering that TOTO-1 was used to respond to G-rich DNA sequences catalyzed by telomerase, poly(dG) was used as the model target in this study. Compared with the result for 15 mM of dGTP, the fluorescence intensity of TOTO-1 increased from 1.5 to 25 when only 86 nM poly(dG) (690 nM of G bases) was present. These data indicated that TOTO-1 is more sensitive to the G bases existing in single-stranded DNA than single dGTP, which is signifcant for the detection of extended DNA sequences catalyzed by telomerase. The absorbance and fluorescence spectra of TOTO-1 are shown in Fig. 1. From curves (a and b) in Fig. 1A-D, it can be seen that their absorption spectra vary signifcantly in the presence of different oligodeoxynucleotides, indicating that the molecular orbitals of TOTO-1 were changed by these oligodeoxynucleotides. The orbital distribution of the excited state transitions for TOTO-1/poly(dG) corresponding to the absorbance at 454 nm and 542 nm were simulated (Fig. S4 †). The fluorescence intensity of TOTO-1 in the presence of different types of oligonucleotides varied greatly (curve c, Fig. 1A-D). Specifcally, TOTO-1 in the presence of poly(dG) yielded much higher signals than those with the three other types of oligonucleotides. The quantum yields in the presence of poly(dG), poly(dA), poly(dC), and poly(dT) were determined to be 0.25, 0.11, 0.05 and 0.02, respectively. These data indicated that TOTO-1 showed strong fluorescence selectivity towards four types of bases: poly(dG) [ poly(dA) > poly(dC) > poly(dT). Fig. 1E shows the comparison of fluorescence intensities of TOTO-1 in the presence of different types of oligonucleotides. ## MD simulation of the interaction between TOTO-1 and poly(dG) MD simulation indicated that intramolecular p-p stacking existed in free TOTO-1, while intermolecular p-p stacking formed between TOTO-1 and poly(dG) (Fig. 2A). As a result, a folding structure similar to that of duplex DNA was formed. The absorption spectrum calculated by M06-2X/6-31G(D) was highly consistent with the experimental spectrum (Fig. 2B), which demonstrated the intermolecular p-p stacking in poly(dG)/TOTO-1. Other fluorescent dyes such as fluorescein sodium, thioflavin T and acriflavine did not have fluorescence selectivity towards the bases (Fig. S5 †). Although TO showed selectivity similar to that of TOTO-1, its PL intensity was much lower (30%) than that of TOTO-1 in the presence of poly(dG) and poly(dA) (Fig. S5C †). ## Fluorescence of TOTO-1 in the presence of G-rich elongated TS primer and PAR The absorbance and fluorescence spectra of TOTO-1 in the presence of G-rich elongated TS primer (Fig. 3A) and PAR (Fig. 3B) were also studied. In the presence of G-rich elongated TS primer, the absorbance and fluorescence spectra of TOTO-1 changed obviously compared with those in the presence of poly(dG) and poly(dA). Compared with the observation for poly(dG), the fluorescence intensity for the elongated TS primer decreased from 25 to 12.6, indicating that thymine and adenine certainly decreased the fluorescence emission of TOTO-1. The corresponding quantum yield decreased to 0.16. Under this circumstance, its fluorescence was still sensitive enough to distinguish the G-rich elongated TS primer, making it possible to detect telomerase activity. PAR, synthesized by PARP-1, is composed of adenosine diphosphate ribose (ADP-ribose) monomers (Fig. S6 †). In the presence of PAR, the absorbance spectrum of TOTO-1 showed signifcant changes with a new absorbance peak at 425 nm and increased absorbance values at 480 nm and 508 nm, indicating that the molecular orbitals of TOTO-1 were also different from those in the presence of poly(dA). Although its maximum fluorescence emission wavelength changed slightly, its fluorescence intensity increased by 2.5 times. The corresponding quantum yield also increased from 0.11 to 0.32, which defnitely contributed to improving PARP-1 activity detection. Furthermore, compared with the result for 15 mM of ADP, the fluorescence intensity of TOTO-1 increased from 0.7 to 7 in the presence of PAR produced by 0.8 U PARP-1. This also indicates that TOTO-1 is more sensitive to polymerized ADP than single ADP. Thus, it is possible to construct a sensitive and simple method to detect telomerase and PARP-1 activities based on the unique fluorescence selectivity and sensitivity of TOTO-1 towards the G-rich elongated primer and PAR. ## The strategy for telomerase activity evaluation The scheme for telomerase activity detection is illustrated in Fig. 4A. The telomerase primer was incubated with dNTPs and telomerase was extracted from tumor A549 cells. The TTAGGG repeat units were continuously synthesized on the 3 0 end of the TS primer to form a G-rich extended primer. In the presence of telomerase, the G-rich elongated primer led to a sharp increase in the fluorescence intensity of TOTO-1 (a, Fig. 4B). In the presence of TS primer without telomerase, the background signal was 2.9 (b, Fig. 4B). A primer was designed to form ds-DNA by incubation with complementary DNA (c-DNA), followed by digesting with Exo III; thus, the background signal decreased to 0.30 (c, Fig. 4B). The signal-to-noise ratio (S/N) increased by nearly 9 times (Fig. 4C). The control experiment results are shown in Fig. 4D. The fluorescence intensities of TOTO-1 in the presence of dNTPs, heated A549 cells and Exo III were negligible. Its fluorescence increased sharply when TOTO-1 was incubated with the primer, dNTPs and A549 cells, indicating that telomerase triggered the fluorescence of the system (f, Fig. 4D). ## The strategy for PARP-1 activity evaluation The scheme for PARP-1 activity detection is illustrated in Fig. 5A. PARP-1 was activated by specifc dsDNA. Then, PARP-1 transferred the frst ADP-ribose unit from nicotinamide adenine dinucleotide (NAD + ) to an acceptor protein and sequentially added multiple ADP-ribose units to the preceding ones to form hyper-branched PAR. However, the background fluorescence produced by the activated dsDNA was very high because TOTO-1 showed strong fluorescence in the presence of dsDNA (curve b, Fig. 5B). It was also necessary to use Exo III to reduce the background signal. From curve a in Fig. 5B, we can see that the background signal is reduced by 38 times, which makes it possible to detect the PARP-1 activity sensitively. The control experiments for PARP-1 detection are shown in Fig. 5C. Only activated PARP-1 could trigger the fluorescence of the system, while other substances such as NAD + , heated PARP-1, and Exo III exhibited negligible effects on the fluorescence intensity. ## Performances for telomerase and PARP-1 analysis Confocal imaging gave further evidence for the feasibility of the proposed method. As is shown in Fig. 6, TOTO-1 exhibits very weak fluorescence in the presence of the telomerase and PARP-1 detection system without telomerase (A) or PARP-1 (A 0 ). The presence of telomerase (B) or PARP-1 (B 0 ) triggered the fluorescence of TOTO-1 sharply due to the formation of G-rich elongated primer and PAR, respectively. The fluorescence spectra for telomerase and PARP-1 detections are shown in Fig. 6C and C 0 , respectively. The optimization of the detection conditions is shown in Fig. S7 and S8. † PL intensity showed a linear correlation with increasing concentration of telomerase from 13 to 4000 A549 cells per mL. The regression equation is Y ¼ 0.75 + 4.79 lg C, (R 2 ¼ 0.997). A linear relationship between the PL intensity and PARP-1 ranging from 0.02 U to 1.5 U (0.04 nM to 3 nM) was obtained Y ¼ 2.66 + 5.23C (R 2 ¼ 0.992), where Y is the PL intensity and C is the content of A549 cells or PARP-1. The detection limit for telomerase and PARP-1 was 13 cells per mL and 0.02 U, respectively. Compared with the previously reported methods for telomerase detection listed in Table 1, it can be concluded that this method is more sensitive than the UV methods. 6,43 It is also comparable to other sensitive methods such as ECL, CL, 9 SERS, 48 and fluorescence methods that use other probes. It can also be seen that the detection sensitivity for PARP-1 is comparable to those of other reported methods (Table S2 †). The practicality and selectivity for telomerase activity detection were verifed by using other kinds of cell lines, proteins and enzymes (Fig. 7A). The PL intensity increased signifcantly for A549, HeLa, and MCF-7 cells and was negligible for heated A549 cells, PARP-1, HRP, BSA, and HSA, indicating that high PL resulted from telomerase in these cancer cell lines. As shown in Fig. 7B, we also studied the selectivity of the biosensor for PARP-1. The experimental results indicate no obvious change in PL when PARP-1 was replaced by BSA, heated PARP-1, and telomerase, suggesting satisfactory selectivity for PARP-1 detection. ## Application of the strategy to detect telomerase in urine and PARP-1 in cells Telomerase was detected by the proposed method in 11 early morning urine samples provided by the Nanjing General Hospital of Chinese People's Liberation Army. A threshold value of 0.59 was calculated from the mean fluorescence intensity of 100 blank samples plus 3 times the relative standard deviation. The detection results are shown in Table 2. Sample 1 is the blank solution and samples 2-8 were obtained from normal, vesical calculus, kidney stone and inflammation patients. All PL intensities were lower than the threshold value. Samples 9-12 were obtained from bladder cancer patients. All their results were far higher than the threshold value. Therefore, this painless method has great potential to be used for bladder cancer diagnosis. The PARP-1 levels in normal cells of IOSE80 and breast cancer cells of MCF-7 and SK-BR-3 were evaluated. The detection results for samples 13-15 showed that PARP-1 was present at a low level in normal human cells, while it was signifcantly up-regulated in the malignant SK-BR-3 cells and MCF-7 cells. PARP-1 was higher in the nucleus than that in the cytoplasm of both SK-BR-3 cells and MCF-7 cells, which was consistent with previous results. 54,55 Recovery experiments were conducted in human serum, normal cells IOSE80 and human ovarian cancer cells A2780 to prove the accuracy of the sensor in complex biological matrices. Recoveries were from 94% to 109% and the relative standard derivations were in the range of 2.48-8.21%, indicating that the method has good accuracy and high precision (Table S4 †). ## Detection of inhibition efficiency on telomerase and PARP-1 To test the inhibitor screening function of the proposed sensor, BIBR1532 and curcumin were chosen as the model inhibitors for telomerase and AG014699 was selected as the model inhibitor for PARP-1. The PL intensities decreased continually to a stable value when the amounts of BIBR 1532 and curcumin increased (Fig. S9A and C †). Decreased percentages of the PL intensity in the presence of various concentrations of BIBR 1532 and curcumin are shown in Fig. S9B and D. † A549 without an inhibitor was used as the control (750 cells per mL). The IC 50 (concentration of inhibitor producing 50% inhibition) values for BIBR1532 and curcumin against telomerase were 251 nM and 8.64 mM, respectively. These results were in accordance with previously reported values. 56,57 The fluorescence spectra in the presence of PARP-1 inhibited by different concentrations of AG014699 are shown in Fig. S9E. † The decreased percentages of the PL intensity are shown in Fig. S9F. † The IC 50 value of AG014699 was determined to be 8.2 nM; it was also in good agreement with the reported IC 50 value in literature. 58 ## Conclusions In summary, the unique fluorescence selectivity and sensitivity of TOTO-1 towards the G bases in single-stranded DNA and PAR were found for the frst time and studied elaborately. MD simulation proved that intermolecular p-p stacking formed between TOTO-1 and poly(dG) and TOTO-1/poly(dG) folded in a way analogous to that of duplex DNA, which explained why TOTO-1 has high selectivity towards poly(dG). TOTO-1 is also the frst probe that has been reported to have high selectivity and sensitivity towards PAR. A label-free, simple and sensitive fluorescence biosensor was constructed for telomerase and PARP-1 sensing based on TOTO-1 and Exo III-assisted background noise reduction. This strategy shows high sensitivity and selectivity towards telomerase and PARP-1. Integrating multiple functions into one simple probe, i.e., detecting not only telomerase but also PARP-1 can signifcantly raise the specifcity of screening cancer and decrease false positive proportion. This makes TOTO-1 a promising candidate probe for clinical diagnosis. However, the probe cannot be used to image telomerase or PARP-1 in vivo because Exo III is necessary to decrease the background signals and it is difficult to be transferred into cells.

chemsum

{"title": "Telomerase and poly(ADP-ribose) polymerase-1 activity sensing based on the high fluorescence selectivity and sensitivity of TOTO-1 towards G bases in single-stranded DNA and poly(ADP-ribose)", "journal": "Royal Society of Chemistry (RSC)"}

fabrication_of_long-term_underwater_superoleophobic_al_surfaces_and_application_on_underwater_lossle

## Abstract: Underwater superoleophobic surfaces have different applications in fields from oil/water separation to underwater lossless manipulation. This kind of surfaces can be easily transformed from superhydrophilic surfaces in air, which means the stability of superhydrophilicity in air determines the stability of underwater superoleophobicity. However, superhydrophilic surfaces fabricated by some existing methods easily become hydrophobic or superhydrophobic in air with time. Here, a facile method combined with electrochemical etching and boiling water immersion is developed to fabricate longterm underwater superoleophobic surfaces. The surface morphologies and chemical compositions are investigated. The results show that the electrochemically etched and boiling-water immersed Al surfaces have excellent long-term superhydrophilicity in air for over 1 year and boehmite plays an important role in maintaining long-term stability of wettability. Based on the fabricated underwater superoleophobic surfaces, a special method and device were developed to realize the underwater lossless manipulation of immiscible organic liquid droplets with a large volume. The capture and release of liquid droplets were realized by controlling the resultant force of the applied driving pressure, gravity and buoyancy. The research has potential application in research-fields such as the transfer of valuable reagents, accurate control of miniature chemical reactions, droplet-based reactors, and eliminates contamination of manipulator components.Lotus leaf has two types of extreme wettability, one is the superhydrophobicity on the upper side, and another is the underwater superoelophobicity on the lower side 1 . Since 2011, several research groups explored the formation mechanism of the underwater superoleophobicity inspired by lotus leaf 2-5 . They found the superhydrophilic surfaces with micro rough structures in air can be easily transformed into underwater superoleophobic surfaces. When the superhydrophilic surfaces immersed in water, the water wetted the whole surfaces first, then the polar water trapped in the micro structures greatly decreased the contact area between the non-polar oil phase and the solid surfaces, resulting in an underwater superoleophobic property with a high contact angle and low adhesion, which is also an underwater Cassie-Baxter state. As mentioned above, the fabrication of underwater superoleophobic surfaces can be transferred into the fabrication of superhydrophilic surfaces in air. The superhydrophilicity on metal substrate can be easily obtained by constructing micro structures. However, many initial superhydrophilic surfaces with the fabricated micro structures easily become hydrophobic or superhydrophobic in air with time 6-11 . This transition is independent of the fabrication methods. The micro structures obtained by chemical etching 12 , electrochemical etching 13 , chemical oxidation 14,15 , laser etching 6,9 , and thermal oxidation 8 all show this transition behavior, resulting in a poor long-term stability of superhydrophilicity. For example, the superhydrophilicity on Al substrates obtained by HCl etching or electrochemical etching can keep no more than 2 days; the superhydrophilicity on Cu substrates obtained by chemical oxidation only keep for 8 days; and the superhydrophilicity on Al substrates obtained by laser etching can keep no more than 8 days. Therefore, a facile method to fabricate superhydrophilic surfaces with long-term stability need to be studied. The underwater superoleophobic surfaces have a great application prospect in the field of underwater lossless manipulation of non-polar organic liquids. To date, although there are many papers about lossless manipulation of liquids in air by using superhydrophobic surfaces , the paper about underwater lossless manipulation of non-polar organic liquids is less. Yong et al. realized the in-situ transfer oil droplets in water based on underwater superoleophobic surfaces by adding sugar in the water and switching the density of the water solution 28 . This method need to change the characteristic of the environment water, resulting in a complex manipulation process. Ding et al. used electrical potential to control the adhesive force of the underwater superoleophobic surfaces and further realize the in-situ capture and release of the oil droplets underwater 29 . However, the main capture force in the Ding's method is the adhesive force between the liquid droplets and the solid surfaces. This kind of adhesive force is very small and no more than 100 μ N, resulting in a small operable limiting volume. Since the underwater lossless manipulation of non-polar organic liquids has potential application in research-fields such as the transfer of valuable reagents, accurate control of miniature chemical reactions, droplet-based reactors, and eliminates contamination of manipulator components, a simple route for the in-situ lossless controllable manipulation of non-polar organic liquid droplet with a big volume need to be studied. Here, we first report a new method to fabricate superhydrophilic surfaces on Al substrates with long-term stability for greater than 1 year. We show that the superhydrophilic surfaces are easily transformed into underwater superoleophobic surfaces when immersed in water. Then, using the non-sticky to non-polar organic liquids of the underwater superoleophobic surfaces, a special transfer-pipette device composed of the fabricated underwater superoleophobic surfaces and pressure-generation device is constructed to realize the underwater lossless manipulation of non-polar organic liquids with a large range of volume. The operable limiting volume of the droplet is shown to be determined by the contact angle, interface tension, and density of organic liquid, which even reaches to 1406 μ L for peanut oil. ## Results Superhydrophilic Al surfaces with long-term stability. Figure 1 shows the micro morphology and chemical composition of the electrochemically etched and boiling-water immersed (EEBWI) Al surfaces. Grain boundaries and dislocations on Al are easily etched by an applied electric field (electrochemical etching) because of their relatively higher energy, forming micrometer-scale rectangular-shaped plateaus and step-like structures with sizes in the range of 1 μ m to 5 μ m 13 . After immersion in boiling water such surfaces were transformed into nanometer-scale needle-like structures with length of 200 nm and width of 30 nm which covered the whole surfaces, as shown in Fig. 1(a,b). Since the electrochemical etching did not change the surface composition, the diffraction peaks from boehmite (γ -AlOOH, Al 2 O 3 •H 2 O) were detected beside the diffraction peaks from Al on the XRD patterns after electrochemical etching and immersion in boiling water, thus we learned that the main compositions of the nanometer-scale needle-like structures were boehmite, as shown in Fig. 1(c). The EEBWI Al surfaces showed excellent superhydrophilicity with a very fast spreading velocity when water touched the surfaces, as shown in Fig. 1(d). To analyze the influence of boehmite on the stability of superhydrophilicity, we measured the change of contact angle of the Al surfaces before and after boiling water immersion with the exposure time in air. As shown in Fig. 1(e), the electrochemically etched Al surface without boehmite only kept superhydrophilicity for 1 day, indicating boehmite plays an important role in maintaining superhydrophilicity over a long time period. The EEBWI Al surfaces have considerably better stability than other usual superhydrophilic surfaces, e.g. the usual Cu(OH) 2 microstructures only kept superhydrophilicity for 8 days 14,15,30 . The EEBWI Al surfaces even have excellent long-time stability as superhydrophilic surfaces in the air for over 1 year, as shown in Fig. 1(f). When the EEBWI Al surfaces were immersed in water, the polar water became entrapped in between the micro/ nanometer-scale structures because of superhydrophilicity, forming a repellent conformal barrier to non-polar organic liquids, resulting in underwater superoleophobicity with small adhesion force (see Supplementary Video S1 and S2). The underwater contact angles of typical organic liquids, e.g. hexane, hexadecane, peanut oil, and dichloromethane, on the EEBWI Al surfaces were larger than 160°. This kind of underwater superoleophobic surface was used in the following transfer-pipette to realize the underwater lossless manipulation of non-polar organic liquids with controllable volume. Underwater lossless manipulation of non-polar organic liquids. Figure 2(a) shows the illustration of a prototype transfer-pipette, which was mainly composited of underwater superoleophobic hole (EEBWI Al surfaces) and pressure-generation device (e.g. a syringe). The micrometer-scale rectangular-shaped plateaus and step-like structures and nanometer-scale needle-like structures were covered on the end face and inner wall of the hole. The transfer-pipette utilized the air pressure to hold up and capture the organic liquid droplets. No organic liquids residues were left on the transfer-pipette because of the underwater superoleophobic properties of the surface. The captured organic liquids can be released on any surfaces including sticky and non-sticky surfaces. The illustration of the complete underwater lossless manipulation processes of non-polar organic liquid droplets are shown in Fig. 2(b). The transfer-pipette first approached the target organic liquid droplet and then captured the organic liquid droplet under the certain pressure. The captured organic liquid droplet was moved to the destination. Finally, the added pressure was removed and the organic liquid droplet was released. The released organic liquid droplets on non-sticky surfaces can be manipulated again for several times as described in the aforementioned processes. The detailed and practical working processes of underwater lossless manipulation of dichloromethane with volume of 2 μ L and 5 μ L are shown in Fig. 2(c). ## Discussion The possible mechanism of the long-term superhydrophilicity was analyzed. Besides N 2 , O 2 , and CO 2 , air also contains a minute amount of organic compounds 9 . Long's research shows that organic compounds from the surrounding atmosphere will adsorb on the metal surface through the interactions with hydroxyl groups, reducing the content of hydroxyl groups on the metal surface 9 . It is well known that hydroxyl groups are hydrophilic and its content decide the wettability of surface. For electrochemically etched Al surfaces in the air, the decrease of surface hydroxyl groups and adsorption of organic compounds reduce the wettability and increase the contact angle. However, the situation for the EEBWI Al surfaces coated with boehmite is different. Boehmite contains water of crystallization, the content of which is small but far bigger than the hydroxyl groups 31 . The water of crystallization is polar and keeps good affinity for water molecules. In addition, the water of crystallization in the boehmite is very stable in air. All these guarantee the long-term supehrydrophilicity of the EEBWI Al surfaces. To realize the lossless manipulation, the applied pressure should be smaller than the threshold pressure of organic liquid placed underwater. The threshold pressure is defined as the pressure under which organic liquids pass through the voids in the microstructures on the EEBWI Al surfaces which are filled with water. The threshold pressure P t is determined by the corresponding Laplace pressure and can be given by 32 where γ ow is the interfacial tensions for organic liquid/water interfaces, θ is the underwater contact angle of organic liquid on the EEBEI Al surfaces, and D s is the spacing of the voids in the microstructures. Since the spacing of the voids in the microstructures, D s , is very small (about 5 μ m), the threshold pressure P t is correspondingly large. Taking dichloromethane as an example of an organic liquid, the values of γ ow and θ are 28.2 mN/m and 164°, respectively 33 . Then, P t = 21.7 kPa. The macro force analysis for a captured droplet is shown in Fig. 3(a,b). For a balance state, the applied force F p is determined by where G and F f are the gravity and buoyancy of organic liquid, respectively. The densities of heavy and light organic liquids are larger or smaller than the density of water, respectively. The F p , G, and F f are also respectively given by where P denotes the applied pressure, D the diameter of the hole in the transfer-pipette, v o the volume of organic liquid, ρ o the density of organic liquid, ρ w the density of water, and g the acceleration due to gravity. According to equations (1) to ( 5), the applied pressure P can be calculated as follows, For 5 μ L dichloromethane, the values of ρ o , ρ w , V o , D, and g are 1325 kg/m 3 , 1000 kg/m 3 , 5 μ L, 0.5 mm, and 10 N/kg, respectively. Then, P = 83 Pa. The applied pressure P is far less than the threshold pressure P t , guarantying the lossless property in the manipulation processes. The transfer-pipette not only can freely and in-situ control the adhesive force to realize the capture and release of liquid droplets, but also can capture the liquid droplets with a large range of volume under the condition of force equilibrium. However, the minimum size of the liquid droplets should be larger than the size of the hole and the value is (π D 3 /6). The maximum size of the liquid droplets is also related with the size of the hole. When the applied force, F p , captures the top layer of the droplet, if the liquid droplet is very heavy, the droplet will break because of elastic deformation, resulting in a failure manipulation. Theoretically, the volume of a quasi-stable droplet that will eventually break from the hole is mainly determined by competition between the gravity and buoyancy of the pendulous droplet and the vertical component of capillary force, F s , around the hole edge, as shown in Fig. 3(c). In the quasi-stable state, the relationship between the aforementioned 3 forces is as follows, According to equations ( 5), ( 6), ( 9) and ( 10), the operable maximum volume V max of the liquid droplet can be predicted as . For dichloromethane and the hole with 0.5 mm, according to equation (11), the theoretical operable maximum volume V max is about 13.08 μ L. Experimentally, the transfer-pipette with a hole of 0.5 mm can easily manipulated the dichloromethane with volume of 12 μ L, but could not capture the dichloromethane with volume of 13 μ L, as shown in Fig. 3(e) and the Supporting Information (see Supplementary Video S3 and S4), indicating the theoretical results agreed well with the experimental ones. We also calculated the operable minimum and maximum volume of the droplet for 4 types of liquids, as shown in Table 1. The relationship between the operable volume of the droplet and the size of the hole is shown in Fig. 3(d). Refering to this Figure, we can choose suitable hole size to manipulate the liquid droplet with different types and volume (just like choose a needle for a syringe). Regardless of the hole, for this method, the operable limiting volume of the droplet can be calculated as follows, Thus, the operable limiting volume, V limit , is , and the corresponding size of the hole, D limit , is . The calculated operable limiting volumes for hexane, hexadecane, peanut oil, and dichloromethane are 187, 776, 395, and 1406 μ L, respectively. The corresponding needed limiting sizes of the hole are 7.1, 11.4, 9.1, and 13.9 mm, respectively. Obviously, the operable range of the volume is very large for this method. The microdroplet-based micromixing was carried out by the developed transfer-pipette, as shown in Fig. 4. The transfer-pipette selectively captured and manipulated one droplet and transferred it to another droplet, completing the mixing of the two microdroplets and demonstrating the potential strategy for quantitative reaction and combinations between microdroplets and components incorporated within microdroplets. This rapid manipulation of the microliter droplets provides us a potential method to save on reactants by ensuring complete liquid transfer and enable microdroplet-based reactions with no loss in sample volumes. In summary, we developed a new method to fabricate long-term superhydrophilic and underwater superoleophobic surfaces on Al substrates by combining electrochemical etching followed by immersion in boiling water. The new surfaces were covered with the micrometer-scale rectangular-shaped plateaus and step-like structures and nanometer-scale needle-like boehmite structures. The boehmite plays a key role in maintaining long-term superhydrophilicity of the surfaces for over 1 year. Based on the modified Al surfaces, a new type of transfer-pipette mainly composed of an underwater superoleophobic hole and pressure-generation device was developed to realize the underwater lossless manipulation of non-polar organic liquids. The operable minimum volume of the droplet is only determined by the hole size, while the operable maximum volume of the droplet is determined by the hole size, contact angle, interface tension, and density of organic liquid. The relationship between the operable volume of the droplet and the size of the hole can be determined and drawn into a Reference Figure . According to the Reference Figure, it is possible to quickly choose the suitable hole size (just like choosing a needle for a syringe) to manipulate the liquid droplet with different types and volume. The operable limiting volume of the droplet for this method is determined by the contact angle, interface tension, and density of organic liquid, which even reaches to 1406 μ L for peanut oil, showing a large operable range in volume. This method can improve the sample volume transfer accuracy, reduce the sample liquid retention, and realize the quantitative microdroplets reaction. This in turn offers the potential for saving expensive reagents, and enabling better quantitative accuracy for transfer of immiscible organic liquids underwater. ## Methods Fabrication of underwater superoleophobic surfaces. Two holes with diameter of 0.5 mm and 2 mm and depth of 5 mm and 15 mm were drilled on the two end faces of an aluminum (Al) rod (purity > 99%, 20 mm length, 10 mm diameter), respectively. The end face with 0.5 mm hole was then polished using #1200 and #1500 abrasive paper and electrochemically etched at 500 mA/cm 2 current density and 6 min processing time in 0.1 mol/L aqueous NaCl solution. Then, the electrochemically etched Al rod was immersed in boiling water for 1 h. The electrochemically etched and boiling water immersed Al surfaces show superhydrophilic in air and superoleophobic under water. To characterize the superhydrophilic and underwater superoleophobic surface better, the plane Al plates with size of 30 × 40 × 2 mm were treated by the aforementioned electrochemical etching and boiling water immersion processes. ## Design of prototype transfer-pipette. A plastic syringe with volume of 1 mL was connected with the Al rod by installing a pressure soft tube into the 2 mm hole to construct a prototype transfer-pipette. The syringe was used as a pressure-generation device to produce negative pressure in the hole of Al rod. In the processes of manipulation of organic liquid droplet, the centerlines of the hole and liquid droplet should be in a straight line as much as possible. Characterization. The microstructures and chemical composition of the sample surfaces were observed by a scanning electron microscope (SEM, JSM.6360LV, Japan) and an X-ray diffractometer (Empyrean, Holland). Water and oil droplet contact angle and sliding angle measurements were performed using an in-house goniometer employing ∼ 5 μ L water and oil droplets. The samples were putted in the culture dish in air for preservation. The sliding angle was defined as the angle at which the liquid drop began to slide on the gradually inclined surface. Hexane, hexadecane, peanut oil and dichloromethane were used in the present study as the non-polar organic liquids.

chemsum

{"title": "Fabrication of Long-Term Underwater Superoleophobic Al Surfaces and Application on Underwater Lossless Manipulation of Non-Polar Organic Liquids", "journal": "Scientific Reports - Nature"}

synthesis_and_evaluation_of_the_biostability_and_cell_compatibility_of_novel_conjugates_of_nucleobas

## Abstract: This article reports the synthesis of a new class of conjugates containing a nucleobase, a peptidic epitope, and a saccharide and the evalution of their gelation, biostability, and cell compatibility. We demonstrate a facile synthetic process, based on solid-phase peptide synthesis of nucleopeptides, to connect a saccharide with the nucleopeptides for producing the target conjugates. All the conjugates themselves (1-8) display excellent solubility in water without forming hydrogels. However, a mixture of 5 and 8 selfassembles to form nanofibers and results in a supramolecular hydrogel. The proteolytic stabilities of the conjugates depend on the functional peptidic epitopes. We found that TTPV is proteolytic resistant and LGFNI is susceptible to proteolysis. In addition, all the conjugates are compatible to the mammalian cells tested. ## Introduction This article describes the synthesis and evaluation of a new class of molecular conjugates that consist of a nucleobase, a peptidic epitope, and a saccharide. Nucleobases, amino acids, and saccharides are part of the unified building blocks of life because they constitute three key types of biomacromolecules-proteins, nucleic acids, and carbohydrates. Inspired by this molecular foundation resulted from evolution, we are developing biomaterials that consist of the covalent conjugates of these three classes of the basic building blocks of life. For example, we found that certain conjugates of nucleobase, amino acid, and saccharide (NAS) or some conjugates of nucleobase, saccharide, and amino acid (NSA) self-assemble in water to form supramolecular hydrogels , but, so far, none of the conjugates of saccharide, amino acid, and nucleobase (SAN) is able to act as hydrogelators . Besides the properties of selfassembly, these conjugates are cell compatible . Moreover, the NAS conjugates promote the proliferation of mES cells and deliver the oligonucleotide into living cells . Particularly, the incorporation of the functional peptidic epitope, RGD , results in a NAS conjugate that self-assembles in water, Scheme 1: Chemical structures of the conjugates (nucleobase-amino acids-saccharide (NAS)), and nucleopeptides (NA). exhibits improved proteolytic stability , and promotes the development of mouse zygotes . These results suggest that it is also worthwhile to incorporate other peptidic epitopes into the NAS conjugates and to evaluate their physiochemical and biological properties. Based on the above rationale, we chose two short peptidic epitopes, TTPV and LGFNI, from two well-characterized proteins as the building blocks for nucleopeptides or NAS conjugates. TTPV is from a calcium channel protein (stargazin) and LGFNI is from a synapse associated protein 102 (SAP102) . We connected these two functional peptide sequences with a nucleobase, and saccharide (or not). After investigating the gelation, biostability, and cell-compatibility properties of these conjugates (1-8), we found that all the conjugates exhibit excellent solubility in water without resulting in hydrogelation or precipitation. We observed that only the mixture of a proper pair of TTPV-containing (e.g., 5) and LGFNI-containing (8) conjugates self-assembles to form nanofibers and results in a supramolecular hydrogel. Moreover, the conjugates containing TTPV or ETPV show excellent proteolytic stability, while the conjugates containing LGFNI undergo complete proteolysis catalyzed by proteinase K after 24 hours, with or without the presence of nucleobase or saccharide in the conjugates. Apparently, the stabilities of the conjugates coincide with their corresponding peptide sequences that TTPV is proteolytic stable and LGFNI proteolytic susceptible. In addition, all the compounds investigated in this work exhibit excellent compatibility to mammalian cells such as HeLa and PC12 cells. This work provides useful insights on the incorporation of peptidic epitopes into molecular conjugates that consist of a nucleobase, amino acids, and a saccharide for potentially developing new supramolecular biomaterials. ## Molecular design Scheme 1 shows the chemical structures of the conjugates explored in this work. In the NAS conjugates 1-4, we chose Scheme 2: The representative synthesis route of conjugates NAS (1, including solid-phase peptide synthesis and liquid-phase synthesis). i) Fmoc-Val-OH, DIPEA; ii) 20% piperidine; iii) Fmoc-Pro-OH, HBTU, DIPEA; iv) Fmoc-Thr(t-Bu)-OH, HBTU, DIPEA; v) thymine-1-acetic acid, HBTU, DIPEA; vi) TFE/DCM 2:8; vii) D-glucosamine hydrochloride, HBTU, DIPEA; viii) TFA/H 2 O 95:5. thymine as the nucleobase, TTPV and LGFNI, which are two well-characterized peptide binding motifs , as the peptidic epitopes, and glucosamine as the saccharide. As a control, we changed the sequence of peptides to ETPV or EPTV. To investigate the function of the saccharide, we designed the nucleopeptides 5 and 6. In addition, we substituted thymine with adenine to generate nucleopeptides 7 and 8 that contain adenine, the nucleobase is complementary of thymine. ## Synthesis The NA conjugates 5-8 were obtained according to a facile method of solid-phase peptide synthesis (SPPS) . The conjugates NAS were produced by a combination of SPPS and liquid phase synthesis. Scheme 2 shows a representative synthesis route of a NAS conjugate (1). We loaded the first amino acid, Fmoc-Val-OH, on 2-chlorotrityl chloride resin, then removed the Fmoc group with 20% piperidine in dimethylformamide (DMF) to expose the amino group. The second amino acid, Fmoc-Pro-OH, was reacted with the free amino group using the coupling reagent N,N,N′,N′-tetramethyl-O-(1H-benzotriazol-1-yl)uronium hexafluorophosphate/N,N-diisopropylethylamine (HBTU/DIPEA). The elongation of the peptide chain was done by repeating the removal of the Fmoc group and sequential addition of Fmoc-Thr(t-Bu)-OH, Fmoc-Thr(t-Bu)-OH and thymine-1-acetic acid. For the final step of the SPPS, we used 2,2,2-trifluoroethanol/dichloromethane (TFE/DCM 2:8) to cleave the fully protected NA from the resin. For conjugates 5-8, we cleaved the chain from the resin with 95% tri-fluoroacetic acid (TFA) without N-protecting groups. Later, NAS was obtained by reacting D-glucosamine hydrochloride with the fully protected NA. After cleaving the protecting group on amino acids with 95% TFA, we used reversed-phase highperformance liquid chromatography (HPLC) to purify the target conjugates. ## Gelation properties Supramolecular hydrogels formed by self-assembly of small molecules in water, as demonstrated previously by us and other researchers, have numerous potential applications, such as encapsulation and delivery of DNA and microRNA , delivery of therapeutic agents , scaffolds for cell culture and spinal arthrodesis , sensor for detection of hyperuricemia disease and diabetes , and matrix for the electrophoresis of acidic native proteins . After obtaining the pure conjugates 1-8, we tested their gelation properties. Conjugates 1-8 show excellent solubility in water. When being mixed in PBS, 5 and 8 ( = = 8.3 mM, pH 6.2) selfassemble to form a hydrogel overnight. The hydrogel of 5 + 8 consists of long and flexible nanofibers (with an average width of 9 ± 2 nm), which entangle to form stable networks (Figure 1B). This result is similar to the hydrogelation when mixing two nucleopeptides of the heterodimer . In contrast, the mixture of 1 and 8 remains a solution. As shown in Figure 1A, the TEM of the solution of 1 + 8 reveals helical nanofibers with an average width of 10 ± 2 nm in the solution. This result indicates that the introduction of the glycan at the C-terminus of 5 increases the solubility of the nanofibers. The mixture of 7 and 8 also fails to result in a hydrogel. Moreover, the TEM of the solution of 7 + 8 hardly exhibits any ordered nanostructure (Figure 1C). This result implies that the base pair interactions between thymine and adenine likely play a critical role for the hydrogelation of the mixture of 5 + 8. In addition, we did gelation tests for 4 + 5 and 4 + 7. Both mixtures are unable to self-assemble to form hydrogels at the same conditions used for 5 + 8. These results illustrate that the subtle change in the molecular structures of the conjugates is able to cause drastically different behaviour of self-assembly . We also investigated the rheological properties of the three mixtures, 1 + 8, 5 + 8, and 7 + 8 in PBS buffer. As shown in Figure 2B, storage modulus (G') is higher than loss modulus (G'') for 5 + 8, confirming that 5 + 8 is a viscoelastic material. Storage moduli (G') overlap with loss moduli (G'') for 1 + 8 and 7 + 8, agreeing with that 1 + 8 and 7 + 8 behave as liquid-like materials. In addition, the maximum storage for 5 + 8 is 3.7 Pa (Figure 2A), indicating that 5 + 8 is a weak hydrogel. When the strain is between 0.8-100%, the storage modulus of 1 + 8 is slightly higher than that of loss modulus (Figure 2A), which is likely due to the existence of nanofibers in the solution of 1 + 8 (Figure 1A). ## Biostability The existence of proteolytic enzymes in organism limits the applications of peptide-based biomaterials in vivo. To evaluate the biostability of the conjugates 1-8, we incubated them with proteinase K (a powerful protease) in HEPES buffer at 37 °C for 24 h. As shown in Figure 3A, conjugates 1, 2, 5, 6, and 7, containing peptidic epitopes TTPV or ETPV, almost 100% remained after incubating with proteinase K for 24 h. When we changed the peptide sequence to EPTV, only 15% of 3 left. Conjugates 4 and 8, containing peptidic epitopes LGFNI, are undetectable after incubating with the proteinase K. The biostability of these conjugates are relevant to their epitopes, since we found that the peptidic epitopes have the same biosta- bility . These results indicate that stable natural peptidic epitopes in the conjugates should be able to improve the biostability of the conjugates. In addition, as shown in Figure 3B, we found that the hydrogel mixture of 5 + 8 promotes the biostability of 8 (about 50% remained at 24 h). The mixture of 7 + 8, being incubated with proteinase K at the same condition as the test of 5 + 8, failed to increase the biostability of 8. The concentration of 8 is slightly increased in the treatment of the mixture of 1 + 8 (about 3% left) with proteinase K, comparing to the case of 8 incubated with proteinase K at 24 h. This result is consistent with TEM investigations showing that there are weak interactions between 1 and 8. ## Cell compatibility Cell compatibility is one of the major considerations for biomaterials . To assess the cell compatibility of the synthesized conjugates, we incubated HeLa and PC12 cells with 1-8 at the concentration range of 20-500 μM. As shown in Figure 4, with (1-4, Figure 4A-D) or without (nucleopeptide, 5-8, Figure 4E-H) glucosamine, conjugates 1-8 are innocuous to HeLa cells for treatment of 3 days. Because of a longer doubling time of PC12 than HeLa cells , we incubated PC12 cells for 7 days. Conjugate 1-8 showed little toxicity to PC12 cells (Figure 5). The mixture of 5 + 8, which has the highest self-assembly ability, also hardly inhibits the prolifera- tion of HeLa and PC12 cells (Figure 6). These results reveal that these conjugates, though having different ability of selfassembly, are cell compatible . The cell compatibility of these molecules and the two component hydrogel 5 + 8 promises them to serve as biomaterials. ## Conclusion In conclusion, we designed eight conjugates by modifying two endogenous binding peptide motifs with nucleobase and glucosamine (or not) and investigated their gelation properties, biostability, and cytotoxicity. Particularly, the mixture of 5 and 8 affords a stable hydrogel, which increases the biostability of 8. Meanwhile, 5, 8, and their mixture show excellent cell compatibility, which is a basic requirement for multi-application in vivo (e.g., wound healing ). This work provides a new approach to develop biocompatible soft materials. ## Experimental Materials Starting materials and reagents were purchased from GL Biochem (Shanghai) Ltd. and Fisher Scientific without further purification unless otherwise noted. Proteinase K was purchased from Sigma (>800 unit/mL). The HeLa cell line (CCL-2) and the PC12 (CRL-1721.1) cell line were purchased from the American Type Culture Collection. All of the media were purchased from Invitrogen. ## Instruments Conjugates were purified with a Water Delta600 HPLC system, equipped with an XTerra C18 RP column and an in-line diode array UV detector. 1 H NMR spectra were obtained on a Varian Unity Inova 400 spectrometer. LC-MS spectra were performed on Waters Acquity Ultra Performance LC with Waters MICRO-MASS detector. TEM images were taken on a Morgagni 268 transmission electron microscope. Rheological data were measured on a TA ARES G2 rheometer with 25 mm cone plate. MTT assay for cell toxicity test were measured on a DTX880 Multimode Detector. ## Synthesis The synthesis procedures for conjugates 1-8 are demonstrated in the main text synthesis part.

chemsum

{"title": "Synthesis and evaluation of the biostability and cell compatibility of novel conjugates of nucleobase, peptidic epitope, and saccharide", "journal": "Beilstein"}

chemical_profiling_analysis_of_maca_using_uhplc-esi-orbitrap_ms_coupled_with_uhplc-esi-qqq_ms_and_th

## Abstract: Lepidium meyenii (Maca), originated from Peru, has been cultivated widely in China as a popular health care food. However, the chemical and effective studies of Maca were less in-depth, which restricted its application seriously. To ensure the quality of Maca, a feasible and accurate strategy was established. One hundred and sixty compounds including 30 reference standards were identified in 6 fractions of methanol extract of Maca by UHPLC-ESI-Orbitrap MS. Among them, 15 representative active compounds were simultaneously determined in 17 samples by UHPLC-ESI-QqQ MS. The results suggested that Maca from Yunnan province was the potential substitute for the one from Peru. Meanwhile, the neuroprotective effects of Maca were investigated. Three fractions and two pure compounds showed strong activities in the 1-methyl-4-phenyl-1, 2, 3, 6-tetrahydropyridine (MPTP)-induced zebrafish model. Among them, 80% methanol elution fraction (Fr 5 ) showed significant neuroprotective activity, followed by 100% part (Fr 6 ). The inhibition of acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) was a possible mechanism of its neuroprotective effect.Maca (Lepdium meyenii), known as "Peruvian ginseng" 1 , has been used as traditional health care food for over 2000 years in South America. According to hypocotyl colors, it was classified in black, purple and yellow varieties 2 . In 1992, Maca was recommended as the safety edible food by Food and Agriculture Organization (FAO). After twenty years of development, it has been considered as one of the star products in the global health care market. Because of its various potential effects, in the early 21 st century, Maca was introduced into China successfully and vigorously promoted the cultivation in Yunnan, Xinjiang and Tibet regions at high altitude similar to Peru.Traditionally, Maca was always used in strengthening body, improving fertility and sexual function. Modern pharmacological studies displayed its effects on depression, rheumatism, premenstrual discomfort and menopausal symptoms 3 . Significantly, along with the increasing risk of neurodegenerative diseases, the neuroprotective effect of Maca has been attracting great concern. The discovery and screening of neuroprotective substances from Maca should be given high priority [4][5][6][7][8][9][10][11][12] .Since 2016, the price of Maca decreased dramatically in China. There were a couple of reasons for this. Firstly, a plenty of inferior products of Maca disturbed the market. Secondly, the basic study of Maca could not meet the requirement of market. Both of them came down to the absence of in-depth research of Maca, including the chemical profiling, the reasonable quality standard and the systematic effects evaluation 13 .Previously, people mainly focused on the nutrient compositions in Maca, such as proteins, amino acids and fatty acids. However, the secondary metabolites were mainly responsible for its multiple functions. The alkaloids, glucosinolates and macaenes should be deserved close attention 14 . UHPLC-ESI-Orbitrap MS was the valid solution for the chemical analysis of secondary metabolites in Maca, which could provide accurate full MS and MS/ MS fragments measurements of the target compounds with high sensitivity and precision. For quantitative analysis, UHPLC-ESI-QqQ MS could determine different types of constitutes irrespective of their ultraviolet absorption and the degree of separation by dynamic multiple reaction monitoring (DMRM) method. In terms of neuroprotective effects evaluation, the zebrafish models of neurodegenerative diseases has been recognized by increasing numbers of researchers . Since zebrafish embryos were susceptible to various toxins such as MPTP, which was used as an inducible model of neuronal loss. Transparency was also a unique attribute of zebrafish that permitted direct assessment of drug effects on the nervous system. Also, the other multiple advantages of zebrafish, such as small size, high fecundity and ease of phenotype recognition, made it well suited for high-throughput screening. Meanwhile, in the evaluation of neurodegenerative diseases, MPTP-induced zebrafish model generated a loss of dopaminergic neurons similar to the mid-brain lesion in the Parkinsonian patients 21 . Generally, in the present study, a sensitive and accurate strategy was developed for the comprehensive chemical analysis of Maca firstly. UHPLC-ESI-Orbitrap MS and UHPLC-ESI-QqQ MS were employed for the qualitative and quantitative analysis respectively. Totally, 160 constituents were detected and identified from 6 fractions of Maca extract. Fifteen of them were selected for the quality control study. Furthermore, the neuroprotective effect of Maca was studied by MPTP-induced zebrafish model with dopaminergic neuronal loss. 80% methanol elution fraction (Fr 5 ) showed significant neuroprotective effects, followed by 100% part (Fr 6 ). Also, the inhibition of acetylcholinesterase (AChE) and butyrylcholinesterase (BuChE) experiments were performed in vitro, which also supported the 80% methanol elution fraction (Fr 5 ) as the strongest neuroprotective fraction. Imidazole alkaloids, macamides and macaenes were predominant constituents in Fr 5 . AChE and BuChE were regarded as the potential targets for the neuroprotective effects of Maca. ## Results and Discussion The chemical profiling analysis of Maca. The use of Octadecylsilyl gel (ODS gel) remarkably increased chromatographic peaks, and thus achieved the enrichment of minor constitutes. UHPLC-ESI-Orbitrap MS was employed for the analysis. The element compositions of reference standards, unknown compounds and their MS n fragments were calculated by accurate high resolution mass measurements. The total ion chromatogram (TIC) of Maca extract was shown in Fig. 1. In 6 enriched fractions, 160 ingredients were observed and identified, which were separated and enriched specifically according to their polarities. The main components in each fraction was illustrated as follows: Fr 1 (organic acids), Fr 2 (glucosinolates), Fr 3 (β -carboline alkaloids), Fr 4 (common amide alkaloids, macaridines), Fr 5 (mainly imidazole alkaloids, macamides, macaenes), Fr 6 (macamides). Among them, 30 compounds were identified unambiguously by comparing with the retention time and MS data of reference standards. The other 130 constituents were deduced by their collision induced dissociation pathways together with literatures. Alkaloids, glucosinolates and macaenes were the major chemical ingredients of Maca. Identification of alkaloids in Maca. Macamides, common amide alkaloids, macaridines, β -carboline alkaloids and imidazole alkaloids were five subtypes of characteristic alkaloids in Maca. As we knew, alkaloids always presented high sensetivity in the positive ion mode. Totally, 121 unknown alkaloids were identified. A 54 and A 32 were two representatives of macamides, the fragmentation patterns of which were shown in Figs 2 and 3 respectively. In the MS/MS analysis of them, the diagnostic fragments of A The benzene ions were also detected in the MS/MS analysis of two compounds. Thus, A 58 and A 55 were identified as N-benzylheptadecanamide (Figure S3) and N-(3-methoxybenzyl)-(9Z)-octadecenamid e (Figure S4). Furthermore, 64 macamides and common alkaloids (A 1 -A 64 ) were deduced according to similar fragmentation pathways (Table S1). S5). The other 2 macaridines (A 66 , A 67 ) were identified following the same dissociation pathways (Table S2). Besides, the MS dissociation of some characteristic β -carboline alkaloids and imidazole alkaloids such as (1 R, 3 S)-1-methyltetrahydro-β -carboline-3-carboxylic acid (A 110 ) and 1, 3-dibenzyl-4, 5-dimethylimidazolium (A 82 ) were investigated as well. A 82 showed the representative fragments [M + H-benzyl] + and [benzyl] + at m/z 185.1075 and m/z 91.0541 according to the literature 22 . In line with the mass data and literatures 23,24 , the cleavage of NH ) were assigned as the successive loss of CH 2 moieties. All the fragmentation pattern of A 84 revealed itself to be 1, 3-dibenzyl-2-ethyl-4, 5-dimethylimidazilium (Figure S6). The MS behaviors of 38 imidazole alkaloids (A 68 -A 105 ) were in line with that of A 84 (Table S3). The ) were generated from the neural loss of phenyl and benzyl group from [M + H] + ion respectively, which suggested that A 111 was (1R, 3S)-1-ethyltetrahydro-β -5,6-carboline-3-carboxylic acid (Figure S7). Sixteen β -carboline alkaloids (A 106 -A 121 ) were also assigned following the same dissociation pathways (Table S4). Identification of glucosinolates and acids in Maca. Glucosinolates and acids displayed the [M-H] − ions with sufficient abundance in the negative ion mode. In agreement with literatures, the cleavage of SO 3 , glucose, C 8 H 7 ON, C 8 H 7 NS, C 6 H 12 O 5 S and H 2 O groups was detected in the MS/MS spectra 25,26 . Take G 5 as a reference compound for example (Fig. 5), the ESI-MS spectrum of G 5 ( 26 . Additionally, the loss of H 2 O group was considered as the representative fragmentation pathway in acids (Table S5). In the negative ion mode, 14 glucosinolates (Table S6) were detected and identified. Among them, 3 glucosinolates (G 3 , G 5 , G 8 ) and 5 acids (C 1 -C 5 ) were isolated and characterized as reference standards. G S8). ## Identification of macaenes in Maca. As reported before, macaenes were also the characteristic compounds from Maca 27 . In the present study, 11 macaenes were observed in the positive ion mode, among of which, 5-oxo-6E, 8E-octadecadienoic acid (M 7 ) has high content. M 7 demonstrated obvious fragments at m/z 277.2165 and m/z 259.2059, indicating the continuous loss of 2H 2 O groups. S7). ## Identification of other compounds in Maca. Nine other compounds were detected and identified including flavonoids, organic esters, pyridine and benzylcyanide constitutes (Table S8). For example, O 6 showed the loss of C 2 H 4 , C 3 H 6 , and C 9 H 10 O 2 clearly in the MS/MS spectrum, which was in accordance with Licochalcone A 28 . Quantitative analysis of samples. Ion polarity switching mode was adopted in the quantitative analysis. Glucosinolates and acids were determined in the negative ion mode, while others were detected in the positive ion mode. ## Method validation. Linearity of calibration curves, limit of detection (LOD) and limit of quantification (LOQ). The internal standard method was employed to calculate the contents of 15 compounds in Maca. The standard solutions with internal standard were diluted with methanol to six different concentrations for the construction of calibration curves. The ratio of peak area to internal standard (Yi/Ys) of each analyte was plotted against the injection concentration (X, ng.mL −1 ). All the calibration curves indicated good linearity with determination coefficients (r) from 0.9951 to 0.9998. The limits of detection (LOD) and the limits of quantification (LOQ) were evaluated at a signal-to-noise ratio (S/N) of 3/1 and 10/1 respectively. The parameters of LOD and LOQ for each constituent in this experiment were from 0.05~164.95 ng.mL −1 and 0.05~824.74 ng.mL −1 (shown in Table 1 29,30 ). Precision, stability and repeatability. The intra-day and inter-day precisions of the present method were calculated by analyzing the standard solution under the optimized experimental conditions. The RSD values of them were 0.41%~2.46% and 0.43%~2.71%. The RSD values of stability of each constituent in 48 hours at room temperature (n = 6) were 1.01%~2.84%. Furthermore, the sample solutions for Maca were prepared in parallel (n = 6) to evaluate the repeatability and achieved the RSD of 1.02%~2.63% (Table S9). Recovery. The recovery was used to evaluate the accuracy of the method. Nine copies of 1 g Maca were taken for recovery test. The mixed standard solutions of 15 constituents were added according to three levels (1:0.8, 1:1, 1:1.2) respectively. The mixtures were treated as the preparation procedure of sample and analyzed using the method described above. Recovery (R) was calculated as R = 100 (Mmeasured− Minitial)/ Madded (Mmeasured = measured amount in the recovery sample, Minitial = initial amount in the sample, Madded = amount in the standard solution used) for each compound. The average recovery rate of each constituent was 96.27%~98.89%, with the RSD values from 0.60% to 3.11% (Table S10), which met the requirements for the determination of 15 constituents in Maca. ## Analysis of samples. This validated UHPLC-ESI-QqQ MS method was used for the quantification analysis of 15 constituents in 17 batches of Maca under the DMRM mode (Fig. 6). The analysis time was shortened to 15 minutes. Each constituent was calculated by their respective calibration curve, and the quantification results were shown in Table 2. 15 markers were identified unambiguously by comparing the retention times and transitions in DMRM mode of reference standards. Two internal standards were employed to guarantee the accuracy of determination. The polarity switching mode of QqQ MS was used to achieve the highest response intensities of each constituent. The quantification of Licochalcone A and 3-benzyl-1, 2-dihydro-N-hydroxypyridine-4-carbaldehyde was investigated as well, which were two representative constituents but with low contents in Maca. The total amounts of investigated 15 compounds in Maca varied from 1.20% to 8.12%. Among them, glucosinolates (10.97~79.84 mg.g −1 ) and alkaloids (0.54~2.99 mg.g −1 ) were the predominant constituents. The contents of glucosinolates were significantly more than the other markers. Significant variations were observed in different alkaloids. The contents of macamides were the highest with the amounts from 0.54 to 2.95 mg.g −1 . Macamides were reported as one of the important secondary metabolites of Maca with neuroprotective, anti-fatigue, improving fertility and sexual functions effects. What's more, β -carboline alkaloids (A 121 ) and macaridines (A 66 , A 67 ) were also detected in present analysis, the contents of which were 0~1.6 mg.g −1 and 1.47~72.61 mg.g −1 . β -carboline alkaloids displayed neuroprotective effect 24 . In general, glucosinolates and β -carboline alkaloids in the samples from Peru were higher than those cultivated in China. However, the contents of organic acids, macaridines, common amide alkaloids, macamides in Maca from China were higher than or similar to that from Peru. The results indicated that Maca cultivated in China especially in Yunnan province could be used as the potential substitute. Among the Chinese samples, the contents of secondary metabolites in Yunnan province (Sample 7-Sample 17) were higher than those from the other two origins. The contents of effective constitutes in the samples from Xinjiang province (Sample1and Sample 2) were relatively lower. Generally, the content tendency of different types constitutes in Maca was similar to each other. Glucosinolates (G 3 , G 5 , G 8 ) and macamides (A 52 , A 54 ) presented large amount (glucosinolates were 8~89 times than macamides), followed by organic acids (C 3 , C 5 ), macaridines (A 66 , A 67 ), common amide alkaloids (A 4 , A 5 ), Licochalcone A (O 6 ). In addition, Maca was always classified according to different hypocotyls colors in the market. In present study, a yellow and a purple Maca samples from the same growing environment and growth cycle were included in the analysis. Interestingly, the content of glucosinolates in 15 yellow Maca samples (Samples 1-5. 8-17) was higher than that in 2 purple Maca samples (Samples 6-7). However, from the data we collected, the types and contents of ingredients in purple and yellow Maca samples had no significant difference with each other, which was also in accordance with the literature 31 . Accurately, the geographical origin played a more important role than color. The results indicated that the contents of secondary metabolites in yellow Maca was in line with those in purple Maca generally. The quality discrepancy of them was not as significant as propagandized in market. Also, the main components in different parts of Maca (roots and up-ground parts) were studied preliminary as well. The contents of active ingredients in the roots were higher than those in leaves. Three experimental samples were collected from the same Maca plant (Sample 12 and Leaves of Maca). To the best of our knowledge, in the biosynthetic pathways of macamides, benzylglucosinolate, free linoleic and linolenic acids, benzyl isothiocyanate and benzylamine were direct or indirect precursors 32,33 . The accumulation of macamides was significantly associated with the contents of fatty acid and benzylamine. Thus, one of the important quality criterions of Maca was the levels of macamides and glucosinolates. In present study, both of them were higher in roots than those in leaves. The up-ground parts were regarded as the potential resources for the enrichment of the effective macamides. In the cultivation of Maca, the harvest time, longitude and altitude affected its quality greatly. The present investigation provided a valuable strategy for the quality evaluation of Maca from their chemical profiling no matter the samples from different regions and varieties. Neuroprotective effect screening of active fractions and pure compounds. Although, Maca has displayed the neuroprotective activity both in vivo and in vitro, its corresponding bioactive components and possible mechanism were still not clearly 8,9,34 . In present study, the zebrafish embryos were treated with MPTP to form the DA neuronal loss model, which was always used for the neuroprotective evaluation. The DA neuronal loss site of zebrafish was located in the ventral diencephalic clusters (indicated by red brackets) (Fig. 7). As a result, by compared with the controls, MPTP model group resulted in 70% reduction of TH-positive neurons in the diencephalic area of the zebrafish embryo. The total Maca methanol extract displayed neuroprotective effect against MPTP-induced toxicity in zebrafish. Then, the activity evaluation of 6 fractions enriched by ODS column and 2 pure compounds was examined. The fractions and pure compounds were able to inhibit DA neuron loss by approximately 30~60% compared with the MPTP group. Consequently, Fr 4 -Fr 6 could increase the amount of dopamine neurons in different degrees with dose-dependent effect, especially for 80% methanol elution fraction (Fr 5 ) (Fig. 8). However, Fr 1 -Fr 3 showed no or little corresponding neuroprotective effects. The LC-MS chromatograms of Fr 4 -Fr 6 were shown in Figures S10-S12. N-benzylhexadecanamide (A 54 ) and N-acetylbenzylamine (A 5 ) were two important active secondary metabolites in Maca with higher abundance. Because of their relative higher contents, A 5 and A 54 were easily to be isolated from the plant. Thus, A 5 and A 54 were chosen as the representative ingredients for the neuroprotective effect evaluation. Combination with chemical profiling analysis, common alkaloids and macaridines were predominant constitutes in 50% methanol elution fraction (Fr 4 ). 80% methanol elution fraction (Fr 5 ) mainly contained imidazole alkaloids, macamides and macaenes. Macamides were also observed in 100% methanol elution fraction (Fr 6 ). Interestingly, both Fr 5 and Fr 6 contained N-benzylhexadecanamide (A 54 ) with high abundance. Therefore, A 54 was selected as the representative pure compound for the effective confirmatory experiment. As the results, at the dose of 30 and 10 μ g mL −1 , A 54 could significantly prevented the MPTP-induced decrease in TH + region area, and showed neuroprotective actions in a dose-dependent manner. Meanwhile, N-acetylbenzylamine (A 5 ) demonstrated noteworthy protections against MPTP-induced toxicity with dose-dependent manner, which was the major constituent in Fr 4 fraction. The results were in accordance with the neuroprotective properties of the pentane extract of Maca reported before. Macamides, a class of benzylated and 3-methoxylbenzylated alkamides, were identified as the major characteristic effective compounds of Maca 35 . Because of its excellent solubility profile, macamides could act on the endocannabinoid system and showed fatty acid amide hydrolase (FAAH) inhibitory activities 5 . In present study, A 54 also showed dose-dependent neuroprotective effect by restoring dopaminergic neuronal loss. The results inspired that Maca was available for the impairment of the endocannabinoid system in dopaminergic neurons, which was closely related to a serious of neurodegenerative diseases, such as Alzheimer's disease, Parkinson's disease and so on. Additional studies of macamides should be carried out on the mechanism of FAAH inhibition involving the endocannabinoid system 36 . Further, macaridines, imidazole alkaloids and macaenes in Maca also should be concerned. Additionally, the results also demonstrated that zebrafish could be useful for the neuroprotective activity evaluation. The brain of zebrafish embryonic contained clusters of dopaminergic with the characteristics of inexpensive, low-maintenance and abundantly produce offspring. Furthermore, a significant inherent advantage was its transparency. Therefore, zebrafish could provide invaluable insights for large scale screening, drug discovery, modeling behavioral and functional parameters of neurodegeneration disorders and preclinical treatments . Parkinson's disease (PD) was a progressive neurodegenerative disorder characterized by the selective loss of nigral dopaminergic neurons and a reduction in striatal dopaminergic fibers 44 . MPTP was metabolized to MPP + , which generated a similar loss of dopaminergic neurons with corresponding Parkinsonian symptoms 45 . Thus, we used zebrafish treated with MPTP as the neuronal loss model. Meanwhile, as reported in patients with Parkinson's disease, both the dopaminergic and cholinergic systems underwent degeneration, which led to deficits in dopamine and acetylcholine at synapses. As for the cholinergic system, MPTP also decreased the gene expression of choline acetyltransferase (ChAT) while increased the expression of acetylcholinesterase (AChE) 46 . In present study, we attempted to bring new evidence supporting the potential neuroprotective action of Maca in PD, focusing on the interaction between dopaminergic and cholinergic systems. Dopamine was supposed to possess modulatory effect in cholinergic transmission, which played a critical role in modulating cortical cholinergic activity by GABAergic intracortical circuits 46 . And the dopaminergic system might be dysfunctional in AD, which was possibly generated by the disrupting of the cholinergic system 47,48 . Thus, after the evaluation of Maca on dopamine neuronal loss model in zebrafish by the method of anti-tyrosine hydroxylase (TH) whole-mount immunostaining, the AChE and BuChE inhibition activities of 6 fractions and pure compound were carried out. Collectively, these findings indicated that Maca showed neuroprotective activity through the synergistic effect of dopaminergic system and cholinergic system, which needed to be further validation. Another important reason for the analysis of AChE and BuChE inhibition was from the structures of the neuroprotective chemicals. As we all know, the structural similarities of endocannabinoids and macamides indicated their potential neuroprotection effects. The purified macamides or its synthetic derivatives suggested highly possible activities on the endocannabinoid system 49 . As we knew, the impairment of the endocannabinoid system in dopaminergic neurons would result in many neurological and psychiatric disorders such as Alzheimer's disease, Parkinson's disease, depression and schizophrenia 50,51 . Meanwhile, dual FAAH/ChE inhibitors, with well-balanced nanomolar activities might be considered as new promising candidates for Alzheimer's disease treatment, which also suggested the close relationship between endocannabinoid system and cholinergic system 52 . This possible collaboration between cholinergic and dopaminergic neurotransmission in the midbrain raised the possibility of targeting both systems simultaneously to treat PD and AD in the future. Therefore, the preliminary analysis of AChE and BuChE were included in the manuscript. The in vitro AChE and BuChE inhibition assays showed (Table 3) that Fr 4 (IC 50 = 50.78 μg.mL −1 ), Fr 5 (IC 50 = 5.37 μ g.mL −1 ) and Fr 6 (IC 50 = 15.77 μg.mL −1 ) displayed significant AChE inhibitory activity. Similarly, Fr 4 (IC 50 = 45.11 μ g.mL −1 ), Fr 5 (IC 50 = 5.41 μg.mL −1 ) and Fr 6 (IC 50 = 23.39 μg.mL −1 ) showed significant BuChE inhibitory activity. The results were in accordance with those in vivo experiments. As the mutual constitute in Fr 5 and Fr 6 , N-benzylhexadecanamide (A 54 ) was selected to validate this mechanism as pure compound, which also displayed high AChE (IC 50 = 14.23 μg.mL −1 ) and BuChE (IC 50 = 17.54 μ g.mL −1 ) inhibitory activities. What's more, A 54 were reported as dual AChE/BuChE inhibitors without remarkable side effects . In conclusion, Fr 5 and macamides produced the healing efficacy by increasing the acetylcholine and butyrylcholine level in brain. ## Conclusions In this paper, a high sensitive and efficient strategy for the integrating quality control of Maca was established. Alkaloids, glucosinolates, and macaenes were detected as the predominant secondary metabolites in this plant. Among them, five types of alkaloids were observed, including macamides, imidazole alkaloids, macaridine, β -carboline alkaloids and common amide alkaloids. According to the quantitative analysis of 15 major chemical markers in Maca, the contents variety of different origins, crust colors and different parts of plant were clarified. Meanwhile, the neuroprotective effects of 6 fractions against MPTP induced neurotoxicity zebrafish model were examined. 80% methanol elution fraction (Fr 5 ) and 100% methanol elution fraction (Fr 6 ) were regarded as the most effective neuroprotective parts. Both of these two parts contained macamides. The verification test by pure compounds also proved that alkaloids were the neuroprotective constituents in Maca. The inhibition of AChE and BuChE was one of the possible mechanisms. These findings suggested that Maca was a valuable health care food for the neurodegeneration disease such as Alzheimer Disease, Parkinson's desease and so on. ## Chemicals and Reagents. Methanol for extraction was obtained from Honeywell Burdick and Jackson (Swedesboro, NJ, USA). HPLC grade acetonitrile was purchased from Merck Company (Rahway, NJ, USA). HPLC grade formic acid was provided by ROE Scientific Inc. (Delaware, USA). MPTP, Dimethyl Sulphoxide (DMSO), Trizma, 5,5′ -Dithiobis (2-nitro-benzoic acid), Acetylthiocholine iodide (ATCI), AChE from Electophorus electricus electric eel Type VI-S (AChE), S-Butyrylthiocholine iodide (BTCI), BuChE from equine serum (BuChE), Physostigmine (eserine) were purchased from Sigma-Aldrich (St. Louis, MO, USA). All other chemicals and reagents were of high analytical grade. The reference standards were isolated and purified by authors, including m-methoxybenzylglucosinolate(G 3 ), benzylglucosinolate(G 5 ), p-hydroxybenzylglucosinolate(G 8 ), N-(3-methoxybenzyl)-hexadecanamid e(A 52 ), N-benzylbenzamide(A 8 ), N-benzylhexadecanamide(A 54 ), LicochalconeA(O 6 ), 3-benzyl-1,2 -dihydro-N-hydroxypyridine-4-carbaldehyde(A 66 ), (3-methoxybenzyl)-N-pyridine-4-carbaldehyde(A 67 ), N-acetylbenzylamine (A 5 ), 2-phenylacetamide(A 4 ), Nicotinic acid(C 3 ), Succinic acid(C 5 ), (1R, 3S)-1-methyl-β -carboline-3-carbaldehyde(A 121 ), Dibutyl phthalate(O 9 ). Silica gel, RP-C 18 (200-300 mesh, Qingdao Marine Chemical Factory, Qingdao, China), sephadex LH-20 column chromatography and preparation liquid chromatography were used in the isolation. Their structures were identified by MS, NMR and UV spectra (Figure S13). 4-Aminohippuric acid (IS 1 ), Evodiamine(IS 2 ) were employed as the internal standards, which were purchased from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). The purities of all the reference standards were over 95%. ## UHPLC-ESI-Orbitrap MS analysis and UHPLC-ESI-QqQ MS analysis. Preparation of samples solu- tions. 3 g powder of Maca was extracted under ultrasonic in 30 mL methanol for 30 min. After centrifugation at 12000 g for 10 min, the supernatant was condensed to 2 mL, and applied to an Octadecylsilyl gel column (ODS, 200 mL, 3.5 cm × 60 cm) eluted with 5%, 10%, 30%, 50%, 80%, 100% (800 mL) methanol. Each 800 mL of the elution was collected as one fraction 56 . All the 6 fractions (Fr 1 -Fr 6 ) were analyzed by LC-MS/MS and employed for neuroprotective evaluation of Maca. Maca samples (fine powder) from different places (Table S11) were weighed 1 g and dissolved in 5 mL methanol. After ultrasonic extraction for 30 min, the samples were centrifuged at 12000 g for 10 min. Then the supernatant of each sample was transferred to a clean test tube. Preparation of standard solutions. The 15 reference standards were dissolved in HPLC grade methanol to achieve a stock solution with concentration of 1.0 mg.mL −1 for each compound respectively, then added corresponding volume according to their proportions in the sample and mixed. An internal standard stock solution was also prepared in a concentration of 0.92 mg.mL −1 for IS 1 and 2.49 mg.mL −1 for IS 2 . The standard stock solutions were kept at 4 °C before analysis. Analytical conditions. In qualitative analysis, the assay was performed using an ultimate 3000 hyperbaric liquid chromatography system coupled to a LTQ Orbitrap mass spectrometer via an ESI interface. The Liquid chromatographic separations of the analytes were performed by a Thermo AcclaimTM120 C18 column (250 mm × 2.1 mm, 3 μ m). The mobile phase consisted of 0.1% formic acid in water (solvent A) and acetonitrile (solvent B). The samples were eluted with the following linear gradient: 1% B at 0-5 min, 1-30% B at 5-20 min, 30-50% B at 20-30 min, 50-70% B at 30-40 min, 70-95% B at 40-60 min, 95% B at 60-80 min. The flow rate was 0.3 mL min −1 . The injection volume was 5 μ L. The temperature-controlled column oven was set at 30 °C and the sampler was set at 4 °C. The ESI source parameters were as follows: Both positive and negative ionization modes were used in the analysis. For positive mode, the capillary temperature was 350 °C, shealth gas (N 2 ) flow rate was 40 psi and aux gas flow rate was 10 psi, ion spray voltage was set at 3.5 kv. While for negative mode, the capillary temperature was 350 °C, shealth gas (N 2 ) flow rate was 35 psi and aux gas flow rate was 10 psi, ion spray voltage was set at − 3.2 kv. In the FT cell, full MS scans were acquired in the range of m/z 50-1500 with a mass resolution of 30,000. The MS/ MS and MS 3 experiments were set as data dependent scan. In the quantitative analysis, an Agilent 6490 A triple quadrupole LC-MS system (Agilent Corporation, MA, USA) equipped with G1311A quaternary pump, G1322A vacuum degasser, G1329A autosampler and G1316A therm was employed. Chromatography was performed on a Thermo Hypersil Gold-C18 column (2.1 mm × 50 mm, 1.9 μ m) with the column temperature at 30 °C. Mobile phase consisted of water containing 0.2% formic acid (A) and acetonitrile (B) and pumped at a flow rate of 0.3 mL.min −1 . A gradient program was used as follows: 1% B at 0-1 min, 1-5% B at 1-2 min, 5-10% B at 2-4 min, 10-50% B at 4-7 min, 50-90% B at 7-10 min, 90-100% B at 10-13 min, 100% B at 13-15 min. The injection volume was 2 μ L. Analytes were quantitated by monitoring the precursor-product combination in the DMRM mode using ion polarity switching mode. To ensure the desired abundance of each compound, the CE values and other parameters were optimized and illustrated as follows: cycle time, 300 ms; For positive mode, capillary voltage, 3 kv, nozzle voltage, 1.5 kv, Delta EMV(+), 200 v. For negative mode, capillary voltage, − 2 kv, nozzle voltage, − 1 kv, Delta EMV(−), 200 v. The optimized mass transition ion pairs (m/z) for analytes and the detection of the conditions of the compounds were shown in Table 4. Neuroprotective effect screening of fractions and compounds in Maca. Stock solutions of MPTP (10 mg.mL −1 ) were made by adding water directly to the bottle. MPTP was diluted in Holtfreter's solution to achieve final concentration of 200 μ M. The samples for neuroprotective effect of Maca in zebrafish including total extract, aforementioned 6 fractions and two pure compounds (A 5 , A 54 ) were dissolved in DMSO to get stock solution according to their toxic limit. Nomifensine, a dopamine transporter (DAT) inhibitor as positive control, was dissolved in DMSO to get the concentration of 3 μ M. The content of DMSO in solutions was not more than 0.5%. The neuroprotective effect assessment was performed on zebrafish model treated with 200 μ M MPTP. Wild-type zebrafish embryos (1 days post fertilization (dpf)) were treated with pure compounds or fractionsin zebrafish model for 2 days. After treatment, anti-tyrosine hydroxylase (TH) whole-mount immunostaining was performed using literature reported method 57,58 . Zebrafish were fixed in 4% paraformaldehyde in PBS for 5 h, rinsed, and stored at − 20 °C in 100% ethanol. Briefly, fixed samples were blocked (2% lamb serum and 0.1% BSA in PBST) for 1 h at room temperature. A mouse monoclonal anti-TH antibody (1:200 diluted in blocking buffer, MAB318, Millipore) was used as the primary antibody and incubated with the sample overnight at 4 °C. The next day, samples were washed 6 times with PBST (30 min each wash), followed by incubation with goat anti-mouse antibody (1: 500 diluted in blocking buffer, Alexa Fluor ® , USA) as secondary antibody for 1 h at room temperature in the dark. After staining, zebrafish were flat-mounted with 3.5% methylcellulose and photographed. The experimental results of each fraction and compound were obtained from the three zebrafish statistics. Every zebrafish was taken a picture and the gray scale was calculated by IMAGE J software. As a result, representative pictures of dopaminergic neurons in the zebrafish brain from different treatment groups were selected. TH + neurons in the diencephalic area of the zebrafish brain were considered as DA neurons. For quantification of neuronal area, the periphery of each cluster was outlined by manually tracing the edge. The area (Am 2 ) of each enclosed region was measured, and the subsections were summed to give the total cluster area. Origin software (MicroCal Software, Inc.) was used to generate graphs. Statistical significance was obtained by performing one-way ANOVA test using SPSS software. Quantitative analysis of TH + neurons was examined in each treatment groups. Values were expressed as a percentage of the control. #p < 0.05 vs. control; *p < 0.05 and **p < 0.01 vs. MPTP group. ## AChE and BuChE inhibition assay. A modified Elman's test was performed to investigate the AChE and BuChE inhibitory potency of fractions and compounds 59 . The data was measured in Microtiter plates 96 wells (Sterilin Art, No.611F96) and recorded on 1420 multilabel counter (Perkin Elmer, Wallac Victor 2 ). The solutions of enzyme (at 2.43 units.mL −1 concentration) were prepared in Trizma buffer (7.09 mg.mL −1 , pH 8). The solutions of reference (eserin), pure compounds and fractions (20 μ L each, at 0.025 to 1000 μ g.mL −1 concentration) in DMSO and BuChE/AChE (40 μ L) were added to buffer (190 μ L) and incubated at 25 °C for 5 min. DTNB (20 μ L) and acetylthiocholine iodide (ATC) (20 μ L) were added to enzyme-inhibitor mixture to investigate the reaction. The production of yellow anion was determined for 10 min at 405 nm. By using same methodology, similar solution of enzyme without the inhibitor was processed, which acted as a control. The blank measurement consisted of substrate (20 μ L), DTNB (20 μ L), DMSO (20 μ L) and buffer (230 μ L). The experiment was done triplicate. The percentage inhibition was calculated using the following equation: Results were expressed as mean ± SD. Differences among the groups were subjected to a one-way ANOVA (analysis of variance) followed by Duncan's multiple range. Statistical significance was accepted when a p-value was less than 0.05. ## Ethics statement. We confirm that all methods were carried out in accordance with relevant guidelines and regulations. We confirm that all experimental protocols were approved by Medicine Ethics Committee in Institute of Chinese Materia Medica, China Academy of Chinese Medical Sciences.

chemsum

{"title": "Chemical profiling analysis of Maca using UHPLC-ESI-Orbitrap MS coupled with UHPLC-ESI-QqQ MS and the neuroprotective study on its active ingredients", "journal": "Scientific Reports - Nature"}

base-catalyzed_aryl_halide_isomerization_enables_the_4-selective_substitution_of_3-bromopyridines

## Abstract: The base-catalyzed isomerization of simple aryl halides is presented and utilized to achieve the 4-selective etherification, hydroxylation and amination of 3bromopyridines. Mechanistic studies support isomerization of 3-bromopyridines to 4-bromopyridines proceeds via pyridyne intermediates and that 4-substitution selectivity is driven by a facile SNAr reaction. Beneficial aspects of a tandem aryl halide isomerization/selective interception approach to aromatic functionalization are demonstrated. The synthetic value of aryl halides derives from their thoroughly studied reactivity that allows reliable and predictable access to functionalized aromatic compounds. 1 The utility of these widely available substrates can be significantly increased as new reactivity modes are discovered and applied. In this regard, catalytic aryl halide isomerization drew our attention as a relatively underdeveloped yet potentially useful process (eq 1). 2 The rearrangement of halogenated arenes under basic conditions has been extensively studied in the context of "halogen dance" chemistry. 3 Early studies by Bunnett on the base-catalyzed isomerization of 1,2,4-tribromobenzene into 1,3,5-tribromobenzene revealed rearrangement occurs via intermolecular halogen transfer, resulting in regioisomeric mixtures and disproportionated side products. 4 This catalytic rearrangement requires an acidic arene that can generate electrophilic halogen transfer intermediates (e.g. tetrabromobenzenes); thus, isomerization is observed for tribromobenzenes but not for simple aryl halides. 5 This insight guided decades of development of modern "halogen dance" methodology, wherein stoichiometric and irreversible metalation of haloarenes can lead to rearrangement through intermolecular metal-halogen transposition. 6 In addition to typically requiring stoichiometric lithium bases under cryogenic conditions, a synthetically useful dance requires a thermodynamic gradient in order to drive a selective rearrangement. 3 In this regard, important achievements have been made in iden-tifying specific classes of metalated haloarenes that rearrange as a strategy for electrophilic functionalization. 7 We sought to identify more mild and general conditions for aryl halide isomerization as an entry to developing new arene functionalization methods. Inspired by sporadic reports of rearranged aryl halide side products 8 in reactions involving aryne 9 intermediates, we hypothesized that nonnucleophilic bases could enable reversible HX elimination/addition as an additional isomerization pathway (Figure 1a). 10 We further proposed that pairing isomerization with a tandem substitution reaction could provide a driving force for nontraditional selectivity in aromatic substitution reactions (Figure 1b). 11 We herein describe initial studies on a general approach to catalytic aryl halide isomerization and demonstrate its utility as a new route to 4-functionalized pyridines. 12 Figure 1. A general approach to aryl halide isomerization and its application to a new selective substitution reaction. We speculated bases that reversibly deprotonate aryl C-H bonds may create conditions capable of isomerizing aryl halides. 13 This led us to investigate the non-nucleophilic organic superbase P4-t-Bu (pKa' 30.2 in DMSO) as a potential isomerization catalyst. 14 In 1,4-dioxane, we discovered P4-t-Bu catalyzes the isomerization of 2-bromobenzotrifluoride (1) into all possible regioisomers (Scheme 1a). Under these conditions, 3-bromobenzotrifluoride (2) and 4bromobenzotrifluoride (3) interconvert but do not form 2bromobenzotrifluoride (Scheme 1b). No protodehalogenated or polyhalogenated side products are observed, and we note isomerization occurs to a lesser extent for 4iodobenzotrifluoride (4). 15 A variety of other bromoarenes (5, 6 and 7) also isomerize, including the formation of 4bromopyridine from 3-bromopyridine (8). 16 Although further studies on the scope of this process are ongoing, these observations suggest P4-t-Bu-catalyzed aryl halide isomerization is a general and reversible process. Scheme 1. P4-t-Bu-catalyzed aryl halide isomerization. a a Yields determined by 1 H NMR spectroscopy; the mass balance is less than 100% with no observed haloarene side products; conditions for Scheme 1b are as shown in Scheme 1a. b Reaction performed in cyclohexane for 14 h. As a broader objective, we questioned if aryl halide isomerization could be utilized to address current challenges in aromatic functionalization. As a halogen migrates around an arene, we reasoned that differing electronic properties of isomeric C-X bonds could provide a source to differentiate interconverting isomers and drive an overall selective transformation. 17 A mechanistic outline for the application of this concept to the 4-substitution of 3-bromopyridines is shown in Scheme 2. This pathway exploits the inherent preference for 4-bromopyridines to undergo nucleophilic aromatic substitution (SNAr) over 3-bromopyridines. 18 However, a likely challenge is avoiding nucleophilic addition to the proposed 3,4-pyridyne intermediate, as this could decrease the desired reaction's yield or regioselectivity. 19 Successful development of this protocol would offer an attractive route to 4functionalized pyridines from 3-bromopyridines, which are more commercially available 20 and stable 21 than 4halogenated congeners. This method would also complement other recently developed methods for 4-selective nu-cleophilic pyridine C-H functionalization, including McNally's powerful heterocyclic phosphonium salt approach. 22,23 Scheme 2. Proposed pathway for the 4-selective substitution of 3-bromopyridine (B = base, NuH = nucleophile). As the proposed process requires stoichiometric base, we first investigated the use of hydroxide bases as more practical reagents for promoting aryl halide isomerization. Hydroxide bases are known to generate and be compatible with aryne intermediates, and we identified that 3-bromopyridines isomerize in the presence of 18-crown-6-ligated KOH in N,N-dimethylacetamide (see Supporting Information). 24 Using these conditions, we then employed two separate strategies for optimizing a 4-selective etherification reaction of 3-bromopyridine (Scheme 3a). When 1 equivalent of 3bromopyridine ( 8) reacts with 4 equivalents of alcohol 9, a 2.4:1 ratio of 4:3-substituted product (10:11) is obtained in 54% overall yield. This ratio is comparable to reported selectivities for nucleophilic additions to 3,4-pyridyne, which typically range from 1:1 to 3:1 for 4:3 addition selectivity. 25 The 4-selectivity increases as higher ratios of pyridine:alcohol (8:9) are used, a result perhaps explainable by less alcohol intercepting a 3,4-pyridyne intermediate. Based on this observation, we hypothesized that added bromide salts may enable more efficient isomerization and prevent undesired side reactions. 26 When 50% KBr is added to a reaction using a 1.5:1 pyridine:alcohol (8:9) ratio, the yield increases to 76% with >14:1 4-selectivity compared to 67% yield and 8.6:1 4-selectivity in the absence of bromide salt. A reaction profile with the optimized conditions shows the rapid formation of a low concentration of 4-bromopyridine (approximately 5%) that decreases as the reaction reaches completion (Scheme 3b). Subjection of the 3-substituted ether product (11) to the reaction conditions does not result in mass balance loss, indicating the high 4-selectivity is not a result of selective decomposition or product rearrangement. 27 To test for the generation of 3,4-pyridyne under these conditions, when the alcohol is replaced with an excess of furan (12) the corresponding cycloadduct 13 forms in 42% yield (Scheme 3c). 28 We also subjected 3-iodopyridine (14) to the reaction conditions in the absence of alcohol; in the presence of furan (12) cycloadduct 13 forms and in the presence of KBr a mixture of 3-and 4-bromopyridine form (8 and 15, Scheme 3d). The observed mixture of 3-and 4bromopyridine supports the proposal of bromide addition to 3,4-pyridyne. Overall, these results are consistent with an isomerization pathway via 3,4-pyridyne and 4-substitution selectivity driven by a facile SNAr reaction. a Yields and selectivities determined by 1 H NMR spectroscopy of the crude reaction mixtures; yields represent total amount of both isomeric products 10 and 11; b 2.0 equiv of KOH used; c Conditions as shown in Scheme 3a using 1.5:1 ratio of 8:9 with 50 mol% KBr additive; see Supporting Information for details. A substrate scope for the 4-selective etherification of 3bromopyridines is provided in Table 1. 29 A range of primary and secondary alcohols are first shown using 1.5 equiv of simple bromopyridines. Sterically hindered alcohols (16) and those containing amino groups (17, 23, and 24), a terminal alkene (18) and a protected sugar (22) are suitable nucleophiles. Pyridines methylated in all positions react in high yield and selectivity (19, 20, 21 and 24), indicating that steric hindrance and acidic C-H bonds are tolerated on the arene. Pyridine biaryl substrates also selectively couple in the 4-position (25 and 26). Both 2-alkoxy (23) and 2-amino (28) substituents on the pyridine are tolerated, although we note the high selectivity observed for these substrates could originate from alcohol addition to a distorted 3,4-pyridyne intermediate. 19 Pyridine substrates with more electronwithdrawing groups undergo direct 3-substitution under the current reaction conditions (e.g. 3-bromo-2-(trifluoromethyl)pyridine). 30 Table 1. Scope of the 4-etherification of 3-bromopyridines. a a Yields are of purified 4-ether product; regioselectivities were determined by 1 H NMR spectroscopy of the crude reaction mixture; b Selectivity >6:1; see Supporting Information for details; c P4-t-Bu (1.3 equiv) used as base in place of KOH. An advantage of this functionalization strategy is demonstrated with substrates 29-33 in Table 1, where the bromopyridine substrate is obtained from commercial pyridines while a 4-halogenated isomer is either not available or significantly more expensive (see Supporting Information for a discussion). Using 1-1.2 equiv of these bromopyridines, 4substituted products featuring a carbamate (29), an acidic 29-33 from bromopyridine substrates where 4-halide not readily available amide (30), a 2,6-disubstituted pyridine (31), a nicotine isomer (32) and a ketal ( 33) can be rapidly accessed. This strategy can also utilize an arene's innate halogenation position as an entry to functionalizing more difficult to access C-H bonds. This is demonstrated in eq 2, where the 2,6-disubstituted pyridine 34 undergoes facile but unselective bromination in the 3-and 5-positions (35a and 35b). Application of conditions from Table 1 provides access to the 4-ether 36 through convergence of the regioisomeric bromopyridine mixture, highlighting an additional benefit of this methodology. 31 We next examined if other nucleophiles can participate in the 4-selective substitution of 3-bromopyridines. We found that indolines are effective coupling partners and this route provides straightforward access to 4-aminopyridines from readily available 3-bromopyridines (37-40, Scheme 4a). The 4-amination of 3-bromopyridine with indoline proceeds on gram scale with excellent selectivity (37). A single isomer of the 5-bromoindoline product 38 is obtained, demonstrating the chemoselectivity of aryl halide isomerization. Scheme 4. The 4-substitution of 3-bromopyridines with additional nucleophiles. a a Isolated yield of purified 4-substituted products; selectivities determined by 1 H NMR spectroscopy of crude reaction mixtures; b 1.5 equiv of 3-bromopyridine used; c Selectivity 9:1. It is interesting to note that 4-hydroxypyridine side products are not typically observed for the reactions in Table 1 even though KOH is used as a base. 32 To develop a 4hydroxylation protocol, we instead hypothesized tandem isomerization/substitution could be further sequenced with a base-promoted elimination reaction. This is demonstrated in Scheme 4b, where the use of β-hydroxyamide 42 as a nucleophile directly delivers the 4-hydroxylated product 43 in 50% yield with >10:1 selectivity. We speculate this reaction proceeds through the standard 4-substitution pathway followed by a facile base-promoted acrylamide elimination reaction. 33 This work demonstrates base-catalyzed aryl halide isomerization can be paired with SNAr reactivity to achieve unconventional substitution selectivity. In contrast, established "halogen dance" methodology relies on the controlled rearrangement of specific classes of stoichiometrically metalated haloarenes prior to treatment with electrophiles. 3 Thus, tandem isomerization/selective interception may be a complementary and general strategy for achieving nontraditional selectivities in aryl halide functionalization chemistry. ## ASSOCIATED CONTENT Supporting Information. The Supporting Information, containing experimental procedures and characterization data for all compounds is available.

chemsum

{"title": "Base-Catalyzed Aryl Halide Isomerization Enables the 4-Selective Substitution of 3-Bromopyridines", "journal": "ChemRxiv"}

recyclable,_biobased_photoresins_for_3d_printing_through_dynamic_imine_exchange

## Abstract: Transimination reactions are highly effective dynamic covalent reactions to enable reprocessability in thermosets, as they can undergo exchange without the need for catalysts, by exposing the materials to external stimuli such as heat. In this work, a series of five biobased vanillin derived resin formulations consisting of vanillin acrylate with vanillin methacrylate functionalized Jeffamines® were synthesized, and 3D printed using digital light projection (DLP). The resulting thermosets produced, displayed a range of mechanical properties (Young's modulus 2.05 -332 MPa) which allow for an array of applications. The materials we obtained have self-healing abilities which were characterized by scratch healing tests. Additionally, dynamic transimination reactions enable these thermosets to be reprocessed when thermally treated above their glass transition temperatures under high pressures using a hotpress. Due to the simple synthetic procedures and the readily available commercial Jeffamines®, these materials will aid in promoting a shift to materials with predominantly biobased content and help drift away from polymers made from nonrenewable resources. ## Introduction Polymeric materials, or plastics, combine unrivalled mechanical properties with low cost, and have become a staple commodity of our modern lifestyles. However, while offering many advantages, the new plastics economy still has major weaknesses that are becoming more evident by the day. Plastics are mainly produced from fossil fuels -a non-renewable resource-with a major carbon footprint that will become much more impactful as the population and demand increase. 1 Furthermore, environmental pollution caused by plastics is creating broad devastation in natural ecosystems and is considered one of the greatest environmental challenges of our time. 2 Based on this, the design of new polymeric materials should be mindful of sustainability practices that consider the process "from cradle to grave". Starting materials should be obtained from renewable feedstocks to reduce the carbon footprint from their production. 3 Vanillin is a biobased, aromatic compound derived from lignin that is widely used as a flavor and fragrance ingredient. However, most industrially produced vanillin is still obtained from petroleum-based sources as it represents a cheaper alternative. Sustainable vanillin production still accounts for a very small portion of the world's supply, but it is expected to continue to grow in the coming years, based on its promise as a potential substitute of petroleum derived monomers like bisphenol A (BPA) 4 and styrene. 5 It is anticipated that the sustainable production of vanillin-derived phenolics will be encouraged in large scale as new markets are developed for its application. 6 In addition to using renewable resources, once the useful life of the material is over, it is important to provide alternative methods for disposal, such as composting and recycling, to prevent them from ending up in landfills. Recycling is a high-value mechanism to repurpose plastics at the end of their life, ideally, into materials for applications of similar quality. The recycling pathway is desirable for most uses since this keeps the material in the economy, reducing the need for additional production. 7 Additionally, the ability to repair high value polymers or objects can extend their lifetimes, resulting in a reduction of single use plastics. Many examples of selfhealing polymers have been produced where covalent bonds can be reformed using a catalyst 8,9 or reconfigured through dynamic covalent exchange reactions. On the other hand, new materials that are compatible with competitive manufacturing methods, such as three-dimensional printing (3DP), are essential 3DP -the process of transforming computer-based designs into 3D structures-is achieved through layer-by-layer deposition or polymerization of a material. 13 The advantages, such as sustainability, 14 low waste production, 15 and customizability 16 have made this a popular technique, which is projected to augment or replace other common manufacturing methods like injection molding in the future. 17 Among 3D printing technologies, vat photopolymerization (VP) methods like stereolithography (SLA) and digital light projection (DLP), provide the highest resolution products and can provide isotropic materials, which are key to obtaining optimal mechanical performance. VP methods produce thermoset polymers, which contain covalent crosslinks between the polymer chains, and cannot be recycled by melt processing. In recent years, this limitation has been addressed through the development of covalent adaptable networks (CANs), and the number of CANs that use VP is rapidly growing. CANs incorporate dynamic functionalities that can be broken and reformed upon exposure to external stimuli, while in some cases, maintaining their overall bond density. 22 CANs are now bridging the gap between thermoplastics and thermosets, enabling thermosets to have the recycling and reprocessing advantages of thermoplastics. A few biobased CANs have been reported, usually consisting of associative bond exchange mechanisms, which are known as vitrimers. Although a minority, there are some reports of 3D printed materials that take advantage of CANs, with direct ink writing (DIW) being the most used technique. 27,28 Nonetheless, a few reports that combine the use of VP and CANs have been documented. 29 There have been a few examples of self-healing polymers used in 3D printing so far, but only a handful are compatible with VP. Most of the reports available combining CANs and VP use transesterification reactions. 38 However, among other areas of dynamic covalent chemistry, imine exchange reactions have numerous advantages. 39,40 They can occur rapidly and without any significant side reactions, there are a wide variety of structures available, simple synthetic approaches, and they have a diverse range of applications. 41,42 Several reports utilizing these chemistries exist, however, they are used mainly to develop materials for extrudable hydrogels, 43 epoxy resins, 44,45 elastomers, 46 or seldomly, methacrylate resins, which have not been 3D printed. Xu et al, 47 reported two biobased, thermally reprocessable, and chemically recyclable imine vitrimers based on methacrylate resins. The use of methacrylates (rather than acrylates) with amine groups can help to avoid aza-Michael side reactions. 48 However, methacrylate containing resins can have reduced printability due to slower photopolymerization kinetics 29 or high viscosities which can compromise their compatibility with photo-based 3D printers. 47 Here, we report a series of five obased, self-healable, and reprocessable resinformulations that consist of an optimized mixture of vanillin methacrylate-functionalized Jeffamines®, and vanillin acrylate. The addition of vanillin acrylate allows for a reduction of the viscosity while providing sufficiently rapid photopolymerization kinetics to allow it to be successfully printed with a DLP 3D printerwithout compromising the amount of bio-based content in the photoresin formulation. Additionally, the presence of imine moieties allow for the self-healing and reprocessing capabilities. ## Materials All chemicals were used as received unless otherwise noted. Vanillin (99%) was purchased from Alfa Aesar. Triethylamine (TEA), sodium chloride (NaCl) sodium bicarbonate (NaHCO3), hydrochloric acid (HCl), sodium hydroxide (NaOH), dichloromethane (DCM), and tetrahydrofuran (THF) were purchased from Fisher Scientific. Acryloyl chloride (≥97% stabilized with phenothiazine) was purchased from Sigma-Aldrich. 4-dimethylaminopyridine (99%) (DMAP) was purchased from Acros Organics. Methacrylic anhydride (94% stabilized with ca. 0.2% 2,4-dimethyl-6-tertbutylphenol) was purchased from Alfa Aesar. Jeffamine® T-403 (J-T-403), Jeffamine® D-400 (J-D-400), Jeffamine® ED-900 (J-ED-900), Jeffamine® D-2000 (J-D-2000), and Jeffamine® T-3000 (J-T-3000) were obtained as samples from Huntsman. Diphenyl(2,4,6-trimethylbenzoyl) phosphine oxide (TPO) was purchased from TCI. ## Synthetic Procedures Vanillin Acrylate (VA). VA was synthesized following previously reported procedures. 49,50 Vanillin (40g, 0.26 mol) was added to a round bottom flask and dissolved in anhydrous DCM, followed by addition of TEA (51.4g, 0.51 mol). The mixture was purged with N2 and cooled to 0-5°C in an ice bath. Acryloyl chloride (16.9g, 0.19mol) was added dropwise over a period of one hour. The reaction contents were stirred for 24 hours, and vacuum filtered to remove precipitates. The filtrate was then washed with water, NaHCO3 (sat. aq.), HCl (0.1M, aq.), brine, and finally dried over Na2SO4. The product was purified using column chromatography (6:4 hexane / ethyl acetate), to obtain the product as a clear and colorless liquid (32.8g, 60%). ( 1 H NMR Figure S1) Vanillin Methacrylate (VMA) VMA was synthesized following previously reported procedures. 47 Vanillin (20g, 0.13 mol) was added to a round bottom flask, followed by addition of DMAP (0.12g, 0.98 mmol), and methacrylic anhydride (22g, 0.14mol). The mixture was refluxed at 60°C for 24 h, then cooled to room temperature. The reaction mixture was washed with water, NaHCO3 (sat. aq.), NaOH (0.5M, aq.), NaOH (1M, aq.), brine, and dried over Na2SO4. A white crystalline product was obtained and used without further purification. (23.5g, 80%). ( 1 H NMR Figure S2) ## VMA-functionalized Jeffamines® The general procedure for the synthesis of the five Jeffamine crosslinkers was the following: VMA (2 equiv for the diamines and 3 equiv for the triamines) was added to a round bottom flask and dissolved in DCM, followed by addition of 1.2 equiv of the different Jeffamines® (J-T-403, J-D-400, J-ED-900, J-D-2000, and J-T-3000). The solution was stirred for 4h at room temperature. The crude product was washed with NaOH (1M, aq.), NaHCO3 (sat. aq.), brine, and dried over Na2SO4. After evaporating the solvent at reduced pressures, the viscous liquids obtained were used without further purification. ( 1 H NMR Figure S3-S7) ## Resin Formulation and 3D Printing All the resin formulations were prepared using 20 mol% of the Jeffamine® crosslinker and 80 mol% of the diluent VA. These components were mixed with 2 wt.% of TPO photoinitiator and ultrasonicated for 30 min to degas the resin and avoid air bubbles while printing, as well as ensure full dissolution of the photoinitiator. After this, the resin was gently warmed with a heat gun to facilitate its flow out of the container, and poured into the vat of the Photon Zero DLP 3D printer. The printed shapes were ASTM D638 standard specimen type V, the raising speed was set to 3 mm/s and the exposure time set to 60 s. Once printed, the specimens were washed twice with isopropanol, the first to remove excess unreacted resin, and the second one while sonicating for 5 min to remove additional unreacted material. After washing, the specimens were post-cured under a 405 nm lamp for 24 h. ## Fourier Transform Infrared Spectroscopy (FTIR) Infrared spectra were obtained with an attenuated total reflection (ATR) accessory coupled to a Fourier transform infrared (FTIR) spectrometer (Cary 600 Series). All spectra were recorded in the 4000−400 cm −1 range with a resolution of 2 cm −1 , accumulating 32 scans. Thermogravimetric Analysis (TGA). TGA of 5-10 mg samples loaded into an alumina crucible, was conducted from room temperature to 700 °C at 10°C/min, with a flow rate of 100 mL/min N2 atmosphere using a Mettler Toledo SDT. ## Tensile Testing Uniaxial tensile testing to failure of ASTM D638 standard type V specimens was performed using an Instron 5500A testing machine with a 50 N load cell at a rate of 10 mm/min until failure. ## Compression Testing Uniaxial compression testing was performed using an Instron 6800 universal testing machine with a 50 kN load cell for 3D printed and reprocessed cylindrical compression samples (10 mm diameter x 10 mm height). The samples tested consisted of printed, annealed, and reprocessed samples. All printed samples were post-cured for 24h in a 405 nm lamp after printing. Annealed samples were thermally treated in a 140 °C oven for 3.5h to replicate the conditions that the reprocessed samples had undergone. Reprocessed samples were compression molded in a hydraulic hot press for 3.5h at 140°C in a cylindrical mold. All compression tests of these samples were conducted at room temperature (25 °C) using a crosshead rate of 10 mm/min until specimen failure. ## Dynamic Mechanical Analysis (DMA) The dynamic mechanical properties of the vanillin-Jeffamine® thermosets were measured using a Discovery DMA 850 from TA Instruments in the tension mode. The samples were tested from -25 to 200°C, or -100 to 200°C accordingly at a frequency of 1 Hz, a 5°C/min heating rate and oscillation amplitude of 15 µm. ## Self-Healing Experiments For the self-healing experiments, a small piece of the printed specimen of each formulation was gently scratched using a razor blade. Optical microscopy images were obtained for each scratched specimen as the "before healing" pictures, and then were placed in between two glass slides, and held together with two binder clips to apply pressure. The chosen temperature for the healing experiments was 80°C as this is above the Tg of all the formulations. The samples were placed in an oven for 3h, and after this, optical microscopy images were obtained to evaluate the healing of each formulation. These tests were carried out in triplicate to ensure reproducibility. ## Reprocessing To evaluate the reprocessability of the samples, the printed specimens were ground into small pieces through mechanical grinding and then placed into a metallic mold and compressed under 1500 psi at 140°C for 3h using a Carver hydraulic press. Scheme 1. General scheme for the synthesis of VA and the five different Jeffamine® crosslinkers. ## Synthesis and photoreactive resin formulation for 3D printing The synthesis of the resin components is illustrated in Scheme 1. All the resin components were functionalized with acrylate and methacrylate groups which are reactive towards radical photopolymerization. VMA was used to generate the methacrylated crosslinkers with the dynamic imine functionalities as methacrylate groups are less reactive than acrylate groups towards the aza-Michael reaction. Therefore, the Jeffamine® amine groups will predominantly react with the phenolic aldehyde group of VMA to generate the imines. To be compatible with DLP 3D printing technologies, photoresins require low viscosity so that the resin can recoat the vat surface completely before the next printed layer. The five different imine-containing crosslinkers synthesized are all viscous liquids. To decrease this viscosity, and to retain fast enough photopolymerization kinetics for the formulation to be 3D printable, VA was used as a reactive diluent. Different formulations were created to qualitatively evaluate the printability and the best performing formulation was the 20 mol% crosslinker, and 80 mol% VA. Therefore, this composition ratio was employed for all the formulations, which showed good printing accuracy using DLP 3D printing as seen in figure 1. ## Structural characterization FTIR was used to confirm the presence of the imine functionalities (1645 cm −1 ) in the cured thermosets, as well as, to evaluate the conversion of the vinyl groups by the reduction of the signal of the =C-H peak at 945 cm −1 (Figure S8). Additionally, gel fractions by swelling in water and ethanol, were calculated, obtaining percentages above 95% for most of the formulations indicating good incorporation of the components into the network (Figure S9-S10). ## Thermal characterization of the thermosets TGA of the five resin formulations were obtained to evaluate their thermal stability (Figure S11). A 5 wt% decomposition temperature was observed above 150°C for all formulations indicating a max temperature range for evaluating the dynamic behaviors. DMA experiments were performed to elucidate the glass transition temperatures (Tg). Obtaining the Tg helps to determine a temperature above which the polymer networks more easily move around each other to participate in the bond rearrangement reactions. The Tg values obtained from the peak of tan δ were 66, 33, 22, -18 and -26°C for J-T-403-MA-VA, J-D-400-MA-VA, J-ED-900-MA-VA, J-D-2000-MA-VA, and J-T-3000-MA-VA respectively (Figure 2, Table 1). As a general trend we observed that the higher Tg corresponded to the lowest molecular weight tri-and diamine Jeffamines® with a polypropylene glycol backbone (J-T-403 and J-D-400). The Tg of the diamine with the polyethylene glycol backbone and intermediate molecular weight resided in the middle (J-ED-900). The lowest Tgs correspond to the di-and triamine with the highest molecular weights (J-D-2000 and J-T-3000). This can be explained by the higher imine density, as well as, the size of the chains in the crosslinker backbones. The longer the chains in the crosslinker, the lower the impediment for the rearrangement of the atoms that need to perform the exchange, resulting in lower Tgs. Additionally, the shape of the Tan δ can help us elucidate the homogeneity of the polymer networks. As seen for J-D-2000-MA-VA we observe a very broad Tan δ peak which suggests heterogeneous networks. ## Table 1. Dynamic Mechanical Analysis of the five formulations Tg = Glass transition temperature (°C), E'= Storage modulus (MPa), E" = Loss modulus (MPa) ## Evaluation of Mechanical Properties Tensile testing was conducted to evaluate the mechanical properties of all the printed formulations. For each formulation, at least five specimens were tested and representative stress-strain curves of all the formulations are shown in Figure 3, as well as a comparison of the calculated ultimate tensile strength (UTS), strain at break, and Young's modulus. From the results, we observed that the J-T-403-MA-VA thermoset showed the highest UTS and Young's modulus since J-T-403 is a triamine which provides the final product with a highly crosslinked network. When compared with its equivalent diamine (in molecular weight) J-D-400-MA-VA, we observed a reduction of about half of the UTS and Young's modulus, but an increase in the strain at break. This can be explained by the fact that the crosslinking density of the diamine is lower than the triamine. This increases the UTS, but the less crosslinked polymer chains have more freedom to move around each other and relax stress which improves the elasticity. When comparing J-T-403-MA-VA and J-D-400-MA-VA with their equivalents J-T-3000-MA-VA and J-D-2000-MA-VA respectively, we observed a similar trend with the strain at break with the diamine having higher elasticity than the triamine. However, in terms of UTS and Young's modulus, J-T-3000-MA-VA showed the lowest values of all the formulations. This is because the J-T-3000 has a high molecular weight, which provides a very soft material when comparing it with lower molecular weight formulations. In the case of J-ED-900-MA-VA, it gives intermediate UTS and Young's modulus values, but the highest strain at break. This can be explained because the structure of the oligomeric J-ED-900 is composed of PEG units that can easily move around each other, being able to relax stress resulting in an elastic material. ## Self-healing and reprocessing To evaluate the healing behavior, small pieces of each polymer were scratched with a razor blade and then monitored through optical microscopy before and after the thermal treatment at 80°C for 16h. The heat triggered imine exchange reactions, which have been reported to occur readily and without side reactions at temperatures between 50 to 130°C, 51 allow for the rearrangement of the polymer chains to form bonds between the interface of the cut material, and therefore healing the inflicted defects. Two controls without the dynamic crosslinkers were prepared to demonstrate that the healing is caused by the dynamic imine moieties. These controls consisted of VA, cured without crosslinker, and VA cured with 20 mol% of ethylene glycol dimethacrylate as a crosslinker. As seen from the images (Figure 4) the controls showed no healing behavior, however, all five thermoset formulations with the dynamic crosslinkers demonstrated excellent self-healing behaviors with complete disappearance of the inflicted scratches. The only formulation that still showed scarring at the site of the cut was the J-T-403-MA-VA and this can be explained because it has the highest crosslink density. Additional experiments to evaluate the healing of the crosslinkers cured without any diluent were performed. For this, 2 mL samples of the five methacrylated Jeffamines® were UV-cured and the same procedure for the healing experiments mentioned above were performed. The results showed that the crosslinkers demonstrated efficient healing behaviors after 16h of heating at 80°C, with complete disappearance of the inflicted scratch (Figure S13). However, these crosslinker formulations were too viscous on their own and required curing times that were too long for 3DP. The presence of the imine moieties also allowed remolding or reprocessing of the materials triggered by heat. All the polymers were readily reprocessed into cylinders with complete incorporation of the material pieces into the new cylinder shape as seen in Figure 5. To qualitatively analyze the incorporation of the fragments into the new shape, microscopy images were obtained of the control and of the five formulations with the dynamic crosslinkers. As seen in Figure S14, the control still showed the small fragments of the material with an uneven surface, whereas the Jeffamine® formulations showed even incorporation of the pieces, which confirmed their reprocessability. The healing and reprocessing efficiency was calculated from the comparison between the reprocessed samples and the annealed samples. The results shown in Figure 6 and Figure S15, showed that the property with the highest recovery is the strain at break. ## Chemical degradation experiments Chemical degradation experiments were performed by placing a small piece (10mg) of each of the five polymers in a 1M solution of hexylamine in THF for three days to break the imine crosslinks and dissolve the polymer. Additionally, a control sample was placed in pure THF which remained practically unchanged (Figure S16). All the polymer samples in hexylamine dissolved and the solution was evaluated through 1 H NMR. The spectra showed the characteristic peaks of vanillin, as well as a peak at 8 ppm which corresponded to an imine, likely due to the transimination of hexylamine with the imine crosslinks of the polymer network (Figure S17). ## Conclusion Five 3D printable, self-healable, reprocessable, and chemically degradable thermoset polymers were fabricated by functionalization of different Jeffamines® with VMA and formulated with VA to allow for printability. The addition of VA to the formulations helped reduce the viscosity of the resin, as well as, provided sufficient photopolymerization kinetics to be compatible with DLP 3DP. The five thermosets showed varied mechanical properties (Young's modulus 2.05 -332 MPa) which indicates a possibility of numerous applications depending on the mechanical requirements. The dynamic imine moieties imparted the polymers with self-healing and reprocessing capabilities through heat-triggered transimination reactions. Microscopy images showed complete disappearance of the inflicted scratches, as well as, complete incorporation of the ground polymer fragments into the new shape after reprocessing. Future work will include more in-depth thermal characterization to quantitatively analyze the vitrimer behavior of these thermoset polymers. Due to the simple synthetic procedures and the commercial availability of Jeffamines®, this series of resins provides a promising alternative to commonly used 3D printable formulations. These advances are necessary in order to shift to materials with predominantly biobased content and to help drift away from polymers made from non-renewable resources. ## ASSOCIATED CONTENT Supporting Information

chemsum

{"title": "Recyclable, Biobased Photoresins for 3D Printing through Dynamic Imine Exchange", "journal": "ChemRxiv"}

interdependent_dynamic_nitroaldol_and_boronic_ester_reactions_for_complex_dynamers_of_different_topo

## Abstract: Complex dynamic systems displaying interdependency between nitroaldol and boronic ester reactions have been demonstrated. Nitroalkane-1,3-diols, generated by the nitroaldol reaction, were susceptible to ester formation with different boronic acids in aprotic solvents, whereas hydrolysis of the esters occurred in the presence of water. The boronic ester formation led to significant stabilization of the nitroaldol adducts under basic conditions. The use of bifunctional building blocks was furthermore established, allowing for main chain nitroaldolboronate dynamers as well as complex network dynamers with distinct topologies. The shape and rigidity of the resulting dynamers showed an apparent dependency on the configuration of the boronic acids. ## Introduction Dynamic covalent reactions enable a wide range of molecular systems and materials. For example, the reversible nature of these bond types, akin to non-covalent supramolecular interactions, leads to different self-assembly-, self-healing-, and self-replication-processes, as well as stimuli-responsive materials for a variety of applications. [16, Systems based on dynamic covalent bonds can furthermore be applied to chemical oscillation, and lead to complex behavior, switching modes, or out-of-equilibrium regimes. Moreover, the unique chemistries of reversible chemical bonds enable transfer, inhibition, or emergence of systemic properties, typically based on multiple dynamic processes operating in conjunction. In this regard, dynamic covalent polymers (dynamers) are gaining a lot of attention due to their many features, such as self-assembly/self-organization, adaptive/responsive functions, and emergent properties. [1,15, This is perhaps especially the case in biomedical applications and materials with functions inspired by biological systems. For example, dynamers mimicking natural biopolymers can be applied to replicate the complexity of natural systems through the exploitation of bonds that are dynamic under ambient conditions. A prevailing challenge with reversible covalent systems is the establishment of multiple, yet individually controllable, dynamics. Such dynamic reactions, operating either independently or in synchrony, rapidly enhance the complexities of the systems and can lead to increased control and unique emergent properties. [9,14, At the macromolecular level, examples of such systems are rare, however enabling the generation of intricate structures of different topology. By controlling the connectivities of building blocks in multiple dimensions, the constitution and overall topology of the resulting entities can be varied with high precision through self-assembly of the components. This challenge has been addressed in the present study where such complex dynamic systems have been established and applied to dynamer formation. In complement to the dynamic nitroaldol (Henry) reaction, for example applied to dynamer formation, reversible boronic ester formation was chosen for the systems. The high chemoselectivity of boronic acids towards diols makes them attractive for a variety of structures, such as macrocycles and cages. [12,23, In this work, we demonstrate a match between boronates and the nitroalkanol functionalities generated through the nitroaldol reaction. The two processes could operate simultaneously under the same conditions, where the boronic ester formation and the nitroaldol reaction complemented each other. This led to a double dynamic process in which the second dynamic reaction is sequentially dependent on the first, while, at the same time, the properties of the first process are altered by the second. The dynamic boronic ester process could thus change the properties of the underlying nitroaldol reaction, e.g., through increased resistance to base and impedance of the reverse nitroaldol reaction. These features could subsequently be extended to the formation and modification of nitroaldol-based dynamers, allowing the incorporation of the two dynamic bond types into topologically distinct dynamers. ## Results and Discussion The boronate formation was initially studied using phenylboronic acid a and nitroalkanediol 1, the reversible addition product between nitromethane (N1) and pyridine-2-carboxaldehyde (A1; Figure 1). The reactions were followed by 1 H NMR in several solvents, displaying that the resulting boronic ester 1a was favored in all tested aprotic solvents, whereas methanol and water disfavored the formation. The 11 B NMR spectrum of the reaction between compounds 1 and a in CD 3 CN indicated the formation of dioxaborinane 1a together with a smaller amount of a potential oxazaborolidine by-product (Figure S28). Even low degrees of water in mixed solvents resulted in low stabilization of the product and the equilibrium in D 2 O/CD 3 CN 10/90 changed to the side of the starting materials (cf. Supporting Information). Acetonitrile was eventually chosen as the most universal solvent, exhibiting high solubilities of the starting materials and the products, while resulting in good stability of the systems over prolonged time periods. Since water-rich media proved to be unfavorable for the reversible boronate process, conditions under which both reactions could be controlled were next explored. Drawing from previous experiences using mild base in organic solvents, the addition of NEt 3 (80 mM, 3 (8) 2 equiv.) to acetonitrile solutions was thus evaluated for the nitroaldol dynamics, and the compatibility of these conditions with the boronate reaction was assessed. As can be seen in Figure 2, these conditions were suitable for the double dynamic system, however leading to low boronate reaction rates. Equilibration could thus be initiated by the addition of base to a solution of nitroalkanediol 1, favoring fragmentation of the diol into the mono-alcohol while releasing one equivalent of aldehyde (Figure 2A). In contrast, when the boronic ester of the nitroalkanediol (1a), preformed in situ, was applied, only a low degree of retro-nitroaldol products was observed (Figure 2B) within 12 h. Similarly, the addition of phenylboronic acid a to a pre-equilibrated mixture of compound 1 and NEt 3 resulted in low equilibration rates (Figure 2C). ## Figure 2: Multidynamic system arising from two-step nitroaldol reaction and formation of boronic ester (top). 1 H NMR spectra recorded at different times during base stability tests in acetonitrile (bottom): A) initial solution of nitroalkanediol 1 and product mixture at different times after base addition; B) initial solution of boronic ester 1a and product mixture at different times after base addition; C) initial mixture of nitroalkanediol 1 after 12 h incubation with base (1*) and at different times after introduction of boronic acid a. A range of boronic acids was subsequently tested in order to evaluate the potential influence of different functional groups on the boronic ester formation. As can be seen in Figure 3, complete or near-complete conversions were obtained in all cases, regardless of boronic acid structure. Having evaluated the formation of the boronate nitroalkanediol esters, the impact of the boronate reaction on the reactivity and topology of nitroaldol dynamers was addressed. In principle, several different dynamer topologies can be be envisaged in this case, involving the reversible formation and interplay of both the nitroaldol and the boronate species (Figure 4). ## Figure 4: Formation of nitroaldol-boronate co-dynamers of different topology: A) mainchain co-dynamer from ditopic nitroalkanediol 2 and ditopic boronic acids p or m; B) modified (graft-on) co-dynamer 3a from polytopic dynamer 3 and monotopic boronic acid a; C) network nitroaldol-boronate co-dynamers 3p or 3m from polytopic dynamer 3 and ditopic boronic acids p or m. A2 = pyridine-2,6-dicarboxaldehyde, N2 = 2-nitroethanol, p = 1,4phenylenediboronic acid; m = 1,3-phenylenediboronic acid. A range of building blocks of each type was evaluated for the co-dynamer formation studies. Thus, compounds with one, two, or multiple reacting groups were selected, including ditopic species with different geometries. For the nitroaldol products, ditopic bisnitroalkanediol 2, produced from pyridine-2,6-dicarboxaldehyde (A2) and 2-nitroethanol (N2), and polytopic dynamer 3 (poly(A2-N1)), produced from aldehyde A2 and nitromethane (N1), were chosen. These nitroaldol species contain two or multiple 1,3-diol functionalities in a linear arrangement, thereby leading to main chain dynamers. Monotopic boronic acid a, and ditopic species p (1,4-phenylenediboronic acid) and m (1,3phenylenediboronic acid) were furthermore included, where the ditopic entities displayed linear and bent configuration, respectively. These structures were combined in stoichiometric amounts, enabling the formation of hybrid nitroaldol-boronate co-dynamers of different topology and molecular weight (Figure 4). The combinations were evaluated by 1 H DOSY NMR and the molecular weights were estimated through comparison with reference compounds of similar composition and known molecular weights (Figure 5). The estimated average molecular weights of the main chain alternating nitroaldol-boronate co-dynamers 2p (poly(N2-A2-N2-co-p)), and 2m (poly((N2-A2-N2-co-m)), formed from bivalent building blocks of both classes, were of comparable magnitude around 2800 g/mol, similar to the main chain nitroaldol dynamer 3 (poly(A2-N1)). Although the geometry of the chains would be influenced by the arrangement of the functional groups on the boronate benzene ring, this effect did not substantially affect the degree of polymerization. However, co-dynamer 2m resulted in a slightly smaller estimated size (2500 g/mol) compared to the more extended co-dynamer 2p (3100 g/mol). More pronounced differences were observed in the reactions with nitroaldol dynamer 3. When dynamer 3 was allowed to react with phenylboronic acid a, the molecular weight of codynamer 3a (poly(A2-N1-mod-a)) increased from 2800 g/mol to 4600 g/mol, albeit without any significant change in the degree of polymerization (~15). This amounts to a modified (graft-)type topology of the co-dynamer, where the boronic acid units are grafted onto the main chain dynamer and cage the basic nitroaldol entities. Similarly, the reactions between nitroaldol dynamer 3 and diboronic acid building blocks m and p led to substantial increases in molecular weight, in these cases associated with noticeable aggregation as expected from the crosslinking nature of the reactions. The NMR signals broadened significantly and the analysis of the average molecular weight of the aggregates resulted in larger errors. Nevertheless, the diffusion coefficient of co-dynamer 3m (net-poly(A2-N1-v-m), M W 5900 g/mol) decreased more than the coefficient of co-dynamer 3p (net-poly(A2-N1-v-p), M W 4700 g/mol). Since the reactivities of the diboronic acids are comparable, as seen from the results with bis-nitroalkanediol 2, this is most likely an effect of the different crosslinking modes of the two diboronic acids. The more bent geometry of diboronic acid m in comparison to that of boronic acid p could thus result in a higher degree of crosslinking (interchain crosslinking or folding), or from lower mono-saturation of the reactive nitroaldol dynamer sites (graft-type topology). The reactions of dynamer 3 with phenylboronic acids a, p, and m were also evaluated by 11 B NMR and FTIR spectroscopy. Although phenylboronic acid derivative 3a and boronic acid a displayed largely coincidental chemical shifts in the 11 B NMR spectra, significant changes were revealed in the FTIR spectra (Figure S34). The major asymmetric B-O stretch thus shifted from 1333 cm -1 in compound a to 1304 cm -1 in modified polymer 3a, consistent with boronic ester formation. On the other hand, the 11 B NMR results of the diboronic acid-based combinations were more clear. The spectrum of network polymer 3p showed a new peak at 12.5 ppm, corresponding to boronate formation, along with a remaining boronic acid signal at 27.9 ppm (Figures S30-S31). In addition, the FTIR spectrum displayed a shift of the asymmetric B-O stretch from 1335 cm -1 in compound p to 1296 cm -1 in structure 3p (Figure S35). The 11 B NMR and FTIR spectra of combination 3m followed a pattern similar to that of structure 3p (Figures S32-S33, S36). These results underscore the dynamic formation of the boronic esters between the dynamer and the three boronic acids. ## Conclusions In summary, dynamic systems based on the combination of the nitroaldol reaction and boronic ester formation have been demonstrated. The esters were thus formed from the nitroalkanediols generated by the nitroaldol reaction in the presence of different boronic acids. The reversible boronic ester formation proved to be favored in most aprotic solvents, whereas hydrolysis was dominant in water-rich media. Both reactions were mutually on each other; nitroalkanediol formation being a prerequisite for boronic ester establishment, and the esters significantly stabilizing the nitroaldol adducts under basic conditions. The systems were furthermore extended to topologically distinct nitroaldol-boronate-dynamers. In these cases, the formation of the esters was dependent on the configuration of the boronic acids, apparently influencing the shape and rigidity of the resulting dynamers. These results show the potential of these systems to enable enhanced stimuli-responsive systems, operating under the influence of either, or both, reaction type. For example, the construction of logic AND-gates can be envisaged, in which the overall assembly would be activated only if the conditions for both reactions were met, i.e. by requiring boronate exchange and the presence of a base. In principle, this effect could also be applied to controlled delivery/release systems. Furthermore, a range of polymeric entities with different topologies can be produced. The system thus enables the grafting to and labeling/functionalization of nitroaldol dynamers, formation of main-chain dynamers involving both reaction types, crosslinking of nitroaldol dynamer chains, establishment of folded geometries, etc. Selective chain cleavage, functionalization by biomimetic hydrogenbonding substituents and extension to stimuli-responsive systems and nanomaterials for biomedical applications are thus under further investigation.

chemsum

{"title": "INTERDEPENDENT DYNAMIC NITROALDOL AND BORONIC ESTER REACTIONS FOR COMPLEX DYNAMERS OF DIFFERENT TOPOLOGIES", "journal": "ChemRxiv"}

photodriven_water_oxidation_initiated_by_a_surface_bound_chromophore-donor-catalyst_assembly

## Abstract: In photosynthesis, solar energy is used to produce solar fuels in the form of new chemical bonds. A critical step to mimic photosystem II (PS II), a key protein in nature's photosynthesis, for artificial photosynthesis is designing devices for efficient light-driven water oxidation. Here, we describe a single molecular assembly electrode that duplicates the key components of PSII. It consists of a polypyridyl light absorber, chemically linked to an intermediate electron donor, with a molecular-based water oxidation catalyst on a SnO 2 /TiO 2 core/shell electrode. The synthetic device mimics PSII in achieving sustained, light-driven water oxidation catalysis. It highlights the value of the tyrosine-histidine pair in PSII in achieving efficient water oxidation catalysis in artificial photosynthetic devices. ## Introduction A central goal in artifcial photosynthesis is storing solar energy from sunlight in chemical bonds. Fujishima and Honda frst demonstrated that direct band gap excitation of the semiconductor, TiO 2 , led to water photolysis and a pathway for solar energy conversion based on water splitting (2H 2 O / O 2 + 2H 2 ). 6 The use of semiconductor electrodes has continued to evolve with progress made in improving light absorption, charge separation, charge transport, and catalysis rates at semiconductor surfaces. The latter includes the development of dye-sensitized photoelectrosynthesis cells (DSPECs) that integrate separate semiconductor electrodes with molecular complexes for light absorption and catalysis. With this approach, each component in a DSPEC can be investigated separately, fnely tuned to maximize performance, and then integrated with a semiconductor(s) in an appropriate surface architecture. DSPEC cells typically utilize chromophores and catalysts that readily attach to oxide surfaces, have high light absorption and strong oxidizing potentials for driving water oxidation at molecular catalysts. 19,20,26 Although signifcant progress has been made in this area, especially with the development of ultra-fast catalysts by Sun and co-workers, stabilization of DSPEC devices may present the most signifcant current bottleneck in practical applications. In Nature, the photosynthetic reaction center evolved over millions of years with water oxidation occurring in the photosystem II (PSII) protein. This protein is responsible for light-driven water oxidation (2H 2 O / O 2 + 4H + + 4e ) in nature. 1, Mimicking the natural system is an inspiration for chemists but given, its high molecular weight and structural complexity, has presented major challenges. 22, PSII is a complex molecular assembly, but its basic components are a spatially extended, light-absorbing chlorophyll array, a P680 chlorophyll acceptor, a tyrosine mediator, and an oxygen-evolving catalyst (OEC), as illustrated in Scheme 1. The relative simplicity of the molecular photoelectrochemical approach described here is notable for achieving many of the key components of PSII. A chromophore, bound to a semiconductor surface, is excited to create a molecular excited state. The excited state then undergoes electron transfer to a SnO 2 /TiO 2 semiconductor electrode, with an internal core/shell structure that facilitates electron transfer to a photocathode for water reduction. The oxidative equivalents from the chromophore undergo intra-assembly electron transfer to a water oxidation catalyst either directly or via a mediator where water oxidation fnally occurs. Because of their relatively high visible absorptivity, and high stabilities in aqueous solutions, polypyridyl Ru(II) complexes have been used as the light-absorbing chromophores (Chrom) in the preparation of these types of assemblies where they are co-loaded with, or chemically linked to catalysts (Cat) for water oxidation. The reaction sequence for water oxidation is illustrated in eqn (1)- (3). It is based on a chromophore-catalyst assembly formed on a SnO 2 /TiO 2 core/shell electrode. In the reaction sequence for water oxidation, the surface-bound chromophore is excited and undergoes electron transfer to TiO 2 followed by electron transfer to an inner SnO 2 layer driven by the lower conduction band of SnO 2 compared to TiO 2 . Electrons removed from SnO 2 at the back contact produce the photocurrent that is transferred to an external cathode for proton or CO 2 reduction. SnO 2 /TiO 2 jChrom,Cat + hn / SnO 2 /TiO 2 (e )jChrom + ,Cat, excitation and injection (1) SnO 2 /TiO 2 (e )jChrom + ,Cat / SnO 2 (e )/TiO 2 jChrom,Cat + , intra-film electron transfer (2) In PSII, a sequence of multi-step electron transfers controls the kinetics and balances the oxidation-reduction reactions. 49 The four underlying redox reactions leading to water oxidation all occur on the millisecond timescale. In comparing the surface activation cycle in eqn ( 1)-( 3) with PSII, a missing component in many artifcial photosynthesis devices is the addition of a mediator that mimics the tyrosine-histidine redox couple in PSII. 21,54 In the analogous reactions in eqn ( 4)-( 6) for the DSPEC phoroanode an additional redox couple (Donor) added to the DSPEC plays a key role mediating electron transfer between the oxidized chromophore and catalyst which mimicks the role of tyrosine as an electron transfer mediator in PSII. In PSII, tyrosine inhibits back electron transfer from the oxidized catalyst and stabilizes the assembly by storing transient oxidative equivalents near the catalyst. In addition the proton-coupled electron transfer (PCET) reaction at the tyrosine-histidine which influences the oxidizing power of the redox couple, also plays a role in the dynamics of charge separation and alters the hydrogen bonding environment near the active site of the catalyst. 56 Nevertheless, in the model here the primary focus is the role as a one-electron transfer mediator. In flling this role in the molecular model, the redox potential for a mediating couple should fall between the ground-state potential for the chromophore and the redox couple(s) of catalyst associated with the rate-limiting step in water oxidation catalysis. Intervention of the mediator, therefore, may occur in any or all of the four photoactivation steps typically associated with the 4e oxidation of water. It is also desired that the different redox states of the mediator be optically transparent in the visible and have good stability in both redox states in aqueous solutions. Triphenylamine (TPA) was chosen here because it meets many of these desired properties. In mimicking PSII, we describe here a chemical approach based on the reactions in eqn ( 4)- (6). It utilizes a semiconductor-surface assembly that mimics PSII by adding an electron transfer mimic for tyrosine to complete the PSII model. In the fnal electrode, an external $5 nm thick TiO 2 shell was used as an external layer on an internal SnO 2 core on a fluorinedoped tin oxide (FTO) electrode. A derivative of the polypyridyl Ru(II) complex, [Ru(bpy) 3 ] 2+ , with bpy ¼ 2,2 0 -bipyridine, was used as the light absorber with a triphenylamine derivative as the electron transfer donor and mediator. 57,58 The catalyst for water oxidation was a derivative of the Ru(II)-2,2 0 -bipyridine-6,6dicarboxylate based, Ru(bda)(py) 2 (py, pyridine), frst described by Sun and co-workers, and, as mentioned above, notable for their rapid rates of oxygen evolution. The fnal assembly was stabilized by adding a 1-2 nm overlayer of the fluorinated DuPont AF polymer, 4,5-difluoro-2,2-bis(trifluoromethyl)-1,3dioxole, which creates an external hydrophobic environment with the structure shown in Fig. S1. † 64 As shown in Scheme 1, in the fnal electrode assembly, FTOjSnO 2 /TiO 2 j-Ru II P(TPA)(Cat) 2+ jAF, the key elements of PSII are included in a working photoelectrode for water oxidation. In a 0.1 M phosphonate buffer solution at pH 7 in 0.4 M NaClO 4 , with an applied bias of 0.6 V vs. NHE, the electrode produced O 2 with an efficiency of 83% and an IPCE value of 10.9% at its absorption maximum of 460 nm. ## Film characterization Mesoporous flms of nanoITO, TiO 2 , ZrO 2 and SnO 2 /TiO 2 , for spectral and electrochemical measurements, were prepared by known literature procedures. 38,65 In brief, a TiO 2 paste, prepared by using a known literature procedure, was deposited on FTO glass with a sheet resistance of 15 U sq 1 by using a doctor blading method on a layer of Scotch tape. Following a heat treatment, flms of 4 micron and 20 nm TiO 2 nanoparticle flms were produced. 66 Four micron, 20 nm particle ITO flms utilized the same procedure but with different paste compositions, as reported in the literature. 67-69 A nanoSnO 2 paste, and flms with core-shell SnO 2 /TiO 2 (4 micron, 20 nm) structures, were coated with $4.5 nm TiO 2 layers by using atomic layer deposition, Fig. S2 and S3. † Absorption spectra Formation of assemblies on oxide surfaces was monitored by UV-visible measurements in air. Results are shown in Fig. 1 for flms of FTOjTiO 2 j-TPA, FTOjTiO 2 j-Ru II P(Cat) 2+ , FTOjTiO 2 j-Ru II P(TPA)(Cat) 2+ , and the electrode substrate, FTOjTiO 2 . The absorption spectrum for FTOjTiO 2 j-Ru II P 2+ includes the expected metal-to-ligand charge-transfer (MLCT) absorption maximum at 460 nm, Fig. S4. † 39,75 Addition of the catalyst to give the assembly, FTOjTiO 2 j-Ru II P(Cat) 2+ , results in additional features in the spectrum from the catalyst. 38,65 The extent of surface loading, G, was determined by absorption measurements with, G ¼ A/(3 1000), and A the absorbance at the wavelength of interest, 3 is the molar extinction coefficient, and G is the surface coverage in mol cm 2 . Following the surface loading procedures described in the Experimental, surface loading of the molecular was, G ¼ 5 10 8 mol cm 2 , based on measurements at 460 nm of chromophore with 3(460 nm) ¼ 1.60 10 4 M 1 cm 1 and catalyst absorptivity at 460 nm of 0.55 10 4 M 1 cm 1 . 38,39,65 The loading level was comparable to surface loading levels for fully loaded surfaces of TiO 2 j-Ru II P 2+ . 76,77 As expected, addition of the TPA electron transfer donor to the assembly to give, FTOjTiO 2 j-Ru II P(TPA)(Cat) 2+ , resulted in no signifcant change in the visible spectrum but with evidence for the added donor in the UV, Fig. 1. 57,58 Electrochemistry Aqueous solution cyclic voltammograms were obtained for the derivatized electrodes at pH 7.0 in 0.1 M phosphonate buffers, in 0.4 M in NaClO 4 on fully loaded planar FTO glass electrodes using a Ag/AgCl (3 M NaCl) as the reference electrode, Fig. S5-S7. † For the electrode FTOj-Ru II P 2+ , a reversible wave appeared for the Ru(III/II) couple at E 1/2 (Ru III/II ) ¼ 1.35 V vs. NHE at a scan rate of 50 mV s 1 . For FTOj-TPA, the TPA/TPA + c couple appeared at 1.08 V vs. NHE. For the catalyst couples in the assembly, FTOj-Ru II P(Cat) 2+ , voltammograms at pH 7 are pH dependent, as they are for model complex Ru(bda)(py) 2 . Oxidation from Ru(II) to Ru(III) occurs with proton loss at a bound aquo ligand to give Ru III -OH 2+ (Cat 0 ) at E 1/2 ¼ 0.7 V; further oxidation to Ru IV ]OH 2+ (Cat 00 ) then occurs at 0.9 V. 65,78 These oxidations are followed by a pH-dependent oxidation of Ru(IV) to Ru(V) at $1.0 V to form Ru V ¼ O (Cat 000 ). 78 The latter triggers water oxidation giving O 2 Fig. 1 Absorption spectra for FTOjTiO 2 j-Ru II P(Cat) 2+ , before (black), and after addition of TPA to give FTOjTiO 2 j-Ru II P(TPA)(Cat) 2+ . Spectra of the related electrodes, FTOjTiO 2 (gray) and FTOjTiO 2 j-TPA, after TPA deposition (blue) are also shown. The spectra were obtained at room temperature in air. with regeneration of the catalyst. 63,79 Water oxidation occurs through an unstable, peroxo-bridged intermediate which decomposes and releases O 2 . 61,80,81 ## X-ray photoelectron spectroscopy (XPS) To further confrm the characterization of the fnal assembly on metal oxide surfaces, X-ray photoelectron spectroscopy (XPS) was used to investigate interfacial elemental compositions for the surface-based structures. Based on the data shown in Fig. S8, † the Ru/P ratio was 0.55 in FTOjSnO 2 /TiO 2 j-Ru II P(Cat) 2+ and 0.32 in FTOjSnO 2 /TiO 2 j-Ru II P(TPA)(Cat) 2+ . Both were consistent with the proposed compositions of the fnal assemblies. ## Photostability The photo-stabilities of the assemblies, with the added 10-20 Dupont (AF) polymer overlayer, were evaluated by procedures described earlier. Derivatized electrodes were exposed to constant irradiation at 455 nm (fwhm $ 30 nm, 475 mW cm 2 ) in aqueous 0.1 M phosphonate at pH 7 solutions, 0.4 M in NaClO 4 . Absorption spectra (360-800 nm) were obtained every 15 min over 16 h periods of irradiation; results are shown in Fig. S9 and S10. † They demonstrate a signifcant enhancement in surface stability for the assemblies with the added aniline donor. As shown in Fig. 2A, following a 16 h irradiation period, the surface coverage of the chromophore FTOjTiO 2 j-Ru II P 2+ jAF had decreased to nearly zero but the decrease was only 50% for FTOjTiO 2 j-Ru II P(TPA) 2+ jAF (Fig. 2A). Addition of the TPA donor stabilizes the excited state at pH 7. Earlier results on the transient FTOjTiO 2 (e )j-Ru III P 3+ , showed that it was unstable toward long term hydrolysis of the bipyridine ligands on Ru(III) based chromophore. 77,85,86 With the added triphenylamine derivative, excitation and quenching give FTOjTiO 2 (e )j-Ru III P(TPA) 3+ jAF. The latter is followed by transfer of the oxidative equivalent to the triphenylamine derivative to give j-Ru II P 2+ (TPA + c) 3+ jAF, with the latter stabilizing the transient excited state. ## Water oxidation Core-shell SnO 2 /TiO 2 electrodes, with 1-2 nm overlayers of the external polymer flm AF, as described above, were used as photoanodes in photoelectrochemical water splitting cells. The photocurrent response with and without the added TPA electron donor was comparable for FTOjSnO 2 /TiO 2 j-Ru II P(Cat) 2+ jAF and FTOjSnO 2 /TiO 2 j-Ru II P(TPA)(Cat) 2+ jAF. Water oxidation was investigated by using a standard three-electrode As shown by the data in Fig. 2B, a comparison of photocurrents for FTOjSnO 2 /TiO 2 j-Ru II P(TPA)(Cat) 2+ jAF and j-Ru II P(Cat) 2+ jAF, at early times, shows that initial photocurrents were higher for the donor-free electrode but that they decreased by a factor of $2 over a period of minutes. With the donor-containing photoelectrode, the photocurrent increased slightly over the initial stages in the water oxidation cycle and reached a maximum at 0.58 mA cm 2 . From the data in Fig. 2C, comparison of long-term photocurrents with and without the added electron transfer donor, is notable. It points toward an important role for the added electron transfer donor to impart an element of stability to the assembly on the electrode surface. The stabilities of the photoanodes and their ability to produce O 2 for extended periods was explored by using a collector-generator, dual working electrode. For FTOjSnO 2 / TiO 2 j-Ru II P(TPA)(Cat) 2+ jAF, O 2 appeared as a product with a 83% efficiency over an electrolysis period of 3 h, Fig. S11. † After 3 h of continuous illumination, the assembly had a photocurrent density of 0.12 AE 0.02 mA cm 2 , Fig. 2. As a control, the electrode FTOjSnO 2 /TiO 2 j-Ru II P(Cat) 2+ jAF had a photocurrent density of 85 mA cm 2 and an O 2 production efficiency of 75% for 1 hour measurements under the conditions described above. A slight photocurrent density increase was noted at the beginning of the test due to local ionic equilibration. Incident photon-to-current efficiency (IPCE) measurements, as a function of excitation wavelength, for FTOjSnO 2 /TiO 2 j-Ru II P(TPA)(Cat) 2+ jAF, at an applied bias of 0.6 V vs. NHE, are shown in Fig. 2D. The IPCE profles overlap with the MLCT absorption profle for the chromophore, consistent with solar conversion initiated by light absorption by the chromophore. Based on the data in Fig. 2D, the IPCE value for FTOjSnO 2 / TiO 2 j-Ru II P(TPA)(Cat) 2+ jAF was 10.9% at the absorption maximum for the assembly at 460 nm. ## Photo-physics Transient absorption measurements were used to understand the events occurring after MLCT excitation of the assemblies: FTOjSnO 2 /TiO 2 j-Ru II P 2+ j(AF), FTOjSnO 2 /TiO 2 j-Ru II P(TPA) 2+j(AF), and FTOjSnO 2 /TiO 2 j-Ru II P(Cat) 2+ j(AF), and of the complete assembly, FTOjSnO 2 /TiO 2 j-Ru II P(TPA)(Cat) 2+ j(AF). As noted below, in analyzing the data, the majority of microscopic events following -Ru II P 2+ excitation occur on the submicrosecond timescale. Based on previous studies on -Ru II P 2+ , and of assemblies on SnO 2 /TiO 2 core-shell electrodes, at the $4-5 nm thickness of the outer TiO 2 shell used in the core-shell experiments, the excited electron is largely trapped in the initial TiO 2 layer on the sub-microsecond timescale. 74,91,92 Transient excitation of FTOjSnO 2 /TiO 2 j-Ru II P 2+ j(AF) at 400 nm occurs with the instantaneous bleach of the metal-toligand charge transfer absorption for the Ru(II) chromophore at 470 nm to give the excited state, -Ru II P 2+ *. On the inert oxide matrix ZrO 2 , free of surface quenching events, the lifetime of the excited state, was $55 ns, Fig. S12 and Table S1 †. On FTOjSnO 2 / TiO 2 j-Ru II P 2+ j(AF), the excited state undergoes rapid electron injection into the oxide flm, followed by back electron transfer, FTOjSnO 2 /TiO 2 (e )j-Ru III P 2+ j(AF) / FTOjSnO 2 /TiO 2 j-Ru II P 2+ j(AF). As found in earlier studies on related surfaces, the kinetics for both electron injection and back electron transfer were biphasic, see below. 70,93 Following excitation, TiO 2 (e ) has an intra-band absorption in the mid infrared. Observations in this spectral region were carried out by transient visible pump/mid-IR probe experiments on CaF 2 jSnO 2 /TiO 2 j-Ru II P 2+ j(AF), Fig. S13. † Quantitative analysis of the kinetics data, probed at 5 mm, gave a time constant of 124 fs for electron injection and 56 ps for decay due to recombination and trapping. As noted in earlier transient studies in the visible, recombination can be monitored by following the recovery of the ground state bleach. 44,58 Transient studies showed that 30% of the bleach recovery occurred with a 56 ps time constant, consistent with the IR decay data (Fig. -S13C †). The remaining decay component that was observed arises from back electron transfer to the oxidized chromophore, j-Ru III P 3+ , on the nanosecond to microsecond timescales, Fig. 3E and S14. † Analysis of the data, based on a KWW analysis, gave a lifetime of, s ¼ 188 AE 5 ms. Fig. 3 Transient absorption spectra for the assemblies, (A) FTOjSnO 2 / TiO 2 j-Ru II P 2+ j(AF), (B) FTOjSnO 2 /TiO 2 j-Ru II P(TPA) 2+ j(AF), (C) FTOjSnO 2 /TiO 2 j-Ru II P(Cat) 2+ j(AF), and (D) FTOjSnO 2 /TiO 2 j-Ru II P(T-PA)(Cat) 2+ j(AF). Kinetics were evaluated at 470 nm at the ground state bleaches, note (E), and at 680 nm, in (F), for absorption by the TPA radical. The samples were excited at 400 nm with pulse energies of 100-300 mJ cm 2 in air. The donor-containing assembly, FTOjSnO 2 /TiO 2 j-Ru II P(T-PA) 2+ jAF, was investigated to explore the kinetics of electron transfer from the oxidized sensitizer to TPA. Ultrafast excitation of FTOj-Ru II P(TPA) 2+ j(AF) results in electron transfer from the sensitizer to the electrode. Excitation is followed by intraassembly electron transfer to the TPA donor to give FTOjSnO 2 / TiO 2 (e )j-Ru II P(TPA + c) 3+ j(AF). The appearance of TPA + c as an intermediate was shown by the appearance of a transient absorption feature with a maximum at $680 nm, Fig. 3B. 57 The sequence of events following excitation is summarized in eqn ( 7)- (9). As noted below, the fnal back electron transfer in eqn ( 9) is sufficiently slow, that it follows after internal electron transfer equilibration in the core/shell. FTOjSnO 2 /TiO 2 j-Ru II P(TPA) 2+ *j(AF) / FTOjSnO 2 /TiO 2 (e )j-Ru III P(TPA) 3+ j(AF) (7) FTOjSnO 2 /TiO 2 (e )j-Ru III P(TPA) 3+ j(AF) / FTOjSnO 2 /TiO 2 (e )j-Ru II P(TPA + c) 3+ j(AF) (8) FTOjSnO 2 /TiO 2 (e )j-Ru II P(TPA + c) 3+ j(AF) / FTOjSnO 2 (e )/TiO 2 )j-Ru II P(TPA + c) 3+ j(AF) / FTOjSnO 2 /TiO 2 j-Ru II P(TPA) 2+ j(AF) In comparing transient results, the appearance of the ground state bleach in FTOjSnO 2 /TiO 2 j-Ru II P(TPA) 2+ jAF, Fig. 3E, is decreased in magnitude compared to FTOjSnO 2 /TiO 2 j-Ru II P 2+ j(AF). The decrease is due to hole transfer in the anilinecontaining assembly from the oxidized chromophore to the aniline donor. Based on an analysis of the data in Fig. S15, † oxidation of the initial transient, FTOjSnO 2 /TiO 2 (e )j-Ru III P(TPA) 3+ to FTOjSnO 2 /TiO 2 (e )j-Ru II P(TPA + c) 3+ , eqn (8), occurs with a lifetime of $830 ps. Back electron transfer from the electrode to the external aniline cation to give the ground state, eqn (9), occurs following internal electron equilibration of the core-shell with a lifetime of $17 ms, Fig. S15C. † Given the spectral properties of the catalyst, in the catalystcontaining assembly, FTOjSnO 2 /TiO 2 j-Ru II P(Cat) 2+ j(AF), there were no spectral probes for the direct observation of hole transfer from the excited state to the catalyst, FTOj-Ru III P(Cat) 3+ j(AF) / FTOj-Ru II P(Cat 0 ) 3+ j(AF). However, the ground state bleach recoveries in Fig. 3B, C, E and S16 † are more rapid than in FTOjSnO 2 /TiO 2 (e )j-Ru III Pj 3+ (AF). The latter is consistent with hole transfer to the catalyst and the reaction sequence in eqn ( 10)- (12). Analysis of the kinetics in Fig. S16B, † assuming return of the added electron from hole transfer to the catalyst, gave a lifetime of $28 ps. The lifetime for back electron transfer to the oxidized catalyst, following internal equilibration of the core/shell, eqn (12), was $64 ms, Fig. S16A. † FTOjSnO 2 /TiO 2 j-Ru II P*(Cat) 2+ j(AF) / FTOjSnO 2 /TiO 2 (e )j-Ru III P(Cat) 3+ j(AF) (10) FTOjSnO 2 /TiO 2 (e )j-Ru III P(Cat) 3+ j(AF) / FTOjSnO 2 /TiO 2 (e )j-Ru II P(Cat 0 ) 3+ j(AF) (11) FTOjSnO 2 /TiO 2 (e )j-Ru II P(Cat 0 ) 3+ j(AF) / FTOjSnO 2 /TiO 2 j-Ru II P(Cat) 2+ j(AF) In the complete assembly, FTOjSnO 2 /TiO 2 j-Ru II P(TPA)(-Cat) 2+ j(AF), transient data were used to investigate hole transfer between TPA + c and the catalyst, eqn (14), by the excitationreaction sequence in eqn ( 13)- (15). In evaluating the dynamics for intra-assembly electron transfer, FTOjSnO 2 /TiO 2 (e )j-Ru III P(TPA)(Cat) 3+ j / FTOjSnO 2 /TiO2(e )j-Ru II P(TPA + c)(Cat) 3+ j, eqn (14), decay of the TPA + c at 680 nm, Fig. S17 † and 3F, occurs by a more rapid, ground state bleach recovery, Fig. 3E. Based on the data, an estimate for the timescale for hole transfer from the catalyst to TPA of $3.6 ns was obtained by analysis of the TPA radical kinetics. The data are shown in Fig. S17. † FTOjSnO 2 /TiO 2 (e )j-Ru III P(TPA)(Cat) 3+ j(AF) / FTOjSnO 2 /TiO 2 (e)j-Ru II P(TPA + c)(Cat) 3+ j(AF) FTOjSnO 2 /TiO 2 (e )j-Ru II P(TPA + c)(Cat) 3+ j(AF) / FTOjSnO 2 /TiO 2 (e )j-Ru II P(TPA)(Cat 0 ) 3+ j(AF) FTOjSnO 2 /TiO 2 (e )j-Ru II P(TPA)(Cat 0 ) 3+ j(AF) / FTOjSnO 2 /TiO 2 j-Ru II P(TPA)(Cat) 2+ j(AF) (15) The kinetics data summarized here were limited to the frst stage in the overall cycle for water oxidation assuming the reactions in eqn ( 13)- (15). Timescales and rate constants for the individual steps, as observed for the individual assemblies by TA measurements, are summarized in Scheme 2 and Table S3. † The exact role of the mediator, which may play a larger role in the 2 nd , 3 rd or 4 th photoactivation steps of the catalyst, are difficult to ascertain from the current experiments. The initial one-electron oxidation of the catalyst from Ru(II) to Ru(III) is likely not a signifcant contributor to the device performance and can be readily achieved by the chromophore alone in the 1 st photoactivation step based on the current kinetic measurements. The TPA mediator likely plays a greater role in either activating the catalyst or storing oxidative equivalents in later stages of the water oxidation catalytic cycle not probed here. Scheme 2 Redox potential diagram based on kinetic studies of the component assemblies and of the final assembly, FTOjSnO 2 /TiO 2 j-Ru II P(TPA)(Cat) 2+ j(AF) vs. NHE, for the first step in the water oxidation cycle. The range of potentials for the three electrons transfer activation of the catalyst is also shown. ## Discussion A DSPEC was prepared and characterized with an integrated semiconductor-molecular assembly that mimicked PSII's ability to use visible light to drive water oxidation to O 2 . The core of the assembly utilized a derivatized polypyridyl complex of Ru(II) which served as both the light absorber and as a scaffold for the assembly of multifunctional units for water oxidation. Flash photolysis experiments on component assemblies, and on the fnal assembly, jSnO 2 /TiO 2 j-Ru II P(TPA)(Cat) 2+ , gave signifcant insights into the microscopic details that occur following MLCT excitation of the -Ru II P 2+ chromophore. Excitation of the assembly was the frst step in its overall activation toward water oxidation. The results of flash photolysis experiments revealed a high level of electron transfer reversibility following the initial 1e oxidation of the -Ru II P 2+ chromophore with no evidence for decomposition of the assembly after repeated transient cycles. Decomposition of the assembly, over extended cycles, occurs following 3-electron oxidation of the catalyst with decomposition occurring in competition with the evolution of O 2 . 97 An energy level diagram for the frst step in the water oxidation cycle by the fnal assembly is shown in Scheme 2. The diagram includes estimates for the individual kinetic steps based on the results of lifetime measurements on the model complexes as discussed above. It also shows the range of redox potentials required for activation of the assembly to its activated 3e form. Based on the scheme, excitation of the -Ru II P 2+ chromophore is followed by excited state injection into the SnO 2 /TiO 2 core/shell electrode with the core/shell inhibiting back electron transfer to the oxidized chromophore on the surface. 98 Following excitation and injection, electron transfer occurs through the core/shell electrode to the cathode where H 2 is formed. Following the initial excitation step, injection and formation of j-Ru III P(TPA)(Cat) 3+ jAF occurs within the assembly, followed by oxidation of the triphenylamine, gives, j-Ru II P(TPA + c)(-Cat) 3+ j(AF). In the frst stage of the water oxidation cycle, the latter undergoes internal electron transfer and loss of a proton to give the singly oxidized, Ru III -OH 2+ form of the catalyst, Cat 0 , in j-Ru II P(TPA)(Cat 0 ) 2+ j(AF). The latter is the frst intermediate in the overall oxidation of water to O 2 . As noted in the Introduction, tyrosine, and its accompanying base, play important roles in PSII as redox mediators between the chromophore and catalyst. A similar redox mediator role may also be played by the triphenylamine cation in j-Ru II P(TPA + c)(Cat) 3+ j(AF). Based on the known mechanistic chemistry of the catalyst in water oxidation in solution, we assumed the catalyst experienced a parallel process on surface. Once oxidized, it undergoes further oxidation through two additional cycles to reach Ru(V). 63,79 Based on previous literature results, in the overall cycle, oxidation to Ru(V) is followed by coordination expansion and O/O bond formation through a transient peroxide intermediate. The latter undergoes further oxidation and loss of O 2 . 33,59,99 For the catalyst in the assembly, 2e oxidation and proton loss give the intermediate, j-Ru II P(TPA)(Cat 00 ) 2+ j(AF), with the catalyst oxidized to Ru IV (O) 2+ . An additional oxidative equivalent is required to give the active form of the catalyst. In the three-electron oxidized form of the assembly, j-Ru II P(TPA + c)(Cat 00 ) 3+ j(AF), the driving force required to reach the activated three electron oxidized form of the catalyst by TPA radical cation is only $0.1 V. Once reached, the active form proceeds through the series of steps required for O/O bond formation and water oxidation. In the overall reaction sequence, the triphenylamine donor redox couple is a kinetically accessible kinetic intermediate. It, and the role that it plays, also provides insight into the role of the tyrosine-histidine acid-base pair in PSII. Intervention of the latter would explain the value of a separate step in which the redox equivalent for the fnal activation step is stored in the tyrosine-histidine acid-base pair with j-Ru II P(TPA + c)(Cat 00 ) 2+ j(AF) as an analog. In the fnal step in the activation of the catalyst, as for tyrosine in PSII, the assembly is converted into an active form that provides access to a reactive form of the catalyst. ## Conclusions The results described here describe a procedure for the preparation of a surface assembly that mimics the ability of PSII for using visible light for water oxidation to O 2 with an electron transfer mediator. Although relatively simple compared to PSII, the assembly includes all of the key functional elements of PSII, including light absorption, electron transfer activation, catalysis, mediation, and the light-induced formation of O 2 . Notable in the results, as revealed by flash photolysis measurements, was the extensive series of electron transfer steps that occur in the light absorption cycle by the molecular chromophore and the creation and storage of redox equivalents in an attached catalyst for water oxidation. North Carolina Research Triangle Nanotechnology Network, which is supported by the NSF, Grant ECCS-1542015, as part of the National Nanotechnology Coordinated Infrastructure. This material is based upon work supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, Solar Photochemistry Program under Award Number (DE-FG02-07ER-15906) for T. L.

chemsum

{"title": "Photodriven water oxidation initiated by a surface bound chromophore-donor-catalyst assembly", "journal": "Royal Society of Chemistry (RSC)"}

study_on_the_cross-linking_process_of_carboxylated_polyaldehyde_sucrose_as_an_anti-wrinkle_finishing

## Abstract: Sucrose was oxidized in a two-step oxidation reaction catalyzed by 2,2,6,6-tetramethyl-1piperidinyloxy (TEMPO)-laccase and sodium periodate (NaIO 4 ). To generate carboxylated polyaldehyde sucrose (openSu) containing multiple aldehyde and carboxyl groups. The amount of TEMPO and laccase used, as well as the temperature and reaction time were optimized for the oxidation reaction. The successful combination of aldehyde and carboxyl groups of openSu with cellulose was achieved by changing the composition, ratio of the catalyst and the curing conditions. Thereafter, we analyzed the structural characteristics of openSu as well as the aldehyde and carboxyl group content using nuclear magnetic resonance carbon spectroscopy ( 13 C NMR). We found that the optimal finishing conditions were a mixture of magnesium chloride and sodium hypophosphite at a mass concentration ratio of 16 g/L:4 g/L, and curing at 150 °C for 3 min followed by curing at 180 °C for 2 min. There was significant improvement in the anti-wrinkle performance of the openSu-finished fabric, with a wrinkle recovery angle of 258°, whiteness index of 72.1, and a tensile strength rate of more than 65%. We also studied the covalent crosslinking mechanism between openSu and the cotton fabrics.High-quality development is the trend of modern textiles, and it is also an effective way and the urgent need to improve the product grade and added value of the textile industry and reform the supply-side structure. This is due to improved living standards and high-paced working environments that demand high quality fabrics and clothing that are easy to take care of and have superior shape retention capacity 1,2 . Cotton fabrics are examples of textile fibers that are popular for their excellent properties such as softness, comfort, breathability, and good moisture absorption 3,4 . However, cotton products have poor resilience, large shrinkage, and are prone to wrinkling thus requiring frequent ironing. These cause a lot of inconveniences in the modern high-paced environment. Therefore, shape retention is a key attribute of high-quality cotton fabrics. The shape retention capacity of cotton fabrics can be improved by subjecting the fabrics to non-iron finishing processes [5][6][7] , an area of active research in the finishing of fabrics.At present, etherified dimethyl dihydroxy ethylene urea resin (DMDHEU) is commonly used as an antiwrinkle finishing agent. However, the use of DMDHEU during the finishing process releases formaldehyde, which is an environmental pollutant and harmful to human health [8][9][10] . As a result, other anti-wrinkle agents such as polycarboxylic acid, amino silicone oil and epoxy resin have been developed. Unfortunately, the finishing effects of the new agents are not as good as DMDHEU, and they are associated with significant loss in fabric strength 1,3,5 . Among the formaldehyde-free anti-wrinkle finishing agents, polycarboxylic acids have the greatest potential to replace urea-formaldehyde resins. This is because there has been a lot of research on polycarboxylic acids and widespread applications leading to their fast development. BTCA is a polycarboxylic acid that has attracted the most attention. However, high production costs, the pressure on the environment caused by the use of phosphorus-containing catalysts, and the significant loss of fabric strength associated with BTCA has limited its industrial applications 1,11 .Sucrose is a disaccharide that is formed through the dehydration of one molecule of glucose and one molecule of fructose. According to previous studies, the primary hydroxyl group of cellulose can be selectively oxidized by the TEMPO-laccase system to obtain carboxyl cellulose 12,13 , while the adjacent hydroxyl group of the glucose ring can be selectively oxidized by sodium periodate to form dialdehyde derivatives [14][15][16][17][18] . Therefore, we designed cross-linking agents based on the molecular structure and cross-linking mechanisms of polyaldehydes and polycarboxylic acids. The aim was to generate a new formaldehyde-free finishing agent that can be used to develop new anti-wrinkle methods and reaction systems. In this paper, we analyzed the selective oxidation of carboxylated polyaldehyde sucrose (openSu) using TEMPO-NaIO 4 . The primary hydroxyl group of sucrose was first oxidized using the TEMPO-laccase system to form 6,6′-carboxy sucrose (oxySu). NaIO 4 was then used to oxidize oxySu to produce openSu. The carboxyl and aldehyde composition of openSu was determined through potentiometric titration with NaOH, while the structural characteristics of openSu were analyzed using 13 C NMR. We also evaluated the effects of different ratios of the components of the catalytic system, and curing conditions (temperature and time) on the wrinkle recovery angle (WRA), whiteness index (WI) and tensile strength (TS) of the fabrics. The effects of DMDHEU, glutaraldehyde (GA) and BTCA on the WRA, WI, TS and hydrophilicity (wetting time) of treated fabrics were also compared under different curing conditions. ## Results and discussion Effect of reaction conditions on the carboxyl groups content. The primary hydroxyl groups of sucrose were first oxidized to aldehyde groups using the TEMPO system, and then further converted to carboxyl groups 13,14,17 . To explore changes in carboxyl group content in the TEMPO oxidation process, a single-factor optimization experiment was carried out. The reaction conditions are shown in Table 1. The changes in the carboxyl content of oxidized sucrose under different conditions are presented in Fig. 1. Figure 1a shows the relationship between the concentration of TEMPO and the carboxyl group content of openSu. There was a significant increase in the number of carboxyl groups in openSu with increase in the amount of TEMPO. The carboxyl group content of openSu reached 1.2 mmol/g when the amount of TEMPO was 8 mg. This was because in the TEMPO oxidation system, the reduced TEMPO reacted with NaClO to generate nitrogen carbonyl cations (TEMPO + ), which then further reacted with the primary hydroxyl groups of the sucrose 16,19 . According to the "effective collision" theory, increasing the amount of TEMPO increased the number of activated molecules per unit volume which in turn increased the effective number of collisions between tempo and sucrose per unit time. Consequently, more nitrogen carbonyl cations were produced leading to accelerated carboxyl formation during the oxidation process. The carboxyl content of openSu was also affected by the amount of laccase as shown in Fig. 1b. There was a rapid increase in the carboxyl group content of openSu as the amount of laccase increased from 0.04 to 0.10 mg. This was because laccase had an important role in the formation of TEMPO +13,14 . An increase in the amount of laccase accelerated the oxidation reaction, since more aldehyde groups were generated in the same reaction time. The generated aldehyde groups were further oxidized to carboxyl groups, thereby increasing the carboxyl group content of openSu. From this experiment, we found that the maximum carboxyl group content of 1.3 mmol/g was attained when the amount of laccase ranged from 0.10 to 0.12 mg. The effect of the reaction temperature on the carboxyl content of openSu was ill in Fig. 1c. As the reaction temperature increased, the carboxyl content was also increased rapidly. This was because biological enzymes have their optimum reaction temperature, and the optimum temperature of laccase was 30-35 °C. At this temperature range, the carboxyl group content can reach the maximum value of 1.7 mmol/g 14 . Figure 1d shows the effect of reaction time on the carboxyl content of openSu. The figure shows that there was an initial increase in the carboxyl group content as the reaction time increased but after a certain time point the content leveled off. During the first 60 min of the oxidation process, the carboxyl content of openSu increased rapidly and reached the maximum value (1.18 mmol/g). However, there was no significant change in the carboxyl content when the reaction time was extended to 90 min. This was because the number of directly accessible groups decreased, and increased steric hindrance prevented the combination of TEMPO + with the primary hydroxyl groups, resulting in decreased oxidation rate. Aldehyde groups content of openSu. NaIO 4 is an inorganic salt containing multiple aldehyde groups. In our previous study 20,21 , we successfully used NaIO4 to selectively oxidize dialdehyde sucrose (OSu). However, the aldehyde groups that were formed self-polymerized during the reaction process, thus reducing the final aldehyde group content of OSu. Therefore, we changed the molecular structure of OSu, by introducing the carboxyl groups to decrease the reaction between the aldehyde and hydroxyl groups, in this study. Figure 2 shows the aldehyde contents of openSu and OSu. Form the Fig. 2, we see that the aldehyde content of openSu was 50.11 mmol/g according to Eq. ( 2), while the aldehyde content of OSu was only 28.14 mmol/g (Fig. 2). These finding indicated that the introduction of carboxyl groups through the TEMPO-laccase system reduced the polymerization of the aldehyde groups during the oxidation of NaIO4, and greatly increased the aldehyde group content of openSu. characteristic chemical shift peak patterns. We compared the carboxyl and aldehyde group contents among sucrose, oxySu and openSu using 13 C NMR spectrum analysis (Fig. 3), the oxidation mechanism was also shown in Fig. 3. From Fig. 3a,b, the characteristic chemical shift at 177.60 ppm was attributed to the carboxyl group (-COONa), which was selectively oxidized by the TEMPO-laccase system 13,14,22 . At 57.60 ppm, the new peak www.nature.com/scientificreports/ was assigned to the CH 2 -O-C group, indicating that the C-6 aldehyde group had reacted with the primary hydroxyl group during the oxidation process. A comparison between Fig. 3b,c showed a characteristic chemical shift at 169.81 ppm which was attributed to the aldehyde group (-CHO) of openSu 14,23 . This proved that oxySu had been successfully oxidized using the NaIO 4 selective oxidation system, and that the expected carboxylated polyaldehyde sucrose had been obtained. ## Effect of the mass concentration ratio of MgCl 2 and SHP. Since openSu contains carboxyl and aldehyde groups, our aim was to combine all the groups with cellulose to improve the anti-wrinkle performance of the openSu-finished fabric. According to literature 1,3,5,20 , SHP catalyze the combination of carboxyl and hydroxyl groups, while MgCl 2 catalyzes the combination of aldehyde and hydroxyl groups. Therefore, to catalyze the combination of openSu and cellulose, a mixture of SHP and MgCl 2 was selected. We then evaluated the anti-wrinkle property of the finished fabric when different mass concentration ratios of MgCl 2 and SHP were used (the curing conditions was 3 min at 160 °C) (Fig. 4). There was only a slight increase in the WRA (183° of openSu-finished fabric compared to 128 ± 5° of the control fabrics) of the finished fabric when the mass concentration ratio of MgCl 2 to SHP was 0:20 g/L (Fig. 4a). This was because the aldehyde groups of openSu could not react with the hydroxyl groups of cellulose in the absence of MgCl 2 . A mass concentration ratio of 12 g/L:8 g/L significantly increased the WRA of the finished fabric to the maximum value of 253°. Under this condition, the aldehyde and carboxyl groups of openSu were able to fully combined with cellulose, and improve the wrinkle property of the finished fabric. On the other hand, a mass ratio of 16 g/L:4 g/L reduced the WRA of the fabric since there was a lack of sufficient SHP to catalyze the combination of the carboxyl groups of openSu with cellulose. There was also a significant reduction in WRA of the fabric when the mass concentration ratio was 20 g/L:0, since the carboxyl groups of openSu could not combined with cellulose. Figure 4b shows the changes in WI and TS of the finished fabric as the ratio of MgCl 2 to SHP changed. An increase in MgCl 2 decreased the WI and TS of the fabric. This was because the presence of MgCl 2 accelerated the combination of aldehyde groups and cellulose which increased the reaction rate of the hydroxyl groups in the cellulose. In addition, the cellulose underwent dehydration and condensation to produce some colored substances, which reduced the WI of the openSu-treated fabric. During the finishing process of openSu and cellulose, the pH of the finishing solution was low (pH was 3-4), and the cloth surface pH of the finished fabric was also low. This resulted in non-uniform hydrolysis of cellulose during the curing process, which significantly reduced the TS. Effect of the curing temperature and time. Curing temperatures of 180 °C and 150 °C are required in the anti-wrinkle finishing of polycarboxylic acids and dialdehydes, respectively, while the curing time ranges from 1 to 3 min 1,5,11 . Since openSu fabrics have polycarboxylic acids and dialdehydes, curing at 150 °C and 180 °C was required. Therefore, combination of the aldehyde and carboxyl groups of openSu with cellulose happened at 150 °C and 180 °C, respectively. This generated openSu that was fully combined with cellulose. The anti-wrinkle properties of openSu-finished fabrics under different curing conditions are shown in Table 2. www.nature.com/scientificreports/ There was an increase in the WRA but a decrease in the WT and TS of the finished fabric when the curing time was increased at a constant temperature of 150 °C (Table 2). This was because the longer curing time allowed openSu to fully combined with cellulose, while the increased hydrolysis of the fabric at high temperatures decreased TS. After curing at 150 °C for 3 min, the WRA of the fabric was higher. The finished fabric has a similar pattern when it is cured at 180 °C. When it was cured at 180 °C for 2 min, the WRA of the fabric was higher than the un-curied fabric. After openSu finishing, the fabric was cured separately at 150 °C for 3 min and 180 °C for 2 min, the WRA of the fabric was improved. These findings indicate that the MgCl 2 and SHPcatalyzed combination of the aldehyde and carboxyl groups of openSu with cellulose improved the anti-wrinkle properties of the openSu-finished fabric. Calculated the P value of WRA statistical data of unfinished fabric, 150 °C finishing and 180 °C finishing fabric respectively, P 1 was the P value of WRA of unfinished fabric and 150 °C finishing, P 2 was the P value of WRA of unfinished fabric and 180 °C finishing, P 3 was the P value of WRA of 150 °C finishing and 180 °C finishing, and P 1 was 0.00003363, P 2 was 0.000246, P 3 was 0.006497. It was shown that the correlation between openSu finishing fabric and curing at 150 °C was higher than the correlation with curing at 180 °C, which was also related to the high content of aldehyde groups in openSu, which was the main crosslinking group. At the same time, P 3 was more than 0.005, indicating that the correlation between 150 and 180 °C was not obvious. Therefore, the curing time at 150 °C should be longer than 180 °C. Combining the WRA, WI and TS of openSu-finished fabrics with different curing temperatures and time in Table 2, we have concluded that the optimal curing method was two-steps curing methods, that was, curing at 150 °C for 3 min and curing at 180 °C for 2 min. The WRA of openSu-treated fabric cured at 150 °C and 180 °C was 258°, which was almost similar to the WRA of DMDHEU and GA finished fabrics, but lower than that of BTCA-finished fabrics (Fig. 5a). This was an indication that the carboxyl and aldehyde groups of openSu cross-linked with the cotton fabric and imparted the anti-wrinkle properties through the two-step curing process 4,5,11 . From The WI of openSu treated fabric was 72.1, which was higher than those of GA (65.5) and BTCA (61.4) (Fig. 5b). This could be attributed to the dehydration and condensation of the finishing agent or the hydroxyl group of cellulose during the curing process to form unsaturated and colored compounds, which affected the whiteness of the fabric 3,8,19 . The openSu-finished fabric had the shortest wetting time (less than 3 s), which was similar to the reference but much shorter than the wetting time of DMDHEU-, BTCA-and GA-treated fabrics (more than 42 s). This suggests that the higher the anti-wrinkle performance of the finished fabrics, the poorer the hydrophilicity performance as seen with the BTCA, DMDHEU and GA finished fabrics. The short wetting time of openSu could be attributed to its molecular structure which contains reactive groups-aldehyde groups, carboxyl groups and hydrophilic groups-hydroxyl groups. The aldehyde and carboxyl groups were used as cross-linking groups to react with the hydroxyl groups of cellulose. During the reaction process, the hydroxyl groups of the cellulose were consumed, and at the same time, the hydroxyl groups of openSu were introduced into the cellulose. As a result, there was little change in the hydroxyl group content of the finished fabric, and the wetting time of the fabric was shorter than the other finishing agents. Therefore, openSu has obvious advantages in improving the WRA, WI, TS and hydrophilicity properties of finished fabric. Fabric stiffness was one of the fabric hand feeling values, which represented the ability of textile materials to resist bending deformation. We compared the stiffness of the fabrics finished with different finishing agents (Fig. 6). In Fig. 6, the stiffness of openSu finished fabric was 5.65 cm, it is 1 cm longer than the reference, which was similar to that of the DMDHEU finished fabric. The stiffness of BTCA and GA finished fabric was 5.81 cm and 5.76 cm, similarly, the WRA of their finished fabrics was also higher. Combined with the WRA of their finished fabric, it can be inferred that the stiffness of the fabric was related to WRA, and the fabric with higher WRA had greater stiffness. ## Crosslinking mechanism between openSu and cellulose. Based on the results of 13 C NMR and antiwrinkle finishing analysis, we proposed a possible crosslinking mechanism between openSu and cotton cellulose (Fig. 7). The SEM images of fabrics finished with oxySu and openSu are also shown in Fig. 7. Figure 7I shows a scheme representing the combination of openSu and cotton fabric under different curing temperatures. From the scheme we see that there was some cross-linking between openSu and cellulose after curing at 150 °C for 3 min, and full cross-linking was achieved after curing at 180 °C for 2 min. Figure 7II shows a scheme representing the combination between the aldehyde groups and carboxyl groups of openSu and hydroxyl groups of the fiber after the two curing processes. The aldehyde groups reacted with the primary hydroxyl groups in the fiber during the curing process at 150 °C, to confer some anti-wrinkle properties to the fabric 4,8,9,22 . At 180 °C, the carboxyl groups of openSu combined with cellulose 5,11 , to further improve the WRA of the finished fabric. Figure 7III shows the SEM images of the cotton fabric treated using openSu in the two-step curing process. From the images, we see that the surface of the openSu treated fabric has obvious attachments, and that the surface of cellulose was rough. In addition, the number of attachments was higher in the openSu treated fabrics cured at 150 °C and 180 °C than in the openSu fabrics cured at only 150 °C. These findings indicated that there were two-step cross-linking reactions between the aldehyde and carboxyl groups with cellulose when openSu was cured at 150 °C and 180 °C. These reaction are the etherification addition reaction and esterification reaction 6,15,21 . ## Conclusions In this paper, openSu was successfully prepared through selective oxidation by the TEMPO-laccase and NaIO 4 system. Results of 13 C NMR analysis showed that openSu had multiple aldehyde and carboxyl groups, which was a reflection of its high reactivity. This article focused on the effects of the catalytic system, and curing conditions on the openSu-finished fabric. The optimal catalytic system was a mixed catalyst of MgCl 2 and SHP, at a ratio of 16 g/L:4 g/L. The optimal curing conditions were 150 °C for 3 min followed by 180 °C for 2 min. Under www.nature.com/scientificreports/ these conditions, the aldehyde and carboxyl groups of openSu were fully cross-linked with cellulose, and greatly improved the anti-wrinkle performance of the finished fabric. The performance of the openSu-finished fabric was similar to that of DMDHEU and GA treated fabrics, but much lower than that of BTCA treated fabric. The hydrophilicity of the openSu-finished fabrics was higher than the other treated fabrics, but similar to that of the reference sample. Therefore, openSu has significant advantages in the anti-wrinkle finishing of cotton fabrics. ## Experimental section Material and methods. Scoured and bleached cotton woven fabric (133 warp yarns and 100 wefts per 10 cm, weighs 133 g/m 2 ) was purchased from Shandong Lutai Group Co., Ltd. Zibo, China. Sucrose (Su), hydroxyl amine hydrochloride, magnesium chloride (MgCl 2 ), sodium hypophosphite (SHP), NaIO 4 , and GA were purchased from Sinopharm Chemical Reagent Co., Ltd. Shanghai, China. TEMPO, Laccase (0.5U/mg), DMDHEU and BTCA were purchased from J&K Scientific Ltd. Beijing, China. All chemical reagents were of reagent grade. Two-steps selectively oxidation of openSu. (1) TEMPO-mediated oxidation of sucrose was carried out as previously described with slight modifications 13,17,24 . Briefly sucrose (1.026 g) was mixed with acetate buffer solution (200 mL, pH was adjusted to 5.0) in a thermostatic magnetic water bath using a stirrer for 2-3 h. Afterwards, 95% ethanol was used to stop the reaction and then the pH was adjusted to 7.0 using 0.5 mol/L HCl. A membrane separation system was used to remove TEMPO and laccase to obtain the 6, 6′-carboxy sucrose (oxySu) solution. The oxySu solution was freeze-dried to obtain oxySu powder. (2) Selective oxidation of oxySu using NaIO 4 was carried out as previously described with slight modifications 7,15,16,21 . OxySu (6.84 g) was mixed with deionized water (200 mL), stirred and bowled in nitrogen (N 2 ) for 0.5 h to remove O 2 . The temperature of the solution was then adjusted to 10-15 °C, NaIO4 added (12.80 g) and the mixture stirred for 20-24 h. After the reaction, BaCl 2 (14.67 g) was added and the mixture stirred for an additional 30 min. The precipitate was filtered to obtain the openSu solution, which was freeze-dried to obtain the openSu powder. Effect of TEMPO-laccase reaction conditions. The "single factor selection method" was used to determine the effect of oxidation reaction conditions on the TEMPO-laccase oxidation process. Four parameters including the amount of TEMPO and Laccase as well as the temperature and reaction time were studied. The optimal values for each parameter were determined and used in subsequent experiments. ## Carboxyl group content. The carboxylate content of openSu was determined through titration. OpenSu powder (0.25 g) was dissolved in 50 mL deionized water, and the pH of the solution adjusted to 2.5 using 0.1066 mol/L hydrochloric acid standard solution. The solution was then titrated by adding 0.50 mL of 0.05 mol/L sodium hydroxide standard solution dropwise until an equilibrium was achieved. The pH value at this point was recorded. The volume of sodium hydroxide solution consumed during titration was taken as the abscissa, while the pH value was taken as the ordinate to draw a graph to obtain a curve with two break points. The first break point corresponded to the volume of the sodium hydroxide solution V 1 (mL), and the second break point was V 2 (mL). The total carboxyl group content A (mmol/g) was calculated using the following formula: where n NaOH was the concentration of NaOH (mol/L), and W was the weight of openSu. Aldehyde group content. Hydroxylamine hydrochloride-point titration was used to measure the aldehyde group content of openSu. OpenSu (0.100 g) powder was mixed with 25 mL of 0.25 mol/L hydroxylamine hydrochloride solution, followed by the addition of 2 drops of 0.05% methyl orange solution. The mixture was left to stand for 2 h to allow all the components to fully dissolve. The mixture was then titrated using 0.1 mol/L NaOH solution. The titration was stopped when the solution turned from red to yellow (pH about 4), and the volume ΔV of NaOH consumed recorded. The aldehyde content of openSu (mmol C = OH/g openSu) was calculated using Eq. ( 2): where ΔV was the volume of NaOH (mL) consumed, n NaOH was the concentration of NaOH (mol/L), w was the weight of openSu. 13 C NMR analysis. Sucrose (20 mg), oxySu and openSu were dissolved in 550 μL deuterium water (D 2 O), and then nuclear magnetic resonance ( 13 C-NMR) detection carried out using a Bruker AduanceIII 400 MHz nuclear magnetic resonance spectrometer. The internal standard was tetramethylsilane (TMS). Anti-wrinkle finishing of cotton fabric. The finishing solution was prepared by mixing a certain concentration of openSu anti-wrinkle finishing solution with penetrant JFC-2 10 g/L and polyethylene softener 10 g/L. Different mass concentration ratios of the catalyst MgCl 2 and NaH 2 PO 2 were used to give a total mass concentration of 20 g/L. (1)

chemsum

{"title": "Study on the cross-linking process of carboxylated polyaldehyde sucrose as an anti-wrinkle finishing agent for cotton fabric", "journal": "Scientific Reports - Nature"}

a_covalent_organic_cage_compound_acting_as_a_supramolecular_shadow_mask_for_the_regioselective_funct

## Abstract: A trigonal-bipyramidal covalent organic cage compound serves as an efficient host to form stable 1:1-complexes with C60 and C70. Fullerene encapsulation has been comprehensively studied by NMR and UV/Vis spectroscopy, mass spectrometry as well as single-crystal X-ray diffraction.Exohedral functionalization of encapsulated C60 via threefold Prato reaction revealed high selectivity for the symmetry-matched all-trans-3 addition pattern. ## Introduction Fullerenes act as spherically arranged electron-deficient polyolefines. 1 Their unique electronic structures facilitate potential applications in medicine, 2 photovoltaics 3 and organic electronics. 4 Synthetically, addition reactions, e.g., Diels-Alder, 5 Bingel, 6 and Prato 7 reactions or other dipolar cycloadditions, 8 exploit the inherent strain energy of the curved double bonds, thus leading to a multitude of exohedrally functionalized derivatives. 9 Ih symmetrical C60 as the most abundant fullerene gives only one monoadduct. However, the number of potential regioisomers quickly raises to 8 9d and 46 9a,e for bis-and trisaddition, respectively. The precise spatial organization of the individual addends identifies such multiple adducts as highly promising building blocks for 3D molecular architectures, e.g., dendritic systems 10 or extended metal-organic 11 and supramolecular 12 frameworks. However, regioselectivity is generally governed by a combination of statistical and kinetic factors. 13 Thereby, the second addition step usually serves as a bottle neck for the selective formation of higher adducts and individual regioisomers have to be purified, if at all possible, by tedious HPLC chromatography. In some cases, specific addition patterns such as C60Cl6, 14 C60Ph5H, 15 or octahedral hexakisaddition 9c,16 have been obtained in surprisingly high purity under thermodynamic control. Following a tether-directed remote functionalization approach that was introduced by the Diederich group, 17 all seven sterically possible bisadducts for methanofullerenes have been obtained by means of a covalent fixation of the two reactive sites in suitable tether systems. 9a For higher adducts however, only a few tethers for privileged addition patterns are so far available. 18 Elaborate tether synthesis and limitations in postysynthetic modification further hamper this approach. In contrast, any supramolecular prealignment of the reactants or an in-situ activation of specific double bonds is much more tempting. Seminal work by Guldi, Torres and coworkers on Prato reactions for a phthalocyanine aldehyde showed the highly selective formation of cis-1 bisadducts mediated by π-π interactions between the chromophores. 19 More recently, von Delius and coworkers reported the preferred formation of trans-bisadducts for Bingel reactions on C60⊂ cycloparaphenylene. 20 When applying more sophisticated cage structures with a spatially precise orientation of multiple pore windows as hosts, the regioselective formation of higher adducts appears feasible (Fig. 1). In recent years, a large variety of organic and metallosupramolecular cage receptors for fullerenes have been reported. 21 Selective binding was utilized for the separation of fullerene mixtures 22 and electron transfer was studied for dye-attached complexes. 23 Nitschke and coworkers reported on the selective encapsulation of Diels-Alder bisadduct mixtures within the cavity of an Fe II 8L6 cage. However, the regioselectivity for the encapsulation has not been investigated. 24 The Clever group used a bowl-shaped [Pd II 2L3(MeCN)2] 4+ cage as a supramolecular protecting group for the selective monofunctionalization of encapsulated C60. 25 The square-planar arrangement of the four cage windows for a metallosupramolecular fullerene sponge was recently utilized for the exclusive formation of all-e Bingel tetrakisadducts of C60, even under catalytic conditions. 26 This masking strategy provides a significant improvement over the multistep orthogonal transposition approach 27 as the so far only practicable synthetic route for this addition pattern. Despite these initial examples, no cage templates to control the inherently less selective Prato reaction have been reported so far. ## Results and Discussion Recently, we utilized the rectangular scaffold of the tribenzotriquinacenes (TBTQs) 28 for the design and synthesis of a series of covalent organic cage compounds 21f,29 with varying sizes and shapes. 30 As fullerene binding was already shown for some π-extended TBTQ derivatives, 31 we tested the host properties of these cages. After saturation with either C60 or C70 by standing over pristine fullerenes for 24 hours at room temperature, fullerene uptake was analysed by UV/Vis absorption spectroscopy. No additional absorption in the range of 350 to 600 nm, which would indicate fullerene complexation, was observed for CHCl3 solutions of a methoxy-protected TBTQ precursor and a large cubic cage (Fig. S5,S6 †). In case of 1 however, a linear correlation between the host concentration and additional absorption features characteristic for C60 and C70, respectively, clearly indicated uptake of both fullerenes (Fig. S7 †). Further evidence for encapsulation was obtained by MALDI-TOF MS. The isotope patterns for the major signals at m/z = 2323.57 and 2442.67 are in excellent agreement with the respective 1:1 host-guest complexes C60⊂1 and C70⊂1 (Fig. 1a). Remarkably, no signals for the empty cages besides small peaks for pristine C60 or C70 are detected for 1:1 mixtures (Fig. 2a,b). Semiempirical molecular modeling on the PM6 level revealed that the two TBTQ units in 1 are indeed perfectly preorganized for efficient encapsulation of one fullerene molecule. The distance of the two TBTQ closo-atoms in 1, C60⊂1 and C70⊂1 only slightly changes from 1.55 to 1.66 and 1.69 nm (Fig. S1 †). In the case of the larger cubic assemblies, only empty cages and no complexes were detected in MALDI measurements (Fig. S10 †). Here, the increased distance between two TBTQs does not allow any favourable interaction of one fullerene with two TBTQs, thus preventing complexation. For a quantitative analysis, CHCl3 solutions of fullerenes C60 and C70 were titrated with a stock solution of 1 and absorption changes were plotted against cage concentration (see ESI † for details). Global fitting to a 1:1 binding model revealed complexation constants of 6.3±0.4⨯10 5 M −1 and 5.3±0.4⨯10 5 M −1 for C60 and C70, respectively (Fig. 2c,d). Apparently, there is no significant preference for either C60 or C70, which is presumably attributed to opposing effects of better size and shape matching for C60 and larger dispersion interactions for C70. Ultimately, single crystals suitable for X-ray diffraction were obtained during reactivity studies (see below) for the C60 complex. C60⊂1 crystallizes in the monoclinic space group C2/c. ‡ Fig. 3b shows an ORTEP representation of the 1:1 inclusion complex. Whereas the host could be nicely refined, Specific shifting of cage protons in 1 H NMR spectra of 1 in CDCl3 in the presence of 3 equivalents of solid C60 revealed instantaneous complex formation (Fig. 3a). The rather small shifts might be explained by the assumption that spherical C60 is freely rotating within the cavities and the fact, that the para-and diamagnetic ring currents within the five-and six-membered rings of C60 are almost cancelling each other out. Directly after mixing, residual signals that can be attributed to free boronic acid and catechol moieties (Fig. 3a) suggest that encapsulation is achieved via partially opened cages and not via direct slipping. After one hour, cages 1 are again fully closed and quantitative complex formation is observed in case of equimolar mixtures or excess C60. For substoichiometric fullerene concentrations, the appearance of two separate sets of signals for both 1 and C60⊂1 suggests a rather high kinetic stability for the assemblies, at least on the NMR time scale. For C70⊂1, chemical shifts are more prominent, with the significant low-field shift for the aromatic TBTQ protons being particularly apparent (Fig. S2 †). Therefore, we speculate that C70 is complexed in a more rigid fashion with the benzenoid hexagons of the equatorial belt being in close proximity to the axial TBTQ units and the fivefold rotational axes located along one of the cage windows. This model is also in accordance with semiempirical PM6 calculations (Fig. S1 †). However, the exchange of the encapsulated C70 between the three windows should be fast, at least on the NMR time scale, as we did not observe any splitting of the signals for 1 in 1 H-NMR. In the 13 C-NMR spectrum of C60⊂1 (Fig. S3d †), a slight upfield shift from 143.24 to 141.65 ppm was observed for C60 upon complexation (Fig. S3e †). Due to the lower solubility of C70, no 13 C-NMR spectrum could be measured in CDCl3. For C70⊂1 however, five signals at 129.01, 143.72, 146.18, 146.63 and 149.63 ppm appeared, which are in very good agreement with values reported for C70 in C6D6 (130.28, 144.77, 146.82, 147.52, 150.07). 32 Based on these analytical data, we concluded that the 1:1 complex C60⊂1 forms with very high thermodynamic and kinetic stability. A Space filling model from the X-ray structure (Fig. 3c) further reveals that the aromatic scaffold of 1 shields a significant part of the C60 surface apart from the three cage windows arranged in a trigonal planar fashion. Therefore, we speculated on the potential of C60⊂1 acting as a supramolecular template for exohedral C60 functionalization. Assuming that each window can only accommodate one addend due to steric constraints, a limitation to trisaddition was anticipated. Furthermore, trans-3 and neighbouring addition patterns, which approach the 120 ° angle between two individual cage windows, should be favoured (Fig. 6 and S16 †). The Bingel 6 and Prato 7 reactions are the two most common methods for exohedral functionalization of fullerenes. For the twofold Bingel reaction, there is an intrinsic preference for trans-3 and e-addition, 13 whereas cis-addition is significantly disfavoured due to the steric demand of the ester substituents at the cyclopropane rings. For higher adducts, any pre-existing addends in e-position increasingly favour further e-addition, ultimately leading to the highly preferred Th symmetrical hexakisaddition pattern. 33 Building upon this inherent selectivity, Ribas and coworkers reported in the exclusive formation of the all-equatorial tetrakisadduct for Bingel reactions on a complex of C60 within a tetragonal prismatic metallosupramolecular cage. 26 For the Prato reaction however, a much broader distribution of all eight possible bisadducts is usually observed (Fig. S12b †). 34 Only trans-1 and cis-1 isomers are formed in lower yields due to statistical (t1) and steric (c1) factors. Substantial cis-additions for smaller Prato addends generally result in very complex mixtures of higher adducts. In contrast to Bingel reactions, the isolation of specific regioisomers has only been reported in rare cases. 34a Initial investigations for Bingel reactions on C60⊂1 gave promising hints for the selective formation of specific bis-and trisadducts at the early stages. At higher conversion however, an increasing proportion of the undesired background reaction on free C60 was observed due to partial decomposition of the cages in the basic medium. Under Prato conditions though, cage 1 is sufficiently stable to restrict functionalization to the encapsulated C60. As a model reaction, we chose the synthesis of N-methylfulleropyrrolidine multiple adducts C60(CH2NCH3CH2)n from sarcosine (Sar, N-methylglycine) and paraformaldehyde (Fig. 4a). The use of formaldehyde as carbonyl component prevents the occurrence of complex mixtures of stereoisomers at the pyrrolidine rings. The implementation of Me as a very small N-substituent also ensures that any selectivity effect is primarily based on the inherent properties of the template and not on the steric demand of the addends. Ultimately, comprehensive data on HPLC eluation profiles and spectroscopic signatures are available 34a,35 as reference for monoadduct 2 (n=1), all regioisomeric bis-(3, n=2) and relevant trisadducts (4, n=3). To test the templating effect of cage 1, varying equivalents of Sar and excess formaldehyde were added to solutions of C60⊂1 in 1:1 dry CHCl3/toluene (or C60 in dry toluene as control) followed by heating under reflux (Fig. 4a, see ESI † for details) and reaction progress was monitored by MALDI-TOF MS. After specific time intervals, methanolic workup resulted in disassembly of the cages and the obtained fulleropyrrolidine mixtures were analyzed by analytical HPLC. As expected, multiple 1,3-dipolar cycloadditions on free C60 proceed rather unselective. After reaction with two equivalents Sar for eight hours, all eight regioisomeric bisadducts 3 and small amounts of trisadducts 4 were observed (Fig. 4b, black) in similar relative yields as previously reported (Fig. 5a and S12 †). 34a By contrast, reactions on C60⊂1 (Fig. 4b, blue) proceed more slowly but with significantly higher selectivity. Reaction with up to 5 equivalents of Sar over 20 hours indicated low conversion with most of the C60 being unreacted (39% based on HPLC integration) or only converted to monoadduct 2 (34%). For the bisadducts 3 (24%), cis-addition is completely prevented and the symmetry-matched 3-t3 is formed as the main regioisomer (Fig. 5b). Intriguingly, even at low conversion significant amounts of D3-symmetrical 4-t3,t3,t3 (3%) are formed as the only detectable trisadduct. For this addition pattern, all three addends can be perfectly centred within the cage windows, thus minimizing any steric interactions between the addends and the cage walls. By contrast, a control reaction on C60 with five equivalents of Sar for 20 hours resulted in a complex and inseparable mixture of many isomers with up to five pyrrolidine addends (Fig. S14a,d †). Reaction on C60⊂1 with 9 equivalents of Sar for 20 hours resulted in higher conversion and the predominant formation of trisadducts 4 (64%). According to MALDI-TOF MS for the reaction solution (Fig. S15a †), all fullerenes are still encapsulated and almost no higher adducts are formed. Remarkably, only four out of the possible 46 regioisomers for 4 were detected (Fig. 5c). Alongside the symmetry-matched 4-t3,t3,t3 (25%) as the main product, only 4-t4,t4,t2 (7%), 4-t4,t3,t3 (8%) and 4-e,t3,t2 (24% together with 3-e) were observed as side products. This selectivity can be nicely explained by molecular models for the trans-and e-bisadduct complexes 3⊂1, since e + e,t3,t2 t4,t3,t3 t4,t4,t2 any addition to double bonds centred in the third window leads exclusively to these four isomers (Fig. 6). Cage 1 acts as a supramolecular shadow mask that prevents over-functionalization and templates the addition in trans-3 and related positions. Indeed, no cis-adducts and only those four trisadducts, which best match the trigonal planar arrangement of the cage windows, were observed throughout the reaction time (Fig. 5). 0.2% yield. When running the cage-templated reaction (Fig. 4a) on a preparative scale, it proved advantageous to add up to 9 equivalents of Sar in three batches and to extend reaction times to 48 hours (see ESI † for details). For these prolonged reactions, the emerging decomposition of cage 1 initiated some overreaction outside of the cages (see Fig. S18a † for MALDI-TOF MS). Apparently, all mismatched isomers react faster, thus further simplifying the remaining mixture and increasing the relative yield for the matching 3-t3 and 4-t3,t3,t3 (Fig. 5d). After methanolic work-up, all five remaining bis-and trisadducts could be purified by one single automated flash chromatography run without the need for tedious HPLC separation. Finally, the four different trisadducts were isolated as pure regioisomers in a one-pot-reaction starting from pristine C60 with a combined yield of 16%. All products were analyzed by MALDI-TOF MS (Fig. S21 †) and comparison of UV/Vis (Fig. S22 †) and 1 H-NMR spectra (Fig. S19,S20 †) to literature data. 34a When compared to the rather nonselective control reaction, the selectivity for the cagetemplated reaction is completely reversed and the isolated yield for 4-t3,t3,t3 as the main product is increased by a factor of 30 (Fig. S23 †). The small pore windows of 1 enforce the formation of the unfavoured t3,t3,t3-addition pattern that would otherwise only be obtained in trace amounts. On the other hand, the cage walls prevent overreaction to higher adducts. In future work, we plan to optimize reaction conditions for a further increase in isolated amounts of pure trisadducts, modify the cage windows to enhance the selectivity and develop more stable cages to perform more sensitive reactions. ## Conclusions In summary, we have demonstrated the efficient fullerene encapsulation within the pores of trigonal bipyramidal covalent organic cage 1. Due to suitable preorganization of its two TBTQ moieties, strong 1:1 complexes C60⊂1 and C70⊂1 are formed in organic solvents and binding constants of 6.3±0.4⨯10 5 and 5.3±0.4⨯10 5 M −1 , respectively, have been determined by UV/Vis titrations. As evidenced by an X-ray structure for C60⊂1, only three distinct parts of the C60 surface are accessible for exohedral functionalization. In prototypical Prato reactions on C60⊂1, remarkable selectivity for the symmetry-matched trisadduct 4-t3,t3,t3, which is formed only in trace amounts in cage-free reactions, was observed. Cage 1 acts as an efficient supramolecular shadow mask, whose small pore windows force the formation of an otherwise unfavoured trisaddition pattern. These exciting findings are the first example for a supramolecular control of the regioselectivity for the intrinsically nonselective Prato reaction. They will pave the way for novel applications of covalent organic cage compounds as effective templates for spatially precise reactions on large spherical πsystems.

chemsum

{"title": "A Covalent Organic Cage Compound Acting as a Supramolecular Shadow Mask for the Regioselective Functionalization of C60", "journal": "ChemRxiv"}

on_the_nature_of_halide-water_interactions:_insights_from_many-body_representations_and_density_func

## Abstract: Interaction energies of halide-water dimers, X − (H 2 O), and trimers, X − (H 2 O) 2 , with X = F, Cl, Br, and I, are investigated using various many-body models and exchange correlation functionals selected across the hierarchy of density functional theory (DFT) approximations. Analysis of the results obtained with the many-body models demonstrates the need to capture important close-range interactions in the regime of large inter-molecular orbital overlap, such as charge transfer and charge penetration. Failure to reproduce these effects can lead to large deviations relative to reference data calculated at the coupled cluster level of theory. Decompositions of interaction energies 1 carried out with the absolutely localized molecular orbital energy decomposition analysis (ALMO-EDA) method demonstrate that permanent and inductive electrostatic energies are accurately reproduced by all classes of XC functionals (from generalized gradient corrected (GGA) to hybrid and range-separated hybrid functionals), while significant variance is found for charge transfer energies predicted by different XC functionals. Since GGA and hybrid XC functionals predict the most and least attractive charge transfer energies, respectively, the large variance is likely due to the delocalization error. In this scenario, the hybrid XC functionals are then expected to provide the most accurate charge transfer energies. The sum of Pauli repulsion and dispersion energies are the most varied among the XC functionals, but it is found that a correspondence between the interaction energy and the ALMO EDA total frozen energy may be used to determine accurate estimates for these contributions. ## Introduction Molecular dynamics (MD) simulations have become an invaluable tool, potentially allowing for detailed investigation of chemical phenomena that are experimentally inaccessible. Presently, the major computational bottleneck associated with MD simulations is handling the electronic degrees of freedom 10 that determine the potential energy landscape of the molecular system of interest. 11 Electronic wave function methods and, in particular, configuration interaction and its approximations, such as the coupled cluster hierarchy, 11,12 offer "chemical" accuracy but scale very poorly with system size, 13 and are therefore computationally intractable in MD simulations. 14,15 One of the most efficient alternatives to wave function methods is the use of analytical potential energy functions (PEFs), 14,15 commonly knows as force fields (FFs), which effectively integrate out the electronic degrees of freedom. The ideal construction of a FF entails generating a lossless representation of the potential energy surface (PES) associated with the molecular system of interest, which is then used to evolve in time the dynamics of the remaining nuclear degrees of freedom. Unfortunately, the construction of accurate PESs suffers from numerous complications due to the difficulties in describing many-electron systems. For example, eliminating the electronic degrees of freedom requires that FFs be specifically tailored to particular molecular systems to achieve sufficient accuracy, and that each atom be assigned a unique identity that is preserved during the MD simulation. Even when remaining within the above-mentioned restrictions, FF construction is not trivial since there does not exist a unique functional representation that accurately describes the target PES. 14 The most common approach to the development of FFs is to express the total potential energy of a molecular system as a sum of contributions arising from different types of interactions. Intramolecular interactions are often partitioned into terms involving stretching, bending, and dihedral coordinates. Intermolecular interactions tend to be much more difficult to represent, especially in configurations involving large inter-monomer orbital overlap (IMOO) at close range. Intermolecular interactions are usually represented as a sum of individual terms describing different types of physical interactions: permanent electrostatics (ELEC), polarization (POL), charge transfer (CT), London dispersion (DISP), and Pauli repulsion (PAULI). Most common FFs use fixed, atom-centered point charges interacting via Coulomb's law to represent the ELEC term and relatively simple and computationallyefficient representations of the combined PAULI and DISP terms, such as Lennard-Jones and Buckingham potentials, hereafter defined as P+D terms. 14 These approximations are known to become less accurate at short intermolecular distances, where point charges fail to reproduce effects due to the diffusivity of electron clouds (e.g., charge penetration, CP, effects), and the DISP term cannot be defined unambiguously, being a component of the electron correlation energy. With advances in both MD algorithms and computational power, it has become more common for a FF to include a POL term, usually through the use of atom-centered point-dipole polarizabilities 16 or Drude oscillators. 17 An explicit representation of CT is in general omitted, although attempts have been made to developing effective functional forms representing CT effects. The advent of explicit many-body PEFs for water, which are rigorously derived from many-body expansions of the interaction energies 35 calculated at the coupled cluster level of theory with single, double, and iterative triple excitations, CCSD(T), the current "gold standard" for molecular interactions, 36 has boosted the prospects for realistic MD simulations of aqueous systems. 37 An efficient and accurate approach to constructing explicit manybody PEFs is represented by the MB-nrg methodology, which has so far been applied to pure water systems through the MB-pol PEF 38 as well as aqueous systems containing halide and alkali-metal ions. The MB-nrg PEFs integrate classical representations of the ELEC, DISP, and POL terms, which correctly reproduce interactions at large molecular separations, with data-driven two-body (2B) and three-body (3B) terms that smoothly turn on as monomers approach each other. These 2B and 3B terms are derived from large sets of CCSD(T) reference data and expressed in terms of permutationally invariant polynomials (PIPs) 45 that effectively represent all non-classical interactions (e.g., PAULI and CT terms) and correct for inadequacies associated with a purely classical representation of electrostatic interactions (e.g., CP effects) at close range. In this context, the PIPs can thus be viewed as corrections to purely classical representations of molecular interactions, which effectively describe non-classical interactions resulting from IMOO (i.e., PAULI and CP contributions) and electron delocalization (i.e., CT contributions). 14,46 Density functional theory (DFT) is another common approach to representing molecular interactions in MD simulations. Although, unlike FFs, density functionals are not constructed to reproduce individual types of interactions, several energy decomposition approximations (EDAs) have been proposed which attempt to determine individual contributions to interaction energies calculated with a given density functional. 47 Since there is no unique way to decompose interaction energies, particularly in the regime of large IMOO, one must interpret any decomposition as being the consequence of the definitions built into the particular EDA scheme used. In spite of the unavoidable ambiguity in the definitions of individual interaction terms, there are a number of properties that can be satisfied by these definitions which effectively bound their variability, including: 1. The terms in the decomposition exactly sum to the total interaction energy 2. Individual terms are calculated variationally 3. Individual terms have meaningful basis set limits 4. Rigorously known asymptotic limits of individual terms are recovered in a smooth way The 2 nd generation Absolutely Localized Molecular Orbital Energy Decomposition Analysis (ALMO-EDA) method 48 follows this approach, 51 defining terms in order to satisfy a number of "desirable qualities" (including those listed above). In order to provide fundamental insights into the nature of ion-water interactions and assess the ability of different models commonly used in MD simulations to describe these interactions, we report here a systematic analysis of halide-water interactions as predicted by three prototypical implicit and explicit many-body PEFs and various XC functionals selected across the Jacob's ladder of DFT approximations. 52,53 This analysis sheds light on both strengths and weaknesses of different many-body PEFs in representing individual contributions to halide-water interactions. In addition, ALMO-EDA calculations carried out with a hierarchy of XC functionals, from generalized gradient corrected (GGA) to hybrid and range-separated hybrid functionals, provide the opportunity to conduct detailed comparisons between different XC functionals, which allows for assessing the relative accuracy of different XC approximations to the description of halide-water interactions. 2 Theoretical and Computational Methodology 2.1 Electronic structure calculations All X − (H 2 O) dimers and X − (H 2 O) 2 trimers were optimized at the RI-MP2 54,55 level of theory with the aug-cc-pVTZ 56 basis set. Calculations involving Br − and I − used effective core potentials with 10 and 28 electrons, respectively. 57 A threshold of 1.0 × 10 −6 a.u., a step size Scans of each halide ion along the direction of the hydrogen bonded O-H bond of the water molecule were performed by optimizing the water monomer at the RI-MP2/aug-cc-pVTZ level of theory and initially placing the ion 2 away from the oxygen atom, with an O-H-X − angle of 180 o (Fig. 1). Each ion was then displaced along the direction of the hydrogen bonded O-H bond (R X − O ) up to 7 from the oxygen, in 0.1 increments. Reference interaction energies for all X − (H 2 O) dimers and X − (H 2 O) 2 trimers were calculated at the explicitly correlated coupled cluster level of theory, CCSD(T)-F12b, 58,59 in the complete basis set (CBS) limit that was achieved via a two-point extrapolation 60 between the values obtained with the aug-cc-pVTZ and aug-cc-pVQZ basis sets, 56 using effective core potentials with 10 and 28 electrons 57 for calculations of clusters containing Br − and I − , respectively. All CCSD(T)-F12b and RI-MP2 calculations were performed using MOLPRO, version 2015.1. 61 ## The AMOEBA force field The Atomic Multipole Optimized Energetics for Biomolecular Applications (AMOEBA) is an implicit many-body (i.e., polarizable) FF commonly used in MD simulations. 16, The AMOEBA functional form can be expressed as V AM OEBA ELEC is modeled using point-multipoles (through the quadrupole) centered on each atom site, which were calculated via distributed multipole analysis from MP2/aug-cc-pVTZ energies. 62 V AM OEBA P OL is treated explicitly by including self-consistent induced dipoles at atomcentered polarizable sites. In order to avoid overpolarization in the close-range, V AM OEBA P OL adopts a Thole-type electrostatic damping scheme, with exponential-3 damping, 62,72 while none of the terms in V AM OEBA ELEC are damped. The remaining term in Eq. 1, V AM OEBA P +D , represented by a buffered 14-7 potential, 73 describes repulsive interactions at close range, which mainly correspond to Pauli repulsion contributions within common energy decomposition schemes, and London dispersion, Here, |R i − R j | is the distance between atoms i and j, and R ij and ij are the minimum energy distance and the depth of the potential well between atoms i and j, respectively. δ and γ are buffering constants, and serve to damp the divergence of the repulsive R −14 term at short distances. 73 Both δ and γ were originally determined from fits to noble gas data. 73 Values of R ij and ij for homonuclear dimers were fitted to MP2/aug-cc-pVTZ energies, and were refined using experimental data. Values for heteronuclear dimers were then determined using the following combination rules: 62 ## The TTM-nrg PEF Within the TTM-nrg PEF, 41,74 water-water interactions are represented by the MB-pol PEF, while, similarly to the AMOEBA FF, ion-water interactions are described by the following classical terms: The TTM-nrg model employs an extended Thole-type model 72 , describing Pauli repulsion between the ion and the water molecule is expressed as a sum of Born-Mayer functions, where A αβ and b αβ are fitted parameters for each pair of atom types (α and β), and R ij is the distance between atoms i and j. V T T M −nrg DISP is represented by a sum of Tang-Toennies where C 6,αβ is the dispersion coefficient for a given pair of atom types (α and β), and δ αβ is a damping parameter, which is set to be equal to the corresponding parameter, b αβ in the corresponding Born-Mayer function (Eq 6). 74 All C 6,αβ coefficients were calculated using the exchange-hole dipole model (XDM), 79 as implemented in the Postg software. ## The MB-nrg PEF In the MB-nrg PEF, the interactions between individual ions and water molecules are by explicit terms rigorously derived from many-body expansions of the corresponding interaction energies calculated at the CCSD(T)-F12b level of theory. 39,40,42 Briefly, the halide-water MB-nrg PEFs adopt the same functional form and parameterization as the corresponding TTM-nrg PEFs to describe interaction terms associated with permanent electrostatics (i.e., ## ELEC ), polarization (i.e., V M B−nrg ), and dispersion energy ## DISP ). 41,74 However, in the MB-nrg PEFs, these classical terms are supplemented with explicit 2B (V M B−nrg ## 2B ) and 3B (V M B−nrg ## 3B ) terms that are expressed by permutationally invariant polynomials 45 in variables that are functions of the distances between the ion and the six sites of an MB-pol water molecule, and smoothly reduce to zero beyond a cutoff oxygen-ion distance defined as the distance at which the interactions can accurately be represented by the remaining classical terms. The coefficients of both 2B and 3B permutationally invariant polynomials were optimized using Tikhonov regression (also known as ridge regression) 82 to reproduce reference 2B and 3B energies calculated at the CCSD(T)-F12b level of theory 58,59 in the CBS limit. 39,40 The 2B and 3B MB-nrg permutationally invariant polynomials effectively recover quantummechanical effects in ion-water interactions (e.g., charge transfer and penetration) which cannot be accounted for by purely classical expressions as those employed in the AMOEBA FF and TTM-nrg PEFs. Specific details on the development of the MB-nrg PEFs can be found in the original references. 39,40 ## ALMO-EDA calculations For a given base functional, E base [•], ALMO EDA decomposes the interaction energy calculated by that functional, V IN T , into a sum of frozen, polarization, and charge transfer energies, V F RZ , V P OL , and V CT , respectively, 48 Additionally, V F RZ can be further decomposed into a sum of permanent electrostatic, Pauli repulsion, and dispersion energies, V ELEC , V DISP , and V P AU LI , respectively, 48,49 Since a detailed description of the ALMO-EDA method is reported in the original references, 48,49,51,83 we only summarize the main aspects of the method here. The three terms on the right-hand side of Eq. 8 are total-system interaction energy contributions, with each term corresponding to its own total-system electron density, each with different degrees of relaxation (optimization). The least-relaxed (highest energy) density, the frozen density, ρ F RZ (r), results from the antisymmetric product of the (fully-relaxed) isolated monomer occupied orbitals, superimposed into the complex geometry. The frozen density is "frozen" in the sense that monomer densities are unchanged relative to their optimized, isolated densities, apart from the antisymmetrization. The frozen energy, V F RZ is taken to be the difference from subtracting the sum of the isolated monomer total energies, E 1B , from the total energy of ρ F RZ (r) calculated with the base functional. The decomposition of the frozen energy (Eq. 9) is described in references 48 and 49, but we point out here that the calculation of the dispersion energy, V DISP , requires the use of a "dispersion-free" functional, specific to the base functional, which ideally would be identical to the base functional apart from lacking any dispersion interaction. Therefore, in the interest of fair comparisons between XC functionals, only the sum is considered in the analyses presented in Section 3. Additionally, we only consider the ALMO EDA 2 "classical" electrostatic energy 48,49 so that decompositions of XC functionals can be compared with those obtained with many-body models. The polarization energy, V P OL , corresponds to the total-system electron density that has been relaxed (with respect to the base functional) such that each monomer polarizes each other monomer (self-consistently) under the constraint that inter-monomer charge transfer does not occur. In an idealized configuration for which it is possible to uniquely partition the complete set of frozen orbitals between the monomers (i.e. in a configuration with no IMOO) this polarization would correspond to orbital-relaxation resulting from mixing each monomer's (frozen) occupied orbitals with the same monomer's (frozen) virtual orbitals in the field of the surrounding monomers. In order to handle real systems, ALMO EDA 2 determines polarization subspaces for each monomer with the use of fragment electric field response functions, which ensures that polarization in the asymptotic limit is reproduced in a smooth way. 46,48,51 V P OL is then computed by subtracting E 1B and V F RZ from the total energy of the polarized density, ρ P OL (r) calculated with the base functional. The charge transfer energy corresponds to the fully-relaxed total-system electron density, ρ SCF (r), that one would obtain from an unconstrained DFT calculation with the base functional. This electron density can be thought of as resulting from mixing each monomer's (polarized) occupied orbitals with its own virtual orbitals, and with the occupied and virtual orbitals of the surrounding monomers. 46,48,51 The total interaction energy, V IN T , results from subtracting E 1B from the total energy of ρ SCF (r) calculated by the base functional, and the charge transfer energy, V CT , is taken to be the remainder from subtracting V F RZ and V P OL The ALMO-EDA 2 method 48 was used in all energy decompositions that were carried out using the following XC functionals: BLYP, B97M-rV, 89 B3LYP, 90 PBE0, 91 revPBE0, M06-2X, 92 ωB97X, 93 ωB97X-D, 94 and ωB97M-V. 95 Excluding B97M-rV, ωB97X, ωB97X-D, and ωB97M-V, all DFT energies were calculated with and without the D3(0) empirical dispersion correction. 96 All ALMO EDA 2 calculations were carried out with Q-Chem and used the aug-cc-pVQZ basis set for all dimers, 56 except for the I − (H 2 O) dimer for which an effective core potential for 28 electrons was also used for I − . 57 3 Results ## Many-body halide-water models In the AMOEBA model and both TTM-nrg and MB-nrg PEFs, the point-multipoles and atomic polarizabilities, which determine the ELEC and POL terms, respectively, were defined independently of any fitting to reference interaction energies. Additionally, in the case of both TTM-nrg and MB-nrg PEFs, the C 6 coefficients, which determine the DISP term, were obtained from XDM calculations, independently of any fitting to reference interaction energies (see Section 2), while in the AMOEBA model, the dispersion energy is included in the fitting. Consequently, by construction, the remaining energy contributions (V REM ) in all three models should recover the difference between the reference interaction energies (V REF ) and the sum of these pre-determined terms according to where the parenthesis around V DISP indicates that this term is pre-determined only in the TTM-nrg and MB-nrg PEFs. In the AMOEBA model and TTM-nrg PEF, V REM is expressed as sums of buffered 14-7 and Born-Mayer potentials between pairs of atoms i and j, respectively, each containing two fitting parameters: ij and ρ ij for each buffered 14 It follows from Eqs. 8, 9, and 14 that V REM contains V P AU LI and V CT as well as charge penetration effects that cannot be captured by classical electrostatic models based on Tholetype expressions. As mentioned above, in the case of the AMOEBA model, V REM also contains V DISP . Although the AMOEBA buffered 14-7 and the TTM-nrg Born-Mayer potentials were designed to be limited in scope-modeling V P +D , and V P AU LI , respectivelyboth potentials were fitted to the total V REM , resulting in the contamination of both potentials with additional close-range interactions that cannot be described by classical expressions (e.g., charge transfer and penetration). This contamination likely imposes a greater burden on the AMOEBA model because of the additional need to simultaneously represent V DISP with a function of equivalent flexibility. In order to assess the ability of the AMOEBA model and both TTM-nrg and MB-nrg PEFs to describe halide-water interactions, Fig. 2 On the other hand, the AMOEBA model has difficulties in reproducing the CCSD(T)-F12b reference values in all four scans, predicting the onset of repulsive interactions at too large intermolecular separations, and often yielding minima that are too attractive. Additionally, the deviations from the CCSD(T)-F12b curves appear to be larger for smaller ions, with the description of the F − (H 2 O) scan being the most inaccurate. These deficiencies are likely due to the inability of the buffered 14-7 potential to accurately capture P+D, CT, and CP contributions. ## 2B interaction energies The accuracy of AMOEBA, TTM-nrg and MB-nrg is further assessed through comparisons with various XC functionals in Fig. 3, which shows deviations relative to CCSD(T)-F12b/CBS interaction energies calculated for each the four X − (H 2 O) dimers in the corresponding minimum-energy geometries obtained at the RI-MP2/aug-cc-pVTZ(-PP) level of theory. As a reference, also shown in Fig. 3 is the ±1 kcal/mol threshold (dashed line) that is generally used to define "chemical accuracy". Similar comparisons for other halide-water dimers are reported in the Supporting Information. General trends can be identified from the analysis of Fig. 3 Among the three meta-GGA functionals, SCAN systematically overestimates the interaction strengths in all halide-water dimers, with the D3 correction being always relatively small (∼0.1 kcal/mol). An opposite trend is predicted by TPSS, another meta-GGA functional, which underestimates the interaction strengths in all halide-water dimers, although the deviations from the CCSD(T)-F12b reference values are always within 1 kcal/mol. Including the D3 correction has relatively large effects on the accuracy of the TPSS functional, leading to overbound dimers, with the exception of F − (H 2 O). The last meta-GGA functional considered in this study, B97M-rV, performs significantly better, on average, than both SCAN(-D3) and TPSS(-D3), with its accuracy improving from ## which provide guidance in In general, all hybrid GGA functionals considered in this study (B3LYP, PBE0, and revPBE0) exhibit relatively higher accuracy when compared to their parent GGA models, although they follow similar trends. Both B3LYP and revPBE0 underestimate the interaction strength in all halide-water dimers, with the associated deviations from the CCSD(T)-F12b reference values being between ∼1.5 kcal/mol and ∼0.5 kcal/mol, respectively, and decreasing significantly after inclusion of the D3 corrections. On the other side, PBE0 systematically overestimates the interaction strengths in all halide-water dimers and inclusion of the D3 corrections increases the deviations from the CCSD(T)-F12b value (up to −1.5 kcal/mol for the fluoride-water dimer). The meta-GGA and hybrid M06-2X functional is not able to correctly reproduce the interaction energy in the F − (H 2 O) dimer, deviating by more than −2.0 kcal/mol from the CCSD(T)-F12b reference value, although its accuracy noticeably improves for the other halide-water dimers. Interestingly, the D3 dispersion correction has little effect on the performance of M06-2X. This can be traced back to the ability of M06-2X to capture "mediumrange dispersion-like" interactions. 92 The range-separated hybrid, meta-GGA functional, AMOEBA and TTM-nrg predict interaction energies that are within "chemical accuracy" for all dimers except F − (H 2 O), which is significantly underbound in both models. These deviations can be attributed to the inability of purely classical induction schemes adopted by both AMOEBA and TTM-nrg to correctly account for quantum-mechanical effects such as charge transfer and penetration, and Pauli repulsion, which are emphasized in F − (H 2 O) by the covalent-like nature of the interactions at short and medium ranges. Fig. 3 shows that the MB-nrg PEF provides the most accurate description of the interaction energies of all X − (H 2 O) dimers among all XC functionals and many-body models analyzed in this study, with the largest deviation being +0.18 kcal/mol for F − (H 2 O). Considering that both TTMnrg and MB-nrg adopt the same description of all classical contributions to the halide-water interactions, the different performance of these two many-body PEFs must be related to their different ability to represent V REM in Eq. 14. ## 3B interaction energies Deviations of 3B interaction energies calculated by each method relative to CCSD(T)-F12b for 3B interaction energies of halide-water global minimum energy trimers are shown in Fig. 4. Similar comparisons for other halide-water trimers are reported in the Supporting Information. All XC functionals display deviations within 1 kcal/mol for each trimer, such that the least accurate 3B energies are more repulsive than CCSD(T)-F12b. Overall, revPBE0, ωB97X-D, and ωB97M-V provide the most accurate 3B energies, while PBE is consistently the least accurate, displaying deviations greater than 0.5 kcal/mol for each trimer. TTM-nrg deviates from CCSD(T)-F12b by +1.2 kcal/mol for F − (H 2 O) 2 , but remains within 0.25 kcal/mol of our reference 3B energies for all other trimers. Similarly, AMOEBA has trouble with F − (H 2 O) 2 , yielding a 3B energy ∼3.0 kcal/mol more repulsive than that of CCSD(T)-F12b. It should be noted that the 3B energy of both AMOEBA and TTM-nrg are due entirely to 3B POL. Among both XC functionals and many-body models included in Fig. 4, MB-nrg provides the closest agreement with the CCSD(T)-F12b reference values, displaying deviations no greater than 0.1 kcal/mol for each trimer. Considering the MBnrg shares the same V ELEC , V P OL , and V DISP as TTM-nrg, this level of accuracy provides further evidence for the ability of the 3B PIP adopted by MB-nrg to quantitatively reproduce non-classical 3B effects at close range. ## Energy decomposition analysis To provide fundamental insights into the ability of both many-body PEFs and DFT models to describe halide-water interactions, the ALMO-EDA method is used to decompose the interaction energies of all X − (H 2 O) dimers, calclulated by each XC functional, into individual contributions defined in Section 2.5. Additionally, ALMO-EDA analysis within families of XC functionals offers some perspective regarding the construction of a DFT model. For example, PBE and revPBE differ from each other by the value of a single parameter in their exchange functionals (specifically in their exchange enhancement factors), 87 with the revPBE exchange energy being more sensitive to the electron density gradient, and having a lower bound in the large-gradient limit, effectively making revPBE more non-local. 97 Consequently, The average values of the charge transfer energy predicted by each class of XC functionals tend to become less negative from the GGA to the hybrid functionals. This progression can be explained by considering the delocalization error, 98 which is particularly significant for the GGA functionals, which suffer from the self-interaction error, as well as a poor description of electron correlation effects. Importantly, all three GGA functionals predict similar values for the CT term, with a standard deviation of at most 0.2 kcal/mol in the case of the F − (H 2 O) dimer. The largest disagreement among all functionals considered in this study is found for the P+D term, and as with all other contributions to the total interaction energy, the values for this term vary the most for the F − (H 2 O) dimer. Notably, differences between PBE and revPBE almost entirely manifest in their P+D energies, with the revPBE P+D term being at least 1 kcal/mol more repulsive than that of PBE for each minimum energy dimer. BLYP and revPBE differ in their P+D terms by at most 0.4 kcal/mol for all dimers. On average, P+D energies from hybrid functionals are less-repulsive than those from the GGAs, but B3LYP and revPBE0 P+D terms are consistently the most repulsive among the hybrids (note that B3LYP uses BLYP exchange scaled by a factor of 0.72 90 ). It is interesting to note that P+D energies calculated with PBE0 and revPBE0 are closer in magnitude than those calculated with PBE and revPBE by a factor of ∼0.75, reflecting the scaling of the PBE and revPBE exchange functionals in PBE0 and revPBE0, respectively. Since, as mentioned above, the disagreement in the CT terms predicted by different classes of XC functionals can be traced back to the delocalization error, it seems reasonable to assume that, among all classes, the hybrid functionals provide the most accurate value for this contribution. Therefore, since both ELEC and POL contributions are of similar magnitude among all XC functionals, it appears that, within the ALMO-EDA scheme, the P+D term is the most uncertain contribution to the interaction energies of the halidewater dimers. Given the similarity between Eqs. 8 and 9, it seems likely that the frozen energy (P+D + ELEC) should be correlated with the total interaction energy. To test this hypothesis, Fig. 6 shows correlation plots between the total frozen energies predicted by the different functionals for each X − (H 2 O) dimer and the corresponding deviations from the CCSD(T)-F12b total interaction energies. A noticeable correlation indeed exists between these two quantities among all functionals, which becomes even more apparent if only the hybrid functionals are considered. From linear fits to the data shown in Fig. 6, it is thus possible to derive an estimate for the "most accurate" total frozen energies associated with each halide-water dimer. Fig. 5 also shows energy decompositions calculated by the three many-body models studied here. Comparing decompositions of XC functionals with those of many-body models term-by-term demonstrates significant differences between the two approaches. Keeping in mind that low-energy halide-water dimers have small intermolecular separations, these differences are attributed to the failure of simple functional forms adopted by AMOEBA and TTM-nrg in capturing complicated close-range interactions. In contrast, the MB-nrg models accurately reproduce CCSD(T)-F12b interaction energies for all four halide-water systems, demonstrating the ability of PIPs to recover all these close-range effects. In order to further emphasize that the differences between many-body models and DFT energy decomposition terms is due to close-range effects such as CT and CP, energy decompositions are calculated for all many-body models as well as ωB97M-V along the scans described in Sections 2.1 These scans show that AMOEBA and MB-nrg converge to the ωB97M-V value for ELEC at large distances but deviate significantly at close range. These discrepencies in the ELEC contribution, where the values produced by both MB-nrg and AMOEBA are too repulsive relative to ωB97M-V, can be explained by attractive close-range effects, like CP, which are absent from the electrostatic terms adopted by many-body models but are included in the ALMO-EDA ELEC energy. It should be noted that the deviations between MB-nrg and AMOEBA ELEC energies below 2.5-3.0 are due to Thole damping which is absent in the AMOEBA model. Comparing the POL energies along these scans, deviations of many-body models from ωB97M-V again occur at close range. However, it appears that these differences are an artifact of the Thole damping implemented in both MB-nrg and AMOEBA to avoid the polarization catastrophe, which is a consequence of using fixed atomic polarizabilities, an approximation which fails to describe the finite extent of the electron density. 99 DFT models do not suffer from the polarization catastrophe given that no such approximation is made. Although Fig. 5 shows that MB-nrg differs greatly from ωB97M-V in its ELEC and POL energies at close range, in addition to the lack of an explicit CT term, Fig. 2 demonstrates that the MB-nrg PIPs provide QCs for all close-range effects, allowing for a lossless representation of CCSD(T)-F12b interaction energies. Furthermore, Fig. 3 shows that MB-nrg is able to accurately reproduce these effects more consistently than most common XC functionals. Statistics for DFT energy decompositions of X − (H 2 O) 2 clusters can be found in the Supporting Information. In general, the contribution energy values within and between all classes agree for all ion configurations. For each trimer, the total 3B interaction energy is predominantly composed of POL and charge transfer. The 3B energy of AMOEBA and TTM-nrg consists entirely of 3B POL, while the MB-nrg 3B energy consists of 3B PIP QCs as well as TTM-nrg POL, so their energy decompositions follow directly from their interaction energies shown in Fig. 4. ## Conclusions We have analyzed the ability of prototypical implicit and explicit many-body models as well as various XC functionals selected across the hierarchy of DFT approximations to describe the interactions between halide ions and water through the decomposition of interaction energies of halide-water dimers and trimers into their fundamental physical contributions. The energy decompositions of XC functionals have been carried out within the ALMO-EDA scheme and the accuracy of the different models has been assessed through comparisons with CCSD(T)-F12b reference interaction energies. Our analysis indicates that many-body models that do not directly attempt to capture close-range effects in the regime of large inter-monomer orbital overlap are unable to accurately reproduce CCSD(T)-F12b interaction energies. In addition, while XC functionals belonging to different classes of DFT approximations tend to provide similar descriptions of permanent electrostatics and polarization energies, they differ significantly in their ability to represent Pauli repulsion, dispersion, and charge transfer contributions. The large variability in representing charger transfer by different XC functionals is largely due to the delocalization error, which is particularly large in GGA functionals. Finally, we have identified a seemingly-meaningful correlation between the total interaction energy and total frozen energy among all XC functionals, which allows for estimating the magnitude of the energy term representing the sum of Pauli repulsion and London dispersion from energy decompositions carried out with various XC functionals simultaneously. ## Supporting Information Available Cartesian coordinates of low-lying isomers of X − (H 2 O) and X − (H 2 O) 2 complexes, with X = F, Cl, Br, and I, along with additional analyses of 2B and 3B interactions. This material is available free of charge via the Internet at http://pubs.acs.org/.

chemsum

{"title": "On the Nature of Halide-Water Interactions: Insights from Many-Body Representations and Density Functional Theory", "journal": "ChemRxiv"}

a_new_library-search_algorithm_for_mixture_analysis_using_dart-ms

## Abstract: This manuscript introduces a new library-search algorithm for identifying components of a mixture using in-source collision-induced dissociation (is-CID) mass spectra. The two-stage search, titled the Inverted Library-Search Algorithm (ILSA), identifies potential components in a mixture by first searching its low fragmentation mass spectrum for target peaks, assuming these peaks are protonated molecules, and then scoring each target peak with possible library matches using one of two schemes. Utility of the ILSA is demonstrated through several example searches of model mixtures of acetyl fentanyl, benzyl fentanyl, amphetamine and methamphetamine searched against a small library of select compounds and the NIST DART-MS Forensics library. Discussion of the search results and several open areas of research to further extend the method are provided. A prototype implementation of the ILSA is available at https://github.com/asm3-nist/DART-MS-DST. ## Introduction Direct Analysis in Real Time Mass Spectrometry (DART-MS) has been a critical technology in the highthroughput analysis of chemicals under ambient conditions , with several important applications being demonstrated in the recent past (See and references therein). Broadly speaking, DART-MS employs a heated stream of metastable gas molecules (typically helium) to both desorb and, through a cascade of steps, ionize a sample. While rapid, because there is no chromatography, resulting mass spectra are often complex as the spectral signatures for all ionizable compounds in mixture are observed simultaneously. DART is considered a soft ionization technique and therefore mainly intact molecular ions (protonated molecules) are produced. While useful for determining potential molecular formulae, in-source collisioninduced dissociation (is-CID) can also be employed to obtain one or more fragmentation spectra that can be used to further assist in structure elucidation and compound identification. These multiple is-CID spectra represent the sample measured with various levels of fragmentation as controlled by settings specific to the employed mass spectrometer. An example set of is-CID spectra for a single sample collected using a JEOL AccuTOF (Peabody, MA, USA) are provided in Figure 1. The National Institute of Standards and Technology (NIST) has a long history of collecting mass spectra for reference compounds, producing high-quality reference mass spectral databases (or libraries) for electron ionization mass spectrometry (EI-MS) and electrospray ionization tandem mass spectrometry (ESI-MS/MS) . These databases of reference spectra can be utilized in several ways, but possibly of most importance is their use in compound identification. An analyst can query the spectrum of an analyte against reference libraries; the mass spectrum of the analyte is compared sequentially to the reference spectra and the most similar entries in the database are returned to the user as an ordered "hit list" (sometimes written as "hits list" or "hitlist"). This process is commonly referred to as mass spectral library searching. An analyst can then decide whether their analyte is one of the compounds in the returned hit list. Recently, NIST has released an updated DART-MS Forensics mass spectral library . The library contains multiple is-CID mass spectra of over 650 pure compounds of interest to the forensics community (e.g., seized drugs, cutting agents, etc.). The spectra in this library can be viewed, interpreted, and even searched against using standard mass spectral library search tools like NIST MS Search . However, there are two major drawbacks to using standards search tools: (1) spectra are searched individually, requiring an analyst to later manually reconcile the search results for each of the multiple is-CID spectra for the same 1) measured using a JEOL AccuTOF (Peabody, MA, USA) mass spectrometer, with orifice 1 energy settings of +30 V, +60 V and +90 V. Potential targets with a relative intensity greater than 25% (i.e., with peak intensities above dashed horizontal line) in +30 V spectrum are highlighted in red. The peaks corresponding to the target(s) identified in the low fragmentation spectrum are highlighted in orange in the higher fragmentation spectra. This figure is further discussed in Example 1 of Section 3.1. ## Low fragmentation (+30 V) Mid fragmentation (+60 V) High fragmentation (+90 V) analyte, and (2) DART-MS is rarely preceded by chromatography and so the query spectrum is usually of a mixture and not pure compound. The objective of this manuscript is to present an algorithmic approach for determining presumptive identifications of mixture components from a DART-MS experiment using a pure compound library. The method leverages the multilevel information contained in centroided is-CID spectra to first match possible library entries to mixture components targeted in the measurement with least fragmentation, and then assign classification indices, or scores, based on the spectral similarity of all centroided is-CID spectra of the sample. Unlike the more traditional search paradigm, where scores are computed to reflect how well peaks in the query (mixture) mass spectra are explained by matching peaks in the library (pure compound) spectra, the scoring in this algorithm reflects how well peaks in the library spectra are explained by matching peaks in the query. We refer to this new approach as the inverted library-search algorithm (ILSA). Utility of the ILSA is demonstrated through several example searches. In particular, the searches are applied to the challenge of identifying components of controlled mixtures containing up to four illicit drugs present at equal concentrations. The drugs considered in the controlled mixtures are acetyl fentanyl, benzyl fentanyl, amphetamine and methamphetamine. These drugs were selected as representatives of two important scenarios: (1) compounds with identical protonated molecules but distinct fragment ions (acetyl fentanyl and benzyl fentanyl), and (2) compounds with unique protonated molecules but similar fragment ions (amphetamine and methamphetamine). The manuscript is structured as follows. Section 2.1 provides a comprehensive description of the ILSA, specifying assumptions and remarking on known limitations. Description of a prototype software implementation and the experimental design employed to test the algorithm follows in Section 2.2. Discussion of selected demonstrative examples are included as Section 3.1 and a discussion of the full search results as well as a discussion of potential extensions and future work comprises Section 3.2. ## Algorithm Details To most intuitively present the inverted library-search algorithm (ILSA), we describe its underlying assumptions, implementation details, and relevant remarks in a comprehensive manner. An extracted is-CID mass spectrum is a list of coupled mass-to-charge ratios (𝑚 𝑧 ⁄ ) and normalized intensities (%) of the ions measured in a sample. Let 𝑥 ! " denote the mass spectrum of a mixture 𝑀 measured with fragmentation condition 𝐸. The mechanisms that impact fragmentation conditions will vary depending on the particular mass spectrometer employed. In this manuscript, we employ a JEOL AccuTOF (Peabody, MA, USA) mass spectrometer where the level of fragmentation increases with increasing orifice 1 energy. For this instrument, typical orifice 1 energy values are ±30 𝑉, ±60 𝑉, and ±90 𝑉, resulting in low, mid, and high fragmentation spectra. Let 𝑙 # " denote a reference mass spectrum of a pure compound 𝑖 measured with fragmentation condition (orifice 1 energy) 𝐸 as previously described for mixtures. Assumption 1: The component molecules contained in a mixture will each present a protonated molecule in 𝑥 ! $% (or the low fragmentation spectrum considered in the experiment) and the relative intensity of these protonated molecules will be greater than a threshold relative intensity denoted 𝜏 &' . Algorithm Steps: 1) Review the 𝑥 ! $% for potential protonated molecules which we can refer to as targets. i. Identify target 𝑚/𝑧 in 𝑥 ! $% by locating peaks with relative intensities greater than 𝜏 &' (e.g., 25 %). ii. Record the target 𝑚/𝑧 values as set 𝑻. Remark 1: While assumption 1 is likely to hold for most mixture components, there are several reasons why certain components may not appear in a mixture spectrum with relative intensity greater than 𝜏 &' : i) The component does not normally produce a substantial protonated molecule peak. ii) The component is at too low of a concentration to be detected. iii) Competitive ionization within the mixture prohibits sufficient ionization of the component. If a mixture component does not present a protonated molecule with relative intensity greater than 𝜏 &' it will not be identified as a target in algorithm step 1. An alternative strategy that has been explored is assuming identified targets are base peak 𝑚/𝑧 values rather than protonated molecules, as will be discussed in the worked examples of Section 3.1. With this assumption, all mentions of protonated molecules in the algorithm can be replaced with base peak. Note that for many molecules, the protonated molecule is the base peak and so the algorithm will perform identically. We should also note that if a mixture is analyzed in negative ion mode, we would see deprotonated molecules in the spectrum rather than protonated molecules. Assumption 2: Reference mass spectra of the component molecules contained in the mixture are available in a searchable database 1 . The difference between protonated molecule 𝑚 𝑧 ⁄ values of database entries and the targets obtained from 𝑥 ! $% in Algorithm Step 1 is accurate to a known instrument resolution ( ±𝜖 % ). The difference in 𝑚 𝑧 ⁄ values of the peaks in the reference spectra and corresponding peaks in the mixture spectrum will be within a defined 𝑚/𝑧 resolution interval (±𝜖 ( ). In most cases, it may be further assumed that 𝜖 ( is a constant and 𝜖 ( = 𝜖 % . ## Algorithm Steps: 2) For each entry 𝑡 in set 𝑻: i. Search the database for entries with a protonated molecule at 𝑚/𝑧 value within ± 𝜖 % units of target 𝑡. ii. Record the index of these database entries as set 𝒓. Remark 2: If a component of mixture 𝑀 does not have a representative spectrum present in the reference library, it will not be directly identifiable through the ILSA. It is important that reference databases are continually developed and improved upon. Additionally, indirect search procedures, such as the Hybrid Similarity Search employed in both electron ionization and electrospray ionization mass spectral library searching, may help mitigate the limitations of incomplete libraries and may be a fruitful avenue for future work. Assuming that 𝜖 ( is constant and 𝜖 ( = 𝜖 % should be reasonable for small molecules, however, we might expect that mass spectra of larger molecules may suffer from mass drift and thus 𝜖 ( will not be constant. Additionally, if both query and library spectra are not measured using high-resolution mass spectrometry, 𝜖 % and 𝜖 ( must be decoupled. In general, one may improve accuracy for any search condition by decoupling 𝜖 % and 𝜖 ( and optimizing values for particular use-cases. Assumption 3: If a molecule from the reference database is a component of the mixture: (a) peaks from its reference (database) mass spectra are likely to be represented within the mass spectra of the mixture, with likelihood of appearance being a function of the relative intensities of the peak (i.e., the reference spectrum base peak is more likely to appear in the mixture spectrum than a low intensity peak), and (b) the 𝑚/𝑧 difference (drift or bias) between the reference spectra peaks and the matched peaks in the mixture mass spectra will be consistent. Algorithm Steps: ## Algorithm Steps: 3) For each entry 𝑖 in set 𝒓: i. Identify the peaks in 𝑙 # $% with 𝑚/𝑧 less than or equal to the protonated molecule 𝑚/𝑧 + 𝜖 % and with corresponding peaks in 𝑥 ! $% within ±𝜖 ( units and denote this set 𝒂 𝟑𝟎 . 2 ii. Compute a spectral similarity score (see Remark 4) for 𝑙 # $% based on the peaks identified in set 𝒂 𝟑𝟎 . Refer to this value as spectral similarity with orifice energy +30 𝑉 and denote it as 𝛼 $% . Repeat process to compute spectral similarity scores for measurements with other orifice energies (e.g., 𝛼 +% and 𝛼 ,% ). iii. Compute the weighted average spectral similarity (𝛼 =) where 𝐸 = {30, 60, 90} and 𝜔 -. are tunable weights between 0 and 1 such that . iv. For each peak identified in set 𝒂 𝟑𝟎 , determine the absolute difference between each 𝑚/𝑧 value in 𝑙 # $% and its closest 𝑚/𝑧 value in 𝑥 ! $% . Refer to this set of absolute differences as 𝒃 𝟑𝟎 . Compute the standard deviation of entries in set 𝒃 𝟑𝟎 . Refer to this value as the mass bias with orifice energy +30 𝑉 and denote it as 𝛽 $% . Repeat process for 𝛽 +% and 𝛽 ,% . v. Compute the weighted average mass bias (𝛽 G ) (2) 2 An upper limit on 𝑚/𝑧 values is employed to prevent computations including obvious noise and/or dimer peaks. where 𝐸 = {30, 60, 90} and 𝜔 -0 are tunable weights between 0 and 1 such that . vi. Compute the absolute mass difference (𝛾) between the target 𝑚/𝑧 value 𝑡 and the known protonated molecule 𝑚/𝑧 value of 𝑙 # . Note that this is a single value and is only dependent on the target 𝑚/𝑧 value identified in 𝑥 ! $% . vii. Compute a classification index (𝜃 # ) for the library entry 𝑙 # as The value 𝜃 # can then be used for classification decisions. Remark 3: Assumption 3 relies on the is-CID spectra of the mixture and reference pure compounds being taken under similar conditions. Characterizing the range of conditions for which one can assume peaks in is-CID spectra are sufficiently reproducible between mixtures and pure compounds is necessary future work. Remark 4: Algorithm step 3ii introduces the concept of a spectral similarity score denoted 𝛼 " where 𝐸 ∈ {30,60,90}. The objective of a spectral similarity score is to provide a numerical value that meaningfully characterizes the similarity between a pair of mass spectra. There are several measures of spectral similarity readily available for comparing spectra of pure compounds as described in and references therein. In this manuscript we consider two scoring methods that estimate the similarity between a pure compound spectrum (𝑙 # " ) and a mixture spectrum (𝑥 ! " ): (1) fraction of relative intensity of peaks in 𝑙 # " with matching peaks within ±𝜖 ( in 𝑥 ! " , and (2) the cosine similarity between the vectors of relative intensity for peaks in 𝑙 # " with the closest peaks within ±𝜖 ( in 𝑥 ! " . Both scoring methods return real numbered values between 0 and 1 inclusive. We refer to the first score as the "fraction of library peak intensity explained" (FPIE) and the second as a "reverse match factor" (RevMF) based on its historical application in mass spectral library searching. Example calculations of both scores are provided as Appendix A. The classification index calculated in equation ( 3), and its dependent parameters computed using assumed weights in algorithm steps 3ii-3vi, is a simple first approximation of likelihood based on mass spectral similarity scores and 𝑚/𝑧 differences between matching peaks. It is possible that optimizing the weights used within equations ( 1) and ( 2) or developing novel similarity scores that more meaningfully measure spectral similarity between pure compounds and mixtures will produce improved approximations of likelihood and subsequently improve algorithm effectiveness. ## Summary of inverted library-search algorithm: The ILSA can be summarized as a two-stage procedure beginning with a target identification stage (steps 1 and 2) and ending with a scoring stage (step 3). The target identification determines matches from the reference library that may be components of the analyte mixture and scoring computes an index for classification/decision-making about each library match. For target identification, our initial implementation uses a relative intensity threshold and assumes that peaks with relative intensity above the threshold are protonated molecules (or base peaks) of library compounds. For scoring, we compute classification indices that combined one of two spectral similarity scores (FPIE or RevMF), the weighted mass bias between corresponding peaks and the absolute mass difference between the identified target 𝑚/𝑧 and the protonated molecule. A graphical summary of the ILSA is provided as Figure 2. Overall improvements with future implementations of the ILSA will be the result of improved target identification and scoring of library matches. Application of the ILSA for pure compound analysis: While exploring the effectiveness of pure compound analysis using is-CID is outside the scope of this manuscript, we note that the only special requirement for pure compound analysis is that the target relative intensity threshold 𝜏 &' is set to 100% in algorithm step 1 such that only a single peak is targeted for further analysis. All other algorithm steps can proceed as described. Application of the ILSA for "mixed resolution" spectral searching: Though not explicitly stated, the discussion thus far has assumed that both query and library mass spectra are measured with high-resolution instrumentation. A detailed discussion of the efficacy of the ILSA when one or both query and library mass spectra are integer resolution is outside of the scope of this manuscript. However, we will briefly summarize our current implementation for this use-case. If both query and library spectra are integer resolution, we set 𝜖 % = 1, allowing a wider range of target 𝑚/𝑧 values to be identified and set 𝜖 ( = 0 to ensure only peaks with exact 𝑚/𝑧 value matches are used in similarity scoring. If only one of the query or library spectra are integer resolution, the high-resolution spectra is approximated by an integer resolution mass spectrum where 𝑚/𝑧 values are rounded and peaks with the same 𝑚/𝑧 are consolidated by summing intensity. The search can then proceed as though both query and library spectra are integer-resolution. ## Application of the ILSA for other non-chromatographic MS techniques: While not evaluated specifically, we believe the outlined approach is applicable with other ambient ionization sources and/or direct infusion sources given appropriate pure compound libraries and algorithm modifications. ## Software Implementation and Test Data The ILSA for mixture and pure compound analysis is implemented within the new NIST DART-MS Database Search Tool (DST). The tool is written in R and follows a Shiny framework. The source code is available at https://github.com/asm3-nist/DART-MS-DST. The search tool requires reference libraries in R data table (.RDS) and search spectra as tab-delineated text files (.txt). Reference libraries can be generated in .RDS format from mass spectra and metadata using the NIST DART-MS Database Builder Program available at https://github.com/asm3-nist/DART-MS-DBB. The example spectra and libraries discussed in this manuscript are available with the NIST DART-MS DST download. To demonstrate the utility of the ILSA for mixture analysis, we created 11 mixtures composed of 2 to 4 drugs in equal concentration: acetyl fentanyl, benzyl fentanyl, amphetamine and methamphetamine. A summary of the components included in each of the 11 mixtures is provided as Table 1. Each mixture was measured using DART-MS at orifice 1 energies of +30 V, +60 V, and +90 V to generate is-CID query mass spectra. The mixtures were searched against a select library of measurements for pure samples of acetyl fentanyl, benzyl fentanyl, amphetamine and methamphetamine, collected at the same time as the mixtures, to confirm algorithm functionality, as well as the NIST DART-MS Forensics Database to investigate more general utility. Mass spectra were collected using a JEOL AccuTOF mass spectrometer (JEOL) coupled with a DART-SVP ion source (IonSense, Saugus, MA, USA). Relevant DART parameters include operation in positive ionization mode with ultra-pure helium as the source gas and a gas temperature of 400 ºC. The mass spectrometer was also operated in positive ionization mode using an orifice 2 and ring lens voltage of +5 V, and a +600 V rf guide voltage. Parameter switching was used to cycle through +30 V, +60 V, and +90 V orifice 1 voltages at a rate of 0.2 s/cycle. Mass spectra were collected from 𝑚/𝑧 80 to 𝑚/𝑧 800 at 2 scan/s. Each drug mixture was made, individually, using 1 mg/mL solutions of the pure drugs purchased from Cayman Chemical (Ann Arbor, MI, USA). For each mixture a total volume of 200 µL was made, with drugs present at a concentration of 0.25 mg/mL in methanol. Samples were introduced to the DART gas stream via dipped glass microcapillaries. This process was completed three times for each mixture and the composite mass spectra from the replicates were extracted. Query spectra were background subtracted against a spectrum of methanol obtained in the same run and then centroided using Mass Center (JEOL). Acetyl fentanyl ## Example Searches Consider the low fragmentation (+30 V) is-CID mass spectrum shown in Figure 1. This is a measurement of Mixture 3 from Table 1 and includes acetyl fentanyl and methamphetamine. Using a relative intensity threshold 𝜏 &' = 25%, 3 targets are identified (step 1): (1) 𝑚/𝑧 323.214 with relative intensity 100 %, (2) 𝑚/𝑧 150.130 with relative intensity 26.0 %, and (3) 𝑚/𝑧 324.217 with relative intensity 25.8 %. For target 1, both acetyl fentanyl and benzyl fentanyl are compounds in the select library with protonated molecules of 𝑚/𝑧 323.214 ± 0.005 (step 2). For the reference acetyl fentanyl, the spectral similarity scores of the +30 V, +60 V and +90 V spectra, using FPIE, are 0.988, 0.990 and 0.966, respectively, resulting in a weighted average spectral similarity score (Equation 1) of 0.981 with equal weights (0. 3 S ) for all three is-CID mass spectra (steps 3i-iii). The weighted mass bias and absolute mass difference (Algorithm steps 3ivvi) minimally impact the computation of the classification index of 0.979 (Algorithm step 3vii). If we use RevMF for scoring spectral similarity, the scores of the +30 V, +60 V and +90 V spectra are 1.000, 0.998 and 0.987, respectively, resulting in a weighted average spectral similarity score of 0.995 with equal weighting. The weighted mass bias and absolute mass difference minimally impact the final computed classification index of 0.992. For the reference benzyl fentanyl, which produces different fragmentation spectra than acetyl fentanyl, the spectral similarity scores of the +30 V, +60 V and +90 V spectra using FPIE are 0.921, 0.496, and 0.516, respectively, resulting in a weighted average spectral similarity score of 0.644. Using RevMF, the spectral similarity scores are 0.983, 0.786, and 0.207, resulting in a weighted average spectral similarity score of 0.659. The weighted mass bias and absolute mass difference minimally impact the computed classification indices of 0.642 and 0.657 using FPIE and RevMF, respectively. For target 2, only methamphetamine from the Select library matches with protonated molecule within ±0.005 𝑚/𝑧 of 𝑚/𝑧 150.130, and the computed classification indices are 0.895 and 0.969 for scoring using FPIE and RevMF, respectively. Target 3 does not match any compounds in the Select library and was an isotope peak of target 1. Example 1 demonstrates that, by using multiple is-CID mass spectra, we are able to correctly assign the reference acetyl fentanyl a greater classification index than the reference benzyl fentanyl. This is true using either of the considered spectral similarity scoring methods. While there are some use cases where knowing the mixture contains one or the other of the fentanyl analogs will be sufficient, the ILSA gives the analyst a more detailed profile of their mixture. For brevity, we will only discuss classification indices using FPIE with equal weighting between spectra as a measure of spectral similarity for the remaining examples. RevMF scores will be discussed again in the complete results presented in the Section 3.2. The impact of weighting on spectral similarity scores has yet to be explored in detail. Exploring weightings, in addition to other questions about spectral similarity scoring with a more diverse test set of data, would be appropriate future work. As a second example, consider Mixture 8 from Table 1 that includes benzyl fentanyl in addition to acetyl fentanyl and methamphetamine. The is-CID mass spectra are shown in mixture. Target 2, with 𝑚/𝑧 324.215 and relative intensity 26.6 %, does not match any compounds in the select library and is an isotope peak of target 1. If we reduce the relative intensity threshold to 𝜏 &' = 13 %, we identify a third target, with 𝑚/𝑧 150.129 and relative intensity 13.1 %, that matches the protonated molecule of methamphetamine from the Select library with a classification index of 0.949. Example 2 demonstrates one of the operational challenges of using the ILSA (see Remark 1 from Section 2.1). By dropping our relatively intensity threshold we are able to identify all three components of the mixture. More discussion of the target identification stage of the algorithm is to follow the last example. As a third and final example, consider Mixture 11 from Table 1 that includes all four of acetyl fentanyl, benzyl fentanyl, amphetamine and methamphetamine. The is-CID mass spectra are shown in Figure 4. In this example, we set the relative intensity threshold to 𝜏 !1 = 9 %, identifying 5 possible target 𝑚/𝑧 values. As with the other examples, target 1, with 𝑚/𝑧 323.214 and relative intensity 100%, matches the protonated molecules for both acetyl fentanyl and benzyl fentanyl from the select library. In this case the classification indices of 0.934 and 0.930 for acetyl fentanyl and benzyl fentanyl, respectively, suggest (correctly) that both compounds are in the mixture. Target 2, with 𝑚/𝑧 value of 324.217 Da and relative intensity of 26.8 %, does not match any compounds in the Select library and is an isotope peak of target 1. Target 3, with 𝑚/𝑧 91.056 and relative intensity 12.5 %, and target 4, with 𝑚/𝑧 119.008 and relative intensity 11.7 % do not match any of the protonated molecules in the Select library. Target 5, with 𝑚/𝑧 150.130 and relative intensity 9.9 %, matches the protonated molecule of methamphetamine with a classification index of 0.934 suggesting methamphetamine is in the mixture. If we drop the relative intensity threshold for target identifications down to 1%, we will identify, as target 15 with 𝑚/𝑧 136.115 and relative intensity 1.6 %, the protonated molecule of amphetamine with a classification index of 0.969, suggesting amphetamine is in the mixture. Targets 6 to 14 were fragment ions that did not match any protonated molecules in the Select library. Example 3 further illustrates the challenge of identifying targets by protonated molecules using a relative intensity threshold. This challenge is heightened as the protonated molecule for amphetamine is often the second or third most intense peak in its +30 V spectrum (with the other prominent peaks observed 𝑚/𝑧 91.056 and 119.008). One work-around, as initially noted in Remark 1 from Section 2.1, is to search the library for spectra with base peaks that match the target 𝑚/𝑧 rather than protonated molecules. Following such a strategy, target 3 in Example 3, with 𝑚/𝑧 91.056 and relative intensity of 12.5 %, matches the base peak of the +30 V amphetamine spectrum contained in the select library. Accordingly, all four compounds would have been identified within the top 5 targets and with a relative intensity threshold of 11 %. Note that for reference compounds where the protonated molecule is the base peak in the +30 V spectrum, with no mass calibration errors, targeting by protonated molecule or base peak will produce equivalent scores. High fragmentation (+90 V) Figure 4: In-source collision-induced dissociation (is-CID) mass spectra of Mixture 11 (Table 1) measured using a JEOL AccuTOF (Peabody,MA, USA) mass spectrometer, with orifice 1 energy settings of +30 V, +60 V and +90 V. Potential targets with a relative intensity greater than 9% (i.e., with peak intensities above dashed horizontal line) in +30 V spectrum are highlighted in red. The peaks corresponding to the target(s) identified in the low fragmentation spectrum are highlighted in orange in the higher fragmentation spectra. This figure is discussed in Example 3 of Section 3.1. ## Complete Results and Discussion The complete results of searching each of the 11 mixtures against both the select and NIST DART-MS Forensics 2020 library is detailed in Table 2. For all searches, the initial relative intensity threshold considered for identifying target molecules was set to 25 %, the tolerance window for 𝑚/𝑧 errors (𝜖 % , 𝜖 ( ) was set to ±0.005 𝑚/𝑧, and targets were matched to protonated molecules from the libraries. Threshold relative intensities were incrementally dropped by 1 % until all expected compounds were identified for the purpose of demonstration. Based on the notes in Remark 1 of Section 2.1 and Examples 2 and 3 in Section 3.1, it was clear that the target identification stage of the ILSA would struggle to detect a mixture component like amphetamine where the protonated molecule is often the second or third most prominent peak in the spectrum. The relative intensity threshold required to detect the protonated molecules for amphetamine was always less than 5 %, and as low as 1 % in the three of the test mixtures. In two of the three test mixtures where we had to drop the relative intensity to 1 %, the ILSA found benzyl acryl fentanyl as a target, from the NIST library, for another low signal peak. The classification indices computed for benzyl acryl fentanyl were between 0.550 and 0.901. We will return to the interpretation of classification indices later in this discussion but note that these indices should not be considered reliable for scenarios where the target is matched to a protonated molecule in the database and yet the target relative intensity is very low. We have mentioned the notion of assuming detected targets are base peaks, which in this set of tests would have ensured we did not explore targets with very low relative intensities. However, it is likely that even more sophisticated methods for target identification are to be discovered that better utilize the metadata available in the database and the multiple is-CID spectra available for the mixture. For example, can we make assumptions about the presence or absence of adducts? Or can we possibly look for common neutral losses to confirm whether an identified target is a protonated molecule? There is a general trend worth noting about the scoring stage of the ILSA. Classification indices computed using FPIE as a spectral similarity score were always greater for correct matches from the select library than the NIST library. This is not at all surprising as the spectra in the select library were collected simultaneously with the mixture spectra. Accordingly, the mixture spectra and select library spectra are likely to have similar features and background ions that are captured through the FPIE computation. Similarly, the computed classification indices leveraging RevMFs as similarity scores were also mostly greater for matches from the select library than the NIST library, with the occasional exception. Interestingly, these exceptions were always observed for matches of either acetyl fentanyl or benzyl fentanyl (see mixtures 1, 4, 5, 7, 8, 10, 11). There are two reasons why RevMF-based scores would be higher for one library spectrum than another: (1) there are more peaks present in the spectrum to be matched by peaks in the mixture, and (2) the relative intensity values of peaks in the spectrum and the matching peaks in the mixture share a similar pattern. With either scoring method (FPIE or RevMF) there are problematic scenarios and edge-cases with less than desirable performance. Identifying and characterizing these cases is something we are presently exploring both mathematically and empirically. In addition to determining the range of utility for each of the considered similarity scores, it may also be worth exploring the impact of weighting on computed parameters, considering other existing similarity scores, developing novel metrics for similarity, or computing agglomerative values that include multiple similarity metrics. In the results presented in this manuscript, the computed classification indices were disaggregated by type of similarity score (FPIE vs RevMF) for exploratory and analytical reasons. However, it may be useful to employ both classification indices when making decisions, either manually or through an automated classification system. One of the least satisfying aspects of the results presented in this manuscript is the unclear distinction between "good" and "bad" classification indices based on the limited test sets of mixtures we considered. As we noted previously in our discussion of benzyl acryl fentanyl, classifier indices should be viewed with caution when computed for targets identified with very low relative intensities. But even for prominent targets some care must be taken when interpreting classification indices. In mixtures that contained one or both of acetyl fentanyl and benzyl fentanyl, the average classification index of the true compound in the mixture using both libraries and both scoring schemes was 0.945 (n=56) 3 , and the average classification index of the compound not in the mixture using both libraries and both scoring schemes was 0.602 (n=24). These substantial differences in classification indices suggest we can determine with confidence which fentanyl analogs are in the mixture. In contrast, for mixtures where both methamphetamine and phentermine were identified as a target using the general NIST library the average classification index using either scoring scheme was 0.897 (n=14) for methamphetamine, the compound in the mixture, and 0.822 (n=14) for phentermine, the compound not in the mixture. Further, if we used only the RevMF scoring scheme to compute classification indices, the discrepancy was even more muted, with the average classification index being 0.858 (n=7) for methamphetamine and 0.832 (n=7) for phentermine. Rather than select a single classification threshold value for all decision making, it may be advantageous to set several decision threshold indices for classes of compounds to try to best characterize mixture components. Determining these thresholds will require a well-designed set of training and testing data that includes a large number of annotated mixtures. With the limited number of mixture components, test mixture compositions, and experimental conditions considered in this manuscript, it is difficult to make universal claims about how well the ILSA will perform in practical situations. Nevertheless, the results of this manuscript suggest that this multi-stage inverted approach to targeting and scoring spectra, especially with further improvements in the areas outlined in this manuscript, has promise as an effective method for identifying mixture components using DART-MS. ## Disclaimer Official contribution of the National Institute of Standards and Technology (NIST); not subject to copyright in the United States. Certain commercial products are identified in order to adequately specify the procedure; this does not imply endorsement or recommendation by NIST, nor does it imply that such products are necessarily the best available for the purpose.

chemsum

{"title": "A new library-search algorithm for mixture analysis using DART-MS", "journal": "ChemRxiv"}

latent_curing_systems_stabilized_by_reaction_equilibrium_in_homogeneous_mixtures_of_benzoxazine_and_

## Abstract: Latent curing systems are widely used in industrial thermosets in applications such as adhesion, coating, and composites. Despite many attempts to improve the practicality of this dormant reaction system, the majority of commercially available latent products still use particulate hardeners or liquid compounds with blocked active groups. These formulations generally lack fluidity or rapid reaction characteristics and thus are problematic in some industry applications. Here we describe a novel concept that stabilizes highly reactive benzoxazine/amine mixtures by reaction equilibrium. These new latent benzoxazine curing systems have a long storable lifetime but very short gel time at 150 °C. The reversible reaction between benzoxazine and amine is further demonstrated by FT-IR spectral measurements and rheological experiments, and it is shown that the overall characteristics of the latent system are promising for many industrial applications.When liquid reactants are mixed homogeneously, chemical reactions occur at a characteristic reaction rate depending on reactant concentrations, activation energy, temperature, and other factors. However, for many industrial polymerizations, this kinetic rule is a problem from chemical engineering and processing perspectives. The adverse examples can be found in a variety of traditional industries such as adhesion, coating, and composites and in numerous modern technologies such as reaction injection molding, resin transfer molding, and 3D printing 1 . These applications demand good fluidity, low viscosity, and sufficient surface wettability for improving formability in the molding and processing stages 2 ; therefore the reactive mixtures should be completely inert under storage conditions for a long time, generally more than three months for their storage and delivery 3 . However, the curing processes of these applications often prefer a rapid curing rate to meet the requirements of high-speed production lines. Combined with another demand that the curing temperature should be as low as possible (ideally at 80 °C) for saving energy and for protecting sensitive parts in automobile and electronic industries, the control of polymerization rate in both storage and curing conditions is very challenging because it is generally out of the range controllable by using suitable reactions with large activation energies.The use of the latent reaction system, which is dormant in storage conditions (often at room temperature) but rapidly reacts in curing condition (commonly > 100 °C), is one of the most powerful methods used to solve this dilemma in the polymer industries [4][5][6][7][8] . Specifically, latent epoxide formulations have led to profound developments in the applications of epoxy resins relevant for adhesion, coating, and composites 9,10 . Traditional epoxy formulations have two basic components, epoxide and hardener reagent (e.g., multivalent amines) that must be stored separately and mixed immediately prior to use 3 . The latent curing systems are pre-mixed formulations containing both epoxide and hardener reagent, and displaying advantages of simple usage and controllable processing window 3 . The use of particulate hardeners (including reactive emulsions) mixed with the liquid epoxide (Fig. 1a) [11][12][13][14] has been the most successful strategy for the preparation of latent epoxy curing system. For example, dicyandiamide (DICY) powders (particle size smaller than ten micrometers) that are nearly insoluble in epoxy compounds at room temperature have been widely used as a thermally latent curing reagent for epoxy resins 15 . Actually, the reactions of DICY and epoxide are effectively suppressed at the solid DICY surface. However, the micro-particles of hardeners have adverse effects on the fluidity and the viscosity of the curing mixture 16 . For this reason, a number of strategies have been tried for establishing a homogeneous system with the desired curing latency. The blocking of the active groups (such as amines, thiols, and phenols) of hardeners with thermally removable groups (Fig. 1b) is the main approach for achieving the liquid latent system. Sometimes, these methods may induce sluggish liberation of the active groups, resulting in slow curing processes that are undesirable in commercial uses. Despite the many published reports, commercial formulations of the latent epoxy systems are much less developed. Similarly, to date, only few strategies are available for the applications of other types of polymers such as polyurethane 24 . Development of new concepts for the preparation of latent curing systems having both good stability at storage condition and rapid curing rate at mild temperatures is still challenging. Polybenzoxazine is a newly developed thermosetting resin which has rich molecular design flexibility . Recently, we have demonstrated that benzoxazines react rapidly with amines upon heating at 120 °C, with the curing reaction mechanism involving several reversible reactions 31 . Herein, the reaction of benzoxazines with amines is extended to create a new latent curing system. At the outset of our investigations, we envisioned an innovative concept for a latent curing system based on reaction equilibrium in a homogeneous liquid. This strategy differs fundamentally from the blocked hardeners proceeding to release active groups via thermal deprotection. The reversible reaction of benzoxazine and amine with the resulting intermediate polymer (IP) leads to reaction equilibrium and results in a stable viscosity in room temperature for a long time; however, the reaction equilibrium would be broken by heating and induce a rapid curing (Figs 1c and 2). Theoretically, the precipitation of the resulting polymer network separated from the solution phase is an important factor for promoting the disequilibrium. ## Results and Discussion To observe the stability of the mixture containing benzoxazine and amine, we initially investigated curing systems of bisphenol-F-benzoxazine (BF) with two amines, m-xylylenediamine (A1) and trimethylhexamethylenediamine (A2). When the solutions of the two mixtures solved in dimethyl formamide (DMF) were stored at 25 °C, their viscosity remained almost constant for a long term (Fig. 3a and b), but increased rapidly over a few minutes after heating at 150 °C. Indeed, despite being stored for one year at 25 °C, the viscosities of the two curing mixtures did not reach the double value. Table S1 presents the latent curing characteristics of the liquid mixtures, including the statistical gel times cured at 120 °C or 150 °C, and the viscosity values stored at 25 °C and 60 °C. Certainly, these parameters of the mixture solutions are suitable for the use as a latent curing system. We have measured the activation energies of the reactions of 6,6′ -(propane-2,2-diyl)bis-(3-phenyl-3,4-dihydro-2H-1,3-benzoxazine) with A1 and A2, and found that these energies are equal 75 KJ/mol and 59 KJ/mol, respectively 31 . It is obvious that the stability of the mixture solutions stored at 25 °C and the high curing rates at 150 °C are inconsistent with the Arrhenius equation. The BF/amine mixtures are homogeneous and transparent (Figure S3), implying a new stabilization mechanism different from that shown in Fig. 1a. The differential scanning calorimetry (DSC) data of the two mixture solutions at the heating rate of 10 °C/min are shown in Figure S1 and Table S2. The BF/amine mixtures showed DSC characteristics that were quite different from those of pure BF. The addition of the amines results in two exothermic peaks at low (< 150 °C), and elevated (> 180 °C) temperatures, suggesting that at least two reaction mechanisms are present in the curing process. The curing enthalpy values of these DSC peaks were approximately 40 J/g and 13 J/g, respectively, indicating that the reactions of BF with the amines are slightly exothermic in contrast to the large enthalpy value of the thermally induced ring-opening polymerization of bulk BF. Furthermore, FT-IR measurements of the A1/BF system indicated a prolonged storability at 25 °C (Fig. 4 and Table S3). The oxazine ring of BF showed three characteristic peaks at 940 cm −1 , 1027 cm -1 , and 1223 cm −1 , assigned to the ring symmetric and anti-symmetric stretching of the C-O-C bond . When the mixture was stored at 25 °C, the three peaks were reduced clearly in the first day, but then become almost constant. These results not only demonstrated the ring-opening reaction of BF but also suggested that the reactive mixture may exhibit a reaction equilibrium. When the curing temperature was fixed at 150 °C, the characteristic peaks of the oxazine rings completely disappeared and characteristic peaks assigned to new chemical structures such as tetra-substituted benzene (1450-1480 cm −1 ) 34,35 , significantly increased (Figure S2). This FT-IR spectroscopic evidence suggests that the reaction equilibrium is broken by heating and new polymers are produced. Measurement of the viscosity of different BF/amine systems enables the observation of the reaction progress and therefore indirectly reflects the reaction equilibrium. Using a rheometer, the BF/amine mixture solutions were isothermally heated at 25 °C, and the rheological data were recorded (Fig. 5). The concentration dependence of the viscosity of polymer solutions is usually represented by the Huggins equation (1). where c is the solution concentration, η is the dynamic viscosity of the solution, η 0 is the dynamic viscosity of the solvent, [η ] is the intrinsic viscosity, and k 1 and k 2 are the Huggins constants. Because the value of c is sufficiently small for the higher-order terms to be negligible, the Huggins equation can be simplified to, 0 Actually, the rheological results for the BF solution were in good agreement with equation ( 2), showing that the relative viscosity increases almost linearly with increasing BF concentration (Fig. 5, line 1). However, when a small amounts of BF or hexamethylenediamine (A3) were increasingly added to the BF/A3 mixture solution, the relationship (Fig. 5, curves 2 and 3) between the relative viscosity and the increased solute concentration upward 1c, the increase of BF or A3 can shift the equilibrium to the right, resulting in an increased polymerization degree of the oligomer adduct. Whereas the addition of p-toluene sulfonic acid (PTSA, consuming some A3 by neutralization with A3) to the BF/A3 mixture drive the reaction equilibrium to the left, causing the polymerization degree of the oligomer adduct to decrease and resulting in the downward curve (curve 5). Because triethylamine (A4) can decrease intermolecular bonds (phenol with amines) between the IP chains, the addition of A4 results in a more downward curve (curve 4) than curve 5. The mechanical and thermal properties of the cured resins are of key importance for their use as a thermoset. Figure 6a and Table 1 summarize the breaking tensile results of the BF/amine mixtures bonding in two aluminum sheets. The DSC results showed that the ring-opening polymerization of the pure BF monomer generally happens after heating to 200 °C. Therefore, the curing of the pure BF monomer at 150 °C produce only a small breaking tensile strength of 44.53 MPa. However, the addition of amine A1 or A2 led to a breaking tensile strength greater than 180 MPa, reaching more than 70% of that of the resin cured at 180 °C. These results demonstrated that the addition of the amines promote the curing process of benzoxazine. Because A1 have the benzene ring structure, the cured BF/A1 mixture showed a stronger breaking strength but a lower stretch rate than that of the cured BF/A2 mixture. Furthermore, the latent mixtures were cured to composite samples with standard filter paper (average pore size is 18 μ m) as a reinforcement filler to reduce the errors generated in the forming process, and were evaluated using dynamic mechanical analyses (DMA) and thermal gravity analysis (TGA). As shown in Fig. 6b, DMA experiments showed that the amine-cured BF resins exhibit high storage modulus values ranging from 4.5 to 5.5 GPa, implying that a highly cross-linked network was formed. The peak that appeared at 150 °C in the tan δ data (Fig. 6c) is in good agreement with optical images of the cured resins (Figure S4), suggesting that a homogeneous polymer network was obtained in this curing process. Figure 6d shows the TGA profiles and Table 2 provides the important data collected from the TGA thermograms. The TGA curves indicated that all thermal degradation temperatures are higher than 150 °C, may be due to the excessive amine. The carbon residue rate of the cured pure BF resin is 46%, whereas the cured BF/A1 and BF/A2 resins are 51% and 24%, respectively. The aromatic amine A1 showed a positive effect on the carbon residue rate due to the increased amount of aromatic rings in the cured resin, but the aliphatic amine A2 had a negative effect that is explained by the decomposition of the aliphatic chains at high temperatures. ## Conclusion In conclusion, we described a new thermal latent curing concept basing on the stabilization caused by reaction equilibrium in a homogeneous solution of benzoxazine and amine. FT-IR spectral measurements and rheological experiments suggest that the reaction between BF and amine is reversible and that the reaction equilibrium stabilizes the curing mixture for a long time at low temperature. By heating up to 120 °C, the reaction equilibrium is broken by the formation of a polymer network, resulting in a rapid cure to thermoset resin. These results represent the first demonstration of a latent curing system based on reaction equilibrium in the homogeneous liquid phase. The stability at room temperature, the reactivity induced by heating, and the material characteristics make the new latent reaction concept highly interesting for various applications such as coatings, adhesives, composites, and healable materials. Preparation of BF/amine mixtures and their cured resins. (1) A BF solution in butanone (75 wt%) was dried in a reduced pressure at 100 °C to obtain BF powder (4.34 g, 0.010 mol), mixed with a diamine (A1 or A2, 0.011 mol), dissolved in DMF (5.0 ml), stored at 25 °C, and used for latent tests (including viscosity and gel time) directly. (2) The A4/BF solutions (10 ml) containing 0.20 mmol BF and A4 were gradually added BF, A3, A4, and PTSA by an amount of 0.20 mmol, respectively. Rheological data were measured using the rheometer before the addition of the additives at 25 °C. The rheological measurement and the addition were repeated for five times. (3) BF power (4.34 g, 0.010 mol) and a diamine (A1 or A2, 0.011 mol) were dissolved in chloroform (8 ml) together. Standard filter papers (with an average pore size of 18 μ m) were immersed in the mixture solutions for one day, weight up to 150%, clamped using polyimide sheets, dried under vacuum to remove the solvent, and progressively cured at 120, 150, 180 °C for 2 h. The resulted composite samples were subjected to DMA and TGA tests. Table 2. TGA data selected from the TGA curves (Fig. 4d) of the cured amine/BF resins.

chemsum

{"title": "Latent curing systems stabilized by reaction equilibrium in homogeneous mixtures of benzoxazine and amine", "journal": "Scientific Reports - Nature"}

near-infrared_photosensitization_via_direct_triplet_energy_transfer_from_lanthanide_nanoparticles

## Abstract: A new strategy to achieve high-performance NIR-light-driven photosensitization is developed, based on lanthanide nanoparticle-organic photosensitizer nanoconjugates, through direct triplet energy transfer from the nanoparticles to the photosensitizers. This method permits efficient NIR photosensitization to generate cytotoxic singlet oxygen under 100-times-lower NIR power than conventional upconversion methods, holding great promising for emerging applications such as deep-tumor photodynamic therapy. ## INTRODUCTION Photosensitization is one of the most important processes to generate reactive oxygen species (ROS) (e.g., single oxygen) and associates with diverse applications in photodynamic therapy (PDT), synthetic chemistry, and organic waste decomposition. Conventional photosensitization reaction for singlet-oxygen generation mostly requires the use of a photosensitizer that can produce a large amount of triplet excited states upon light absorption. This is because the triplet energy has higher probabilities to transform ground-state oxygen molecules to excitedstate singlet oxygen species. 7 However, except under certain stringent conditions (e.g., collision perturbations 8 ), a spin-triplet excited state can be hardly generated via the optical transition from a spin-singlet ground state due to the spin-forbidden nature. As a result, most of the photosensitizers must first be excited with ultraviolet or visible light to their higher-lying excited singlet states. The lower-lying triplet states can then be populated through singlet-triplet intersystem crossing, consequently initiating the production of reactive singlet oxygen. This process causes inevitable photon energy loss because the lowest photon energy required for exciting the singlet states of most photosensitizers is around 1.8-3.5 eV 9 , much higher than the minimum energy needed for generating a singlet oxygen ( 3 S g / 1 D g ; 0.98 eV). ## The bigger picture Photosensitization reactions for the generation of reactive oxygen species (ROS) have played important roles in many applications, including energy conversion, organic waste decomposition, and photodynamic therapy. Using near-infrared (NIR) light for photosensitization is attractive because NIR excitation offers deeper penetration depth through various media such as wastewater and biological tissues. However, conventional approaches for NIR photosensitization are hampered by low conversion efficiency and difficulties in molecular designs. Herein, we developed a facile lanthanide-triplet sensitization method to realize highperformance NIR photosensitization by adopting lanthanide nanoparticle-organic photosensitizer nanoconjugates. This method enables efficient ROS generation at ultralow NIR irradiance, which is potentially useful for diverse applications related to an area such as human health, pollution abatement, and efficient energy utilization. To harness NIR light for singlet-oxygen generation, one method is to develop sensitizer molecules with red-shifted optical absorption. 10,11 This method is constrained by stringent molecular designs and complicates synthetic approaches. Besides, it is still challenging to develop molecular sensitizers with maximum absorption wavelength above 800 nm, due to the largely decreased efficiency and increased structural instability for chromophores absorbing at longer wavelengths. Alternatively, non-linear upconversion processes can be adopted to initiate NIR photosensitization. However, these processes generally show very low conversion efficiency (<1%) and must be excited at relatively high-power density. Considering the triplet electronic energy levels of many conventional organic photosensitizers (e.g., porphyrins and phthalocyanines) are at an infrared region of around 1.0-1.5 eV 9 , we reasoned that NIR photosensitization may be achievable if these triplet levels could be directly populated without going through singlet-triplet intersystem crossing. This is principally viable if we could find a rational donor with appropriate energy coupling to the triplet states of the organics. Nevertheless, because the optical transitions to triplet levels contain negligible oscillator strength, this energy coupling should not lead to direct spectral overlaps between donor's emission and acceptor's absorption as that in typical Fo ¨rster resonance energy transfer (Figure 1A). Here, we propose the use of antibody-sized (sub-10 nm) lanthanide inorganic nanocrystals to sensitize organic photosensitizers through direct triplet energy transfer. We expect a ''dark'' electronic coupling that collects the 4f states of trivalent lanthanide ions with the triplet levels of organic sensitizers, thereby leveraging NIR-light-activated photosensitization with largely improved performance (Figure 1A). ## RESULTS As a proof of principle, we prepared a NaGdF 4 :Nd (50 mol %)-Chlorin e6 (Ce6) donor-acceptor hybrid system and systematically investigated its photophysical properties (Figures S1-S5). Ce6 is a classical photosensitizer that has been widely applied in PDT studies for the treatment of multiple types of cancer tumors. It was chosen as the energy acceptor because the lowest triplet state of Ce6 is at 1.14 eV, making it possible for NIR photosensitization. 9 We employed an 8 nm-sized NaGdF 4 nanocrystal as a host material and incorporated Nd 3+ as a light absorber to sensitize the triplets of Ce6. Due to the efficient 4 I 9/2 / 4 F 5/2 transition, Nd 3+ is capable of absorbing at NIR region at $800 nm. Note that the spontaneous emission of Nd 3+ upon such absorption lays at a much lower photon energy region than the optically allowed singlet absorption of Ce6 (Figure 1B). Despite having no spectral overlap, the energy of spin-forbidden S 0 /T 1 transition of Ce6 matches with several 4f-4f transitions of Nd 3+ (e.g., 4 F 3/2 / 4 I 11/2 ), offering the possibility of energy transfer from Nd 3+ to Ce6 through direct lanthanide-triplet sensitization (Figure 1B). We first implemented transient absorption spectroscopy to demonstrate the sensitization of Ce6 molecules by nanocrystals under NIR illumination (Figure S6). When excited by a continuous-wave diode laser at 808 nm, the deoxygenated tetrahydrofuran (THF) suspension of NaGdF 4 :Nd-Ce6 exhibited a narrow bleaching band at 660 nm and a broad absorption band between 420-600 nm (Figure 2A). These two bands can be assigned to the characteristic ground-state bleaching and photoinduced triplet absorption of Ce6, respectively. 25 It clearly indicates the formation of excited triplets in the hybrid system. By contrast, transient absorption spectra of Ce6 and NaGdF 4 :Nd alone do not show any noticeable absorption band in the range of 420-680 nm, suggesting the absence of direct light excitation to Ce6 under such illumination (Figures 2A and S7). Moreover, the sensitized photon energy on excited Ce6 triplets can transfer to freely diffused oxygen molecules and generate cytotoxic singlet oxygen. This process is evidenced by the observation of featured phosphorescence of singlet oxygen ( 1 D g / 3 S g ; at $1,275 nm 11 ) in the luminescence spectrum of air-saturated suspension of NaGdF 4 :Nd-Ce6 at 808 nm excitation (Figure 2B). We observed a sharp decline of such phosphorescence when the colloidal suspension was purged with nitrogen. The phosphorescence was also significantly quenched by the addition of violanthrone-79 (V79), a singlet-oxygen quencher known to have triplet states below singlet oxygen. 26 These results confirm the emission comes from singlet oxygen, which supports the direct triplet sensitization by NaGdF 4 :Nd nanocrystals. Besides, we also observed substantial changes in the transient absorption spectrum of NaGdF 4 :Nd-Ce6 after the addition of V79 (Figure S8). It may suggest the triplet energy transfer from Ce6 to V79. We also performed time-resolved luminescence spectroscopy to probe the dynamics of triplet sensitization by the nanocrystals. We monitored the 4 F 3/2 state relaxation of Nd 3+ in the presence and absence of Ce6 by measuring the decay function of 4 F 3/2 / 4 I 9/2 transition at 899 nm. While the luminescence decay of Nd 3+ is mono-exponential in free NaGdF 4 :Nd nanocrystals with a characteristic lifetime of 2.23 ms, it becomes multi-exponential and significantly shortens at the early stage to less than 0.1 ms (the measured shortest lifetime was limited by the response of the microsecond-scale pulse excitation source) after the nanocrystals were modified with Ce6 (Figure 2C). This is a clear indication of non-radiative energy transfer from nanocrystals to adjacent Ce6 molecules, which is also in agreement with the observation of photoluminescence quenching of NaGdF 4 :Nd nanocrystals after conjugating with Ce6 (Figure S9). By isolating the energy transfer dynamics at the early decay stage, we derived a dominant energy transfer rate of 9.6 3 10 6 s 1 . The transfer rate is much faster than the intrinsic decay dynamics of free NaGdF 4 :Nd nanocrystals, resulting in an extremely high energy transfer efficiency of >95%. Besides, there is also a longer decay component with a characteristic time of 1.07 ms, suggesting the existence of some less efficient energy transfer paths, probably from the buried Nd 3+ ions deep inside the nanocrystals. By calculating the total luminescence lifetime changes of Nd 3+ with/without Ce6, we have obtained an overall energy transfer efficiency of $60% from NaGdF 4 :Nd to Ce6. Since spin-singlet-triplet transitions of Ce6 molecules have a negligible dipole moment, the sensitization of triplet states by lanthanide nanocrystals should not Article follow typical resonant dipole-dipole coupling mechanisms. 27 Therefore, we speculated that a short-range electronic interaction (e.g., Dexter-type exchange coupling) should be a dominant concept for energy transfer from lanthanide energy donors to Ce6 acceptors. To validate the hypothesis, we investigated the effect of donoracceptor separation on singlet-oxygen yields using 1,3-diphenylisobenzofuran (DPBF) as an indicator. We synthesized core-shell NaGdF 4 :Nd@NaGdF 4 nanocrystals with different shell thickness (0-2.5 nm) and modified them with the same amount of Ce6 molecules (Figure S10). As shown in Figures 2D and S11, the generating rate of singlet oxygen by NaGdF 4 :Nd@NaGdF 4 -Ce6 drastically decreases while the photoluminescence increases as the thickness of the shell increases. It was found that even a 0.7 nm thin layer of inert NaGdF 4 shell can strongly prohibit the formation of singlet oxygen by a factor of $70%. This is in stark contrast to recent studies on resonance energy transfer to surface defects/ligands from similar lanthanide-doped core-shell nanocrystals that show a much longer effective distance over 2 nm via dipole-dipole coupling. 28,29 The observations reveal the triplet transfer occurs only in a narrow interfacial region near the surface of the nanocrystals, which is consistent with the characteristic short-range energy transfer of exchange interaction. 30 We also confirm that a higher concentration of Nd dopants would contribute to enhanced singlet oxgen generation, while the particle size and matrix have a negligible effect on the energy transfer process (Figures S12-S14). To shed more light on the origin of the short-range coupling, we performed firstprinciples calculations for the coupled nanosystem (see supplemental information). We simulated a typical situation when a Ce6 molecule is attached on the {001} facet of hexagonal NaGdF 4 :Nd substrate. As illustrated in Figures 2E and S15, the calculation confirms the occurrence of charge density redistribution when the organic molecule is approaching to the nanocrystal surface. The differential charge density profile indicates that the coupling is mainly between the conjugated polycyclic carbons of Ce6 and the trivalent lanthanide ions right below the organic molecule. To verify the impact of energy coupling on triplet transfer, we examined photosensitization by lanthanide nanocrystals in combination with a series of porphyrin and phthalocyanine derivatives (PpIX, TCPP, Ce6, and ZnPcS) with various triplet energy (Figures 3A and S16; Table S1). Interestingly, we found that only TCPP (T 1 =1.43 eV 9 ), Ce6 (T 1 = 1.14 eV), and ZnPcS (T 1 = 1.13 eV 9 ) can be efficiently sensitized by NaGdF 4 :Nd nanocrystals under 808 nm illumination (Figures 3B and 3C). The NaGdF 4 :Nd-PpIX nanoconjugates show little effect in the degradation of DPBF, indicating a poor ability for generating singlet oxygen by the system. This is presumably because of the large endothermic barrier for energy transfer from the lower Nd 3+ excited 4 F 3/2 state (1.43 eV) to the higher PpIX triplet level (1.56 eV 9 ). The results indicate the importance of an appropriate energy match of donor and acceptor for efficient triplet sensitization. More interestingly, we further confirmed the generality of the lanthanide-triplet sensitization approach toward other photosensitizers and lanthanide ions. We found the NaGdF 4 :Nd is also able to sensitize molecules other than porphyrins such as 5-carboxylic-tetracene (Figure S17). We also replaced Nd 3+ with Yb 3+ whose excited energy state ( 2 F 5/2 ; 1.26 eV) is lower than the lowest triplet levels of PpIX and TCPP but higher than those of Ce6 and ZnPcS (Figure 3D). In a similar manner, we observed efficient NIR photosensitization from Ce6 and ZnPcS in conjugation with NaGdF 4 :Yb (50 mol%) nanocrystals after pumping the 2 F 5/2 state of Yb 3+ through 980 nm excitation (Figure 3E). Compared with the conventional upconverted photosensitization, the lanthanide-triplet strategy is not compromised with energy loss through multiphoton accumulation or intersystem crossing (Figure 3F). It lowers the excitation threshold required for populating the triplet states. To demonstrate such effect, we examined the excitation power-dependent singlet-oxygen generation by the systems taking the widely studied Ce6-functionalized NaYF 4 :Yb,Er(18,2 mol %)@NaYF 4 upconversion nanoparticles (UCNP-Ce6) as control 31,32 (Figure S18). As shown in Figure 3F, the irradiation power required for the lanthanide-triplet systems is generally two-orders of magnitude lower than that for the canonical upconversion standard. For instance, we found that at least 435 times higher irradiance (8.7 W/cm 2 ) must be used for the UCNP-Ce6 to maintain similar singlet-oxygen yields as that of NaGdF 4 :Nd-ZnPcS nanoconjugates under 20 mW/cm 2 of irradiation. Moreover, the singlet-oxygen generation rate was still 60 times lower even when the UCNP-Ce6 was excited by a power density of 3.7 W/cm 2 , $9 times stronger than that was used for NaGdF 4 :Nd-ZnPcS (400 mW/cm 2 ). It is worthwhile noting that without the need of populating higher singlet states our approach can in principle generate 100% of excited triplets in organic photosensitizers. By comparing a standard photosensitizer (methylene blue; singlet-oxygen quantum yield 0.56), an ll Chem 7, 1615-1625, June 10, 2021 1619 Article observed singlet-oxygen quantum yield of 0.49G0.12 was obtained for NaGdF 4 :Nd-Ce6 hybrids under 808 nm irradiation (20 mW/cm 2 ). It suggests an oxygen sensitization efficiency of 0.49/0.6 z 0.82 by the Ce6 molecules, considering the energy transfer efficiency from the nanoparticles to Ce6 is around 60%. This value is higher than the singlet-oxygen yield of pristine Ce6 (quantum yield $0.63) through direct excitation. The result implies that our method can be more efficient compared with the conventional photosensitization processes adopting singlet-triplet intersystem crossing. Given the ability of efficient photosensitization at ultralow irradiation, we further investigated the feasibility of NaGdF 4 :Nd-Ce6 as a PDT agent for treating cancer tumors at the deep lesion. Prior to the investigation, we employed Pluronic F127 (PF127), a biocompatible polymer widely applied in theranostic applications, to functionalize the surface of the as-prepared nanoconjugates to render them good stability under physiological conditions (Figures S19-S25). We next assessed the dark cytotoxicity of the nanoconjugates to three different cancer cell lines (4T1, CT26, and SKOV3) using MTT assay. It was found that the Ce6-conjugated nanocrystals show negligible inhibition to cell growth within 24 h at the concentration range of 10-200 mg/mL (Figures 4A and S26). By contrast, even when the cells were covered by 4-mm-thick chicken tissue, the nanoconjugates exhibited markedly cytotoxicity to all of the three cell lines after exposure to NIR irradiation (790 mW/cm 2 ) for 3 min. In all cases, the NaGdF 4 :Nd-Ce6 conjugates show apparently higher lightinduced cytotoxicity than that of the UCNP-Ce6, owing to more efficient singlet-oxygen production through direct triplet sensitization (Figures 4B and S26). The flow cytometry analysis reveals that 50 mg/mL of NaGdF 4 :Nd-Ce6 causes more than 80% of cell death, while the same amount of UCNP-Ce6 results in only around 30% of cell death under the same dosage of irradiation (Figure S27). It is worth mentioning that a higher concentration of NaGdF 4 :Nd-Ce6 can work for PDT at much lower irradiations. We demonstrated that even an irradiation as low as 10 mW/cm 2 can induce substantial cytotoxicity to CT26 cells when the cells were treated with 400 mg/mL of NaGdF 4 :Nd-Ce6 (Figures S28 and S29). The generation of singlet oxygen in situ in living cells was further verified by confocal fluorescence microscopic imaging using 2 0 ,7 0 -dichlorodi-hydrofluorescein diacetate (DCFH-DA), a chemical sensor widely used for the determination of all kinds of ROS. As expected, we observed the fluorescence of the NaGdF 4 :Nd groups were always more intense than the UCNP groups, indicating the superiority of NaGdF 4 :Nd in PDT for generating cytotoxic reactive oxygen species (Figures 4C and S30-S35). We then evaluated the efficacy of the nanoconjugate as a PDT agent in vivo for treating deep-tumor models at low irradiance of 660 and 808 nm laser with different penetration capacities (Figure S36). To maximize the PDT effect, we modified the nanoconjugates with folic acid (FA) to aid their specific targeting to the overexpressed folate receptors in SKOV3 human ovarian cancer cells. 33 In a typical experiment, we first intravenously injected the SKOV3-bearing mice with 200 mL of the nanoconjugates (10 mg/mL) or PBS. By tracking both the fluorescence distribution of Ce6 and T 1 -weighted magnetic resonance imaging (MRI) of Gd 3+ in the bodies, we confirmed that the nanoconjugates can accumulate at tumor sites in 24 h (Figures 4D, S37, and S38). Moreover, compared with the group with FA modification, the NaGdF 4 :Nd-Ce6 nanoconjugates without FA modification showed weaker ex vivo fluorescence signal in tumor tissues 24 h post the injection indicating less nanoconjugates had been delivered to the tumor sites (Figure 4E). This finding supports FA can assist preferential localization of NaGdF 4 :Nd-Ce6 nanoconjugates in targeted tumor tissues. After the injection for 24 h, we treated the mice with an 808 nm laser for 30 min with a relatively low irradiance of $80 mW/cm 2 . During the light exposure, the tumor sites were blocked by a pork tissue with a thickness of 4 mm to mimic a condition when a deep lesion was treated by PDT. Note that such irradiance intensity is significantly lower than those adopted in conventional nanoparticle-based deep-tissue PDT (Table S2). We also confirmed that there is no obvious heating effect for the mice under such irradiation (Figure S39). In order to compare the performance of our direct triplet sensitization method with conventional single-triplet PDT approaches, we also treated control groups with the same light dosage by 650 nm far red light to directly excite the singlet state of Ce6. After the treatment, we assessed the PDT effect by monitoring the changes in tumor volumes and weights in different groups (Figures 4F and 4G). Whereas the subcutaneously injected SKOV3 cells in control groups rapidly grew into solid tumors, the growth of the tumors in all NaGdF 4 :Nd-Ce6-FA + light treatment groups were significantly inhibited (p < 0.001 compare to control groups). Although the progression of tumors in NaGdF 4 :Nd-Ce6-FA + 650 nm laser groups also showed inhibition of growth, the growth rate was statistically higher than that of NaGdF 4 :Nd + 808 nm laser group when the tumor sites were covered with 4 mm pork tissues (p = 0.014 at 28 days after the treatment) (Figure 4F). Notably, even without the pork tissue coverage (groups 5 and 6), the 808 nm lighttreated group shows greater antitumor effect than the 650 nm light-treated group (p = 0.009 at 28 days after treatment). This is more likely due to the higher penetration depth of 808 nm light in solid tumors. We also conducted histopathology H&E staining and immunohistochemical TUNEL analysis of tumor paraffin slices for the model mice at 28 days after the PDT treatment. We observed a greater population of apoptotic cells in histological sections from NaGdF 4 :Nd-Ce6-FA + 808 nm laser group (Figures 4G and S40-S42). Blood biochemistry tests of mice and H&E staining of the major organs were performed, indicating minimal adverse effects for normal organs after different treatments (Figures S43-S44). The metabolic clearance investigation after the treatment indicated Ce6 were cleared through both the hepatic and renal systems after injection for 72 h, while the NaGdF 4 :Nd nanoparticles can be completely removed from the body in 30 days (Figures S45-S47). We also confirmed both Ce6 and NaGdF 4 :Nd nanoparticles exhibited negligible accumulation to normal tissues such as brain, muscle, and bone at the therapeutic dose. Taken together, these results clearly suggest the superiority of the direct triplet sensitization approach for deep-tissue PDT. ## DISCUSSION Conventional approaches for screening new sensitizers for singlet-oxygen production typically require molecules with a large light-absorption coefficient and a high ratio of the singlet-triplet intersystem crossing. 34 The results demonstrated in this work point to a radically different design strategy for photosensitization of which the intersystem crossing process can be circumvented by using lanthanide inorganic nanocrystals to direct sensitize the triplet excited states of photosensitizers. This approach, by minimizing photon energy loss during activation of sensitizer molecules, enables us to realize high-performance NIR photosensitization systems capable of working at ultralow radiation energy with much-improved efficacy for deep-tissue ablation and cancer therapy. Given the abilities of activating triplet states of a wide range of photosensitizers by near-infrared light, our findings may assist exploitation of photocatalysis reaction in new fields of application, promoting progress for areas relevant to human health, pollution abatement, and efficient energy utilization. 35 ## EXPERIMENTAL PROCEDURES Resource availability Lead contact Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Renren Deng (rdeng@zju.edu.cn).

chemsum

{"title": "Near-infrared photosensitization via direct triplet energy transfer from lanthanide nanoparticles", "journal": "Chem Cell"}

antibacterial_activity_and_mechanism_of_plant_flavonoids_to_gram-positive_bacteria_predicted_from_th

## Abstract: Antimicrobial resistance seriously threatened human health, and new antimicrobial agents are desperately needed. As one of the largest classes of plant secondary metabolite, flavonoids can be widely found in various parts of the plant, and their antibacterial activities have been increasingly paid attention to. Based on the physicochemical parameters and antibacterial activities of sixty-six flavonoids reported, two regression equations between their ACD/LogP or LogD 7.40 and their minimum inhibitory concentrations (MICs) to gram-positive bacteria were established with the correlation coefficients above 0.93, and then were verified by another sixty-eight flavonoids reported. From these two equations, the MICs of most flavonoids against gram-positive bacteria could be roughly calculated from their ACD/LogP or LogD 7.40 , and the minimum MIC was predicted as approximately 10.2 or 4.8 μM, more likely falls into the range from 2.6 to 10.2 μM, or from 1.2 to 4.8 μM. Simultaneously, both tendentiously concave regression curves indicated that the lipophilicity is a key factor for flavonoids against gram-positive bacteria. Combined with the literature analyses, the results also suggested that the cell membrane is the main site of flavonoids acting on gram-positive bacteria, and which likely involves the damage of phospholipid bilayers, the inhibition of the respiratory chain or the ATP synthesis, or some others. Antimicrobial resistance (AMR) has been seriously threatened human public health and global economic development, and new antimicrobial agents are desperately needed 1,2 . Antibiotics, as the secondary metabolites produced by many bacteria, actinomycetes and fungi, showed remarkably antimicrobial activities, while they also bring some toxic side effects to human body, and are unavoidable to lead to the resistance 3 . Many plant ingredients present weaker antimicrobial activities, while some of them can reverse the resistance of antimicrobial agents 4 . Simultaneously, most of them are considered nontoxic to human body because of their ubiquity in all sorts of plant derived foods and beverages. As one of the largest classes of plant secondary metabolite, flavonoids can be widely found in various parts of the plants, such as fruit, vegetables, nuts and tea 4 . These compounds have a wide range of pharmacological activities including antibiosis, antioxidation, and coronary heart disease prevention, etc. It is worth noting that some flavonoids can enhance the sensitivity of bacteria to antibiotics, and even reverse the AMR 4,5 . Thereout, the antibacterial activities of flavonoids have been paid more and more attention to. Recently, several investigations were performed for the antimicrobial activities of flavonoids, and the probable relationships between their chemical structures and antimicrobial activities were also summarized . However, the regularity conclusions on the structure-activity relationships of flavonoids against bacteria still need to be further explored. During our researches on antimicrobial agents , it is vaguely found that the antimicrobial activities of flavonoids are not related to their special structure, while may be related to their polarities or lipid-water partition coefficients. Many data of plant flavonoids, involving their chemical structures and antibacterial activities reported in previous papers, were searched and analyzed for proving it. The inhibitory activities of plant flavonoids against gram-positive bacteria especially Staphylococcus aureus can be widely searched, while those against gram-negative ones and fungi were reported too few to carry out statistical analyses 4,6 . Thereby, the former was our focus in this research. As the inhibitory activities of a compound against different pathogenic bacteria are Data analysis and correlation establishment. The regression analyses for the physicochemical parameters CLogP, ACD/LogP, or LogD 7. 40 and the antimicrobial activities (MIC or MIC 90 ) of these flavonoids to a certain pathogenic bacterium were respectively performed, and their regression curves were showed on Fig. S1 to S6 in Supplementary Information. From these figures, nearly all regression curves indicate that the antibacterial activities of these flavonoids present similar change characteristics along with the increase of their LogP or LogD 7.40 . First, the antibacterial activities will dramatically increase when the LogP or LogD 7. 40 www.nature.com/scientificreports/ to a specific value. Along with the further increase of LogP or LogD 7.40 , the antibacterial activities will first increase tendentiously and then decrease. Simultaneously, their regression equations between the physicochemical parameter (x) and the MIC (y), together with the correlation coefficients (r), were respectively presented on Fig. S1 to S6, and summarily listed in Table 7. Most correlation coefficients (r) were more than 0.90 (Table 7). This indicated that there is a good correlation between the physicochemical parameter CLogP, ACD/LogP, or LogD 7.40 and the antimicrobial activities (MIC), of these flavonoids to a certain pathogenic bacterium. www.nature.com/scientificreports/ As we pointed out above, the antimicrobial activities of a compound against different pathogenic bacteria were varied, and even against the same one in different determination conditions. Thereby, the regression analyses were respectively performed for these flavonoids reported in different papers. Considering that the pathogenic bacteria used for antibacterial experiments mainly involved S. aureus, S. epidermidis, and B. subtilis, the same compound should present similar inhibitory activities and identical antibacterial mechanism to these gram-positive bacteria. Thereby, we put the physicochemical parameters and the average MICs to S. aureus, S. epidermidis, or/and B. subtilis (Tables 1, 2, 3, 4, 5 and 6), of these flavonoids together for further regression analyses. The results indicated that the correlation between CLogP and antibacterial activities (MICs) is weak with a correlation coefficient of 0.8412, while that between ACD/LogP or LogD 7.40 (x) and MICs (y) is more reliable (Fig. 7). The regression equations were respectively expressed as y = − 1.6745x 5 + 56.143x 4 − 741.93x 3 + 4831.8x 2 − 15531x + 19,805 and y = − 1.1474x 5 + 38.802x 4 − 515.39x 3 + 3361.9x 2 − 10789x + 13,706, with the correlation coefficients of 0.9349 and 0.9309, respectively. These further proved, by a larger sample, that the inhibitory activities of these flavonoids to gram-positive bacteria will nonlinearly increase as the ACD/LogP or LogD 7.40 increase to approximately 7.0, and then decrease along with the further increase of ACD/LogP or LogD 7.40 . Verification. To verify the above correlations, other sixty-eight flavonoids (Fig. 8) including flavone, isoflavone, flavonol, flavanonol, dihydroflavone, dihydroisoflavone, flavane, and chalcone subclasses etc., reported in seven papers 4, , were selected for the comparison of theoretical and reported MICs. Using above two regres- ), the theoretical MICs of these flavonoids can be calculated. Considering that many factors, such as determination method, concentration of bacterial suspension, and test medium used, may influence on the determination of MIC 5 , the results reported would fluctuate within a reasonable range of the actual values. Thereout, the predicted MICs ranged from 1/4 × to 4 × the determined one were acceptable (marked as A), especially those ranged from 1/2 × to 2 × the determined one, were considered as complete coincidence (marked as C) since the MICs were generally determined by double dilution method 22 . Simultaneously, those more than or equal to the minimum value when the determined MICs were no upper limit were also regarded as complete coincidence (marked as C). Otherwise, those were unacceptable (marked as U). The results (Table 8) indicated that the predicted MICs were in acceptable or complete coincidence with the measured ones for approximate 85.3% flavonoids. Although the antibacterial activities of ten flavonoids (14.7%) are unsatisfactorily predicted, there are six compounds with the predicted MICs falling into the range of 1/8 × to 8 × determined ones. This together indicated that the MICs of most flavonoids against gram-positive bacteria can be roughly calculated from their ACD/LogP or LogD 7.40 although the predicted values are not in accordance with their tested ones for a few flavonoids. At least, these indicated that the ACD/LogP or LogD 7.40 is a key factor for the inhibitory activities of plant flavonoids against gram-positive bacteria. ## Discussion and conclusion Flavonoids can be widely found in various parts of the plant, and their antibacterial activities have been paid more and more attention to, especially after some of them were discovered to have the potency to enhance the susceptibility of some antibiotics to bacteria 4,5 . Based on the related data of plant flavonoids reported, many related physicochemical parameters were calculated, using software ChemBioDraw Ultra 12.0 and ACD/Labs 6.0, for the discovery of the correlations between the physicochemical parameters and the MICs of flavonoids against gram-positive bacteria. Two regression equations between the ACD/LogP or LogD 7. Considering that the experimental MICs would fluctuate within a reasonable range 5 , the minimum MIC of plant flavonoids will likely fall into the range from 2.6 to 10.2 μM, or from 1.2 to 4.8 μM, predicted from their ACD/ LogP or LogD 7.40 . After all, the antibacterial activities of a compound to different pathogens are varied, and so these two regression equations, mainly valuable for Staphylococcus and Bacillus, may not always be suitable for flavonoids to other gram-positive bacteria. However, the acceptable range from 1/4 × to 4 × the determined MICs will increase applicability of these two equations used for the prediction of plant flavonoids to other gram-positive bacteria. To say the least, if necessary, similar regression equations can be also established from the physicochemical parameters and the MICs to other gram-positive bacteria, of flavonoids. Thereby, we concluded that the MICs of most flavonoids against gram-positive bacteria can be roughly calculated from their physicochemical parameters ACD/LogP or LogD7. 40. Lipophilicity is a very important descriptor indicating membrane permeation 23 , and generally expressed as LogP which is valid only for a single electrical species. For ionizable drugs, LogD that refers to a pH-dependent mixture of all electrical species presented at any given pH was regarded as a better descriptor reflecting the actual partitioning and lipophilicity 24,25 . Generally, most flavonoids contain two or more phenolic hydroxyl groups , and present similar weak acidity with the pKa of 7.0 to 10.0. Thereby, their LogD will correspondingly decrease along with the increase of environmental pH from about 5.0. Considering the pH in human blood or in the media of MIC determination was approximately 7.40, their LogD at pH 7.40 were selected. These together above indicate that the lipophilicity of plant flavonoids is a key factor for their inhibitory activities to gram-positive bacteria. As the lipophilicity is closely related to membrane permeability 26 , the tendentiously concave regression curves between the antibacterial activity and the LogP or LogD 7.40 also indicate that the cell membrane is probably an important site of flavonoids acting on gram-positive bacteria. Different antibacterial mechanisms of plant flavonoids were reported , such as causing cell-membrane damage, inhibition on various synthase involving the nucleic acid synthesis, the bacterial respiratory chain, or the cell envelope synthesis. However, the results above suggested that the antibacterial activities of these plant flavonoids had no obvious relationship with the specific fragments of their structures, while presented great relationship with their lipophilicities. Simultaneously, the antibacterial activities of plant flavonoids will dramatically increase as the LogP or LogD increases from 2.5 to 4.0 which range the membrane permeability remarkably decrease while the affinity to lipid bilayer greatly increase . According to this, plant flavonoids may not target specific synthases, but more likely to nonspecifically act on the cell-membrane bilayer or the respiratory chain to kill Table 6. Physico-chemical parameters and antimicrobial activities of compounds 55 to 66 15 . a The CLogP values were calculated using software ChemBioDraw Ultra 12.0. b The ACD/Log P and LogD 7.40 values were calculated using software ACD/Labs 6.0. c MIC, minimum inhibitory concentration; MRSA G31 and G47, methicillin-resistant Staphylococcus aureus G31 and G47. d Both MICs of compound 65 against MRSA G31 and G47 were more than 25 μg/mL (61.2 μM). As microdilution broth method was used to test MIC, we set 50 μg/ mL (122.4 μM) as their MICs. www.nature.com/scientificreports/ bacteria. This deduction was indirectly supported by many researches which were reviewed in three paper , such as follows: (1) two mechanisms may be involved the interactions of flavonoids with lipid bilayers, which include the interactions at the membrane interface between the polar heads of phospholipids and the more hydrophilic flavonoids, and the partition of the more hydrophobic flavonoids in the interior of the lipid bilayer 30 ; (2) nonspecific interactions of flavonoids with phospholipids can lead to the changes of the membrane properties 31 ; (3) The increased activities of more lipophilic flavonoids are due to the enhanced membrane affinity of their long acyl chains 32 ; (4) Some lipophilic flavonoids can decrease the fluidity and integrity of cellular membrane to inhibit gram-positive bacteria 33,34 , such as sophoraflavanone G and 3-arylideneflavanones. www.nature.com/scientificreports/ Although many other antibacterial mechanisms acting on various synthase for the nucleic acid or cell envelope syntheses were mentioned in these reviews 4,6 , two facts found from the researches of the cited literature are worth further discussing. First, most flavonoids used for mechanism exploration have the cLogP ranged from about 2.0 to 4.0, and are easy to infiltrate into the bacterial cell, while they present very weak antibacterial activities with the MICs more than 250 μg/mL. Second, most experiments were achieved by the determination of enzyme activities in vitro 35,36 , the molecular docking of flavonoids with various synthases 37 , the proteomics technology without the combination of related experiments and the consideration of first the chicken or the egg 38 . Another thing should be considered is whether some molecules can pass through the cell membrane and infiltrate into the bacterial cell or not. Moreover, previous works indicated the antibacterial activity to gram-positive bacteria was observed only four of fourteen flavonoids, while only four of seven flavonoids with DNA gyrase inhibition showed weak inhibitory activity to gram-positive bacteria 20 . Simultaneously, the authors pointed out that mechanisms other than DNA gyrase inhibition may also play a role in the antibacterial activity. Thereby, the conclusion that some of these flavonoids studied are potent inhibitors of DNA gyrase is worth reconsidering 20 . In fact, this work just right indicated that the inhibitory activity of flavonoids against gram-positive bacteria did not correlate with their in vitro DNA gyrase inhibition to a large extent. This was also supported by previous publication 39 . These together further confirmed that the cell-membrane should be the main region of plant flavonoids acting on Gram-positive bacteria, and which likely involving the disruption or damage of phospholipid bilayers, the inhibition of the respiratory chain or ATP synthesis, or some others. ## Compounds According to the regression equations and above conclusions, the inhibitory activities of flavonoids to grampositive bacteria will increase when the alkyl especially isopentyl were introduced into the structures of flavonoids no matter carbon position it is introduced into. This can be interpreted that the introduction of alkyl would increase the lipophilicity of flavonoids or the LogP, and thereout increase their interactions with phospholipids of cell membrane. However, the introduction of too many alkyls will overmuch increase the LogP of these flavonoids, and which will lead their lipophilicities too large to pass through the hydrophilic region of phospholipid bilayers. This was proved by previous similar work 26,32,40 . On the contrary, the inhibitory activities of flavonoids to gram-positive bacteria will decrease when polar groups, such as hydroxyl and glycosyl, were introduced into their structures. This can be interpreted as that the excessive hydrophilicity of flavonoids will hinder its infiltration into phospholipid bilayers and interaction with hydrophobic region of cell membrane. Based on the physicochemical parameters and MICs of various flavonoids, the regression equations and above conclusions were achieved. For a certain subclass of flavonoids, the regression equations with larger correlation coefficient can be established for their more accurate MIC predictions, and then can be further used for the structural design and optimization to obtain more efficient antibacterial activity. As the inhibitory activities of plant flavonoids against gram-negative bacteria were reported less, it is difficult to draw a statistical conclusion. Considering that the cell envelope of gram-negative bacteria was different from that of gram-positive ones, it is worth further exploring whether the above regression equations and above conclusions are suitable for plant flavonoids against gram-negative bacteria. However, these can provide good references for their related researches. Referring to the above conclusions, the anti-MRSA activities of trimethylhydroquinone, vitamin K 3 and carnosic acid were successfully predicted and verified by our laboratory 9,41 . In conclusion, the MICs of most flavonoids against gram-positive bacteria can be roughly calculated from their physicochemical parameters ACD/LogP or LogD 7.40 , and the lipophilicity is a key factor of plant flavonoids against gram-positive bacteria. Combined with the analyses of previous publications, the results also suggest that the cell membrane may be the main site of plant flavonoids acting on gram-positive bacteria, and which likely involves the damage of phospholipid bilayers, the inhibition of the respiratory chain or ATP synthesis, or some others. Base on this, the inhibitory activities and mechanisms of plant flavonoids to gram-positive bacteria were diagrammatically presented as Fig. 9. ## Methods Information and data. The structures, antimicrobial activities and other related information of plant flavonoids were unsystematically searched from Google academic search engine, and several databases SciFinder, Medline, Elsevier, ACS, ScienceDirect, Wiley Online Library, Springer-Link, and RSC, using keywords flavonoid and antimicrobial, or and antibacterial, and or and anti-MRSA. Furthermore, the relevant references in the obtained literature were also tracked. The structures, antibacterial activities, and other related information of flavonoids were collected from the obtained literature that can provide more than five or more flavonoids. As the antimicrobial activities of a certain compound against different pathogenic strains were varied, compounds reported in different papers were independently collected for the following analyses. Finally, the structures of selected compounds were drawn using software ChemBioDraw Ultra 12.0. ## Simulation calculation of physicochemical parameters. The physicochemical parameters Gibbs energy, LogP, CLogP, MR, CMR and tPSA were calculated using software ChemBioDraw Ultra 12.0. Moreover, another software ACD/Labs 6.0 was also used for the calculations of physicochemical parameters LogP, LogD 7.40 and solubility (SolDB). ## Data analysis and correlation establishment. The physicochemical parameters and antibacterial activities of flavonoids reported in the same paper were respectively listed in a table, even those of the same compound. The regression analyses between the calculated values of each parameter and the antimicrobial activities (expressed as MICs) of all compounds in a table were respectively performed using Microsoft Excel software. It is noting that compounds without related antimicrobial information were not considered for the regression analyses, while they can be used for the following discussion. The physicochemical parameters significantly correlating with the antimicrobial activities were selected for the further analyses of correlations between the physicochemical parameters and antimicrobial activities of flavonoids. Verification. Some other flavonoids were searched from above several databases, and the chemical structures of various flavonoids presented in previous publications were also drawn using software ChemBioDraw Ultra 12.0. The physicochemical parameters LogP and LogD 7.40 of these flavonoids were respectively calculated by software ACD/Labs 6.0, and then their antimicrobial activities (MICs) were respectively predicted using the above regression equations. Comparing with the predicted MICs with the determined one, the regression equations can be verified.

chemsum

{"title": "Antibacterial activity and mechanism of plant flavonoids to gram-positive bacteria predicted from their lipophilicities", "journal": "Scientific Reports - Nature"}

thermal-electrochemical_parameters_of_a_high_energy_lithiumion_cylindrical_battery

## Abstract: To accurately predict the lifetime of commercial cells, multi-physics models can be used, however the accuracy of the model is heavily reliant upon the quality of the input thermodynamics and kinetic parameters. The thermal properties and the variability of the transport and thermodynamic properties with temperature and state-of-charge (SoC) in a high energy 21700 cylindrical cell were measured. The parameters are used in a DFN and 0D thermal model, and the model was tested against experimental data from the commercial cell. The results demonstrate an improved model fit by 27% when including the parameter dependency upon SoC and temperature, compared to without. The maximum power is limited by the negative electrode, which has lower diffusion coefficients and current exchange density over the full SOC window compared to the positive electrode, particularly at 50% and 80% SoC (x=0.45 and 0.85), reflected in high activation energies of up to 60 kJK -1 and low diffusion coefficients of 5 x 10 -13 cm -2 s -1 at 25 °C. At 45 °C, the reaction rate increases to greater than that of the positive, diffusion also increases, 2 x10 -12 cm -2 s -1 , but is still limiting. This work provides for the first time an electrochemical and thermal experimental dataset for a high energy cell, and provides insights into the rate limitations and prediction errors. ## Introduction Lithium-ion batteries are becoming a preferred technology for energy storage, particularly within the automotive industry due to a transition towards electric vehicles. 1,2 Significant improvements in battery technology have been made, including reducing cost and increasing energy density. 3 However, improving battery performance has an impact upon safety considerations due to increased heat generation inside the cell, which in turn increases the probability of thermal runaway. 4,5 Therefore, the thermal characteristics of a cell are an important consideration during cell and pack design. 6 Models can aid this design process by simulating the heat generation and electrochemical behaviour of a battery. 7 Increasingly, physics-based models are being used for predictive purposes, providing insights into the internal states of a battery and more accurate predictions compared to equivalent circuit models. 8 Through the porous electrode theory introduced in the work of Newman and Tiedemann, physics-based electrochemical models became popularised for predicting the internal states of a battery. 9,10 These predictions can be further improved by coupling to a thermal model to capture the thermal-dependency. 11,12 Higher energy density materials have been developed to meet the automotive specifications for lower cost vehicles, these include nickel-rich layered oxides and silicon-doped graphite electrodes. 13,14 The development of new materials and cell types means that these systems need to be parameterised to enable accurate model predictions for these applications. This is because electrodes vary in composition and microstructure, factors that have significant influence on the resulting electrochemical and thermal properties. 15 Commonly used parameter sets for commercial cells are not for high energy systems and do not include the information required to extend to a 3D thermal model. Recent parameterisations of commercial cells only considered batteries with electrodes less than 55 µm. 17,19 Previous work has compared differences in energy vs power cells and their physical and electrochemical properties, these differences identified a need to parameterise high energy cells as literature has focussed on high power. 20,21 Additionally, research that has parameterised thermal-electrochemical models has not involved measuring the specific heat capacities and thermal conductivities of the individual materials needed to describe thermal performance beyond 0D (Table 1). 18,22 The requirements of the different thermal model definitions are as follows (the electrochemical parameters also need to be defined): • 0D Thermal: Activation energy, entropic term, and lumped (volumetric) heat capacity. • 1D Thermal: Thermal conductivity, heat capacities, and 1D cell geometry. • 2D Thermal: 2D cell geometry. • 3D Thermal: 3D cell geometry, tab locations, and inner structure. In this paper we parameterise the LG M50, a cell that with a very high energy density 267 Wh kg -1 , attributed to high electrode coat weight and its composition of Li[Ni1-x-yMnxCoy]O2 (NMC) and SiOy materials. To our knowledge, this is to date the highest energy cell reported in literature, for which parameterisation has been performed, and the only cell to have the thermal characteristics for the electrodes and cell (267 Wh kg -1 ). The parameterisation provides the modelling community with the data to predict thermal inhomogeneities within the cell by detailing the cell anatomy and the thermal transport properties required to extend to 3D. Including the information allows better predictions about battery performance to be made, therefore allowing more efficient thermal management systems to be designed. The M50 has become popular in the academic battery modelling community. The thermal parameters for this cell have not been outlined, meaning research has neglected the thermal behaviour or used properties not specific to the M50. 23,25,26 Presently, the influence of temperature on the electrochemical behaviour has not been included or these properties required have been taken from a different cell. 27 This work details the complete experimental design for a thermal-electrochemical parameterisation of a 21700 cylindrical cell to evaluate the geometric, electrochemical, and thermal properties of the electrodes, separator, and current collectors. In addition to providing the information necessary for a 3D thermal model, the lithium concentration and temperature parameter dependencies are documented to enable more accurate model predictions by accounting for the local variability in performance during cell operation. Models often neglect the effect of lithium concentration and temperature on cell properties, 28,29 despite the parameters being significantly influenced by these variables. 17 This includes the experimental methodology and the mathematical analysis to assess these parameter-dependencies, which can be applied to commonly used NMC and graphite materials. The parameter requirements were based on the most commonly utilised electrochemical model developed by Doyle, Fuller, and Newman (DFN) outlined in Table S2. 30 This paper focusses on presenting parameters relevant to capture 3D thermal behaviour, with equations outlined in Table S3. However, to validate the parameters against experimental data, a pseudo-two-dimensional (P2D) electrochemical coupled to a lumped (0D) thermal model was used. ## Experimental 1.2.1 Teardown Procedure The battery investigated was a 5 Ah M50 21700 cylindrical cell manufactured by LG Chem. This cell utilises nickel-rich NMC811 and SiOy-graphite active materials. To extract the components the cell was discharged and disassembled in a glovebox. The teardown methodology has been described previously and detailed chemical and physical composition can be found in Table S1. 23 During the teardown the gravimetric and volumetric contribution of each component was measured: the jellyroll was weighed immediately after disassembly, and again after the electrolyte had been evaporated to evaluate solvent content. The anatomy of the cell was detailed and the components were separated for individual analysis. To evaluate the mass of electrode coating the current collectors were delaminated. The positive electrode was soaked in N-methyl-2-pyrrolidone (NMP) and sonicated at 70 °C to remove the coating. The negative electrode coating was removed using water. The measured weight of the bare current collectors was used to calculate the total black mass for each electrode. The electrode black mass was used to measure specific heat capacity. For further characterisation, a pristine cell was dismantled, and fresh electrodes extracted. For the electrochemical testing, one side of the coating had to be removed and to measure the thermal conductivity the double-sided electrode was left intact. ## Thermophysical Characterisation Differential Scanning Calorimetry (DSC) The specific heat capacities of the dried composite powders and the separator were evaluated by applying a continuous method on a DSC 1 from Mettler Toledo between 25 °C and 100 °C. The heating rate was set at 10 K min -1 with a sampling interval of 0.1 s. Three repeats of the blank pans, sapphire reference, and the samples were measured. The sample weights were approximately 10 mg. ## Laser Flash Analysis (LFA) The through-plane thermal diffusivity was measured for the electrodes using laser flash analysis (LFA 467 HyperFlash, Netzsch). This measurement was carried out at temperatures of -5 °C, 10 °C, 25 °C, 40 °C, and 55 °C using a nitrogen purge gas. For this measurement samples of 20 mm x 20 mm were used, each sample was measured five times at each temperature. ## Thermal-electrochemical Characterisation To deconvolute the behaviours of the negative and positive electrode a three-electrode configured PAT-Cell (EL-Cell) and a perfluoroalkoxy alkane Swagelok TM half-cell (using lithium metal as the counter electrode) were utilised for the electrochemical testing. The three-electrode cell was comprised of an 18 mm negative and positive electrode, a 21.6 mm double layered separator comprised of 180 µm polypropylene woven layer and a 38 µm polyethylene membrane (EL-Cell), with 100 µl of electrolyte. The half-cell was comprised of a 11 mm working electrode, a 12 mm lithium counter electrode, 12.8 mm Celgard 2325 tri-layer separator (polypropylene/polyethylene/polypropylene) and 50 µl of electrolyte. The electrolyte used was 1 mol dm -3 LiPF6 in ethylene carbonate: ethylmethylcarbonate (3:7, v:v, Soulbrain). The electrochemical protocols were programmed on a VMP3 potentiostat (Bio-Logic). Electrochemical testing was preceded by two cycles of C/20 CC CV charge (CV cut-off was C/50) and CC discharge between 2.5 V and 4.2 V. The C-rate was based upon the discharge capacity for the second cycle. For temperature control a programmable climatic chamber (Temperature Applied Sciences) with an accuracy of ± 1 °C and a fan i.e. forced convection. ## Galvanostatic Intermittent Titration Technique (GITT) GITT was conducted at temperatures of 5 °C, 15 °C, 25 °C, 35 °C, and 45 °C in a three-electrode cell between 2.5 V and 4.2 V. Transients were C/10 CC for 150 seconds and the relaxation period was limited to a duration of 2 hours or when the voltage decay with time was dE/dt < 0.1 mV h -1 . ## Entropy Determination (Potentiostatic method) OCV measurements were carried out at temperatures of 25 °C, 15 °C, 5 °C, -5 °C at SoCs between 0% and 100% (10% intervals) in a three-electrode cell. The cell was initially charged to 100% SoC with C/5 CC CV (C/50 cut-off), then was discharged for 1 h by a C/10 CC step at 25 °C i.e. to 90% SoC. The battery was subsequently allowed to relax for 15 h at the same temperature, after which the thermal cycle (15°C, 2 h; 5 °C, 2 h; -5 °C, 2 h) was applied. This process was repeated until a final SoC of 0% was attained. ## Potentiostatic Electrochemical Impedance Spectroscopy (PEIS) PEIS measurements were conducted at SoCs between 10% and 100% for temperatures 15 °C, 25 °C, 35 °C, and 45 °C in a Swagelok half-cell. The sinusoidal current applied had an amplitude of 10 mV and a 10 mHz -100 kHz frequency range. The data was analysed by fitting to an equivalent circuit model in Zview (Ametek). ## Four-point Probe The electronic conductivity of the positive electrode was evaluated at temperatures 15 °C, 20 °C, 25 °C, 30 °C, and 35 °C using a four-point probe (Ossila Instruments). The positive electrode coating was delaminated using liquid gallium to dissolve the aluminium current collector and obtain the electrode coating undamaged. Small quantities of 1 mol dm -3 hydrochloric acid and deionised water were used to remove the gallium alloy. This methodology has been described in detail previously. 31 To measure the electronic conductivity of the positive electrode a target current of 100 µA was used, the voltage was stepped by 0.1 V until the target current had been reached. ## Model Validation The thermal-electrochemical model was validated for discharge rate capability tests at various temperatures, the temperature during these experiments was monitored to validate the heat generation component of the model. The heat generation was measured using an external thermocouple on the cylindrical cell. Before the experiment there is a two hour rest period to record the initial state. Then the cell was charged at 0.3C with a C/100 current cutoff for the CV step, and discharged at C/10, C/2, 1C, and 2C using a Maccor battery tester. The voltage window used was 2.5 V -4.2 V. Between each charge and discharge there was a two hour rest period. The testing protocol was carried out at 0 °C, 10 °C, and 25 °C with the chamber temperature being measured throughout the experiment. ## Simulations Simulations were conducted in the Python Battery Mathematical Modelling (PyBaMM) software package (using v0.4.0). 32 The equations for the thermal-electrochemical model are summarised in Table S2 and Table S3. In order to solve the model a finite volume scheme was used, with 30 grid points for each electrode and the separator, and 150 grid points for each particle; resulting in a system of 9092 ODEs and 150 algebraic equations. An exponential mesh was used to help with the convergence of the solver. In order to solve the system, a CasADI solver was used. 33 Each simulation of discharge plus relaxation takes 10 to 20 minutes using an Intel Core i7-7660U (2.50GHz) processor and 16 GB RAM. This is because the nonlinear diffusion takes the solver many iterations to converge. The computational time could be reduced by using more sophisticated numerical methods, but this is out of the scope of this work. It should be noted that reduced models such as the Thermal Single Particle Model with electrolyte (TSPMe) 25 yield very similar results with a significant reduction of the computational time. ## Cell Structure The cell structure describes the geometry and the anatomy, this information is needed to resolve a 3D thermal model. 27 The structure can be used to build a complex model to predict inhomogeneities due to detail relating to the internal cell layered-structure, including tab location, winding structure, and gravimetric contributions. The cell is comprised of a of high nickel Li[Ni1-x-yMnxCoy]O2 positive electrode material and a SiOy-graphite negative electrode, with a ceramic-coated polyolefin separator. 23 This information and the specifications provided by the manufacturer are summarised in Table S1. The dimensions of the cell including and tab locations are illustrated in Figure 1, with the tabs are located at opposite ends; from the top view, the positive tab is positioned 90° clockwise from the negative tab. The positive tab is 7.0 cm in length and is visible from both sides, whereas the negative tab is 5.0 cm in length and only visible from the side that has the electrode coating facing the inside of the jellyroll. Photographs of these tabs can be found in the Figure S2 with their dimensions are outlined in Table 3. The cell jellyroll is electrically isolated from the cell casing by a thin plastic layer (thickness 20 µm) in the radial direction. In the axial direction, there is a thick plastic disc (thickness 0.22 mm) at the bottom and a perforated fibrous membrane (thickness 0.2 mm) at the top (Figure S1). These components significantly lower the axial thermal conductivity by forming a barrier for heat conduction through the current collectors to the casing. The jellyroll is comprised of an electrode stack of two separators, a double-side coated positive electrode, and a double-side coated negative electrode. The copper current collector was not completely coated (Figure S1). The jellyroll consisted of ca. 23 windings of the electrode stack, with the schematic of a single stack illustrated in Figure 2. The gravimetric and volumetric contributions of the components are summarised in Table 4. Immediately after disassembly, all the components of the cell were measured to account for solvent lost via evaporation. The jellyroll mass included any electrolyte still present and that lost during disassembly. The remaining solvent was evaporated to obtain a total electrolyte mass of 3.33 g. For the electrolyte assumed in this investigation, 1 mol dm -3 LiPF6 in 3:7 EC:EMC (v:v), this corresponded to a salt content of 0.47 g, and a total electrolyte mass of 3.80 g. It was assumed this salt content was equally distributed within the separator and electrode coatings, these contributions were subtracted from the components. The mass of the separator, positive electrode, and negative electrode windings were 1.96 g, 28.96 g, and 22.36 g. After delaminating the coatings, the black mass of the individual components was evaluated. The volumetric contribution of each component could then be calculated, these are outlined in Table 4, with the densities of copper, aluminium, the stainless steel cell casing, and the electrolyte taken from literature. 34,35 ## Thermophysical Characterisation This section outlines the properties needed to accurately predict thermal transport and local temperature inhomogeneities in a cell. Information including the specific heat capacities and thermal conductivity of each material is often not outlined or measured in cell parameterisations. 16,17 This information can be combined with details of cell anatomy to construct an accurate 3D thermal model. Table 7 describes the thicknesses, densities, specific heat capacities, and thermal conductivities for all the individual cell components. This allows model to capture this detail rather than using macro-properties for thermal transport. The specific heat capacity of the active materials was measured from the composite powders, whereas the thermal conductivity could not be measured directly so thermal diffusivity is measured first to calculate it. The thermal properties of the components should be considered with the presence of electrolyte as the commercial cell is comprised this way and the electrolyte significantly effects thermal transport. 19,36 However, due to the volatility of the electrolyte, ex situ measurements could not be carried out for the separator or electrode and in situ measurements do not allow the deconvolution of the thermal properties of individual components-this is needed to enable physics-based models to predict thermal inhomogeneities. Here we outline a method to calculate the thermal properties of the wetted electrode using experimental data from the extracted electrodes. It is important to directly measure the thermal properties of the electrodes as microstructure significantly effects heat transport and generation within the battery. 15,37 However, due to difficulty extracting any usable quantity of electrolyte, the thermal properties of a known electrolyte 1 mol dm -3 LiPF6 in 1:1:1 EC:EMC:DMC (v:v) was used (this is dissimilar from the electrolyte used in the electrochemical tests and model which is 1 mol dm -3 LiPF6 in 3:7 EC:EMC (v:v)). 35 The heat capacity and thermal conductivity of this electrolyte were used in the following section as LiPF6 in carbonate electrolytes are assumed to have similar properties and the model is not significantly sensitive to electrolyte parameters, the information is summarised Table 7. 24 The following measurements were carried out on materials extracted from a cell fully discharged to 2.5 V i.e. 0% SoC. The effect of lithiation on the thermophysical properties of these materials was not considered here due to the difficulty in maintaining air stability at higher states of charge. It should be noted that the state of lithiation does significantly influence heat transport properties as described in a previous work, it was shown that the thermal diffusivity only changed by 15% for LiNi1-x-yMnxCoyO2 across the whole lithium stoichiometry range. 38 In practice, the change would be less significant as the materials are never fully lithiated/delithiated. ## Specific Heat Capacity The specific heat capacity describes the heat energy required to raise a material by a unit of energy, this relates to how easily the temperature rises within a cell and helps the model to predict temperature gradients. Selection of materials with a high specific heat capacity means more energy is needed to raise the internal temperature of a cell reducing the presence of internal gradients. This property was measured for the delaminated electrode powders and the separator as it is not dependent on microstructure (Figure 3). The values for the black masses provide an aggregate heat capacity for the active material, binder, and carbon black. The reported specific capacities corroborate the values reported previously for a NMC/graphite battery. 19 Figure 3. Specific heat capacities for the NMC composite powder (red), the graphite composite powder (black), and the separator (purple). The specific heat capacities for the NMC powder, graphite powder, and separator are represented by Equations [S1] to [S3]. To capture the change in specific heat capacity due to temperature in the model a function is needed to describe this relationship. The Debye theory of specific heat states that the specific heat of solids is proportional to 𝑇𝑇 3 if the temperature is below the Debye temperature, which is reported to be 141 °C for graphite and between 136 °C -204 °C for the transition metals in NMC. 39 We can therefore fit the specific heat capacity against temperature data using third order polynomial according to Debye theory, it is important to capture these phenomena using scientifically-robust descriptions rather than arbitrary functions. The fits for the electrode powders and separator are illustrated in Figure 3 with the comparison to raw data in Figure S3. The current collectors also make a notable proportion of the cell mass (>7%) so the variability of the specific heat capacities can also be captured to better predict local inhomogeneities in temperature. The heat capacities across the normal temperature operating range of a battery for aluminium and copper were taken from literature and fitted to third order polynomials to capture the parameter-dependency using Debye theory (where 𝑇𝑇 is in Kelvin and units are J kg -1 K -1 ). These relationships are outlined in Equations [S4] and [S5]. 34 The decreasing NMC specific heat capacity above 80 °C could be due to decomposition of the binder within the sample, as there are no phase transitions within this range. 40 However, this is beyond the normal operating range of a battery cell so this information would not be captured by the simulations. The effect of the electrolyte is included by calculating the volumetric heat capacity for the wetted electrode from the volumetric heat capacities of each bulk material as shown in Equation : 𝜃𝜃 ̅ is the averaged volumetric heat capacity; while 𝜌𝜌, 𝐶𝐶 𝑝𝑝 , and 𝜀𝜀 are the density, gravimetric heat capacity and volume fraction for the electrode and electrolyte, respectively. The densities and the specific heat capacities of the electrolytes and the dry electrodes can be found in Table 7, and the electrolyte volume fraction in the positive and negative electrode is 0.335 and 0.25 respectively. Similarly, the averaged density 𝜌𝜌̅ of the saturated electrodes can be calculated from the densities of the electrode solid and liquid components: The volumetric heat capacity for the wetted positive electrode and wetted negative electrode are 3335500 J m -3 K -1 and 1743680 J m -3 K -1 . Then the gravimetric heat capacity of the wetted electrodes is calculated by dividing the averaged volumetric heat capacity (calculated above) by the averaged density of the saturated electrodes (3700 kg m -3 and 2060 kg m -3 for positive and negative electrode). This provides a gravimetric heat capacity of 900 J kg -1 K -1 and 845 J kg -1 K -1 for the wetted positive electrode and negative electrode (without the current collectors) at 25 °C. As the specific heat capacity of the dry electrode bulk (powder) varies with temperature, this dependency can be included for the wetted electrode: Here 𝜃𝜃 ̅ and 𝜌𝜌̅ are calculated from and respectively, the other values are summarised in Table 6. Thermal Conductivity The thermal conductivity describes the ability of a material to conduct heat and heat transport within the cell. This is influenced significantly by the anisotropic structure of a cell and means that the thermal conductivities of the individual components can be evaluated to predict thermal gradients. For the electrode materials the thermal conductivity λ was obtained from the following Equation : Here 𝐶𝐶 𝑝𝑝 , is the specific heat capacity, 𝜌𝜌 is the density, and 𝛼𝛼 is the thermal diffusivity. It was not possible to measure the thermal diffusivity of the separator as it is too thin to be characterised using laser flash analysis (LFA). The electrolyte volume fraction of traditionally used PP/PE/PP separator does significantly affect the thermal conductivity, therefore the thermal diffusivity is assumed to be similar across commercially used separator materials. 42 The thermal conductivity of a wetted separator with similar properties is taken from literature, this is 0.3344 W m -1 K -1 . 43 The density and the volumetric specific heat capacity of the wetted separator can be evaluated using Eq. . At 25 °C, 𝜌𝜌, 𝐶𝐶 𝑝𝑝 , and 𝜀𝜀 𝑙𝑙 are 946 kg m -3 , 1700 J kg -1 K -1 and 0.47 for the dry separator. The volumetric heat capacity is 1745970 J m -3 K -1 . This corresponds to a density and gravimetric specific heat capacity of 1620 kg m -3 and 1080 J kg -1 K -1 for the wetted separator. Therefore the heat capacity for the wetted separator is represented by Equation , where 𝐶𝐶 𝑝𝑝,𝑠𝑠 is defined by Equation [S3], and the other values are summarised in Table S5. For the electrodes, this property cannot be determined from the powder as it is influenced by microstructure. 15 Instead, the through-plane thermal diffusivity of the double-side coated electrode is measured. The through-plane thermal diffusivity and calculated specific heat capacities of the double-sided electrodes are illustrated in Table 6. The overall specific heat capacity of the electrodes were calculated using the heat capacities and mass fractions (Figure 3) of the black mass and current collectors. The area density of the positive electrode was attributed to 8.1% aluminium and 91.9% NMC coating, whereas the area density of the negative electrode was attributed to 26.6% copper and 73.4% graphite coating. The thermal diffusivity decreases with temperature for both electrodes, although as the specific heat capacity relationship with temperature is inversely proportional, the thermal conductivity does not change significantly with temperature. The through-plane thermal conductivity of the graphite-SiOy electrode increases slightly, whereas the NMC electrode decreases slightly. The thermal diffusivity and the specific heat capacity of the double-sided electrodes are used to calculate the overall thermal conductivity for the positive and negative electrode respectively. As the electrode is a layered material the conductivity of the electrode lamina can be deconvoluted. The conductivity in the in-plane direction is the harmonic average of the conductivities (weighted by their thickness). In this case, there were three layers, two layers of the electrode laminate and one layer of the current collector. Therefore, the thermal conductivity perpendicular to the electrode can be written as: 4 The thickness of the electrode 𝐿𝐿 𝑒𝑒 and current collector 𝐿𝐿 𝑐𝑐𝑐𝑐 , the conductivity of the current collector λ 𝑐𝑐𝑐𝑐 and the total conductivity λ 𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑡𝑙𝑙 are known, therefore the conductivity of the electrode laminate λ 𝑒𝑒 can be calculated. This corresponds to 0.807 W m -1 K -1 and 3.793 W m -1 K -1 at 25 °C for the NMC811 and graphite composite laminates respectively, values measured between -5 °C and 55 °C are reported in Table 6. These thermal conductivities are in a similar range to reported values of 1.31 to 2.62 W m -1 K -1 and 0.68 to 2.62 W m -1 K -1 for NMC811 and graphite respectively. 42 The thermal conductivity measurements were for the dry porous electrode, although the properties of the wetted electrode are required. The measurement included the contribution of the nitrogen purge gas and not the electrolyte, however N2 exhibits similar thermal conductivity (0.025 W m -1 K -1 ) to the electrolyte (0.03 W m -1 K -1 ) so the thermal conductivity of the wetted and dry electrode are assumed the same. The graphite and NMC thermal conductivities were fitted as interpolants in the model as low order polynomials did not provide good agreement. The thermal conductivity for aluminium and copper current collectors were taken from literature as in batteries they appear in their pure elemental form and the thermal properties do not vary. The thermal conductivity of aluminium does not change appreciably within the operating range of the battery so the constant value of 237 in W m -1 K -1 is used. However, the thermal conductivity for the copper decreases with temperature. The dependency of thermal conductivity on temperature is fitted to a third degree polynomial to capture this change in the model (where 𝑥𝑥 is in Kelvin and in λ is in W m -1 K -1 ), see Equation S6. 34 These measurements allow prediction of thermal inhomogeneity within a cell, but combining these values do not provide an accurate value for the overall cell thermal conductivity. This is because of the cumulative error in measuring several materials separately and by not accounting for the thermal contact resistances between the layers in a cell. For accurate 3D thermal models and to predict the performance of larger battery systems the macro-thermal properties need to be measured. Cylindrical cells are a composite and not a pure material, therefore the conductivity of each cell type has to be measured to improve prediction of the cell temperature. Effective Thermal Properties Lumped thermal models require the effective thermal properties of the jellyroll. The thermal properties of the jellyroll can be measured directly, or they can be calculated using the individual heat capacities of the components. The jellyroll is considered to be a homogenous material so it has a uniform temperature, and the thermal conductivity can be neglected. The jellyroll excludes the casing and the miscellaneous components such as the hard plastic and perforated fibrous membrane at the axial sites. An overall specific heat capacity of 887 J kg -1 K -1 (2.38•10 6 J K -1 m -3 ) was calculated using the gravimetric contributions and the heat capacities of the individual components in Table 7. The 2-sided electrodes relate to the whole double sided coating, including the current collector and both coated electrodes, whereas the coatings relates to the single sided coating thickness only (not including the current collector). The total specific heat capacity is similar to value reported previously, 850 J kg -1 K -1 for a jellyroll with density 2400 kg m -3 . 28,44 34 16 m 2702 l 876 l 2366952 l 100 l 237 l Copper foil 34 12 m 8933 l 383 l 3421339 l 117 l 401 l Separator (dry porous) 43 12 ## Thermal-electrochemical Characterisation In this section, the temperature dependencies of the parameters were outlined to capture the influence of temperature on electrochemical behaviour, batteries undergo self-heating even at moderate C-rates so this information is essential to accurately predict these changes. These dependencies were modelled using an Arrhenius relationship: Here 𝑛𝑛 represents a temperature dependent parameter, 𝑛𝑛 𝑟𝑟𝑒𝑒𝑟𝑟 is its value at reference temperature namely 𝑇𝑇 𝑟𝑟𝑒𝑒𝑟𝑟 , 𝐸𝐸 𝑛𝑛 𝑡𝑡𝑐𝑐𝑡𝑡 is the activation energy, 𝑅𝑅 is the gas constant, and 𝑇𝑇 is the temperature in Kelvin. The activation energy can be evaluated by finding the slope 𝑐𝑐 of plot ln (𝑛𝑛) vs. 1/𝑇𝑇: To describe the heat generation of the cell, the entropic coefficients were measured to capture the reversible heat and the irreversible heat is accounted for in the DFN model definition. The effect of voltage hysteresis is often not considered in models, these properties are therefore evaluated during cell discharge only i.e., positive electrode lithiation and negative electrode delithiation. It has been observed that the hysteresis effect is minimal for parameters including the entropic coefficients and diffusion. 23,45 The properties are mapped as a function of lithium concentration, the stoichiometric ranges were mapped previously using half-cells. 23 For the positive electrode this corresponds to crystal lattice lithium concentrations of x=0.2567 to 0.9072 and for the negative electrode, x=0.0279 to 0.9014 (Figure 4). However, after a battery is discharged to a specific SOC; electrode stoichiometry varies with time as lithium diffuses in/out of the active materials. To correct for this, the open circuit voltages during the test procedure were aligned to the OCV data (Figure 5). It should be noted, that the capacity of the negative electrode is oversized to the positive electrode for safety, reducing lithium plating and ensuring enough capacity for the lithium removed from the positive electrode. 46 The positive electrode capacity is not defined by the complete stoichiometric range, but a lithium content which is reversible in the voltage window specified, e.g. LixMO2 where 0.26≤x≤91. In this case NMC811 has a practical upper voltage of 4.4 V. 47 ## Diffusion The activation energy of solid-phase diffusion can be calculated by repeating GITT at different temperatures. 18,48 This enables us to find the concentration/temperature dependency of solid-phase within the active material. The diffusion coefficients were analysed from the transients and steady-phase regions during GITT. The analytical approach involves evaluating changes in lithium concentration at the particle surface as a function of time. This can be described by an expression relating to the Sand equation: 49 Here 𝐼𝐼 is the applied current, 𝐹𝐹 is the Faraday constant, 𝐹𝐹 is the effective area of the porous electrode-electrolyte interface, and 𝑐𝑐 𝑠𝑠 𝑚𝑚𝑡𝑡𝑚𝑚 is the maximum concentration in the electrode. 𝐹𝐹 can be calculated from the total number of particles in the electrode and the average surface area of a particle: 𝜀𝜀 𝑡𝑡𝑐𝑐𝑡𝑡 is the active material volume fraction, 𝜕𝜕 𝑒𝑒𝑙𝑙𝑒𝑒𝑐𝑐𝑡𝑡𝑟𝑟𝑡𝑡𝑒𝑒𝑒𝑒 is the electrode volume, 𝜕𝜕 𝑝𝑝𝑡𝑡𝑟𝑟𝑡𝑡𝑝𝑝𝑐𝑐𝑙𝑙𝑒𝑒 is particle volume, 𝐹𝐹 𝑝𝑝𝑡𝑡𝑟𝑟𝑡𝑡𝑝𝑝𝑐𝑐𝑙𝑙𝑒𝑒 is the particle surface area. For this the average particle sizes of 5.22 µm and 5.86 µm were used for the positive and negative electrode respectively, these parameters were measured previously using SEM. 23 For the analysis of the positive electrode diffusivity it is assumed 𝐹𝐹 remains unchanged as NMC volume expansions is negligible, but for the negative electrode we account for the volume expansion of the graphite-SiOy. 23,50,51 This involves applying a linear scaling factor across the diffusivity to capture the 4x volume expansion across the SOC range of the battery. 23 The solid-phase diffusion coefficients were measured using GITT in a three-electrode cell comprising of the NMC811 positive electrode and the graphite-SiOy negative electrode (Figure 6). The three-electrode set-up probed a limited stoichiometric range for each electrode, which was near equivalent to that observed under operation in the cylindrical cell. The diffusion coefficients vary significantly with lithium concentration, distinct regions can be identified in these profiles and attributed to distinct thermodynamic phases. 52 For the NMC-based electrode these phases relate to the different crystal structures of LiNi0.8Mn0.2Co0.2O2 e.g. hexagonal and monoclinic. 47,53 For graphite-SiOy, the Li insertion stages of graphite are responsible for the changes in solid-phase diffusivity e.g. I to IV. 52,54 For the temperature range 5 °C to 45 °C, the diffusivity of the NMC811 was approximately 10 -10 cm 2 s -1 to 10 -11 cm 2 s -1 and for graphite-SiOy it was 10 -14 cm 2 s -1 to 10 -10 cm 2 s -1 , see Figure S4 and S5 respectively for the raw data. The solid-phase diffusion coefficients for graphite-SiOy between 10 -14 cm 2 s -1 to 10 -12 cm 2 s -1 are underestimates due to difficultly observing the minor voltage changes that occur during the transient and relaxation periods of GITT for the plateau regions in graphite. In plateau regions the reaction overpotentials dominate and produce inaccuracies for the solid-phase diffusivity. The diffusion coefficients of NMC materials have been reported in the range of 10 -10 cm 2 s -1 to 10 -8 cm 2 s -1 previously, corroborating fairy well with the results here. 55,56 However, the diffusion values for SiOy and graphite have had values reported much higher than observed here, between 10 -11 cm 2 s -1 to 10 -8 cm 2 s -1 and 10 -10 cm 2 s -1 to 10 -8 cm 2 s -1 for the materials respectively. 51,57,58 Diffusion coefficients are often reported over several orders of magnitude due to differences in experimental set-ups and analysis. The solid-phase diffusivity often has to be tuned to a higher value for to provide reasonable simulation values as if the low diffusivity values are included in the fits, the simulations either do not converge or provide unrealistic results. This tuning was carried out here and the fitted profiles are illustrated in Figure 6. Equation was used to account for the dependency of solid-phase diffusivity on stoichiometry, this equation is illustrated in Figure 6. The equation is a linear combination of Gaussian functions and captures the features of the profile accurately, meaning non-linear solid-phase diffusivity can be captured by the simulations (the fitting parameters can be found in Table S7): These fits are later used in the model and shown in Figure S4 and S5. Figure 6 also describes the change in diffusivity across the measured stoichiometric range as a function of temperature, demonstrating the Arrhenius-type behaviour of this parameter. The average value of solid-phase diffusivity of the 'fitted' curves for the graphite-SiOy is lower than the experimental average for about a factor of three, because we are fitting to the log10(D) rather than D. The reason not to fit to D directly is because that would not capture the features in low diffusivity zones, plus adds the risk of obtaining negative values for diffusivity. Fitting to to the log10(D) ensures diffusivity is always positive and captures the features at different orders of magnitude, at the expense of underestimating the average diffusivity (especially in the negative electrode). Therefore, a scaling factor of 3.03 for the graphite-SiOy diffusivity was introduced in the diffusivity used in the final model so the fitted average diffusivity matched the experimental average diffusivity across all temperatures, as shown in Figure 6 (bottom right). An Arrhenius relationship (Eq. ) can be used to find the activation energies across the stoichiometric range of each electrode (Figure 7). Similarly to the solid-phase diffusivity, the activation energies are influenced by thermodynamic phases and their transitions. For NMC the activation energy is in the range of 10 kJ mol -1 to 20 kJ mol -1 . This compares to reported values of 15 kJ mol -1 to 30 kJ mol -1 for the lower nickel content NMC electrodes. 59 For graphite-SiOy the activation energy has a wider range of 5 kJ mol -1 to 60 kJ mol -1 . Similar to values up to 50 kJ mol -1 for graphite reported by Ecker et al. 16 For graphitic materials the consensus is that values are generally between 20 kJ mol -1 and 40 kJ mol -1 . The range in reported values is attributed to the variability and high uncertainty for the analytical approaches used in determining the diffusion coefficients. 52 Significant temperature dependence is observed for the diffusivity between stoichiometries of x=0.3 to 0.5 and at approximately x=0.8, corresponding to the plateau regions of graphite. Previously the exchange current density and its activation energy were evaluated at a single stoichiometry. 23 Here, the exchange current density and its dependency were mapped on temperature and lithium concentration at various stoichiometries using a half cell (Figure 9). The exchange current density 𝑗𝑗 𝑡𝑡 can be evaluated by measuring the charge-transfer resistance 𝑅𝑅 𝑐𝑐𝑡𝑡 during EIS: Here 𝑅𝑅, 𝐹𝐹, and 𝐹𝐹 are the gas constant, the electrode-electrolyte interfacial area, and the Faradaic constant. For these calculations S was calculated from the geometrical electrode volume 𝜕𝜕, the active material volume fraction 𝜀𝜀 𝑡𝑡𝑐𝑐𝑡𝑡 , and particle radius: This was determined to be 3.266076•10 -3 m 2 and 4.151297•10 -3 m 2 for the positive and negative electrode respectively. The Nyquist plot for the positive electrode only shows one semi-circle that can be attributed to the charge transfer process. This is because the SEI resistance shares a similar time constant with the double layer and therefore is difficult to visually discern the phenomena (see Figure S6). However, two RC elements should be included in the equivalent circuit model to account for the charge transfer and SEI resistance. The Nyquist plots are fitted to an equivalent circuit model (Figure 8) to evaluate 𝑅𝑅 𝑐𝑐𝑡𝑡 and then used to determine 𝑗𝑗 𝑡𝑡 from Eq. . The exchange current density demonstrates a dependency on lithium concentration that can be described by a form of the Butler-Volmer equation (see Figure 9): 16 Here 𝑐𝑐 𝑠𝑠 and 𝑐𝑐 𝑒𝑒 refer to lithium concentration in the solid lattice and electrolyte, respectively, 𝑐𝑐 𝑠𝑠,𝑚𝑚𝑡𝑡𝑚𝑚 is the maximum concentration in the electrode particles and 𝑐𝑐 𝑒𝑒0 is the reference concentration in the electrolyte. The parameter 𝑘𝑘 0 is the reference current of the reaction. As shown in Figure 9, the exchange current density for the NMC811 positive electrode was measured at 25 °C between 0.5•10 -4 A cm -2 and 3.0•10 -4 A cm -2 with a mean value of 2.01•10 -4 A cm -2 . For the graphite-SiOy electrode the exchange current density was measured at 25 °C between 2.0•10 -5 A cm -2 and 9.0•10 -5 A cm -2 with a mean value of 7.1•10 -5 A cm -2 . The observed trends are similar to the results reported by Ecker et al. who reported a value of 2.23•10 -4 A cm -2 and 7.05 10 -5 A cm -2 corresponding to x=0.5 in Lix(Ni0.4Co0.6)O2 and LixC6. 16 The Butler-Volmer equation (Eq. ) fits the negative electrode exchange current density vs stoichiometry well, this fit is worse for the positive electrode (R 2 =0.63) and a semi-ellipse with an exchange current value of 0 at stoichiometries at ~ 0.2 and 0.9 (at 25 °C) provides a better fit (R 2 =0.9, Fig S7). The question arises whether using the stoichiometry of 0 to 1 to represent degree of lithiation with respect to the molar lithium concentration within the crystal structure is the correct assumption to use in the model. There is most probably 'inactive' or non-mobile lithium below x=0.2 in this case as observed in Figure 9, where it is impractical for LixNi0.8Mn0.1Co0.1O2 to be delithiated further due to collapse of the crystal structure. 47 There is little sensitivity in this particular parameter, 63 and therefore have assumed Butler-Volmer kinetics over the entire lithium stoichiometry in this work which is consistent with the model inputs. The activation energy of the exchange current density was evaluated using an Arrhenius relationship (Equation ) across the stoichiometric range of each electrode (Figure 10). The activation energies for the positive electrode range between 20 kJ mol -1 to 50 kJ mol -1 and for the negative electrode between 45 kJ mol -1 to 65 kJ mol -1 . The mean values for the Ni-rich and graphite based electrodes were 31.1 kJ mol -1 and 54.8 kJ mol -1 respectively. Ecker et al. reported activation energies of 43.6 kJ mol -1 for Li(Ni0.4Co0.6)O2 and 53.4 kJ mol -1 for graphite. 16 Jow et al. reported activation energies for graphite and NCA as 68 kJ mol -1 and 50 kJ mol -1 respectively. 64 Smart et al. reported activation energies for graphite and Li(Ni0.8Co0.2)O2 with different electrolyte systems in the ranges 45 kJ mol -1 to 60 kJ mol -1 and -34 kJ mol -1 to 48 kJ mol -1 . 65 Similar ranges have been reported for graphite and Li(Ni0.5Mn0.3Co0.2)O2, these were 56 kJ mol -1 to 72 kJ mol -1 and 58 kJ mol -1 to 69 kJ mol -1 . 66 Previously reported values for this cell was 17.8 kJ mol -1 and 35 kJ mol -1 for NMC and graphite respectively. 23 However, these activation energies are appreciably lower than the values reported here and for similar materials. In this case, the parameter table is updated with the newly evaluated activation energies that have been measured at various lithium stoichiometries and corroborate with literature. The dependency of the exchange current density on temperature and lithium concentration can be described as: ��. Here 𝑘𝑘 0 is the reference current, 𝛼𝛼 is the activity coefficient, and 𝐸𝐸 𝑡𝑡 is the activation energy; and the variables are the stoichiometry and temperature. Because the stoichiometry is defined as 𝑥𝑥 = , this equation combines Butler-Volmer (Eq. ) and Arrhenius behaviours (Eq. ). These values are outlined in Table S8. ## Electronic Conductivity The electronic conductivity is also a temperature-dependent process, however intrinsic material properties determine whether there is a corresponding activation energy. The semi-metallic properties of graphite relate to an inversely proportional relationship with temperature and electronic conductivity. However, in the normal operating temperature range of a battery, the change in graphite electronic conductivity is considered negligible. In contrast, NMC exhibits semiconducting properties (owing to its non-zero band gap energy) this causes electron conduction to be significant dependent on temperature. As the electron conduction in NMC is a thermally activated process the corresponding activation energy can be evaluated using an Arrhenius type relationship (Equation ). 67 The NMC811 electrode was extracted from a cell that was discharged to 2.5 V. This relates to a lithium content of approximately x=0.9. At this state of lithiation, the solid-phase electronic conductivity was evaluated to be 0.847 S m -1 at 25 °C. This value is four times higher than the 0.18 S m -1 activation energy calculated by Chen et al., demonstrating that using liquid gallium to delaminate the electrode rather than adhesive tape preserves the electrode structure. The corresponding activation energy for the positive electrode electronic conductivity was determined to be 3.5 kJ mol -1 (Figure 11). The effect of lithium concentration was not evaluated due to the stability of the partially lithiated NMC materials in ambient conditions. Elsewhere, this relationship has been investigated previously by Amin et al. by pelletizing the pure active material and conducting EIS measurements in an electrochemical cell. 56 Amin et al. studied NMC532 and NMC111 at lithium stoichiometries between x=0.25 and x=1.0. 56 For these materials the activation energy of the electronic conductivity decreased from 40 kJ mol -1 to 4.8 kJ mol -1 and 46.3 kJ mol -1 to 9.6 kJ mol -1 respectively. The latter values for the lithiated materials are similar to the value reported in this paper. For states of lithiation below x=0.25 the material exhibits metallic properties and electronic conduction is not thermally activated. The electronic conductivity is the least sensitive parameter in the DFN model and therefore it is less critical to describe its dependency on lithium concentration. ## Entropic Term There are several sources of heat generation in batteries. 37 The irreversible and reversible heat components are considered important as the active materials' dominating heat generation. 68 For high C-rates (most scenarios) more than half of the heat generation can be ascribed to irreversible heat, known as ohmic heat loss. 69 At low C-rates (1C or less) the reversible heat contribution from the material phase changes becomes more significant, this heat generation is due to the entropy changes that occur as a result of intercalation reactions, and this property depends on the internal temperature and OCV of the system. The entropy change can account for over half the total heat generated at the rates typically used in electric vehicles. 70,71 Parameterisations that outline activation energies do not characterise the reversible heat of the battery, this is important to predict temperature and the influenced electrochemical performance correctly. 16,18 The change in entropy ∆𝐹𝐹 can be determined through the slope of the OCV with temperature: 72 The entropic term is measured using a potentiostatic method which involves measuring the dependency of OCV on temperature (Figure 12). 73 The temperature is changed three times at each lithium concentration. This provides four OCVs at the different temperatures; the entropic term can be calculated from the gradient of a line through the points. This process is repeated to map the stoichiometric range of the electrodes. It can take many hours to attain OCV at a particular SOC and temperature, therefore the thermal stability of the electrolyte needs to be considered when choosing the temperature regime. This is because at high temperatures due to electrochemical instability, particularly at the graphite interface, it is more difficult to attain OCV. 74 Choosing a lower temperature regime avoids instabilities, while allowing the thermodynamic behaviour to be measured. Figure 12. Potentiometric profile illustrating the initial SOC change, followed by a long period to attain OCV and a temperature-cycling regime to observe voltage change (left). The entropic term is calculated from the gradient of the OCV points at each temperature (right). Using a three-electrode configuration for the experiment allows us to measure the entropic terms for SoCs between 0% and 100%, see Figure 13. This means that the entropic term for both electrodes is mapped at the same SOCs for both electrodes. At low states of lithiation the graphite OCV decreases as temperature increases, this corresponds to a ∆𝐹𝐹 < 0. As the graphite electrode is delithiated the entropic term becomes positive. This change occurs at x = 0.6. These observations agree with research by Reynier et al. on a pure graphite electrode. 75 This suggests that less than 10wt% of SiOy has a minimal effect on the entropic term. This results are also consistent with the entropic term of a silicon-graphite material that was reported for lithium stoichiometries less than 0.7. 76 The entropic term of the NMC electrode is negligible at several states of lithiation. For the other states of lithiation it does not show an appreciable value. This means the full cell behaviour is dominated by the negative electrode and the trend is the same, although the opposite magnitude. This is due to the definition of the full cell potential (Ecell = Ewe-Ece). The reversible heat generation in the cell is determined by the graphite-based electrode. These values are in good agreement with published results; the entropic term of NMC-type electrodes have been shown to be negligible in comparison to other positive electrode chemistries. 71 The variation of entropic term with stoichiometry has been captured by fitting functions to the experimental data. For the negative and positive electrode these functions are (fitting parameters outlined in Table S9): However, the function for the negative electrode entropic term was chosen to exclude the points at intermediate stoichiometries. This is because including these points inadequately describes heat generation at 0.5C, see Figure 14. This is despite the values reported here being similar as previously reported for this material, it is not clear why a discrepancy arises when the negative entropic coefficients for the negative electrode are included. 22 Figure 13. Entropic term and polynomial fits for the negative electrode (left) and positive electrode (right). This term has been represented by high order even polynomial functions in the past, however this fitting is often inadequate outside the stoichiometric range. 77 It is difficult to predict the entropic term value outside the measured range, for this reason it is assumed the parameter tends to zero, rather than assigning it a non-zero value that may overestimate heat generation. This section provides the information needed to construct an accurate thermally-coupled electrochemical model by outlining the activation energies and the reversible heat of the battery. This information is often not measured in parameterisations and is critical to predicting the internal temperature and its influence on thermal performance. 16,17 The methodologies also describe the experimental and mathematical approach to quantify the parameter-dependencies of several electrochemical parameters, enabling the changes in performance during battery operation to be documented. The methods can be applied similarly to the widely-used materials, graphite and NMC. ## Electrolyte Properties Ion-transport models for concentrated binary electrolyte solutions depend on the ionic conductivity, ionic diffusivity, and the transference number. 78 The thermodynamic factor (TDF) is also required to describe the thermodynamic behaviour of the electrolyte system, this parameter is dependent on the mean molar activity coefficient. To be consistent with the electrochemical parameterization of this cell it is assumed the electrolyte was 1 mol dm -3 LiPF6 in EC:EMC (3:7, v:v). 23 This was assumed to the difficulty determining the electrolyte composition. The temperature and concentration dependence of the electrolyte properties has been determined previously by Gasteiger et al. 79 The dependencies for the ionic conductivity, ionic diffusivity, thermodynamic factor, and the transference number can be described as the following empirically-derived relationships: The fitting parameters values 𝑝𝑝 𝑝𝑝 have been outlined elsewhere. 79 These functions have been previously included in the PyBaMM software used for the simulations. ## Validation The parameters outlined in Table 8 have been made available in PyBaMM and can used as inputs for different physics-based models to predict battery behaviour in various conditions. To validate the determined parameters, we compared experimental data to simulations using the DFN model coupled to a thermal model. These equations are outlined in Tables S1 and S2. The model was used to predict the voltage and temperature profiles at C-rates of 0.5C, 1C, and 2C, at 0 °C, 10 °C and 25 °C. These 9 datasets are available in a data repository, however for this paper we study five cases: (i) 0.5C|25 °C, (ii) 0.5C|0 °C, (iii) 0.5C|10 °C, (iv) 1C|25 °C, and (v) 2C|25 °C. 25 The initial concentrations for the positive and negative electrodes were set to 13975 mol m -3 and 28866 mol m -3 , respectively, which correspond to stoichiometries of 0.27 and 0.98. Note that the initial concentrations can vary significantly from cell to cell, so their values were determined by manually adjusting the rest voltage at the beginning of the simulation to the experimental data. To achieve good agreement between the simulated and experimental data adjusting of a few other parameters values is needed. This was initially carried out based on the voltage profile of the 1C|25 °C case, for this only one parameter needed manual tuning: the positive electrode diffusivity. This tuning was done manually by trial and error by setting a multiplicative factor to the diffusivity function (illustrated in Fig S8 ) until a good qualitative agreement was observed with the experimental data. We found that a factor of 2.7, the same as used in a previous work, 23 gave good agreement with experimental data, and even though this can seem a significant adjustment, note that it is within the typical variability between different cells (see Fig S8). The negative electrode diffusivity, on the other hand, was adjusted but instead of manual tuning we used the factor of 3.03 found earlier, which gives a good agreement with experimental data. This contrasts to the previous electrochemical parameterisation for this cell, requiring the negative electrode diffusion coefficient to be increased 1800% from the experimentally determined value, which demonstrates the improvement in simulated data when parameter value variability is considered and not taken to be a constant. 23 The temperature profiles for the 1C and 2C cases demonstrated good agreement, although as the 0.5C is dominated by reversible heat it is more sensitive to entropic term. Therefore, in the 0.5C|25 °C case the negative electrode entropic term had to be tuned. This adjustment involved excluding the lower entropic term values, see Figure 13. In summary, the values of three parameters had to be adjusted to achieve good agreement for the temperature and voltage profiles. The quality of tuning was confirmed by comparing agreement of the different C-rates at 25 °C, see Figure 15. Since the parameters were tuned based on the 1C case there is excellent agreement here, the voltage profiles for 0.5C and 2C have disagreement for the final voltage during relaxation. However, this could be improved by adjusting the electrode diffusivities explicitly for these cases-tuning is likely needed depending on the operational conditions being used in the simulation. Next, we compare the data for the 1C case at various temperatures, see Figure 16. This allows us to observe whether the temperature dependencies of the electrochemical parameters have been mapped adequately. The comparison at 10 °C and 25 °C illustrates good agreement for both the voltage and temperature profiles. However, the agreement between experimental and simulated data is worse at 0 °C, this is due to this temperature being outside the range for measured parameter values-there is likely an interplay of different effects meaning that the Arrhenius relationship cannot be applied for the entire operating temperature range. The heat transfer coefficient was adjusted manually to a value of 15 W m -2 K -1 , which is within the expected range of values. To verify the improvement in simulations accounting for the parameter-dependencies and thermal behaviour for the M50, the simulations were compared to a C-rate discharge that the parameters were not tuned to. The diffusivity needs tuning due to underestimation in the solid-state coefficient during GITT, and in this case we have tuned to 1C for both the Chen et al. parameter set and those outlined in this paper, they both provide good agreement for this C-rate, see Figure 17. However, if these simulations are used to observe cell behave under various C-rates then it is not possible to tune to each C-rate, to observe how these parameters compare to the experimental data for those C-rates not specifically tuned for we repeat the simulations at C/2 and compare to the experimental data. For this case the new parameter set reduces the RMSE (root mean square error) by 27% (RMSE is 0.14 for Chen2020 parameter set and 0.10 for ORegan2021 parameter set). This demonstrates that including these parameter dependencies improves the prediction under conditions that could not be specifically tuned to. This relaxation can also be captured better in the simulation by capturing particle size distributions in the simulations. 80 Reducing the number of parameters needed to be tuned and reducing the magnitude of tuning needed by including non-constant values. There is always deviation between the experimentally measured parameters and the values needed to provide agreement with simulations, so tuning is a necessary step. This is due to errors introduced by the unknown cell composition, damage to materials during teardown, errors introduced in the analysis and the simplicity and assumptions of the models in capturing the full kinetic and thermodynamic data. It should be noted that the computational time required in PyBaMM can be reduced by removing parameter granularity, improving the solver methods, or using a model type with lower complexity. For example, the thermal single particle model with electrolyte (TSPMe) demonstrates similar quality results and is an order of magnitude faster. 25 The parameters in both cases have been tuned to the data for the 1C discharge and then these same parameters are used for the C/2 simulation to demonstrate that the discharge behaviour and relaxation is captured better for C-rates that deviate from the tuned values. C/2 RMSE (root mean square error) for 0.10 for Chen2020 and 0.14 for this work. Future work to improve the accuracy of the parameterisation includes mapping the lithium concentration of the thermal parameters (e.g. specific heat capacity and thermal conductivity), rather than at a single stoichiometry. This relationship was not considered here due to the stability of active materials in ambient conditions. In situ methods that allow measurement as a function of lithium concentration would allow the changes in these parameters that occur during battery operation to be captured. Additionally, the requirement to increase the solid-phase diffusivity value of each electrode highlights limitations in the analytical approach and DFN theory. It is assumed that particles are spherical and monodisperse despite the electrode microstructure being heterogeneous. Ignoring these effects in parameter evaluation and the model is one of the main reasons for disagreement with experimental data. Accounting for electrode inhomogeneity will improve model accuracy, for example including size distribution and non-spherical morphologies. ## Conclusions This paper outlines the parameterisation methodology for a 3D thermal-electrochemical model for a high-energy lithium-ion battery. The electrochemical and thermal relationships in a high energy density cylindrical cell (21700) and the electrodes have been mapped through electrochemical testing at different temperatures, to provide diffusivity, exchange current and electronic conductivity profiles. Additional thermal properties, specific heat capacity, thermal conductivity and the entropic terms are measured using thermal characterisation techniques. The thermal parameters of the cell provide information for a 3D thermal model, whilst the electrode parameters provide information for 0D, 1D and 2D models. A 0D electrochemical-thermal model has been derived with the obtained parameters, providing improved fit to the validation data performed on the cell. Further improvements are likely through expansion of this model to 3D, but further modelling work is required. The physical parameters of the negative and positive electrode are very similar, the positive electrode had slightly lower thickness, at 76 µm rather than 84 µm, a higher pore volume of 33,5% compared to 25% and a slightly lower mean particle radius of 5.2 µm rather than 5.9 µm. Whereas the thermal properties are also very similar with the Specific hear capacity (Cp) at 990 and 950 J kg -1 K -1 respectively, thermal diffusivity, ( α) is 0.282 compared to 2.266 mm 2 s -1 and the thermal conductivity 0.892 compared to 4.058 for the dry electrodes at room temperature, which corresponds to the difference in electronic conductivity of 0.847 and 215 Sm -1 . Indicating that the thermal and electronic conductivities can be linked. The negative electrode likely limits the maximum power observed by the cell, as observed from the lower diffusion coefficient and current exchange density compared to the positive electrode over the full SOC window. At stoichiometries of LixC6, where x=0.45 and 0.85 activation energies of up to 60 kJK -1 and low diffusion coefficients of 5 x 10 -13 cm -2 s -1 at 25 °C were observed. Some of these limitations may be compensated for at 45 °C as the exchange current in the negative electrode surpasses that of the positive electrode and the diffusion coefficient increases in the negative by an order of magnitude to 2 x10 -12 cm -2 s -1 . Whereas for the positive electrode the lowest diffusion coefficients were observed for LixMO2, at x=0.32 and 0.81, which are just within the full cell cycling window (0.26≤x≤91), 7 x10 -12 cm -2 s -1 was obtained at room temperature which increased to 2 x 10 -11 at 45 o C, above stoichiometry of x=0.8 the activation energy also doubled to 24 kJK -1 . The changes in temperature change the ionic transport by orders of magnitude and the reaction rates increase. In terms of application to the modelling, the incorporation of state of charge or stoichiometry and temperature variable conductivities and diffusivities have improved the model fit before tuning. The electrochemical parameters tuning was reduced from four parameters; diffusivities and maximum concentrations, to only the solid-phase diffusivities. The magnitude of tuning was also reduced, the tuning needed of the negative electrode diffusivity was decreased by 303%. The diffusivity of the negative electrode is key to improving the models for high energy cells, large changes in magnitude of the diffusivity with temperature variation with only small changes in SOC, cause difficulties in fitting. To improve the fits at higher rates, the effect of the ohmic resistance and heating must be taken into consideration. As observed by the fit of the 2D discharge and relaxation, the actual observed voltage is significantly lower than the estimated. This is likely because the diffusion coefficient is being underestimated and Ohmic heating is causing faster movement of the lithium ions in the solid, resulting in more lithium transport over that time frame. In summary, a parameterisation methodology is outlined, which uses electrochemical and thermal techniques, illustrating the parameter variability caused by local and global changes in temperature or lithium concentration. This methodology is chemistry and format-agnostic and can be applied to different cell types to increase the availability of 3D thermal-electrochemical parameters. Insight into the diffusion and reaction rate kinetics show the limiting electrodes. Further work with these parameters would be to design, predict

chemsum

{"title": "Thermal-electrochemical parameters of a high energy lithiumion cylindrical battery", "journal": "ChemRxiv"}

formation_of_xenon-nitrogen_compounds_at_high_pressure

## Abstract: Molecular nitrogen exhibits one of the strongest known interatomic bonds, while xenon possesses a closed-shell electronic structure: a direct consequence of which renders both chemically unreactive. Through a series of optical spectroscopy and x-ray diffraction experiments, we demonstrate the formation of a novel van der Waals compound formed from binary Xe-N 2 mixtures at pressures as low as 5 GPa. At 300 K and 5 GPa Xe(N 2 ) 2 -I is synthesised, and if further compressed, undergoes a transition to a tetragonal Xe(N 2 ) 2 -II phase at 14 GPa; this phase appears to be unexpectedly stable at least up to 180 GPa even after heating to above 2000 K. Raman spectroscopy measurements indicate a distinct weakening of the intramolecular bond of the nitrogen molecule above 60 GPa, while transmission measurements in the visible and mid-infrared regime suggest the metallisation of the compound at ~100 GPa.Nitrogen is the most abundant element in the terrestrial atmosphere, existing as a diatomic molecule with one of the strongest known triple bonds and as a result is unreactive at ambient conditions. Under high compression, molecular nitrogen exhibits a rich polymorphism 1-7 and significant overlap of thermodynamically competing phases, dependent on formation conditions 8 . The application of high pressure can also provide new synthesis routes, initiating chemical processes that would not happen otherwise, such as N 2 becoming reactive with the noble metals, as in the formation of platinum or iridium nitrides 9,10 . Xenon, an archetypical inert gas due to its closed shell system, has long been known to form stable halide and oxide compounds through chemical synthesis 11,12 . The reactivity can also be fundamentally altered with the application of high pressure, the process which has produced van der Waals compounds composed of Xe-H 2 13 and Xe-O 2 . Xenon has also been shown to be inserted into both quartz 18 , and a small-pore zeolite at high pressure and temperature 19 . Theoretical studies also suggest the increased reactivity of xenon at high pressures with the formation of binary solids Xe-O 20,21 , Xe-Fe/Ni 22 , and Xe-Mg 23 synthesised solely from their constituent elements. Such studies on the reactivity of xenon, especially with terrestrially abundant elements, could provide an explanation into the significant under-abundance of xenon detectable in the Earth's atmosphere. The direct reaction of N 2 and Xe would seem unlikely due to the relative inertness of both materials. Nevertheless, a recent theoretical study predicts the formation of novel xenon nitride compounds above 146 GPa with stoichiometry -XeN 6 24 . Possible interactions between Xe and N 2 have been explored experimentally at low pressures investigating mutual solubility 25,26 . Through Raman spectroscopic measurements those studies inferred the formation of an orientationally disordered van der Waals compound but were limited up to pressures of 13 GPa at 408 K with no structural investigation. It is known that at high pressures both xenon and nitrogen exhibit (semi-)conducting phases. Xenon has been shown to transform to metallic state at pressures between 130-150 GPa, giving it the lowest pressure of metallisation amongst the rare gas solids and nitrogen becomes semiconducting with band gap of 0.4 eV at 240 GPa 2,3 . Previous studies have claimed that by doping Xe with O 2 , the metallisation pressure is drastically reduced 30 . Therefore it is of significant interest to investigate pressure-induced electronic effects of any formed Xe-N 2 compound. Here, we report the synthesis and characterisation of a Xe-N 2 van der Waals compound through x-ray diffraction, Raman and transmission spectroscopies. We show that two inert condensed gases form a Xe(N 2 ) 2 compound at pressures as low as 5 GPa at room temperature. When the novel compound is formed in a xenon medium, it becomes metallic at around 100 GPa, whilst Xe(N 2 ) 2 with an abundance of nitrogen demonstrates metallic behaviour above ~140 GPa. Mixtures of Xe-N 2 at various concentration were loaded into diamond-anvil cells (DAC) using a combination of cryogenic and high-pressure gas-loading techniques (see Methods section). Compressing the mixture above 2 GPa leads to the formation of a xenon single crystal surrounded by liquid N 2 as seen visually and in x-ray diffraction measurements (see Figs S1 and S2). At pressures above 5 GPa we observe the formation of a N 2 -rich compound in the media surrounding the xenon single crystal (Fig. S2). Through x-ray powder diffraction analysis we have identified this phase as having a fcc structure, with a = 9.2361(3) at 5.6 GPa (Fig. 1), indexing with space group Fd m 3 or Fd3 accounts for all observed Bragg peaks. Several patterns were of sufficient quality to allow for Rietveld refinement, otherwise Le Bail fitting was used to extract unit-cell dimensions. Solution of the structure by charge-flipping suggests space group Fd m 3 . Two atomic sites could be refined; Xe(0, 0, 0) and N ( ) resulting in a cubic Laves Cu 2 Mg-type structure (Fig. 1(a)). From both the structure type and unit-cell dimensions we determine the stoichiometry as Xe(N 2 ) 2 , designated Xe(N 2 ) 2 -I herein, which is in excellent agreement with the calculated equation-of-state data for Xe + 4N (Fig. 1(b), see also below). Both the structure type and the stoichiometry are identical to that proposed for oxygen-rich xenon mixtures 14 . The N-N site distances of 3.2655(1) are clearly too long to be bonded, these sites therefore represent scattering from disordered N 2 molecules. N 2 molecules have been found to adopt both spherical and disk-like rotational disordering in the solid state 31 , and refinement of both disorder types was attempted, with a spherical disorder model (i.e. with the N-site occupancy equal to 2) resulting in the best fit to the data (see table in SM for more details on the structure refinement). The structure of this phase can be considered as a diamond-type host lattice of Xe atoms with four rotationally disordered N 2 molecules forming a tetrahedron within each vacancy. The N-N site distance of 3.2655(1) implies a N… N closest-contact distance of 2.1655(1) . Raman spectroscopy measurements of the formed single crystal at 2 GPa reveals the appearance of a weak vibrational mode, which is lower in frequency than the fluid N 2 vibrational mode by 10 cm −1 (compare red and black spectra in Fig. 2). This mode has been observed in a previous high-temperature study and attributed to fluid N 2 dissolved in the Xe crystal lattice 26 . By contrast, in xenon-rich samples (ca. 4:1 concentration), we observe the complete transformation of the sample, evident through only the low-frequency vibrational mode and no evidence of excess N 2 (see SM). In Raman measurements of the surrounding media (see blue spectra in Fig. 2), we observe a broad N 2 mode at 5 GPa, which consists of overlapping modes of Xe(N 2 ) 2 -I, as determined by x-ray diffraction, and pure N 2 that increasingly separate in frequency space at higher pressure. The vibrational mode of Xe(N 2 ) 2 -I (blue spectra in Fig. 2) and the vibrational mode attributed to N 2 in Xe (red spectra in Fig. 2), exhibit identical behaviour with pressure (see red and blue points in Fig. 3a), suggesting that the latter is most likely due to small crystallites of Xe(N 2 ) 2 -I that form within a Xe matrix. It should be noted that the prescence of this N-containing dopant does not significantly affect the measured unit-cell volume which agrees with the literature to within experimental error 32 . Above pressures of 14 GPa, we observe a phase transition from the low-pressure Xe(N 2 ) 2 -I to a high-pressure Xe(N 2 ) 2 -II phase. This transition pressure corresponds approximately to the critical pressure of the δ to  transition in pure molecular N 2 . Xe(N 2 ) 2 -II adopts a body-centered tetragonal cell with a = 5.7228(3), c = 9.2134(10) at 18.7 GPa (Fig. 1). Systematic-absence analysis unambiguously confirm space-group symmetry I4 1 /amd. Again Xe is located at the origin, with one N position refined to (0.5, 0.721(2), 0.179 (1)). This position lies displaced by 0.52(2) from an inversion centre resulting in four ordered N 2 molecules aligned along the c-axis. Final Rietveld agreement factors are R p = 0.015 and R = 0.094. The origin of this transition lies in the ordered orientation of N 2 molecules within the vacancy, corroborated by the poorer fit to the data (R = 0.1672) with a spherically-disordered N 2 molecule model. Shortest N… N interatomic distances are now 2.5238(1) and 2.610(12) at 18.7 GPa. Recalling that the shortest N… N interatomic distances at 5.6 GPa were 2.1655 (1) in phase I, the alignment of N 2 molecules relieves unfavourable close N… N contacts while maintaing the same coordination number for each N 2 molecule. Over the I-II phase transition the unit cell undergoes a tetragonal distortion elongating by 0.323(3) (+ 3.6%) along c accompanied by a reduction of − 0.519(1) (− 8.1%) along tetragonal a, corresponding to 〈 110〉 in phase I (see table in SM for more details on the structure refinement). We tracked unit-cell dimensions for Xe(N 2 ) 2 -II up to 58 GPa (Fig. 1(c)), confirming again the stoichiometry of the compound (Fig. 1(a)) and allowing the determination of equation-of-state parameters for both Xe(N 2 ) 2 phases (see methods section). At pressures of 38 GPa and above there were clear signs of the incipient high-pressure hcp phase of Xe accompanied by strong diffuse scattering and increased background at d-spacings overlapping with a significant number of Xe(N 2 ) 2 reflections and above 58 GPa unit-cell dimensions could not be reliably extracted from the data. However the low-angle (101) reflection could be observed up to 103 GPa (see Fig. S4). Above 40 GPa, the frequency dependence with pressure of the vibrational mode of Xe(N 2 ) 2 deviates greatly from that of pure N 2 (Fig. 3). The maximum in the vibrational frequency vs. pressure is shifted from 80 GPa in pure N 2 to 30 GPa. In the sample in Xe matrix, we observe splitting of the vibrational band (see Fig. 2) up to 70 GPa, after which the splitting is not distinguishable due to the enhanced broadening of the modes. At 140 GPa the N 2 vibrational frequencies of Xe(N 2 ) 2 are 2161 cm −1 and 2212 cm −1 , considerably lower frequencies than either those of κ-N 2 (2376 cm −1 ) or λ-N 2 (2320 cm −1 , 2400 cm −1 ). Interestingly, at 178 GPa, we observe the persistence of molecular nitrogen, which is above the pressure at which pure N 2 is claimed to become non-molecular (η-N 2 ) . We note that although we observe a much softer N 2 molecular mode than that just before ζ transforms to the non-molecular amorphous η phase in pure N 2 , there is no evidence that the N 2 molecules in Xe(N 2 ) 2 dissociate to form Xe-N bonding. However, there is a clear reduction in intensity (see Fig. 2) together with a marked increase in the FWHH (see Fig. S5) indicating that the molecular N-N bond is weakening. Up to highest pressure studied (180 GPa) we see no evidence of Xe-N bonded compounds predicted by theory 24 . In an attempt to promote synthesis of such compounds, we performed laser heating of the sample to temperatures of 3000 K at 120, 150, 160 and 180 GPa but no transition was observed in either Raman spectroscopy or x-ray diffraction. It is remarkable that a van der Waals solid, the components of which are inert materials, can remain stable to such extreme conditions. Figure 3(b) shows the transmission spectra collected from two samples with different initial ratio of Xe and N 2 . The spectra were collected with both visible and mid infrared light sources which allow the coverage of energy region between ~3 to 0.6 eV. The samples in a Xe matrix (black), appear to exhibit metallic behaviour evident by the sharp rise in the absorption in the near-IR, which shifts with pressure. By 120 GPa, no detectable transmission was observed in the visible, the sample appearance became shiny and reflected red laser light (see photomicrographs in Fig. S6). Samples of Xe(N 2 ) 2 with higher N 2 concentrations do exhibit absorption (Fig. 3 green) but not to the same extent as in the Xe matrix, which could be due to the excess of N 2 . Pure xenon has been shown to be conductive at above 135 GPa through both absorption/reflectivity and electrical measurements . The mechanism of conductivity is an indirect overlap of the 5p valence and 5d conduction bands. Although determining the mechanism was beyond the scope of this study, our results indicate that by doping Xe with N 2 , or Xe(N 2 ) 2 , we are able to tune the conductive properties of Xe and lower the pressure of metallisation. Our results demonstrate that xenon can form compounds not only with chemically reactive gases such as hydrogen or oxygen, but also with unreactive nitrogen. That such a compound forms at low pressure, exhibits metallic properties, and stable to both high-pressure and high-temperature conditions will no doubt stimulate further research in the reactivity of xenon, an element which now appears to be substantially less inert than previously thought. . Right Panel: Optical absorption as a function of energy for Xe-rich (black) and N 2 -rich (green) samples. The reference spectra were taken at 50 GPa in both experiments.

chemsum

{"title": "Formation of xenon-nitrogen compounds at high pressure", "journal": "Scientific Reports - Nature"}

development_of_a_practical_synthesis_of_the_8-fdc_fragment_of_opc-167832

## Abstract: A concise and practical synthesis has been developed to provide the 8-fluoro-5-hydroxy-3,4-diydrocarbostyril (8-FDC) fragment of OPC-167832 in 41 % yield and in > 99 % purity over 4 steps from 3-amino-4-fluorophenol. The key feature of this process is the development of a telescoped one pot synthesis of the quinolone via a chemoselective amidation and easier product isolation without the need for a column chromatography.Tuberculosis (TB) is a contagious bacterium infection caused by Mycobacterium tuburculosis (mtb) and is the leading cause of death worldwide. In 2017, 10 million people were infected by TB and 1.6 million deaths by TB occurred including 230,000 children. 1 Treatments for TB are available; however, drug resistance to these treatments is an ongoing problem. First-line treatments were discovered as early as 1952. Drug resistant strains of TB include multidrug resistant (MDR-TB) and extensively drug-resistant (XDR-TB). The only current compounds available to treat MDR-TB are delamanid, pretomanid, and bedaquiline in combination with other TB drugs. 2 However, identification of novel compounds with unique mechanisms of action are needed to combat drug-resistance and develop shorter less toxic treatment regimens. In this regard, OPC-167832 3 (1, Figure 1) is as a promising compound for the treatment of MDR-TB strains in combination therapy developed by Otsuka Pharmaceuticals Co., Ltd. and has received fast track status for development by the US FDA. 4 Due to the biological significance of 1, efficient synthetic access to this compound is important to enable supply. The current reported synthesis of 1 3b,c utilizes a final coupling of two fragments (2 and 8-FDC, Figure 1). In collaboration with Otsuka Pharmaceuticals Co., Ltd., our group became interested in investigating improved synthetic routes to one of the key coupling fragments, 8-FDC. Access to 8-FDC suffers from a long synthetic sequence (9 chemical steps, 8 "reaction pots") starting from fluorinated nitrobenzene 3 (Scheme 1). 3b This approach mainly suffers from multiple functional group interconversion steps, including replacement of the 5-fluoro group of 3 with the requisite OH-group of 8-FDC. This F to O swap leads to the incorporation of several additional steps in the synthetic route. As a result, we decided to investigate an alternative synthetic design from aniline 4 with the requisite 5-OH group already present in reaction with methyl 3,3-dimethoxypropionate (5) to access the desired pyridone scaffold of 8-FDC. Importantly, both 4 and 5 are commercially available, and synthetic procedures to access 4 on up to 5 kg scale have been reported. 5 Herein, we describe the successful synthesis of 8-FDC utilizing this approach. Scheme 1. Proposed Synthetic Strategy to 8-FDC. ## Scheme 2. Planned Forward Synthesis of 8-FDC. To realize our proposed synthesis plan in Scheme 1, the envisioned forward synthesis is given in Scheme 2. Chemoselective amide formation between 4 and 5 was envisioned to provide 6 that may be converted to quinolone 7 by a Friedel-Crafts type process. Recently, this approach for the synthesis of quinolones was reported. 6 However, we were unsuccessful in identifying chemoselective conditions for formation of amide 6 over the ester formed from reaction of 5 with the phenol-group of 4. As a result, we next investigated the synthesis of amide 6 from the coupling of acid 8 prepared from the hydrolysis of 5 with aqueous NaOH (Table 1). Of the various carboxylic acid activating agents studied, MsCl and PivCl appeared to be the most promising for formation of the product 6 (entries 1 -4). A subsequent survey of bases and solvents utilizing PivCl as the activating agent was then carried out (entries 4 -15). Amongst the bases analyzed, DBU and Hunig's base afforded the highest amounts of the desired amide 6 (entries 8 and 9), but the latter was preferred for its lower cost. In regards to reaction solvent, toluene provided the maximum conversion to 6 and also accounted for the highest yield (entry 13). With the intent to telescope the overall synthesis, a solvent study for the cyclization of amide 6 to the quinolone 7 was next performed. A comparative analysis of the efficiency of the reaction in various solvents was made on the basis of the amount of the product isolated post precipitation from the reaction mixture by pouring into ice cold water, and the results are as captured in Figure 2. Although CPME and Toluene gave excellent mass recovery, toluene was preferred for telescoping the reaction due to its better performance than CPME in the amidation step. In toluene, 67 % assay yield of the quinolone was obtained in 75 wt% purity by quantitative 1 HNMR analysis after direct precipitation from the reaction mixture using water (normal quench). Telescoping the process into a one-pot synthesis of quinolone 7 was found to be successful (Scheme 3), and the quinolone was isolated in 65 wt% purity (quantitative 1 HNMR analysis) after precipitation. Due to the significant exotherm observed during the normal quenching procedure, a reverse-quench by transferring the reaction mixture to ice cold water was next tested in an effort to increase the reaction yield. Gratifyingly, this led to an improved yield of 75% on a 10 g scale. The stages of this one-pot sequence are shown in Figure 3. The initial reaction solution of acid 8 in toluene was homogeneous, however, upon addition of the pivaloyl chloride, the reaction becomes biphasic. After addition of DIPEA, the reaction remained biphasic and resulted in the formation of precipitated DIPEAꞏHCL and amide 6 (solubility of 6 in toluene is ~0.3 mg/mL) after overnight agitation. H2SO4 was then slowly added by addition funnel at 0 o C to control the exotherm leading to the formation of a triphasic reaction mixture. Final reverse-quenching into water induces the precipitation of the quinolone 7 that was isolated by filtration. Losses of 7 to the filtrate were determined to be 348 mg in 390 mL (~3 %) when using a normal quench mode and 656 mg in 448 mL (~6 %) when a reverse quench was employed as determined by quantitative reverse phase HPLC analysis. ## Figure 3. Reaction progress at different stages utilizing reverse-quenching Purification of the resultant quinolone obtained from direct precipitation with water can be achieved using either recrystallization from 80:20 MeOH:H2O or by treatment with 10 V aqueous sodium bicarbonate solution. Using these procedures, 7 could be obtained in 100 wt% purity with 78% recovery after recrystallization with MeOH/H2O or in 93 wt% purity in 88% recovery if treated with aqueous NaHCO3. As a result, the overall yield of 7 from acid 8 was 59% when recrystallization was employed or 62% when utilizing a NaHCO3 reslurry. Since an increased overall yield of 7 was obtained when purifying the material from aqueous NaHCO3, this quality material was attempted to be converted to 8-FDC by the established literature procedure 3b,c to ascertain if 93 wt% material was acceptable (Scheme 4). Gratifyingly, 7 was smoothly converted to 8-FDC in comparable yields without issue. ## Scheme 4. Conversion of quinolone 7 to 8-FDC In conclusion, we have developed a concise process using cheap reagents and starting materials accessible on a bulk scale for the preparation of the key dihydroquinolone (8-FDC) fragment of OPC-167832. The novel process described herein features a telescoped one-pot operation for the preparation of quinolone 7 from 3-amino-4-fluorophenol and 3,3-dimethoxy propionic acid without the need for isolation of any of the intermediates. The method described herein should in principle be broadly applicable for selective acylation of a wide variety of aminophenols, and would therefore be significant for efficient synthesis of various biologically active molecules. ## EXPERIMENTAL SECTION General. All reactions were carried out under nitrogen atmosphere unless otherwise indicated. Glassware was pre-dried in an oven prior to use. 3-Amino-4-fluorophenol was purchased from Oakwood, Methyl 3,3-dimethoxy propionate was purchased from TCI chemicals, acetic anhydride from Chem Impex, trimethylacetyl chloride, 10 % Pd/C and N,N-Diisopropylethylamine from Sigma Aldrich. Toluene and Methanol reagent grade were purchased from J T Baker whereas MTBE, NaOH (pellets), dimethyl fumarate, mesitylene, Hydrochloric acid, sulfuric acid, Acetic acid were purchased from Sigma Aldrich. ## 3,3-dimethoxpropanoic acid (8). To a 100 mL round bottom flask with stir bar is charged 20 mL of water followed by 8.1 mL of 10 M NaOH (30%). The ester 5 (10 g) was charged and stirred at 55 o C for 2 h. TLC (30% EtOAc/hex) and NMR of a worked-up aliquot (obtained by quenching 20 µl of solution with a few drops of concentrated HCl until pH was acidic, extraction with MTBE (2 ml) and concentration) showedconsumption of the starting ester. The reaction mixture was then cooled to rt and washed with MTBE (1x30 mL). To the aqueous layer was then added 6.7 mL of conc. HCl over ~10min keeping the internal temperature below 18 o C using an ice bath. The batch was warmed up to room temperature and 4.5 g of sodium chloride was added. The reaction mixture was stirred for about 10 min to dissolve the salt, and the organic layer was extracted with MTBE (3x50mL). The organics were further dried with Na2SO4, filtered and concentrated in vacuo to provide 9.72 g (99%) of 8 as a near colorless oil in 91.7 wt % purity as determined by quantitative 1 HNMR spectroscopic analysis using dimethyl fumarate as the analytical standard. Spectral data was identical to the literature. 7 8-fluoro-5-hydroxyquinolin-2(1H)-one (7). 3,3-dimethoxy propionic acid 8 (9.5 g, 63 mmol, 89 wt %) was weighed into a three neck round bottom flask having mechanical stirring and under an atmosphere of N2. Reagent grade toluene (95 ml) was added followed by N,N-diisopropyl ethylamine (14 ml, 1.25 eq, 79 mmol) at room temperature. The reaction mixture was then cooled to 0 o C and trimethylacetyl chloride (9.7 ml, 1.25 eq, 79 mmol) was added dropwise over 5 minutes. The reaction mixture became cloudy post addition with the formation of a precipitate. The reaction was then warmed to room temperature and stirred for 4 h. 3-Amino-4-fluorophenol (4, 8.9 g, 63 mmol, 1 eq, 90 wt %) was added neatly as a solid to the reaction mixture at room temperature. The heterogeneous biphasic reaction mixture was then allowed to stir for overnight. HPLC analysis post overnight showed clean formation of 6. At this stage, the reaction was again cooled to 0 o C and H2SO4 (50 ml, 15 eq, 950 mmol) was added via an addition funnel dropwise over 45 minutes. Post addition, the reaction was warmed up again to room temperature and stirred for 45 minutes. At this stage, clear formation of two layers was observed. The reaction mixture was transferred carefully into cold water (~30 V, 300 ml) in a 500 ml Erlenmeyer flask kept in an ice bath. To the viscous and oily nature of the bottom sulfuric acid layer that still remained in the round bottom flask was added additional water (10 V, 100 ml) successively in portions dropwise in cold condition and the precipitate obtained was combined with the material in the Erlenmeyer flask above. After stirring the contents for 30 minutes at room temperature, the obtained precipitate was filtered and successively washed with water and toluene (5 V each). Lastly, the precipitate was washed with MTBE (5 V) and dried for 30 minutes under house vacuum. The buchner funnel was kept for drying inside a beaker in a vacuum oven at 60 o C for 15 h to provide 12.8 g (75%) of 7 as a light-yellow solid in 66 wt% purity determined by quantitative 1 HNMR spectroscopic analysis using mesitylene as the analytical standard. This crude material was added to 130 ml, 10 V of 9 % aqueous sodium bicarbonate solution and stirred at room temperature for 1 h. After gas evolution subsided (~40 min), the solid was collected by filtration and washed with water (20mLx2 times) and dried in a vacuum oven at 60 o C for 15 h under a gentle stream of N2 to afford 7.5 g (62% overall, 82% recovery) of 7 as a beige colored solid in 93 wt% purity determined by quantitative 1 HNMR spectroscopic analysis using mesitylene as the analytical standard. Spectral data was consistent with the literature. Synthesis of 8-fluoro-5-hydroxy-3,4-diydrocarbostyril (8-Fluoro-5-hydroxy-3,4dihydroquinolin-2(1H)-one, 8-FDC). To a suspension of 7 (2.00 g, 10.4 mmol, 93 wt %) in a 20 ml reaction vial was added 12 ml of acetic anhydride and heated to 120 ˚C for 2 h. The reaction mixture was cooled to room temperature (Upon cooling precipitation occurred) and poured into 10 V of iced water. The reaction mixture was stirred at this temperature for 1 h. The precipitate was collected and washed with 3 V of water to afford 10 as a beige colored solid (2.1 g, 90 % yield, 98 wt %). The analytical data was identical in all respects to the literature. 3b,c In a mini autoclave vessel, 10% Pd/C (20 mg, 10 wt % as a dry powder) was added to a suspension of 10 (0.200 g, 0.832 mmol) in 10V of AcOH. The vessel was backfilled and vented with nitrogen followed by H2 at 60 psi (4 atm). The reaction mixture was then heated at 75 ˚C for 8 h. The vessel was then cooled to 40 deg. C and vented/replaced autoclave with N2. The residue was filtered through 2 wt % celite. Filtrate was evaporated to a white solid under reduced pressure to provide 214 mg (97 % yield, 84 wt %) of acetate-protected 8-FDC as a white solid. To a suspension of this material (214 mg) in 1 ml, 5V of MeOH was added conc. HCl (5V), 1 ml. The reaction mixture was heated at 100˚C for 1 h. It was then cooled to 40 ˚C and water (2 ml) was added (precipitation occurred) followed by stirring at 30˚C for 1 h. The reaction mixture was then cooled to 0 ˚C and stirred an additional 1 h. The precipitate was then filtered and dried to provide 113 mg (77%, 75% over two steps) of 8-FDC as a white solid in 99.3 wt% purity by quantitative 1 HNMR spectroscopic analysis using mesitylene as the analytical standard. Analytical data of 8-FDC exactly matched the original literature. 3b,c ASSOCIATED CONTENT ## Supporting Information The Supporting Information is available free of charge on the ACS Publication website at DOI: Copies of 1 H and 13 C NMR spectra of all new compounds (below).

chemsum

{"title": "Development of a Practical Synthesis of the 8-FDC Fragment of OPC-167832", "journal": "ChemRxiv"}

the_s∴π_hemibond_and_its_competition_with_the_s∴s_hemibond_in_the_simplest_model_system:_infrared_sp

## Abstract: The Srp hemibond (two-center three-electron, 2c-3e, bond) is an attractive interaction between a sulfur atom and p electrons. The Srp hemibond is of essential importance in understanding chemistry of sulfur radical cations, and its roles in biochemistry have recently attracted much interest. In the present study, we observe the Srp hemibond in the simplest model system in the gas phase. Infrared spectroscopy is applied to the [benzene-(H 2 S) n ] + (n ¼ 1-4) radical cation clusters. In n ¼ 1, the CH stretch and SH stretch bands of the benzene and H 2 S moieties, respectively, are clearly different from those of the neutral molecules but similar to those of the ionic species. These vibrational features show that the positive charge is delocalized over the cluster due to the Srp hemibond formation. In n ¼ 2-4, the SrS hemibond and Sp-S multicenter hemibond (three-center five-electron, 3c-5e, bond) can compete with the Srp hemibond. The observed vibrational features clearly indicate that the SrS hemibond formation is superior to the Srp hemibond and S-p-S multicenter hemibond. Calculations of several dispersioncorrected density functionals are compared with the observations. While all the tested functionals qualitatively catch the feature of the Srp hemibond, the energy order among the isomers of the different hemibond motifs strongly depends on the functionals. These results demonstrate that the [benzene-(H 2 S) n ] + clusters can be a benchmark of density functionals to evaluate the sulfur hemibonds. ## Introduction The Srp hemibond (S-p interaction or two-center threeelectron, 2c-3e, bond) is an attractive interaction between a singly occupied lone pair orbital of sulfur and a doubly occupied p orbital, or vice versa (here, we should note that the whole benzene molecule with the delocalized p-orbital is regarded as one "center" in the 2c-3e bond while another center is the sulfur atom). The Srp hemibond has attracted great interest because it plays crucial roles in chemistry and biochemistry of sulfur radical cations. For example, it has been pointed out that increased strength of interactions of thioethers and arenes, e.g., methionine-phenyl, facilitates oxidation through the stabilization by the Srp hemibond formation. The Srp hemibond between a sulfur-containing residue and a phenyl ring in oxidized protein has been supposed to serve as a potent relay station in multistep electron hopping processes. Despite the increasing interest to the Srp hemibond, especially in biochemistry, its spectroscopic evidence has been rather scarce. Experimental confrmation of the Srp hemibond has been pioneered by Werst with EPR spectroscopy. 8 Broad electronic transitions due to the excitation to the hole in the anti-bonding orbital have also been used to confrm the formation of the Srp hemibond in the condensed phase, 9,10 as well as many reports on SrS hemibonds. 11 Bally, Glass, and coworkers have performed photoelectron spectroscopy of thioether compounds carefully designed to explore the Srp hemibond, and the comparison of the spectra of the compounds with and without the phenyl group, as well as with and without the sulfde group, has clearly evidenced the orbital interaction between the sulfur and phenyl ring. 9 For characterization of an intermolecular interaction, gas phase spectroscopy of its model clusters can provide the most reliable data, which can be also compared directly with high precision quantum chemical computations. 12 With the proper choice of the combination of molecules in the cluster, factors other than the intermolecular interaction of interest can be excluded. Moreover, competition among multiple intermolecular interactions can be also studied by the choice of the component of the cluster. While for the SrS hemibond and prp hemibond (ordinarily called "charge resonance"), detailed spectroscopic studies on their simple model cluster systems have been reported, no such a study has been performed for the Srp hemibond. Then, in the present work, the radical cation clusters of benzene (Bz) and hydrogen sulfde, [Bz-(H 2 S) n ] + , (n ¼ 1-4), are studied by infrared (IR) spectroscopy in the SH and CH stretch regions. Since IR spectroscopy is sensitive to molecular structures and intermolecular interactions, IR spectroscopy of Srp hemibonded systems would provide us rich information on the nature of the Srp hemibond, which is complementary to the previous electronic and photoelectron studies. 9,10 The n ¼ 1 cluster can be the simplest prototype of the Srp hemibond in the gas phase. The observed vibrational features clearly reveal the Srp hemibond formation in the radical cation cluster. In the n ¼ 2-4 clusters, other hemibond motifs, the SrS hemibond and S-p-S multicenter hemibond (three-center fve-electron, 3c-5e, bond) can compete with the Srp hemibond. The SrS hemibond formation in (H 2 S) n + has been observed by the transient electronic absorption in aqueous solution, 19 and has recently been confrmed by IR spectroscopy in the gas phase clusters. 13 Formation of a multicenter hemibond (3c-5e bond) has been reported for rare gas atom clusters, alkaline earth atom clusters, and boryl radicals. However, such charge delocalization over two molecules has never been discussed on the Srp hemibonded system, to our best knowledge. The observed IR spectra of the n ¼ 2-4 clusters show that SrS hemibond formation among H 2 S molecules is superior to the Srp hemibond and S-p-S multicenter hemibond. The observed spectra are also compared with the spectral simulations by density functional theory (DFT) calculations. As has been pointed out, the validity of DFT calculations of hemibonds strongly depends on functionals. Several dispersioncorrected functionals are tested, and it is demonstrated that the [Bz-(H 2 S) n ] + clusters can be a benchmark to evaluate the performance of functionals on the simulation of the hemibond motifs. ## Experimental Two different preparation methods were applied for [Bz-(H 2 S) 1 ] + to test the existence of its stable isomers. One method is resonance-enhanced multiphoton ionization (REMPI) of the neutral Bz-(H 2 S) 1 cluster under the collision free condition. The gaseous mixture of He/H 2 S/Bz was expanded to a vacuum chamber, and the resultant supersonic jet was skimmed to form a molecular beam. The [Bz-(H 2 S) 1 ] + radical cation was prepared by one-color REMPI of neutral Bz-(H 2 S) 1 via its S 1 -S 0 6 1 0 band. 31 The produced ion structure can be restricted by the vertical ionization from the structure of the neutral cluster. The microwave and IR spectroscopies of jet-cooled Bz-(H 2 S) 1 have revealed that H 2 S locates on the C 6 axis of the aromatic ring in S 0 , 31,32 and this structure of neutral cluster might be advantageous to preferentially form the Srp hemibonded structure of the cation, in which the H 2 S molecule should locate on the phenyl ring. The produced ions were detected by a timeof-flight mass spectrometer. An IR spectrum of the [Bz-(H 2 S) 1 ] + radical cation was measured by IR dissociation spectroscopy. The IR light pulse was introduced 50 ns after the ionization light pulse, and the IR light frequency was scanned. The depletion of the [Bz-(H 2 S) 1 ] + signal due to the vibrational predissociation was detected as a measure of the IR absorption. Another method is ionization of bare Bz molecules in the collisional region of the jet expansion. Bare Bz cations were frst produced by REMPI of bare neutral Bz, and the [Bz-(H 2 S) 1 ] + cluster ion was generated by following collisions with H 2 S in the supersonic jet expansion. In this "pick-up" method, the most stable structure tends to be produced. 33 The produced ions were introduced into a tandem type quadrupole mass spectrometer. 34 The cluster ion mass-selected by the frst mass spectrometer was irradiated by the IR light in the octopole ion guide. The fragment ion was produced by predissociation following the IR absorption, and was detected by the second mass spectrometer. By measuring the fragment ion intensity while the IR frequency was scanned, an IR spectrum of the parent ion was obtained. The fragment detection can be free from the background signal. Therefore, the quality of observed spectra is less sensitive to the fluctuation of the parent ion intensity in the fragment ion detection than in the depletion detection of the parent ion. The n ¼ 2-4 cluster ions were also produced by the pick-up method and their IR spectra were measured by photodissociation spectroscopy using the tandem quadrupole mass spectrometer. In all the IR spectral measurements of n ¼ 1-4 by using the tandem quadrupole mass spectrometer, the [Bz-(H 2 S) n1 ] + fragment cation was monitored. No signal was detected in the (H 2 S) n + fragment channel. In our previous studies on the sulfur-containing charged clusters, (H 2 S) n + (n ¼ 3-6) and H + (H 2 S) n + (n ¼ 3-9), 13,35 we have demonstrated that MP2 calculations show the best performance to reproduce their observed IR spectra. In the present study on [Bz-(H 2 S) n ] + , however, we failed in MP2 calculations because of the signifcant spin contamination. Therefore, we employed DFT to calculate energy-optimized structures and their IR spectra. The four dispersion-corrected functionals, B3LYP-D3, M06-2X, M06-L, and uB97X-D, were used with the 6-311G+(3df,2p) and aug-cc-pVDZ basis sets. Energy-optimized structure search and harmonic vibrational simulations were performed by the Gaussian 09 and 16 program suites. 36,37 Results and discussion in the SH region are also shown in Fig. 1. It has been known that bare Bz in the neutral ground state shows three bands in the CH stretch region because of the Fermi mixing. 38 In the spectrum of (Bz-H 2 S) + , this Fermi triad disappears and only a single weak band is found at 3084 cm 1 . This feature rather resembles that in Bz + -Ar at 3095 cm 1 , which is supposed to be essentially identical with the bare Bz cation. 39,40 Thus, the CH stretch region suggests that the Bz moiety in [Bz-(H 2 S) 1 ] + should be charged, as easily expected by the lower ionization energy of Bz (9.24 eV) than H 2 S (10.46 eV). 42 In the SH stretch region, a single band is observed at 2560 cm 1 and this band is largely red-shifted from those of neutral H 2 S. 41 This shift cannot be attributed to the p-hydrogen bond formation between H 2 S and Bz. This is because the aromatic ring is positively charged, as shown by its CH stretch band, and the aromatic ring rather repels the proton (hydrogen) of H 2 S. 43,44 When H 2 S directly solvates a charged site, charge transfer occurs more or less, and it lowers the SH stretch frequencies because of the partial reduction of the electron density in the SH bonds. The IR spectra of (H 2 S) n + (n ¼ 3-6) and H + (H 2 S) n + (n ¼ 3-9) 13,35 have showed that free SH stretch bands of an essentially neutral H 2 S molecule in the frst solvation shell of a charged site are seen only in the region higher than 2585 cm 1 . The 2593 cm 1 band of (H 2 S) 4 + reproduced in Fig. 1 is assigned to the free SH stretch band of such neutral H 2 S molecules in the frst solvation shell to the ion core. The observed SH band frequency of [Bz-(H 2 S) 1 ] + is too low to be assigned to neutral H 2 S, but it is very close to that of the free SH stretch (2565 cm 1 ) of the ion core moiety in (H 2 S) 4 + . 13 It has been shown that (H 2 S) 4 + has the hemibonded ion core, (H 2 SrSH 2 ) + , in which the positive charge is equally shared by the two H 2 S molecules. 13 Hence, the SH stretch feature suggests that the positive charge in [Bz-(H 2 S) 1 ] + is largely shared by the H 2 S moiety, i.e., an Srp hemibond is formed in the cation. The CH stretch frequency in (Bz) 2 + , in which the two Bz molecules are prp hemibonded (in the charge resonance state), 15 has not yet been clearly determined, but it has been estimated to be 3066 cm 1 from the dimer ion core component in the IR spectra of (Bz) n + (n $ 3). 16 The observed CH stretch of [Bz-(H 2 S) 1 ] + is actually located in between those of bare Bz + and (Bz) 2 + , and this is consistent with the Srp "hetero" hemibond, in which the excess charge cannot be equally shared by the two different molecules. 1, which is generated by the vertical ionization of the on-top p-hydrogen-bonded neutral cluster. Fig. 2(b) is the spectrum of the ion produced by the pick-up type source. This type of ion source tends to produce more stable ions because of the collisional cooling during the cluster production and no restriction of the initial cluster geometry. 33 It is clearly seen that the two spectra are essentially identical. This means that there exist no apparent isomers and the Srp hemibonded structure suggested by the IR spectra would be the most stable structure of [Bz-(H 2 S) 1 ] + . This Srp hemibond formation would be a unique interpretation to reasonably explain the observed IR spectrum of [Bz-(H 2 S) 1 ] + . To confrm the above qualitative discussion on the observed IR spectrum by the comparison with the related species, dispersion-corrected DFT calculations were performed with four functionals (B3LYP-D3, M06-2X, M06-L, and uB97X-D) and two basis sets (6-311G+(3df,2p) and aug-cc-pVDZ). Both the basis sets provided similar results for each functionals. Then, in the following, we focus on the results with the 6-311G+(3df,2p) basis set, and those of the aug-cc-pVDZ basis set are summarized in the ESI. † Several initial on-top structures as well as in-plane (CH-S hydrogen-bonded type) structures were tried in the energy optimization of [Bz-(H 2 S) 1 ] + , and all of them fnally converged to a unique on-top structure. This is consistent with the missing of apparent isomers in the observed spectra. The stable structure (1-1) at the uB97X-D/6-311G+(3df,2p) level is shown in the left column of Fig. 3(a). The essentially same structure was obtained by the other functionals and basis set, and the results are summarized in ESI. † In this structure, the sulfur atom locates right above a carbon atom and the SH bonds are parallel to the aromatic ring plane. Since the non-bonding orbitals of H 2 S have the 3p character and are almost perpendicular to the SH bonds, the optimized structure suggests large overlap between the nonbonding orbital of H 2 S and the p orbital of Bz. The spin density of this structure is also shown in the right column of Fig. 3(a), and it clearly proves the Srp hemibond formation, in which the unpaired electron is delocalized over the Bz and H 2 S moieties. The ionization energy of H 2 S is, however, about 1.2 eV higher than Bz, and the charge distribution cannot be equivalent in the two moieties. The natural charges on the H 2 S and Bz moieties are calculated to be 0.438 and 0.562, respectively. The harmonic vibrational spectra of the Srp hemibonded structure of [Bz-(H 2 S) 1 ] + by the four functionals with the 6-311G+(3df,2p) basis set are shown in Fig. 2(c)-(f). The simulated spectra were scaled to adjust the strongest SH stretch band to the observed band at 2560 cm 1 , and were also normalized to have the same peak maximum of the SH band. Agreement between the observed and simulated spectra is not perfect; the position and relative peak intensity of the CH stretch band show small differences from the observed ones. However, all the simulated spectra reproduce well the gross features of the SH and CH band positions and their relative intensities of the observed spectra. These simulations strongly support the Srp hemibond formation in [Bz-(H 2 S) 1 ] + . Moreover, these simulations demonstrate that all the four functionals are useful to catch the physical essence of the Srp hemibond. It is worth to note that the present result on [Bz-(H 2 S) 1 ] + is quite different from that of the water analogue cluster, [Bz-(H 2 O) 1 ] + . 43,44 Also in [Bz-(H 2 O) 1 ] + , stable on-top structures attributed to the Orp hemibond have been predicted in the theoretical computations. However, the in-plane structure formed by the CH-O hydrogen bonds is much more stable, and only this structure has been experimentally identifed. The fnding of the stable in-plane structure in [Bz-(H 2 O) 1 ] + , which is missing in [Bz-(H 2 S) 1 ] + , can be attributed the fact that the difference of the ionization energy between Bz and H 2 O (3.4 eV) is much larger than that between Bz and H 2 S (1.2 eV). 42 Since the exponential dependence of the hemibond strength on the ionization energy difference has been shown, 11,45,46 the Orp hemibond might be much weaker than the Srp hemibond (therefore, the on-top structure of [Bz-(H 2 O) 1 ] + might be essentially regarded as a charge-dipole complex), and the potential minimum in the aromatic ring plane can independently exist. In the in-plane structure of [Bz-(H 2 O) 1 ] + , the CH stretch band intensity is remarkably enhanced by the CH-O hydrogen bond, and the CH band appears as strong as the OH stretch bands. 44 This contrasts with the observed intensity distribution of [Bz-(H 2 S) 1 ] + , in which the SH band is much stronger than the CH band, and also supports the Srp hemibonded structure of [Bz-(H 2 S) 1 ] + . Fig. 4 shows the IR spectra of [Bz-(H 2 S) n ] + (n ¼ 1-4), which are produced by the pick-up type ion source. In the following, the cluster size is simply denoted only with n, the number of H 2 S molecules in [Bz-(H 2 S) n ] + . In the spectra of the n > 1 clusters, several new features are seen, in addition to the free SH of the ion core ($2560 cm 1 ) and CH stretches of Bz (3000-3100 cm 1 ). By the comparison with the previously reported IR spectra of (H 2 S) n + and H + (H 2 S) n , 13,35 these new features are unequivocally assigned even without help of quantum chemical computations. The intense and broadened features below 2400 cm 1 are attributed to H-bonded SH stretches of the ion core (note that H-bonded SH stretches of neutral H 2 S moieties are generally seen in 2500-2600 cm 1 ). 13,47 The sharp feature ($2600 cm 1 ) at the high frequency side of the free SH of the ion core is assigned to free SH stretches of the neutral H 2 S moiety. 13,35 In the n > 1 clusters, multiple H 2 S molecules enable several different hemibond motifs. The observed IR spectral features provide us rich information on the competition among the hemibond motifs. Four different types of hemibond motifs are found in the stable structure search of the n ¼ 2 cluster at the uB97X-D/6-311G+(3df,2p) level, and these stable structures and their spin density plots are shown in Fig. 3(b). The essentially same structures were found also with other functionals and basis set, and they are summarized in ESI. † In structure 2-1, Bz is sandwiched by two H 2 S molecules. The spin density plot shows that the charge is delocalized to all the three molecules, and this means that the S-p-S multicenter hemibond (3c-5e bond) is formed. Structure 2-2 is a variation of the multicenter hemibond; two Srp hemibonds are formed on the carbon atoms in the diagonal position, and the charge is delocalized to all the molecules. Structure 2-3 holds the single Srp hemibond. The second H 2 S molecule is essentially neutral and is H-bonded to the hemibonded H 2 S (ion core). In structure 2-4, the two H 2 S molecules form an SrS hemibonded ion core, (H 2 SrSH 2 ) + , and the ion core is solvated by the neutral Bz molecule. These stable structures demonstrate that three different hemibond motifs, S-p-S, Srp, and SrS, can compete in the present system. The relative energy of each isomer structure is summarized in Table 1. For each functional (row in the table), the energy of the most stable isomer is set to zero. The relative energies of the isomers strongly depend on the functionals, and no common trend among all the functionals can be seen in the table. This means that most of these functionals have serious problems to quantitatively evaluate the hemibond though all of them can qualitatively illustrate the nature of the hemibond motifs. In the following, the reliability of these functionals to evaluate the hemibond motifs is examined by comparison with the isomer distribution suggested by the observed IR spectra. The simulated IR spectra of the four isomers of the n ¼ 2 cluster at the uB97X-D/6-311G+(3df,2p) and B3LYP-D3/6-311G+(3df,2p) levels are shown in Fig. 5 with the reproduction of the observed spectrum. The simulations by the M06-2X and M0-6L are summarized in ESI † because their energy evaluations obviously conflict with the observation, as described below. The band assignments are presented by colored arrows; violet: free SH stretch of the ion core, blue: free SH stretch of the neutral H 2 S moiety, and green: CH stretch. Note that each arrow indicates a peak of an envelope in which contribution of multiple vibrational modes can be involved. Bands without an arrow are attributed to H-bonded SH stretch of the ion core, and they appear only below 2400 cm 1 . The free SH bands of the ion core and neutral H 2 S moiety appear in 2540-2600 cm 1 , and the band patterns (symmetric/antisymmetric SH stretches or dangling SH stretch) depend on the structures. The splitting between the symmetric and antisymmetric SH stretch bands and their intensity distributions depend also on the functionals. The observed IR spectra of the n ¼ 2 cluster shows two Hbonded SH bands (2235 and 2340 cm 1 ) of the ion core (charged) moiety. To reproduce these two bands in the low frequency region, coexistence of structures 2-3 and 2-4, each of which shows a single strong H-bonded SH band of the ion core, is clearly requested. In both the M06-2X and M06-L calculations, however, the energy of structure 2-3 relative to structure 2-4 is much higher, and its coexistence is practically excluded. Therefore, the M06-2X and M06-L results conflict with the observation, and these two functionals are omitted in the following discussion. The remaining two functionals, uB97X-D and B3LYP-D3, provide the totally different energy evaluations of the hemibond motifs. While structures 2-4 and 2-3 are the most stable isomers in uB97X-D, structure 2-1 is most stable and structure 2-4 is a rather high energy isomer in B3LYP-D3. Structure 2-1 shows only the free SH stretch band of the ion core, which can correspond to the band at 2564 cm 1 in the observed spectrum. But this observed band can be also attributed to structures 2-3 and 2-4. The correct evaluation of the relative intensities of the free SH stretch and H-bonded SH stretch bands is practically difficult from the observed spectrum because of the large intensity difference between these bands. Table 1 Relative energies of the stable structures of [Bz-(H 2 S) 2 ] + calculated by four different functionals with the 6-311G+(3df,2p) basis set. In the calculations by each functional, the energy of the most stable isomer is set to zero. The zero point energy (ZPE) correction is applied. All units are kJ mol 1 Therefore, we cannot estimate the contribution of structure 2-1 (and/or 2-2) in the observed spectrum of n ¼ 2. The matching of the spectral simulation with the observed spectrum is slightly better in B3LYP-D3 than uB97X-D. However, the minor difference in the spectral simulations does not affect the assignments of the observed bands, and this cannot be a critical issue in the present case. Further examination of the reliability of the functionals is difficult for n ¼ 2. Thus, we examine the spectra of n ¼ 3 and 4. In the observed spectra of n ¼ 3 and 4 shown in Fig. 4, the prominent change of the feature is seen in the free SH stretch band of the ion core around 2560 cm 1 , which is highlighted by the dashed line. This band becomes weaker with increasing cluster size, and fnally disappears at n ¼ 4, while the free SH stretch band of the neutral moiety is still seen at 2599 cm 1 . This weakening of the free SH band of the ion core reflects the progress of its solvation (H-bond formation) by H 2 S. The ion core of structure 2-1 (2-2) is the whole cluster. Both the two H 2 S molecules are not H-bonded at all, and they have totally 4 free SH bonds. In structure 2-3, H 2 S molecule in the ion core (H 2 S bound by the Srp hemibond) is H-bonded to a neutral H 2 S molecule. Therefore, the ion core has only 1 free SH bond. In structure 2-4, the ion core is the SrS hemibonded (H 2 S) 2 + dimer moiety. This ion core is bound to the Bz moiety with the two SH/ p H-bonds, and the core has totally 2 free SH bonds. Therefore, each ion core of structures 2-1, 2-3, and 2-4 has 4, 1, and 2 free SH bonds, respectively (structures with the labels of the free and H-bonded SH bonds are summarized in Fig. S6 The preference of the structure 2-4 type ion core is also evidenced by the spectral change in the CH stretch region. The CH stretch region of the observed spectra of the n ¼ 1-4 clusters is reproduced in Fig. 6 in the expanded scale. While n ¼ 1 shows the single and broadened band, which is quite similar to the CH stretch band of bare Bz + (Bz + -Ar), 39,40 the n ¼ 3 cluster shows the clear Fermi triad structure, which indicates that the Bz moiety is essentially neutral. 38 As seen in the spin density plots in Fig. 3(b), among the four ion core structures (stable structures of n ¼ 2), only structure 2-4 has the neutral Bz moiety. Therefore the spectra of the CH stretch region demonstrate that the major isomer of n ¼ 3 has the structure 2-4 type ion core. Unfortunately, the CH stretch band of the n ¼ 4 cluster is hardly seen because of its poor signal to noise ratio. Computations of stable structures of n ¼ 3 and 4 were performed by the uB97X-D and B3LYP-D3 functionals with the 6-311G+(3df,2p) basis set. As for n ¼ 3, the energy optimized structures and their relative energies are summarized in 2, respectively (in Fig. 7(a), only schematic structures obtained at uB97X-D/6-311G+(3df,2p) are shown. The corresponding stable structures were also obtained at B3LYP-D3/6-311G+(3df,2p) and they were summarized in ESI †). Structures 3-1, 3-2, 3-3, and 3-4 shown in Fig. 7(a) are based on structures 2-1, 2-2, 2-3, and 2-4, respectively, and one more H 2 S molecule solvates the ion core moiety by an H-bond. In both the computational levels, structure 3-4 of the SrS hemibonded motif is the most stable isomer. However, other isomers are higher in energy in uB97X-D, while structures 3-1, 3-3, and 3-4 are almost degenerated in B3LYP-D3. The vibrational simulations based on these structures are summarized in ESI. † Because of the appearance of the free SH stretch band of the ion core (2567 cm 1 ) in the observed spectrum, dominant population of 3-3, which lacks free SH in the ion core, is easily excluded. As the case of n ¼ 2, it is practically difficult to uniquely identify the contribution of isomers by the comparison of the SH stretch region. As shown in Fig. 6, however, the CH stretch region of the observed spectrum clearly demonstrates that the Bz moiety is essentially neutral, and this means exclusive population of structure 3-4, in which the Bz moiety is not involved in the ion core (here, we should note that the characteristic Fermi triad structure of the CH stretch of neutral Bz cannot be reproduced by the simple harmonic vibrational simulation). Therefore, the uB97X-D computation well reproduces the observed spectral feature, and the energetics evaluated by B3LYP-D3, which predicts coexistence of structures 3-1, 3-3, and 3-4, is not consistent with the observation. The energy optimized structures of n ¼ 4 and their relative energies are summarized in Fig. 7(b) and Table 3, respectively. Structures 4-1, 4-2, and 4-4 are based on structures 3-1 (2-1), 3-2 (2-2), and 3-4 (2-4), respectively. Both structures 4-3a and 4-3b are based on 3-3 (2-3). In all these structures, the ion core moiety is solvated by one more H 2 S molecule than n ¼ 3. In n ¼ 4, both the calculation levels show that structure 4-4 which has the SrS hemibonded ion core is the most stable structure. Other isomers are much higher in energy, and the exclusive population of structure 4-4 is predicted. The vibrational simulations by these structures are also summarized in ESI. † Structures 4-3 and 4-4 have no free SH in the ion core, and well reproduce the missing of the free SH band of the ion core in the observed spectrum. As was already pointed out in the above qualitative discussion, the missing of the free SH band of the ion core occurs at n ¼ 4, not at n ¼ 3. Moreover, an H-bonded SH stretch band of the neutral moiety is expected in 2500-2600 cm 1 region for structure 4-3 (both a and b), but such a band is missing in the observed spectrum. All these observed features The observed IR spectrum of n ¼ 2 shows the coexistence of structures 2-3 (Srp hemibonded) and 2-4 (SrS hemibonded), and those of n ¼ 3 and 4 demonstrate the clear preference of the SrS hemibonded type ion core in these sizes. These observations agree well with the energy evaluation of the uB97X-D functional, but conflict with those of other functionals. Therefore, it is concluded that uB97X-D is the best performance functional among the present four functionals to evaluate the hemibond motifs. Moreover, the uB97X-D calculations predict that the S-p-S multicenter hemibond is much less stable than the Srp and SrS hemibonds. This less preference of the multicenter hemibond is also supported by the present experimental spectra of n ¼ 3 and 4. ## Conclusions The Srp hemibond formation and its competition with other sulfur hemibond motifs in the model clusters, [Bz-(H 2 S) n ] + (n ¼ 1-4), were studied by IR spectroscopy combined with the DFT computations. The IR spectrum of n ¼ 1 clearly demonstrated the formation of the Srp hemibond in this simplest model system. In n ¼ 2-4, the S-p-S multicenter hemibond and SrS hemibond can compete with the Srp hemibond. The IR spectrum of n ¼ 2 showed the coexistence of the Srp and SrS hemibond motifs. The spectral features in n ¼ 3 and 4 indicated the SrS hemibond motif is superior to other hemibond motifs. The relative energy evaluations of these hemibond motifs strongly depend on the DFT functionals, and the uB97X-D functional showed the best performance to reproduce the observed trend. It should be noted that the charge accommodation motifs in [Bz-(H 2 S) n ] + (n ¼ 1-4) can be uniquely assigned by their spectral features essentially without help of theoretical computations. Therefore, these [Bz-(H 2 S) n ] + clusters can be a benchmark system to evaluate performance of DFT functionals to handle sulfur hemibonds. Moreover, they would be highly helpful to explore the nature of the Srp hemibond by both more extensive experimental and theoretical approaches, and obtained knowledge will be the basis to discuss roles of the Srp hemibond in biological functions. ## Conflicts of interest There are no conflicts to declare.

chemsum

{"title": "The S\u2234\u03c0 hemibond and its competition with the S\u2234S hemibond in the simplest model system: infrared spectroscopy of the [benzene-(H₂S)_n]⁺ (<i>n</i> = 1\u20134) radical cation clusters", "journal": "Royal Society of Chemistry (RSC)"}

the_electrochemical_measurement_of_salt_diffusion_coefficient,_apparent_cation_transference_number_a

## Abstract: Further development of the electrochemical measurement procedure, namely the Symmetric Polarization Procedure (SPP), is described. The SPP allows for additional verification of measurement assumptions through a quick and simple inspection of a symmetry between the procedure outcomes. It also improves the precision of estimated transport properties. A considerable emphasis was also put on detailed outcomes analysis. In particular, a newly developed approach to the analysis of restricted diffusion data is proposed. This approach is based on a power law which is a common form for the rate equation, therefore it allows for a precise estimation of the salt diffusion coefficient and, innovatively, the diffusion domain. Importantly, as a result of this approach, the subdiffusive motion of species is recognized in every electrolyte examined herein. Additionally, four approaches that lead to a quantification of the apparent cation transference number are also extensively discussed. It has been demonstrated, how a knowledge about the salt diffusion coefficient, the apparent cation transference number and the ionic conductivity can be used to evaluate the electrochemical performance of an electrolyte. It is also shown, how these properties can be utilized to approximate the limiting-current density and the deviation from Nernst-Einstein equation. ## Highlights: • electrochemical measurement procedure with additional outcomes verification by a visual inspection of a symmetry • comprehensive approach to a battery electrolyte characterization • newly developed approach to the estimation of the salt diffusion coefficient and the diffusion domain • subdiffusive motion of ions has been confirmed experimentally ## INTRODUCTION Market demand for portable electronic devices, electric vehicles and energy storage facilities is higher than ever and everything indicates that it will grow even further. This demand prioritizes the research over batteries which would store more energy per volume, recharge faster, withstand more chargingdischarging cycles, provide higher instantaneous power and above all be safe. To fulfill these expectations, new electrolytes must be engineered. The task is not trivial, considering almost an infinite number of substrate combinations. Therefore, in a search of the most optimal solution, the electrolytes should not be evaluated based on a single property (i.e. the ionic conductivity). In theory, as long as ion exchange and aggregation reactions are fast enough to be in equilibrium, three transport properties are necessary to properly describe a binary electrolyte. These properties are: the ionic conductivity, the salt diffusion coefficient and the cation transference number . There are many techniques that are used to estimate these properties . But, as it can be concluded based on examination of Fig. 1 and Fig. 2, the value of a transport property may differ substantially, even though it was estimated with similar or essentially the same technique. Such discrepancies may be attributed to many techniquespecific assumptions, which are not always fulfilled or to a research & development stage of the materials being examined. Human made mistakes, which are sometimes very hard to spot, probably also contribute substantially. A proper design of the measuring routine together with a validation of outcomes should reduce or maybe even eradicate the burden of such mistakes. Accurate data about already-known-compounds is also of great importance to harvest the power of the Artificial Intelligence and Molecular Dynamics calculations. With help of AI and MD, subtle patterns could be spotted, which will presumably lead to many improvements or even to some spectacular breakthroughs. In this work the electrochemical techniques are chosen to estimate the transport properties. These techniques should output the ensemble-averaged properties, which could differ from the temporal properties of relatively small number of markers, especially, in systems where the ergodicity is not fulfilled . In this paper a set of the electrochemical measurements together with the analysis of outcomes is described. The article is structured as follows: In Section 2. all aspects related to the investigated materials, the equipment and the measurements are talked over. The details of performed calculations are embedded in Section 3. Section 4. contains the results together with the discussion. Lastly, the Appendices include explanations of abbreviations used on Fig. 1 and Fig. 2 and derivations of some equations showed in Section 3. ## Materials Ten electrolytes based on four variants of lithium salt were electrochemically characterized in this work. The lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) is a first salt variant (see Fig. 3a). It was purchased from the Sigma-Aldrich/Merck. The LiTFSI salt was chosen because it is well recognized by the scientific community. It is also a commercially available chemical compound of great purity. Polymer electrolytes containing it were extensively studied, thus the results of those studies may serve as a reference. The remaining three lithium salt variants are a part of the Lithium AlkylTriAlkoxyBorates (LiATAB) (Fig. 3b) family. They were engineered at the Warsaw University of Technology by E. Zygadło-Monikowska et al. . The main future of these salts is the ability to obtain a salt variant with the desired length of the oligomer arms (i.e. proportional to the number of ethylene oxide repeat units -n) by controlling the conditions of the synthesis process. The reader is kindly referred to the inventor's publication to see the details of synthesis process, as well as some properties obtained by means of the DSC, the NMR and the EIS measurements. In the , the three borate salt variants used herein are called: "salt 1", "salt 3" and "salt 7", respectively. Lastly, the high molecular weight (Mv ~ 5×10 6 g/mol) poly(ethylene oxide) was obtained from Sigma-Aldrich/Merck (abbreviated here as PEO). Solid polymer electrolytes (SPE) studied here were prepared by a casting from solution technique. Weighted portions of the PEO and the anhydrous salt variant were dissolved in acetonitrile solvent. A magnetic stirrer was used to thoroughly mix the substrates. Then, the solution was poured onto flat PTFE dishes and subjected to a two-step drying process. The first step resulted in formation of a thin foil/film. It lasted up to a day and consisted of evaporating solvent under low vacuum. In the second step, which lasted at least one week, the remaining solvent was removed under a high vacuum. Prior to measurement, a dry polymer electrolyte foil (0.150 to 0.400 mm thick) was cut into discs, 16 mm in diameter. The thickness of every disc was estimated at a room temperature (RT, ca. 25 °C) with a micrometer. Prior to measurements, to minimize the influence of convection, the electrolytes that were liquid at RT, were soaked into the battery separator discs. The discs were cut out of the Celgard® 2500 battery separator sheet (type: monolayer PP, thickness: d=25 µm, porosity: ε=55%, PP pore size: 0.064 µm, tortuosity: δ=2.5, MacMullin number: NM=4.5 ). In the measurements a single fully soaked disc was used. To provide a good surface contact, the lithium-metal electrodes were gently pressing on the disc from both sides. ## Electrolytes consisting of LiTFSI and PEO Four SPE samples (see Tab. 1.) were studied. They were prepared by mixing the weighted amounts of the PEO and the LiTFSI. The weight of the compounds was calculated to obtain samples with a certain molar proportion of the ether oxygen (EO) atoms to the lithium (Li) atoms. The 16:1, 10:1, 5:1 and 3:1 EO:Li proportions were chosen. These proportions are equivalent to the salt mass fractions (WS [wt%]) 28.9%, 39.5%, 56.6% and 68.5%, respectively. Additionally, as a handy indicator of an electroactive specie content, the lithium mass permille (WLi) is also shown in Tab. 1. It is a ratio of the lithium mass to the combined mass of all electrolyte constituents. To shorten the notation, it is expressed in permilles. At RT all samples had a form of semitransparent foil. For convenience, the designation, composition and graph symbol of each measured electrolyte is summarized in Tab. 1. ## Table 1. Designation and composition of the LiTFSI+PEO electrolytes. ## Electrolytes containing a LiATAB salt variant Six electrolytes (see Tab. 2.), each containing one of three LiATAB salt variants, were measured. At RT two samples were in a form of a viscous liquid (i.e. AB3 and AB7), thus they were soaked into the battery separator. Four remaining samples were made by mixing a chosen LiATAB salt variant with the PEO. At a RT, the EO29AB1 polymer electrolyte (the LiATAB salt variant with three oligomer arms, each arm contains a single oxyethylene substituent, n=1) is in a form of a semitransparent, white foil. The AB3 electrolyte (neat LiATAB salt variant with three oxyethylene substituents, n=3) is in a form of a clear, viscous, yellow liquid. The EO23AB3 polymer electrolyte (n=3) has a form of a semitransparent, white foil. The AB7, a neat LiATAB salt electrolyte (seven oxyethylene substituents on average, n≈7), is in a form of transparent, brown and highly viscus liquid. The EO9AB7 polymer electrolyte (n≈7) has a form of semitransparent, light-brown wax. Lastly, the EO27AB7 polymer electrolyte (n≈7) has a form of semitransparent, white foil, with very good mechanical properties. ## Apparatus The main part of the experimental setup consisted of an examined electrolyte sandwiched between two, parallel, equalarea, mirror-symmetry, cation-reversible (e.g. lithium) electrodes. The electrolyte was in a form of a thin-film polymer electrolyte or a liquid electrolyte soaked into the battery separator. From an experimental-effort point of view, such setup is fairly easy to assemble. Also, from the application perspective, it mimic conditions that are likely to be encounter in batteries. Each electrode had 16 mm in diameter and it was prepared by flattening a slice of a lithium rod (Aldrich) on a stainless-steel current collector (SS), with help of a custom-made press. Active parts of the press included (bottom to top): the SS, then the slice, and lastly the piston. Prior to the flattening process, a thin (ca. 50 µm) PTFE tape was placed between the piston and the slice. Then, the force, acting on the piston, smeared the slice over the top surface of the mushroom-shaped SS. By this operation, the top of the SS was covered by a few hundred micrometers thick layer of lithium metal, creating a lithium electrode. After the compression, the PTFE tape was gently exfoliated from the lithium electrode surface. The exfoliation produced a clean, metallic surface. A lithium excess, extruded beyond the SS circumference, was gently cut-off. For each sample two such electrodes were prepared. Then a stack consisting of a sample sandwiched between the electrodes was mounted into a custom-made, gas-tight measuring cell (to see construction details reader is kindly referred to Fig. 2 in ). After being fully assembled, the cell was placed on a custombuild Peltier thermoelectric cooling and heating device, located outside of the glove-box and connected to the Potentiostat/Galvanostat Autolab PGSTAT 30 in a 2-electrode configuration (see Fig. 4). The sample temperature was controlled with ±0.1 °C precision by a thermocouple placed close to the sample. Every stage of the materials preparation and the cell assembly was performed in an argon-filled glove-box. Relatively large area of the electrodes assured a minimal distortion from any edge effects. ## Measuring Procedure The Symmetric Polarization Procedure (SPP) is an ordered set of particular electrochemical measurements (measuring steps). A graphical representation of the SPP is shown in the upper part of Fig. 5 in a form of a flow diagram. The measuring steps that are essential for calculations are marked by a pair of curly brackets with single digit in between (e.g. {1}). Such nomenclature will be used consistently throughout this paper. The lower part of Fig. 5 contains graphs with an exemplary data, which was obtained by the essential measuring steps. The main future of the SPP, compared to approaches presented in the scientific literature, is the ability to quickly and easily spot common experimental mistakes or inconsistencies. This is done by a preliminary verification of the outcomes. The verification is based on an anticipated symmetry between the outcomes of two halves of the SPP. To clarify further, a current-time representation of the outcomes of a chronoamperometric (CA) measurement marked by {2} should be a mirror reflection of the same-type-measurement outcomes, namely the {5}. Similarly, the mirror symmetry should be observed between the electromotive force (EMF) measurement outcomes, the {3} and {6} (compare the graphs in the lower part of Fig. 5). A relatively small deviations from the mirror symmetry are acceptable because they can arise from the noise or unavoidable imperfections in the experimental setup. A detailed description of every measuring step constituting the SPP (the boxes sketched on Fig. 5) is presented below: • Then the VZC is determined relative to the lithium anode. In the subsequent measuring steps, the VZC is used as "a reference potential", meaning that all the potential perturbation values are set relative to the VZC value. • Electrochemical Impedance Spectroscopy (EIS; {1}, {4}, {7}) The amplitude of the sinusoidal potential perturbation is not greater than 10 mV root mean square (rms). For every sample, the value of that amplitude is tuned to be similar in magnitude to |ΔV| (see (1)). The VZC is set as the DC-potential-bias (the offset potential), thus the impedance measure- Where: • Idb = I(t=1/fdb) -electric current flowing at the beginning of diffusion (A), • Rb -electrolyte resistance (Ω), • RPEI -Polymer Electrolyte Interphase (PEI) layer resistance (Ω), • R -gas constant (~8.31 J⋅K −1 ⋅mol −1 ), • T -absolute temperature (K), • F -Faraday constant (~96485 C⋅mol −1 ). It is assumed that for experimental conditions discussed here, a change of the perturbation value does not affect the salt concentration gradients that exist in the sample until the time t=1/fdb. Thus, the mass transport is not affected until the inverse of this system characteristic frequency (fdb). The fdb can be estimated based on an EIS measurement (see Fig. 7b). The "SPE overpotential" accounts for the Ohmic/IR loss component in the SPE. For samples presented here, the "SPE overpotential" did not exceed several dozens millivolts (mV). Also, the Idb should be small enough to observe a correspondingly low current density but high enough to minimize the noise to signal ratio. That is why, initially, the Idb corresponds to a current density, which does not exceed 0.1 mA/cm 2 . At a later stage of outcomes analysis, the Idb is additionally tested if it is lower than 5% of the limiting current (Ilim, see (29)). Additionally, the CA measuring steps are tweak to improve measurements and facilitate data analysis. Each CA measuring step consists of 2 parts, differing in the interval between sampling. The duration of parts and sampling rates given below are arbitrary. The first part has a fast sampling (1 reading per 10ms, I fast ) and it lasts 80 seconds. During first 20 seconds the perturbation is equal to the VZC. In this period none or negligibly small current should be measured. At the 20 th second since beginning of the CA measuring step, an instantaneous change of the perturbtion is imposed. The perturbation changes from VZC to VZC+ΔV during CA {2} and from VZC to VZC-ΔV during CA {5}. In response, the absolute magnitude of current raises to a maximum value and then starts to attenuate. This attenuation is registered for a remaining 60 seconds of the first part. Immediately following the first, the second part begins, where the sampling rate is lowered in order to save the disk space (1 reading per 1s, I slow ). The slow sampling rate persists until the system reaches a steady state. The steady state is inferred when the current does not change in time for at least 50% of the overall CA measuring step duration. • ElectroMotive Force (EMF; {3}, {6}) Immediately after each CA measuring step there is a potential relaxation measurement in open-circuit conditions (the current passing through the Potentiostat/Galvanostat is zero). During these measuring steps, the ElectroMotive Force (EMF) of the measuring cell is registered as a function of time. Here the sampling rate is deliberately set to 0.2 second. The duration of the EMF measuring steps is correlated with the duration of the preceding CA measuring steps. • High Frequency EIS with DC bias (HF EIS) On the Fig. 5, the HF EIS measuring steps are marked by a dashed border, because they are treated as being optional. They are meant to be performed during the CA measuring steps when the system is in a steady state. The HF EIS have the DC potential bias set to VZC+ΔV during the CA {2} or to VZC−ΔV during the CA {5}. The amplitude of an AC signal is not greater than 10 mV rms. In those measurements the frequency range is deliberately set from 1kHz to 1MHz. After each HF EIS the respective CA measuring step is continued to assure that the sample is brought back to the steady state. ## CALCULATIONS During the work with the experimental setup, it was noticed that certain electrochemical states of the sample are more easily induced and more precisely measurable (i.e. a build up of a salt concentration gradient). In practice, it is very difficult to completely get rid of the concentration fluctuations in the sample. The probable causes are small deviations from the experimental setup symmetry, non-homogeneity, the existence of interfaces between the diffuse layer and the bulk or hysteresis effects. To mitigate the influence of departure from pristine state the idea of symmetric perturbation and the averaging was introduced. In case of the effective salt diffusion coefficient the averaging is expressed by formula: And for the apparent cation transference number the suitable equation is: The results presented on figures in sections 4.1. and 4.2. are the outcome of such averaging. ## Effective salt diffusion coefficient (DS) The effective salt diffusion coefficient (DS) of an electrolyte consisting of a net neutral solvent, a univalent cation and a univalent anion is given by the formula : Where: • D0+ -diffusion coefficient for interaction of solvent and cation (cm 2 /s), • D0--diffusion coefficient for interaction of solvent and anion (cm 2 /s), • cT -total solution concentration, sum of ionic species and monomers in solvent (mol/cm 3 ), • c0 -monomers in solvent concentration (mol/cm 3 ), • γ± -mean molal activity coefficient of the mobile species, • m -molality of electrolyte (mol/kg). Experimentally, DS can be estimated by the method of restricted diffusion, developed by Harned and French . This method was later refined by Newman and Chapman . Additionally, Thomson and Newman pointed out that the EMF induced at cation-reversible electrodes reflects the salt concentration gradient near electrodes. In the sample, a monotonic, one-dimensional salt concentration gradient is introduced by a potentiostatic polarization imposed until at least the steady state is reached, that is by the CA measuring step. Each restricted diffusion measurement (aka EMF) begins immediately after the CA step (see Fig. 5). During the restricted diffusion measurement a flow of current is forbidden (the open-circuit condition is impose no the cell). A decline of the EMF is registered as a function of time (as in EMF{3} and EMF{6} measuring step). The data from both EMF decays/relaxations is used to estimate the DS. Prior to the proper outcome analysis, a validation of shape-symmetry between EMF{3} and EMF{6} decay is performed. Shapes of the EMF decays must be acceptably symmetrical to each other, with a line of symmetry lying in the vicinity of zero volts. In cases were the shapes show high degree of symmetry but the line of symmetry is slightly shifted from zero volts (less than 0.5 mV), the EMF decays data is altered by adding or subtracting the value of such shift, causing the line to move to zero volts. Additionally, prior to the calculation of the numerical derivatives, the long time part of the EMF decays was smoothed by averaging a number of EMF values. That number was proportional to time since beginning of the decay. Such treatment calmed noise and equipment-introduced analog-todigital conversion artifacts. Under the "ideal" conditions (a Fickian, one-dimensional diffusion, with the reflective boundaries, no diffuse layers) the EMF decays can be expressed by an infinite series : Where: • E -electromotive force measured at any given time (V), a function, • E0 -electromotive force measured at the beginning of the EMF decay/relaxation (V), herein -the first value measured at the beginning of the EMF decay, a scalar, • Ê -rescaled/normalized electromotive force, a function. For a polymer electrolyte the restricted diffusion length is equal to the sample thickness, and for a liquid electrolyte soaked into a battery separator, the length is calculated according to the formula: Where: • d -thickness of the battery separator (cm), • δ -tortuosity of the battery separator, • ε -porosity of the battery separator, • NM -the MacMullin number. It is troublesome to fit (5) to measured EMF decay, therefore, in order to extract the DS, the approximation at long time is commonly in use: But, to make a proper use of equation (7), the time at equivalence point (τEP) needs to be known. Without this knowledge it is tricky to recognize which part of the relaxation data is the "long time" part. Under the "ideal" conditions, the τEP would be the time since beginning of the relaxation until which the salt concentration gradient in the middle cross section of the sample remains constant (how long it takes for the diffusion fronts to reach the half length). The value of τEP is proportional to the restricted-length squared divided by the DS (see (B.1)). Because the DS is about to be estimated, the τEP cannot be known a priori. But, it should be possible to approximate τEP based on the experiment data. Herein it was assumed that the τEP is equal to the time coordinate of the maximum of the plot shown on Fig. 6b. The ordinate of each point at this plot was Fig. 6. Data representations suitable for the estimation of the effective salt diffusion coefficient (DS), the time at equivalence point (τEP) and the relaxation order (n). Data from EMF{3} and EMF{6} measured for the PEO+LiTFSI electrolyte, (EO16TF, WS=28.9%, EO:Li=16:1) at 105 °C by means of the SPP. equal to the Ê value multiplied by a square root of the Ê-corresponding time. For an unbiased and diffusion-only data a single maximum is expected to appear on such plot. The time coordinate of this maximum is herein inferred to be the τEP. In the present paper, two approaches were deployed to obtain the DS from the analysis of EMF decays, namely: • the "common" approach, which is very similar to one described in the literature , • the "new" approach, first time proposed herein. The whole process of EMF data analysis by both approaches is illustrated on Fig. 6 and described in details below. ## The "common" approach ( █ ) The data representation showed on the Fig. 6c is a consequence of a rearrangement of the equation ( 7) to a time-linearized form: This representation is only useful for the long time part of the EMF decay, thus for the part measured some time after the τEP. Such representation facilitates the estimation of the "slope(α)", thus of the bracket underscored in the equation ( 8). This slope is obtained by fitting a linear function to a subrange of transformed EMF decay. Herein, the subrange was deliberately chosen to span from 1.5•τEP to 2.0•τEP due to of the data within a wide temperature range. Additionally, under the "ideal" conditions, the DS, estimated using the equation ( 7) based on that subrange, differs by less than 0.2% from the DS derived by fitting the infinite series (5) to the "ideal" condition data. Knowing "slope(α)", two values of DS by the "common" approach were calculated according to below formula: and then those values were averaged, as in (2), to give the final result. ## The "new" approach ( █ ) Similarly to the "common" approach, the "new" approach is only suitable for the EMF data registered after the τEP. The "new" approach is based on the assumption, that after a sufficiently long time the EMF decay has the form of a power law (10) which is a common form for the rate equation. According to this law the relaxation rate of the Ê is proportional to the Ê raised to a system-specific power: In present paper that system-specific exponent will be called the "relaxation order (n)". It is worth noticing that the "com-mon" approach is a special case of the "new" approach, in which the exponent is equal to one (n=1). A plot used to evaluate the relaxation order is presented on Fig. 6d and described below. It is a "log-log" plot with the same logarithm base on both axes (here the natural logarithm "ln" was chosen). The values of rescaled/normalized EMF are marked by a hat/circumflex above a capital E letter (Ê) and are obtained by dividing the EMF value by the first value of the EMF decay (E0). To obtain Fig. 6d each Ê value is transformed into two coordinates. The first coordinate, the abscissa, is a logarithm of the Ê value. The second coordinate, the ordinate, is a logarithm of the relaxation rate of the Ê, calculated at the same point in time as the value chosen for abscissa. The relaxation rate of the Ê is mathematically expressed as a derivative with respect to time of the Ê. Herein, the derivatives were calculated numerically. Finally, the relaxation order is the "slope(n)" of a straight line fitted to the long time part (the part at times greater than 1.5•τEP) of this "log-log" plot. Since in every case measured herein the relaxation order was greater than one, the equation ( 8) is replaced by: The relaxation order estimated with help of Fig. 6d enables a transformation of the EMF decay to the representation presented on Fig. 6e. Such time-linearized representation facilitates the determination of the "slope(β)", underscored in the equation (11). To allow a comparison with the "common" approach, the same subrange (from 1.5•τEP to 2.0•τEP) was used to fit a straight line by means of least squares method, and thus to obtain the DS. Two values of DS by the "new" approach were calculated according to the formula: and then those values were averaged, as in (2), to give the final result. ## Model of the electrolyte-electrode interface In this section an attempt will be made to translate known micro-and macroscopic phenomena happening on the polymer electrolyte-lithium metal electrode interface into an electrochemical model. Later, this model will be frequently used in the analysis of the outcomes. Additionally, other concepts that will be used in the subsequent sections are also introduced and justified herein. A schematic representation of the electrolyte-electrode interface together with the electrochemical equivalent circuit is showed on Fig. 7a. The representation renders a LiTFSI+PEO polymer electrolyte, being in contact with a lithium metal 9/25 electrode. The schema depicts a cross-section, which is perpendicular to electrode surface. The interface structure is approximated by four layers, differing in electrochemical properties. This Gedanken cross-section consists of (right to left): • solid, lithium-metal electrode, • compact layer, • diffuse or Polymer Electrolyte Interphase (PEI) layer, • bulk electrolyte. The solid, lithium-metal electrode is the source/sink of the lithium. The compact layer, according to studies of Zaban et al. and Eshetu et al. , contains Li2O, Li2CO3, LiOH together with other products of LiTFSI salt decomposition. Presumably, this layer is very thin, dense and ionically conductive. It forms almost immediately, due to very high reactivity of the lithium. Next, there is the Polymer Electrolyte Interphase (PEI) layer. This layer is electronically insulating, but it allows for diffusion of ionic species. High viscosity of polymer electrolytes contributes to poor electrode wettability, thus many empty voids are expected to exist in this layer , These voids are the major factor affecting the character of ion transport in this layer. Additionally to the voids, other factors may also affect the electrochemical properties of this layer. According to Thiam et al. , during a sample preparation process (i.e. during stirring in the casting from solution) some of monodisperse, long PEO chains may break into much shorter peaces. Depending on the degree of such breaking the polymer chains in the sample may become polydisperse. Moreover, such shorter chains may accumulate in excess at the surface of polymer electrolyte sample due to excluded-volume/entropic-segregation effects . Also, the end groups of the PEO chains (those present at the surface of a polymer electrolyte sample) may react with lithium and anchor-in/adhere-to the compact layer (see pink dots on Fig. 7a). In consequence, the conformation freedom of a part of the chain that is directly connected to such end group may be reduced, what should affect the migration and diffusion of ions in the PEI layer. Such effect presumably intensifies with decreasing length of polymer chains (due to higher concentration of the end groups). Such effect could explain vary interface properties of samples with different PEO chain lengths as observed by Pesko et al. . The bulk electrolyte is listed as last, but the estimation of its electrochemical properties is the main goal of this work. In general, it is worth mentioning that the thickness and active contact area of the compact and PEI layers may evolve with time, temperature and the magnitude of an external perturbation. Also, throughout this work, the assumption is made that the thickness of the compact layer and the PEI layer is much smaller (much less than 5%) than the thickness of the bulk electrolyte, thus the sample thickness is approximately equal to the bulk thickness. The lower part of Fig. 7a contains the electrochemical equivalent circuit, which is intended to mimic an electric response of the upper schematic representation. The circuit consists of two Randles circuits , which are connected in series. The resistances are denoted by capital "R". The Constant Phase Elements (denoted as "CPE"), are used to model a space-separated charge at non-ideal interfaces. The non-idealities in the system most likely arise from a distribution of resistances . The Generalized Finite Warburg elements (denoted as "W s ") served as an electric analogue of the anomalous diffusion (the Warburg-type mass transport impedance, bounded/finite-length anomalous diffusion with an absorbing/ transmissive boundary condition). The diffusive part of the impedance spectra is well modeled by equation (35) presented in the work of Bisquert and Compte : The anomalous resistance parameter of W s is denoted as RW or R(W s ). For γ=1 the bounded diffusion impedance is obtained. The equation ( 13) is probably more suitable for well dissociated electrolytes. In case of systems with poorly dissociated salt, thus an extensive ion aggregation, the modified Gerischer impedance seems a better choice: The subscripts next to the letters denoting an electric analogue on the circuit have the meaning: • dl -double layer, • ct -faradaic charge transfer, • PEI -Polymer Electrolyte Interphase, • b -bulk/solution, • mt -mass transport, Theoretically, an EIS measurement in a wide range of frequencies (usually from MHz to µHz) should allow to distinguish the transport processes occurring in a sample. But, in case of polymer electrolytes, such measurement would usually last for a very long time. Moreover, the impedance measured at very low frequencies could be tampered by an external noise or by a time evolution of the system. Such tampered impedance spectrum could be very hard to recognize and to analyze. Therefore in this work, the chronoamperometric (CA) measurements are utilized to obtain a very low frequency part of the impedance spectrum. The CA outcome, after being checked for any artifacts, is Numerically Laplace Transformed (NLT) from a time-domain to the frequency-domain, using an algorithm similar to that proposed by Takano et al. . The EIS technique is used to measure and deconvolute fast electrochemical processes occurring at high to medium frequencies. Ultimately, the CA NLT outcomes are merged with the EIS outcomes to give impedance spectrum from very wide frequency range. Exemplary spectrum together with data transformation process is shown on Fig. 7b. Based on a set of impedance spectra obtained at a range of temperatures, the electrochemical equivalent circuit, sketched on Fig. 7a and Fig. 7b., was proposed. The circuit was fitted to each impedance spectrum by means of non-linear least squares method to extract the parameters of the electric analogues. Another concept, namely the "optimal time" (τopt), is introduced here. In essence, this concept is similar to the concept proposed by Wohde et al. . The τopt is utilized to estimate the value of I0 (as in ( 16) and ( 18)). Herein, the τopt is the time that it takes to affect the salt concentration gradient in the bulk electrolyte by applying the potential perturbation. The τopt is estimated based on the optimal frequency (fopt). The relation between the two is: The fopt (see Fig. 7b) is the frequency at which the real part of impedance (Z', resistance) is equal to the sum of electrolyte bulk resistance (Rb) and interface resistance (Ri). Whereas, the Ri is a sum of the PEI layer resistance (RPEI), the charge transfer resistance (Rct) and the PEI layer anomalous diffusion resistance (R(W s PEI)) (see Fig. 7b). The content of the subsequent sections should be more clear knowing all the concepts mentioned in this section. ## Apparent cation transference number (T+) The apparent cation transference number (T+) is the fraction of the overall current that flows through an electrolyte when only one electroactive constituent of the salt, e.g. lithium is allowed to be reduced or oxidized at the electrodes. In practice, the overall current should be small enough not to cause large salt concentration gradients, thus the system should obey the Ohm's law. The T+ is also called the limiting current fraction . For an ideal electrolyte with fully dissociated salt, the T+ is equal to the cation transference number . But, if mobile associated species (like ion pairs, triplets etc.) exist in an appreciable amount than those two quantities may differ substantially. Nevertheless, a high value (ideally equal to one) of T+ is desired for an effective battery operation. In the upcoming subsections four approaches to an estimation of T+ are presented. The outcomes of the chronoamperometric and the EIS measurements (from measuring steps marked by {1}, {2}, {4}, {5} and {7} on Fig. 5) serve as the input to these approaches. ## Steady-state to initial current ratio (ISS █) In this approach the T+ of the bulk is calculated as a ratio of the steady-state to the initial current . It is the most primitive approach. Mathematically, it is expressed by the formula: Where: • IS -"steady-state" or "stationary" electric current. It is a steady in time electric current measured after applying a small potential difference (ΔV in (1)) to the system. For a typical polymer electrolyte, it takes quite some time for transient effects to die out, and thus to converge to the steady state (see current vs time plot on Fig. 7b). To esti-mate the value a statistically significant number of the steady-state readings was averaged to obtained more precise value of IS. • I0 -"initial" electric current. It is an electric current measured at the beginning of the salt concentration gradient build up in the bulk of the electrolyte. To assess this moment, the analysis of the EIS data was carried out (see Fig. 7b and description of the τopt at the end of the preceding section). Remark: Typically, by this approach, the analysis of EIS spectra was not carried out and thus, the I0 was equal to the first/maximal output of an ammeter (or an oscilloscope or a potentiostat/galvanostat), that is, the current value measured immediately after the application of the perturbation. Clearly, this methodology introduces additional errors resulting from maximal sampling rate of the equipment, thus it leads to the indeterminacy of such calculated T+. The ISS approach neglects/disregards the influence of the electrode-electrolyte interface on the value of T+. It also assumes that the electrochemical characteristic of that interface is unaffected by a change in the external conditions i.e. by the perturbation or the temperature. ## By Sørensen-Jacobsen & NLT (SJ NLT █) Equation ( 17) is the SPP adjusted version of the formula used by Sørensen and Jacobsen in 1982: k=2 , 5 ; m=4 , 7 ; ( Where: • Rb -bulk electrolyte resistance, • Rd -diffusional resistance of the bulk electrolyte. The Rd is an additional resistance of the bulk electrolyte due to the steady-state salt concentration gradient induced by the potentiostatic polarization. It is represented by an anomalous resistance parameter of the W s b denoted as a R(W s b) (see Fig. 7b and description in the previous section). In practice, the R(W s b) should be approximately equal to (ΔV/IS -Rb -Ri). The Rd value is evaluated based on the data from the CA & the EIS measuring steps and the Rb is estimated based on EIS data only. The superscripts in formula (17) point to a source of input data (the measuring steps). As indicated, the CA measuring step always precedes the EIS measuring step. Such order is intended to minimize the influence of possible changes in the system properties (i.e. due to very long CA measuring step). In the author's opinion, the SJ NLT approach is the most appropriate way to calculate the T+. It verifies the agreement between the EIS and the CA data, and is based on a thorough analysis of the impedance spectrum covering a broad frequency range. ## By Bruce-Vincent (BV █) In 1989 the T+ measurement approach, combining EIS and chronoamperometry, was described . The EIS was used to deconvolute the contributions of high rate processes, like those occurring at the electrolyte-electrode interface. The chronoam-perometry was applied to characterize more sluggish processes, namely the diffusion. The formula for T+ calculation was also presented: Where: • ΔV -potential perturbation imposed upon the measuring cell during a chronoamperometric measurement (same as in (1)), • I0 and IS are described in the ISS section (3.3.1.). ## • Ri0 and RiS are the interface resistances (similarly to subscripts in I0 and IS, characters "0" and "S" mean "initial" and "steady state" respectively). Formula ( 17) accounts for the influence of the electrolyte-electrode interface, thus is a much more advanced approach than the equation ( 16). The major difficulty of this approach lies in proper estimation of the RiS. The EIS measurements in conditions other than open circuit potential (i.e. with a bias/offset potential as in the HF EIS) are much more difficult to interpret. Moreover, such biased EIS distorts the steady state, especially when it is performed to low frequencies. This in turn may affect the validity of IS estimation or considerably elongate the duration of the measurement routine. Remark: For systems and experimental conditions discussed herein (i.e. a small perturbation and salt concentration gradient, no convection, a linear, stable and causal system response) the relations based on the Ohm's law should be valid: The (19) and (20) relations make the equations ( 17) and (18) equivalent to each other. ## By Watanabe (W █) One can omit an estimation of the I0 and Ri0 by merging (17) together with (20), to obtain the formula similar to one presented by Watanabe et al. : In general, for systems in which properties may evolve/fluctuate with time the interface resistance should be inferred based on EIS measured at an open-circuit potential and after the potentiostatic polarization from which the IS was obtained. It is also of great importance to precondition the measuring cell with fresh sample to accelerate convergence to a relatively stable thermodynamic state. ## Ionic conductivity (σ █ ) The ionic conductivity of electrolyte is estimated based on data obtained by EIS (from measuring steps {4} and {7}, see Fig. 5) and the geometrical dimensions of the sample. In case of polymer electrolyte the bulk solution resistance (Rb) is obtained by fitting an electrochemical equivalent circuit showed on Fig. 7a and Fig. 7b. Then the average of two resistances is calculated, as it is expressed by the formula: Finally, the ionic conductivity is calculated according to the formula: (23) Where: • L -thickness of the sample (cm), • A -contact area (apparent) between a single electrode and the sample (cm 2 ). For a liquid electrolyte soaked into the battery separator, the ionic conductivity is calculated according to the formula: (24) Remark: In previous work , it was shown that the bulk conductivities measured with use of cation (i.e. lithium) reversible/transmissive electrodes were similar to those estimated with blocking/reflective (i.e. stainless steel) electrodes. ## Apparent diffusional conductivity (σd █ ) The apparent diffusional conductivity (conductivity based on the DS) is calculated via: Where: • F -Faraday constant (96485.33212 C⋅mol −1 ), • R -gas constant (8.31446261815324 J⋅K −1 ⋅mol −1 ), • e -elementary charge (1.602176634×10 −19 C), • kB -Boltzmann constant (1.380649×10 -23 J⋅K −1 ), • NA -Avogadro constant (6.02214076×10 23 mol −1 ), • T -absolute temperature (K), • cS -molar concentration of salt (mol⋅cm −3 ), • DS -effective salt diffusion coefficient calculated by the "new" approach via (12) and then averaged as in (2) (cm 2 /s). Additionally, the apparent diffusional conductivity amendment (χd) is defined as: Where: • T+ -apparent cation transference number calculated by (17) and then averaged as in (3), • cT -total solution concentration, sum of ionic species and monomers in solvent (mol/cm 3 ), • c0 -monomers in solvent concentration (mol/cm 3 ), • γ± -mean molal activity coefficient of the mobile species, • m -molality of electrolyte (mol/kg). The derivation of the equation (26) and the underlying assumptions are presented in Appendix C. ## Effective conductivity (σ+ █ ) The effective (cationic) conductivity is calculated according to the formula: Where: • T+ -apparent cation transference number calculated by (17) and then averaged as in (3). ## Limiting-current density (Jlim) Formula ( 28) was used to calculate the limiting-current density: Where: • L * = 100 µm -reference thickness of an electrolyte sample. The value was deliberately chosen to facilitate possible conversions. Originally, this formula was derived to be used with the lithium transference number rather than T+ and for a Fickiantype diffusion . Based on below presented results and considerations, this is apparently not the case, but the Jlim should still be a good and useful estimator. The limiting current (Ilim) could be calculated based on Jlim: The Ilim calculated via ( 29) is used to check whether acceptably small current (less than 5% of the Ilim) flowed through each sample during the potentiostatic polarization measuring steps. Fulfillment of this condition allows to neglect salt concentration variations near the electrodes. In such cases only the ohmic potential drop and the interface overpotentials determine the distribution of currents . ## RESULTS AND DISCUSSION In the present work the measurements were performed at temperatures between 60 °C and 110 °C. This temperature range was deliberately chosen to minimize the influence of heterogeneity (i.e. crystal grains) or phase transitions in the electrolyte structure. For some samples, temperature range was narrower because of materials used or experimental errors. To facilitate comparisons between examined electrolytes, ten unique symbols are used to present the results on the following figures. Each symbol shape is associated with one composition, as summarized in Tab. 1. and Tab. 2.. Dotted or dashed lines linking data points are guides for the eye only. For every data point, the experimental uncertainty has been calculated and expressed by the error bars (sometimes barely visible due to small error). ## Effective salt diffusion coefficient (DS) Based on the EMF decays, the "common" and the "new" approaches (see section 3.1.) were used to estimate the effective salt diffusion coefficients (DS). The difference between outcomes of these approaches can be compared by studding pairs of figures. Figures 8a and 9a depict results by the "common" approach, while figures 8b and 9b are showing results by the "new" approach. Each DS on the "a" figure has its equivalent on the "b" figure. Recalling: the equivalents resulted from the same input data i.e. the same subrange of each EMF decay. The difference in value between the equivalents is solely caused by the approach used to process the data. It is also worth noticing, that the DS obtained by the "new" approach are independent of a deliberately chosen subrange, in contrary to the DS evaluated by the "common" approach. Such independence is due to the linearity of the long-time-part of the plot shown on on Fig. 6e. Therefore, only the DS by the "new" approach were used to calculate σd, Jlim and Ilim. The secondary ordinate axis on the "a" figures is related to the time at equivalence point (τEP, light-gray points) determined in a way described in the section 3.1. Whereas, the secondary ordinate axis on the "b" figures refers to the relaxation order values (n, gray points). ## Electrolytes consisting of LiTFSI and PEO The DS, τEP and n of four polymer electrolytes consisting of the LiTFSI salt and the PEO polymer (see Tab. 1.) are presented in this part. The results are grouped by the two approaches. • The "common" approach ( █ ) The results of the "common" approach are depicted on Fig. 8a and listed below. At 70 °C on Fig. 8a, the DS values, ordered from highest to lowest, are: • EO16TF -30.97 ± 1.02 × 10 -9 cm 2 s -1 , • EO10TF -25.63 ± 5.28 × 10 -9 cm 2 s -1 , • EO5TF -17.25 ± 2.73 × 10 -9 cm 2 s -1 , • EO3TF -6.72 ± 0.80 × 10 -9 cm 2 s -1 . The DS of the EO16TF sample are changing the least with temperature and they have the highest values at low temperatures. Apparently a trend can be recognized, the more LiTFSI salt in polymer electrolyte contains, the greater the variation of DS with temperature is. The τEP plots have a linear character, similar to the DS, but the dependence on temperature is reversed. The overlapping of the τEP plots is smaller compared to the DS, because the τEP additionally account the diffusion length of the sample. • The "new" approach ( █ ) Next, on Fig. 8b and below, the results due to the "new" approach are presented. At 70 °C the DS values, ordered from highest to lowest, are: • EO16TF -71.13 ± 1.26 × 10 -9 cm 2 s -1 , • EO10TF -39.62 ± 2.13 × 10 -9 cm 2 s -1 , • EO5TF -26.43 ± 3.04 × 10 -9 cm 2 s -1 , • EO3TF -8.69 ± 0.89 × 10 -9 cm 2 s -1 . In general, the order and the temperature dependencies of DS between the samples and the approaches are quite similar, but the values calculated by the "new" approach are larger. The relaxation orders of the EO16TF sample are the highest among the LiTFSI containing electrolytes. Apparently, the temperature variation of the relaxation orders increases with the salt content. This probably indicates a gradual change in the ion transport mechanism. The relaxation orders of all measured samples are greater than one, what appears to be an indicator of the anomalous diffusion i.e. the subdiffusion. The subdiffusion of polymer segments in polymer melts has been recognized and measured long time ago . The physics behind those observations was also figured out and descried mathematically. At short times (between the shortest and the longest relaxation time of a Rouse chain, the latter often called the Rouse time) the subdiffusive motion of the segment results from a Rouse-like relaxation process of a linear polymer chain. At longer times (between the Rouse time and the disentanglement time) and for long, entangled chains the exponent of subdiffusion is better quantified by the Tube Model . However, according to the author's knowledge, the subdiffusive motion of ions in polymer electrolytes was not recognized experimentally. However, the Molecular Dynamics simulations (MD) showed that the subdiffusion of ions should exist in short-chain-PEO and LiTFSI system. By these simulations, it was discovered that the motion of the lithium cation (Li + ) should be in part subdiffusive. It was postulated that Li + subdiffusion is due to a Rouse-like relaxations of a polymer segment with which the cation is coordinated. Two mechanisms of Li + motion have been identified. The first one was due to breaking of "most outer" coordination bond, one out of five, and then formation a new bond on the "other side". The breaking and formation is probably triggered by the longitudinal Rouse-like modes. By this mechanism Li + subdif-fused along the polymer chain. The second one was an effect of Li + being transported by the coordinated segment. During such transport the coordination bonds between Li + and the segment did not change. This motion is likely caused by the transverse Rouse-like modes. These two subdiffusion mechanisms are intertwined by diffusive-in-character hops of Li + to other segments or to low energy sites present in the cations surrounding. For electrolytes with high salt concentration those hops were predominantly assisted by anions. It was also pointed out that the motion of anions is strongly coupled with PEO ether oxygen atoms (EO) displacements. Herein, for samples with high salt concentration, the relaxation order rapidly decreases with temperature towards value equal to one, thus towards the diffusive behavior. Such trend presumably indicates that when sample crystallizes the Rouse-like conformations are frozen and the motion of ions becomes purely diffusive. Clearly, there are additional factors that are not accounted in the MD simulations, but may contribute to the subdiffusive behavior. Herein, the polymer matrix consists of much longer (than in the MD) polymer chains, which are entangled. The center of mass of these chains is practically immobile. The system is not as homogeneous and "pure" as the one modeled in the MD, thus it is possible that some of the ionic species get immobilized (trapped) in local electrochemical potential minima (i.e. percolation network cul-de-sacs/pockets etc.) for waiting-times that constitute the probability density function with long tail. Such shape of this function is characteristic for a subdiffusion . For polymer electrolytes the longest waiting time should be limited by a time evolution of the polymer matrix (i.e. by the "renewal time") as predicted by the Dynamic Bond Percolation theory or, indirectly, by the Reptation Theory . It is also possible that the aforementioned interfaces (i.e. the layered structures or the PEO matrix polydispersity described in section 3.2.) may mimic the subdiffusive response. Nonetheless, it is assumed that the contributions by these additional factors are of little account. To the author's knowledge, this is the first communication that presents a measure of diffusion domain in the restricted diffusion experiment, thus the results could not be compared. Also a more detailed studies are needed to discover causes of described behavior. ## Electrolytes containing a LiATAB salt variant Similarly, the DS, τEP and n of six electrolytes based on three borate salt variants are shown. The electrolytes consist of the LiATAB salts with or without the PEO polymer (see Tab. 2.). ## • The "common" approach ( █ ) The results obtained by the "common" approach are summarized on Fig. 9a. At 70 °C the EO27AB7 sample has the highest value of DS which is equal to 13.78 ± 0.51 × 10 -9 cm 2 s -1 . At this temperature the DS of remaining samples are aligned in descending order: • EO29AB1 -12.03 ± 1.41 × 10 -9 cm 2 s -1 , • EO23AB3 -9.33 ± 1.02 × 10 -9 cm 2 s -1 at 73 °C, • AB3 -2.50 ± 0.21 × 10 -9 cm 2 s -1 , • AB7 -1.89 ± 0.10 × 10 -9 cm 2 s -1 , • EO9AB7 -3.31 × 10 -9 cm 2 s -1 (This DS value was obtained by a linear extrapolation based on two data points plotted in the Arrhenius representation). Remark: Postmortem inspections of the cells with EO9AB7 sample (LiATAB+PEO n=7 EO:Li=9:1) exposed a greatly deteriorated surface of the electrodes. Probably, not-fully evaporated acetonitrile in the sample was the culprit. The acetonitrile leftovers continued to evaporate from the sample inside the assembled measuring cell and they reacted with the lithium electrodes. Additionally, at higher temperatures and during the electric perturbations the evaporation and the unwanted reactions proceeded at a faster pace. The persistently increasing resistivity of the system and the vast corrosion of the electrodes seem to confirm the above scenario. Even during extremely long potentiostatic polarizations, this system did not reach the steady state. Only the outcomes measured at the highest temperatures showed an acceptable level of uncertainty. The correlation between the τEP and the DS has similar character to the one described in the subsection 4.1.1. • The "new" approach ( █ ) The results due to the "new" approach are depicted on Fig. 9b. At 70 °C EO27AB7 sample has the highest value of DS which is equal to 85.94 ± 0.54 × 10 -9 cm 2 s -1 . At this temperature the DS of remaining samples are aligned in descending order: • EO29AB1 -28.41 ± 3.26 × 10 -9 cm 2 s -1 , • EO23AB3 -23.63 ± 0.66 × 10 -9 cm 2 s -1 at 73 °C, • AB3 -4.41 ± 0.21 × 10 -9 cm 2 s -1 , • AB7 -3.99 ± 0.47 × 10 -9 cm 2 s -1 , • EO9AB7 -2.95 × 10 -9 cm 2 s -1 (obtained by linear extrapolation in Arrhenius representation to 2.91 × 10 3 K -1 ). ## Fig. 9b. Arrhenius-type plot of the salt diffusion coefficients (DS, violet, the "new" approach, see 3.1.2., left ordinate, decimal logarithm) and the order of relaxation (n, gray, right ordinate) for the electrolytes containing LiATAB salts. Each symbol is related to an electrolyte composition (see Tab. 2.). Before discussing the results showed on Fig. 9b in general, let us focus our attention on two data points of the EO29AB1 sample which are indicated by arrows. These points were obtained from two successive SPPs both performed at 80 °C. During the first SPP the absolute value of perturbations (|ΔV| 1st ) was set to 30 mV and the absolute value of the first EMF values (|E0| 1st ) was approximately equal to 26 mV. During the second SPP the |ΔV| 2nd was equal to 10 mV and |E0| 2nd ≈ 9 mV respectively. When we compare the |E0| 1st to the |E0| 2nd in case of those two SPPs, we notice that the ratio |E0| 1st / |E0| 2nd ≈ 289%, but the ratio DS 1st / DS 2nd ≈ 108%. Based on this observation a conclusion can be made: The DS by the "new" approach is a property which is independent of potential-perturbation magnitude (for perturbations fulfilling condition (1)). For comparison, the ratio DS 1st / DS 2nd by the "common" approach (Fig. 9a) is on the order of 110%. It is also worth mentioning that the first SPP was preceded by the SPP during which the temperature of the sample was set to 90 °C. In contrast, in case of the second SPP, the preconditions were different. There, the Temperature Setting measuring step could be omitted because the sample was already well equilibrated at 80 °C. This difference in preconditions should explain why the ratio DS 1st / DS 2nd is not equal to 100%. Additionally, the relaxation orders of these successive SPPs are very similar, what may be viewed as an another argument in favor of the "new" approach. Same as in the case of the DS, the relaxation order is a property which is independent of a magnitude of |ΔV|. In general it could be concluded that dissolving the Li- ATAB salt in the PEO increases the DS. This effect is clearly visible by comparing electrolytes containing the same salt variant, i.e. AB3 vs. EO23AB3 or AB7 vs. EO27AB7. Presumably, the PEO chains compete with the anions in the Li + coordination. If the concentration of the chains is sufficiently high, the electrostatic interactions between ions are efficiently shielded. Also, long chains constitute an immobile 3D mesh which hampers the motion of big ionic species. The relaxation orders of the EO27AB7 sample are considerably greater than those of the remaining LiATAB samples. It is reasonable to assume that for this sample a 3D mesh made of PEO is dense enough to successfully block the motion of fairly large and bulky ionic species. High concentration of PEO provides an environment for Li + to percolate and effectively shields the cations from interactions with other ionic species. The relaxation order can also be used to explain the influence of the anion size on the diffusion. The EO27AB7 sample has n approximately equal to 2.2 and for the EO29AB1 sample n≈1.6. These samples have similar molar content of PEO (EO:Li values are 27.5:1 and 29:1 respectively) but they differ in the borate salt variant. Recalling, in EO27AB7 the anion has long oligomer arms that comprise approximately n≈7 oxyethylene substituents and in EO29AB1 the anion has the arms that are much shorter (n=1, see Fig. 3. and Tab. 2.). Two phenomena may be responsible for the difference in the subdiffusion extent. First, at a given PEO concentration, the longer the oligomer arms of the anion are, the more prone the anion is to be slowed down or even trapped by a dense PEO mesh. Second, such trapped anion or an ion cluster may additionally confine Li + in pocket/cul-de-sac in between the oligomer arms. The probability of such confinement should increase with the length of the arms. The confinement of Li + may be particularly effective in the vicinity of the boron atom. The relaxation orders of the neat salts (AB3 n=3 and AB7 n≈7) look particularly puzzling. Surprisingly, they are considerably greater than one (n≈1.5 and n≈1.6 respectively). Similarly as mentioned above, the Li + in these neat salts may be confined/"slowed-down" by the oligomer arms. Also, in addition to the translational motion, the anions may have to rotate to move, what may contribute to the subdiffusion. The structure of the battery separator may also be the culprit. In particular, its greater than one tortuosity or the Knudsen friction which differs between ions may have an impact on the motion. The trapping caused by an aggregation and clustering seems also a plausible cause. Broader and more advanced measurements, reaching beyond the scope of the present work, are needed to clarify the microscopic causes of such behavior. ## Apparent cation transference number (T+) The T+ presented in this section were calculated according to the four approaches described in section 3.3. On the plots below, a color of the symbol corresponds to the approach, namely: • yellow (ISS █) as in ( 16), • orange (SJ NLT █) as in ( 17), • purple (BV █) as in ( 18), • green (W █) as in (21). Figures have a form of the apparent cation transference number plotted against the reciprocal of the absolute temperature. ## Electrolytes consisting of LiTFSI and PEO The T+ of four polymer electrolytes consisting of the LiTFSI salt and the PEO polymer (see Tab. 1.) are presented on Fig. 10 and discussed bellow. The T+ values obtained by the ISS approach are significantly greater than those obtained by the three remaining approaches. Moreover, in case of the EO16TF and the EO10TF samples the T+ by the ISS approach show a reverse temperature dependence. The dissimilarity of the T+ by the ISS approach occurs, even though the influence of the interface is partially taken into account (see (15) and ( 16)). At 70 °C the EO3TF sample gives the highest values of T+: • SJ NLT -0.278 ± 0.023, • W -0.282 ± 0.021, • BV -0.289 ± 0.015. At the same temperature the T+ values of the remaining samples are more more similar in value: • EO16TF: • SJ NLT -0.112 ± 0.004, • W -0.100 ± 0.005, • BV -0.166 ± 0.080, • EO5TF: • SJ NLT -0.072 ± 0.006, • W -0.072 ± 0.009, • BV -0.104 ± 0.001, • EO10TF: • SJ NLT -0.076 ± 0.003, • W -0.066 ± 0.001, • BV -0.094 ± 0.004. The T+ values obtained by the SJ NLT approach are quite similar to those derived by the W approach. The BV approach outputs somewhat greater T+ values, and the discrepancy in EO16TF case is quite significant. This particularly big discrepancy is prescribed to human errors in the chronoamperometric measurement settings, which lead to an inaccurate "initial" current values. The errors were discovered in the course of much later analysis of the measurement outcomes, when there was no possibility to repeat the measurements. In general, T+ obtained herein (by further modified approaches) are somewhat lower than those from the previous work but the trend in concentration dependence is similar. At first, with increasing salt concentration, the T+ gradually decreases until it reaches a minimum at around EO:Li=6:1. From there, the values of T+ start to grow rapidly with rising salt content. Based on above results and a scientific literature study, it can be concluded what follows. At low salt concentrations, each Li + is solely coordinated by five ether oxygen (EO) atoms (on average) of the PEO chains . The Li + and TFSI − are the dominant ionic species. High flexibility of the chains is maintained because of relatively small number of Li + coordination bonds. Interactions between ions are greatly diminished owing to a substantial spatial separation and the screening effect provided by EO . Ions migrate almost independently from each other. The motion of cations is mainly induced by the PEO conformations. Although, at low salt concentration regime, the transport of cations is relatively very effective due to high elasticity of the chains, a low T+ is observed. The reasoning behind such observation is that the interactions of PEO with TFSI − are much weaker than those with Li + . Also, the anion has more than six times greater molecular mass than the PEO monomer, thus it has much greater inertia (oddly enough, the molecular mass of 6 PEO monomers and Li + is almost equal to one of TFSI − ). The direction of Li + motion is to a great extent determined by the arrangement of PEO segments surrounding the cation. Therefore, the migration path of the cations is much more tortuous than the one of the anions, what likely translates to a considerably higher mobility of the latter. With increasing salt concentration, a PEO chain capability of Li + coordination gets progressively depleted. At around EO:Li=6:1 the chains are saturated, thus cannot solely complex more lithium. At this concentration the motion of Li + along the chains is almost brought to a halt. The coordination bonds affect the effective segment length and decrease chain conformational freedom , thus the characteristics of Rouse-like relaxations are changed. Further concentration rise perpetuates ion clusters emergence and their gradually growth in size. Also, at higher salt concentrations, the clusters have overall negative charge . The structure becomes more compact and rigid, what in turn affects at most the motion of spices with large size. Presumably at very high salt concentrations due to crystallization the hopping of cations becomes the dominating mechanism of charge transport. For the LiATAB electrolytes, the T+ calculated by the BV, W and SJ NLT approach are quite similar to each other. Also, the influence of the interface on T+ value is not as severe as it is in the case of LiTFSI+PEO systems. To facilitate the recognition of samples, the T+ values obtained by SJ NLT approach at 70 °C are listed (aligned in a descending order): • EO27AB7 -0.331 ± 0.014, • EO23AB3 -0.149 ± 0.011 at 73 °C, • EO29AB1 -0.066 ± 0.002, • AB7 -0.048 ± 0.010, • AB3 -0.035 ± 0.002. Additionally, based on the linear extrapolation, the T+ of the EO9AB7 sample has been evaluated to be equal to ca. 0.303. The order in which the T+ values are aligned supports the following hypothesis: The greater the size of the anion and the denser the network of an entangled, thus immobile, PEO chains, the higher the T+. Anions with long oligomer arms seem to have a great difficulty in switching between positions due to necessity of squeezing through holes in polymer mesh. In case of the "big" anions (see Fig. 3) in a dense PEO matrix, the rise in temperature causes substantial increase in the T+. Such effect is probably caused by a decrease in enthalpy of mixing between anion and cation. Also, the pace of conformation change, especially of PEO chains, but also of the oligomer arms is more rapid. Presumably, polymer electrolytes with short-oligomer-arms anions show low T+ values caused by a high mobility of anionic species. ## Ionic conductivity (σ █ ), Apparent diffusional conductivity (σd █ ), Effective conductivity (σ+ █ ) The ionic conductivities (σ) were calculated according to equations ( 23) and (24). The apparent diffusional conductivities (σd), thus the conductivities based on the DS by the "new" approach, were evaluated by using (25). The effective (cationic) conductivities (σ+) were obtained using the SJ NLT approach T+, thus by (27). ## Electrolytes consisting of LiTFSI and PEO The σ, σd, σ+ of four polymer electrolytes consisting of the LiTFSI salt and the PEO polymer (see Tab. 1.) are presented on Fig. 12 and discussed bellow. To facilitate the recognition of the results on Fig. 12, the values obtained at 70 °C are listed (aligned in descending order): • σ: • EO16TF -4.94 ± 0.06 × 10 -4 Ω -1 cm -1 • EO10TF -4.61 ± 0.05 × 10 -4 Ω -1 cm -1 • EO5TF -1.90 ± 0.02 × 10 -4 Ω -1 cm -1 • EO3TF -0.45 ± 0.01 × 10 -4 Ω -1 cm -1 • σd: • EO16TF -5.85 ± 0.10 × 10 -4 Ω -1 cm -1 • EO5TF -5.02 ± 0.58 × 10 -4 Ω -1 cm -1 • EO10TF -4.62 ± 0.25 × 10 -4 Ω -1 cm -1 • EO3TF -2.16 ± 0.22 × 10 -4 Ω -1 cm -1 • σ+: • EO16TF -0.55 ± 0.04 × 10 -4 Ω -1 cm -1 • EO10TF -0.35 ± 0.04 × 10 -4 Ω -1 cm -1 • EO5TF -0.14 ± 0.03 × 10 -4 Ω -1 cm -1 • EO3TF -0.13 ± 0.01 × 10 -4 Ω -1 cm -1 The σ on Fig. 12 and those from the previous work show that a maximal value of the ionic conductivity is presumably located somewhere between EO:Li=16:1 and 10:1. The location of this maximum is supported by the Flory-Huggins mean-field theory . According to this theory, the change in the entropy of mixing (per lattice site and for flexible, monodisperse, linear polymer chains mixed with solvent molecules of a monomer size) is proportional to: Where: • Φ -volume fraction of the polymer, • N -number of lattice sites taken by the polymer. Let us assume (very roughly) that: • Φ is related to the salt mass fraction as Φ ≈1-WS, • N is proportional to the number of monomers in the chain, and that: • the chain has the entanglement molecular weight (Me). For the PEO, the Me=2000g/mol . Thus, the N is approximately equal to 45. By performing mathematical operations on the (30), we obtain that the maximum of the entropy is located at Φ≈0.636, thus WS≈0.364 or EO:Li≈11:1. Clearly, this value is a very rough estimate, because of the approximations, and because the influence of the enthalpy of mixing was not comprised here. In polymer electrolytes, in addition to the theory, the enthalpy should also account for the electrostatic ion-ion interactions. For the EO16TF sample the σ are quite similar to the σd. Such similarity is apparently an indicator of a well dissociated salt. Additionally, the molar conductivities of this sample are the highest among the samples. The motion of ions in this sample appears to be uncorrelated. Also, in case of the EO10TF sample the σ and σd are alike, but the molar conductivities for this sample are somewhat lower. For the EO5TF sample the differences are much bigger (the σd to σ ratio is equal to ca. 3). Presumably, ion aggregates are additionally present in this sample. For the EO3TF sample the ratio is equal to about 7. This suggests that the level of ion association is high, thus to the correlation between the motion of ions is strong. Additionally, the χd was calculated (according to (26)) in an attempt to have a better sense of the deviation from the Nernst-Einstein relation. The estimation was made only for PEO+LiTFSI system and only at one temperature equal to 90 °C. Because the evaluation of the thermodynamic factor is not incorporated in the present work, the values (these in the square brackets below) were extracted from Fig. 3b of Pesko et al. • EO3TF -0.76 • [5.0] -1 = 0.15 Noticeably, all χd values are less than one. This apparently means that the motion of ions is less correlated than concluded based on the σd to σ ratio. But, because of many simplifying assumptions (see Appendix C.) and the thermodynamic factor measured elsewhere, high uncertainties of these χd are expected. As stated above, significant differences between σ and σd presumably indicate a high concentration of ion pairs or other ion aggregates. In such cases additional measurements can be made at varied perturbation values (|ΔV|). The results of such measurements should provide better insight into the type of the dominating species . The plots on Fig. 12 look linear, thus a thermally activated charge transport can be recognized (i.e. Arrhenius-type transport). To facilitate the comparison of an energy landscape, the apparent activation energies (Ea) were extracted from Fig. 12 into Tab. 3. Each Ea listed in that table is proportional to the slope of a linear function fitted to the data points. Clearly, the Ea values shown in Tab. 3. are only a rough estimate of an actual energy landscape. The Vogel-Fulcher-Tamman fittings should give a more accurate view. Nevertheless, due to a relatively small number of data points such analysis was omitted. To further clarify, for every type of conductivity presented on Fig. 12, a type of Ea was extracted. The ionic Ea values stem from the fittings to the σ data points, the diffusional Ea are based on the σd data and the cationic Ea is related to the σ+. The EO16TF sample has the lowest Ea values and the values between the types are also very similar. Such observations additionally confirm the superiority of this electrolyte. A gradual increase of each type of Ea with increasing salt concentration is also noticeable. Based on possessed knowledge, the influence of salt content in LiTFSI+PEO polymer electrolytes can be summarized as follows. Long and entangled polymer chains constitute a relatively immobile matrix. At low salt concentrations, the LiTFSI salt is well dissociated by the PEO. Each of the Li + is preferably coordinated to 5 ether oxygen atoms (EO) , therefore there are plenty of "free" EO available for Li + coordination at low salt concentrations. Above the melting temperature the PEO, the chains are very flexible, but the Li + coordination bonds make the section of the chain more rigid . Such section has presumably the length of 6 consecutive EO for a coordination by a single chain. The Li + shifts position mostly with assist of the chains via the subdiffusion. The direction of Li + motion is highly influenced by the alignment of a chain segment to which the Li + is currently bonded to. A large spatial separation and a the screening effect of the EOs diminish the ion-ion electrostatic interactions . A relatively small positive charge of PEO hydrogen atoms and a considerable charge delocalization on TFSI − contribute to a more fluid and less tortuous motion of the TFSI − . Additionally, there is a substantial difference in the inertia because of more than six times higher molecular mass of a TFSI − compared to a PEO monomer. Therefore, the majority of TFSI − move swiftly in the direction of the electric field. Nevertheless, some of the TFSI − may be trapped by the polymer matrix (i.e. especially in the cross-linking points) for long waiting times. The trapping could be caused by a large size and bulkiness of the TFSI − , and thus may contribute to the subdiffusion. With increasing salt concentration the influence of Rousetype conformations on the Li + motion is fading, but the volumetric number of ions rises. A high number of ions is a prerequisite of a high ionic conductivity. At EO:Li=6:1 ratio, energetically favorable structures consisting of pairs of PEO coils which interlock to form cylinders with inner cation channels are believed to be predominant . Such chains arrangement apparently hampers an efficient Li + motion by impairing the effect of Rouse-type chain conformations. A constructive synchronization of chain conformations in both coils seems possible but unlikely, also the orientation and "clogging" of the channels may be an issue. Above the salt concentrations of EO:Li=6:1, additionally to Li + and TFSI − , other ionic species (i.e. ion pairs, triplets etc.) begin to emerge and play an important role in charge transfer between electrodes. In this concentration regime an addition of salt not necessarily translates to an increase of number of charge carries. A sole Li + transport via a percolating network of the chains becomes rare. The Li + are forced to frequently modify their coordination environment between the EO atoms and the oxygen atoms belonging to the TFSI − . Frequent changes of the coordination character further impedes a motion of the Li + , thus most of the charge is transported with assist of anionic species. ## Electrolytes containing a LiATAB salt variant At 70 °C the AB3 sample has the highest ionic conductivity (σ), equal to 4.84 ± 0.29 × 10 -5 Ω -1 cm -1 . At this temperature σ of the remaining samples are aligned accordingly: • AB7 -1.25 ± 0.16 × 10 -5 Ω -1 cm -1 , • EO29AB1 -9.02 ± 1.10 × 10 -7 Ω -1 cm -1 , • EO27AB7 -5.90 ± 0.29 × 10 -7 Ω -1 cm -1 , • EO23AB3 -1.29 ± 0.10 × 10 -7 Ω -1 cm -1 at 73 °C, • EO9AB7 -2.81 ± 0.72 × 10 -8 Ω -1 cm -1 . Substantial differences between σ and σd of the polymer electrolytes (EO29AB1, EO27AB7, EO23AB3, EO9AB7) can be noticed on Fig. 13. As already mentioned above, such differences likely indicate a poorly dissolved salt. In contrast, σ and σd of the neat borate salts (AB3, AB7) are relatively similar to each other. Probably, the PEO chains disturb a spatially symmetric disposition of identical interactions acting upon each ion in the neat salts. Thus, the chains facilitate an emer-gence of ion pairs or even larger aggregates . In particular, species that have no net charge (like ion pairs, etc.), don't contribute to the ionic conductivity, but they should participate in the diffusional conductivity (σd). For the case of the borate salts, and especially for variants with long oligomer arms, ion pairs should have similar size as ATABanions. Furthermore, the denser the PEO matrix is, the better it is in restraining the motion of species based on their size. Probably, for LiATAB polymer electrolytes, the electrode reactions are the main source of relatively small ionic species i.e. Li + and ATAB -. Also, the battery separator, used in case of AB3 and AB7, should be quite effective in diminishing convection but it may not inhibit the motion of large size ionic species. The EO27AB7 sample has overall the lowest Ea values among the LiATAB salt containing samples, thus it has overall the best transport properties among LiATAB electrolytes examined here. ## Limiting-current density The limiting-current densities (Jlim) are presented on Fig. 14. They were calculated using (28). To facilitate possible conversions, in those calculations the reference electrolyte thickness was deliberately chosen to have 100 µm. Overall, as a result of better salt dissociation, the LiTFSI based polymer electrolytes have higher limiting-current densities than electrolytes based on the LiATAB salts. Apparently, the high amount of large ionic aggregates significantly reduces the number of charge carriers in the latter electrolytes. ## CONCLUSIONS In the beginning of the present work, the details of Symmetric Polarization Procedure were described. The SPP is an ordered set of measurements, which in addition to a precise estimation of the properties, facilitates a simple and fast detection of experimental mistakes. These mistakes are discovered by a visual examination of a symmetry in the SPP outcomes. Thus, an information about any inconsistency is available instantaneously and the SPP can be repeated or the additional measurements can be performed to explain the source of confusion. Thanks to this simple preliminary validation task an ad hoc advanced data analysis can be avoided and the outcomes inspection can be made by less qualified staff. Also, a new approach to the analysis of restricted diffusion measurement was proposed, together with a quantitative measure of the diffusion domain, namely the relaxation order. This approach not only improves the precision of evaluated salt diffusion coefficient, but also it provides an important information on the character of the salt gradient relaxation. Such accurate and complete information should lead to the design of more optimal battery charging/discharging protocols as well as more precise electrochemical sensors. Herein, in case of all examined electrolytes, the subdiffusive ion motion was observed. Two presumable reasons were given as the sources of such motion: the motion of lithium cations coordinated by wriggling polymer chains and the trapping events of ions. Finally, it has been shown how the combination of the apparent cation transference number, the salt diffusion coefficient and the ionic conductivity, can be used as a proper indicator of an electrochemical fitness of an electrolyte. Another characteristic time is correlated with a salt diffusion at the restricted length (τd) given by formula below: The τEP derived from the short-time approximation: ## C. Diffusional conductivity Based on the Nernst-Einstein relation we can write the diffusional conductivity expression: Additionally, the transference numbers (with respect to solvent) in concentrated solution theory are defined as: Combining ( 4) with (C.1) and (C.2) gives: Assuming that t+ can be approximated by T+ in (C.3) and utilizing (25) we get:

chemsum

{"title": "The electrochemical measurement of salt diffusion coefficient, apparent cation transference number and ionic conductivity for a thin-film electrolyte", "journal": "ChemRxiv"}

an_activity-based_fluorescent_sensor_for_the_detection_of_the_phenol_sulfotransferase_sult1a1_in_liv

## Abstract: The biological activation and incorporation of inorganic sulfate proceeds via a process known as sulfurylation. Transfer of a sulfuryl moiety from the activated sulfate donor, 3'-phosphoadenosine-5'-phosphosulfate (PAPS), to hydroxy-containing substrates by human phenol sulfotransferases (SULT1 family) alters substrate solubility and charge to affect the metabolism of endogenous metabolites, xenobiotics, and drugs. Current methods to monitor SULT1 activity in living cells primarily rely on radiolabeling and/or cell extractions, but these methods do not provide a direct readout of enzyme activity with a dynamic, temporally resolved spatial map in live, intact cells. To fill this gap, here, we present the development, computational modeling, in vitro enzymology, and biological application of Sulfotransferase Sensor-3, STS-3, an activity-based fluorescent sensor for SULT1A1, the most widely expressed and promiscuous SULT1 isoform. ## Main Text Human phenol sulfotransferases (SULT1 family) are essential phase II metabolic enzymes that mediate sulfuryl group transfer from the activated sulfate donor 3'-phosphoadenosine-5'phosphosulfate (PAPS) primarily to hydroxy-containing small molecules. Sulfurylation enhances the water solubility to increase the clearance of xenobiotics, to recycle endogenous metabolites (e.g. estrogen, dopamine), and, in some cases, to (in)activate drugs. 6,7 As such, phenol sulfotransferases are not only linked to cellular signaling in normal physiology but also in disease states ranging from cancer to neurodegeneration. 1,3,4, Our understanding of the substrate scope and activities of these enzymes in biological contexts has been significantly advanced by parallel efforts using computational modeling, structural characterization, mechanistic enzymology, and assay methods. 10, Along these lines, one of the most widely used approaches to monitor phenol sulfotransferase activity with purified protein or cell lysates relies on radiolabeling of substrates, PAP 35 S, or sulfate ( 35 SO4 2-), coupled to chromatographic detection of the sulfurylated products. 24,25 Alternatively, coupled-enzyme assays with colorimetric and fluorescent substrates provide a safer, cost-effective, and continuous readout of enzyme activity. If radiolabeling is not preferred or if the sulfurylated product does not have an optical signature, mass spectrometry or nuclear magnetic resonance spectroscopy can be used. 20,24,25,29,30 A limited number of these approaches have been translated to living cells but do not provide a direct readout with a spatially and temporally-resolved map of activity. 31 We envision that activity-based fluorescent sensors can address this gap. This strategy affords the ability to chemically tune and transform an enzyme's substrate into a fluorescent imaging platform for live-cell applications. To our knowledge, activity-based sensing has not been widely exploited for phenol sulfotransferases in living cells. 31 The SULT1 family consists of 9 isoforms, all with a high degree of sequence and structural similarity (≥ 55%). 15 Of these isoforms, SULT1A1 is the most widely expressed in the human body and promiscuous, thus making it an ideal target for this proof-of-concept study. 38,39 To create an activity-based fluorescent sensor, we first selected the substrate 2-naphthol as it has been demonstrated to undergo sulfurylation by SULT1A1. 15,18 Even though 2-naphthol and the resulting 2-naphthyl sulfate product are fluorescent in water, there are negligible differences in the emission spectra of these compounds at physiologically relevant pH. 15,16 As such, our strategy to convert 2-naphthol into an activity-based fluorescent sensor relied on its structural similarity with 3-hydroxy-1,8-naphthalic anhydride (Compound 1, Figure 1). The latter can be readily functionalized with primary amines to generate naphthalimide fluorophores. Specifically, we selected three electronically distinct amines: butylamine (STS-1), 3-aminopropanoic acid (STS-2), and N,N-dimethylethylenediamine (STS-3) (Figure 1). We reasoned that all three sensors would be weakly fluorescent because the non-bonding electrons on the oxygen atom of the phenol could quench the excited state of the naphthalimide fluorophore. However, the quenched state could be relieved upon sulfurylation to generate a turn-on or ratiometric fluorescence response, as previously reported for 3-and 4-substituted-1,8-naphthalimide-based sensors. 46 We do note that STS-1 has been evaluated as a fluorescent sensor for a plant glucosyltransferase in vitro, 46 and STS-3 has been tested as an anti-cancer agent. 47 Preparatory docking calculations and constrained MD simulations were carried out, followed by extensive equilibrium MD simulations to see if the sensors could bind SULT1A1 in a similar fashion to 2-naphthol. 21,48 During the course of the MD simulations, two stabilizing interactions were monitored: one between the phenol group of the substrate and the sulfur atom of the PAPS cofactor (S-O distance) and one between the phenol group of the substrate and the δ-nitrogen atom (N-H distance) of the catalytic histidine residue (His108) in the active site (Figures 2, S9-S11, Movies smov1-smov8). Like 2-naphthol, STS-1, STS-2, and STS-3 were maintained in catalytically productive orientations and had stabilizing interactions with PAPS and His108 (distances < 6 ). 21,48 For each sensor, the functionalized naphthalimide backbone was exposed to the bulk water and did not interfere with the orientation of the phenol group in proximity to the site of catalysis. The importance of the phenol group for optimal substrate positioning and reactivity was further demonstrated with control substrates lacking the hydroxy functional groups, namely naphthalene and STS-3C (Figures S9-S11). The simulations revealed that water was able to interact with the PAPS sulfuryl group, the polar N and O atoms of the catalytic His108, and the control substrates in such a way that both naphthalene and STS-3C were separated from the binding pockets. In parallel, the spectroscopic properties of STS-1, STS-2, and STS-3 were evaluated with purified SULT1A1. Spectral changes were observed when 10 µM of each sensor is incubated with 0.015 µg/µL (0.4 µM) of purified SULT1A1 and 60 µM of the PAPS cofactor at 37 ºC in 50 mM Tris buffer at pH 7.4 (Figure 3). At t = 0 min, the absorption spectrum for STS-1 was featureless, but STS-2 and STS-3 showed two broad maxima at ~340 nm and ~390 nm. Excitation at both absorption maxima resulted in no emission above the background of the buffer for all three sensors, which is consistent with a quenching mechanism. However, clear spectral changes were observed within t = 10 min of incubation with STS-2 and STS-3. Notably, the absorption maxima at ~340 nm increased in intensity whereas the absorption maxima at ~390 nm decreased in intensity, suggesting the formation of new products. Upon excitation at 340 nm, robust turn-on emission responses were observed at ~415-420 nm within t = 1 h (average ± standard deviation): STS-1 (5.7 ± 0.9), STS-2 (6.3 ± 1.7), STS-3 (7.1 ± 0.2) (Figure 3, Table S1). These spectral changes did not occur in the absence of the enzyme or PAPS cofactor (Figure S14-S22). Similarly, coincubation with 10 µM of 2,6-dichloro-4-nitrophenol (DCNP), a substrate inhibitor for SULT1A1 (Ki = 2 µM) 49 attenuated the turn-on fluorescence response of each sensor to varying degrees (Figure S14-S22). 15, Encouraged by these results, we used liquid chromatography mass spectrometry (LC-MS) to confirm that the turn-on fluorescence response for each sensor was attributed to the formation of a sulfurylated product (Figure S23). To better understand the differences in the observed emission responses, we determined the kinetic parameters for each sensor with SULT1A1 using a Michaelis-Menten model (Figure S24). The low solubility of STS-1 above 10 µM prevented accurate measurements, so it was not further evaluated. The average KM values of STS-2 and STS-3 were determined to be 10 µM and 4 µM, respectively. Based on the kcat values (average ± standard deviation), STS-3 (21 ± 5 s -1 ) reacted faster than STS-2 (8 ± 4 s -1 ) with SULT1A1, thus allowing for the quick buildup of fluorescence signal (Table S1). These differences also translated to the overall catalytic efficiencies. Based on these parameters and the robust turn-on fluorescence response of STS-3, we next established the ability of STS-3 to detect SULT1A1 activity in living cells. To validate STS-3, we selected the human SK-N-MC neuroepithelial cancer cell line, which has been previously reported to express SULT1A1. 52,53 First, SK-N-MC cells were treated with either a DMSO vehicle control or 10 μM of STS-3 for 4 h at 37 °C in serum free media, followed by lysis and analysis with fluorescence-based high-performance liquid chromatography (HPLC). Extracts from STS-3 treated cells revealed a single fluorescent product with emission at 420 nm, whereas extracts from DMSO treated cells did not (Figures 4, S25-S30). Co-injection with the authentic sulfurylated product standard, STS-3P, confirmed that STS-3 does indeed undergo sulfurylation in living cells. Encouraged by these results, we turned to live-cell fluorescence microscopy to visualize endogenous SULT1A1 activity. Cells treated with STS-3 and excited at 340 nm showed measurable levels of fluorescence signal above the autofluorescence of cells treated with DMSO. Co-incubation with the substrate inhibitor DCNP attenuated the intracellular fluorescence signal by ~51%. Moreover, this change did not arise from any cytotoxic effects of cotreatment with STS-3 and DCNP (Figures 4, S31-S34). It is not surprising that STS-3 lowers cell viability as related derivatives have been used as anti-cancer agents. 44,45,47 Interestingly, DCNP is known to lower PAPS levels 54 and inhibit the SULT1A1 isoform, but not the SULT1A3 isoform, 15,49,55 which is also expressed in SK-N-MC cells. 12,52,53 As such, this provides strong evidence that STS-3 is a reporter of SULT1A1 activity in this cell line. Indeed, the fluorescence imaging also highlights how the reported activity can vary from cell-to-cell. This could arise from differences in the uptake, distribution, or reactivity of DCNP or STS-3 with SULT1A1 or other sulfotransferases, and even cofactor availability, all of which will be the subject of future investigations. In summary, we have presented the development, computational modeling, in vitro enzymology, and biological application of STS-3, a first-generation activity-based fluorescent sensor for SULT1A1. This proof-of-concept study sets the stage to further develop and apply activity-based fluorescent sensors to discover how phenol sulfotransferase activity can intersect competitive metabolic pathways to modify endogenous metabolites, xenobiotics, or drugs. Along these lines, efforts are currently underway to develop high-throughput screening methods with STS-3 and generate a palette of sensors with improved reaction kinetics and expanded isoform preferences for applications in a range of cell types.

chemsum

{"title": "An activity-based fluorescent sensor for the detection of the phenol sulfotransferase SULT1A1 in living cells", "journal": "ChemRxiv"}

co3o4_nanoparticles_anchored_on_nitrogen-doped_reduced_graphene_oxide_as_a_multifunctional_catalyst_

## Abstract: This study describes a facile and effective route to synthesize hybrid material consisting of Co 3 O 4 nanoparticles anchored on nitrogen-doped reduced graphene oxide (Co 3 O 4 /N-rGO) as a highperformance tri-functional catalyst for oxygen reduction reaction (ORR), oxygen evolution reaction (OER) and H 2 O 2 sensing. Electrocatalytic activity of Co 3 O 4 /N-rGO to hydrogen peroxide reduction was tested by cyclic voltammetry (CV), linear sweep voltammetry (LSV) and chronoamperometry. Under a reduction potential at −0.6 V to H 2 O 2 , this constructing H 2 O 2 sensor exhibits a linear response ranging from 0.2 to 17.5 mM with a detection limit to be 0.1 mM. Although Co 3 O 4 /rGO or nitrogen-doped reduced graphene oxide (N-rGO) alone has little catalytic activity, the Co 3 O 4 /N-rGO exhibits high ORR activity. The Co 3 O 4 /N-rGO hybrid demonstrates satisfied catalytic activity with ORR peak potential to be −0.26 V (vs. Ag/AgCl) and the number of electron transfer number is 3.4, but superior stability to Pt/C in alkaline solutions. The same hybrid is also highly active for OER with the onset potential, current density and Tafel slope to be better than Pt/C. The unusual catalytic activity of Co 3 O 4 /N-rGO for hydrogen peroxide reduction, ORR and OER may be ascribed to synergetic chemical coupling effects between Co 3 O 4 , nitrogen and graphene. With the transition from traditional fossil fuels to clean and sustainable energy, lots of attentions have been paid on storage systems with environmental benignity, high efficiency and alternative energy conversion. Fuel cells have been considered as the most efficient and clean energy conversion device because fuels react with oxygen via mild electrochemical processes without combustion and the overall fuel-conversion efficiency is not limited by the Carnot cycle laws 1,2 . Designing bifunctional catalysts with good oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) activities would be highly beneficial to the development of metal-air batteries. However, developing catalysts for ORR and OER with high activity at low cost remain great challenges 3,4 . Platinum-based materials are known to be the most active electrocatalysts for the ORR and the OER. However, the limited reserves of Pt, high cost, the activity deterioration with time and poor durability severely hinder the large-scale applications of Pt in ORR and OER . On the other hand, the determination of hydrogen peroxide (H 2 O 2 ) has aroused more and more interests of researchers as its significance in the fields of applications in industry as well as biological reactions. Therefore, a rapid, accurate and reliable method to detect H 2 O 2 is of highly demanded. Among various techniques for H 2 O 2 detection, electrochemical H 2 O 2 electrocatalysts are promising due to their high sensitivity, low cost, good selectivity, easy for automation and operational simplicity . Catalysts for hydrogen peroxide reduction, oxygen reduction and oxygen evolution reactions are vital in biological assay and renewable-energy technologies including fuel cells and water splitting. Recently, transition metal oxides including magnanimous oxide, cobalt oxide, iron oxide and nickel oxide as promising materials have received considerableattention due to their low cost, high abundance and perfect catalytic activity for the ORR, OER and immobilizing enzymes for further applications in fabrication of hydrogen peroxide biosensor . Among them, Co 3 O 4 with spinel crystal structure is beneficial to electron transportation between Co 2+ and Co 3+ ions, which has been extensively considered as an efficient electrocatalyst for OER and ORR . Previous studies reported that the efficiency of cobalt oxide as an OER catalyst could be ascribed to the increasing population of Co IV centers at the oxide surface during electrochemical oxidation 19 . More interesting, Co 3 O 4 , which exhibits catalase-like activity for the decomposition of H 2 O 2 , can be applied to the detection of H 2 O 2 in aqueous medium 20 . However, these Co 3 O 4 -based catalysts usually suffer from the poor electrical conductivity, short active site density and the dissolution or agglomeration during electrochemical processes. Co 3 O 4 itself is a material with a little ORR, OER and H 2 O 2 sensing activity and further studies exhibit that synergy between the carbon materials and Co 3 O 4 can give a huge promotion of the electrocatalytic activity 21,22 . Therefore, lots of researches have used conductive carbon nanomaterials such as carbon nanotubes (CNTs), carbon foam and graphene etc. To improve the conductivity of Co 3 O 4 based hybrid catalysts as well as obtain uniformly dispersed Co 3 O 4 nanoparticles and thus to improve the electrocatalyst activity. Graphene, a two-dimensional layer framework of sp2-hybridized carbon with outstanding chemical and physical properties, has attracted a lot of attention in the last years 21,23,24 . Graphene could be an attractive support for metal oxides to form a new class of nanocomposites for ORR due to their notable electronic conductivity and high surface area 25,26 . Dai's group reported a hybrid material consisting of Co 3 O 4 nanocrystals grown on reduced graphene oxide as a high-performance bi-functional catalyst for the ORR and OER 27 . Wang etc. synthesized a novel multifunctional nanohybrid by chemically coupling ultrafine metal oxide nanoparticles to reduced grapheme oxide (rGO) as an effective catalyst for oxygen reduction reaction 28 . Anchoring Co 3 O 4 nanocrystals on carbon-based supports could significantly improve their electrocatalytic activity contributed by the small crystalline size and conductive support 29 . What's more, chemical doping with hetero atoms is an efficacious method to regulate electronic properties and surface chemistry of assembled graphene by the modulation of the carbon-carbon bonds 30,31 . It has been also reported that nitrogen-doped graphene can promote the electrochemical reduction of H 2 O 2 ## 32 . As previous study, the introduction of the Co− N 4 complex onto the graphene basal plane facilitates the activation of O 2 dissociation and the desorption of H 2 O during the ORR 33 . Nitrogen-graphene can produce the synergistic support effect because the reactive intermediates such as hydrogen peroxide are known to decomposed by nitrogen doped carbon nanostructures. However, so far, few researches have reported catalyst which has three functions for H 2 O 2 reduction, ORR and OER. We report herein the synthesis of Co 3 O 4 nanoparticles anchored on nitrogen-doped reduced graphene oxide (Co 3 O 4 /N-rGO) through a simple and scalable method as tri-functional catalysts for H 2 O 2 reduction, ORR and the OER, as shown in Fig. 1. Co 3 O 4 anchored uniformly into laminar nitrogen-doped reduced graphene oxide was confirmed by scanning electron microscopy (SEM). Co 3 O 4 /N-rGO possesses a good electrocatalytic activity toward H 2 O 2 reduction by enhancing the current response and decreasing H 2 O 2 reduction over potential. The electrochemical results demonstrate that the Co 3 O 4 /N-rGO can exhibit higher activity for both the ORR and the OER and better durability than a commercial carbon-supported Pt catalyst. The strong coupling between Co 3 O 4 , nitrogen and reduced graphene oxide (rGO) is found to play an important role in the high electrocatalytic activities of the Co 3 O 4 /N-rGO. This synthesis route can be easily adopted for large-scale manufacturing due to its process simplicity and the accessibility of precursor materials. ## Results and Discussion Characterization. Figure 2A and B illustrated field emission scanning electron microscopy (FE-SEM) images of Co 3 O 4 /N-rGO. We can clearly see from SEM images in Fig. 2B that Co 3 O 4 nanoparticles are uniformly anchored on the rGO substrate with an approximate average diameter of 150 nm. This may be attributed that Co 2+ ion was coordinated with negatively charged oxygen-containing functional groups on N-rGO sheets. During the hydrothermal process, Co 2+ was oxidized into Co 3+ by oxygen-containing groups, and crystallized to form Co 3 O 4 nanoparticles anchored into N-rGO sheets 34 . However, Fig. 2C demonstrate Co 3 O 4 /rGO does not exhibit such a uniform morphology distribution of Co 3 O 4 . In addition, we only found that a typical corrugated structure in Fig. 2D, suggesting there is no Co 3 O 4 particles nucleate on N-rGO surface. The oxygen-containing functional groups of rGO were beneficial for the nucleation and anchoring of nanocrystals on the sheets to achieve covalent attachments, which help to shape the uniform formation of Co 3 O 4 35,36 . In addition, these uniform structures of Co 3 O 4 particles anchored into N-rGO can also be ascribed to NH 3 together with oxygen-containing functional group coordinating with cobalt cations and thus reducing Co 3 O 4 particles size and enhancing particles nucleation on N-rGO 37 . XRD was performed to investigate the phase structure of Co 3 O 4 /N-rGO. As shown in Fig. 3A, the diffraction peaks of the pristine Co 3 O 4 was consistent with the standard Co 3 O 4 (JCPDS card: 42-1467). The major diffraction peaks of Co 3 O 4 /N-rGO were well in agreement with those of Co 3 O 4 except for the broad (002) peak at approximately 25°, which can be ascribed to disordered stacked graphitic sheets 30 . This manifest that the original GO has been reduced to rGO during the hydrothermal process, again confirming we have successfully incorporated Co 3 O 4 into rGO 38 . BET experiments was conducted to obtain specific surface area of as-prepared samples and the isotherms exhibit typical IV isotherms where the recorded BET surface area of 39 . It is noted that there are two remarkable peaks around 1339 and 1591 cm −1 refer to the D-band (arising from the edge or defect sites of carbon) and G band (representing the sp2 carbon) of the graphene domain, respectively 40 . , respectively, with a spin energy separation of 15 eV, which is attributed to the Co 2+ oxidation state, indicating that a portion of Co 3+ is reduced to Co 2+ with generating oxygen vacancies 17 . These results again confirmed that Co 3 O 4 nanoparticles have been anchored on N-rGO, Co 2+ and Co 3+ in the crystal structure of Co 3 O 4 are being considered to be playing a vital role in improving catalytic performance of oxygen reduction reaction and oxygen evolution reaction 37 . Furthermore, the main beautiful structure of Co 3 O 4 is the peculiar cation distribution in the face centered cubic (FCC) crystal where the Co 2+ ions reside on the 1/8 th of the tetrahedral A sites while the Co 3+ ions occupy 1/2 of the octahedral B sites 11 , endow the system viable for electrocatalytic applications. The high-resolution N 1s XPS spectrum of Co 3 O 4 /N-rGO was used primarily to determine the bonding configurations of N atoms in the composite, as seen in Fig. 3E, The peak deconvolution suggests four components were centered at about 398, 400, 401, and 403 eV, corresponding to pyridinic N, pyrrolic N, quaternary N, and oxidized N, respectively. N atom have the lone electron pairs which can hybridize with sp 2 carbon atoms to celebrate oxygen reduction reaction. The performance of ORR depends on the bonding configuration of N atoms in carbon materials. It has been reported that the onset potential of a nitrogen-doped catalyst has strong relation with pyridinic form nitrogen, but little effect by pyrrolic nitrogen and oxidized type nitrogen 41 . due to the advantages of high sensitivity and good selectivity, these sensors often suffer from unstable response due to the intrinsic nature of enzymes 42 . Therefore, it is necessary to develop a simple non-enzymatic strategy for sensing H 2 O 2 with high sensitivity. To date, electrocatalysts for design H 2 O 2 sensors with high sensitivity, good selectivity and easy regulation properties hold leading position among various sensors 12 . It has been proved that functional nano-structured transition-metal oxides exibit good electrocatalytic activity toward the H 2 O 2 reduction, which provides valuable strategy for the nonenzymatic determination of H 2 O 2 43,44 . Among various kinds of transition metal oxides, Co 3 O 4 shows attracting electronic and electrocatalytic properties. Particularly, its normal spinel crystal structure is favorable for electron transportation between Co 2+ and Co 3+ ions and Co 3 O 4 possess catalase-like activity, which is benefit to sensing H 2 O 2 . Therefore, Co 3 O 4 have been extensively explored as the sensing materials for developing enzyme-free H 2 O 2 sensors. However, Co 3 O 4 -based catalysts usually suffer from the poor electrical conductivity, low active site density and the dissolution or agglomeration during electrochemical processes 45,46 . On the other hand, graphene has the ability to promote electron transfer rates and graphene-based modified electrode had much better electrocatalysis toward H 2 O 2 47 . ## Electrocatalytic activity of Co In our study, Co 3 O 4 nanoparticles were incorporated into nitrogen doped graphene, leading to improved conductivity, enhanced catalytic activity and stability of the metal oxide nanocatalyst, and thus a better catalytic effect to H 2 O 2 reduction due to the synergistic effect. To investigate the electrocatalytic characteristics to H 2 O 2 reduction of Co 3 O 4 /N-rGO, voltammetric measurements were performed using the Co 3 O 4 /N-rGO, Co 3 O 4 /rGO and N-rGO modified GC electrodes in the presence of 5 mM H 2 O 2 at a scan rate of 0.05 V s −1 . Figure 4A It is a significant way for amperometric technique to test the sensing property of Co 3 O 4 /N-rGO modified GC electrode. We studied the effect of the applied potential in order to improve the Co 3 O 4 /N-rGO modified GC electrode performance towards non-enzymatic H 2 O 2 sensing. We investigated applied potential on the amperometric response on the Co 3 O 4 /N-rGO modified GC electrode towards sequential addition of 0.5 mM H 2 O 2 by varying the potential between− 0.6 V and − 0.3 V. As shown in Fig. 5A and B, as amperometric response of the Co 3 O 4 /N-rGO modified GC electrode has the optimal sensitivity, the applied potential at − 0.6 V was selected. Figure 5A shows a typical current-time plot of the Co 3 O 4 /N-rGO modified GC electrode on successive addition of H 2 O 2 at an applied potential of − 0.6 V. Catalytic currents showed linear response to H 2 O 2 from 0.5 mM to 17.5 mM (R 2 = 0.994) with a detection limit (S/N = 3) to be 0.1 mM, which are comparable to or even better than those of the other metal-free or enzyme based H 2 O 2 biosensors 9,32,48 . To investigate the selectivity for H 2 O 2 sensing, the amperometric responses of ascorbic acid (AA), dopamine (DA), uric acid (UA) and glucose (Glu) are investigated on the Co 3 O 4 /N-rGO modified GC electrode. As shown in Fig. 5D, when the Co 3 O 4 /N-rGO-modified GC electrode was polarized at − 0.6 V, the addition of 0.2 mM AA, 0.02 mM DA, 0.2 mM UA and 5 mM Glu did not produce an observable current response while the addition of H 2 O 2 induced obvious reduction currents reponse, indicating that the measurements of H 2 O 2 are essentially interference-free from other relevant electroactive species. Therefore, the as-prepared the Co 3 O 4 /N-rGO modified GC electrode is a good ## The performance of oxygen reduction reaction. To evaluate the ORR catalytic activity of Co 3 O 4 /N-rGO, N-rGO and Co 3 O 4 /rGO, CV measurements were performed in both O 2 and N 2 -saturated 0.1 M KOH solution. As shown in Fig. 6A, CV of N-rGO and Co 3 O 4 /rGO in the O 2 -saturated electrolyte shows a reduction peak at − 0.34 V and − 0.33 V respectively, suggesting their electrochemical catalytic activity for ORR. As for Co 3 O 4 /N-rGO composite modified electrode, a reduction peak at ca. − 0.26 V is observed, which is more positive than those of N-rGO and Co 3 O 4 /rGO while it also has a highest current density, suggesting a great improvement of catalytic activity, which is better than tri-functional carbon materials in previous study 32 . Previous have reported that the electrocatalytic activity of Co 3 O 4 was mainly affected by structure 46 . Co 3 O 4 particles have a spinel structure and the direct Co-Co interactions across shared octahedral edges of its spinel framework can enhance the electronic conductivity which is beneficial to the ORR catalytic activity. On the other hand, the N-graphene also exhibited a much better electrocatalytic activity, long-term operation stability for oxygen reduction reaction 22 . Therefore, such an excellent electrocatalytic activity of the Co 3 O 4 /N-rGO toward ORR can be ascribed to the synergetic chemical coupling effects of Co 3 O 4 and N-graphene 18,49,50 . To investigate the oxygen reduction mechanism of Co where I d is the disk current, I r is the ring current, and N is the current collection efficiency of the Pt ring, which was determined to be 0.4 21,46 . From the Fig. 6B, it was calculated H 2 O 2 % value for the Co 3 O 4 /N-rGO during ORR process is about 63.5-32.2% at potentials ranging from − 0.3 to − 0.8 V. The calculated n value for the Co 3 O 4 /N-rGO is about 2.9 to 3.4 from − 0.3 to − 0.8 V. These results reveal that the electrocatalytic process of Co 3 O 4 /N-rGO is an improved four-electron pathway and a two-electron transfer pathway occurred simultaneously for ORR. Methanol poisoning and stability are key issues challenging the cathode materials in current fuel cell techniques 51 . The effect of methanol poisoning and stability on the Co 3 O 4 /N-rGO was investigated in Fig. 7A and B by current-time (i-t) chronoamperometry. As shown in Fig. 7A, when methanol was injected, a significant decrease (90.7%) in current was observed for the Pt/C electrode, whereas only a slight decrease (13.3%) was observed for the Co 3 O 4 /N-rGO, suggesting poor tolerance of Pt/C to methanol compared with the Co 3 O 4 /N-rGO material. Figure 7 Bshows that the amperometric response of ORR on the Co 3 O 4 /N-rGO which exhibits a very slow attenuation of relative current, after 2000 s i.e. a current loss of approximately 31.37%. In contrast, the Pt/C reveals degraded stability with a current loss (39.58%) after 2000 s, indicating the Co 3 O 4 /N-rGO has a better stability than Pt/C. The catalytic property of oxygen evolution reaction. Previous studies have reported Co 3 O 4 particles deposited on stable supporting and conducting substrates can be used as effective electrode materials for both the oxygen reduction (ORR) and evolution (OER) reactions via decreasing overpotential in fuel cells and water electrolyzers 17,52 . The good catalytic performance of OER can be ascribe to the small crystalline size and the mixed valences Co 2+ and Co 3+ of Co 3 O 4 as well with conductive support substrates 17 . Sun and his groups synthesize Co 3 O 4 nanorod-multiwalled carbon nanotube hybrid with a onset potential of about 0.47 V vs. Ag/AgCl and Tafel slope of 65 mV/dec 37 . In our work, a rotational disk electrode (RDE) tests were also carried out in alkaline solution to further evaluate the OER catalytic activity of the Co 3 O 4 /N-rGO. Figure 8A showed the typical linear sweep voltammograms using the RDE at an electrode rotating speed of 1600 rpm and a potential scanning rate of 5 mV s −1 . From the OER region, the Co 3 O 4 /N-rGO afforded a sharp onset potential at 1.54 V, which is worse than that of RuO 2 at 1.49 V and better than Pt/C. The OER over potential at current density of 10 mA cm ## Conclusions In summary, this study describes a facile and effective route to synthesize hybrid material consisting of Co 3 O 4 nanoparticles anchored on nitrogen-doped reduced graphene oxide (Co 3 O 4 /N-rGO) as a high-performance tri-functional catalyst for ORR, OER and H 2 O 2 sensing. Owing to the synergetic chemical coupling effects between Co 3 O 4 and graphene, the Co 3 O 4 /N-rGO exhibited excellent electrocatalytic activity with a direct reduction to H 2 O 2 at − 0.6 V and sensing ability towards H 2 O 2 . Although Co 3 O 4 /rGO or N-rGO alone has little catalytic activity, the Co 3 O 4 /N-rGO exhibits high ORR activity with ORR peak potential to be − 0.26 V (vs. Ag/AgCl) and the number of electron transfer number is 3.4, excellent tolerance to methanol crossover and exceptionally good stability to Pt/C (20%) in alkaline solutions. Catalytic studies of Co 3 O 4 /N-rGO for OER display a better onset potential, overpotential under the current density of 10 mA cm −2 and a smaller Tafel slope with Pt/C (20%). Due to the ease of synthesis and electrode fabrication, the method developed by this study could be used for large-scale synthesis of non-precious metal-based trifunctional metal catalyst for hydrogen peroxide reduction, ORR and OER. ## Experimental Chemicals and materials. Nafion perfuorinated resin solution (5 wt% in a mixture of lower aliphatic alcohols and water) and commercial platinum/carbon (Pt/C) 20 wt% (Pt loading: 20 wt%, Pt on carbon black) were obtained from Sigma-Aldrich. All other chemicals (analytical grade) were purchased from Beijing Chemical Reagent Company (Beijing, China) and used without further purification. Ultra-pure water was obtained with a Milli-Q plus water purification system (Milli-pore Co. Ltd., USA). ## Materials characterization. Scanning electron microscopy (SEM) images were obtained on a Hitachi S-2600N scanning electron microscope. Elemental analysis data were obtained through Flash EA 1112. The X-ray photoelectronspectra (XPS) spectra were obtained using a VG Micro-tech ESCA 2000 using a monochromic 15 Al X-ray source. For rotating ring-disk electrode (RRDE) measurements, a bipotentiostat (CHI 832, Shanghai Chenhua Instrument Co. Ltd.) and a rotating ring-disk electrode with a rotating GC disk electrode and a platinum ring electrode (ALS RRDE-2) were used. The collection efficiency of the ring-disk electrode was evaluated with the Fe(CN) 6 3− / 4− redox couple and was calculated to be 0.4. Electrochemical measurements were performed with a computer-controlled Electrochemical analyzer (CHI600E, Chenhua, China) in a two-compartment electrochemical cell with as-prepared material modified on a glassy carbon electrode (3 mm in diameter) as working electrode, a platinum wire as counter electrode, and a Ag/AgCl (3 M KCl) electrode as reference electrode. All electrochemical experiments were performed at room temperature. (GO) were redispersed in 50 mL anhydrous ethanol to form GO anhydrous ethanol suspension with concentration to be 0.33 mg/mL. The first step to prepare Co 3 O 4 /N-rGO was performed by adding 3.6 ml of 0.2 M Co(Ac) 2 aqueous solution to 72 ml of GO anhydrous ethanol suspension, followed by the addition of 1.8 ml of NH 4 OH (30% solution) and 2.1 ml of water, consequently. The reaction was kept at 80 °C with stirring for 10 h. After that, the reaction mixture from the first step was transferred to a 100 mL autoclave for hydrothermal reaction at 180 °C for 12 h. Co 3 O 4 /GO hybrid was made by the same steps without adding NH 4 OH (30% solution) in the first step 26 . N-rGO hybrid was also made by the same steps just as making Co 3 O 4 /N-rGO preparation without adding Co(Ac) 2 aqueous solution. The fabrication of as-prepared materials modified electrodes. A rotating ring-disk electrode (RRDE) with a rotating glassy carbon (GC) disk electrode (4 mm diameter) and a platinum ring electrode (ALS RRDE-2), and a GC electrode with a diameter of 3 mm working electrode were used as working electrode in this study. Prior to the surface modification, the delectrode were polished with 1.0, 0.3, and 0.05 μ m alumina slurries, and finally rinsed with Milli-Q water under an ultrasonic bath for 1 min. A Co 3 O 4 /N-rGO modified GC electrode was prepared by casting the 4 μ L of 2 mg/mL Co 3 O 4 /N-rGO suspension on the disk electrode surface and drying in air to evaporate the solvent. Similarly, 4 μ Lof 2 mg/mL N-rGO solution and 4 μ L of 2 mg/mL Co 3 O 4 /GO suspension were dropped on GC electrodes, respectively and dried in air to evaporate the solvent for control experiment. Finally, 5 μ L nafion (0.5%) solution (diluted 10 times with deionized water) was covered onto electrode surface and dried to form modified working electrode. All of the electrochemistry experiments were performed at room temperature. The Co 3 O 4 /N-rGO modified GC electrode was pretreated by electrochemical oxidation in a phosphate buffered solution (pH = 6.8) at a potential of 1.7 V (vs. Ag/AgCl) for 300 s at room temperature, followed by potential sweeping from 0.0 V to 1.4 V in 0.5 M H 2 SO 4 until a stable voltammogram was achieved, the purpose of electrochemical oxidation in phosphate buffered solution and H 2 SO 4 is increased more oxygen containing functional group in carbon materials to increase the active site in oxygen reduction reaction. some Co 3 O 4 nanoparticles may dissolv in H 2 SO 4 thus leaves more active sites on grapheme. For linear sweep voltammetry (LSV) from 0.2 to − 1.0 V, The Co 3 O 4 /N-rGO modified GC was scanned at a scan rate of 10 mV•s −1 to measure the surface behavior of the ORR activity of the catalyst in O 2 -saturated 0.1 M KOH. For more quantitative measurements of the ORR activity, LSV was conducted on the catalyst-coated RRDE at a scan rate of 5 mV•s −1 in O 2 -saturated KOH solution at various rotation rates from 400 to 2025 r•min −1 .

chemsum

{"title": "Co3O4 nanoparticles anchored on nitrogen-doped reduced graphene oxide as a multifunctional catalyst for H2O2 reduction, oxygen reduction and evolution reaction", "journal": "Scientific Reports - Nature"}

isolation_and_characterization_of_chitosan_from_ugandan_edible_mushrooms,_nile_perch_scales_and_bana

## Abstract: Of recent, immense attention has been given to chitosan in the biomedical field due to its valuable biochemical and physiological properties. Traditionally, the chief source of chitosan is chitin from crab and shrimp shells. Chitin is also an important component of fish scales, insects and fungal cell walls. Thus, the aim of this study was to isolate and characterize chitosan from locally available material for potential use in the biomedical field. Chitosan ash and nitrogen contents ranged from 1.55 to 3.5% and 6.6 to 7.0% respectively. Molecular weight varied from 291 to 348KDa. FTIR spectra revealed high degree of similarity between locally isolated chitosan and commercial chitosan with DD ranging from 77.8 to 79.1%. XRD patterns exhibited peaks at 2θ values of 19.5° for both mushroom and banana weevil chitosan while Nile perch scales chitosan registered 3 peaks at 2θ angles of 12.3°, 20.1° and 21.3° comparable to the established commercial chitosan XRD pattern. Locally isolated chitosan exhibited antimicrobial activity at a very high concentration. Ash content, moisture content, DD, FTIR spectra and XRD patterns revealed that chitosan isolated from locally available materials has physiochemical properties comparable to conventional chitosan and therefore it can be used in the biomedical field. ## Results Composition by dry weight of chitin and chitosan. The composition of dry weight chitin was determined by using the ratio of the starting dry weight of the raw material (10 g) and the obtained chitin dry weight after demineralization and deproteinization. The dry weight of chitin obtained was 11.8%, 9.9% and 39% for banana weevils (BW), mushrooms (MSR) and Nile perch scales (NS) respectively with P values > 0.05 indicating variability between chitin yield from each raw material. The percentage yield of chitosan from chitin ranged from 70.2 to 82% with P value > 0.05 between BW and MSR chitosan, P values < 0.05 among BW and NS chitosan; MSR and NS chitosan. The chitosan yield from commercial chitin was statistically comparable to that of NS and significantly different from the chitosan yield of BW and MSR, Table 1. Ash content and moisture content. In general, Chitin and chitosan from each raw material registered statistically similar ash contents. However, chitin and chitosan samples from MSR had the lowest ash content statistically similar to the ash content of commercial chitosan but significantly different (P values < 0.05) from the BW and NS chitin and chitosan ash content, Table 2. The moisture content varied from 3.5 to 6.4%, with P values > 0.05 for MSR and NS chitosan signifying analogous moisture content but < 0.05 for BW and commercial chitosan indicating considerable moisture content difference between BW and commercial chitosan and the other chitosan obtained from different locally available materials, Table 2. XRD analysis. XRD patterns exhibited peaks at 2θ values of 19.5° for both chitosan extracted from BW (corresponding to different reticular plans with lattice periodicities of 4.570 and 4.530 ) and MSR (corresponding to 4.407 and 4.136 reticular plans), Fig. 2A,B. Chitosan from NS registered 3 peaks at 2θ values of 12.3°, 20.1° and 21.3° with periodic reticular atomic plans lattice periodicities of 2.350 and 5.581 respectively, Fig. 2C, whereas commercial chitosan (control) scored 2 peaks at 2θ values of 9.6° and 20.2° with intereticular atomic periodicities of 4.414 and 9.201 , Fig. 2D. However, the peak intensities of chitosan isolated from locally available materials were lower than those attained by the control (23,061 a.u and 41,664 a.u), Fig. 2. Furthermore, XRD patterns exhibited by chitin are comparable to the XRD pattern of chitosan but with high intensity pointed peaks, Fig. 3. CrI calculated using XRD pattern ranged from 41% to 51.1% for BW and NS chitosan and Tukey multiple comparison revealed that the CrI for chitosan obtained from locally available materials was substantially different from that of commercial chitosan, Table 2. www.nature.com/scientificreports/ Bactericidal activity of chitosan. Chitosan isolated from locally available materials and commercial chitosan revealed concentration dependent antibacterial activity against carbapenem resistant and sensitive bacteria. No antibacterial activity was registered at all concentrations below 3000 µg/ml. Chitosan exhibited bactericidal activity when the concentration was increased above 3000 µg/ml. The mean inhibition zones increased with increasing chitosan concentration. MSR chitosan exhibited potent antibacterial activity with mean growth inhibition zones of 7 mm (3 mg/ml) and 11 mm (4 mg/ml) similar to those of commercial chitosan. BW chitosan followed with mean growth inhibition zones of 5 mm (3 mg/ml) and 7 mm (4 mg/ml). NS chitosan recorded the least antibacterial activity with mean growth suppression zones of 2 mm (3 mg/ml) and 4 mm (4 mg/ml), Table 4. ## Discussion The percentage of chitin obtained in previous studies ranged from 2.5 to 12.2% for insects 21 , 7.9-11.4% for mushrooms 22 and 33-45% in fish scales 23 . The dry weight of chitin obtained in this study was 11.8%, 9.9% and 39% for banana weevils (BW), mushrooms (MSR) and Nile perch scales (NS) respectively falling within the ranges of the previous studies. Furthermore, the percentage yield of chitosan from chitin in this study ranged from 70.2 to 82%. These results corroborate well with the chitosan yield reported by Erdogan et al. 22 . The residue that remains after complete pyrolysis of the material in the presence of air is termed as ash and is inorganic in nature 24 . Therefore, chitin and chitosan ash contents were determined gravimetrically and the ratio of chitosan weight burnt to the weight of inorganic residue was computed into percentages. Determination of ash content in chitin and chitosan is a vital litmus to assess the effectiveness of the demineralization process. The remnant minerals may include toxic inorganic elements such as Cadmium, Lead and Mercury that could pose health risks if such chitosan is used for biomedical applications as they are extremely hazardous at even very low levels of exposure . Furthermore, solubility of chitosan is greatly affected by the presence of inorganic minerals as this subsequently lowers viscosity 24 . This greatly affects fabrication of chitosan-based drug delivery systems. Furthermore, the level of demineralization and deproteination determines the purity of chitosan which in turn affects its biological properties like immunogenicity, biocompatibility and biodegradability 28 . Chitosan with an ash content lower than 1% possesses superior biological properties and is recommended for biomedical applications 29 . Contrarily, several studies have used chitosan with ash content higher than 1% for biological applications . Thus, chitosan generated by this study with ash contents ranging from 1.5 to 3% is fit for medical use. However, the demineralization step needs to be improved to reduce the ash content further to meet the regulatory requirements if the chitosan isolated from locally available materials is to be used in medical applications. Chitosan has a great capacity to form hydrogen bonds with water through both its hydroxyl and amino groups hence its hygroscopic in nature. The quantity of adsorbed moisture relies on the initial moisture content of the raw materials and storage environmental conditions 33 . The moisture value of commercial chitosan powder ranges from 7 to 11% (w/w) and not influenced by degree of deacetylation or molecular weight 34 . Moisture content is one of the most important factors which influence the usability of chitosan powder during drug carrier and tablet preparations. Moisture content level should be put into consideration when formulating chitosan-based drugs to reduce pharmaceutical powder faults especially after storage as water content above 6% affects powder www.nature.com/scientificreports/ www.nature.com/scientificreports/ flow properties, compressibility and tensile strength of the tablets 28 . The moisture content of the chitosan isolated from locally available material was within the recommended range hence suitable for pharmaceutical use. Chitosan isolated from locally available materials exhibited moderate solubility ranging from 69 to 86%. Contrary to this, Nessa et al. 29 reported excellent solubility ranging from 96.0 to 97.2% of chitosan isolated in-house from prawn shells. Low to moderate solubility values of chitosan are attributed to high protein content and low DD 35 . Nevertheless, this study achieved high DD comparable to levels reported by other studies and commercial chitosan. Therefore, this moderate solubility may be attributed to low demineralization 35 and distribution of the remaining acetyl groups (glucosamine and N-acetylglucosamine units) along the polymer chain which is termed as the pattern of deacetylation (PA) 36 . PA substantially impacts on the charge density which in turn influences the solubility of chitosan regardless of the DD and molecular weight 37 . Indeed, analysis of the residues that remained after solubilizing chitosan in 1% acetic acid revealed that the residues mainly contained inorganic materials with www.nature.com/scientificreports/ ash contents of over 81.1% for BW, MSR and commercial chitosan and the remaining percentages majorly contributed by proteins as depicted by the nitrogen content 38 . Furthermore, this study revealed nitrogen content of 7.0%, 6.8%, 8.2% and 6.8% for BW, MSR, NS and commercial chitins respectively. Statistically similar nitrogen contents were registered by the respective chitosans. The nitrogen content of chitin and chitosan is extremely a vital measure of purity. The nitrogen level of fully acetylated chitin is 6.89% 39,40 . Nitrogen values greater than 6.89% suggest presence of proteins hence low level of deproteination, whereas nitrogen content below 6.89% www.nature.com/scientificreports/ postulates ineffective demineralization step 41 . This explains why the insoluble chitosan residues had considerably higher ash content and to some extent nitrogen bearing compounds. High mineral and residual proteins contents may cause complications in chitosan dissolution and hinder designing and development of chitosan matrix-based drug delivery systems. Total elimination of minerals and proteins during chitosan isolation is impractical as the process requires use of extremely concentrated acids and bases respectively at higher temperatures which yields degraded chitosan. However, increasing the duration of the demineralization and deproteination steps only possibly may farther decrease the ash and protein contents to the recommended levels. Molecular weight is one of the most important physiochemical properties that influences other physicochemical and biological behaviors such as hydrophilicity, viscosity, moisture absorption, biodegradability, antimicrobial activity and mucoadhesion of chitosan 6 . With respect to the raw material and isolation method, the molecular weight of commercial chitosan ranges from 10 KDa to 100,000 kDa. For example, the process of deacetylation may lower the molecular weight of the polymer 28 . Pharmaceutical industries have widely exploited chitosan in several forms of drug delivery systems such as tablets, nanocarriers, hydrogels microspheres, micelles among others. Due to the high viscosity, high molecular weight chitosan based drug carriers discharge the active ingredient gradually and in a controlled manner prolonging the duration of drug activity hence improving treatment outcomes as well as decreasing the drug side effects 42 . In contrast, low molecular weight chitosan possesses high penetrative power than high molecular weight chitosan, thus, can effectively infiltrate bacterial cell walls, bind DNA, block the process of transcription hence inhibit the protein synthesis. Thus, low molecular weight chitosan possesses potent antimicrobial activity 43,44 . On the other hand, high molecular weight chitosan at higher concentrations can exhibit antibacterial activity through binding to the negatively charged bacterial cell wall parts through electrostatic interactions forming an impermeable coating around the cell, hence blocking movement of materials into and out of the cell. Thus, the molecular weight of chitosan should be determined to ensure that it meets the quality for biomedical application. In this study, a relatively high molecular weight chitosan varying between 291 and 348 KDa was obtained. This fairly high molecular weight chitosan isolated from locally available materials falls within the range of chitosan recommended for designing and development of drug delivery system. Fourier Transform Infrared spectrophotometric examination of chitosan isolated from BW, MSR, NS and commercial chitosan (Sigma Aldrich) yielded comparable spectra an indication that locally isolated chitosan has similar physicochemical properties due to the presence of almost similar functional groups. Similar FTIR results were obtained for chitin. However, chitin samples were more hydrated than their respective chitosan as shown by presence of extra 2 bands signifying OH groups between 4000 and 3700 cm −1 . Chitin from various sources is mainly grouped into α and β polymorphs and rarely the γ type. Chitin is made of fibres that are arranged in layers. In α-chitin the adjacent chains are arranged in opposite directions and in an anti-parallel arrangement whereas in β-chitin, the adjacent layers are in the same direction and parallel. In γ-chitin, every 3rd layer is in the opposite direction as compared to the two precedent layers 45 . Beta chitin display bands for CH X deformation at a wavelength of approximately 1455 cm −1 and 1374 cm −1 and several narrow peaks in the C-O-C and C-O stretching region of 1200-950 cm −1 not present in α chitin 46 . Comparable FTIR spectra were observed in this study indicating presence of α chitin in BW, MSR and NS. The percentage chitosan DD varied from77.8 to 79.1%. These DD values are consistent with the shrimp commercial chitosan DD value (76%) used as a control in this study and DD values reported by other studies. Liu et al. 47 and Santos et al. 48 reported DD values of 73.1% and 76% respectively. Furthermore, the DD values obtained in this study are within the range of 75 and 90% deacetylation degree in industrial processing 49 . One of the most important factors that should be considered when isolating chitosan in-house for biomedical application is DD. Degree of deacetylation influences several chitosan traits that include biological, physicochemical and mechanical properties. It was reported that chitosan polymer with low DD disintegrated fast and induced an acute inflammatory response while highly deacetylated chitosan induced negligible inflammation hence biocompatible 50 . Furthermore, swelling property is one of the most vital factors that impact the chitosan performance in the biomedical field. High level of swelling reduces the elasticity and tensile strength of chitosan pharmaceutical materials which increases the risk of collapsing. Chitosan with low DD has a higher swelling index. When water molecules are absorbed and combine with the polar groups in the material molecules, the material swells. Pharmaceutical materials designed from chitosan with higher DD exhibited lower swelling index an indication that highly deacetylated chitosans are suitable for clinical applications 51 . Additionally, chitosan chains with higher DD are more flexible and flexible chains will enhance the formation of hydrogen bonds, boosting the tensile strength and elasticity of chitosan material as a whole 51,52 . Thus, the DD influences the mechanical property of chitosan. Preparation of chitosan therapeutic formulations involves use of solvents and the mechanical strength is compromised by absorption of water by the hydrophilic regions of chitosan. However, this can be overcome by use of chitosan with higher DD. Thus, cautious isolation and purification of chitosan with appropriate DD specifically for fabrication of chitosan-based formulations for parenteral biomedical application should be of great interest. The 2θ values obtained from chitosan isolated from locally available materials were within the same range with the XRD patterns (2θ angles 9.6° and 20.2°) registered by commercial chitosan (control) used in this study. Furthermore, peaks at 2θ = 10° and 2θ = 20° of commercial chitosan (Sigma Aldrich) have been exhibited by X-ray diffraction studies 53 . The results from this study are comparable to the established chitosan XRD pattern as peaks at 2θ values of 12.3°, 20.1° and 21.3° for Nile perch scales chitosan, 19.5° for banana weevil and mushroom chitosan were registered. In banana weevil and mushroom chitosan, the weak peak at 2θ = 10° disappeared. Similar deviations were registered by other studies which attempted to extract chitosan from locally available materials . However, high peak intensity for Nile perch scale chitosan and a slight shift in 2θ = 20° diffractive angle for all extracted chitosan indicates that this study achieved highly crystalline chitosan. The XRD patterns of commercial chitin and that isolated from different locally available are analogous to those of their respective chitosan. However, a slight variation exists in the intensity of peaks. Chitin XRD pattern displayed higher intensity of peaks than chitosan. Similar results were observed in NS chitosan XRD spectrum. High intensity www.nature.com/scientificreports/ peaks in NS chitosan and chitin is mainly as a result of high concentration of impurities such as minerals and proteins. Furthermore, XRD analysis showed only peaks at 2θ values associated with hydrated polymorphs for all the chitosans isolated from locally available materials and the commercial chitosan. This is in agreement with the XRD spectrum of the hydrated chitosan obtained by other studies 57,58 . The CrI values were obtained by Focher et al. 59 methods using formula (9). However, the CrI estimated by this method might be too low as the procedure has been implicated to underestimate CrI, because of overestimation of the input of the amorphous phase 57,60 . Indeed, this study achieved low CrI values but higher than the CrI values reported by De Queiroz Antonino et al. 60 who used Osorio-Madrazo et al. 57 improved method. A substantial variation was observed in the yield, ash content, moisture content, solubility and crystallinity of chitosan isolated from locally available materials. The variation in physicochemical properties among chitosan obtained from BW, MSR and NS is in line with other studies. Szymańska and Winnicka 28 observed that a variety of chitosan raw materials lead to considerable dissimilarities in the quality and properties of chitosan and its products. Thus, significant deviations from the pharmacopeial recommendation might be registered by chitosan obtained from different sources. Recently, substantial research attempts have been made to investigate the antimicrobial activity of chitosan. It has been reported that chitosan possesses potent antibacterial and antifungal activity . Vilar Junior et al. 61 reported that chitosan exhibited minimum inhibitory concentration ranging from 78 to 625 µg/ml in in vitro studies. However, the antibacterial activity of chitosan isolated from locally available materials and commercial chitosan in this study was attained at very high concentrations of 3000 µg/ml and 4000 µg/ml. Kamjumphol et al. 66 reported similar results where chitosan antibacterial activity was dose dependent and the most efficacious concentration was 5000 µg/ml. The low antibacterial activity may be attributed to fairly high molecular weight of chitosan. In general, the antimicrobial activity of chitosan against E. coli increases with increasing molecular weight but up to a certain level. Tanigawa et al. 67 observed that Chitosan of 80 KDa exhibited superior antimicrobial activity against E. coli as compared to 166 KDa, 190 KDa and very low molecular weight chitosan of 2 to 12 KDa. Similar results were reported by several studies . Indeed, this study isolated fairly high molecular weight chitosan. Furthermore, low antibacterial activity of chitosan in this study may also be associated to the average DD as antibacterial activity of chitosan increases with increase in the DD. Chitosan with high DD of over 90% possesses a higher positive charge density that facilitates electrostatic interaction with the negatively charged bacterial cell thereby conferring more potent bactericidal activity than chitosan with moderate DD 71,72 . Additionally, the dose dependent antibacterial activity achieved in this study may be due to increase in the net positive charge as the concentration of chitosan increases. Moreover, the low antibacterial activity in this study may due to high ash content above the recommended value (1%) which affects the physicochemical properties of chitosan such as solubility which in turn negatively affects its bioavailability. Indeed, low solubility affects the bioavailability of chitosan. NS chitosan with the lowest solubility exhibited the least antibacterial activity. ## Conclusion The purity level of chitosan and its physicochemical properties affect its biological parameters such as biodegradability, biocompatibility and antimicrobial activity. These physicochemical characteristics are influenced by raw materials and the method used in chitosan isolation. Most literature documented shrimp shells and other crustaceans as the main raw materials for high grade chitosan. Basing on this background, this study isolated chitosan from banana weevils, mushrooms and Nile perch scales. Chitosan isolated from the locally available materials exhibited moderately high DD and other physicochemical properties corroborating with commercial chitosan (Sigma Aldrich) but with moderate solubility and antibacterial activity. Therefore, attempts should be made to improve the chitosan isolation methods so that the DD and solubility are further increased while the inorganic and protein contaminants are completely eliminated. This should result into optimal chitosan isolation suitable for pharmaceutical and biomedical applications. Furthermore, the cell membrane of bacterial cells is negatively charged. Thus, the zeta potential of chitosan intended for antibacterial application should be determined as only positively charged materials with a pH lower than 6.5 interact with the negatively charged components of the bacterial cell wall. ## Methods and materials Source of materials. Nile perch scale wastes were obtained from a local fish market while banana weevils and edible mushroom were collected from National Agricultural Research Laboratories, Kawanda. Shrimp shell chitosan (CAS number: 9012-76-4; 448877-50G) and chitin (CAS number: 1398-61-4; C7170-100G) were purchased from Sigma Aldrich. Carbapenem resistant E. coli and K. pneumoniae were a kind donation from Department of Microbiology, College of Health Sciences, Makerere University. ## Study design and site. This was a laboratory-based study conducted from College of Veterinary Medicine Animal Resources and Biosecurity, Makerere University, iThemba LABs, Cape Town and University of South Africa (UNISA). Isolation of chitin and chitosan was performed from the Pharmacology Laboratory, Makerere University, while chitosan antibacterial activity was evaluated from the Central Diagnostic Laboratory, Makerere University. Characterization of chitin and chitosan was conducted from iThemba LABs and UNISA. Commercial shrimp shell chitin and chitosan were used as controls in all characterization experiments. Chitin extraction. Banana weevils, Nile perch scales and mushrooms were cleaned using running tap water and finally rinsed in distilled water. The cleaned weevils, scales and mushrooms were oven dried at 60 °C for 1 week and then ground to powder using an electric miller. Chitin was extracted from the resultant powder following Mohammed et al. 73 www.nature.com/scientificreports/ weevil, Mushroom and Nile perch scale powders with 1.0 M HCl solution at 50 °C in a water bath for 24 h with a solution to solid ratio of 15 mL/g. This step was replicated ten times. The mixture was centrifuged at a speed of 4000 × g for 10 min using Thermo Scientific™ Fiberlite™F6-10 × 1000 LEX roto centrifuge. The resultant sediment was washed with distilled deionized water until neutral pH was achieved. The sediment was deproteinized by adding 1.0 M sodium hydroxide at a ratio of 15 mL:1 g and then heated at 80 °C for 8 h in a water bath. This treatment was repeated four times. The resultant chitin was then washed with distilled deionized water to neutrality. Finally, chitin was washed by boiling in hot absolute ethanol and later in absolute acetone in a water bath for 10 min to remove any impurities. The purified chitin was dried in a vacuum oven at 50 °C to constant weight. The chitin content was determined by computing the weight differences between the raw materials and that of the chitin obtained after acid and alkaline treatments. Chitosan preparation. Chitin was treated with 50% NaOH (15 mL/g) at 90 °C in a water bath for 10 h with continuous mixing using a magnetic stirrer after which the resultant mixture was centrifuged at 4000 × g for 10 min using a Thermo Scientific™ Fiberlite™F6-10 × 1000 LEX roto centrifuge. The residue was washed with hot distilled deionized water until neutrality. The obtained chitosan was dried in a vacuum oven at 40 °C for 48 h. All the chitosan samples were purified by dissolving in 1% acetic acid and reprecipitated in 20% NaOH solution followed by centrifugation at 6000 × g for 10 min using a Thermo Scientific™ Fiberlite™F6-10 × 1000 LEX roto centrifuge to sediment chitosan. The sedimented chitosan was washed with distilled deionized water until a neutral pH, lyophilized and stored at − 20 °C until further use. The percentage chitosan yield was computed as a fraction of weight of dry chitosan and dry chitin from which it was generated. Characterization of chitin and chitosan. Estimation of the ash content of chitin and chitosan. The ash content of each chitin and chitosan sample was gravimetrically estimated after the pyrolysis of 1 g in a muffle furnace at 650 °C for 5 h. This procedure was done in triplicates and the mean ash content computed. The ash content was computed as a fraction of mass of the residue (MR) and mass of the sample (MS) using the formula (1) that follows; where MS and MR are the weights (in grams) of the initial sample of the sample and residue respectively 74 . Nitrogen content of chitin and chitosan. Amino acids are building blocks of proteins and contain nitrogen. Thus, nitrogen content is representative of protein content as the percentage of nitrogen present in a sample is directly proportional to the percentage of proteins. It is estimated that 1 g of a given protein sample contains 0.16 g of nitrogen. However, this value varies greatly depending on the protein source 75,76 . Therefore, nitrogen content can be used to infer the amount of protein in a sample. Nitrogen content was estimated by the Kjeldahl method 77,78 . Briefly, 1 g of each chitin and chitosan samples was hydrolyzed at 420 °C for 2 h in 15 ml of concentrated sulphuric acid (98% W/W) holding two copper catalyst tablets using a DT 220 digestor™, heat block (FOSS analytical, Denmark). After cooling, distilled deionized water (60 ml) was added to the hydrolysate followed by 50 ml of 60% NaOH to liberate ammonia, then distillation to recover the ammonia in 4% boric acid receiver. To quantify the amount of ammonia trapped, the receiving solution was titrated with 0.1 M HCl and the amount of nitrogen calculated using the formula (2) below; Moisture content of chitin and chitosan. The water content of chitin and chitosan samples was assessed by gravimetric technique. This method involved drying of the samples until a constant mass in a vacuum oven at 105 °C for 24 h. This experiment was done thrice and the average moisture content was calculated. The water content was computed as the difference between the wet weight (WW) and dry weight (DW) of samples per gram using the formula (3) that follows: where WW is the wet weight of samples and DW is the dry weight of samples after oven drying 16 . Determination of chitosan solubility. A 1% solution of chitosan was constituted by adding 0.1 g (W1) of each chitosan sample previously dried at 105 °C for 24 h into 10 ml of 1% acetic acid in 15 ml falcon tube. The tubes were sealed and placed in an overhead shaker running at 60 rpm for 48 h. The solution was centrifuged at10,000 × g for 15 min using a Thermo Scientific™ Fiberlite™F6-10 × 1000 LEX roto centrifuge. The liquid phase was poured off and the sedimented residue was washed with 10 ml of distilled deionized water and centrifuged at 10,000 rpm for 15 min. The supernatant was decanted and the residue dried at 105 °C for 24 h (W2). This experiment was done three times and mean dry residue calculated. The dry residue was weighed and the percentage of solubility was determined using the formula (4) that follows; (1) Percentage ash content = MR MS × 100; (2) 1.0 ml 0.010 N HCl = 10 µMol N www.nature.com/scientificreports/ where; W1 was the initial weight of dry chitosan and W2 was the weight of the dried residue. Furthermore, the ash and nitrogen contents of the residues were determined. Chitosan molecular weight estimation. The molecular weight of chitosan was determined using the intrinsic viscosity (ƞ) following Costa et al. 79 adjusted method. A solvent medium was constituted by mixing 0.25 M acetic acid and 0.25 M sodium acetate at 1:1 ratio. Five hundred milligram (500 mg) of each chitosan sample was dissolved in 100 ml of the solvent medium to attain a chitosan concentration of 0.005 g/ml. The mixtures were left to stand for 24 h under constant stirring for complete solubilization of chitosan. Intrinsic viscosity was estimated using an automated Ubbelohde-type glass capillary with capillary tube diameter of 0.63 mm at 25 ± 01 °C. Determination of the intrinsic viscosity was achieved by recording the time of the solvents flow which included the flow of the solvent and the four chitosan solutions. This step was repeated three times to obtain the average flow rate for the solvent and the chitosan solutions. The rate of solvent flow was used to calculate intrinsic viscosities values by means of a single concentration value using Solomon and Ciuta 80 Eq. ( 5). Obtained intrinsic viscosities were employed to estimate the molar mass or molecular weight of different chitosan samples using Mark-Houwink formula (6) 81 . where [ƞ] SC is the intrinsic viscosity from Solomon and Ciuta equation, ƞ r is the relative viscosity; (ƞ r = t/t 0 where t 0 is the efflux time of the solvent and t is the efflux time of chitosan solution of a given concentration), ƞ sp is specific viscosity (ƞ sp = ƞ r -1), ln is natural log and C is solvent concentration. where M v is the viscosity average molecular weight of polymer, α and k are constants (α = 0.83 and k = 1.4 × 10 −4 for 0.25 M acetic acid and 0.25 M sodium acetate solvent system 82 and [η] is the intrinsic viscosity. Fourier transform infra-red spectroscopy (FTIR). Three milligrams (3 mg) of each sample (chitin and chitosan) and 5 g of Potassium bromide (KBr) were dried at 60 °C and 120 °C respectively under reduced pressure for 12 h. Each dried chitin and chitosan sample was homogenized with 100 mg of KBr and then compressed to form very thin discs of approximately 0.2 mm thickness. The chitin and chitosan samples were examined at 4000-400 cm −1 Wavenumber range using a PerkinElmer FT-IR Spectrometer. The spectrometer was set to perform at least 64 scans per sample. A KBr disc was used as reference. Functional group assigning to the generated FTIR spectra bands was done using documented literature 46, . Determination of the degree of deacetylation (DD%). The acetylation and deacetylation percentage of chitosan samples was determined by Fourier Transform Infrared Spectroscopy (FTIR). This was done through the correlation of some absorbance bands linked to some of amide, methyl and hydroxyl bands registered by the FTIR spectra. Vilar Junior et al. 61 used the Amide-I band with a wavenumber of 1655 cm −1 and the hydroxyl group band at 3450 cm −1 using the formulae (7 and 8) that follow to determine the degree of acetylation (DA) and then the DD 90 ; where A1655 was the absorbance at 1655 cm −1 of the Amide-I band which is measure of the N-acetyl group content, A3450 was the absorbance at 3450 cm −1 corresponding to the hydroxyl band as an internal standard to correct the disc thickness, factor 1.33 is the ratio of A1655 and A3450 for fully N-acetylated chitosan. X-ray diffraction analysis (XRD). X-ray diffraction was used to determine the crystallinity of the isolated chitin and chitosan where 500 mg of each sample chitosan powder were analyzed employing BRUKER AXS diffractometer, D8 Advance (Germany) fitted with Cu-Kα radiation (λKα1 = 1.5406 ) from 2θ = 0.5° to 130°, with increment ∆2ϑ: (0.034°), voltage of 40 kV, current of 40 mA, power of 1.6 kW and counting time of 0.5 s/step. Generated data was analyzed by OriginPro Version 8.5 and resultant peaks 2θ values were compared with the commercial shrimp chitosan from Sigma Aldrich. The crystalline Index (CrI) values were determined from the XRD pattern following Focher et al. 59 methods using formula 9. where I 200 is the maximum peak intensity for each chitosan at 2Ɵ-20° and I am is the intensity of amorphous diffraction at 2Ɵ-16°. Antimicrobial susceptibility assay of chitosan. Antibacterial activity of the chitosan was evaluated using standardized inocula of 1 × 10 7 CFU/mL with 0.5 McFarland standards streaked onto the surface of sterile agar plates. Carbapenem resistant E. coli and K. pneumoniae suspended in Brain Heart Infusion Broth were inoculated onto the Mueller Hinton Agar plates and round wells of diameter 6 mm, depth 3 mm were prepared using a sterile

chemsum

{"title": "Isolation and characterization of chitosan from Ugandan edible mushrooms, Nile perch scales and banana weevils for biomedical applications", "journal": "Scientific Reports - Nature"}

diverse_protein_manipulations_with_genetically_encoded_glutamic_acid_benzyl_ester

## Abstract: Site-specific modification of proteins has significantly advanced the use of proteins in biological research and therapeutics development. Among various strategies aimed at this end, genetic code expansion (GCE) allows structurally and functionally distinct non-canonical amino acids (ncAAs) to be incorporated into specific sites of a protein. Herein, we genetically encode an esterified glutamic acid analogue (BnE) into proteins, and demonstrate that BnE can be applied in different types of site-specific protein modifications, including N-terminal pyroglutamation, caging Glu in the active site of a toxic protein, and endowing proteins with metal chelator hydroxamic acid and versatile reactive handle acyl hydrazide.Importantly, novel epigenetic mark Gln methylation is generated on histones via the derived acyl hydrazide handle. This work provides useful and unique tools to modify proteins at specific Glu or Gln residues, and complements the toolbox of GCE. ## Introduction There is an increasing demand for precisely decorated proteins in academic research and the pharmaceutical industry, so methods that enable the site-specifc introduction of various functional entities into proteins are under extensive research and development. 1 Typical strategies include the fusion of protein/peptide tags with catalytic activities (e.g. SNAP tags, HaloTags and SpyTags) 2 or chemical reactivities (e.g. tetracysteine, aldehyde, CBT tags and p-clamps), 3 enzyme-catalyzed protein modifcations (e.g. sortase, lipoic acid ligase, transglutaminase and butelase 1), 4 the use of elaborately designed chemical reagents to react with rare residues (e.g. Cys, Met and N-terminal residues), 5 ligand-directed protein labelling, 6 and protein synthesis via expressed protein ligation, 7 native chemical ligation, 8 or most recently KAHA ligation, 9 etc. These technologies have signifcantly advanced our capabilities in functionalizing proteins with photophysical probes 2b,4a,4c and bioorthogonal handles, 10 preparing proteins with post-translational modifcations, 11 as well as producing high-quality biologics. 12 Genetic code expansion (GCE) is a biotechnology whereby a non-canonical amino acid (ncAA) could be inserted into proteins through suppressing an artifcially introduced stop codon or quadruplet codon with an orthogonal aminoacyl-tRNA synthetase (aaRS)/tRNA pair. 13 After two decades of research and development, 14 GCE is becoming one of the most powerful strategies in protein engineering, and provides unique tools to many research felds, such as bioorthogonal chemistry, 15 arti-fcial enzyme design, 16 biocontainment development, 17 vaccine development, 18 protein/peptide-based drug discovery, 19 and so on. We are interested in further expanding the GCE toolbox to address important and yet unmet needs in protein engineering. When carefully examining the repertoire of genetically encoded ncAAs, 20 it is not difficult to fnd out that ncAA designed as Glu or Gln derivative is very rare, thus rendering the manipulation of proteins at specifc Glu or Gln residues challenging or even unlikely. Glu/Gln residues are often essential for proper protein folding, maturation, and activity, and there are increasing reports of post-translational modifcations occurring on them. 21 Additionally, ester, as one of the most basic functional groups, remains largely unexplored in the history of GCE, likely due to the concern of ncAA integrity in intracellular environments. With these considerations in mind, herein we would like to genetically encode a glutamic acid analogue with an esterifed side-chain to enable certain protein manipulations at Glu and Gln residues. The ester side-chain is more amenable to nucleophilic attack as compared to the natural counterpart amide and carboxylate side-chains, and therefore holds great potential to realize several types of largely unexplored protein manipulations, including (1) protein N-terminal pyroglutamation by placing ncAA at the N-terminus; (2) a chemical cage of Glu; (3) site-specifcally conferring hydroxamic acid (HA) on proteins; (4) generating an acyl hydrazide handle for diverse protein modifcations (Scheme 1). ## Results and discussion Genetic incorporation of L-glutamic acid g-benzyl ester (BnE) Firstly, we chose the commercially available L-glutamic acid gbenzyl ester as the target of investigation (BnE, Fig. 1A), as the unnatural benzyl ester is less likely to experience enzymatic hydrolysis in living cells, and empirically the moderate-sized hydrophobic side-chain is favourable for genetic incorporation. To encode BnE, a pyrrolysyl-tRNA synthetase/tRNA pair (PylRS/PylT) was employed based on its adaptability to recognize a large number of ncAAs and well-demonstrated orthogonality in different organisms. In detail, a Methanosarcina barkeri PylRS library with residues A267, Y271, N311, C313 and Y349 randomized to NNK (N ¼ A, T, G or C; K ¼ T or G), 22 was subject to multiple rounds of positive and negative selection in E. coli DH10B. Gratifyingly, colonies with BnE dependent growth were obtained after the last round of positive selection, and DNA sequencing revealed three unique variants (BnERS1-3, Table S1 †). A GFP-based assay revealed that all variants could enhance fluorescence intensity in the presence of BnE (Fig. S1 †). BnERS1 harboring N311S, C313A and Y349F showed the highest amber suppression efficiency, and was used in the following studies (thereafter denoted as BnERS). To confrm the selective incorporation of BnE, a C-terminal His-tagged super-folder GFP variant with an amber mutation at a permissive site (sfGFP-Y151TAG) was expressed in E. coli DH10B in LB media supplemented with 5 mM BnE, and purifed through a Ni-NTA column in 56 mg L 1 isolated yield. This efficiency is comparable to N-Boc-L-lysine in our study (Fig. S2 †). The purifed sfGFP-Y151BnE was analyzed by SDS-PAGE gel (Fig. 1B) and electrospray ionization quadrupole time of flight (ESI-QTOF) mass spectrometry (Fig. 1C). The observed mass peak (27 653 Da) is consistent with the expected mass, thus confrming BnE incorporation. Of note, BnE at Y151 is located on the external surface of sfGFP, and no ester bond hydrolysis was observed, suggesting sufficient stability in an intracellular environment and during the protein purifcation processes. We also attempted to encode BnE in mammalian cells. An EGFP variant (EGFP-Y39TAG) along with BnERS/PylT was transiently expressed in HEK293T cells, and obvious green fluorescence was observable only in the presence of BnE, indicating the BnE-dependent full-length expression of EGFP. The result was further validated by western blot analysis (Fig. S3 †). ## BnE enabling the recombinant expression of proteins with Nterminal pyroglutamation N-terminal pyroglutamate (pGlu) frequently occurs in proteins and peptides, such as ribonuclease, 23 amyloid-b 24 and antibodies, 25 and as reported pGlu formation can reduce the susceptibility of proteins to aminopeptidase digestion, and affect protein's structural integrity. 26 pGlu is derived from Nterminal glutaminyl and glutamyl precursors, usually catalyzed by glutaminyl cyclase, and in certain cases through a very sluggish intramolecular cyclization. 25b We reasoned that BnE at the N-terminus of a protein should accelerate the spontaneous intramolecular cyclization to form pGlu. The initial sfGFP construct (sfGFP-S2BnE) revealed failed removal of the starting Met in E. coli DH10B (Fig. S4 †), and then a BnE-exposing strategy was designed. Specifcally, the protein of interest (POI) with N-terminal BnE was fused to the Cterminus of a small ubiquitin-like modifer protein (SUMO), which can be removed by SUMO-specifc protease (Ulp1), thus exposing BnE (Fig. 2A). Experimentally, SUMO-sfGFP fusion containing BnE right after the SUMO tag was expressed in E. coli Scheme 1 Genetic incorporation of side-chain esterified glutamic acid analogue enables multiple types of protein engineering. DH10B, and purifed through the C-terminal His-Tag in 10 mg L 1 isolated yield. This protein was treated with recombinantly expressed Ulp1 in pH 8 buffer at 30 C for 2.5 h, and analyzed by SDS-PAGE and ESI-QTOF mass spectrometry (Fig. 2B-D). The results show that the SUMO tag was completely removed, and mass peaks corresponding to BnE (27 729 Da) and pGlu (27 621 Da) respectively at the N-terminus of sfGFP were observed (Fig. 2D). In contrast, SUMO-sfGFP-S2E was expressed and treated with Ulp1, and no pGlu formation was observed in mass spectrometry (Fig. S5 †). The result indicates that BnE at the Nterminus of a protein undergoes an obviously accelerated cyclization reaction. To complete the pGlu formation, Ulp1 was co-expressed with the SUMO-BnE-POI fusion, as in situ generated N-terminal BnE would allow instant pGlu formation. Specifcally, Ulp1 was coexpressed along with BnERS/PylT and SUMO-sfGFP-S2BnE in a standard expression process, and the affinity purifcation by the C-terminal His-Tag afforded 5 mg L 1 of the recombinant protein. SDS-PAGE revealed a complete removal of the SUMO tag, and mass spectrometry indicated complete pGlu formation (Fig. 2E and F). As a control, co-expression of SUMO-sfGFP-S2Q with Ulp1 hardly produced pGlu-sfGFP based on mass spectrometry analysis of the purifed protein (Fig. S6 †). To demonstrate the generality of this method, thioredoxin and 14-3-3g with an N-terminal pGlu were facilely prepared and confrmed by SDS-PAGE and mass spectrometry (Fig. S7 †). The heavy chain of Herceptin Fab has an N-terminal Glu residue, and by the method here, a pGlu was quantitatively generated at this position (Fig. S8 †). Recently, Ball et al. demonstrated that pGlu-His dipeptide at the N-terminus of a protein could serve as an efficient reactive handle for Chan-Lam coupling. 27 By the method developed in this work, the pGlu-His tag was installed at the N-terminus of sfGFP, and successfully reacted with an arylboronic acid reagent; without the adjacent His, no labelling product was observed (Fig. S9 †). In short, this work provides an alternative method to equipping proteins with N-terminal pGlu, 28 possibly promoting the relevant biological research and applications. ## BnE can serve as a chemical cage of Glu The ester bond is susceptible to alkaline hydrolysis (Fig. 3A). Indeed, after incubation in a pH 11 Tris buffer solution for 7 h, sfGFP-Y151BnE was completely converted to sfGFP-Y151E based on SDS-PAGE and ESI-QTOF analysis (Fig. 3B and C). Thus, BnE could be regarded as a chemical cage of Glu. Barnase is a secreted ribonuclease from Bacillus amyloliquefaciens, and its recombinant expression in a bacterial host requires the presence of cytoplasmic inhibitor barstar, with which a complex can be formed (K D ¼ $10 14 M). 29 To obtain an active barnase protein, additional steps including denaturation and refolding are thus inevitable to get rid of barstar, and to some extent impede the application of barnase. 30 There is an essential Glu (E73, Fig. 3D) at the active site of barnase, and we wondered if E73 can be caged with BnE to temporarily abate its toxicity to the expression host. To prove this concept, barnase-E73BnE under the control of a T5 promoter was expressed in E. coli DH10B in the presence of 5 mM BnE at 37 C. The expression after induction hardly affected E. coli growth (Fig. S10 †), and the affinity purifcation through the C-terminal His-Tag afforded barnase-E73BnE at a yield of 2.5 mg L 1 . Then the purifed protein was incubated in pH 11 Tris buffer at 37 C for 7 h, and analyzed by SDS-PAGE (Fig. 3E) and ESI-QTOF. Mass spectra indicated a complete conversion of barnase-E73BnE to wild type barnase (Fig. 3F). Furthermore, a yeast RNA-based assay confrmed the enzymatic activity of generated barnase (Fig. 3G), and the calculated enzymatic activity (4 10 6 units per mg) is consistent with the reported value of wild type barnase. 30 Herein, the benzyl cage can be quantitatively removed under relatively mild conditions, indicating potential for future applications. ## Site-specically generating acyl hydrazide and hydroxamic acid on proteins with BnE Esters can react with hydrazine and hydroxylamine to form acyl hydrazide and HA respectively. HA has a high chelating power to many metal ions, and serves as reactive warheads in a number of bioactive compounds, such as natural iron chelator siderophores and inhibitors of Zn 2+ dependent enzymes (e.g. matrix metalloproteinases and HDACs). 31 Also, acyl hydrazide has widespread applications in bioorthogonal conjugation reactions 32 and chemical protein synthesis. 33 Considering the attractive features, we attempted to equip proteins with acyl hydrazide and HA through derivatizing the encoded BnE (Fig. 4A). To prove this method, the purifed sfGFP-Y151BnE was incubated with an excessive amount of hydrazine or hydroxylamine (4%, v/v) in pH 7.4 PBS solutions at 37 C for 1 h, and then analyzed by SDS-PAGE (Fig. 4B) and ESI-QTOF. The observed mass peaks for both reactions are in good agreement with the expected values (Fig. 4C). As a control, no reaction occurred when wild type sfGFP was treated with hydrazine or hydroxylamine. This confrmed that BnE on a protein could be efficiently converted into acyl hydrazide or HA. ## Constructing HA-based articial metal-binding centers on proteins Metal-chelating ability is of great interest in protein engineering, and herein the site-specifcally generated HA sidechains on a protein would be a unique tool in metalloprotein design. Natural siderophores usually employ three HA moieties to realize the extremely high binding affinities for Fe 3+ , thus making us wonder if multiple HA moieties could be constructed on a single protein. Experimentally, three closely spaced residues on sfGFP including K45, D210 and N212, were chosen to construct a tridentate metal-binding center (Fig. 5A). As a control, single and double HA installations were also included (sfGFP-N212, and sfGFP-D210-N212). To enhance amber suppression, sfGFP variants with two and three BnE were expressed in a RF1 knockout strain (C321.deltaA), and purifed at a yield of 47 and 20 mg L 1 respectively. sfGFP-N212BnE was purifed in E. coli BL21(DE3) at a yield of 53 mg L 1 . All sfGFP variants were treated with hydroxylamine, and then subject to SDS-PAGE and ESI-QTOF analysis (Fig. 5B, C and S11 †). Mass spectrometry revealed nearly quantitative conversion of BnE to HA in all cases. Metal cations can affect the fluorescence intensity of GFP, 34 so the fluorescence intensity of these sfGFP variants was investigated in the presence of Cu 2+ and Fe 3+ respectively. Relative to the wild type, single or double HA installation moderately increased the sensitivity of sfGFP to Cu 2+ , while triple HA installation displayed much enhanced sensitivity, implying a synergistic effect of triple HA moieties (Fig. 5D). A similar phenomenon was also observed for Fe 3+ (Fig. S12 †). This experiment indicated the feasibility of constructing HA-based artifcial metal binding centers on proteins by encoding BnE. ## BnE-derived acyl hydrazide on proteins is a versatile reactive handle Lastly, we would like to explore the potential of BnE-derived acyl hydrazide in protein engineering. Liu et al. developed a chemoselective reaction between N-terminal Cys and C-terminal hydrazide of two different peptides for chemical protein synthesis, and demonstrated its application in many cases. 8b,33a This inspired us to know whether the BnE-derived acyl hydrazide can be used in site-specifc protein labelling (Fig. 6A). Specifcally, sfGFP harboring an acyl hydrazide at Y151 was treated with 1 mM NaNO 2 in 20 mM pH 3 phosphate buffer for 20 min, and then incubated with 2 mM Fl-cys at pH 7 for 4 h. The labelling product was subject to SDS-PAGE analysis, and a bright band was observed by fluorescence (Fig. 6B). Furthermore, ESI-QTOF analysis confrmed the formation of the desired product; the observed mass (27 996 Da) is in good agreement with the expected mass (27 995 Da), and only a small amount of hydrolyzed product was observed (Fig. 6C). This experiment verifed the usefulness of BnE-derived acyl hydrazide in site-specifc protein labelling. A recent study revealed that Gln methylation on histones is a novel epigenetic mark and RNA-polymerase-I-dedicated modifcation. 21b Ambitiously, we attempted to use BnE-derived acyl hydrazide to prepare histone variants with site-specifc Gln methylation. To start with, a H3 variant with BnE at Q56 (H3-Q56BnE) was expressed in E. coli, and verifed by mass spectrometry (Fig. 7B). Initial efforts in adjusting Liu's method by using methylamine instead of a Cys derivative did not afford the desired product, possibly due to the short half-life of acyl azide in aqueous solutions. Recently, Dawson et al. reported the use of Knorr pyrazole as the thioester surrogate in native chemical ligation. 33b Derived from acyl hydrazide, the acyl pyrazole exhibits weaker electrophilicity, and hopefully would react with methylamine in aqueous solutions (Fig. 7A). BnE at H3-Q56 was converted into the corresponding acyl hydrazide, then reacted with acetyl acetone (acac) in a pH 3 guanidine solution at 37 C for 1.5 h. ESI-QTOF analysis revealed an efficient formation of the acyl pyrazole intermediate (Fig. 7B). Methylamine was added into the acyl pyrazole solution with subsequently pH adjusted to 9, and then the mixture was incubated at 37 C for 5 h. The fnal product was analyzed by SDS-PAGE (Fig. 7C) and ESI-QTOF. The mass spectra confrmed the formation of H3-Q56Q me , and only a small amount of hydrolyzed product was observed (Fig. 7B). To prove the generality of this method, Gln methylation was also generated at Q5 and Q94 on H3 with similar efficiency (Fig. S13 and S14 †). Finally, enzymatic digestion and collision-induced fragmentation were applied to H3-Q56Q me and H3-Q94Q me , and confrmed the site-specifc Gln methylation (Fig. 7D, and Fig. S14 †). Notably, this work provided a useful method to construct amide bonds on protein side-chains, and may offer solutions to the generation of other PTMs on proteins, such as N-glycosylation. ## Conclusions In summary, a side-chain esterifed glutamic acid analogue (BnE) has been genetically encoded into proteins in bacteria and mammalian cells. With this tool, we demonstrated the generation of N-terminal pGlu on recombinant proteins expressed in E. coli, and proved that BnE can serve as a chemical cage of Glu, offering a method for the expression of cytotoxic barnase. Furthermore, we showed that BnE on proteins can be facilely converted into HA, a useful metal chelator, and into acyl hydrazide, a handle that allows versatile protein modifcations, including histone Gln methylation. We believe this work provides unique tools to manipulate proteins, and would be useful in a number of felds.

chemsum

{"title": "Diverse protein manipulations with genetically encoded glutamic acid benzyl ester", "journal": "Royal Society of Chemistry (RSC)"}

the_production_of_biodiesel_from_plum_waste_oil_using_nano-structured_catalyst_loaded_into_supports

## Abstract: w www.nature.com/scientificreports/ in the production of biodiesel. Bentonite is an absorbent swelling clay consisting mostly of montmorillonite that been used previously to enhance catalytic activity of the catalysts to produce biodiesel [2][3][4][5][6][7][8] .Plum seeds are waste product that is generated from juice, food processing and cosmetic industries. Plum seeds produce over 50% of oil, this fruit is an important feed stock to produce biodiesel oil. The quality of biodiesel depends on the composition profile of fatty acids produced from the feedstock and is also important when we decide future biodiesel production possibilities 9,10 . The plum processing industry can be an interesting source of alternative, inexpensive oily raw materials for biodiesel synthesis due to the world production and processing of plums and high oil content in grain. Total global plum production in 2013 amounted to 11.5 million tons and about 12 million tons in 2020. The oil content in plum grains (PK) ranges from 32% to even 45.9%, which is similar to sunflower and rapeseed oil (about 40%) but is higher than in soybean (about 20%). Biodiesel is best alternative available today to conventional diesel. Everyday introduction of improved methods of production in new alternative feedstocks are ever rapidly cementing biodiesel strong position as a mainstream alternative. There are several benefits of biodiesel including local production, rural development, environment friendly fuel and economically sound product as compared to conventional diesel. Why biodiesel did not attain that popularity which was expected due to its well-known so many benefits. Biodiesel production is viable only if produced from non-food crops. To make it further popular and competitive it is necessary to use non-food crops. It will not only reduce production costs but will also end food verses fuel debate [11][12][13][14][15][16] . There are several clays including Bentonite, White pocha, Kolten and Sindh clays which are used in ceramic industry and available at very low cost. Granite stone powder is a waste product of stone processing industries and is a material available free of cost. All these materials were selected in the present study to prepare composite supports to hold catalyst potassium ferricyanide to produce biodiesel from non-edible waste seed oil of plum. The various parameters affecting the yield were tested for optimum production of biodiesel. The percentage biodiesel yield depends on different parameters like effect of catalysts concentrations, effect of reaction time and temperature, and oil to methanol ratio were optimized during the present study. The quality of biodiesel samples was monitored through standard procedures including iodine value, cetane number, specific gravity, density, and acid value. Finally, advanced instrumental techniques were used for characterization of catalysts, supports and produced biodiesel of zeolite with different catalysts and composite supports were prepared and also characterized. Materials and methodsMaterials and oil extraction. Plum stones were collected from local market, juice, and fruit shops of Faisalabad City, Pakistan. Plum stones were broken to obtain seeds. The collected seeds were dried to remove moisture 17 . Oil was extracted from crushed dried seeds by using of expeller machine. The extracted oil was allowed to stand for 24 h to settle down of impurities and particles. Free fatty acid contents (FFA) of oil was determined using acid-base titration. Bentonite, White pocha, Kolten and Sindh clays were obtained from Ceramic Industry, Gujranwala, Pakistan. All chemical used in the present study were of analytical grade. The production of biodiesel from plum waste oil using nano-structured catalyst loaded into supports Aasma Saeed 1 , Muhammad Asif Hanif 1* , Haq Nawaz 1 & Rashad Waseem Khan Qadri 2 The present study was undertaken with aims to produced catalyst loaded on low-cost clay supports and to utilize plum waste seed oil for the production of biodiesel. For this purpose, Bentonitepotassium ferricyanide, White pocha-potassium ferricyanide, Granite-potassium ferricyanide, Sindh clay-potassium ferricyanide, and Kolten-potassium ferricyanide composites were prepared. Transesterification of plum oil under the different conditions of reactions like catalysts concentrations (0.15, 0.3 and 0.6 g), temperature (50, 60, 70 and 80 °C), reaction time (2, 4 and 6 h) and oil to methanol ratio (1:10) was conducted. The maximum biodiesel yield was recorded for Bentonitepotassium ferricyanide composite. This composite was subjected to calcination process to produce Calcinized bentonite-potassium ferricyanide composite and a further improvement in biodiesel amount was recorded. The fuel quality parameters of all biodiesel samples were in standard range. Gas chromatographic mass spectrometric analysis confirmed the presence of oleic and linoleic acids in the plum seed oil. The characterization of composite was done using FTIR, SEM and EDX. Two infrared bands are observed in the spectrum from 1650 to 1630 cm −1 indicates that the composite materials contained highly hydrogen bonded water. The presence of surface hydroxyls groups can also be confirmed from FTIR data. SEM image clearly show the presence of nano-rods on the surface of Granite-potassium ferricyanide and Kolten-potassium ferricyanide composites. Another interesting observation that can be recorded from SEM images is the changes in surface characteristic of Bentonite-potassium ferricyanide composite after calcination (at 750 °C, 1 atm for 4 h). Calcinized bentonite-potassium ferricyanide composite found to contain more nano rod like structures at its surface as compared to Bentonite-potassium ferricyanide composite which contained spherical particles. EDX data of Bentonite-potassium ferricyanide composite and Calcinized bentonitepotassium ferricyanide composite show that after calcination carbon and oxygen was reduced. The other lost volatile compounds after calcination were of Na, Mg, Al, Si, and S. The XRD spectrum of pure bentonite showed the average crystal size of 24.46 nm and calcinized bentonite of 25.59 nm. The average crystal size of bentonite and potassium ferricyanide composite and its calcinized form was around 33.76 nm and 41.05 nm, respectively. Clay based minerals materials are commonly found around us and have applications as catalysts in the organic synthesis since long time. Several types of clay-based catalysts have designed and applied for use in organic synthesis including ion-exchanged, acidic, and basic clay catalysts. The use of abundant and commonly available natural kaolin is one such example that has been used for preparing precursors and catalytic supports. The clay mineral's original crystalline structure can be altered by means of different treatments in a controlled way to enhance their use as catalyst. The treatments that are most commonly performed for modification are done with pillaring, impregnation, intercalation, and acids that are effective in modifying the surface area, in optimization of the active sites, and to facilitate the attachment of the reagent molecules to active sites through the mesopores (average diameter of 20 to 500 ) 1 . The inorganic salts in the form of heterogenous catalysts such as KF impregnated on γ-Al 2 O 3 , Zn-Al(O) Ca-Al mixed oxides, CaO, and CaO-Fe 3 O 4 supports has been successfully utilized Preparation of catalyst and transesterification of oil. Bentonite potassium ferricyanide composite was prepared by mixing 25 g of potassium ferricyanide and 25 g of bentonite clay in 250 mL distilled water. The mixture was stirred for 5 min at 100 rpm and filtered. The obtained solid material was dried at 60 °C in an electric oven. By following similar procedure potassium ferricyanide catalyst was mixed with different support materials including White pocha, Granite, Sindh clay, and Kolten. Bentonite-potassium ferricyanide composite have shown maximum biodiesel yield during the present study. To improve catalytic activity of Bentonite-potassium ferricyanide composite further, this composite was subjected to at 750 °C for 4 h. Transesterification of plum oil under the different conditions of reactions like catalysts concentrations (0.15, 0.3 and 0.6 g), temperature (50, 60, 70 and 80 °C), reaction time (2, 4 and 6 h) and oil to methanol ratio (1:10) was conducted. Magnetic stirring was maintained at 300 rpm during all experiments. Glycerol was formed as a byproduct during biodiesel production. The upper biodiesel was separated from glycerol and washed with hot water until clear biodiesel layer was obtained. The biodiesel quality was accessed by the determination of density, specific gravity, pH, saponification value, acid value, cetane number, iodine value, free fatty acids contents and acid value 13, . Characterization of plum oil. GC-MS (gas chromatographic-mass spectrometric analysis) was conducted for the quantification of methyl esters present in the biodiesel. For this purpose, three samples were selected. GC-MS analysis was performed on a Perkin Elmer Clarus 600 GC System, fitted with an Elite-5MS capillary column (30 m × 0.25 mm i.d. × 0.25 μm film thickness; maximum temperature, 350 °C), coupled to a Perkin Elmer Clarus 600C MS. Ultra-high purity helium (99.999%) was used as a carrier gas at a constant flow of 0.2 ml/min. The injection, transfer line and ion source temperatures were 220, 200 and 200 °C, respectively. At the ionizing energy of 70 eVthe data was collected from 10 to 600 m/z by using 0.1 μL of sample with 50:1 spilt ratio. The temperature program for oven was as follows: 35 °C holds for 10 min, 10 °C/min 200 °C hold for 10 min. The unknown compounds were identified by the use of NIST 2011 (v.2.3 and Wiley, 9th edition). ## Results and discussion Optimization of biodiesel yield. The plum oil was transesterified into fatty acid methyl esters in the presence of following composite materials: (a) Bentonite-potassium ferricyanide composite (b) Calcinized bentonite-potassium ferricyanide composite (c) White pocha-potassium ferricyanide composite (d) Granitepotassium ferricyanide composite (e) Sindh clay-potassium ferricyanide composite (f) Kolten-potassium ferricyanide composite. Effect of catalyst concentration on yield of biodiesel produced from plum waste oil was studied at three catalyst concentrations 0.15, 0.3 and 0.6%. The biodiesel yield increased on increasing catalyst amount from 0.15 to 0.30%. The maximum biodiesel yield for all composite catalysts was obtained at 0.30%. A further increase in the catalyst amount decreased the biodiesel yield. The decrease in the biodiesel yield at increased catalyst concentrations might be due to increase in the viscosity of reaction mixture (Fig. 1). The effect of reaction times on biodiesel yield was evaluated from to 2 to 6 h (Fig. 2) at methanol to oil molar ratio 8:1, 0.3% catalyst, and 60 °C reaction temperature. The maximum biodiesel yield for (a) Bentonite-potassium ferricyanide composite (b) Calcinized bentonite-potassium ferricyanide composite (e) Sindh clay-potassium ferricyanide composite and (f) Kolten-potassium ferricyanide composite was obtained after 4 h. A further increase in the reaction time resulted in the loss of biodiesel yield may be due to breakdown of fatty acid methyl esters. However, (c) White pocha-potassium ferricyanide composite and (d) Granite-potassium ferricyanide composite have produced maximum of biodiesel after 6 h of reaction time. The determination of an optimum time to produce biodiesel is essential as it contributes to calculate the cost on pilot and commercial scales. The impact of transesterification reaction temperature was studied on three temperatures (50, 60, 70 and 80 °C) by keeping other variables constant as follows: reaction time of 4 h, catalyst amount 0.3%, and methanol to oil ratio of 10:1 (Fig. 3). The rate of transesterification reaction increased as the reaction temperature increased from 50 °C to 60 °C. On increasing reaction temperature from 60 to 80 °C, a decrease in the biodiesel yield was observed. Although, it is expected that biodiesel yield increases with the temperature, however, increasing temperature above 60 °C could result in the decrease of biodiesel yield due gasification of methanol 21 . Among all used composites as catalysts, the maximum biodiesel yield was obtained for Bentonite-potassium ferricyanide. The calcination of Bentonite-potassium ferricyanide composite have further increased the biodiesel yield. The highest biodiesel yield observed using calcinized Bentonite-potassium ferricyanide composite was due to increase in the average crystal size (as supported by XRD results) that has provided comparatively greater surface area for reactants for successful conversion into products. The XRD spectrum of pure bentonite showed the average crystal size of 24.46 nm and calcinized bentonite of 25.59 nm. The average crystal size of bentonite and potassium ferricyanide composite and its calcinized form was around 33.76 nm and 41.05 nm, respectively. ## Determination of fuel properties. The estimated values of various fuel quality parameters are tabulated in Table 1. Biodiesel density is important parameters as the fuel working in the fuel injector system and engine is strongly related to density value 22 . Amount of weight comprised in a unit volume is referred as density. Denser the oil more the energy it contains. Standards used to measure density of biofuels are 3675/12185 in European Union and D1298 in USA and are measured at reference temperature of 15 or 20 °C. Density which is measured by comparing with water's density is called relative density. The density of biodiesel measured relatively necessary for calculating the conversion of mass to volume and for determination of flow rate (Sanford et al., 2009). The densities of different biodiesel samples determined in the present study was between 0.856 to 0.877 kg/L. The recommended range of density lies between 0.86 and 0.90 g/cm 3 by EN 14214:2003 for a B100 type biodiesel. www.nature.com/scientificreports/ All biodiesel samples have density in the recommended range. These results reveal that produced biodiesel may be suitable for optimal performance. Acid value is defined "as the number of milligrams of potassium hydroxide (KOH) required to neutralize the free fatty acid in oil". The acid values of all biodiesel samples produced in the present study were in the standard range. Iodine value is "the measure of the total degree of unsaturation, and it provides useful guidance for preventing various problems in engines". The iodine value tells about stability and the presence of double bonds in the fatty acid methyl esters . The iodine values measured during the present study ranged from 140.2 to 168.3 g I 2 /100 g. Catalyst concentration is an important factor that affect the iodine value of produced biodiesel samples. Saponification value is "the amount of alkali required to saponify a given quantity of oil sample, which is expressed as the number of milligrams of KOH required to saponify 1 g of oil sample and is inversely proportion to the molecular weight of fatty acid of the biodiesel" 22 . The saponification value of plum oil biodiesel ranged from 103.78 to 186.23 mg/g. Cetane number is a fuel quality parameter related to the ignition delay time and combustion quality. According to UNE-EN 14214 (2003) specification, biodiesel should have minimum Cetane number of 51, while ASTM D6751-02 assigns 47 as the minimum cetane number for biodiesel. The cetane number of all biodiesel sample were greater than 50. Cetane values obtained in the present study has higher value as compared to the previous study on soya bean oil that ranged from 45 to 60 22 . ## Gas chromatographic analysis (GC-MS analysis). The fatty acid composition of plum seed oil is given in the Table 2. The chemical composition of biodiesel determines its fuel stability, while the stability of the fatty acid methyl ester depends on its number of double bonds, polyunsaturated fatty acids, which are more suscepti- www.nature.com/scientificreports/ ble to oxidation than the fatty acid having single bond 26 . The major fatty acids present in the plum oil were oleic acid and linoleic acid. According to a previous study, the oils having fatty acids with more than 15 carbon atoms as major components could be explored to produce good quality biodiesel 13 . Characterization of composite supports. FTIR spectra of various composites used in the present study including (a) Bentonite-potassium ferricyanide composite (b) Calcinized bentonite-potassium ferricyanide composite (c) White pocha-potassium ferricyanide composite (d) Granite-potassium ferricyanide composite (e) Sindh clay-potassium ferricyanide composite (f) Kolten-potassium ferricyanide composite are presented in the Fig. 1. The peaks in functional group region (4000-1500 cm −1 ) are characteristic of specific kinds of bonds, and therefore can be used to identify whether a specific functional group is present. The peaks in fingerprint region (1500-400 cm −1 ) arise from complex deformations of the molecule. They may be characteristic of molecular symmetry, or combination bands arising from multiple bonds deforming simultaneously. It can be seen from Fig. 1 that overall FTIR spectra of all composites significantly vary from each other although peaks in some regions matches too. However, frequencies of absorbing functional groups were different in all cases. A broad band of different intensities seen in the all-composite catalytic materials in the range of 3550-3200 cm −1 is due to O-H stretching vibrations. A clear C≡C band in the range of 2300-2100 cm −1 was observed for Bentonite-potassium ferricyanide composite, Calcinized bentonite-potassium ferricyanide composite, White pochapotassium ferricyanide composite and Granite-potassium ferricyanide composite. Sindh clay-potassium ferricyanide composite and Kolten-potassium ferricyanide composite did not show any clear band. However, a weak band was present for C≡C. The bands in the region of 1650-1600 cm −1 (C=C stretching), 1650-1580 cm −1 (N-H bending), 1550-1500 cm −1 (N-O stretching) and 1000-1200 cm −1 (C-O stretching) were observed in all composite materials, however, they have shown variable transmittance intensities. A clear difference between bentonite-potassium ferricyanide composite and calcinized bentonite-potassium ferricyanide composite band intensity can be seen in FTIR spectra (Fig. 4). Calcination, which refers to the heating of inorganic materials to remove volatile components. The release of volatile matter during calcination minimizes internal shrinkage in later processing steps that can lead to the development of internal stresses and, eventually, cracking or warping. Calcination treatment is an integral part during fabrication and activation of the heterogeneous catalysts 27 . The OH-bending shows vibrations of the inner surface OH groups were observed at 913 cm −1 and that of the surface OH groups near 936 cm −1 ; the surface hydroxyls are associated with additional bands near 701 and 755 cm −1 . Iron-bearing composite materials show typical of bands due to Fe(AlFeOH) at 865-875 cm −1 and compressing at 3607 cm −128 . Two infrared bands are observed in the spectrum from 1650-1630 cm −1 indicates that the composite materials contained highly hydrogen bonded water 29 . The band just above 3600 cm −1 (at 3620 cm −1 ) corresponds to the "inner hydroxyls" located on the plane common to octahedral and tetahedral sheets. The vibration of the "outer hydroxyls" located on the surface and along the broken edges of composites may be attributed to bands recorded at 3668 and 3652 cm −1 . Adsorption of ions and complexes on clay minerals is considered to occur as a result of surface complexation, ion exchange, electrostatic and hydrophobic interaction 30 . The adsorption capacity modes on mineral surfaces are primarily divided into complexes of the outer sphere and inner-sphere surface. In general, in the inner-sphere complexes, chemical interactions are stronger than in the outer-sphere complexes. The mobility of ionic species in the environment influences these differences in binding strengths . Scanning electron microscopy (SEM) images were recorded to study surface morphology of different catalytic materials (Fig. 5). SEM image clearly show the presence of nano-rods on the surface of Granite-potassium ferricyanide composite and Kolten-potassium ferricyanide composite. Another interesting observation that can be recorded from SEM images is the changes in surface characteristic of Bentonite-potassium ferricyanide composite after calcination. Calcinized bentonite-potassium ferricyanide composite found to contain more nano rod like structures at its surface as compared to Bentonite-potassium ferricyanide composite which contained spherical particles. In broader sense, SEM images show that catalyst loaded composite materials surface particles were different in size and of variable shape 34 . Energy-dispersive X-ray spectroscopy (EDX) is a powerful tool for the analysis of fine-grained clay mineral components, and Al-pillared clays in particular. It can be seen that Al, Si, Co, Ni and Fe are the primary elements of the untreated clay. The exchange process resulted in an increase in the composite's aluminum, iron and nickel content. EDX spectra for all composite materials were recorded and is presented in Fig. 6. The S, Al, S, Mg, P, Ca and O were recorded in the element analysis EDX, and various peaks of Fe and C of carbon with Fe 3 O 4 elements were also observed. Oxygen was present as a major element and shows the presence of most other compounds as oxygen derivatives. Bentonite-potassium ferricyanide composite was calcinized to produce Calcinized bentonite-potassium ferricyanide composite. EDX data of Bentonite-potassium ferricyanide composite and Calcinized www.nature.com/scientificreports/ bentonite-potassium ferricyanide composite show that after calcination carbon and oxygen was reduced. The other lost volatile compounds after calcination were of Na, Mg, Al, Si, and S. Within the clay structure, the content of these elements was not constant, and the pillaring process impacted it. The increase of Al-Si is due to presence of Al 13 poly cation in the Al-pillared clay particle. Which can be seen in the curves, the reflectance of the non-treated clay is lower than that of the particles containing Al-integrated clays and in the considered spectral region varies between 50 and 60%. The findings also showed that there are more EM waves reflected from the Al-exchanged clays than those from the Al-pillared ones. This could be due to the contribution of electrical dipoles or multiple reflection phenomena in Al-exchanged clays from the front face of the first layer 35 . The optimized catalytic materials that have shown highest transesterification ability were subjected to XRD studies (Fig. 7) including bentonite, calcinized bentonite, bentonite, and potassium ferricyanide composite and calcinized bentonite and potassium ferricyanide composite (Tables 3, 4, 5 and 6). Bentonite is a potential adsorbent and a swelling clay having montmorillonite as a major constituent. It generally consists of silicon dioxide (SiO 2 ), aluminum trioxide (Al 2 O 3 ), ferric oxide (Fe 2 O 3 ), calcium oxide (CaO), sodium oxide (Na 2 O), magnesium oxide (MgO) and potassium oxide (K 2 O). The bentonite consists of silicon (Si), aluminum (Al), oxygen (O), sulphur (S) and carbon (C) as major chemical elements. The XRD spectrum of pure bentonite showed the sharp These results are in close agreement with the international standard compound under JCPDS Card No. 01-088-0891. The average crystal size of crystallites was found to be around 24.46 nm 36 . Calcination is the process of high temperature heating, for the activation of natural clays and stony powdered materials, to enhance the catalytic potentials and adoption properties. The high temperature heating helps to reduce the overall moisture contents and liberate the entrapped associated gases from the deeper layers of material. The structural properties and morphological characteristics of the bentonite clay can significantly be improved by high temperature heating 37 . In the present study, the calcinized bentonite showed more sharp and clear peaks with less background noises as compared to the un-calcinized bentonite. This calcinized bentonite showed characteristic peaks at angle 2θ as follows: 21.03° (110), 26.77° (210), 36.84° (124) and 50.26° (144). These diffraction angles and hkl planes were in close agreement with the standard JCPDS Card No. 01-088-0891. The average crystallite size of calcinized bentonite as calculated by the Debye Sherrer Formula was 25.59 nm 36 . The modification of bentonite clay by the potassium ferricyanide has proved to be helpful in altering the structural properties, crystal lattices and catalytic potentials of composite material. In a recent study, 38 reported the use of bentonite and potassium ferricyanide as a catalyst, for the cost efficient production of biodiesel. In the XRD spectrum of bentonite and potassium ferricyanide, some sharp and pointed peaks were obtained at the angle 2θ: 21.76°, 26.79°, 27.96°, 29.29°, 34.35°, 55.24° and 60.02° corresponding to the following hkl planes: (110), ( 220), ( 210), ( 124), (400), ( 144), ( 600) and (440). These results are in close agreement with the internally available standard compounds under JCPDS Card No. 01-088-0891 and 73-0689. As per the Debye Sherrer Formula, the average crystal size of modified bentonite sample was around 33.76 nm 36,39 . In the present study, the combination of calcinized bentonite and potassium ferricyanide showed no significant variations as compared to the un-calcinized modified bentonite clay. Only the slight differences appeared in the clarity and sharpness of peaks as calcination improves the purity of sample 124), (400), (420), ( 144) and (600). These results are in close agreement with the international standards under JCPDS Card No. 01-088-0891 and 73-0689. The average crystallite size of the samples was found to be around 41.05 nm 36,39 . ## Conclusions Following important conclusions can be withdrawn from the present study. Plum seed oil is of toxic nature and can be added to list of those waste oils which can be further explored to produce biodiesel. The present study reported the use of Bentonite-potassium ferricyanide, White pocha-potassium ferricyanide, Granitepotassium ferricyanide, Sindh clay-potassium ferricyanide, and Kolten-potassium ferricyanide composites to produce biodiesel from plum seed oil. The maximum biodiesel yield for all composite catalysts was obtained at 0.30% catalyst concentration and 60 °C. The maximum biodiesel yield was recorded for Bentonite-potassium ferricyanide composite which further increased after calcination of the composite. The fuel quality parameters of all biodiesel samples were found in the standard range. The calcination process was remarkably effective in the removal volatile compounds from composite materials to generate further active sites to enhance biodiesel yield.

chemsum

{"title": "The production of biodiesel from plum waste oil using nano-structured catalyst loaded into supports", "journal": "Scientific Reports - Nature"}

synthesis_and_binding_studies_of_two_new_macrocyclic_receptors_for_the_stereoselective_recognition_o

## Abstract: We present here the design, synthesis, and analysis of a series of receptors for peptide ligands inspired by the hydrogen-bonding pattern of protein β-sheets. The receptors themselves can be regarded as strands 1 and 3 of a three-stranded β-sheet, with crosslinking between the chains through the 4-position of adjacent phenylalanine residues. We also report on the conformational equilibria of these receptors in solution as well as on their tendency to dimerize. 1 H NMR titration experiments are used to quantify the dimerization constants, as well as the association constant values of the 1:1 complexes formed between the receptors and a series of diamides and dipeptides. The receptors show moderate levels of selectivity in the molecular recognition of the hydrogen-bonding pattern present in the diamide series, selecting the α-amino acid-related hydrogen-bonding functionality. Only one of the two cyclic receptors shows modest signs of enantioselectivity and moderate diastereoselectivity in the recognition of the enantiomers and diastereoisomers of the Ala-Ala dipeptide (ΔΔG 0 1 (DD-DL) = −1.08 kcal/mol and ΔΔG 0 1 (DD-LD) = −0.89 kcal/mol). Surprisingly, the linear synthetic precursors show higher levels of stereoselectivity than their cyclic counterparts. ## Introduction Manipulation of protein-protein interactions is gaining interest as they are known to play a critical role in important biological processes such as the normal function of cellular/organelle structure, immune response, enzyme inhibitors, signal transduc-tion, and apoptosis. Rational protein surface recognition poses a challenging test to our actual knowledge of molecular design. Nevertheless, its practice and developments will provide a better understanding of protein-protein interactions. Interest- ingly, in molecule-based disease therapy, the disruption of protein-protein interactions by small molecules constitutes an alternative approach to the classical active-site enzyme inhibition design. One of the strategies employed for binding protein surfaces relies on the use of arrays of synthetic receptors originally designed for the recognition of oligopeptides. Consequently, the selective recognition of oligopeptides represents an intermediate step toward the recognition of protein surfaces . The studies of host-guest complexes as model systems of peptide-peptide interactions are of particular interest because they may provide insight into the structural basis of the high size/shape specificities and enantioselectivities exhibited by the complex protein-protein recognition processes that occur in biology. Moreover, short oligopeptides are themselves worthwhile targets for recognition and their conformational flexibility represents an added challenge to achieve selective binding. The preparation of synthetic receptors for the selective binding of short oligopeptides has potential applications in the development of diagnostic sensors, separation techniques, and therapeutic agents. With respect to this latter point, there is a significant interest in the advance of receptors that selectively bind the D-Ala-D-Ala dipeptide, the common target for the vancomycin antibiotics. This group of antibiotics is active against certain aerobic and anaerobic Gram-positive bacteria, and has been used for many years as treatment of last resort in clinical wards . However, vancomycin resistance has recently been identified among clinical isolates of several Gram-positive species . Therefore, although many examples already exist in the literature , the design and synthesis of new synthetic receptors for this dipeptide is still a relevant endeavor not only in terms of understanding the interactions that take place during vancomycin action, but also because the structures of the most efficient receptors prepared might be useful as scaffolds for future antibiotics. Herein, we report the design and synthesis of two new macrocyclic receptors, 1 and 2, conceived for the binding of dipeptides, in particular for the selective recognition of D-Ala-D-Ala. We also report on the studies performed using these two macrocyclic receptors, as well as their linear precursors, in the molecular recognition of a series of dipeptides and diamides with diverse hydrogen-bonding patterns. We rationalize the observed modulation of their binding affinity as a function of the hydrogen-bonding pattern exhibited by the target molecule. We also describe the levels of stereoselectivity displayed by these receptors in the recognition of the diastereoisomers and enantioisomers of Ala-Ala dipeptide. We explain the differences observed in their binding abilities as a function of conformational rigidity (macrocyclic vs linear receptors). ## Design of the synthetic receptors: general considerations The design of the receptors described in this article is based on the interactions that occur in the β-sheets commonly found in the secondary structure of many biologically relevant proteins. We start from a schematic termolecular complex mimicking a three-stranded β-sheet in which the central strand corresponds to the target guest peptide and the two outer strands constitute the structure of the host (Figure 1). In this design, we employ some of the properties of the β-sheet structure -the convergence of hydrogen-bonding patterns and the presence of exposed side chains. It is worth mentioning that the β-sheet structure has already been used as the inspirational binding motif for the preparation of other synthetic receptors for peptides . However, we believe that our design includes some novelties. To reduce the conformational flexibility and confer a certain degree of preorganization to this type of receptor, the use of one or two linkers connecting the two peptide strands is mandatory. The principal difference with respect to previous designs of β-sheet-based synthetic receptors is that the connection between the two peptide strands, used as the receptor's binding sites, emerges from their side chains and not from their C-or N-terminus. In our final design, we propose the introduction of two linkers connecting the two peptide strands affording, a macrocyclic structure. In doing so, we expect that the molecular recognition properties of the designed receptor will also benefit from the macrocyclic effect . Simple molecular modeling studies revealed that a benzophenone unit would be ideally suited to span the gap between two methyl side chains emerging from alanyl residues of the outer strands in the three-stranded β-sheet complex (Figure 1). In proteins, adjacent β-strands can form hydrogen bonds in antiparallel, parallel, or mixed arrangements. In an antiparallel arrangement, the successive β-strands alternate directions so that the N-termini of two adjacent strands are at opposite ends. In a parallel arrangement, all of the N-termini of successive strands are oriented in the same direction . In contrast, successive strands in a mixed-mode arrangement may be parallel or antiparallel to each other. To examine the influence of the relative orientation of the two receptor strands on their binding abilities, we conceived and synthesized two analogous receptors mimicking these two types of arrangements present in the β-sheet structure (Figure 2). The outer peptide strands of receptor 1 are arranged antiparallel to each other ("antiparallel receptor"). That is, the stereogenic center of the C-terminus of one strand is covalently connected to the stereogenic center of the N-terminus of the other strand. Receptor 1 is anticipated to form a mixed-mode sheet structure with an included peptide ligand. Conversely, the two outer strands of receptor 2 are oriented parallel to each other ("parallel receptor"), such that the covalent connections between strands join similar stereogenic centers, C-terminus with C-terminus and N-terminus with N-terminus. Receptor 2 is anticipated to form an antiparallel sheet structure with the included peptide ligand. This change in connectivity does not involve any inversion of the stereogenic centers but only a modification in the sequence of peptide-coupling reactions that yield the cyclic structure, and will be explained below. The exploration of the conformational space of both macrocycles, using molecular modeling, indicated the existence of a built-in cavity. These studies also suggested that the reduced conformational flexibility of the receptors avoids the complete collapse of the cavity through the formation of intramolecular hydrogen bonds. Moreover, we were able to minimize structures for the complexes formed between both receptors and n-C 6 H 13 CO-D-Ala-D-Ala-NH 2 in which the dipeptide is threaded through the macrocycle (Figure 3). In these minimized structures, the hydrogen-bonding groups of the receptor converge toward the center of the macrocycle. The macrocycle is large enough to accommodate the threading dipeptide without incurring any substantial steric clashes. We also observed appropriate complementarity between the hydrogen-bonding groups of substrate and receptor (Figure 3). The analysis of the structures of the minimized complexes revealed that they are stabilized by the formation of the same number of hydrogen bonds, that is, five. The hypothesized "endo" structure for the complexes of 1 and 2 with n-C 6 H 13 CO-D-Ala-D-Ala-NH 2 , in which the ligands thread through the receptor's macrocycle, also allows for the possibility of binding short amino acid sequences not necessarily located on the edges of larger peptides. Other considerations, apart from preventing intramolecular hydrogen bond formation, related to the use of bis(alanyl)benzophenone rigid linkers include: a) to avoid steric clashes between the methyl groups of the target peptide and the benzophenone linking chains, the stereogenic centers in the linkers must have the (S) configuration, opposite to that of the bound peptide (R); b) the benzophenone aromatic ring will also provide a hydrophobic pocket for the neighboring side chains of the target peptide, and may promote the formation of additional CH-π and π-π intermolecular interactions. The pleat of the bound D-Ala-D-Ala peptide is inverted because of the unnatural stereochemistry and the resulting complex is not exactly a β-sheet. Attaching the linking groups in the side chains leaves the ends of the receptor strands open, allowing the introduction of additional interactions between the receptor and the chain of a larger peptide. As schematically depicted in Figure 2, we planned that both receptors could be obtained through cyclic dimerization, through the formation of two peptide bonds, of two S,Sbis(alanyl)benzophenone units 3. In turn, the synthesis of the protected S,S-bis(alanyl)benzophenone units 3 could be easily achieved by means of Stille carbonylative cross-coupling reactions of two adequately bisprotected S-phenylalanine derivatives, iodo-aryl 4 and trimethylstannyl-aryl 5, following experimental procedures described in recent literature reports . ## Synthesis The synthetic strategy designed for the construction of receptors 1 and 2 involves the use of a carbonylative cross-coupling reaction between two aryl derivatives (iodo-aryl 4 and trimethylstannyl-aryl 5) to prepare 4,4'-bis(alanyl)benzophenones 3, followed by macrocyclization of two molecular units of 3. The macrocyclization reaction of two 4,4'-bis(alanyl)benzophenones 3 will be promoted by the sequential and regioselective formation of two peptide bonds between them, the first one through an intermolecular reaction and the second one intramolecularly, affording the desired macrocyclic structures 1 and 2. The main dissimilarity between the two synthetic Scheme 1: Synthesis of tetraprotected bis(alanyl)benzophenones 3 from L-phenylalanine 7. strategies resides in the type of functional groups that each 4,4'bis(alanyl)benzophenone 3 supplies to the macrocyclization reaction. Thus, for the synthesis of antiparallel receptor 1 each benzophenone unit will provide, in an alternative way, one carboxylic and one amino function to the final macrocyclic skeleton. Conversely, for the synthesis of antiparallel receptor 2, one benzophenone unit will donate its two carboxylic acid functions while the other will participate with its two amino groups. To achieve the regioselective control demanded in the macrocyclization reactions, a precise selection of the orthogonal protecting groups to be included in the bis-amino acid functionalities of the benzophenone derivatives 3 is needed. The starting material for both synthetic routes is 4-iodo-Lphenylalanine ( 6). We prepared 6 in multigram scale starting from commercial L-phenylalanine (7) by following a described procedure consisting in the iodination of 7 in acetic acid solution in the presence of I 2 , NaIO 3 , and sulfuric acid (Scheme 1). We obtained (S)-6 in enantiomerically pure form in 50% yield. Since we plan to assemble the 4,4'-bis(alanyl)benzophenones 3 by a Stille carbonylative cross-coupling reaction, the required trimethylstannyl derivatives 5 should be easily prepared from adequately diprotected phenylalanine iodides 4 (Scheme 1). The mild reaction conditions used in the carbonylative crosscoupling permit the use of common protecting groups of peptide synthesis . This characteristic of the carbonylative cross-coupling reaction allowed us to achieve the differential protection of the two amino acid moieties present in the 4,4'bis(alanyl)benzophenones 3 by protecting separately the functional groups in the reaction partners, 4 and 5, before attempting the cross-coupling. We prepared a single orthogonally protected benzophenone 3a for the synthesis of the antiparallel macrocycle 1. In contrast, the synthesis of parallel receptor 2 called for the preparation of two differently protected bis(alanyl)benzophenone units, 3b and 3c. All iodo-phenyl derivatives 4 were prepared in high yields using standard procedures (Scheme 1). Thus, 4-iodo-L-phenylalanine 6 (I-Phe) was converted into the methyl ester hydrochloride by treatment with thionyl chloride in methanol followed by acylation of the amino group with tert-butyl dicarbonate to yield Boc-I-Phe-OMe, 4a. Iodo-L-phenylalanine 6 was acylated under Schotten-Baumann conditions with benzyl chloroformate to obtain the N-protected amino acid Cbz-I-Phe. This compound was esterified with 2-(trimethylsilyl)ethanol using DCC as coupling agent, providing Cbz-I-Phe-TMSE, 4c. In a different reaction, Cbz-I-Phe was treated with 4-nitrobenzyl bromide and triethylamine to afford Cbz-I-Phe-PNB, 4b . Finally, 6 was treated with Fmoc hydroxysuccinimide (Fmoc-OSu) to obtain the Fmoc N-protected amino acid that was subsequently esterified with diazomethane affording Fmoc-I-Phe-OMe, 4d. The diprotected aryl iodides, 4a and 4b, were converted uneventfully to the corresponding diprotected aryl trimethylstannane derivatives, 5a and 5b, by reaction with hexamethylditin catalyzed by Pd(0) under inert atmosphere. The organostannanes 5 showed signs of decomposition with time, and they were freshly prepared just before being used in the cross-coupling reaction. The 4,4'-bis(alanyl)benzophenones 3 were prepared in an orthogonally protected form by carbonylative coupling between diprotected iodo-aryl derivatives 4 and diprotected aryl trimethylstannanes 5, using the experimental conditions described by Morera and Ortar for similar substrates . The reactions were performed at 90 °C under atmospheric CO pressure in the presence of PdCl 2 /PPh 3 , proceeding smoothly to give derivatives 3 with isolated yields, after column chromatography, in the range of 53-61%. This complete synthetic sequence is reminiscent of the work of Lei et al. for the preparation of phosphinate bis-amino acids . This convergent route allows the installation of diverse and differentiable functionality in a small molecule like 3. The sequential peptide coupling of two units of 4,4'bis(alanyl)benzophenones 3a should lead to the construction of the designed macrocyclic bis-dipeptide receptor 1. Benzophenone 3a was converted into the carboxylic acid 8 by treatment with tetrabutylammonium fluoride (Scheme 2). In a separate reaction, 3a was treated with trifluoroacetic acid to remove the Boc group and produce the trifluoroacetic salt of amine 9 . Both deprotection reactions proceeded uneventfully in almost quantitative yields. The PNB group in 3b was removed using a mixture of SnCl 2 and phenol in acid media and the Fmoc in 3c using piperidine to obtain 10 and 11, respectively (Scheme 2). Next, we carried out intermolecular peptide-coupling reactions between 8 and 9, as well as between 10 and 11 to obtain the linear tetrapeptides 12 and 13, direct precursors of receptors 1 and 2, respectively. The best results for the coupling reactions were obtained when using a combination of HATU/NMM in DMF at room temperature (Scheme 3). The analysis of the crude reaction mixtures by HPLC and 1 H NMR spectroscopy revealed that both tetrapeptides, 12 and 13, were obtained as mixtures of two diastereoisomers. Most likely, the stereogenic center in the α-position with respect to the carboxylic group undergoing activation during peptide coupling was partially epimerized. The all-S diastereoisomers, (S)-12 and (S)-13, were the major products detected in the crude reaction mixture. They were isolated as pure compounds using preparative reverse-phase HPLC and fully characterized by a complete set of high-resolution spectra. Scheme 2: Deprotection reactions of bis(alanyl)benzophenone units 3. However, the subsequent sequence of reactions directed toward the macrocyclic receptors utilized, as starting material, the diastereomeric mixture of 12 or 13 obtained by flash chromatography purification of the reaction crude. The deprotections of the diastereoisomeric mixtures were carried out using standard methods. First, we used fluoride to cleave the TMSE group, and subsequently, we removed the Boc group by the action of TFA. We obtained the bis-deprotected tetrapeptides 15 and 17 in high yield (70-80%). The macrocyclization reactions of the linear tetrapeptides, 15 and 17, were carried out under high-dilution conditions. Using a syringe pump and under inert atmosphere, a DMF solution of the corresponding linear tetrapeptide was added dropwise, over a period of 12 h, to a stirred DMF solution containing the coupling agent and the base. The purification of the crude macrocyclization reactions using flash chromatography afforded the expected macrocyclic products in acceptable yields but as complex mixtures of diastereoisomers. HPLC-MS analysis of the isolated fraction showed the presence of four different peaks in the chromatogram producing ions with molecular mass value corresponding to the expected cyclic structure. We tentatively assigned the two major peaks to cyclic diastereoisomers formed during the intramolecular peptide-coupling reaction of the all-S linear tetrapeptide. As discussed above, one of the two diastereoisomers is probably the outcome of the epimerization reac- tion experienced by the stereogenic center in the α-position with respect to the carboxylic group undergoing activation. Likewise, the two minor peaks should correspond to cyclic diastereoisomers formed from macrocyclization and concomitant epimerization reactions experienced by the minor linear tetrapeptide S,S,S,R also incorporated into the starting material. Figure 4b depicts the HPLC chromatogram obtained from the analysis of the purified fraction containing the mixture of diastereoisomers of receptor 1. Using normal-phase preparative HPLC, we isolated the two major products of the macrocyclization reaction of 15 as pure compounds. The structures of the isolated products were assigned by means of standard spectroscopic techniques and symmetry considerations to cyclic diastereoisomers of receptor 1. Furthermore, the structure of the major product of the cyclization of 15 was also characterized in the solid state by X-ray diffraction and proved to be the desired all-S antiparallel cyclic receptor 1. The results obtained in the macrocyclization of tetrapeptide 17 were completely analogous. The all-S diastereoisomer corresponds to macrocyclic receptor 2, and was the major product isolated from the purification of the reaction mixture using normal-phase preparative HPLC. Receptor 2 was fully characterized by means of standard spectroscopic techniques. Initially, we used HBTU/NMM for activation of the intramolecular peptide bond formation. We observed considerable epimerization at the stereogenic α-carbon. We assessed the coupling reaction using different coupling methods, HATU/ NMM and PyAOP/DIEA , and found that although the overall reaction yields were independent of the coupling method, the epimerization diminished substantially when the PyAOP/DIEA combination was used (Figure 5c). ## Conformational studies The 1 H NMR spectra of chloroform-d solutions of the diastereomerically pure all-S cyclic receptors 1 and 2, as well as those of their linear tetraprotected precursors, 15 and 17, were temperature-dependent (Figure 6). We attribute this temperature dependence to the existence of conformational equilibria that are in a slow chemical exchange regime with respect to the NMR time scale, i.e., the rotation of the C-N single bond in the carbamate protecting groups . Upon increasing the temperature of chloroform-d solutions of 1, 2, 15, and 17, the proton signals became sharper and well defined, which is indicative that the chemical exchange due to the conformational equilibria has been accelerated. Conversely, cooling the samples slows down the rate of the chemical exchange. Thus, at low temperature, we observed the appearance of new proton signals that were assigned to different conformations. We observed another general trend in the variable-temperature 1 H NMR spectra, that is, as the temperature was lowered, the NH signals shifted downfield. This behavior suggested that the cyclic and acyclic peptides may dimerize or oligomerize in chloroform solution through the formation of intermolecular NH•••O hydrogen bonds. We have already observed the formation of intermolecular hydrogen bonds in the solid-state structure of receptor 1 (Figure 7). Before undertaking the study of the binding and molecular recognition properties of the receptor series, and due to their tendency to aggregate in solution, we quantified their dimeriza-tion constants in chloroform. The calculation of the dimerization constants relies on the chemical shift changes observed for certain proton signals of the receptors when their 1 H NMR spectra are acquired at different concentrations. In particular, the receptors' NH signals shift downfield when the concentration of the solution is increased, indicating the formation of aggregates in the solution that are stabilized through hydrogen bonding. The observed chemical shifts for the NH signals were analyzed mathematically using the HypNMR software and a simple theoretical dimerization binding model . We obtained a good fit between the experimental and theoretical data. Additional conclusions can be drawn from the data presented in Table 1. Macrocyclic receptors 1 and 2 show greater tendency to dimerize than their linear precursors. A stronger dimerization tendency for the antiparallel cyclic receptor 1 also becomes apparent. We used a wide range of guest molecular structures to examine the molecular recognition properties of receptors 1, 2, 15, and 17 (Figure 8). We selected a series of diamides to evaluate the effect that the hydrogen-bonding pattern produces in the binding affinity. We also investigated the molecular recognition properties of the receptor series with a set of dipeptides. The effect of the size of the amino acid substituents in the dipeptide series (Ala-Ala vs L-Phe-L-Phe) was investigated to shed some light on the geometry of the complexes formed with the cyclic receptors. Finally, the stereoselective recognition properties of the receptors were derived from their binding interactions with the four diastereoisomers of Ala-Ala. The molecular structures of all selected guests have several hydrogen-bonding groups, making them natural candidates to dimerize in solution. Therefore, before studying the interactions of these guests with the receptors, we studied their dimerization behavior in chloroform solution. Using the same methodology described above for the receptors, we calculated the dimerization constants of all guest molecules. The values obtained are summarized in Table 2. With an additional amide group with respect to diamides, the dipeptide dimers can be stabilized by a higher number of hydrogen bonds. The value of the dimerization constant of the diamide of fumaric acid stands out from the rest, likely due to the higher conformational rigidity of this compound (preorganization). Figure 9 depicts the 1 H NMR spectra acquired in the variable-concentration experiments used for the calculation of the dimerization constant of fumaramide. The NH proton signals experience a significant downfield shift on increasing the concentration of fumaramide. Having determined the dimerization tendency of host and guest molecules, we initiated the study of the molecular recognition properties of the receptors toward the different guests. All binding constants were determined using 1 H NMR titration experiments. As an example, Figure 10a shows a series of spectra acquired during the titration of receptor 2 with n-C 6 H 13 -D-Ala-D-Ala-NH 2 . We monitored the chemical shift changes experienced by the NH proton signals of the receptor and of the guest when a 1 mM chloroform-d solution of the receptor is treated with incremental amounts of the guest. The titration data were fitted to a theoretical binding model considering the exclusive formation of a 1:1 complex, and the existence of dimeric aggregates of both the receptor and the guest. Figure 10b depicts the experimental data of the titration fitted to the theoretical binding isotherm derived from the abovementioned theoretical model. The values of the calculated stability constants for the 1:1 complexes are summarized in Table 3 and Table 4. The analysis of the tabulated data allowed us to draw several conclusions (Table 3 and Table 4). The macrocyclic receptors do not show any affinity for the complexation of diamides in which the two amide groups are spanned by three methylene groups (glutaramide and propane-1,3-diamine). However, these receptors do form complexes with the rest of diamides showing certain degree of selectivity in response to the hydrogenbonding pattern (Table 3). The antiparallel macrocycle 1 exhibits a moderate preference for the hydrogen-bonding pattern DAAD (D = hydrogen bond donor, A = hydrogen bond acceptor) instead of ADAD when just one methylene group spans the two amide groups (ΔΔG 0 (malonamide-Gly) = −0.96 kcal/mol). In contrast, parallel receptor 2 effectively discriminates in favor of the hydrogen-bonding pattern ADDA when two methylene groups span the amide groups (ΔΔG 0 (ethylenediamine-succinamide) = −2.35 kcal/mol). The calculated stability constants are, in general, lower than the values expected for a complex that can be stabilized by an array of not adjacent four hydrogen bonds in chloroform solution (K ≈ 10 4 M −1 ). The stability constant values determined for the complexes formed by both cyclic receptors and fumaramide are more consistent with our estimate. Most likely, the high thermodynamic stability calculated for the complexes of fumaramide in comparison with the rest of diamides resides in the reduced conformational flexibility of the substrate (ΔΔG 0 1 (fumaramide-succinamide) = −2.46 kcal/mol and ΔΔG 0 2 (fumaramidesuccinamide) = −2.79 kcal/mol). When the association constant values obtained for the DAAD hydrogen-bonding pattern are compared, it becomes evident that both cyclic receptors exhibited a marked preference for the diamides in which the NH-CO groups are separated by just one methylene group (ΔΔG 0 1 (malonamide-succinamide) = −1.56 kcal/mol and ΔΔG 0 2 (malonamide-succinamide) = −1.01 kcal/mol). Not surprisingly, the binding affinities calculated for the cyclic and acyclic receptors toward the dipeptide series were higher than those for the diamides. Dipeptides have an additional amide hydrogen-bonding group. The degree of stereoselectivity displayed by the cyclic and acyclic receptors was low (two possible binding geometries for the complexes formed between the macrocyclic receptors and n-C 6 H 13 -L-Phe-L-Phe-NH 2 are shown in Figure 11). The cyclic antiparallel receptor 1 showed reduced signs of enantioselectivity and moderate diastereoselectivity in the recognition of the enantiomers and diastereoisomers of the Ala- 3). The difference in free energy measured for the complexes of 2 with the four diastereoisomers of Ala-Ala was in the order of 0.3 kcal/mol. We also investigated the complexation affinity of the cyclic receptors toward n-C 6 H 13 -L-Phe-L-Phe-NH 2 , with the aim of gaining some information about the geometry of the complex. Molecular modeling suggested that although the formation of an endo-complex in which n-C 6 H 13 -L-Phe-L-Phe-NH 2 is threaded through the macrocyclic ring of the receptor is plausible, the steric clashes detected between the dipeptide side chains and the benzophenone linking units should significantly reduce the binding affinity of the cyclic receptors for this substrate or even favor the formation of an alternative complex with exogeometry. Unexpectedly, the stability constant values that we calculated for the 1:1 complexes of the cyclic receptors and n-C 6 H 13 -L-Phe-L-Phe-NH 2 were higher than those for any of the complexes with Ala-Ala (Table 3). Probably, additional intermolecular interactions between the receptors and the phenyl side chains are responsible for the increase in affinity. The low stereoselectivity exhibited by the cyclic receptors, together with the lack of selectivity for the size of the amino acid side chain, encourages us to propose that the geometry of the 1:1 complex is, most likely, exo-cyclic. In other words, the dipeptide is not threaded through the cyclophane skeleton of the receptor but bound externally. This hypothesis is also supported by the fact that we were unable to observe upfield shifts in any of the protons of the dipeptide during the binding experiments. The inclusion of the dipeptide in the aromatic cavity of the receptor should produce the shielding of some of its protons due to the anisotropic magnetic current produced by the aromatic rings. The linear receptors 15 and 17 seem to be more promiscuous in the interaction with the diamides (Table 4). In general the binding affinities are low, except for the fumaramide. The linear receptor 17 shows moderate selectivity for the hydrogenbonding pattern DAAD instead of ADAD when n = 2 (ΔΔG 0 (succinamide-ethylenediamine) = −0.92 kcal/mol) but selects the hydrogen-bonding pattern ADAD when n = 1 (ΔΔG 0 (Glymalonamide) = −0.90 kcal/mol). Surprisingly, linear receptors 15 and 17 exhibited higher levels of stereoselectivity than their cyclic counterparts (Table 4). Receptor 15 displayed the highest enantioselectivity we have measured in the molecular recognition of the D-Ala-D-Ala dipeptide (ΔΔG 0 15 (DD-LL) = −1 kcal/mol) and an acceptable level of diastereoselectivity (ΔΔG 0 15 (DD-DL) = −1.60 kcal/ mol). Even higher values of diastereoselectivity were obtained when studying the interaction between the linear receptor 17 and Ala-Ala diastereomers (ΔΔG 0 17 (DD-DL) = −2.18 kcal/mol and ΔΔG 0 17 (DD-LD) = −1.70 kcal/mol). We attribute the surprising and superior stereoselectivity measured for the linear receptors to their higher conformational flexibility compared with the cyclic analogs. This enhanced conformational flexibility allows them to adopt a more effective binding conformation for the sensing of the substrate's chirality. ## Conclusion We have designed two macrocyclic receptors for the stereoselective recognition of dipeptides on the basis of the interactions that occur in the β-sheets commonly found in the secondary structure of many biologically relevant proteins. The geometry of the putative complex used in the design of the receptors implies the threading of the dipeptide guest through the macrocyclic skeleton of the receptor. The two designed macrocycles, 1 and 2, have been synthesized and fully characterized. One of the key synthetic steps, which is common to both synthetic routes, consists in the use of a Stille carbonylative cross-coupling reaction that affords orthogonally tetraprotected 4,4'bis(alanyl)benzophenone units in good to acceptable yields. Sequential deprotection reactions combined with the formation of two consecutive amide bonds between two units of 4,4'bis(alanyl)benzophenone produced the macrocyclic receptors in low yield. Notwithstanding the epimerization reactions observed in the formation of the peptide bonds of the macrocyclic structures, both receptors have been isolated as single diastereoisomers. The molecular structure of receptor 1 has been confirmed by single-crystal X-ray diffraction analysis. Although molecular modeling suggested that the cyclic receptors can adopt a conformation with a cavity size large enough to include a peptidic substrate, the X-ray structure obtained for antiparallel receptor 1 shows the collapse of the designed cavity. Although crystal packing may contribute to this conformational change to some degree, the solid-state structure of 1 suggests that the optimal conformation for binding is probably not the lowest-energy conformation. The prepared macrocyclic receptors 1 and 2 as well as their acyclic tetraprotected precursors 15 and 17 show a moderate tendency to aggregate in chloroform solution. Dilution studies carried out at room temperature show that the variation in chemical shift fits a simple theoretical dimerization model, although higher order aggregation cannot be ruled out. Using 1 H NMR titration experiments we have determined the association constant values of the 1:1 complexes formed between receptors 1, 2, 15, and 17 and a series of diamides and dipeptides. We have observed that each receptor shows different selectivities in the recognition of the hydrogen-bonding patterns present in the diamide series as well as of the number of methylene groups used to separate the two amide functions. However, when the association constant values obtained for the DAAD hydrogen-bonding pattern are compared, it becomes clear that both cyclic receptors exhibited a marked preference for the diamides in which the NH-CO groups are separated by just one methylene group. It is worth noting that a single methylene unit was used as the spacer for the diamide guest used in the receptors' design. We also investigated the stereoselective recognition properties of the synthesized receptors using the four diastereoisomers of the Ala-Ala dipeptide as guests. The low stereoselectivity displayed by the cyclic receptors, together with their insensitivity to the size of the amino acid chain of the dipeptide guest, allows us to propose that the topology of the 1:1 complexes is not a pseudorotaxane as initially proposed in our design. Most likely, the guests, dipeptides and diamides, bind to the hydrogen-bonding groups that are directed toward the exterior of the aromatic cavity. If macrocyclization results in the receptor adopting a low-energy conformation different from that envisioned in the modeled structures, then preorganization will have created an additional energetic barrier to endo-complexation. Finally, the affinity and surprising stereoselectivity exhibited by the linear receptors 15 and 17 are very difficult to rationalize with an endo-complex geometry. We conclude with the caveat that the analysis here pre-supposes that the receptors respond to different ligands with similar binding modes. Due to the complexity of the system, we have not attempted to analyze the possibility that multiple binding modes -exo-binding, endo-binding -all operate simultaneously and to varying degrees depending on the ligand.

chemsum

{"title": "Synthesis and binding studies of two new macrocyclic receptors for the stereoselective recognition of dipeptides", "journal": "Beilstein"}

controlled_growth_of_the_non-centrosymmetric_zn(3-ptz)_2_and_zn(oh)(3-ptz)_metal-organic_frameworks

## Abstract: Non-centrosymmetric single-crystal metal-organic frameworks (MOF) are promising candidates for phase-matched nonlinear optical communication. However, the typical hydrothermal synthesis conditions produce small crystals with relatively low transmittance and poor phase matching. In the search for optimal crystal morphology we study the effect of the metal-to-ligand molar ratio and reaction pH on the hydro-thermal synthesis of the non-centrosymmetric Zn(3-ptz) 2 and Zn(OH)(3-ptz) MOFs with in-situ ligand formation. In acidic environments, we find that decreasing the amount of ligand below the stoichiometric molar ratio 1:2 also produces highly transparent single-crystal octahedrons of Zn(3-ptz) 2 . In alkaline environments, we obtain long rod-like Zn(OH)(3-ptz) crystals whose length exceeds previous reports by up to four orders of magnitude. All reaction products are characterized by using p-XRD, FTIR and optical and scanning electron microscope. Additionally, we find an alternative synthesis route for the recently reported high-energy MOF Zn(3-ptz)N 3 . Potential applications of these results in the development of MOF-based nonlinear optical devices are discussed. ## Introduction Functional crystalline materials based on metal-organic frameworks (MOF) have been intensely studied over the past decade due to the diverse set of applications enabled by properties such as ultra-high porosity (1)(2)(3)(4)(5)(6), record-high pore aperture size (5,7), high thermal and chemical stability (8)(9)(10)(11)(12)(13) and nonlinear optical activity (14)(15)(16)(17)(18)(19)(20)(21)(22)(23). MOF materials have been successfully used for applications in gas storage and separation (24)(25)(26)(27), fuel cells (28)(29)(30)(31), chemical sensing (32)(33)(34), biomedical imaging (35)(36)(37) and drug delivery (38,39). Several methods can be used to synthesize MOFs including microwave-assisted synthesis (40,41), slow evaporation synthesis (42,43), sonochemical synthesis (44,45), mechanochemical synthesis (46,47), electrochemical synthesis (48,49) and hydro/solvo-thermal synthesis (50,51). Different synthesis methods produce structures with characteristic size, homogeneity and morphology, which are directly related with the performance of the grown MOF crystals in applications. Therefore for each synthesis method, it is necessary to understand and control key synthesis parameters such as pH, temperature, pressure and metal-to-ligand molar ratio, in order to optimize the desired MOF structure for a target application. The variation of the metal-to-ligand molar ratio in a MOF synthesis has been shown to affect the topology and dimensionality of the crystal structures. Wu et al. (52) reported the synthesis of [Zn(hfipbb)(H 2 hfipbb) 0,5 ] n , a 3D two fold parallel interpenetrating pillared network, and [Zn 2 (hfipbb) Our work focuses on the synthesis of non-centrosymmetric MOFs with tetrazole-based ligands for which nonlinear optical activity is reported in powder form, for which phase-matching is not possible (55). Nonlinear optical MOF crystals could be used for applications in quantum cryptography based on entangled photon generation via spontaneous parametric down conversion (SPDC), a second-order nonlinear quantum optical process, provided that phase-matching conditions can be achieved. Tetrazole compounds are promising ligands for nonlinear optical MOFs formation due to their strong dipolar characteristics (56) and large optical response (57). From the synthetic point of view, tetrazole-based ligands are interesting because they possess a large number of coordination modes that can be selectively activated by varying the reaction pH, as recently demonstrated in Ref. (58) for the non-centrosymmetric MOF bis [5-(3-pyridyl)tetrazolato]zinc(II) Zn(3-ptz) 2 . This MOF crystal was first obtained by Wang et al. (59) using the Demko-Sharpless method for in-situ formation of ligand 3-ptz starting from sodium azide (NaN 3 ) and 3-cyanopyridine in the presence of a suitable Lewis acid, giving [Zn(3-ptz) 2 ] under hydro-thermal conditions for 24 hours at 105°C, using ZnCl 2 as the Lewis acid catalyzer. We also produce the noncentrosymmetric MOF catena-((μ 3 -5-(3-Pyridyl)tetrazol-N,N',N'')-(μ 2 -hydroxo)-zinc) [Zn(OH)(3-ptz)], obtained previously under different conditions (60). The zinc tetra-aquo coordination compound Zn(H 2 O) 4 (3-ptz) 2 (59) is also obtained as a by-product. Besides, we report an alternative synthesis route for the non-porous 3D framework [Zn(3-ptz)N 3 ], which exhibits energetic behavior due to the presence of structural azide groups in the unit cell (61). In Fig. 1 we plot the set of pH and metal-to-ligand molar ratios used in this and previous work. In this work, we study the effect of varying the molar ratio of the metal salt Zn(NO 3 ) 2 and the ligand precursor 3cyanopyridine on the non-centrosymmetric MOF structures and other compounds obtained. Keeping constant the molar ratios at which pure MOFs crystals are obtained we also studied the variation of the mixing pH and corresponding structures are reported and analyzed. We summarize the reaction parameter space explored in this work in ## Results and Discussion We first searched for an optimal metal-to-ligand molar ratio to obtain high-quality [Zn(3-ptz) 2 ] crystals, by keeping the pH unaltered relative to previous work in Refs. ( 59) and (58). For the metal-to-ligand molar ratios 1:1 and 1:2, we varied the pH to search for improved crystal quality in terms of purity, yield and optical transparency. Figure 2 shows the products obtained at different metal-to-ligand molar ratios such as 1:4, 1:2, 1:1 and 1:0.5, and further allowing the samples to cool down up to room temperature inside the furnace upon completion of the reaction time (24 h). In the stoichiometric ratio of 1:2 the pH was not controlled externally. The target compound is obtained for all molar ratio but the metal-to-ligand ratio of 1:0.5 favors the high purity and maximum growth of the crystals. The by-products obtained corresponding to each molar ratio are analyzed and summarized in Table 1. The pH reported are the pH measured immediately after mixing the reactants, without any additional steps of pH control. The corresponding optical images and powder x-ray diffraction patterns for each entry in Table 1 are shown in Figs. 2 and 3, respectively. For the molar ratio 1:4, we obtain our target compound Zn(3-ptz) 2 as well as the zinc tetra aquo complex Zn(H 2 O) 4 (3ptz) 2 (ZAC), which is obtained in relatively large quantities. The same products were obtained previously in Ref. (58) using half the amount of ligand. At the stoichiometric ratio 1:2, we obtain Zn(3-ptz) 2 along with very small quantities of non-reacting 3cyanopyridine. This result should be compared with Ref. (58), which at same ratio (1:2) co-crystallization of ZAC and Zn(3-ptz) 2 is obtained without pH control. The difference in results in spite of having same reaction time, temperature and pressure, can only be attributed to the use of zinc nitrate (this work) salt instead of zinc chloride (58). In general, we obtain Zn(3-ptz) 2 crystals with improved size and transparency with zinc nitrate than with zinc chloride. As we decrease the metal-to-ligand ratio below 1:2, we obtain increasing amounts of the recently reported energetic metal-organic framework Zn(3-ptz)N 3 (mu-N)2-azido-[5-(3-pyridyl)tetrazolato]zinc(II), a non-porous MOF with 8 azide groups per unit cell (61). This energetic MOF does not exhibit explosive behavior at the conditions used for its synthesis, filtration, and its spectroscopic characterization. For the molar ratio 1:1, we obtain Zn(3-ptz)N 3 along with our target compound, while decreasing the amount of ligand to the molar ratio 1:0.5 gives only Zn(3-ptz)N 3 in high yield and high purity, which represents an alternative synthesis route in comparison with Ref. (61). Optical images of the products listed in Table 1. The powder XRD patterns in Figure 3 show that our target MOF Zn(3-ptz) 2 is present for metal-to-ligand molar ratios 1:4, 1:2 and 1:1. As expected, panel 3(b) shows that the stoichiometric ratio 1:2 gives a Zn(3-ptz) 2 crystal phase with higher purity. Residual peaks correspond to unreacted 3-cyanopyridine. Increasing the relative abundance of the ligand relative to the stoichiometric ratio, results in the formation of the aquo complex Zn(H 2 O) 4 (3-ptz) 2 along with Zn(3-ptz) 2 . On the other hand, decreasing the relative abundance of the ligand relative to the stoichiometric ratio, enhances the growth of Zn(3-ptz)N 3. For molar ratio 1:1, Zn(3-ptz)N 3 is minority product. For molar ratio 1:0.5, the compound Zn(3-ptz)N 3 (stoichiometric ratio 1:1) is the main reaction product. For such a high molar ratio, there is no sufficient availability of 3-ptz ligands to form our target MOF. Figures 4 shows the SEM images for the samples listed in Table 1. Figure 4b clearly shows the octahedron morphology of the Zn(3-ptz) 2 crystals formed. Crystal sizes obtained for Zn(3-ptz) 2 in all experiments vary in the range 100-300 µm, which are typical sizes for reactions without pH control, all other reaction parameters the same as in Refs. (58,59). For the stoichiometric metal-to-ligand ratio, the size of Zn(3-ptz) 2 crystals is in general larger than those obtained at other molar ratios. This can be explained by the absence of competing nucleation events leading to the growth of by-product crystals, which also consume 3-ptz ligands. Figure 4a 1. We also measure the FTIR spectra of the products listed in Table 1. We observe the characteristic peaks for tetrazole near 1500 cm -1 in all samples. The broad OH band peaked at 3250 cm -1 is also present for the molar ratio 1:4, associated with the aquo complex Zn(H 2 O) 4 (3-ptz) 2 . Fig. 5a shows, that at the stoichiometric ratio 1:2, the FTIR spectrum exhibits a weak peak at 2073 cm -1 , which can be assigned to the -CN bond of unreacted 3-cyanopyridine. For the molar ratio 1:1, a strong and narrow peak is observed at 2070 cm -1 , which can be assigned to the azido group (-N 3 ) that forms the energetic MOF Zn(3ptz)N 3. The strong azido IR peak (-N 3 ) is also present for repetitions for the synthesis at molar ratio 1:0.5, for which Zn(3ptz)N 3 is the main product. ## Controlling the reaction pH at fixed molar ratio In addition to variation of molar ratio, we explore the effect of changing the pH of the solution at the stage of mixing the reactants whereas during the reaction in the furnace the pH is unmonitored. We vary the pH as described in Ref. (58) for two sets of experiments. One set for the molar ratio 1:2, which is the stoichiometric ratio of Zn(3-ptz) 2, and another set for the molar ratio 1:1, which is the stoichiometric ratio of Zn(3-ptz)N 3 . The main products obtained in each set of experiments are summarized in Tables 2 and 3. For the molar ratio 1:2 (Table 2), our results are qualitatively in agreement with those in Ref. (58). In general, increasing the mixing pH above the uncontrolled pH level (5.81) favors the formation of the two-dimensional MOF Zn(OH)(3ptz) (58,60), as the ions Zn(OH) + and Zn(OH) +2 become the most abundant species (>99%) in the pH range 6 -11, at 75ºC (62). Near pH 13 we obtain an almost exclusive formation of zinc oxide (ZnO), in agreement with Ref. (58). Powder XRD and FTIR spectra confirm these results as shown in the Supplementary Material (SM). We obtain roughly the same total crystal yield and reaction products in the case of uncontrolled pH (5.81) and at pH 4.67, controlled with nitric acid. The powder XRD spectra for both samples are indistinguishable (see SM), both confirming the presence of Zn(3-ptz) 2 and 3-cyanopyridine. The optical images in Fig. 6 show that at pH 4.67, the Zn(3-ptz) 2 octahedrons coexist with large 3-cyanopyridine molecular crystals, whose observed habit is consistent with simulations. The more intense -CN peak in the IR spectrum shown in Fig. 7a, confirms the greater abundance of unreacted 3-cyanopyridine at lower pH in comparison with pH 5.81. Unreacted 3-cyanopyridine can result from the protonation of azide ions (N 3 -) at low pH. As the azide ion captures a proton, its availability to participate in the cycloaddition reaction that forms 3-ptz ligands ( 63) is significantly reduced. ------------------- ZnO -------------------- 2, for metal-to-ligand molar ratio 1:2. Figure 8 shows the SEM images of the samples in Table 2. Panels a-d show the Zn(3-ptz) 2 MOF octahedrons obtained at the low pH, whose size vary in the range 100 -300 µm. For alkaline environments (pH 7.77 and 10.01), we obtain the large rod-like structures characteristic of MOF crystal Zn(OH)(3-ptz), which has a 2D coordination framework (58,60). At pH 10.01, we obtain large micro-rods with lengths in the range 300 µm -1.8 mm, which is 10 4 times higher than previously reported (58,60), for the same reaction times and temperatures. The rod widths vary in the range 2 -30 µm, which is up to ten times higher as previous reports. We attribute this improvement in size to the use of zinc nitrate instead of zinc chloride (58,60). In a second set of experiments, we fixed to 1:1 the metal-to-ligand molar ratio, which is the stoichiometric ratio for the energetic MOF Zn(3-ptz)N 3 (61), varying the mixing pH as specified in Table 3. Figure 9 shows the optical images of the products obtained at pH 3.90 and 5.49. The corresponding powder XRD patterns are given in Fig. 10a,b. Despite obtaining similar total crystal yields, and having generated our target MOF Zn(3-ptz) 2 for the two acidic conditions explored, the overall results are qualitatively different. While Fig. 10b shows that without active pH control (pH 5.49), a significant amount of Zn(3ptz)N 3 is formed in a mixed phase with Zn(3-ptz) 2 , at pH 3.90 there are no traces of the energetic compound. We understand this in terms of a lower availability of azide ions N 3 to form coordination bonds with zinc at lower pH. At the mixing temperature (25ºC), the relative abundance of azide ions in solution is less than 13% (pKa=4.72 ( 64)). On the other hand, for pH 5.49, the relative abundance of azide ions is about 85%. Figure 10c shows that at pH 7.96, we obtain predominantly Zn(OH)(3-ptz) and ZnO. Additional XRD data confirms (not shown) that as the pH is increased beyond 8.0, the growth of ZnO is favored over Zn(OH)(3-ptz). ## CONCLUSIONS We explored a set of hydrothermal synthesis parameters to optimize the production of the transparent noncentrosymmetric MOFs [Zn(3-ptz) 2 ] and [Zn(OH)(3-ptz)], which have been proposed for applications in nonlinear optical signal processing with phase matching. By varying the metal-to-ligand molar ratio and the mixing pH, we find that highquality [Zn(3-ptz) 2 ] can be obtained in large quantities at the molar ratios 1:2 and 1:1. For the molar ratio 1:2 and pH around 4.2 the [Zn(3-ptz) 2 ] crystals with improved transparency and purity is obtained. In alkaline environments (pH 7 -10), we obtain large rod-like [Zn(OH)(3-ptz)] crystals with lengths reaching up to the millimeter regime. This is a fourorder of magnitude improvement over previous reports. All samples were obtained using nitrate (NO 3 2-) as the zinc counter ion, which is found to give higher quality crystals in comparison with previously used zinc chloride salts. As an additional result, we found a simple alternative synthesis route for [Zn(3-ptz)N 3 ], a high-energy MOF that has been recently reported (61). The results obtained are promising for the development of MOF-based nonlinear optical devices. The high transparency and increasing size of the [Zn(3-ptz) 2 ] crystals obtained at lower mixing pH show the potential for further optimization of the reaction acidity. Reaching single crystal sizes in the millimeter regime would enable optical characterization such as Mueller polarimetry as well second-harmonic generation (SHG) with phased-matching. Moreover, the record lengths obtained for the [Zn(OH)(3-ptz)] crystals would facilitate the simple alignment of the rod-like structures on a substrate using mechanical stress. Such aligned quasi-1D crystals can be used as nonlinear waveguides for efficient generation of SHG signals using standard microscopy techniques. Nonlinear optical MOF waveguides would allow the development of alternative materials for scalable integrated quantum photonics (65). ## Methods The Zn-to-(3-ptz) molar ratio (metal-to-ligand) was varied by using 1 mmol, 2 mmol, 4 mmol and 8 mmol of Zn(NO) 3* 6H 2 O, keeping the amount of 3-Cyanopyridine and NaN 3 fixed to 4 mmol and 6 mmol, respectively. The corresponding molar ratios are 1:4, 1:2, 1:1 and 1:0.5. All reagents were mixed in a bottle with a screw cap, 6 ml of distilled water were added at room temperature. After measuring the pH of the solution, this is put to the oven for 24 hours at 105 °C. The samples were left for another 24 hours inside the oven for cooling to room temperature. After cooling, the obtained samples were washed with ethanol and left for another 24 hours to dessicate in vacuum. For the set of experiments at fixed molar ratios 1:2 and 1:1, the pH is fixed by using nitric acid at 70% (Sigma Aldrich) and KOH 18,8 [M] after mixing the reactants with distilled water at room temperature. The reaction procedure follows as described above. For the set of experiments with molar ratio 1:1, the samples are quickly filtered in hot with a kitasato and a vacuum pump, then washed with ethanol and vacuum-stored in the desiccator for 24 hours. X-ray diffraction analysis was done with a Shimazdu XRD 6000 diffractometer using Cu Kα (l = 1.5418 ) radiation. Microstructural characterization of the synthesized materials was done by scanning electron microscope Zeiss EVO MA10. pH was measured with pH meter 2700 OAKTON. FTIR measurements were taken with a JASCO FTIR-4600 Spectrophotometer equipped with an ATR PRO ONE.

chemsum

{"title": "Controlled growth of the non-centrosymmetric Zn(3-ptz) 2 and Zn(OH)(3-ptz) metal-organic frameworks", "journal": "ChemRxiv"}

synthesis_of_sulfoquinovose_and_sulfoquinovosyl_diacylglycerides,_and_a_fluorogenic_substrate_for_su

## Abstract: The sulfolipid sulfoquinovosyl diacylglycerol (SQDG) and its headgroup, the sulfosugar sulfoquinovose (SQ), are estimated to harbour up to half of all organosulfur in the biosphere.SQ is liberated from SQDG and related glycosides by the action of sulfoquinovosidases (SQases). We report a 10-step synthesis of SQDG that we apply to the preparation of saturated and unsaturated lipoforms. We also report an expeditious synthesis of SQ and ( 13 C6)SQ, and X-ray crystal structures of sodium and potassium salts of SQ. Finally, we report the synthesis of a fluorogenic SQase substrate, methylumbelliferyl α-D-sulfoquinovoside, and examination of its cleavage kinetics by two recombinant SQases. ## Introduction: The sulfosugar sulfoquinovose (SQ, 1) is produced by photosynthetic organisms at a rate of 10 billion tonnes annually (Figure 1). 1 Approximately half of all organosulfur in terrestrial biomass is estimated to reside within SQ and consequently SQ metabolism comprises a significant arm of the global biogeochemical sulfur cycle. SQ is rarely found as the free sugar and is mainly present as the head group of the plant and cyanobacterial sulfolipid sulfoquinovosyl diacylglycerol (SQDG, 2) (Figure 1). 2,3 SQ is liberated from SQDG (and its delipidated form sulfoquinovosyl glycerol, SQGro, 3) by the action of specialized glycoside hydrolases known as sulfoquinovosidases (Figure 1). 4,5 Fig. 1 Sulfoquinovosyl diacylglycerol (SQDG) and sulfoquinovosyl glycerol (SQGro) undergo sulfoquinovosidase (SQase) catalysed hydrolysis to sulfoquinovose (SQ) and the corresponding aglycon. The biosynthesis of SQDG has been well-characterized in plants, photosynthetic protists and cyanobacteria, and involves synthesis of the sugar nucleotide UDP-SQ from UDP-Glc and its transfer to diacylglycerol. 2 In the last five years two pathways for catabolism of SQ, a process known as sulfoglycolysis, 6 have been identified in bacteria. Both pathways involve the initial cleavage of SQ glycosides by SQases, which are typically encoded within sulfoglycolytic operons. The two pathways can be considered variants of classical glycolytic pathways and are termed the sulfoglycolytic Entner-Douderoff 7 and the sulfoglycolytic Embden-Meyerhof-Parnas 8 pathways. Bioinformatics analysis suggests that sulfoglycolysis plays important roles in sulfur-cycling in complex microbial ecosystems, including those in soils, the gut and the ocean. Our knowledge of the biochemical properties of enzymes involved in SQDG degradation and sulfoglycolysis remains in its infancy. Little is known about the broader metabolic consequences of sulfoglycolysis on cellular metabolism. 9 Such studies could benefit new methods for the preparation of SQDG and from large-scale methods for the synthesis of SQ and 13 C-labelled isotopologues. As the operons encoding the sulfo-ED and sulfo-EMP pathways typically encode an SQase, detection of the activity of this highly specific class of enzyme 4,5 can therefore be considered a marker for the possible presence of a sulfoglycolytic pathway. Consequently, sensitive substrates for SQases that can detect enzymatic activity in cell lysates could allow study of expression of sulfoglycolytic pathways and support the untargeted discovery of new sulfoglycolytic organisms. Here we disclose a new synthesis of SQDG, the shortest yet reported, enabling the synthesis of saturated and unsaturated lipoforms. We disclose an improved method for the synthesis of SQ and ( 13 C6)SQ, facilitated by the identification of a crystalline sodium salt. We also report the first single-crystal X-ray structures of SQ. Finally, as part of an effort to develop sensitive fluorogenic substrates for SQase, we report the synthesis and kinetic characterization of the methylumbelliferyl glycoside of SQ and its kinetic parameters for two recombinant SQases. ## Results and Discussion: 1. Synthesis of SQDG Three approaches have been reported for the synthesis of SQDG as single diastereoisomers but all are lengthy and involve multiple protecting group interchanges. Gigg and co-workers developed a seventeen-step route, 10 and two independent fourteen-step routes have been disclosed by Danishefsky 11 and by Hanashima and co-workers. 12 All three approaches involve the glycosylation of a glycerol moiety but owing to the nature of the protecting groups, only the approach of Danishefsky provides access to unsaturated lipoforms. Hanashima and coworkers subsequently developed an alternative approach involving dihydroxylation of an allyl α-D-glucoside; this method, while more direct, affords SQDG as a mixture of diastereomers epimeric at C2 of the glycerol moiety. 13 We have previously developed an efficient approach to synthesis of unsaturated α-glucosyl diglycerides through a non-stereoselective epoxidation of allyl α-D-glucoside followed by a Jacobsen hydrolytic kinetic resolution that delivers a 2'Rglyceryl α-D-glucoside. 14 We proposed to apply this approach to synthesize a naturally occurring lipoform of SQDG-(sn1/sn2), SQDG-(C18:1/C16:0) 5 4 as well as the simplified analogue SQDG-(C4:0/C16:0) 4. The first four steps proceeded in an identical fashion to that originally reported 14 and are reproduced here for completeness (Figure 2): (1) methoxyacetylation of allyl αglucopyranoside (→ 6); (2) oxidation with mCPBA to afford epoxide 7 as a mixture of 2'epimers; (3) Jacobsen hydrolytic kinetic resolution of 7 using S,S-C1.OTs to afford 2'R-epoxide 8 (and the 2'S-diol, not shown); and (4) ring-opening of the epoxide 8 to the bromohydrin 9 using Li2NiBr4. Towards the SQDG target, we next treated bromohydrin 9 with palmityl chloride afforded the bromo-ester 10. Nucleophilic substitution of the bromide of 10 using the tetrabutylammonium salts of butyric and oleic acids in toluene 14 gave the diglycerides 11 and 12, respectively. Deprotection using tBuNH2/MeOH 15 afforded the glucosyl diglycerides 13 and 14. A thioacetate was introduced at C6 using a Mitsunobu reaction with AcSH and DIAD/Ph3P to give the thioacetates 15 and 16. Oxidation of the thioacetate group of 15 was achieved using Oxone buffered with KOAc/AcOH. 16 This provided the target SQDG analogue, contaminated with a less polar material tentatively assigned as the 4-O-acetyl-SQDG derivative. Careful deacetylation was achieved using 0.04 mM hydrazine in EtOH/H2O to give SQDG-(C4:0/C16:0) 4. Oxidation of the thioacetate 16 proved more challenging. Oxidation with Oxone under the same conditions used for 15 resulted in epoxidation of the alkene in addition to oxidation to the sulfonate. After considerable experimentation we established an alternative protocol for oxidation of 16 using H2O2 in potassium phosphate buffer (pH 7). Again, contamination with the 4-O-acetyl SQDG was observed but was readily overcome by deacetylation with 0.04 mM hydrazine in EtOH/H2O to afford SQDG-(C18:1/C16:0) 5. ## Synthesis of SQ (1) Multiple publications have reported the synthesis of SQ. However, most reported methods can trace their roots to early work by either Ohle and Mertens 17 or Helferich and Ost. 18 Ohle and Mertens utilized 1,2-O-isopropylidene-α-D-glucofuranose as starting material and introduced divalent sulfur at C6, which was then oxidized to the sulfonate, followed by deprotection. Subsequent workers have modified this procedure to directly introduce the sulfonate moiety using a sulfite substitution reaction. 17,19,20 Helferich and Ost used methyl α-D-glucopyranoside as a substrate and introduced a sulfonate directly by substitution with sulfite. 18,21 In early work crystalline potassium 17,22 and sodium salts, 18 were identified. However, subsequent studies did not take advantage of these isolation methods; in some cases there was either no purification of the final SQ isolated, or it was purified through tedious methods involving interconversion of toxic salts. 22 We have explored an alternative approach to SQ that involved introduction of sulfur into methyl α-D-glucopyranoside as a thioacetate and then oxidation to the sulfonate. 23 This approach involved use of acetate protecting groups throughout most of the sequence to allow normal-phase chromatography of advanced intermediates, and of SQ sodium salt, an approach that we successfully applied to the synthesis of ( 13 C6)SQ. 23 However, this route is long, and purification of the final SQ product via chromatography proved challenging to scale up. Here we have revisited the method of Helferich and Ost, and report a streamlined synthesis of SQ that involves just two chromatographic steps and direct crystallization of sodium salt of SQ, which we also use for the synthesis of ( 13 C6)SQ. Our approach commenced with methyl α-D-glucopyranoside, which was converted to the acetylated iodide 17 using the method of Garegg and Samuelsson (Figure 3a). 24 Zemplen deacetylation using NaOMe/MeOH afforded the known iodide 18. 25 The sulfonate group was introduced by nucleophilic substitution using sodium sulfite to give the methyl glycoside 19. Hydrolysis using aqueous HCl afforded crude salt of SQ 1, which crystallized upon prolonged storage. With access to seed crystals we were able to develop a reliable crystallization of the sodium salt of SQ from MeOH/H2O on a multigram scale in 73% yield. Fig. 3 (a) Synthesis of sulfoquinovose (SQ) and (b) synthesis of ( 13 C6)-sulfoquinovose (( 13 C6)SQ). • = 13 C. This approach was applied to the synthesis of ( 13 C6)SQ. D-( 13 C6)glucose was converted to the acetylated iodide 20 as previously reported (Figure 3b). 23 Deacetylation using NaOMe/MeOH gave the iodide 21 and the sulfonate group was introduced by sulfite substitution to produce 22. Hydrolysis of 22 under acidic conditions, followed by crystallization from MeOH/H2O afforded ( 13 C6)SQ 23 as the sodium salt. This route consists of five chemical steps from D-( 13 C6)glucose and involves three chromatographic purifications, an improvement over our previous method. 23 The acquisition of several crystalline forms of SQ prompted us to determine their single crystal X-ray structure of this compound. SQ potassium salt (prepared by ion-exchange) crystallizes as a monohydrate that forms a co-crystal consisting of the α-and β-anomers in the ratio 0.70(2): 0.30(2) respectively, which occupy the same crystallographic site in the crystal, with one molecule in the asymmetric unit. The structure is shown in Figure 4 (left) and shows the major α-anomer form. Molecules in the crystal array are extensively crosslinked by coordination of the sulfonate, hemiacetal and hydroxyl oxygens to the potassium counterion. SQ sodium salt also crystallises as a monohydrate with two independent molecules in the asymmetric unit linked by coordination to one of the two independent sodium counterions. One of the molecules (defined by carbon atoms C1-C6) exists entirely as the α-anomer, while the second molecule (defined by carbon atoms C7-C12) exists as a mixture of α-and β-anomers in the ratio 0.34(1):0.66 (1), respectively, at the same crystallographic site in the crystal. The structure is presented in Figure 4 (right) and shows the major β-anomer at the variable site. As for the potassium salt, the sodium salt structure is extensively crosslinked by coordination of sulfonate, hemiacetal and hydroxyl oxygens to the sodium counterion. ## A fluorogenic substrate for SQases Aryl glycosides are widely used as substrates for glycosidases due to their potential to release coloured, chromophoric or fluorescent phenol/phenolate dyes. We have previously reported that PNPSQ is an effective substrate for SQases, allowing real-time measurement of enzyme rates in a UV/Vis spectrophotometer. 4,5 However, chromogenic substrates have limited utility in crude cell lysates where background absorption and low rates provide poor sensitivity. 4-Methylumbelliferyl glycosides are widely used as sensitive substrate for a range of glycosidases. For this reason we set out to synthesize methylumbelliferyl α-sulfoquinovoside (MUSQ) 24. Our approach sought to first prepare methylumbelliferyl α-glucoside (MUGlc) 25, then to introduce the sulfonate group at the 6-position. The synthesis of MUGlc has been reported in the peer-reviewed 26 and patent literature 27, 28 but all methods suffer from low yields. We report a new method for its preparation via the β-chloride 26. 29 Nucleophilic substitution of 26 using the tetrabutylammonium salt of methylumbelliferone in MeCN afforded the tetraacetate 27 in modest but reproducible yield of 15%. While low, this represents an improvement on the literature approach, which reports just 10% yield. 26 Deacetylation under Zemplen conditions gave 25. ## Fig. 5 Synthesis of methylumbelliferyl α-sulfoquinovoside (MUSQ). Introduction of the sulfonate into 25 required some exploration. Initially, treatment of MUGlc with CBr4/Ph3P under Appel conditions afforded the 6-bromide in good yield. However, attempted substitution with Na2SO3 in H2O led to cleavage of the glycoside. Instead, Mitsunobu reaction of 25 using BzSH and DIAD/Ph3P afforded the thiobenzoate, which was cleanly oxidized using H2O2 buffered with KOAc/AcOH. Chromatography, followed by recrystallization afforded the potassium salt of MUSQ 24. In order to explore the potential for use of MUSQ as a fluorogenic substrate for SQase, we measured the kinetic parameters for cleavage of MUSQ by two recombinant SQase, EcYihQ from Escherichia coli and the sulfoquinovosidase AtSQase (NCBI accession WP_035199431) from Agrobacterium tumefaciens. 5 We used a stopped assay involving incubation of varying concentrations of MUSQ with each SQase over a period in which rates were linear, and then quenching of the reaction with a high pH stop buffer, followed by measurement of fluorescence, which was quantified by reference to a calibration curve. Michaelis-Menten and Lineweaver-Burk plots are shown in Figure 6. In both cases limited saturation was observed up to 1 mM MUSQ preventing accurate determination of KM and Vmax values. However, kcat/KM values could be accurately determined and are shown in Table 1 versus the equivalent values for PNPSQ with the same enzymes. This data reveals that MUSQ is approximately 10 4 -10 5 -fold poorer as a substrate for these two SQases. This data is unexpected given that the pKa values of 4-nitrophenol (7.15) and 4-methylumbelliferone (7.79) are similar, and so should have similar nucleofugacity. Possibly, the large differences in kcat/KM values between PNPSQ and MUSQ arise from the geometry of the active site. The active site of SQases have evolved to accommodate the aglycon of the natural substrate SQDG, a slender glycerol/diacylglyceride, meaning that the active site may have difficulty accommodating the more sterically demanding bicyclic aromatic system of methylumbelliferone. This interesting observation is, however, disappointing as the low kcat/KM values means that MUSQ is unlikely to be suitable as a sensitive substrate for SQases. ## Conclusion As major species within the biogeochemical sulfur cycle, the biosynthesis and catabolism of SQDG and its headgroup SQ, are of considerable interest. SQDG can only be obtained from natural sources as mixtures of lipoforms, and consequently there is a need for streamlined methods for its synthesis. The route described herein is compatible for the preparation of both saturated and unsaturated lipoforms and at 10-steps is considerably shorter than previously reported methods. Practical access to large quantities of SQ and the ( 13 C6) isotopologue, facilitated by the identification of a readily crystallized sodium salt, will support future metabolic studies where this compound is used as a carbon source for microbial culture, as well as metabolomic studies of the effects of sulfoglycolysis on cellular metabolism. Finally, we report the first efforts to develop fluorogenic substrates of SQases. Disappointingly, MUSQ lacks the reactivity necessary to support its use in crude cell extracts, possibly as a result of the preference of SQases for elongated glycolipid substrates. Future design of fluorogenic SQase substrates will need to overcome the tendency of these enzymes to discriminate against bulky aglycons. ## Experimental General Pyridine was distilled over KOH before use. Dichloromethane and THF were dried over alumina according to the method of Pangborn et al. 30 Reactions were monitored using thin layer chromatography (TLC), performed with silica gel 60 F254. Detection was effected by charring in a mixture of 5% sulfuric acid in methanol, ceric ammonium molybdate, and/or with UV light. Flash chromatography was performed according to the method of Still et al. 31 ( ## 2'R)-1'-O-Butyl-2'-O-palmitoyl-glyceryl α-D-glucopyranoside (13) A solution of t-butylamine (0.154 mL, 1.48 mmol) and 11 (0.120 g, 0.141 mmol) in CHCl3 (0.32 mL) and MeOH (0.8 mL) was stirred at 0 °C for 10 min and then at 10 °C for 3 h at which time tlc indicated that the starting material was completely consumed. The solvents were evaporated under high vacuum without heating. Flash chromatography of the residue (pet. ether to EtOAc/pet. ether 50:50) afforded 13 as a white semisolid (0.071 g, 90% A solution of triphenylphosphine (0.072 g, 0.279 mmol) and diethyl azodicarboxylate (0.044 mL, 0.279 mmol) in THF (10 mL) was prepared and held at 0 °C under nitrogen for 30 min. This solution was then transferred by cannula to a solution of thioacetic acid (0.0063 mL, 0.09 mmol) and 13 (0.050 g, 0.089 mmol) in THF (10 mL) at 0 °C under nitrogen. ( ## 2'R)-1'-O-Oleoyl-2'-O-palmitoyl-glyceryl α-D-glucopyranoside (14) A solution of t-butylamine (0.154 mL, 1.48 mmol) and 12 (0.20 g, 0.082 mmol) in CHCl3 (0.32 mL) and MeOH (0.80 mL) was stirred at 0 °C for 10 min and then at 10 °C for 3 h at which time tlc indicated that the starting material was completely consumed. The solvent was evaporated under high vacuum without heating. Flash chromatography of the residue (CH2Cl2 to 5:95 MeOH/CH2Cl2) afforded 14 as a white syrup (0.060 g, 98% ( ## 2'R)-1'-O-oleoyl-2'-O-palmitoyl-glyceryl 6-S-acetyl-6-deoxy-6-thio-α-D-glucopyranoside (16) A solution of triphenylphosphine (0.048 g, 0.186 mmol) and diethyl azodicarboxylate (0.029 mL, 0.186 mmol) in THF (10 mL) was prepared and held at 0 °C under nitrogen for 30 min. This solution was then transferred by cannula into a solution of thioacetic acid (0.0039 mL, 0.056 mmol) and 14 (0.040 g, 0.053 mmol) in THF (5 mL) at 0 °C under nitrogen. The resultant reaction mixture was stirred at 4 °C under nitrogen overnight. The solvents were evaporated and flash chromatography of the residue (CH2Cl2 to 3:97 MeOH/CH2Cl2) afforded 16 as a white oil (0.027 g, 73% ## Sodium methyl 6-deoxy-6-sulfonato-α-D-glucopyranoside (19) A solution of 18 25 (3.04 g, 10.0 mmol) and Na2SO3 (5.06 g, 40.2 mmol) in H2O (90 mL) was stirred at 80 °C for 4 h, then evaporated to dryness under reduced pressure. MeOH (75 mL) was added to the residue and the mixture was stirred at rt for 16 h. The filtrate was evaporated to dryness under reduced pressure and the residue purified by flash chromatography (EtOAc/MeOH/H2O, 9:2:1 → 3:2:1) to afford 19 as a colourless solid (2.72 g, 97% A solution of methyl 2,3,4-tri-O-acetyl-6-deoxy-6-iodo-D-( 13 C6)glucopyranoside 20 23 ## Kinetic analysis of sulfoquinovosidases using MUSQ AtSQase and YihQ enzymes were obtained as previously described. 4,5 Reaction buffer contains 50 mM NaH2PO4/Na2HPO4 and 150 mM of NaCl at pH 7.4. Stop buffer contains 1 M NaOH and 1 M glycine at pH 10. ## Calibration curve for 4-methylumbelliferone A calibration curve for the 4-methylumbelliferone fluorophore was constructed as follows. Solutions of 4-methylumbelliferone were prepared at 2. performed to demonstrate that reaction rates were linear over the chosen time period to ensure measurement of initial rates. ## X-ray crystallography Crystals of SQ sodium and potassium salts were mounted in low temperature oil then flash cooled using an Oxford low temperature device. Intensity data were collected at 100 K on the MX2 beamline at the Australian Synchrotron. 33 The structures were solved by direct methods and difference Fourier synthesis using the SHELX suite of programs 34 as implemented within the WINGX 35 software. Thermal ellipsoid plots were generated using the program ORTEP-3. ## Conflicts of interest There are no conflicts to declare.

chemsum

{"title": "Synthesis of sulfoquinovose and sulfoquinovosyl diacylglycerides, and a fluorogenic substrate for sulfoquinovosidases", "journal": "ChemRxiv"}

the_chirality_origin_of_retinal-carotenoid_complex_in_gloeobacter_rhodopsin:_a_temperature-dependent

## Abstract: Retinal proteins play significant roles in light-induced protons/ions transport across the cell membrane. A recent studied retinal protein, gloeobacter rhodopsin (gR), functions as a proton pump, and binds the carotenoid salinixanthin (sal) in addition to the retinal chromophore. We have studied the interactions between the two chromophores as reflected in the circular dichroism (CD) spectrum of gR complex. gR exhibits a weak CD spectrum but following binding of sal, it exhibits a significant enhancement of the CD bands. To examine the CD origin, we have substituted the retinal chromophore of gR by synthetic retinal analogues, and have concluded that the CD bands originated from excitonic interaction between sal and the retinal chromophore as well as the sal chirality induced by binding to the protein. Temperature increase significantly affected the CD spectra, due to vanishing of excitonic coupling. A similar phenomenon of excitonic interaction lose between chromophores was recently reported for a photosynthetic pigment-protein complex (Nature Commmun, 9, 2018, 99). We propose that the excitonic interaction in gR is weaker due to protein conformational alterations. The excitonic interaction is further diminished following reduction of the retinal protonated Schiff base double bond. Furthermore, the intact structure of the retinal ring is necessary for obtaining the excitonic interaction. AbbreviationsgR Gloeobacter rhodopsin Apo-gR/apo Apo-protein of gR Sal Salinixanthin CD Circular Dichroism xR Xanthorhodopsin bR Bacteriorhodopsin wt-bR Wild type bacteriorhodopsin DDM N-Dodecyl β-d-maltoside SDS Sodium dodecyl sulphate EC Excitonic coupling CE Cotton effect NaBH 4 Sodium borohydride PBS Protonated Schiff base Gloeobacter rhodopsin (gR) is a recently studied retinal protein which was found in thylakoid-less unicellular cyanobacterium Gloeobacter violaceus Pcc 7421 1,2 , heterogeneously expressed in E. coli. Similarly to other retinal proteins, it transfers protons from the cytoplasmic region to the extracellular region and acts as a light-activated proton pump 3,4 . Limited studies were carried out related to gR structure, function and spectral properties 1, 3,5-14 . gR is a type I rhodopsin in which the retinal chromophore is bound to Lys-257 amino acid residue of the seventh transmembrane G-helix by a protonated Schiff base formation (Fig. 1a-c) 10,15 . It was revealed that the gR protein is capable of binding salinaxanthin carotenoid (sal, Fig. 1d) which transfers at least 40% of its absorbed energy to the retinal chromophore following light absorption 6 . gR has 50% residue identity and 42% sequence similarity to xanthorhodopsin (xR) 1 . CD measurements have been widely used as a sensitive tool to investigate protein conformational changes, and to analyse chromophore-chromophore and chromophore-protein interactions in various proteins 16,17 . The CD spectrum of native xR, which contains two chromophores (retinal and sal), exhibits sharp positive bands at 513, 478, and 455 nm, and a negative band at 530 nm, which formed a biphasic shaped spectrum 18,19 . These CD bands arise only when the carotenoid sal is bound to the protein and forms the protein-carotenoid complex. Sal itself as an isolated chromophore has very weak CD bands in the visible region 20 . The CD bands of the complex vanished when the protein is denatured at high pH or the retinal protonated Schiff base linkage is cleaved following a reaction with hydroxylamine 19 . Different possibilities were proposed to explain the origin of the CD spectrum of xR. Balashov et al. proposed that mainly two factors contribute to the CD spectrum of xR. One is the excitonic coupling between the carotenoid and the retinal, and another is the induced chirality of the chromophores due to their binding to the protein 18,21 . CD excitonic coupling (EC) is originated from light excitation of two non-conjugated but energetically close chromophores, which their excited states are coupled, causing an energy splitting of the excited states into two different energy states. The transition to these states with different rotational strength generates a bisignate CD spectra (with opposite sign CD lobes) and observed Cotton effect (CE) 22 . We have proposed that the CD spectrum of xR originated from an excitonic coupling interaction between two sal molecules located in subunits of XR 19,23 , and from a retinal band. The CD of native bR is composed from bisignate curves of opposite signs at 593 nm, and at 531 nm. The cross-point is located at 560 nm close to the corresponding UV-Vis absorption maxima (557 nm). The biphasic CD spectrum of bR has been interpreted as a superposition of an intrinsic positive CD arising from the protein environment and the exciton coupling of the retinal chromophores within the bR trimer . Alternatively, it was proposed that it is arising only due to superposition of the intrinsic CD spectra . Although the CD spectra of native bR, and xR-sal complex were widely studied, the exact origin of the CD spectra is not completely clear 18,19,21,23,24, . The CD spectra of native gR as well as in the presence of different types of carotenoids were studied previously, and it was proposed that the CD of the gR-carotenoid complex is due to induced chirality of the carotenoid 10,11 . pH-dependent oligomer-monomer transition, and their CD spectra were reported by Demura et al. 9 Recently it was reported that the pentameric form of the pigment predominant at pH 8 while the monomeric form at pH 3 15,33 . However, in presence of sal, the exact origin of the observed positive and negative CD lobes with the enhancement of intensity and reversal of the CD band sign compared to the native gR, is not clear. Besides, it is also unclear how the CD spectral nature of gR and its complex with sal is affected by temperature and pH of the medium 9 . In this work, we have examined the origin of the gR chirality, and the role that excitonic coupling and temperature play in affecting the CD spectra of gR, and its complex with sal. We propose that the CD spectrum of the gR-sal complex can be attributed mainly to the excitonic interaction between sal and the retinal chromophore. Temperature elevation affects the CD spectral shape and abolishes the excitonic coupling between the retinal and sal. ## Materials and methods Protein preparation, Sal extraction and purification. gR was grown and purified as described elsewhere 3,6,12 . Sal was extracted from cell membranes of Salinibacter ruber 23,34 . Following the addition of 10 mg/ OD sodium dodecyl sulphate (SDS), the mixture was lyophilized overnight. Acetone was added and vortexed for five minutes to extract the sal chromophore, which was purified using a silica gel column, and eluted with 25:75 acetone: n-hexane and, finally, with pure acetone. Ethanol solutions of all-trans retinal, synthetic retinal analogues, and sal were incubated with Apo-gR to carry out the binding process. To avoid protein denaturation, the ethanol volume was kept less than 2% of the protein solution volume 6 . To maintain the required pH of Apo-gR samples, we used a 50 mM citrate/phosphate/Tris buffer in 0.06% DDM (n-dodecyl β-d-maltoside). Preparation of Apo-gR and artificial pigments. Apo-gR was prepared by incubating gR with 0.5 M freshly prepared hydroxylamine (pH 7.5) and irradiation for 20 min with a Schott 250 W cold light source (Carl Zeiss Microscopy, Jena, Germany) equipped with a heat-absorbing filter and an optic fibre (level 4B). The light was filtered through a long pass cut off filter (Schott, Mainz, Germany) with λ > 520 nm. The bleaching process was monitored by UV-Vis absorption spectral measurements. Next, the sample was filtered through a membrane filter (centricon of 10 000 MW) by centrifugation and washed with 0.02% DDM for 5-6 times to remove the retinal oxime and unreacted hydroxylamine. The sample was then stored at 4 °C to avoid reconstitution with all-trans retinal originating from residual retinal oxime 12 . The retinal analogues were synthesized as previously described 23, . Artificial pigments of gR were prepared by overnight incubation with Apo-gR (absorbance of 0.1-0.2 OD) with two equivalents of the synthetic retinal analogues (1.5 equivalents for all-trans retinal) in presence of 50 mM citrate/phosphate/Tris buffer (according to the required pH), and 300 mM NaCl in 0.06% DDM at 25 °C. ## CD spectral measurement. A Chirascan CD spectrometer (Applied Photophysics) was used for CD spectral measurements. All the CD spectra are given here in ellipticity θ in millidegree. A quartz cuvette of 10 mm path length was used for the measurements. The CD spectra were recorded with a 1 nm bandwidth resolution and in 1 nm steps at different temperatures using Peltier temperature controller. For temperature variation measurement, the sample was allowed to thermal equilibrate for 10 min at each temperature. The CD spectra were measured in the dark state using a 150 W air-cooled Xe arc lamp as light source and baselines were corrected by subtracting reference spectra of the corresponding buffer solution. We have checked that there are no photoexcitation during the CD measurement under dark state for the gR and its artificial pigments. During processing the CD spectra after internal reference correction (690-700 nm wavelength region was taken as a reference where no CD intensity was detected, except for 13-CF 3 retinal analogue (Table 1) where we used 745-750 nm range), each spectrum was smoothed by 5 points using the adjacent averaging method. During the formation of the gR and artificial gR pigments in presence of sal, the CD spectra were recorded at 5 min interval for the first one hour followed by one-hour interval for the remainder of the experiment (i.e., overnight). ## UV-Vis spectral measurement. The UV-Vis spectral measurements were done using an Agilent 8453 diode-array spectrophotometer (Agilent Technologies, Palo Alto, CA) equipped with an Agilent 89090A thermostated cuvette holder in a 10 mm quartz cuvette at 25 °C in the dark state. The spectra were recorded after proper background correction with the corresponding buffer solutions using Tungsten/Deuterium lamp as the light source. The processed absorption and difference spectra were generated after internal reference correction (780-800 nm wavelength region was taken as a reference where no absorption was detected). ## Results We have studied the CD and corresponding absorption spectroscopies of gR and its artificial pigments derived from synthetic retinal analogues following sal binding, to shed light on the origin of the CD spectra in the visible region, and its enhancement by sal binding. The binding of all trans retinal to apo-gR was studied previousely, and indicated high binding efficiency (ca. 85%) 12 . Similarly, the binding of retinal analogues to the apo-gR was studied. The binding of sal and other carotenoids to gR and CD spectra were studied by Balashov et al. 10,11 and also indicated high efficiency of binding. We studied the effect of temperature on the CD Spectra. The CD bands and UV-Vis absorption maxima of gR pigments and its artificial pigments in their dark-adapted state are summarized in Table 1. pH and temperature effects on the CD spectra of gR. Figure 2a shows the CD spectra of gR at different pH values at 25 °C. At pH 3, gR has very low-intensity CD bands at 540 (+) and 330 nm (−), whereas, at pH 5, the CD spectrum is prominent with bands at 566 (+), 500 (−) and 321 nm (−). The intensity of the CD spectrum is further enhanced at pH 8 9 . Absorption spectra of gR at similar conditions 26 (Fig. S1a,b), show pigment bands maxima at 546, 542 and 540 nm for pH 3, 5 and 8, respectively. Figure 2b shows the CD spectra of gR at pH 3, 5 and 8 at 45 °C temperature (gR CD spectral reversibility was observed until this temperature). Figure S1c,d represents the recorded absorption spectra at the same time and same condition as applied for CD measurement. Gradual temperature increase from 25 to 45 °C, indicates a shift of the CD band position from 566 to 547 nm at pH 5 (Fig. 2c), whereas, a slight change of CD intensity without any change of spectral band position at pH 3 and 8 was observed (Fig. 2d,e). Corresponding absorption and difference absorption spectra are presented in Fig. S2a-d. CD spectra of gR following binding of sal. Figure 3a represents the CD spectra for reconstitution of gR with the carotenoid sal at 25 °C. gR itself has very weak CD spectrum at pH 5 (Fig. 2c) with bands at 566 (+), 500 (−) and 321 nm (−). Following binding of sal, negative CD bands were observed at 538, 367 nm and positive CD bands at 514, 481, 456 nm. The positive part of the CD spectrum has a similar spectral shape to that of sal absorption spectrum except for a small blue shift of the band positions (Table 1). The complex formation and the formation of the CD spectrum were almost completed within 1 h 12 . Fig. 3b represents the difference CD spectra obtained by subtraction of the spectra recorded after 5 min of sal addition from the other spectra. To check whether a similar complex is formed by reconstitution of Apo-gR and sal mixture with all-trans retinal (1), we monitored this process with CD spectroscopy at pH 5. As shown in Fig. 3c, the Apo-gR has almost no CD bands in the visible range, whereas, sal itself has a very weak CD spectrum in the presence of Apo-gR. Fol- www.nature.com/scientificreports/ lowing reconstitution with the retinal chromophore, the CD spectrum of the complex was strongly enhanced, and yielded a negative band at 538 nm and a positive band at 481 nm. The difference spectra (Fig. 3d) show that the positive and negative bands intensities are similar. We have also monitored the absorption spectra recorded at similar conditions (Fig. S3a) 12 . We note that the CD spectrum of sal did not change after addition of the sal to the Apo-gR, (data not shown). We have further studied the temperature effect on the CD spectra of the gR-sal complex (Fig. 3e). Increasing the temperature was accompanied by a decrease of the CD bands intensities with a slight change of the positive band position, as demonstrated by the difference spectra (Fig. 3f). The difference spectrum is not identical to the original spectrum indicating that the original spectrum is composed of at least two components, in which only one is temperature dependent. We have calculated the temperature dependent component of the CD spectrum of gR-sal complex from the CD intensity at 538 nm due to intensity change by increasing the temperature from 25 to 60 °C which is approximately 65%. The corresponding absorption and difference absorption spectral changes are shown in Fig. S3b,c. It clearly indicates that sal experiences a conformational change, which is reflected in disappearance of a part of the vibrational fine structure at 486 and 518 nm absorption spectrum. Following cooling the system to 25 °C the CD spectrum regained its original bands, which indicates that the thermal process is reversible (Fig. S3d). ## CD spectra of artificial gR pigments in presence of sal. To further study the origin of the CD spectrum of gR, we have used artificial gR pigments derived from synthetic retinal analogues in which the native retinal ring and side-chain were modified. Consequently, the pigments absorption maxima were shifted, and the effect of these shifts on the CD spectrum was evaluated. First, we have incubated 14-fluoro (2) and 13-CF 3 (3) retinal analogues with Apo-gR in the presence of sal to form the two artificial pigments (Table 1). As shown in Fig. 4a,b (difference spectra), the negative CD band of the artificial pigment derived from 14-fluoro retinal is at 548 nm which is 10 nm redshifted relative to the negative CD band of the gR-sal complex. The band is also wider than the CD band of the gR-sal complex, and the amplitude of the positive and the negative bands are not equal. To compare the spectral shape and band position of the negative band, the reconstituted final spectra of gR-sal complex and other synthetic retinal analogues are shown in Fig. S4a. For a better comparison of the redshifted negative band position, we have normalized the bands by dividing the spectra by the maximum negative amplitude (Fig. S4b). The artificial pigment derived from 13-CF 3 retinal analogue (3) and its sal complex, has a wide negative CD band at 580 nm (Fig. 4c,d). The corresponding absorption spectra are shown in Fig. S5a-b. A temperature increase of the two gR artificial pigments induces a decrease of the CD bands intensities (Fig. S5c-f). The negative band almost disappeared at 55 °C but the positive band still has a significant intensity (Fig. S5c,e). We have studied the CD spectra of artificial pigments derived from synthetic retinal analogues modified at the retinal ring. The retinal analogues included retinals with one additional ring double bond (4), without any ring double bond (5), without the retinal ring (6), and with aromatic core substituting the retinal ring ( 7) www.nature.com/scientificreports/ (Table 1). The CD and its difference spectra of retinal analogue 4 (Table 1) are shown in Fig. 5a,b. The 543 nm negative band is slightly redshifted (5 nm), but the other bands remain almost at the same positions compared to gR-sal complex. In addition, the negative CD band is wider than gR-sal complex (Fig. S4b). Here, the amplitude ratio between the positive and negative CD bands is 1.6. Following a gradual increase in temperature from 25 to 60 °C, both bands lost intensity at 60 °C temperature (Fig. 5c,d). The negative band at 543 nm almost disappeared whereas the positive band still had substantial intensity. This observation indicates that the original CD spectrum consists of at least two components. The absorption spectra obtained in similar conditions (Fig. S6a,b), clearly indicate almost no change of the original pigments at 60 °C temperature. The CD spectra of the artificial pigment derived from retinal analogue 5, which lacks the ring double bond (Table 1), show almost no negative CD band (Fig. 5e,f). A temperature increase led to an intensity decrease of the positive band and similar to gR-sal complex still the positive band maintains a significant intensity (Fig. S6c,d). The reversibility of the CD spectra was detected following increasing and decreasing temperature. The artificial pigment derived from retinal analogue 6 lacks the retinal ring but has an additional double bond instead of the retinal ring (Table 1). As shown in Fig. 6a,b a narrow negative CD band at 534 nm was observed accompanied by positive bands at 510, 478, 454 nm. Comparison with the CD spectra of gR-sal complex (Fig. S4b), indicates that the bands are slightly blue-shifted relative to the gR-sal complex. A temperature increase indicated similar results to that detected in the native gR-sal complex. We note that the negative CD band completely disappeared at 60 °C temperature (Fig. 6c,d), unlike the gR-sal complex. Next, we used the aromatic substituted retinal analogue 7, where the retinal ring is substituted by a N,N dimethyl amine aromatic ring (Table 1). The pigment derived from retinal analogue 7, absorbs at 594 nm in the presence of sal 12 . The CD spectra of the artificial pigment 7 (Fig. 6e,f), exhibit a relatively weak broad CD negative band at around 565 nm and positive bands at 517, 486, 459 nm. The temperature effect on the CD spectra (Fig. S7a,b), showed similar results to the other artificial pigments. ## CD and absorption spectra following reduction with NaBH 4 . We reduced the protonated Schiff base double bond, which links the retinal chromophore to Lys-257 using NaBH 4 to evaluate its effect on the CD spectra. As shown in Fig. 7a,b; the positive and negative CD bands of gR-sal complex almost vanished with a residual CD of sal which is similar to that observed in the presence of Apo-gR (Fig. S4a (for sal + Apo-gR)). The disappearance of the positive band is in contrast to the temperature effect on the gR-Sal complex CD spectrum in which the positive band maintained significant intensity even though the negative band almost disappeared. www.nature.com/scientificreports/ We also monitored the absorption spectra recorded during CD spectral measurements. The absorption and difference spectra (Fig. S7c,d), clearly represent the change mainly at the pigment band position with the loss of sal fine structure bands. The reduction experiment was performed as well with the artificial pigments derived from the synthetic retinal analogues. Reduction of the gR-4-sal complex led to complete disappearance of the negative CD band while a small intensity of the positive band remained probably due to the sal chromophore (Fig. 7c,d). The absorption and its difference spectra (Fig. S7e,f) show that except intensity decrease of the pigment absorption at 585 nm, sal bands decreased as well during the reduction process. The 13-CF 3 retinal analogue (3) shows a complete loss of negative CD band and partial decrease of the positive CD band (Fig. S8a,b). The absorption and difference absorption spectra (Fig. S8c,d) clearly show the loss of the redshifted pigment band at 630 nm. ## Discussion The origin of gR CD spectrum and its pH and temperature dependence. The gR pigment contains only one chromophore (retinal) attached to Lys-257 by a protonated Schiff base double bond 1,8-12 . Therefore, the gR CD spectrum bands in the 300-700 nm region are attributed to the retinal induced chirality, and/or to excitonic interaction between retinal chromophores located in different proteins which are arranged in pentamers as previously reported 9,13,39 . At pH 5 and 8, it is evident that there are positive and negative CD bands at 566 and 500 nm, respectively, with Δε = 0 at 534 nm (Fig. 2a), which is close to the absorption maximum (540 nm) of wt-gR pigment 7,9,11 . Similar phenomenon was also reported for bR which has a negative band at 600 nm and positive band at 534 nm with Δε = 0 at 574 nm comparable with the wt-bR pigment absorption maximum at 567 nm 26 . The gR CD spectrum has opposite sign CD bands than bR and the first positive band is at 566 nm followed by a negative band at 500 nm 9 . Previously studied size-dependent chromatography proposed that at pH 3, gR adopts a monomeric form, whereas, at pH 8 only oligomers are present probably in the form www.nature.com/scientificreports/ of pentamers 9,15,33 . Our results clearly show that at pH 3, gR has a weak CD spectrum without characteristic positive and negative CD lobes at 25 °C (Fig. 2a) in keeping with monomer formation and lack of excitonic interaction between the retinal chromophores (Fig. 8a). The CD spectrum of gR at pH 5 was altered upon raising the temperature above 45 °C, and it resembled the CD spectrum at pH 3 and 25 °C. It may be explained by pKa alteration of the His87-Asp121 pair due to temperature increase. This pKa change can lead to pentamer to monomer transition probably by breaking the salt bridge, thereby effecting the CD spectrum due to vanishing of the excitonic interaction between the retinal chromophores (Fig. 8a,b) 9 . Therefore, the 547 nm CD spectral band of gR at pH 5 at 45 °C was similar to that observed at pH 3 at 25 °C (Fig. 2b). The CD spectrum was not altered at 45 °C at pH 8 even though the pKa of the His87-Asp121 pair was altered, still the protonation state of this pair was not changed since the pKa is still well below 8. Still, the possibility that the CD spectrum is altered following increasing the temperature to 45 °C at pH 5 due to protein conformation changes and not due to pKa change of the His87-Asp121 pair, cannot be completely excluded, but it is unlikely since at pH 8 the CD spectrum does not change at 45 °C. chiral origin of the gR-sal complex and temperature-dependent excitonic coupling. The gRsal complex exhibits much stronger CD spectrum than that of gR or sal in the presence of Apo-gR (Fig. 3a,c). The CD spectrum which is composed of negative and positive lobes, can be accounted for by several possibilities. The first possibility is based on excitonic interaction between sal and the retinal chromophore 21 . The second stems from the chiral conformation of the retinal in the presence of sal, while retinal-retinal excitonic coupling in a pigment pentamer form can be an additional cause for the CD spectrum 15,24,26,27 . Another possibility is based on a chiral conformation of sal in the presence of the retinal chromophore and nearby protein residues 19 . An excitonic coupling between two sal in the pentameric form cannot be excluded as well. A recent study suggested that excitonic interaction among chromophores in a photosynthetic protein strongly depends on temperature, and it is inversely affected by the temperature 39 . The present studies indicated that increasing the temperature of the gR-sal complex from 25 to 60 °C led to equally intensity decrease of the positive and the negative CD bands (Fig. 3f), which is completely reversible. We propose that these bands originated from excitonic interaction between the retinal and the sal chromophores. The excitonic interaction is lost at 60 °C possibly due to alteration of protein conformation, which may trigger sal and retinal conformational changes and change of the regular www.nature.com/scientificreports/ spatial organization, thereby eliminating the excitonic interaction. The intensity of both the negative and positive CD bands are reduced by warming to 60 °C, but still, a significant intensity of the positive band remains with a somewhat different shape relative to the original spectrum (Fig. 8c,d). We propose that the positive band is composed of an excitonic component of the interaction of sal with the retinal, as well as of a second sal component, which originates from the chirality of sal due to its specific chiral conformation in the protein, and/or the induced chirality by the protein. The absorption spectrum of the gR-Sal complex remains unchanged following temperature increase, which suggests that the complex is stable at 60 0 C, and supports the suggestion that the CD spectrum is altered at 60 0 C due to protein conformational changes and not protonation of the His87-Asp121 pair which is likely to affect the absorption maximum. Reduction of the PSB , and disappearance of the induced CD (Fig. 7a,b), while detecting a residual CD band of sal, is in keeping with retinal-sal excitonic interaction in the retinal-sal complex (Fig. 8e). The presence of retinal-sal excitonic interaction may gain further support from the blue shift of the sal CD bands compared to the absorption band of the gR-sal complex (from 518, 486 and 460 to 514, 481, and 456 nm, respectively; Table 1). The blue shifts were observed as well in the CD spectra of the artificial pigments-sal complexes. Further support that the negative CD lobe is associated with the excitonic interaction between the retinal and the sal and not due to sal band, can be derived from the CD spectra of the artificial pigments. Substitution of the retinal by 14-fluoro retinal analogue (2) redshifts the negative band of the CD spectrum with decreasing intensity compared to gR-sal complex. Whereas in the case of 13-CF 3 retinal analogue (3), the negative CD band is redshifted as well, with a significant decrease of intensity. The effect is probably due to the bulky group, which may affect the retinal conformation, thereby significantly decreases the excitonic interaction with the sal chromophore. Therefore, the redshifting, as well as the intensity decrease of the negative CD band, support as well the proposal that the negative band is associated with the retinal-sal excitonic interaction. The absorption maximum of the artificial pigment derived from retinal analogue 5 is significantly blueshifted (485 nm), and therefore the negative band of the CD spectrum which should be associated with the retinal analogue is possibly masked and neutralized by the opposite sign intense sal positive band (Fig. 5e,f). The protein-retinal ring interactions are important for imposing the appropriate retinal and sal conformation for efficient excitonic interaction. It can be concluded from the CD spectra of the artificial pigments derived from chromophores 6 and 7. The linear chromophore 6 lacks part of the retinal ring while chromophore 7 has an aromatic core, which drastically alters the ring structure. While the CD spectrum of the artificial pigment derived from retinal analogue 6 still exhibits a negative band (somewhat weaker relative to native gR), the CD spectrum of the pigment derived from chromophore 7 lacks the negative band almost completely. It was proposed that sal ring carbonyl group is important for binding of sal to gR 10,11 . Our present studies indicate that retinal ring structure controls the specific retinal-sal interactions thereby affecting the excitonic interaction of the two chromophores. The CD spectra of the artificial pigments are affected less by the temperature increase relative to the native gR, since the excitonic interaction is weaker. Still, the positive band of sal is maintained at high temperature, indicating that the sal gains its chirality also in the artificial pigments and it is affected significantly by modifying the retinal structure. Reduction of the PSB of the gR-sal complex with NaBH 4 decreases much of the positive band and all the negative band CD intensities (Fig. 7a). Since the negative band is associated with the retinal chromophore it completely disappeared (Fig. 8e). The retinal absorption is significantly blueshifted to 357 nm, and therefore, the excitonic interaction between the two chromophores is abolished. Since the positive band lost significant intensity (beyond the excitonic interaction) we propose that the reduction of the Schiff base linkage affected the retinal conformation and thereby the conformation of the sal as well (Fig. 8e). We note that the absorption of the sal chromophore did not change following the protonated Schiff base reduction, which indicates that the sal conformation did not drastically change (Fig. S7c). The present results indicate that the possibility of the CD origin from excitonic interaction between two sal in a protein pentameric form is very unlikely. The shift of the CD negative band in the artificial pigments and its disappearance following the PSB reduction strongly indicates that the negative band is associated with the retinal chromophore (Figs 4d, 7c and S4b). We recently studied the origin of the CD spectra in xR 19 , but it appears that the situation in gR is completely different. We have proposed that the CD spectrum of xR mainly originated from the exciton interaction between two salalinixanthin chromophores located in different subunits. In addition, two contributions stem from the chiral conformation of the salalinixanthin within its binding site probably due the fixation of the 4-keto ring in a specific twisted conformation, and the contribution of the retinal chromophore to a negative lobe located at 550 nm. xR lacks a major excitonic interaction between the retinal and salalinixanthin probably due to unfavourable angle between the planes of retinal and salalinixanthin. The situation in gR is different and the CD of gR-sal complex has a www.nature.com/scientificreports/ component originated from excitonic interaction between retinal and salinixanthin which is possibly due to the favourable angle between the planes of retinal and salinixanthin. Unlike xR, gR-sal complex lacks excitonic interaction between the sal chromophores located at different gR monomers in the aggregated cluster possibly due to unfavoarble spatial arrangement required for such excitonic interaction. We have quantitatively estimated the components of the CD spectrum of gR-sal complex from the intensity of the CD spectra at 538 nm (CD intensity of gR-sal complex at 25 and 60 °C, Fig. 8c-e), which indicates that retinal-sal exciton contributes approximately 65% and remaining 35% from the sal chirality acquired by sal following binding to gR. In summary, it can be concluded that the CD spectrum of gR originated from the excitonic coupling between retinal chromophores located in a protein pentameric form (Fig. 8b) 15,33 . Due to change of temperature, the His87-Asp121 pair arrangement is altered leading to pentamer to monomer transition and the corresponding change of CD spectra at pH 5 and vanishing of CD exciton 9 . In the presence of sal, the gR CD spectrum gains significant intensity. The spectrum is composed of two main contributions: (1) The chirality that sal acquires following retinal binding either by its specific conformation that may be associated with sal twisted ring-chain conformation, and/or induced chirality by the protein environment. (2) Excitonic coupling between sal and the retinal chromophore. Increasing the temperature to 60 °C eliminates the excitonic coupling, which may be due to conformational alteration.

chemsum

{"title": "The chirality origin of retinal-carotenoid complex in gloeobacter rhodopsin: a temperature-dependent excitonic coupling", "journal": "Scientific Reports - Nature"}

mechanism_of_tubulin_oligomers_and_single-rings_disassembly_catastrophe

## Abstract: Cold tubulin dimers coexist with tubulin oligomers and single-rings. These structures are involved in microtubule assembly, however, their dynamics are poorly understood. Using state-of-the-art solution synchrotron time-resolved small-angle X-ray scattering we discovered a disassembly catastrophe (half-life of about 0.1 sec) of tubulin rings and oligomers upon dilution or addition of guanosine triphosphate. A slower disassembly (half-life of about 38 sec) was observed following a temperature increase. Our analysis showed that the assembly and disassembly processes were consistent with an isodesmic mechanism, involving a sequence of reversible reactions at which dimers were rapidly added/removed one at a time, terminated by a two orders-of-magnitude slower ring-closing/opening step. We revealed how assembly conditions varied the mass fraction of tubulin in each of the coexisting structures, the rate constants, and the standard Helmholtz free energies for closing a ring and for longitudinal dimer-dimer associations. ## Introduction Microtubules (MTs) are filamentous protein nanotubes (25 nm in diameter) with walls comprised of assembled protofilaments, built from tubulin dimers. MTs are involved in vital cellular processes including cell division and intracellular trafficking. A growing MT can abruptly shrink even when there is plenty of free tubulin in the solution, in a process called catastrophe. 1 Moreover, while some MTs may completely disassemble, others may continue to grow. The activity and stability of MT, its dynamic assembly and disassembly processes, mostly depend on whether guanosine triphosphate (GTP) or guanosine diphosphate (GDP) molecules are bound to the tubulin dimers, and are facilitated by the GTPase activity of tubulin. GTP binds tubulin in two sites, a non-exchangeable (N) and an exchangeable (E) site. α−tubulin binds GTP that is buried at the monomer-monomer interface (N-site), whereas β−tubulin binds GTP that is exposed on the monomer surface (E-site), and can be readily hydrolyzed into GDP upon MT polymerization, predominantly at the interface with GDPtubulin. 3 Tubulin dimer with GTP at the E-site (GTP-tubulin) can initiate and promote protofilaments and MT assembly. 3, Tubulin dimer with GDP at the E-site (GDP-tubulin), assumes a kinked conformation between its α and β-tubulin subunits. Recent studies suggest that in the soluble state, both GDP-and GTP-tubulin adopt, on average, similar kinked conformations and straightening occurs when GTP-tubulin polymerizes into MT. GDP bound to tubulin is not exchangeable back to GTP as long as the dimer is part of the MT polymeric form. 2 The assembled GDP-tubulin (whose E-site is buried within the polymer) is under conformational tension that can promote MT disassembly catastrophe. 1, 12 The protofilaments that are disassembling from MT edges, are curving outwards and were shown to bend in a direction perpendicular to the curvature plane of the polymeric tube, suggesting a specifically curved dimer symmetry. 13 The resulting disassembled curved GDP-rich tubulin dimers and curved one-dimensional (1D) tubulin oligomers can promote the assembly of tubulin single-rings (38 nm in diameter). 3,9,14,15 Free GDP-tubulin dimers may readily exchange their bound GDP for GTP. 16 However, the GDP to GTP exchange reaction equilibrium constant is an order-of-magnitude smaller for dimers whose E-sites are buried in oligomers or rings. 17 Tubulin oligomers and single-rings are dynamic structures that coexist with MTs and involved in MT assembly and disassembly processes. 18 The rings are a storage form of active tubulin subunits. 18 The initial phase of MT assembly is accompanied by a simultaneous ring disassembly, 14 providing most (85 -90%) of the tubulin subunits incorporated in the initial stages of MT assembly. 15 During MT disassembly, the assembly of rings is delayed until the concentration of dimers is sufficiently high. 14 Despite their involvement in MT assembly, the mechanism of tubulin ring assembly and disassembly remained poorly understood. Solutions of ice-cold GTP-tubulin are rich in dimers, oligomers, and rings and do not form MTs. MT assembles in-vitro only after the temperature is increased. To resolve the mechanism of ring assembly and disassembly, it is of utmost importance to characterize cold GDP-and GTP-tubulin solutions, the precursor solutions at the onset of MT assembly at elevated temperatures. We have therefore analyzed the structure, interactions, and kinetics of cold GDP-tubulin and cold GTP-tubulin solutions using solution small-angle X-ray scattering (SAXS). Steady-state SAXS data were analyzed using atomic structural models of tubulin assemblies whose mass fraction was determined based on an isodesmic thermodynamic model of tubulin self-association. The analyses revealed the structure and mass fractions of tubulin dimer, tubulin single-rings, and tubulin oligomers (ring fragments), as a function of the total tubulin concentration. Additionally, we determined the longitudinal association standard Helmholtz free energies between GTP-or GDP-tubulin dimers in cold solutions. Using time-resolved SAXS, we analyzed the disassembly kinetics of GDP-tubulin single-rings upon dilution, GTP addition, or a temperature jump. We discovered a rapid isodesmic disassembly catastrophe (half-life of about 0.1 sec) of cold tubulin single-rings upon the addition of GTP or sample dilution. Temperature elevation induced a slower ring isodesmic dissociation kinetics (half-life of about 38 sec). ## Results and Discussion Tubulin Single Rings Coexisted With Tubulin Oligomers Cryo-TEM images (Fig- ure S1) reveal that free tubulin dimers coexisted with tubulin single-rings, and small tubulin oligomers in curved conformations (ring fragments). In excess GTP, the fraction of tubulin single-rings was smaller than in excess GDP. 14,19 HPLC analysis showed that seven Heat-Cool Cycles (see Materials and Methods in the supporting information) or incubation in 10 ± 0.5 mM GDP, led to 94% GDP-tubulin and 6% GTP-tubulin at the E-site. 17 After seven heat-cool cycles tubulin retains its MT assembly capability, 17 however, the fraction of tubulin single-rings was larger than after incubation in 10 ± 0.5 mM GDP. The excess of rings is most likely kinetically-trapped (stable) rings. A small fraction of stable tubulin single-rings remains even after SEC elution experiments, 19 suggesting that even after about 100-fold dilution, some of the rings do not completely disassemble. Mass fraction Distribution of Tubulin Assemblies Steady-state SAXS measurements at different concentrations of GDP-and GTP-tubulin (Figure 1) were performed below the critical temperature for MT assembly. The two data sets were fitted to a linear combination of tubulin single-rings and oligomers (ring fragments), whose mass fractions were based on a thermodynamic model of tubulin self-association (Equations 1 and 6). The best fit was obtained after assuming that 10 % of the tubulin mass fraction formed kinetically-trapped stable tubulin single-rings. The rest of the tubulin mass was distributed according to the thermodynamic model (Equations 4 -6). In a separate experiment, a GDP-tubulin sample at a higher concentration adequately fitted the same model and further supported these results (Figure S3). GTP Acts as a Hydrotrope Panels C and D of Figure 1 show the distribution of tubulin oligomers (ring fragments) that best-fitted the data at each GDP-and GTP-tubulin concentration, respectively. The best-fitted standard Helmholtz dimer-dimer association free energy per longitudinal contact, ∆F • c , and standard Helmholtz free energy cost for closing a curved oligomer into a ring, ∆F • RC , are provided in Table 1. GTP-tubulin had a weaker standard Helmholtz free energy values, suggesting that GTP acts as a hydrotrope that increases the solubility of tubulin, similarly to ATP. 20,21 The thermodynamic analysis is applicable for GDP-tubulin, where there is no hydrolysis reaction. The fact that the same analysis explained the data of cold GTP-tubulin solutions (after adjusting the dimer-dimer association energy) is consistent with our recent HPLC experiments that showed that at low temperatures the hydrolysis of GTP is very slow, when free in the buffer or bound to the tubulin dimers. 17 after dilution series, as indicated in the figure (in units of mg/mL). The data were fitted (red curves) to a linear combination of scattering curves from atomic models of tubulin rings and oligomers (see Figure S2 and subsection SAXS Models in Materials and Methods in the supporting information), whose mass fraction was determined by a thermodynamic model of tubulin self-association (Equations 1 and 6), with the best-fitted standard Helmholtz free energies from Table 1. The measured intensity at q < 0.08 nm −1 of 2.4 mg/mL GTP-tubulin were omitted owing to a technical measurement error. Mass fraction distribution of GDPtubulin (C) and GTP-tubulin (D) rings and oligomers (ring fragments) as a function of size (number of tubulin dimers), based on the standard Helmholtz free energies that bestfitted the data (Equation 7)) and the measured tubulin concentration. The mass fraction of 13 dimers, includes open-, closed-, and stable-rings. The contribution of open rings was negligible, the contribution of stable-rings was typically 10% of the total tubulin mass fraction. The remaining mass fraction of 13 dimers was owing to closed-rings, as predicted by the thermodynamic model. Heat map of the GDP-tubulin (E) and GTP-tubulin (F) mass fraction, plotted in the plane of the number of dimers in assembly versus the total tubulin concentration, computed according to Equation 1, using the parameters from Table 1. Data were measured at the P12 EMBL BioSAXS Beamline in PETRA III (DESY, Hamburg). GDP-Tubulin Single Rings Catastrophe Time-resolved SAXS (TR-SAXS) measurements followed tubulin-ring disassembly kinetics, triggered by either dilution (Figure 2), GTP addition (Figure 3), or a temperature jump (Figure 4). GTP addition includes dilution of the tubulin solution, hence it was important to separate the effect of dilution from the addition of GTP. The dilution and the GTP addition experiments also simulated the dilution and nucleotide exchange reactions that may occur in SEC elution experiments. 17,19 The TR-SAXS data were fitted to an isodesmic kinetic model (Equations 12 and 13), in which dimers were rapidly added/removed one at a time and rings were closed/opened at a two orders-of-magnitude slower rate (Table 2). The thermodynamic parameters determined by the steady-state measurements (Table 1) and SEC-SAXS chromatogram analysis, 19 were used to estimate the initial size distribution, and to calculate the ratio between the assembly and disassembly rate constants, according to the detailed balance conditions (Equations 10 and 11). The rate constants and the standard Helmholtz free energies used to analyze the TR-SAXS data are summarized in Table 2. Very rapid (half-life of about 0.1 sec) ring and oligomers disassembly catastrophe kinetics were observed upon dilution or GTP addition. Steady-states were attained within about 1 s (Figure S5). The observed disassembly products were dimers, tetramers (dimer of dimers), and hexamers (trimer of dimers). Larger oligomers did not accumulate to detectable amounts. The fraction of hexamers at steady-state was similar upon dilution or GTP addition. Upon dilution, however, the fraction of rings was higher and consequently, the fractions of dimers and tetramers were lower. 2) determined the mass fraction of rings and oligomers (ring fragments) as a function of time (right panel), which in turn were used to compute the red curves, using Equation 6. The error in the mass fractions (right panel) are indicated by shaded colored areas, surrounding the solid curves. Data were measured at the ID02 beamline (ESRF, Grenoble). 2 shows the best-fitted model parameters. Data were measured at the ID02 beamline (ESRF, Grenoble). 23 Disassembly of GDP-Tubulin Rings Following a Temperature Jump The fraction of tubulin rings decreases when increasing the temperature. 14,18 We, however, observed a slower GDP-tubulin ring disassembly rate (half-life of about 38 sec) following a temperature jump (Figure 4). Initially, the net amount of tetramers and hexamers decreased because they disassembled at 36 • C, but then the rings continued to disassemble and hence the mass fraction of tetramers and hexamers increased. Furthermore, the mass fractions of tubulin in free dimers and in tetramers were comparable after ≈ 40 s (unlike the low-temperature results, in which free dimer dominated). These data show that at a higher temperature, larger oligomers were more stable, suggesting that water molecules were released upon association and increased the entropy of GDP-tubulin oligomerization. 24 Similar steady-state results were obtained from a 1 h incubation of GDP-tubulin at 36 • C, where strong longitudinal association standard Helmholtz free energy (∆F was fitted to the data (Figure S4). The data in Figure S4, however, included tubulin aggregates, hence the fit was limited to a smaller q-range (0.2 nm −1 < q < 3 nm −1 ), and was of somewhat lower precision. The characteristic oscillation pattern of ring and ring fragments was substantially reduced compared with the 9 • C scattering curves (Figure 1). We attribute the change to the increased flexibility and thermal perturbations owing to the temperature increase. An earlier study also found that above ≈25 • C the fraction of tubulin rings at steady-state decreases with increasing temperature. 15 In our earlier study, 14 we showed that a similar decrease in the fraction of GTP-tubulin single-rings was observed after the temperature was increased. Table 2: The best-fitted rate constants and their associated thermodynamic parameters, used to analyze the TR-SAXS data. The standard Helmholtz free energies were obtained in the molar fraction scale. The standard Gibbs free energies in the concentration scale can be obtained by adding 4 k B T (or 2.3 kcal • mol −1 ). 19 Two orders of magnitude larger rate constants for ring closure did not change the fitting results, suggesting that our data were insensitive to these rate constants. 2 shows the best-fitted model parameters. Data were measured at the ID02 beamline (ESRF, Grenoble). 23 ## Conclusions In this paper, we examined the disassembly catastrophe mechanism of tubulin oligomers and single-rings, involved in the early steps of microtubule assembly. We showed that at low temperatures and over a wide range of GDP-and GTP-tubulin concentrations, the distribution of tubulin single-rings and 1D curved oligomers (ring fragments) is consistent with a thermo-dynamic theory of isodesmic tubulin self-association. GTP acts as an effective hydrotrope that increases the solubility of tubulin, reduces the longitudinal dimer-dimer Helmholtz standard association free energy, and the free-energy of ring closure. Therefore, solutions of GTP-tubulin contained higher concentrations of tubulin dimers and smaller assemblies, compared with the corresponding GDP-tubulin solutions. GDP-tubulin single-rings rapidly destabilized upon dilution or GTP addition. Time-resolved experiments discovered ring disassembly catastrophe (half-life of about 0.1 sec) and were consistent with an isodesmic disassembly mechanism, involving ring-opening followed by two-orders-of-magnitude faster consecutive single dimer removal steps. A similar disassembly mechanism explained the disassembly of cold GDP-tubulin rings following a temperature jump to 36 • C, however, at a significantly slower rate (half-life of about 38 s). ## Materials and Methods Materials were purchased from Sigma-Aldrich Co. The accuracy of the reported temperatures was ±1 • C. ## Tubulin Purification Tubulin was purified from porcine brains, using a modified version of the high-molarity buffer purification method, 25 as described earlier. 14 Purified tubulin was in BRB80 (80 mM 1,4-piperazineediethanesulfonic acid (PIPES), 1 mM ethylene glycol-bis(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), 1 mM MgCl 2 , at pH 6.9 (adjusted with KOH)). Immediately following purification, aliquots were flash-frozen in liquid nitrogen, stored at −80 • C, and used within the following two weeks. 26,27 Before used in experiments, tubulin was thawed on ice. ## Heat-Cool Cycles To hydrolyze the residual GTP, seven heat-cool cycles were applied, which led to tubulinbound nucleotide compositions at the E-site of 94% GDP-tubulin, and 6% GTP-tubulin. 17 We verified that this tubulin retained its activity and was able to form MTs to a similar extent before and after seven heat-cool cycles. 17,28 Similar nucleotide composition was achieved by incubation of fresh tubulin samples (i.e. a sample that did not undergo seven heat-cool cycles) in the presence of 10 ± 0.5 mM GDP. 19 ## Sample Preparations Before any measurement, the tubulin samples were spun-down for 30 min at 4 • C and 20,800 g. The top 60%, taken from the supernatant, was then used for the experiments. Steady-state SAXS 42 ± 3 mg/mL purified tubulin, containing a total GTP and GDP concentration of 1.3 ± 0.1 mM and 0.9 ± 0.1 mM ATP, was used for the steady-state SAXS experiments. GTPtubulin was obtained by diluting the tubulin sample in BRB80 and adding 3 ± 0.1 mM GTP. GDP-tubulin was obtained by applying seven heat-cool cycles, diluting the sample in BRB80, and adding 3 ± 0.1 mM GDP. Steady-state SAXS measurements were performed at 9 • C. ## Time-resolved SAXS 24 ± 2 mg/mL tubulin samples, containing a total GTP and GDP concentration of 0.7 ± 0.1 mM and 0.4 ± 0.05 mM ATP, underwent seven additional heat-cool cycles and spin-down as described above. Samples were then transferred to the stopped-flow reservoir, kept at 9 • C, from which they were injected into the stopped-flow quartz capillary, kept at either ## Cryo-TEM GTP-and GDP-tubulin. 36 ± 3 mg/mL purified tubulin, containing 0.6 ± 0.1 mM GTP, 0.4 ± 0.05 mM GDP, and 0.7 ± 0.1 mM ATP, were thawed on ice and spun-down as described above. Either 10 ± 0.5 mM GTP or 10 ± 0.5 mM GDP were added, and the samples were incubated on ice for 1.5 h. The incubated samples were then 10-fold diluted in BRB80 supplemented with either 1 ± 0.1 mM GTP or 1 ± 0.1 mM GDP, incubated on ice for 20 min, and flash-frozen. Cycled tubulin. 44±4 mg/mL purified tubulin, containing a total GTP and GDP concentration of 1.3 ± 0.1 mM and 0.7 ± 0.1 mM ATP, underwent seven additional heat-cool cycles and spun down, as described above. The supernatant was then incubated for 2 h on ice, and 10-fold diluted in BRB80, supplemented with 1.3 ± 0.1 mM GDP. 1 h after the dilution, the samples were flash-frozen. ## Solution SAXS and Time-Resolved SAXS Measurements and Data Reduction SAXS experiments were performed at P12 EMBL BioSAXS Beamline (headed by D. Svergun) in PETRA III (DESY, Hamburg), 22 and in ID02 beamline (headed by T. Narayanan) in the European synchrotron radiation facility (ESRF, Grenoble). 23 At P12, measurements were taken using an automated temperature controlled sample changer setup. 29 Samples were injected into a temperature controlled 2 mm quartz capillary. The sample-to-detector distance was 3 m, resulting in a q-range of 0.05 − 5 nm −1 . The size of the beam was 0.2×0.12 mm 2 and the exposure time per frame was 45 ms. The X-ray wavelength was 1.24 , at a flux of 5 × 10 12 photons • s −1 , and the scattered intensity was recorded on a Dectris Pilatus 2M detector. At ID02, time-resolved SAXS (TR-SAXS) data were measured using a BioLogic SFM-400 stopped-flow apparatus, as explained in our earlier publications. The capillary outer diameter was 1.4 mm, and the sample-to-detector distance was 2.5 m, resulting in a q-range of 0.03 − 3 nm −1 . The size of the beam was 0.3×0.2 mm 2 and the exposure time per frame was 20 ms. The intervals between frames was 160 ms or longer. The X-ray wavelength was 0.995 , at a flux of 1 × 10 13 photons • s −1 , and the scattered intensity was recorded on a Rayonix MX170-HS detector. 23,33 The 2D scattering images were normalized to the intensity of the transmitted beam and azimuthally averaged to yield the scattering intensity as a function of the magnitude of the scattering vector, q, using the integrated analysis pipeline SASFLOW 34 at the P12 beamline, and the online SAXS/WAXS data reduction package (SPD) 35 at ID02. Background measurements, before and after each sample, were performed from the solvent of each sample, under identical measurement conditions. Background scattering curves were averaged and subtracted from the averaged sample signal, resulting in the final background-subtracted scattering intensity curve, as explained in our earlier papers. 14, Absolute intensity scales were obtained using water, whose differential cross section is 0.0164 cm −1 , when 1 < q < 4 nm −1 . 39 ## TR-SAXS Measurement protocol and Analysis The disassembly kinetics following dilution were measured by mixing 100 µL of 13±1 mg/mL GDP-tubulin (following seven heat-cool cycles) and 100 µL of BRB80, supplemented with 0.7 ± 0.1 mM GDP (GDP was added to the diluting solution to maintain a constant concentration of unbound GDP in the sample). The disassembly kinetics following GTP addition, was measured by mixing 100 µL of 13 ± 1 mg/mL GDP-tubulin and 100 µL of BRB80, supplemented with 0.7 ± 0.1 mM GDP and 8 ± 0.5 mM GTP. Both experiments were done at k −1 and k −2 were fitted to the TR-SAXS data (using Equations 12 and 13)), whereas k 1 and k 2 where calculated based on the detailed balance conditions (Equations 10 and 11), using thermodynamic parameters close to that of the steady-state data analysis (see Table 2). ## Cryo-TEM Tubulin solutions were directly imaged using transmission electron microscopy at cryogenic temperatures (cryo-TEM). A droplet of 3 µL was deposited on a 300 mesh Cu Lacey grid (Ted Pella Ltd.) and blotted using Vitrobot Mark IV (FEI Co.). Ultra-thin films (∼ 20 − 200 nm thick) were formed following removing of excess solution by blotting with filter papers. Specimens were vitrified by rapid plunging into liquid ethane pre-cooled with liquid nitrogen at controlled temperature and relative humidity. The vitrified samples were transferred to a cryo-specimen holder (Gatan model 626; Gatan Inc.) and imaged at −177 • C using a Tecnai ## SAXS Models SAXS models were calculated using our home-developed data analysis software D+. 36,40 The scattering amplitude, F sol dimer ( q), from atomic models of the tubulin dimer subunit in solution were modeled as explained. 14,19,36 The atomic tubulin models were based on PDB ID 5JQG. 41 Missing residues were added according to the published tubulin sequence, downloaded from UniProt, 45 and refined using Modeller, 46 as explained. 19 Following refinement, hydrogens were added using MolProbity. 47 Tubulin assemblies were modeled as fragments of the tubulin single-ring, according to the following symmetry: Chains A and B of PDB 5JQG were translated by vector R 0 = {x 0 , y 0 , z 0 } = {−2.14 nm, 8.19 nm, 1.73 nm} and rotated according to the Tait-Bryan rotation angle convention used in D+, 36,40 where {α, β, γ} = {58.16 • , −25.15 • , −125.65 • }. The resulting PDB was placed in a ring symmetry, using the rotation matrix A j (0, 0, γ j ), which rotates the dimer around the z-axis by γ j , defined as: for which θ j = 2πi /n Dimers , where n Dimers is the number of tubulin dimers in the ring. The dimer was then translated by: where R Ring is the ring radius from the origin to the center of mass of the tubulin dimer. The scattering amplitude of a linear tubulin oligomer, containing n dimers was: The contribution of the hydration layer of the dimer was taken into account in D+, 36,40 by computing the contribution of voxels around the protein using the following D+ parameters: solvent voxel size of 0.05 nm, solvent probe radius of 0.14 nm, solvation thickness of 0.2 nm, and hydration layer electron density of 364 e/nm 3 . 36 The scattering amplitudes of the hydration layer of the ring or its fragments were calculated using the hydrated tubulin dimer subunit and removing the overlapping hydration areas, as described in our earlier publication. 36 To optimize the computation efficiency, all the models were computed after translating their center of mass to the origin. The radius of ring fragments that best fitted the data was 19.2 nm, whereas the radius of the complete ring that best fitted the data was 18.5 nm. The modeled scattering intensity of the ring or its linear tubulin oligomer fragments, containing n dimers are: where, F hydrated n is the scattering amplitude of the hydrated F n structure, and ... Ωq represents the orientation averaging of the scattering intensity in reciprocal ( q) space. All the models were then rescaled to an absolute scale, as described. 48 A Thermodynamic Model of Tubulin Self-Association To account for the equilibrium between free dimers, curved oligomers, and tubulin singlerings, we have applied a thermodynamic model to determine the mole fractions of the different tubulin assemblies. The derivation of the model was described elsewhere. 19 The resulting mole fractions, at temperature T and total tubulin concentration C Tubulin Total , depends on the standard Helmholtz dimer-dimer association free energy per longitudinal contact, ∆F • c , and the standard Helmholtz free energy cost for closing a curved oligomer into a ring, beyond the energy gain from the formation of an additional longitudinal dimerdimer contract, ∆F • RC . We have shown 19 that the mole fractions that minimize the total Helmholtz free energy in the grand canonical ensemble are where k B is Boltzmann's constant. Solving the conservation of mass, for X 1 , using the experimental value of C Tubulin The total mole fraction of tubulin dimers in solution was defined as: 49 where N Total was the total number of molecules in solution (N Total = N Tubulin + N Water ≈ N Water ), and N AV was the Avogadro number. To account for stable (kinetically trapped) tubulin rings, that did not disassemble owing to slow kinetics, the total mole fraction of tubulin dimers was defined as: where x Stable Rings is the mole fraction of stable rings out of the total tubulin concentration: for which C Stable Rings is the concentration of stable rings in units of mg/mL. ## Integrating the Thermodynamic model into the SAXS data analysis The tubulin oligomers mole fractions in solution were determined by the above thermodynamic model. The expected ensemble scattering intensity was a weighted sum of the oligomers modeled scattering curves, according to the calculated mole fractions: I Ring , I Stable Ring , and I Model,n were the modeled scattering intensity curves of the closed ring, stable rings, and of a curved oligomer, containing n tubulin dimers, respectively. N is the number of dimers in a full ring. The modeled scattering curves were computed using D+ software (https://scholars.huji.ac.il/uriraviv/book/d-0) 36,40 according to the hydrated atomic model described above, then normalized to absolute units of cm 2 /mg (Figure S2). For each set of ∆F was minimized. n q is the number of q points within the fitted q-range. Minimizing χ 2 with respect to Const produced the analytic solution, for each experimental scattering curve: I Signal (q) is the measured scattering intensity, and σ(q) is the measured scattering intensity error at a scattering vector whose magnitude is q. q min and q max define the q-range used to fit the data, which was set to 0.05 − 3 nm −1 , unless otherwise indicated. Each fit was repeated 10 times, using only one-tenth of the data points, uniformly distributed (within the fitted q-range), and randomly selected. For each fit, the values of χ 2 were computed according to Eqs. 7 and 8 for the entire fitted q-range. A k-means clustering algorithm was then applied on the resulting χ 2 values, to separate the results into two clusters. 31,50 Within the largest cluster, the model with the mean χ 2 value was selected. The errors of the total and the n-th model concentrations were calculated based on the standard deviation around the mean χ 2 value, within the largest cluster. The global parameters, F • RC , ∆F • c , C Tubulin Total and x Stable Rings , were fitted by minimizing χ 2 Mean of the concentration series, given by: where χ 2 j is the χ 2 value calculated according to Equation 7 for signal j of the concentration series, and J is the number of concentrations measured in the series. The fitting and clustering algorithms were applied using a Matlab program, which can be found online (at https://scholars.huji.ac.il/uriraviv/thermodynamic-model-tubulin). . ## Kinetic Model of Tubulin Single Ring Assembly/Disassembly Ring assembly kinetics was assumed to follow a consecutive second-order kinetic steps in the forward direction, and a first-order kinetics in the backward direction. Ring-closing kinetics was assumed to be first order in both directions. where D n represents an oligomer containing n tubulin dimers, and R 13 , corresponds to a tubulin single-ring, containing 13 tubulin dimers. Given k 1 and ∆F • c , k −1 can be directly calculated from detailed balance: where K C and K X are the equilibrium constants on the concentration and molar fraction scales, respectively. Similarly, k −2 can be derived from k 2 and ∆F • RC : The series of modeled signals corresponding to the time-resolved experiments, we used the following rate equations: for n ∈ {2, ..., 12} The rate equation for [D 1 ] was calculated from mass conservation: This set of ordinary differential equations (ode) was numerically solved to propagate in time the initial conditions, described by the thermodynamic model. The solution was obtained by the Dormand-Prince method (an explicit member of the Runga-Kutta (4,5) family), using the ode45 function of Matlab . The calculated concentrations were set into Equation 6, following units conversion:

chemsum

{"title": "Mechanism of Tubulin Oligomers and Single-Rings Disassembly Catastrophe", "journal": "ChemRxiv"}

hybrid_photocathode_consisting_of_a_cugao₂_p-type_semiconductor_and_a_ru(<scp>ii</scp>)–r

## Abstract: A CuGaO 2 p-type semiconductor electrode was successfully employed for constructing a new hybrid photocathode with a Ru(II)-Re(I) supramolecular photocatalyst (RuRe/CuGaO 2 ). The RuRe/CuGaO 2 photocathode displayed photoelectrochemical activity for the conversion of CO 2 to CO in an aqueous electrolyte solution with a positive onset potential of +0.3 V vs. Ag/AgCl, which is 0.4 V more positive in comparison to a previously reported hybrid photocathode that used a NiO electrode instead of CuGaO 2 . A photoelectrochemical cell comprising this RuRe/CuGaO 2 photocathode and a CoO x /TaON photoanode enabled the visible-light-driven catalytic reduction of CO 2 using water as a reductant to give CO and O 2 without applying any external bias. This is the first self-driven photoelectrochemical cell constructed with the molecular photocatalyst to achieve the reduction of CO 2 by only using visible light as the energy source and water as a reductant. ## Introduction The use of artifcial photosynthesis, particularly for the photochemical reduction of CO 2 , has gained much attention because it might provide solutions for overcoming both the problem of global warming and the shortage of fossil resources. Metal complex photocatalysts have been intensively developed for achieving the reduction of CO 2 using the energy of visible light with a high selectivity for the products. In particular, supramolecular photocatalysts composed of both redox photosensitizer and catalyst units have exhibited high selectivity, durability, and efficiency for the reduction of CO 2 under visible light conditions even in aqueous solutions, 4,5 as well as in organic solvents. In these photocatalytic reactions, sacrifcial reductants such as 1-benzyl-1,4-dihydronicotinamide, 7 sodium ascorbate, 4 and 2-(1,3-dimethyl-2,3-dihydro-1H-benzimidazol-2yl)benzoic acid 5 are necessary because of the low oxidizing power of the photosensitizer unit, which cannot sufficiently oxidize water, an abundant electron donor. Hybrid photocatalysts and photoelectrodes composed of metal complexes and semiconductor materials with relatively high photochemical oxidizing power, in which the metal complexes act as catalysts alone in some cases but as photocatalysts in others, have recently been studied for the visible-light-driven photocatalytic reduction of CO 2 . In the cases where metal complexes are used as photocatalysts, hybrid photocatalysts with semiconductor particles can cope with both the high selectivity of the CO 2 reduction and the strong oxidation power via the step-by-step excitation of both the redox photosensitizer unit of the metal complex photocatalyst and the semiconductor, so-called Z-scheme mechanism. However, visible-light-driven CO 2 reduction accompanied by water oxidation has not been achieved yet using hybrids with these particle systems. We have previously reported a hybrid photocathode consisting of a Ru(II)-Re(I) supramolecular photocatalyst (RuRe) fxed on a NiO p-type semiconductor electrode for the photoelectrochemical reduction of CO 2 . 31 In this system, the metal complex photocatalyst can drive the reduction of CO 2 by the injection of electrons from an external electric circuit through the NiO semiconductor electrode without the addition of any sacri-fcial electron sources. This RuRe/NiO photoelectrochemical system was used to successfully achieve the catalytic reduction of CO 2 using water as an electron source, combined with a CoO x / TaON n-type semiconductor photoanode for the oxidation of water. 32 However, the performance of this photoelectrochemical cell was still insufficient, i.e., the combined photoelectrochemical cell needed the assistance of an external electrical bias (0.3 V) and a chemical bias (0.10 V) produced by a difference in pH, and the photoelectrochemical cell generated much lower yields of the reduction products (CO and H 2 ) compared to the oxidation one (O 2 ) from the viewpoint of electron balance. One of the main reasons for these insufficiencies was the low activity of the NiO electrode. Although NiO has been widely used as an electrode material for p-type dye-sensitized photocathodes, the flat band position of NiO is relatively negative (+0.34 V vs. Ag/ AgCl in a saturated aqueous solution of KCl at a pH of 7). 33 This could cause a large loss of energy for passing an electron to the excited state of the Ru(II) photosensitizer unit (E 1/2 ¼ +0.51 V vs. Ag/AgCl) 31 through the interface, leading to the requirement of a large external bias (totally 0.4 V) when this photocathode is used for constructing a photoelectrochemical cell combined with a photoanode. The consumption of electrons for reducing trivalent nickel ions (Ni 3+ ) that originally existed in NiO could also be problematic because it lowers the Faraday efficiencies of the reduction reactions on the photocathode. 31,32,34 Therefore, the development of new p-type semiconductor electrodes is necessary for the feld of hybrid photoelectrochemical cells with supramolecular photocatalysts for reduction of CO 2 . Here, we report a novel hybrid photocathode (RuRe/CuGaO 2 ) consisting of CuGaO 2 as the p-type semiconductor electrode and RuRe as the photocatalyst for the reduction of CO 2 . CuGaO 2 with a delafossite crystal structure is known to exhibit p-type semiconducting properties that are derived from native Cu + vacancies. 35 It has received attention as a transparent conducting oxide material and has also been studied as an alternative to NiO for dye-sensitized photocathodes due to both its high conductivity (10 1 -10 2 S cm 1 ) 38 and positive flat band potential, which was reported to be approximately 0.16 V more positive compared to that of NiO. 39,40 Although CuGaO 2 electrodes have been investigated as parts of dye-sensitized solar cells, to the best of our knowledge there are no reports on the application of CuGaO 2 electrodes for photocatalytic reactions such as CO 2 reduction and H 2 evolution. The onset potential for the reduction of CO 2 by the as-synthesized RuRe/CuGaO 2 was revealed to be 0.4 V more positive in comparison to that of the RuRe/NiO electrode. A photoelectrochemical cell consisting of RuRe/CuGaO 2 and a CoO x /TaON photoanode enabled the visible-light-driven catalytic reduction of CO 2 using water as the reductant without applying any external bias. ## Results and discussion Preparation of the RuRe/CuGaO 2 photocathode CuGaO 2 powder was prepared using a solid-state reaction method. The XRD pattern of the synthesized powder revealed that the delafossite structure of the CuGaO 2 was obtained with no obvious impurity phase (Fig. S1 †). The UV-vis diffuse reflectance spectrum of the synthesized CuGaO 2 powder indicated that the powder absorbs light at l < 460 nm, which has been reported in previous research into CuGaO 2 powder (Fig. S2 †). 41,42 The CuGaO 2 electrodes were prepared by dropcasting a powder suspension onto FTO glass substrates following annealing with a N 2 flow. SEM observations revealed that the polycrystalline CuGaO 2 particles had rod-like shapes on the micron scale and the thickness of the stacked CuGaO 2 particle layer of the electrode was approximately 15 mm (Fig. S3 †). The dominant material at the solid-liquid interface of the electrode should be the deposited CuGaO 2 particles and not the underlying flat FTO flm. RuRe and its model complexes (Ru and Re) (see Chart 1) were synthesized in accordance with reported procedures. 43 A CuGaO 2 electrode was immersed in an acetonitrile solution containing the metal complex overnight to obtain hybridized photocathodes. ## Photoelectrochemical properties of the RuRe/CuGaO 2 photocathode The photoelectrochemical properties of the synthesized RuRe/ CuGaO 2 electrode were investigated under irradiation at l ex > 460 nm, which can be selectively absorbed by the Ru photosensitizer unit of RuRe, in an aqueous solution containing NaHCO 3 (50 mM) saturated with CO 2 , which was used as the supporting electrolyte. Fig. 1(a) shows the current-potential curves of the RuRe/CuGaO 2 electrode under continuous visiblelight irradiation (l ex > 460 nm) and under dark conditions. This clearly indicates that the RuRe/CuGaO 2 electrode generated a cathodic photocurrent under irradiation. The difference in the observed current between the irradiation and dark conditions indicated that the cathodic photoresponse of the RuRe/CuGaO 2 electrode started at approximately +0.3 V vs. Ag/AgCl (equivalent to +0.9 V vs. RHE). It should be noted that this onset potential of the photocurrent was approximately 0.4 V more positive than that of RuRe/NiO (ca. 0.1 V, Fig. 1(b)). 31 A pristine CuGaO 2 electrode showed a slight cathodic photoresponse (ca. 1 mA cm 2 ) under irradiation at l ex > 460 nm (Fig. S4 †), which was almost negligible compared to the photocurrent from the RuRe/ CuGaO 2 electrode, as shown in Fig. 1(a). The anodic current obtained from ca. +0.2 V vs. Ag/AgCl in the dark was derived from the self-oxidation of the CuGaO 2 surface 42 but not from the redox reaction of the immobilized RuRe. RuRe. In the case of the RuRe/CuGaO 2 electrode, the dependence agreed well with the absorption spectrum of the electrode, whereas the bare CuGaO 2 electrode exhibited almost no photoresponse under irradiation at l ex > 460 nm. The flat band potential of the CuGaO 2 electrode in the reaction solution was estimated using electrochemical impedance spectroscopy (Fig. S5 †) to be +0.47 V, which was more negative than the reduction potential of the excited state of RuRe ðE *red 1=2 ¼ þ0:49 VÞ. 44 These results obviously indicate that the photocurrent was induced by the injection of the electrons from the CuGaO 2 electrode into the excited Ru photosensitizer unit of RuRe, as shown in Fig. 3. To conclude this section, we successfully synthesized the new hybrid photocathode (RuRe/ CuGaO 2 ) for the reduction of CO 2 with an onset potential that was approximately 0.4 V more positive in comparison to that reported for the RuRe/NiO photocathode. ## Photoelectrochemical reduction of CO 2 using the RuRe/ CuGaO 2 photocathode The gas products were analyzed during the continuous visiblelight irradiation (l ex > 460 nm) of the RuRe/CuGaO 2 photocathode in an aqueous solution saturated with CO 2 in the presence of an applied potential. Fig. 4 shows the time courses of both the produced CO and H 2 , and a half amount of electrons passing at an applied potential of 0.3 V vs. Ag/AgCl. After irradiating for 15 h, 966 nmol of CO and 622 nmol of H 2 were detected as the reduction products when the turnover number for the formation of CO (TON CO ), which was based on the RuRe deposited on the electrode, was 125, and the total faradaic efficiency for the production of the reduced products (CO + H 2 ) was 81%. In the irradiated solution, HCOOH was not detected (the detection limit of the capillary electrophoresis system used was 2 mmol). An isotope tracer experiment using 13 CO 2 and NaH 13 CO 3 was also conducted to confrm the source of the carbon for the produced CO. GC-MS analysis using NaH 13 CO 3 as the electrolyte under a 13 CO 2 atmosphere revealed that 13 CO was the main product of the reaction with a very small proportion of 12 CO, where the 13 C content was 99% in both 13 CO 2 and NaH 13 CO 3 (Fig. 5(a)). It is worth noting that the Re unit of RuRe has three 12 CO ligands, and the exchange of the 12 CO ligands with 13 CO proceeded in the photocatalytic 13 CO 2 reduction reactions in homogeneous systems. On the other hand, only 12 CO was detected in the gas phase of the reaction using ordinary CO 2 and NaHCO 3 (Fig. 5(b)). These results clearly indicate that the source of carbon for CO was CO 2 . Table 1 shows the results of the photoelectrochemical reactions that were conducted under various conditions. The RuRe/ CuGaO 2 electrode could produce CO as the main product with H 2 under visible-light irradiation at an applied potential from 0.1 V to 0.7 V vs. Ag/AgCl (Table 1, entries 1-3). The TON CO of the RuRe/CuGaO 2 electrode was more than twice the reported value of the RuRe/NiO electrode (32; entry 10), even at more positive potentials, i.e., 0.3 V and 0.1 V. Carbon monoxide was not produced under an Ar atmosphere even when the same photocathode was used, whereas H 2 was produced (entry 4). The reactions using the bare CuGaO 2 electrode or electrodes with only one of the model mononuclear complexes, i.e., Ru or Re, did not lead to the formation of any CO (entries 5, 6, 7 and 8). In these cases, almost negligible or small amounts of H 2 were produced during the irradiation. Therefore, the CuGaO 2 electrode, the RuRe supramolecular photocatalyst, and a CO 2 atmosphere may all be required for the photoelectrochemical reduction of CO 2 . We note that an electrode that was loaded with both the model complexes (Re and Ru) at random produced a much smaller amount of CO (entry 9), which indicates that the molecular design of RuRe, which enables the sufficiently efficient transfer of electrons from the reduced photosensitizer unit to the catalyst unit, is quite important for developing photocathodes for the reduction of CO 2 . As was described above, the production of H 2 was detected using the RuRe/CuGaO 2 photocathode in reactions under both Ar and CO 2 atmospheres, whereas only a small amount of H 2 was produced in the case of RuRe/NiO. It has been reported that the decomposition products of both the Ru photosensitizer unit in systems using RuRe supramolecular photocatalysts and Ru in a mixed system of Ru and Re, i.e., a complex of the form [Ru II (N^N) 2 (solvent) 2 ] n+ (N^N ¼ diimine ligand), were gradually produced during the photocatalytic reactions, and these acted as catalysts for both the evolution of H 2 and the formation of HCOOH from CO 2 . 5,45 However, the Ru/CuGaO 2 photocathode produced a smaller amount of H 2 than that of the RuRe/CuGaO 2 photocathode. Therefore, the generation of H 2 might result not only from the decomposition products of RuRe but also from the coexistence of photoexcited RuRe on the CuGaO 2 surface. Because it has been reported that CuGaO 2 worked as a H 2evolving photocathode under UV light irradiation (CuGaO 2 itself can absorb UV light), 42 back electron transfer from photoexcited RuRe to the CuGaO 2 surface, which could act as an active site of H 2 evolution, might induce the production of H 2 . Detailed studies of the reaction mechanism of the production of H 2 and its suppression are in progress in our laboratory. Fig. 6 shows the time courses of the photocurrent using the RuRe/CuGaO 2 photocathode at the potentials of 0.1 V, 0.3 V, and 0.7 V vs. Ag/AgCl under irradiation at l ex > 460 nm, which correspond to entries 1-3 in Table 1. In all of the cases, a photocurrent was observed even after irradiation for 15 h. However, the photocurrent decreased in the frst stage of the photoelectrochemical reaction, and the rates of formation of CO and H 2 also became slower in accordance with the decline in the photocurrent (Fig. 4 for the case at 0.3 V; Fig. S6 † for the other cases). To confrm the reasons for the decline in the photocurrent and the rates of product formation, an estimation of the amount of "electrochemically active" RuRe species on the CuGaO 2 surface was conducted using cyclic voltammetry in an Fig. 5 GC-MS chromatograms of the gas products in the reaction chamber after irradiation (the detected m/z ¼ 28 for 12 CO and 29 for 13 CO). The RuRe/CuGaO 2 photocathode was irradiated at l ex > 460 nm in an aqueous solution containing 50 mM sodium bicarbonate (pH 6.6) at 0.7 V vs. Ag/AgCl for 5 h using 13 acetonitrile solution (Fig. S7 and Table S1 †) according to reported procedures, where the areas of the oxidation peaks of Ru II /Ru III were used for the estimation. 32 This indicated that about 80% of the RuRe lost its electrochemical activity after reacting for 15 h, which was probably due to (photo)desorption and/or photodecomposition of RuRe and could account for the decline in the photocurrent and the photocatalytic activity. In addition, the difference between the peak potentials of the redox reaction of Ru II /Ru III was greater after the reaction (Fig. S7 †), which indicates that the ohmic resistance of the electrode increased. Because the XPS measurements of the electrodes showed no obvious change in the electronic states of the Cu and Ga on the surface of the CuGaO 2 particles during the reaction (Fig. S8 †), the increase in the internal resistance of the CuGaO 2 electrode might not be caused by the collapse of the GuGaO 2 surface. An increase in inter-particle resistance probably proceeded in the electrode owing to the gradual deterioration of physical contact among the GuGaO 2 particles. This might induce the deactivation of the photocathode. We note that the peaks of Ru and Re in the XPS measurements were too small to be identifed, possibly due to the low loading density of RuRe on CuGaO 2 . ## Photoelectrochemical reduction of CO 2 using a hybrid electrochemical cell with a Z-scheme conguration To couple the photoelectrochemical reduction of CO 2 to the oxidation of water, the CoO x /TaON photoanode, which has been reported to be an efficient photoanode for the water oxidation reaction, was examined as the counter photoanode in a full photoelectrochemical cell using the RuRe/CuGaO 2 hybrid photocathode. The CoO x /TaON electrode was prepared according to a previously reported method. 46 Fig. 7 shows the currentpotential curve of the CoO x /TaON photoanode in a 50 mM aqueous solution of NaHCO 3 with a pH of 6.6 that was purged with CO 2 , which is the same electrolyte as was used for the photocathode. An anodic photocurrent was clearly observed, as in the cases of the aqueous solutions of Na 2 SO 4 (pH 8) and NaHCO 3 (pH 8.3), that were both purged with Ar. 46 In addition, the evolution of O 2 under visible-light irradiation in a 50 mM aqueous solution of NaHCO 3 purged with CO 2 was confrmed by the analysis of the gas products at an applied potential of +0.2 V vs. Ag/AgCl, and its total faradaic efficiency was 93% after irradiating for 60 min (Fig. S9 †). Fig. 8 shows a schematic representation of the hybrid photoelectrochemical cell consisting of the RuRe/CuGaO 2 photocathode and the CoO x /TaON photoanode in a Z-scheme confguration. Two chambers containing the electrolyte solution purged with CO 2 were separated by a Nafon membrane, and one photoelectrode was installed in each chamber. Visible light with l ex > 460 nm and l ex > 400 nm was used to irradiate the RuRe/CuGaO 2 photocathode and the CoO x /TaON photoanode, respectively. It is worth nothing that the photoelectrochemical reaction was conducted under short-circuit conditions with the same electrolyte for both electrodes, i.e., there was no external electrical or chemical bias between the electrodes. In addition, a higher intensity of irradiated light was utilized for the photocathode (4.5 10 17 photon cm 2 s 1 in the wavelength range 460-600 nm) than that for the photoanode (4.1 10 16 photon cm 2 s 1 in the wavelength range 400-600 nm) to achieve sufficient photoexcitation of the loaded RuRe on the photocathode. Photoelectrolysis using the constructed photoelectrochemical cell was conducted with intermittent visible-Fig. 6 Time courses of the photocurrent using the RuRe/CuGaO 2 photocathode at various potentials (corresponding to entries 1-3 in Table 1) under irradiation at l ex > 460 nm. light irradiation for a total of 2 h. Fig. 9 shows the time courses of the photocurrent and electrode potentials (a) and the chemical products (b). The generation of a photocurrent and a corresponding increase in the products were clearly observed under irradiation, and the working potential of the electrodes was confrmed to be around +0.15 V vs. Ag/AgCl, at which point both of the photoelectrodes could generate a photocurrent under irradiation via the respective half reactions. The products of the reduction (CO and H 2 ) and oxidation (O 2 ) were detected in the cathode and anode chambers, respectively (Fig. S10 †). A total of 232 nmol of CO was produced, and the TON CO reached 22 based on the amount of immobilized RuRe before the irradiation. The total faradaic efficiency of the cathodic reaction, including the evolution of H 2 (311 nmol), was 72%, whereas that of the anodic reaction was 70% (266 nmol O 2 ). Notably, an electron balance between the reduction products (CO and H 2 ) and the oxidation product (O 2 ) was almost acquired in this photoelectrochemical reaction, while in the case of the corresponding system using the NiO electrode instead of CuGaO 2 , there were much less reduction products than the produced O 2 . 32 The slight decrease in the half amount of electrons when the light was cut in Fig. 9(b) means that a reverse current flowed, which was possibly derived from the discharge of the electrical double layers of the electrodes. When only the cathode or the anode was irradiated, the photocurrent was relatively small in comparison to that when both of the electrodes were under irradiation (Fig. S11 †), which suggests that photoexcitation of both of the electrodes is required for the subsequent progress of the photochemical reduction of CO 2 with water as the reductant. On the basis of these data, we conclude that this hybrid photoelectrochemical cell can drive both the reduction of CO 2 and the oxidation of water by the stepwise photoexcitation of RuRe and TaON, i.e., via Z-scheme-type electron transfer, which gives CO and O 2 under visible-light irradiation. This is the frst successful instance of the reduction of CO 2 in the absence of electrical or chemical bias using visible light and water alone by a hybrid photocatalytic system consisting of a molecular photocatalyst and a semiconductor photocatalyst. Further investigations for the improvement of the efficiency and durability of the photoelectrochemical cell are currently under way in our laboratory. ## Synthesis of CuGaO 2 particles Particles of CuGaO 2 powder were prepared using a solid-state reaction method. The precursor materials, i.e., Cu 2 O (Kanto Chemicals, >92%) and Ga 2 O 3 (Wako Chemicals, 99.99%), were mixed using an agate mortar. The molar ratio of the precursors was set at Cu : Ga ¼ 1 : 1. The obtained mixture was calcined with a N 2 flow of 100 mL min 1 at 1373 K for 5 h to give a polycrystalline powder of CuGaO 2 . ## Fabrication of CuGaO 2 electrodes Fluorine-doped tin oxide (FTO) glass (AGC Fabritech, 15 50 mm, 12 U sq 1 ) was sonicated in acetone for 10 min and in methanol for 10 min before use. A suspension of 30 mg of CuGaO 2 powder in 300 mL of isopropanol was prepared with sonication for 10 min. Then, 100 mL of the suspension was dropped onto an exposed area of the FTO glass (ca. 5 cm 2 ) and dried in air at room temperature using a drop-casting method and Scotch mending tape as a masking material. The obtained sample was calcined at 773 K for 3 h with a N 2 flow of 100 mL min 1 . The electrode was cut in half before use. The amount of powder that was deposited was estimated to be approx. 5 mg for each electrode (ca. 2.5 cm 2 ). Preparation of the hybrid photocathode by the adsorption of the metal complex on the CuGaO 2 electrode RuRe and its model complexes (Ru and Re) (Chart 1) were synthesized in accordance with reported procedures. 43 A CuGaO 2 electrode was immersed in an acetonitrile solution containing the metal complex (4 mL, 5 mM) overnight. The electrode was washed with acetonitrile after the adsorption procedure. The amount of the metal complex that was adsorbed was estimated from the difference in the absorbance between the solutions at 461 nm before and after the hybridization procedures. ## Preparation of the CoO x /TaON photoanode The CoO x /TaON electrode was prepared in accordance with a reported procedure. 46 The outline of the procedure is as follows. TaON powder was prepared by heating Ta 2 O 5 powder with an NH 3 flow (20 mL min 1 ) at 1123 K for 15 h. CoO x nanoparticles (5 wt% as the metal) were loaded onto the TaON particles via impregnation from an aqueous Co(NO 3 ) 2 solution, followed by heating at 673 K for 30 min in air. The as-prepared CoO x -loaded TaON particles were deposited on a Ti substrate using electrophoretic deposition. The electrodes were treated with 50 mL of a solution of TaCl 5 in methanol (10 mM) and then dried in air at room temperature. After this process had been performed fve times, the electrode was heated in an NH 3 flow (10 mL min 1 ) at 723 K for 30 min. The details of the structural characterization of the electrode are shown in our previous paper. ## View Article Online Photoelectrochemical measurements and reduction of CO 2 using the RuRe/CuGaO 2 photocathode A three-electrode setup with an HZ-7000 potentiostat (Hokuto Denko) was used throughout the photoelectrochemical measurements and half reactions. A 50 mM aqueous solution of NaHCO 3 saturated with CO 2 (pH 6.6) was used as an electrolyte. A Pt wire and Ag/AgCl in a saturated aqueous solution of KCl were employed as the counter and reference electrodes, respectively. The counter electrode was separated from the reaction solution using Vycor glass to avoid the influence of the oxidation reaction taking place on the counter electrode. The working electrode (ca. 2.5 cm 2 ) was irradiated using a 300 W Xe lamp (Asahi Spectra MAX-302) with an IR-blocking mirror module. A cutoff flter (HOYA Y48 for irradiation at l ex > 460 nm) or a band-pass flter (Asahi Spectra, for the IPCE measurements) was employed to control the irradiation wavelength. The potential against a reversible hydrogen electrode (RHE) (see Fig. 1) was calculated using the Nernst equation (eqn (1)): The dependence on the wavelength of the IPCE at a wavelength was calculated using eqn (2): where l ex and P light are the wavelength and power density of the incident light (nm and mW cm 2 ), respectively, I light is the current density under irradiation (mA cm 2 ), and I dark is the current density in the dark (mA cm 2 ). The photoelectrochemical reduction of CO 2 was conducted using a Pyrex cell sealed with an O-ring and a stirring tip. The total volume of the cell without the stirring tip and electrodes was ca. 130 mL, and 84 mL of the electrolyte solution was utilized; therefore, the estimated volume of the gas phase was ca. 46 mL. The reaction was conducted after purging the system with CO 2 for more than 40 min. A cutoff flter (HOYA Y48) was employed for irradiation at l ex > 460 nm. Product analyses of CO and H 2 in the gas phase and HCOOH in the liquid phase were performed using a gas chromatograph (Infcon MGC3000A) and a Capi-3300I capillary electrophoresis system (Otsuka Electronics), respectively. 13 CO 2 (CIL, 13 C ¼ 99%) and NaH 13 CO 3 (Aldrich, 13 C ¼ 99%) were utilized for the labeling experiments. Gas chromatography-mass spectrometry (GC-MS) analysis of the gas phase was conducted using a gas chromatograph equipped with a mass spectrometer (Shimadzu GCMS-QP2010 Ultra). Photoelectrochemical reduction of CO 2 using the hybrid electrochemical cell in the Z-scheme confguration The reaction was conducted using a Pyrex cell with two chambers (ca. 27 mL for each chamber) that were divided by a Nafon 117 membrane (Aldrich), and the photoelectrodes were installed into each chamber and the cell was sealed with an Oring. Ag/AgCl in a saturated aqueous solution of KCl was employed as the reference electrode and was placed into the cathode chamber. A 14.5 mL portion of a 50 mM aqueous solution of NaHCO 3 was added as an electrolyte to each chamber, which was purged with CO 2 . A three-electrode setup with an HZ-7000 potentiostat (Hokuto Denko) was used in nonresistance ammeter mode. The RuRe/CuGaO 2 photocathode (ca. 3.2 cm 2 , RuRe 10.5 nmol) was irradiated at l ex > 460 nm using a 300 W Xe lamp (Eagle Engineering) with a cutoff flter (HOYA Y48) and a cold mirror (CM-1). The CoO x /TaON photoanode (ca. 3.3 cm 2 ) was irradiated at l ex > 400 nm using a 300 W Xe lamp (Asahi Spectra MAX-302) with a cutoff flter (HOYA L42) and an IR-blocking mirror module. The analysis of gas products was performed using gas chromatography (Infcon MGC3000A). ## Characterizations UV-vis diffuse reflectance spectra were recorded using a V-565 spectrophotometer (JASCO) that was equipped with an integral sphere unit. X-ray diffraction (XRD) patterns were recorded using a MiniFlex600 X-ray diffractometer (Rigaku) that was equipped with monochromatic Cu Ka radiation and operated at 15 mA and 40 kV. The scanning electron microscopy (SEM) observations were conducted using an S-4700 microscope (Hitachi High-Tech). An ECSA-3400 X-ray photoelectron spectrometer (Shimadzu) was utilized for the X-ray photoelectron spectroscopy (XPS) measurements. The binding energies were corrected using the C 1s peak (284.6 eV) for each sample. ## Conclusions A novel hybrid photocathode consisting of a CuGaO 2 p-type semiconductor and a RuRe supramolecular complex has been developed for the photoelectrochemical reduction of CO 2 in aqueous solution. The RuRe/CuGaO 2 photocathode displayed photoelectrochemical activity for the conversion of CO 2 to CO in an aqueous electrolyte solution with an onset potential of +0.3 V vs. Ag/AgCl. A photoelectrochemical cell comprising the RuRe/ CuGaO 2 photocathode and a CoO x /TaON photoanode exhibited activity for the visible-light-driven reduction of CO 2 using water as a reductant to generate CO, H 2 , and O 2 with no external bias.

chemsum

{"title": "Hybrid photocathode consisting of a CuGaO₂ p-type semiconductor and a Ru(<scp>ii</scp>)\u2013Re(<scp>i</scp>) supramolecular photocatalyst: non-biased visible-light-driven CO₂ reduction with water oxidation", "journal": "Royal Society of Chemistry (RSC)"}

taming_ros:_mitochondria-targeted_aiegen_for_neuron_protection_via_photosensitization-triggered_auto

## Abstract: Oxidative damages lead to accumulated harmful wastes, which in turn aggravate the related diseases and ROS imbalance. Therefore, provoking the defense system against severe oxidation and maintaining ROS homeostasis are desired. Herein, we used a mitochondria-targeted aggregationinduced emission luminogen (AIEgen) as a phototherapy agent for neuron protection by virtue of its efficient ROS generation in aggregates and mitochondrial delivery. It is demonstrated that controllable ROS generation within mitochondria can trigger defensive autophagy against oxidative damages in neuron cells. This work not only verifies the concept that taming ROS can be used for cell protection, but also provides a promising method to trigger autophagy against destructive oxidation, displaying broad prospects for alleviating oxidation-related diseases and promoting cell-based therapy. Life is a contradictory organism full of dynamic balance, such as immune homeostasis, energy equilibrium, reactive oxygen species (ROS) balance, etc. The balance of these dynamic systems is finely tuned in a normal range to keep fit. Among various dynamic systems, ROS balance has drawn great attention due to its critical role and extensive participation in life and disease process. For example, ROS with normal level is closely involved in cell differentiation and development, whereas too much ROS will lead to superfluous oxidized substances, which in turn exacerbate the ROS imbalance and elicit many severe diseases (e.g., cancer, inflammation, neurodegenerative diseases and ischemia-reperfusion injury). Thereby, developing agents or methods to maintain the ROS homeostasis and timely clean the harmful wastes is highly desired. At present, antioxidants (such as N-Acety-L-Cysteine (NAC), Vitamin C, Vitamin E, etc.) are commonly used to maintain the ROS balance. 7 Although antioxidants can quickly scavenge redundant ROS, they can neither remove the accumulated harmful wastes nor generate durable protection. In contrast, autophagy inducers (e.g., rapamycin, resveratrol and many natural compounds) can improve the level of autophagy, which enables cell to digest harmful substances via cell's own defense system to achieve a long-term protection. 8 However, most autophagy inducers belong to chemotherapy agents, and the lack of specific and controllable delivery and visual tracking contribute to some common drawbacks, such as multiple targets, organ toxicity, and elusive mechanism, limiting their further application. Compared to chemotherapy, phototherapy owns inherent advantages of high selectivity and super temporal-spatial resolution, exhibiting promising prospect for specific delivery and accurate treatment. Photodynamic therapy (PDT) is a successful demonstration of phototherapy in precise treatment of diseases and is easy to achieve localized delivery of massive ROS via engineering control of light. Inspired by this, we wonder whether phototherapy can be used to trigger cell's own protective system to remove harmful oxidized components and maintain the ROS balance. According to literatures, moderate ROS derived from mitochondria is crucial for stimulating the cell's defense system, especially for autophagy to remove harmful wastes. Thereby, provoking cell's defensive autophagy via photo-controlled ROS generation within mitochondria was proposed. Traditional photosensitizers (PS) usually possess conjugated planar structures and they are easy to form aggregates in physiological environment. After excitation, these PS will predominantly consume energy through heat production due to strong intermolecular π-π interactions in the aggregate state, leading to reduced ROS generation as well as fluorescence emission. Therefore, it's hard to accurately and effectively control ROS generation. 25 Contrary to traditional PS, AIEgenbased photosensitizers (AIE-PS) are able to effectively produce ROS and fluorescence in aggregates thanks to the twisted structures, which strongly inhibits the thermal decay. Thus, we can conveniently control ROS at different amount and exact location via AIE-PS based phototherapy. Based on these considerations, two mitochondria-targeted AIE-PS (DTCSPY and DTCSPE) with twisted structures were designed. As both intermolecular interactions (e.g., π-π interaction) and intramolecular motions are major pathway to consume energy by heat generation, comparatively rigid alkyne group was introduced into the structure of DTCSPY to further hinder thermal dissipation through intramolecular motions. For comparison, a more flexible alkene-based πbridge was introduced into DTCSPE. As expected, the ROS generation efficiency of DTCSPY is much better than that of DTCSPE both in PBS solutions and live cells. Particularly, taming DICSPY-generated ROS for cell protection against severe oxidation via photosensitization-triggered autophagy was confirmed. The experimental results also showed that AIE-PS exhibited a better protective effect than the widely used antioxidants (NAC and Vitamin C). DTCSPY and DTCSPE were synthesized through Knoevenagel condensation and methylation as shown in Figure 1A and Scheme S1. The chemical structures of the two molecules were fully characterized by 1 H NMR, 13 C NMR, and HRMS. DTCSPY and DTCSPE share the same electron donating and accepting skeleton but different π-bridges. Pyridinium was introduced as mitochondria-targeted moiety. Then photophysical properties of DTCSPY and DTCSPE were investigated. As shown in Figure 1B and Figure S2, the absorption maxima of DTCSPY and DTCSPE were 490 nm, and the emission spectra were centered at 700 nm and 710 nm of DTCSPY and DTCSPE in DMSO/water mixtures, respectively. Figure 1C and D showed that both of them exhibited typical AIE activity and remarkable ROS generation capability much better than rose bengal (a well-known PS) under the same condition. Furthermore, DTCSPY decorated with rigid alkyne-based πbridge displayed better performance in both AIE activity and ROS production capability than that of DTCSPE featured with flexible alkenebased π-bridge (Figure 1C, D, and S3). These results indicated that rigid alkyne in AIEgens could further improve ROS generation ability by hindering thermal dissipation from intramolecular motions. Theoretical calculation was conducted to understand the relationship between chemical structures and different ROS generation performance by density functional theory (DFT) at the level of CAM-B3LYP/6-31G(d,p). 33 Optimized geometries and electron cloud distribution separately depicted comparable large dihedral angle of stilbene segment ( > 21°) and relatively distant HOMO (highest occupied molecular orbital)-LUMO (lowest unoccupied molecular orbital) separation in both AIE-PS (Figure 2A), which contributed to twisted structures, AIE activity and efficient ROS generation in aggregates. To further explain the different ROS performance of DTCSPY and DTCSPE, the energy gap (E) and spin-orbital coupling (SOC) between singlet and triplet states were analyzed. 34 Although the results showed almost the same level of SOC efficiency of the two compounds, the E between S1 and T2 state of DTCSPY is much smaller than that of DTCSPE, which is mainly responsible for its higher ROS generation efficiency (Figure 2B). From the results, HT22 cells themselves cannot oxidize the DCF-DA to highly green-emissive DCF (Figure 3D), and the cells without light irradiation (Figure 3E) or AIE-PS (Figure 3F) can also hardly be collected DCF fluoresce signals. In contrast, after treatment with AIE-PS (DTCSPY and DTCSPE) and light irradiation, significant signals could be observed, indicating that efficient ROS was generated in HT22 neuron cells (Figure 3G-H). Figure 3I showed that NAC (a ROS scavenger) could suppress the strong signals, further demonstrating that it was indeed the ROS generated by AIE-PS that oxidized DCF-DA to DCF. Additionally, compared with DTCSPE-treated group, DTCSPY-treated cells exhibited much stronger fluorescence, displaying more efficient ROS generation, which was consistent with the results of theoretical calculation and in vitro tests in PBS solutions. After verifying the mitochondria-targeted and ROS generation capabilities of AIE-PS in HT22 neuron cells, we examined whether controllable ROS generation within mitochondria can trigger a defensive autophagy against oxidative damage. Firstly, we confirmed that both DTCSPY and DTCSPE owned good biocompatibility in darkness by MTT [3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay (Figure 4A). Then, we optimized experimental conditions to control the amount of the generated ROS via treating the cells with AIE-PS in different concentrations under light irradiation. As displayed in Figure 4B, when incubated concentrations of DTCSPY and DTCSPE are lower than or equal to 0.5 μM and 2.5 μM, respectively, the cell viability of HT22 neuron cells was higher than 85%, indicating that the amount of ROS generated within these ranges (0-0.5 μM for DTCSPY and 0-2.5 μM for DTCSPE) was tolerable in neuron cells. Next, for comparison, the concentration of 0.5 μM was selected for DTCSPY and DTCSPE to explore the protective effect by adjusting the irradiation time. Briefly, HT22 neuron cells were incubated with 0.5 μM DTCSPY and DTCSPE for 90 min, then treated with light irradiation for different time (0, 1, 3, and 5 min). After further 6 h incubation, the cells were subsequently exposed to 100 μM of hydrogen peroxide (H2O2) for 1 h and finally measured at 48 h. As shown in Figure 4C, the protective effect of neuron cells increased with prolonged irradiation time, suggesting that the amount of ROS generated by AIE-PS needs to reach a certain level to trigger cell's defense system. Compared with DTCSPE-treated cells, DTCSPY-treated cells exhibited much stronger protective effect as DTCSPY was easier to reach the ROS threshold of provoking defense system owing to its more efficient ROS generation upon light irradiation. Moreover, bright field pictures and CLSM images also described the differences of healthy state and quantity of live or dead cells between DTCSPY-treated group and control group, respectively (Figure S5), which were consistent with the protective effect in Figure 4C. Finally, both NAC and Vitamin C, the widely used agents for resisting oxidative damages, were selected for comparison. As shown in Figure 4D, DTCSPY can trigger a more effective protection against (H2O2) than NAC and Vitamin C, demonstrating the importance of long-term defense triggered by DTCSPY. In addition, in CQ (chloroquine, an autophagy inhibitor)-added group, the protective effect of DTCSPY treatment was lost, suggesting this protection was mediated by autophagy. Further experiments were conducted to confirm that the protective effect was definitely from DTCSPY-triggered autophagy at this optimized condition. As the punctate distribution of LC3B is one of the best hallmarks of autophagy induction, immunofluorescence staining of LC3B were analyzed using CLSM. 37 As shown in Figure 5A, "Control" group without any treatment and "DTCSPY" group without light irradiation exhibited diffused distribution of LC3B, while massive LC3B puncta were observed in "DTCSPY + Light" group at 6 h after irradiation, which indicated that effective autophagy was induced by DTCSPY-generated ROS. Quantified results also revealed a 6.4-fold and a 5.0-fold increase of LC3B puncta in "DTCSPY + Light" group than that of "Control" and "DTCSPY" groups. Next, the LC3B conversion from full-length LC3B-I to LC3B-II was detected at 6 h by western blots, as it is another important hallmark of autophagy. As depicted in Figure 5B and S6, the ratio of LC3B-II / LC3B-I in "DTCSPY + Light" group was 1.8-fold higher than that of "Control" and "DTCSPY" group, suggesting activated autophagy in experimental group. Moreover, two other necessary proteins (ATG5 and Beclin 1) involved in autophagy were also elevated in "DTCSPY + Light" group compared with the other two groups. 38 These results together verified that controlling the ROS generation within mitochondria using DTCSPY can effectively trigger autophagy, which contributed to the protective effect of neuron cells from severely oxidative damage. In summary, two highly effective and mitochondria-targeted AIE-PS (DTCSPY and DTCSPE) were designed and synthesized. DTCSPY constructed with alkyne-containing πbridge, exhibited better performance than DTCSPE featured with alkane-containing πbridge in AIE activity and ROS generation efficiency due to less thermal dissipation in relatively rigid skeleton. HT22 neuron cells were used as a model to demonstrate the concept that taming ROS generation can trigger cell protection against severely oxidative damages. To the best of our knowledge, this demonstration is the first attempt to trigger neuron protection by control of ROS via phototherapy. Besides, we have applied this method to protect mesenchymal stem cells from oxidation-induced aging and death. Not only does it provide a useful strategy against oxidative damages caused by acute inflammation, neurodegenerative diseases, ischemia-reperfusion injury, serious side-effect of radiotherapy, and chemotherapy, but also exhibits great potential to improve cell survival in oxidative environment during stem-cell transplantation and chimeric antigen receptor T (CAR-T)-cell based therapy.  ASSOCIATED CONTENT Supporting Information. The Supporting Information is available free of charge on the ACS Publications website at DOI. Experimental procedures, chemical synthesis and characterization of compounds, and Supplementary figures.

chemsum

{"title": "Taming ROS: Mitochondria-Targeted AIEgen for Neuron Protection via Photosensitization-Triggered Autophagy", "journal": "ChemRxiv"}

acid-durable_electride_with_layered_ruthenium_for_ammonia_synthesis:_boosting_the_activity_<i>via</i

## Abstract: Ruthenium (Ru) loaded catalysts are of significant interest for ammonia synthesis under mild reaction conditions. The B 5 sites have been reported as the active sites for ammonia formation, i.e., Ru with other coordinations were inactive, which has limited the utilization efficiency of Ru metal. The implantation of Ru into intermetallic compounds is considered to be a promising approach to tune the catalytic activity and utilization efficiency of Ru. Here we report an acid-durable electride, LnRuSi (Ln ¼ La, Ce, Pr and Nd), as a B 5 -site-free Ru catalyst. The active Ru plane with a negative charge is selectively exposed by chemical etching using disodium dihydrogen ethylenediaminetetraacetate (EDTA-2Na) acid, which leads to 2-4-fold enhancement in the ammonia formation rate compared with that of the original catalyst.The turnover frequency (TOF) of LnRuSi is estimated to be approximately 0.06 s À1 , which is 600 times higher than that of pure Ru powder. Density functional theory (DFT) calculations revealed that the dissociation of N 2 occurs easily on the exposed Ru plane of LaRuSi. This systematic study provides firm evidence that layered Ru with a negative charge in LnRuSi is a new type of active site that differs significantly from B 5 sites. ## Introduction Ammonia synthesis is one of the most important processes in industry. Ruthenium catalysts are of particular interest for ammonia synthesis and have been widely investigated as they function under mild conditions and exhibit a tenfold increase in the rate of ammonia synthesis compared with the conventional promoted iron catalyst. The energy barrier for N 2 dissociation has been decreased with promotion by alkali/alkali earth metal oxides, which have electron donating ability, and this has led to signifcant enhancements in ammonia formation rates. Therefore, a support or promoter with stronger electron donation ability may further enhance the activity of Ru catalysts. This has been achieved using electride materials such as C12A7:e and Y 5 Si 3 :e , where the electrons are localized in the cages/cavities as anions and can migrate onto the supported Ru to decrease the N 2 dissociation barrier. For instance, the activation energy of Ru/C12A7:e for ammonia synthesis is 49 kJ mol 1 , which is much lower than that of the conventional Ru-Cs/MgO catalyst (120 kJ mol 1 ), and the turnover frequency (TOF; 0.2 s 1 ) is one order of magnitude higher than that of the Ru-Cs/MgO catalyst (0.008 s 1 ). 9,14 The anionic electron in electride easily traps dissociated H atoms at the same site to form hydride ions (H ) with the reversible reaction of e + H 4 H , which suppresses hydrogen poisoning of the Ru surfaces that always occurs with conventional Ru based catalysts. 14 However, various investigations reveal that the active sites of Ru particles consist of fve Ru atoms exposing a threefold hollow hexagonal close-packed (hcp) site and a bridge site close together, i.e., a B 5 -type site, where part of the atoms are edge/ step atoms. Both experimental results and theoretical calculations have shown that the barrier for N 2 dissociation at a step site (B 5 site) of a Ru (0001) face is approximately 0.4 eV, which is much lower than that at a terrace site (ca. 1.8-1.9 eV). Thus, the dissociation rate of N 2 at steps is at least 9 orders of magnitude higher than that on the terraces at 500 K. The number of B 5 sites is sensitive to the particle size; the optimized Ru particle size for ammonia synthesis is in the range of 1.8-3.5 nm. As a result, the fraction of B 5 sites in total Ru atoms is rather small, which indicates that most Ru atoms are not active and are unused for the ammonia synthesis. A B 5 -site-independent Ru catalyst is thus desirable to improve the utilization rate of Ru. Transition-metal (TM)-containing intermetallic compounds are promising candidates to realize high activity and high utilization of active TMs for various heterogeneous reactions, including ammonia synthesis. With active TM atoms implanted into the lattice of an intermetallic compound, both the geometric confguration and charge state of TM atoms are modifed via the crystal structure and chemical bonding effects, and the stability of the catalyst would be improved by preventing the aggregation of active sites. For instance, intermetallic LaCoSi with negatively charged Co has shown good activity and stability for ammonia synthesis. 21 Ru exhibits higher activity for ammonia synthesis than Co, which suggests that Ru-containing intermetallics may also exhibit higher activity than LaCoSi, which was confrmed by the LaRuSi electride and YRu 2 catalyst. 26,27 These studies are strongly indicative of the possibility to realize active Ru sites other than B 5 sites through the use of intermetallics. Following this suggestion, not only is the space for the exploration of new active catalysts expanded, but the approaches to improve catalytic activity will also be enriched. Herein, we propose LnRuSi with layered Ru as a new type catalyst that differs from conventional Ru catalysts with B 5 sites as active centers for ammonia synthesis. These intermetallic electrides are robust in terms of stability, even against some acids such as HNO 3 and EDTA, in contrast to conventional electrides, which are unstable in the presence of air and moisture. According to the acid durability of these electrides, we have developed a simple selective etching method using EDTA-2Na solution to remove surface Ln and Si, which enables the exposure rate of active Ru sites to be enhanced. The catalytic activity of etched LnRuSi toward ammonia synthesis is enhanced 2-to 4-fold compared with that of the original sample. ## Results and discussion LnRuSi electrides crystallize in a tetragonal structure with the space group P4/nmm (Fig. 1a), where Ru and Si layers are separated and sandwiched by double layers of Ln. The Ln layers are composed of edge-shared Ln 4 tetrahedra with a vacant space (V site) located at the center (depicted as light blue balls in Fig. 1a) that can accommodate an electron or hydrogen anion. 27 The crystal quality of LnRuSi was confrmed by Rietveld refnements of the obtained powder XRD data (Fig. 1b and S1 †). The 111 target phases with purity higher than 95 wt% were obtained for LaRuSi, CeRuSi and PrRuSi samples; however, the purity of NdRuSi was only as high as 83.5 wt%, and the 122 impurity phase was difficult to remove completely, even after long-time annealing. It is noteworthy that the XRD patterns show that the (00l)-orientation is favored for all ground LnRuSi powders, which indicates that the exposed surfaces of the present catalysts are dominated by (00l)-Ln/Ru/Si layers. This was confrmed by SEM measurements, as shown in Fig. S2, † where clear layered structures were observed for all LnRuSi compounds. The cleavage energy between different layers along the z direction was calculated for LaRuSi using DFT to determine the most possible exposed surface. As shown in Fig. S3, † the cleavage energy between La-La layers (0.051 eV 1 ) was much lower than that between La-Si layers (0.211 eV 1 ) and Ru-Si layers (0.312 eV 1 ). Therefore, cleavage between the La-La layers is most likely to occur during the milling process, i.e., (00l)-La terminated surfaces should be the major exposed surfaces for the hand-milled powder of LaRuSi, and accordingly, for other LnRuSi compounds. The catalytic performance for ammonia synthesis over the LnRuSi electride catalysts is shown in Fig. 2 and summarized in Table S2. † The LnRuSi catalysts exhibited much higher activity than pure Ru powder with a similar surface area (as high as 30-80 times), and were comparable to conventional iron or ruthenium based catalysts. For instance, the ammonia yield over LaRuSi was 0.492%, which was much higher than that over commercial Fe-K 2 O-Al 2 O 3 (0.164%) under same operation condition. 31 On the other hand, no ammonia formation was observed over the Ru-free compounds, including the LaScSi electride, LaSi and La 5 Si 3 (Table S3 †), which indicates that lattice Ru atoms in LaRuSi are the active sites for ammonia synthesis. Moreover, present LnRuSi electrides are robust to air and water. The ammonia synthesis over LaRuSi catalysts that were stored in air for half year and treated with water for 24 h. The ammonia formation rate for these samples were 1800 and 1760 mmol g cat 1 h 1 , respectively, almost identical to that for the as prepared catalyst of 1810 mmol g cat 1 h 1 (Table S4 †). The XRD measurements were performed for these samples, and we did not fnd any clear phase change (Fig. S4 †). Moreover, even after the used catalyst is stored in air for 1 year, the catalytic activity is not decreased (1780 mmol g 1 h 1 ) without any phase change. These observations demonstrate the excellent stability of the present catalyst. An attempt was made to increase the surface area of the LnRuSi catalysts by vigorous milling for longer times to enhance the catalytic activity. For the LaRuSi catalyst, the surface area was increased from 1.0 to 1.4 and 1.8 m 2 g 1 by increasing the milling time from 0.5 to 1.0 and 2.0 h, respectively. However, the ammonia formation rates for these two samples were almost the same as that for the low surface area sample (Fig. 2a, blue symbols and Table S2 †). These results suggest that the exposed Ru active sites did not change as the surface area increased. The La-La layer is most likely to be cleaved (Fig. S3 †); therefore, longer milling times should result in an increase of the inactive La exposed layer rather than desired Ru layers. Therefore, although the surface area was increased, the exposure of Ru active sites was not signifcantly enhanced (shown schematically in Fig. S5a †), which resulted no signifcant change of the ammonia synthesis rates. Chemical etching is one of the most powerful routes to modify the surface structure of intermetallics. For example, selective removal of Gd atoms on the surface layers was reported by etching intermetallic Pt 5 Gd in HClO 4 solution, which led to a 5-fold increase in the oxygen reduction reaction relative to pure Pt. 32 We propose that selective chemical etching with acid to remove the surface Ln and Si of LnRuSi could increase the exposure of Ru sites, which would enhance the activity for ammonia synthesis (Fig. S5b †). LaRuSi was used to determine the optimal conditions for chemical etching (Table 1). The results showed that HNO 3 acid did not signifcantly change either the activity or surface area. XRD and SEM measurements (Fig. 3) showed that there was no phase change during HNO 3 treatment. This is mainly due to the oxidation ability of HNO 3 , which may passivate the surface by the formation of oxides. Highly concentrated HCl acid reacted with LaRuSi rapidly; La was completely dissolved and amorphous Ru-Si and H 2 were formed as the products (Fig. S6 and S7 †). Although with a high surface area, the amorphous Ru-Si obtained had low catalytic activity. With diluted HCl acid (0.02 M), La could be partially removed, but the reaction was still very strong, whereby H 2 was rapidly formed, and the catalytic activity was only slightly enhanced. The H 2 formation rate was decreased signifcantly using 50% HCOOH solution, and the ammonia formation rate for the resultant sample was increased from 1100 to 1966 mmol g cat 1 h 1 . These results indicate that control of the etching strength is a key factor to optimize exposure of the Ru sites. EDTA is a well-known chelating acid for fxing Ca 2+ , Ga 3+ and La 3+ ions; therefore, EDTA-2Na solution was used for the selective etching of LaRuSi. XRD measurement shows that EDTA treated LaRuSi maintained the original phase with the minor formation of hydride, while the surface was etched (see the SEM images Fig. 3 and S9 †). The concentration of the EDTA-2Na solution and the etching time were changed, and the activity of the EDTA-treated LaRuSi increased linearly with the surface area, unlike the sample not treated with EDTA (Fig. 2a), which demonstrates that the present etching method is a promising route to enhance ammonia synthesis activity, and suggests that higher ammonia formation rates can be achieved when a fner catalyst powder is prepared. The optimized condition for etching was: 0.2 g LaRuSi powder with 5 mM EDTA-2Na (10 mL) for 5 h (Fig. S8, † etching time longer than 5 h did not further change the specifc surface area and activity). This treatment resulted in ammonia formation rates for LaRuSi at steady state of 3020 and 5340 mmol g cat 1 h 1 at 340 C and 400 C, respectively (Table 2), which were approximately 3 times higher than that of the original catalyst. These reaction rates were much higher than other unloaded catalysts, such as LaCoSi, 21 Co 3 Mo 3 N 31 and YRu 2 , 26 and even higher than that over the Ru/C12A7:e catalyst. 14 The same EDTA treatment process (0.2 g of catalyst placed into 10 mL of 5 mM EDTA-2Na solution for 5 h) was performed for the other LnRuSi catalysts and the etching of surfaces were confrmed by SEM measurement (Fig. S10-S12 †). As shown in Fig. 2b and Table 2, the ammonia formation rates of all the catalysts were increased 2 to 4-fold by the EDTA treatment. The TOF values for LnRuSi were as high as ca. 0.06 s 1 at 400 C, which is 600 times higher than that for pure Ru powder and much higher than other reported unloaded catalysts and even the Cs-Ru/MgO catalyst, which demonstrates the superior performance of the present catalysts. Fig. 2c shows the time course for NH 3 formation over LaRuSi before and after treatment with EDTA. The ammonia formation rate increased initially and then became stable for the untreated sample, while it was steady for the EDTA-treated sample. This indicated that no aggregation and/or leaching occurred during ammonia synthesis. The temperature dependence was also measured for all the LnRuSi catalysts before and after EDTA treatment (Table 2). The EDTA-treated catalysts showed much higher activity than the untreated catalysts in the range examined (260-400 C) with similar apparent activation energies (E a ) ranging from 40 to 55 kJ mol 1 (Fig. S13 †). For instance, the E a values for LaRuSi before and after EDTA treatment were 40 and 48 kJ mol 1 , respectively, which was similar to that for the a Operation condition: 0.1 g of catalyst at 0.1 MPa with H 2 /N 2 ¼ 45/15 mL min 1 flow. The ammonia formation rates were measured after holding at 400 C for 24 h. b CO is not easily adsorbed on the bulk catalysts; therefore, the estimation was conducted using the surface area (S BET ) and the covalent radii of active atoms, and the Wigner-Seitz radii as the averaged radii for the bulk compounds. c 0.2 g of catalyst. d 0.5 g of catalyst. e 0.4 g of catalyst. This journal is © The Royal Society of Chemistry 2019 Chem. Sci., 2019, 10, 5712-5718 | 5715 electride-supported Ru catalysts. 4,9,13 The reaction orders for LaRuSi before and after EDTA treatment were almost same (Table S6 †). The positive hydrogen reaction order of around 0.6 indicates that the hydrogen poisoning effect was avoided via reversible hydrogen absorption/desorption over LaRuSi due to the intrinsic nature of the anionic electron. 27 Fig. S14 † shows the effect of pressure for the LaRuSi catalysts, where the ammonia formation rate increased approximately 2.6-fold for both samples when the pressure was raised from 0.1 to 0.9 MPa. XRD measurements were performed for the EDTA treated LnRuSi catalysts to check the stability (Fig. S15 †). For all catalysts after ammonia synthesis, the H ions, which are crucial for the formation of ammonia, 27 were incorporated and LnRuSiH x hydrides were formed as the dominant phase. Since LnRu 2 Si 2 cannot incorporate H ions, 27 its phase should not be changed during ammonia synthesis, which was confrmed for the samples without EDTA treatment. However, the 122 phase increased slightly for the EDTA-treated catalysts after ammonia synthesis, which will be discussed later. Consequently, LnRuSiH x as the actual catalyst remained stable during ammonia synthesis for EDTA-treated catalysts, i.e., EDTA treatment did not affect the robust ability of LnRuSi electrides. ICP-AES measurement of the solution after EDTA treatment of LaRuSi was performed. As shown in Table S7, † the dissolved species were predominantly La and Si, while only a small amount of Ru was detected (ca. 3% of La), i.e., surface La and Si were removed during the EDTA etching process, so that the Ru layer emerged to the surface. The amount of La removed was slightly higher than that of Si, which suggests the possibility for the formation of LaRu 2 Si 2 at Si-rich sites, which was confrmed by XRD measurements (Fig. S15 †). XPS measurements were also conducted to evaluate the amounts and valence states of surface species. Fig. 4 shows that the intensity of the La 3d and Si 2s signals decreased signifcantly after EDTA treatment, while that of Ru 3p was signifcantly increased. The intensities of these signals changed slightly after ammonia synthesis. Fig. 4d shows the estimated ratios of surface La, Ru and Si. The Ru ratio increased from 0.15 to 0.54 after EDTA treatment, and decreased slightly to 0.46 after ammonia synthesis, probably due to the reconstruction process. The La to Si ratio was almost constant for all three samples, which indicates that the surface La and Si layers were simultaneously removed during EDTA treatment, as supported by the ICP-AES results. Both the ICP-AES and XPS results revealed that the amount of exposed Ru sites on the surface was increased after EDTA treatment, which was the key reason for enhancement of the catalytic activity. On the other hand, the valence state of Ru was initially negative (460.6 eV present, and 461.4 eV for Ru metal); it became positive after EDTA treatment (461.9 eV) due to the partially oxidation of exposed Ru on the surface, and then returned to negative (460.8 eV) with the reduction of hydrogen during ammonia synthesis. This indicates that during ammonia synthesis, the exposed Ru was still in a negatively charged state for the EDTA-treated catalysts. Most of the catalytic properties of LaRuSi, including the activation energy, reaction orders and pressure effect, remain unchanged before and after EDTA treatment, which suggests that the active sites for both samples are layered Ru in the negative charge state; however, the possibility for the formation of Ru nanoparticles via aggregation on the surface during reaction cannot be excluded. Therefore, HAADF-STEM measurements were performed to determine whether Ru nanoparticles were formed on EDTA-treated LaRuSi. The homogeneous lattice fringes were present in both bright and dark areas (Fig. 5), which demonstrates that there were no Ru nanoparticles on the surface. This was further supported by the FFT pictures in Fig. 5 and the selected area electron diffraction pattern in Fig. S16: † the tetragonal structure corresponding to LaRuSi lattice was clearly observed, while the hexagonal structure that corresponds to Ru metal was not observed. In addition, no Ru peaks were evident in the powder XRD measurements (Fig. S15 †). Therefore, it was concluded that there were no Ru particles on the EDTA-treated LaRuSi surface. We have recently reported that the LaRuSi electride exhibited good activity for ammonia synthesis due to its electride character; negatively charged Ru together with lattice H ions, which can reversibly exchange with anionic electrons, are the key factors in the promotion of ammonia formation. 27 A large difference in the activation energy for ammonia formation (40 kJ mol 1 ) and N 2 isotope exchange reaction (157 kJ mol 1 ) indicated the strong adsorption of N 2 or N atoms on the LaRuSi surface, so that N 2 would be activated through the hot-atom mechanism suggested for LaCoSi. 21 The heat released from adsorption leads to the simultaneous desorption of H from lattice H , which immediately combines with adsorbed N to form NH x . The Ru-N bond is thus weakened through the hydrogenation process, and ammonia is fnally formed and desorbed. However, how planar Ru in LnRuSi acts to activate N 2 has not been discussed previously. It is well-accepted that terrace Ru atoms are not active for ammonia synthesis because both experiments and DFT calculations showed that the apparent activation energy for N 2 dissociation at terrace sites of the Ru (0001) face was as high as 1.8-1.9 eV. Considering the large Ru-Ru distance in LnRuSi (ca. 0.30 nm, Table S8 †), which differs from Ru metal or YRu 2 (ca. 0.26 nm), together with the layered geometric confguration, N 2 activation over LnRuSi should be different from that over conventional Ru catalysts. Here DFT calculations were performed to investigate how the N 2 was adsorbed and activated on the LaRuSi(001)-Ru surface; the reaction path for N 2 dissociation and the adsorption states are shown in Fig. 6 and S17-S19, † respectively. The most stable adsorption confguration of N 2 on LaRuSi(001)-Ru is where a N 2 molecule is adsorbed parallel on the hollow site of La and bonded with four surface Ru atoms, and La is beneath the surface Ru layer and centered within a square of Ru atoms (Fig. S17a †). The deformation charge density for this adsorption state (Fig. S19a †) clearly showed that electrons are transferred from Ru to N 2 , which weakens the N^N triple bond with an elongated bond length (from 0.111 nm to 0.127 nm). As a result, the negatively charged nitrogen has a strong coulombic attraction with the positively charged lanthanum atom, which leads to a larger adsorption energy of 4.21 eV. The N 2 dissociation barrier is 1.30 eV, and two dissociated N atoms locate at the hollow sites of La and Si with an adsorption energy of 4.44 eV (Fig. S18a †). The strong exothermic adsorption of N 2 and the relatively low dissociation barrier indicate that the adsorbed N 2 can be easily dissociated into N atoms without further addition of energy. 21 In this model, all Ru atoms on the surface are active for ammonia synthesis. This is quite different from that of the conventional Ru catalyst. For instance, N 2 activation only occurs at B 5 sites in the conventional Ru catalyst, and the adsorption energies for N 2 and 2 N are respectively 0.7 and 0.8 eV, 17,18 which are much lower than that on the LaRuSi surface. In the conventional Ru catalyst, N 2 is adsorbed perpendicular on the step site and the overall dissociation barrier is 0.4 eV with the dissociated N adsorbed at the 3-fold hcp site. 19 These differences clearly demonstrate that layered Ru in LnRuSi is independent from B 5 site active centers for ammonia synthesis. Further calculations and related studies to clarify the detailed reaction mechanism are in progress. ## Conclusions In summary, we have reported a B 5 -site-independent Ru catalyst, the LnRuSi electride, which exhibits good performance for ammonia synthesis. With respect to their robust durability toward acids, surface Ln and Si are removed while more Ru sites emerge by selective etching using EDTA-2Na acid, which results in a 2-to 4-fold increase in the activity compared to that of the original samples. The estimated TOFs for the LnRuSi catalysts under mild conditions (0.1 MPa, 400 C) were approximately 0.06 s 1 , which is 600 times higher than that of Ru powder. DFT calculations revealed that the specifc LnRuSi structure stabilized N 2 adsorption with a strong exothermic effect, which decreased the apparent activation energy for N 2 dissociation. Therefore, N 2 activation is much easier on the Ru plane of LnRuSi than that on conventional Ru catalysts. We propose that the present air-and acid-durable LnRuSi electride catalysts are applicable to catalytic reactions in water/acid atmospheres, in which conventional electrides do not function well. The present EDTA etching method would also be useful to enhance the activities of other noble metal-containing intermetallic catalyst systems. ## Conflicts of interest There are no conflicts to declare.

chemsum

{"title": "Acid-durable electride with layered ruthenium for ammonia synthesis: boosting the activity <i>via</i> selective etching", "journal": "Royal Society of Chemistry (RSC)"}

multinuclear_iron–phenyl_species_in_reactions_of_simple_iron_salts_with_phmgbr:_identification_of_fe

## Abstract: The first direct syntheses, structural characterizations, and reactivity studies of iron-phenyl species formed upon reaction of Fe(acac) 3 and PhMgBr in THF are presented. Reaction of Fe(acac) 3 with 4 equiv. PhMgBr in THF leads to the formation of [FePh 2 (m-Ph)] 2 2À at À80 C, which can be stabilized through the addition of N-methylpyrrolidone. Alternatively, at À30 C this reaction leads to the formation of the tetranuclear ironphenyl cluster, Fe 4 (m-Ph) 6 (THF) 4 . Further synthetic studies demonstrate that analogous tetranuclear iron clusters can be formed with both 4-F-PhMgBr and p-tolylMgBr, illustrating the generality of this structural motif for reactions of simple ferric salts and aryl Grignard reagents in THF. Additional studies isolate and define key iron species involved in the synthetic pathway leading to the formation of the tetranuclear iron-aryl species. While reaction studies demonstrate that [FePh 2 (m-Ph)] 2 2À is unreactive towards electrophile, Fe 4 (m-Ph) 6 (THF) 4 is found to rapidly react with bromocyclohexane to selectively form phenylcyclohexane. Based on this reactivity, a new catalytic reaction protocol has been developed that enables efficient cross-couplings using Fe 4 (m-Ph) 6 (THF) 4 , circumventing the current need for additives such as TMEDA or supporting ligands to achieve effective cross-coupling of PhMgBr and a secondary alkyl halide. ## Introduction Iron-catalyzed organic transformations continue to attract signifcant interest due to the low cost, improved sustainability and potential for novel reactivity of iron compared to more traditional precious metal catalytic systems. Catalytic reactions involving simple ferric salts and phenyl nucleophiles (most extensively phenylmagnesium bromide, PhMgBr) are of particular interest as this combination has shown to be effective in catalysis in both iron-catalyzed cross-coupling and ironcatalyzed C-H functionalization systems. Unfortunately, the nature of the in situ formed and reactive iron species generated from simple ferric salts and PhMgBr in catalysis remains ambiguous. Such structural insight is essential for defning the role of various iron-phenyl species for both productive and off-cycle reactivity, and for the development of more efficient catalytic methodologies using these reagents. Motivated by the critical need to defne the iron-phenyl species involved in catalysis, several recent studies have employed NMR and electron paramagnetic resonance (EPR) spectroscopies, as well as density functional theory (DFT) calculations to investigate iron speciation in reactions of simple ferric salts and PhMgBr. Such studies have hypothesized the formation of mononuclear Fe I and Fe 0 species in situ. Alternatively, extended X-ray absorption fne structure (EXAFS) studies have led to the proposed formation of Fe II dimers in situ in these reactions, whereas mass spectrometry studies have suggested that iron-phenyl-ate species of even higher nuclearity can form. 26,27 While these studies represent important contributions towards the identifcation of some of the in situ formed iron-phenyl species, the resulting lack of consensus on the nature of these species (e.g. oxidation state, nuclearity) reflects the limited availability of isolated and structurally defned ironphenyl species. To date, the only structurally defned species reported have been a mononuclear iron(II)-phenyl-ate species formed from the reaction between simple iron salts and PhLi, as well as a reduced iron(I)-phenyl species formed using the same nucleophile (Scheme 1). 18,24 The latter was found to be unreactive towards electrophile, likely representing an unproductive iron-phenyl species for cross-coupling. 24 Notably, no structurally defned iron-phenyl species generated from the reaction of simple ferric salts with PhMgBr in THF, the reagents and solvents employed in both cross-coupling and C-H functionalization reactions, have been reported. This lack of structural insight greatly contrasts analogous reactions with the more sterically encumbered mesitylmagnesium bromide (MesMgBr), where both FeMes 3 and Fe 2 Mes 4 have been isolated. 28 In this study, we report the isolation and characterization of the frst multinuclear iron-phenyl species formed from the reaction of Fe(acac) 3 and ArMgBr in THF. Insight into the reaction pathway of formation of both di-and tetranuclear ironcomplexes is presented, as well as the generality of the tetranuclear structure across several aryl Grignard reagents. Stoichiometric reaction studies are utilized to evaluate the potential roles of the multinuclear iron-phenyl complexes in catalysis. These combined synthetic and reaction studies have led to the identifcation of a highly reactive iron-phenyl species for the selective formation of cross-coupled product, resulting in a new ligand-and additive-free reaction protocol for cross-coupling with Fe(acac) 3 and PhMgBr. Previous iron-catalyzed cross-coupling studies have indicated that a minimum of 4 equiv. of Grignard reagent is required to achieve effective catalysis using simple ferric salts and PhMgBr. 29,30 Therefore, synthetic studies focused on the isolation of iron-phenyl species formed using this ratio of iron to Grignard reagent in THF. While viable single crystals could not be isolated directly from reactions at 80 C, a modifed synthetic procedure involving layering and warming of 4 equiv. of PhMgBr on top of Fe(acac) 3 (see Experimental section for further details) produced orange crystalline material suitable for single crystal X-ray diffraction (SC-XRD). The orange crystals were found to have extreme air-, moisture-, and temperaturesensitivities; crystalline material could only be handled at or below 80 C under a nitrogen atmosphere to prevent decomposition. Despite these handling limitations, the crystalline material was successfully characterized by SC-XRD which revealed the identity of the orange crystals as the dinuclear iron species, [Mg(acac)(THF) 4 ] 2 [FePh 2 (m-Ph)] 2 $4THF (1a) (Fig. 1). This complex exhibits an Fe-Fe distance of . Based upon the crystal structure, both iron sites are formally iron(II), where the presence of distorted tetrahedral iron sites suggest the presence of high-spin iron(II) centers due to the tetrahedral ligand felds. ## Results and discussion The 57 Fe Mössbauer spectrum of crystalline 1a was characterized by a single doublet with an isomer shift (d) of 0.34 mm s 1 and a quadrupole splitting (DE Q ) of 2.28 mm s 1 . The observed quadrupole splitting is consistent with the presence of highspin iron(II) sites and the low isomer shift is in agreement with previously reported transmetalled iron(II) complexes corresponding to the highly covalent Fe-C bonds. 36,37 Unfortunately, further characterization of pure 1a was not possible due to the high thermal instability of the complex. N-Methylpyrrolidone (NMP) has been previously employed as an additive in cross-coupling reactions with simple ferric salts, and has been demonstrated in the case of MeMgBr to stabilize [FeMe 3 ] , favoring its formation over [Fe 8 Me 12 ] . 38,39 Interestingly, NMP was found to bind to the Mg cation in this case 38 as opposed to binding directly to the iron center as suggested in earlier studies by Holland and co-workers. 40 39 the change of counterion in the complex signifcantly improved the thermal stability of the iron-phenyl dimer at 30 C, enabling additional characterization of 1b. While Evans method NMR coupled with atomic absorption spectroscopy (AAS; to quantify the amount of 1b dissolved in solution as the material is too thermally unstable to weigh) would be an ideal way to calculate the spin state of 1b, the complex was found to be too unstable to obtain meaningful NMR data. Instead, magnetic circular dichroism (MCD) was employed to determine the spin state of 1b. No MCD signal is observed in both mull and solution samples at 5 K which, combined with the absence of an EPR signal at 10 K, is consistent with the assignment of a S ¼ 0 ground state due to antiferromagnetic coupling of the iron(II) sites in the dimer. Lastly, it was important to evaluate whether 1b could undergo additional transformations upon reaction with additional PhMgBr. 57 Fe Mössbauer spectroscopy confrmed that addition of excess PhMgBr (>20 equiv.) directly to a solution of 1b at 80 C resulted in no reaction within 5 min (a catalytically relevant time frame, vide infra). The formation of both 1a and 1b differs drastically from the formation of Fe 2 Mes 4 ; Fe 2 Mes 4 is accessible at RT, illustrating how the sterically bulkyl mesityl ligands modulate the formation and stability of the resulting iron dimer. 28 ## Isolation and characterization of tetranuclear iron-aryl species Due to the thermal instability of 1a, it was critical to determine the nature of the iron species formed at elevated temperatures as a more reduced iron-phenyl species might be obtainable. Reaction of Fe(acac) 3 with 4 equiv. of PhMgBr in THF at 30 C, in the absence of NMP, yielded a brown solution from which brown crystalline material could be obtained. The overall connectivity, geometry, and chemical formulation could be unambiguously assigned by SC-XRD as the tetranuclear iron cluster Fe 4 (m-Ph) 6 (THF) 4 $2THF (2a) (Fig. 3). This cluster is more reduced than 1a or 1b, formally containing two iron(I) and two iron(II) ions. Unfortunately, these crystals diffracted very weakly and, hence, a more detailed discussion of the structural parameters of 2a is not possible. It was hypothesized that changing the aryl group of the Grignard reagent might enable access to higher quality crystalline material for SC-XRD analysis. , where the Fe-Fe bond distances typically range from 2.469 to 2.618 . Interestingly, a [Fe 8 Me 12 ] cluster has been previously isolated and characterized as a key intermediate in reactions involving simple ferric salts and MeMgBr. 45 The Fe-Fe distances in [Fe 8 Me 12 ] are closer to the Fe-Fe distances in 2b and 2c, ranging from 2.4188(15) to 2.4514(15) . 45 The bridging Fe-Ph bonds range from 2.139(3) to 2.266(3) for 2b and from 2.078(5) to 2.412(5) for 2c. Lastly, an analogous tetranuclear iron complex could also be synthesized using 4-fluorophenylmagnesium bromide (4-F-PhMgBr), identifed by SC-XRD as Fe 4 (m-4-F-Ph) 6 (THF) 4 (2d). As with 2a, the isolated crystalline material was weakly diffracting, and hence, the SC-XRD data was solely used for the unambiguous assignment of connectivity. Overall, the ability to access analogous tetranuclear iron-aryl clusters across both electron withdrawing and electron donating substituents on the aryl ligands demonstrates the generality of this structural motif in reactions of simple ferric salts and aryl Grignard reagents. Complex 2a was the most synthetically robust tetranuclear complex for larger scale isolation and was utilized for further characterization and reactivity (vide infra) studies. The 57 Fe Mössbauer spectrum of 2a is a broad quadrupole doublet with d ¼ 0.60 mm s 1 and DE Q ¼ 0.84 mm s 1 (see ESI †). Analogous broad features were also previously observed for [Fe 8 Me 12 ] . 45 The tetranuclear complex is not EPR active, but exhibits an intense C-term MCD spectra in both the near-infrared and UVvis regions (see ESI †) consistent with the presence of an integer spin paramagnetic complex. Unfortunately, variabletemperature, variable-feld (VTVH) MCD studies on the ground state of 2a did not enable the assignment of the spin state of this complex, likely due to complications from decay to additional paramagnetic iron species which complicates both these measurements and attempted Evans studies. ## Investigation of the iron species involved in the synthetic pathway for formation of di-and tetranuclear iron-aryl species Stoichiometric reactions of Fe(acac) 3 with varying equivalents of 4-F-PhMgBr enabled the further investigation into the underlying synthetic pathway leading to formation of the dinuclear and tetranuclear iron species. When 1 equiv. of 4-F-PhMgBr is slowly added to Fe(acac) 3 in THF at 30 C, a red color persists. From this reaction, crystalline material of [trans-Fe(acac) 2 (THF) 2 ] 0.58 $[trans-Mg(acac) 2 (THF) 2 ] 0.42 (3) was isolated and characterized by SC-XRD (Fig. 4). Similar crystals were independently isolated from reactions of Fe(acac) 3 with 1 equiv. of three different aryl Grignard reagents (PhMgBr, 4-F-PhMgBr, and p-tolylMgBr), consistent with the frst equivalent of ArMgBr reducing the iron(III) starting material to iron(II). Addition of a second equivalent of aryl Grignard reagent to the iron solution at 30 C results in a color change from red to yellow. Although no usable crystals of the yellow material could be obtained, one could envision a complex analogous to the structurally characterized [(di-tBu-acac)Fe(m-Mes)] 2 . 46 The yellow iron solution turns orange upon the slow addition of a third equivalent of aryl Grignard reagent at 30 C. From this solution, orange crystals were obtained at 80 C suitable for SC-XRD and determined to be [Mg(acac)(THF) 4 ] 2 [FeBr(4-F-Ph)(m-4-F-Ph)] 2 $2.5THF (4). The Fe-Fe distance was found to be 2.5903(11) , which is longer than previously identifed homoleptic iron(II)-phenyl dimers 1a and 1b. The unique terminal and two bridging Fe-Ph bond lengths, based on the independent dianion without disorder, are 2.072(4), 2.166(4) and 2.212( 4 re-dissolution of 1b in THF at 20 C, only 70% of 1b was found to remain in solution after 45 s. As expected, repeating this experiment at 0 C shows more rapid decay of the dimer with only $50% of 1b remaining in solution after 45 s. Therefore, reaction studies with electrophile were performed at 20 C, focusing on a 45 s reaction window in order to minimize contributions from the decomposition of 1b. Bromocyclohexane was selected as an example electrophile for the reaction studies due to its common use in ferric salt catalyzed crosscoupling reactions with PhMgBr. 47,48 GC-MS reaction studies showed no consumption of electrophile or generation of phenylcyclohexane within 45 s of reaction at 20 C. Thus, the [FePh 2 (m-Ph)] 2 2 dimer exhibits no reactivity towards electrophile prior to its thermal decomposition in THF. Furthermore, the observation that NMP stabilizes the formation of this unreactive dimer is consistent with previous studies by Nakamura and co-workers, where NMP was shown to be an unfavorable co-solvent for iron-catalyzed cross-coupling reactions involving simple ferric salts and aryl Grignard reagents. 47 In contrast to the lack of reactivity of 1b, formation of crosscoupled product was observed to be generated, albeit in low yield (24%), from the reaction of Fe 2 Mes 4 and electrophile. 28 In order to evaluate the potential reactivity of the more reduced tetranuclear iron complex 2a with bromocyclohexane, it was again critical to frst establish the thermal stability of 2a in solution at catalytically relevant temperatures. Fortunately, 2a was found to be stable at RT for up to 5 min in THF, enabling stoichiometric reactions to be performed within this time frame. Reactions of 2a with 15 equiv. bromocyclohexane at RT resulted in the rapid and selective formation of phenylcyclohexane (0.95 equiv. with respect to 2a within 5 s). Thus, 2a is a highly reactive species for the selective formation of crosscoupled product (k obs $ 12 min 1 for the initial turnover). Prolonged reaction times led to the generation of additional cross-coupled product, indicating that the iron products of each cross-coupling are capable of further reaction with electrophile ($4 equiv. phenylcycohexane after 1 min of reaction (see ESI †); note that the reaction rate decreases for subsequent turnovers). It is noteworthy that 2a can directly react with electrophile to form cross-coupled product, whereas [Fe 8 Me 12 ] requires the addition of MeMgBr following initial reaction with electrophile to form product, 45 indicating the presence of different underlying reaction mechanisms for iron-phenyl and iron-methyl clusters. Cross-coupling catalysis using Fe 4 (m-Ph) 6 (THF) 4 Simple ferric salts were previously found by Nakamura and coworkers to perform poorly for catalytic cross-couplings of PhMgBr with secondary alkyl halides in the absence of TMEDA. 47 Interestingly, Bedford and co-workers observed similar reactivity in the presence and absence of TMEDA when MesMgBr was employed, though this system was low yielding ($35%). 28 Hence, we were motivated by the observed reactivity of 2a to explore its potential effectiveness for catalytic cross- coupling in the absence of TMEDA. The utilization of the same reaction protocol as described for stoichiometric reactions of 2a but with the addition of PhMgBr (1 : 1 with respect to electrophile) resulted in the formation of >95% cross-coupled product (Scheme 2a). Because 2a is challenging to synthesize and handle, a modifed catalytic method targeting the formation of 2a in situ at 30 C (a temperature where 2a is stable for days) was also evaluated as a potentially more convenient protocol that utilizes the selective reactivity of 2a without the need to isolate it (Scheme 2b). With this method, >95% of cross-coupled product could be obtained. Interestingly, removal of magnesium salts by fltration at 30 C following in situ formation of 2a was found to be critical to achieve high yields of product, likely indicating an important role of cations on iron speciation during the initial synthesis of 2a or during catalysis. While a proof of concept, this initial evaluation of in situ generated 2a for catalysis will hopefully inspire future studies in the area of ligandless iron cross-coupling catalysis. While the current study has demonstrated the importance of dinuclear and tetranuclear iron-phenyl species in reactions of simple ferric salts and aryl Grignard reagents, additional ironphenyl species beyond those isolated herein might also be accessible in such reactions. For example, previous EPR studies by Bedford indicated the in situ formation of a S ¼ 1/2 species in reactions of FeCl 3 and p-tolylMgBr at 30 C though, unfortunately, this species was never spin quantifed. 21,28 We have observed the formation of the same S ¼ 1/2 species in reaction of FeCl 2 and 4 equiv. PhMgBr at 30 C, though spin quantitated EPR indicated that it is a very minor species in solution (<5% of all iron) (see ESI †). 21,28 Additionally, Bedford and coworkers also suggested that a monomeric iron(II) species, [Fe(p-tolyl) 3 ] can also form based on 1 H NMR studies of reactions of FeCl 2 with p-tolylMgBr, 28 where the resonances assigned to this mononuclear species are signifcantly down-feld shifted compared to those observed for 2b (see ESI †). Again, however, the amount of the mononuclear species in solution was not quantifed and it remains unclear whether it is formed signifcantly in solution compared to other iron-p-tolyl species. Future studies should continue to defne the diverse iron-aryl species accessible in such reactions as a function of concentration, solvent, aryl nucleophile (e.g. ArMgBr, ArLi, Ar 2 Zn, etc.), temperature, and reaction time. Lastly, it is interesting to consider the origin of the reactivity differences observed for [FePh 2 (m-Ph)] 2 2 and Fe 4 (m-Ph) 6 (THF) 4 . Since previous studies have proposed an Fe(I) active species for cross-coupling with PhMgBr and simple ferric salts (such as [PhFe I (acac)(THF)] ), the presence of two formally iron(I) sites in the mixed valent tetranuclear iron complex 2a might suggest iron reduced below iron(II) is important for reactivity in the isolated multimetallic complexes. Specifcally, the THF ligation differences between complexes 2b and 2c demonstrate the ability of the tetranuclear complexes to lose a THF ligand to generate an open coordination position for reaction with electrophile. This ability to readily form an open coordination site might be equally signifcant in facilitating reactivity, whereas the dinuclear complexes would require a more signifcant geometric distortion in order to react with electrophile. ## Conclusions In this study, the frst direct syntheses, structural characterizations, and reactivity studies of iron-phenyl species formed upon reaction of Fe(acac) 3 and PhMgBr in THF have been presented. At 80 C, this reaction leads to formation of [FePh 2 (m-Ph)] 2 2 , which was found to be unreactive towards electrophile. Alternatively, at 30 C the formation of a more reduced, tetranuclear iron-phenyl cluster, Fe 4 (m-Ph) 6 (THF) 4 , is observed, where this species is found to rapidly react with bromocyclohexane to selectively form cross-coupled product. Further synthetic studies demonstrate that analogous tetranuclear iron clusters can be formed with both 4-F-PhMgBr and p-tolylMgBr, illustrating the generality of this structural motif for reactions of simple ferric salts and aryl Grignard reagents in THF. Lastly, Fe 4 (m-Ph) 6 (THF) 4 can be utilized for efficient catalytic crosscoupling of PhMgBr and bromocyclohexane, circumventing the current need for additives such as TMEDA or supporting ligands to achieve high yields of cross-coupled product in this reaction. ## General considerations All reagents were purchased from commercial sources. All airand moisture-sensitive manipulations were carried out in an MBraun inert-atmosphere (N 2 ) glovebox equipped with a direct liquid nitrogen feed through inlet line. All anhydrous solvents were freshly dried using activated alumina/4 molecular sieves and stored under an inert atmosphere. Gas chromatography mass spectrometry was performed using a Shimadzu GCMS QP 2010. Atomic absorption spectroscopy (AAS) analysis was performed using a Shimadzu AAS 7000. Details on low temperature crystal manipulations, sample preparations for spectroscopy and MCD and EPR spectroscopy are given in the ESI. † 57 Fe Mössbauer spectroscopy Solution samples for 57 Fe Mössbauer spectroscopy were prepared from 57 Fe(acac) 3 to enable data collection from dilute, freeze-trapped solution samples; solid samples were made from non-enriched Fe(acac) 3 . All samples were prepared in an inert atmosphere glovebox equipped with a liquid nitrogen fll port to enable sample freezing to 77 K within the glovebox. Each sample was loaded into a Delrin Mössbauer cup for measurements and loaded under liquid nitrogen. Low temperature 57 Fe Mössbauer measurements were performed using a See Co. MS4 Mössbauer spectrometer integrated with a Janis SVT-400T He/ N 2 cryostat for measurements at 5 K, 80 K, and 150 K. Isomer shifts were determined relative to an a-Fe at 298 K. All Mössbauer spectra were ft using the program WMoss (SeeCo). Errors of the analyses are d AE 0.02 mm s 1 and DE Q AE 3%. ## Magnetic circular dichroism spectroscopy All samples were prepared in an inert atmosphere glovebox equipped with a liquid nitrogen flling port to enable sample freezing to 77 K. Low temperature near-infrared (NIR) MCD experiments were conducted using a JASCO J-730 spectropolarimeter and a shielded S-20 photomultiplier tube. Both instruments have a modifed sample compartment, which incorporates focusing optics and an Oxford Instruments SM4000-7T superconducting magnetic/cryostat. This set-up allows for measurements from 1.6 K to 290 K, with magnetic felds up to 7 T. A calibrated Cernox sensor directly inserted in the copper sample holder is used to measure the temperature at the sample to 0.001 K. All MCD spectra were baseline-corrected against zero-feld scans. ## Electron paramagnetic resonance spectroscopy A cold spatula was used to transfer material to a vial containing a known amount of THF at 80 C. A cold pipette was then used to transfer the redissolved crystalline material to a precooled (in liquid nitrogen) 4 mm OD suprasil quartz EPR tube from Wilmad Labglass. The solution in the EPR tube was immediately frozen in liquid nitrogen. The remaining solution not used for the EPR sample was saved for AAS, so that spin integration of any EPR signal could be completed. All X-band EPR spectra were collected on a Bruker EMXplus spectrometer containing a 4119HS cavity and an Oxford ESR-900 helium flow cryostat. All EPR spectra were collected at 10 K, 9.38 GHz. Preparation of Fe 4 (m-4-F-Ph) 6 (THF) 4 (2d) Solid Fe(acac) 3 (70 mg, 0.2 mmol) was dissolved in THF (2 mL). The solution was then cooled to 0 C, where 4-F-PhMgBr in THF (1.0 M, 794 mL, 0.794 mmol) was added dropwise at 0.33 mmol min 1 . The brown solution was then fltered through cold Celite. Cold toluene (1 mL) was then layered on top of the THF solution, and the sample was stored at 30 C until crystalline material formed. Preparation of [trans-Fe(acac) 2 (THF) 2 ] 0.58 $[trans-Mg(acac) 2 (THF) 2 ] 0.42 (3) Solid Fe(acac) 3 (66 mg, 0.19 mmol) was dissolved in THF (5 mL). The solution was then cooled to 30 C, where 4-F-PhMgBr in THF (1.0 M, 186 mL, 0.186 mmol) was added dropwise at 0.33 mmol min 1 . The orange solution was allowed to stir at 30 C for 5 minutes at 620 rpm prior to fltering through cold Celite. Cold pentane (10 mL) was then layered on top of the solution, and the sample was stored at 80 C until crystalline material was formed. Solid Fe(acac) 3 (63 mg, 0.18 mmol) was dissolved in THF (5 mL). The solution was then cooled to 30 C, where 4-F-PhMgBr in THF (1.0 M, 535 mL, 0.535 mmol) was added dropwise at 0.33 mmol min 1 . The dark orange solution was allowed to stir at 30 C for 5 minutes at 620 rpm. The dark orange was then fltered through cold Celite. Cold pentane (2 mL) was then layered on top of the solution, and the sample was stored at 80 C until crystalline material was formed. ## Thermal stability of 1b and 2a in solution Crystalline material was collected as described in the ESI. † Crystalline material was transferred to a vial containing a known amount of THF at 80 C for 1b and at 30 C for 2a using a cold spatula. Once the crystalline material was completely redissolved, the solution was transferred to a vial containing a known amount of THF and stir bar at a warmer temperature using a cold pipette. Aliquots of the decaying solution were taken at various points over 20 minutes. Aliquots were taken using a cold pipette, and were transferred into a Delrin Mössbauer sample cup. Samples were immediately frozen in liquid nitrogen. This process was repeated for various temperatures, including, 20 C and 0 C. AAS was used to determined to the concentration of the respective complexes in solution. ## Reaction of 1b with bromocyclohexane Dark red blocks of 1b were collected as described in the ESI. † Crystalline material was transferred to a vial containing a known amount of THF at 80 C using a cold spatula. Once the crystalline material was completely redissolved, the solution was transferred to a vial containing THF, bromocyclohexane, PhMgBr, and stir bar at a warmer temperature using a cold pipette. The crystalline material was allowed to react, and aliquots were taken at various points over 20 minutes. The aliquots were quenched in 1 : 1 (v/v) THF : MeOH solution. A known amount of dodecane was then added to the quenched samples, and the samples were diluted to 1 mM prior to fltering through silica. AAS was used to determine the concentration of 1b in solution. ## Reaction of 2a with bromocyclohexane Crystalline material was collected as described in the ESI. † Crystalline material was transferred to a vial containing a known amount of THF at 30 C using a cold spatula. Once the crystalline material was completely dissolved, the solution was transferred to a vial containing THF, bromocyclohexane (15 equiv. wrt 2a), and a stir bar at a warmer temperature using a cold pipette. The crystalline material was allowed to react, and aliquots were taken at various points over 20 minutes. The aliquots were quenched in MeOH (50 mL). A known amount of dodecane was then added to the quenched samples, and the samples were diluted to 1 mM prior to fltering through silica. AAS was used to determine the concentration of 2a in solution. ## Catalytic reaction protocol using isolated 2a as catalyst Crystalline material was collected as described in the ESI. † Crystalline material was transferred to a vial containing a known amount of THF at 30 C using a cold spatula. Once the crystallined was completely dissolved, the solution was quickly transferred to a vial containing THF, bromocyclohexane (13 equiv. wrt 2a), PhMgBr (1.0 M in THF, 13 equiv. wrt 2a), and a stir bar at RT. The crystalline material was allowed to react with electrophile and excess nucleophile, and aliquots were taken over the course of 20 minutes. The aliquots were quenching in MeOH (50 mL). A known amount of dodecane was then added to the quenched samples, and the samples were diluted to 1 mM prior to fltering though silica. AAS was used to determine the concentration of 2a in solution. Catalytic reaction protocol targeting 2a formation in situ

chemsum

{"title": "Multinuclear iron\u2013phenyl species in reactions of simple iron salts with PhMgBr: identification of Fe₄(\u03bc-Ph)₆(THF)₄ as a key reactive species for cross-coupling catalysis", "journal": "Royal Society of Chemistry (RSC)"}

a_diastereodivergent_and_enantioselective_approach_to_syn-and_anti-diamines:_development_of_2-azatri

## Abstract: We introduce a new reagent class, 2-azatrienes, as a platform for catalytic enantioselective synthesis of allylic amines. Herein, we demonstrate their promise by a diastereodivergent synthesis of synand anti-1,2-diamines through their Cu-bis(phosphine)catalyzed reductive couplings with imines. With Ph-BPE as the supporting ligand, anti-diamines are obtained (up to 91% yield, >20:1 dr, and >99:1 er), and with the rarely utilized t-Bu-BDPP, syn-diamines are generated (up to 76% yield, >20:1 dr, and 97:3 er). ## INTRODUCTION Chiral 1,2-diols, amino alcohols, and diamines are important targets for organic synthesis as these motifs are ubiquitous in natural products and drugs, as ligands for metal-based catalysts, and as catalysts themselves. Several approaches to these scaffolds have been established; however, the invention of carbon-carbon bond-forming reactions that directly set these vicinal heteroatomsubstituted stereogenic centers is underdeveloped. A recent elegant report was disclosed by the Krische group, utilizing their hydrogen auto-transfer technology to couple an allenimide with a primary alcohol-derived aldehyde to afford 1,2-amino alcohols where the amino group is allylic (Scheme 1). Allylic amines are important structural features in numerous bioactive molecules and natural products. Furthermore, the unsaturation may serve as a functional group handle for downstream transformations. Although having excellent scope in the alcohol partner, the reactions were limited to terminal allenes, giving rise to terminal allyl groups; moreover, the anti-amino alcohol was the only stereoisomer accessible. ## Scheme 1. Catalytic Reductive and Borylative Processes that Set Vicinal Stereogenic Centers Our group has investigated the synthesis of both 1,2-diamines (Scheme 1) and amino alcohols by reductive couplings of 2-azadienes. These transformations proceed by means of a copper-hydride intermediate with the bis(phospholane) Ph-BPE as the ligand. In both cases, the product amines bear an -alkyl group. Furthermore, the diamines were generated solely as the anti diastereomer in every case. These examples highlight an often encountered situation in enantioselective reactions that afford more than one stereogenic center: the ability to access only one diastereoisomer. One strategy that addresses this shortcoming is a dual catalyst approach wherein each catalyst acts cooperatively but independently to activate two reaction components individually, thereby enabling each to control stereochemistry at its respective fragment. An alternative is the use of two related single catalysts for transformations that individually afford opposite diastereomers with high enantioselectivity. Such an approach has recently been illustrated in copper-phosphine-catalyzed borylative couplings (Scheme 1). Shimizu, Kanai, and co-workers demonstrated Cu-B(pin) addition to styrene followed by coupling with N-thiophosphinoylimines. -Arylamines are obtained as the syn-isomer with a Josiphos ligand whereas Ph-BPE delivers the anti-diastereomer. Similarly, the Ostreich group discovered that 2-substituted dienes yield homoallylic alcohols as the anti-diastereomer with Josiphos but the syn-diastereomer with a phosphoramidite ligand. To our knowledge, no examples of diastereodivergent behavior in copper-catalyzed reductive couplings of olefins with electrophiles have been reported. We have developed 2-azatrienes as new reagents for the synthesis of substituted allylic amines. Herein we illustrate their reductive coupling with N-phosphinoylimines to afford 1,2diamines with high chemo-, regio-, diastereo-, and enantioselectivity (Scheme 1). Cu-Ph-BPE promotes the formation of anti-diamines. Unexpectedly, and in stark contrast to our findings with azadiene reagents, we discovered that several other ligands enable the cross-coupling and favor the syn-diamine product. We disclose the first examples of reductive coupling using t-Bu-BDPP, an uncommon ligand in catalysis, to achieve good to excellent levels of diastereo-and enantioselectivity for syn-diamine production. ## RESULTS AND DISCUSSION 2.1. Method Development. We began by examining the coupling of terminal 2-azatriene 1 with imine 2a, employing Cu(OAc)2 and Ph-BPE (L1) under the conditions established for azadiene addition to these imines (Table 1, entry 1). The transformation generates the anti-diamine 3a as the sole stereoisomer, isolated in 90% yield and 99:1 er. Regioselectivity for the 6,3-addition product over the isomeric azadiene 4a (6,5-addition) is excellent. Furthermore, chemoselectivity for reductive coupling over imine reduction (3a/4a:5a >20:1) is considerably greater than in our previous azadiene-imine coupling (coupling/reduction = 5:1), which might be attributed to the LUMO-lowering effect of extra conjugation in 1 plus its decreased sterics over an azadiene (cf. ## Scheme 1). Unexpectedly, we discovered that syn-diamine 3a is the major product (1:3.5 anti:syn-3a) with achiral DCyPe (L2, entry 2) when attempting to prepare the authentic racemic material for entry 1. This finding stands in contrast to azadiene reductive couplings with imine 2a, where Ph-BPE and DCyPE both preferentially furnish the anti-diamine product. Although selectivity metrics were modest for DCyPE in the azatriene coupling, this result prompted us to explore whether a chiral ligand could be found that would lead to enantioselective formation of the syn-3a diastereomer. [a] Reaction with 0.1 mmol imine 2a. [b] Determined by 500 MHz 1 H or 162 or 202 MHz 31 P NMR of the unpurified mixture. [c] Determined by HPLC analysis of purified 3. [d] Isolated yield of diamine 3a. [e] (L)Cu(OAc)2 complex formed from L•2BH3. [f] 2.0 equiv TMDS. [g] In CH2Cl2 with 10 mol % catalyst. DMMS = Me(MeO)2SiH; TMDS = [(Me)2HSi]2O. With Chiraphos (L3), the reaction is reasonably efficient but poorly selective in all categories, generating syn-3a as a racemate (entry 3). In contrast, spacing the phosphino groups farther apart by turning to BDPP (L4) leads to markedly improved stereoselectivity (1:6 anti:syn-3a, 83:17 er, entry 4). Replacing the methyl groups of BDPP with phenyl substituents (L5) significantly erodes stereoselectivity (1:1.5 dr, 50:50 er) and leads to a large quantity of imine reduction (entry 5). Similarly, changing the diphenylphosphino groups to dicyclohexylphophino (L6) abolishes stereoselectivity (entry 6); regio-and chemoselectivity are also poor. Fortunately, modification of the aryl groups of the phosphine within the BDPP structure proved more fruitful. Introduction of a tert-butyl group at the arene's para position (herein called t-Bu-BDPP, L7, entry 7) restores diastereoselectivity (1:6 dr), increases the proportion of diamine 3a, and significantly improves the enantioselectivity (94:6 er). Switching the silane to TMDS further increased the quantity of syn-diamine 3a (1:8.5 dr, entry 8). Finally, changing the solvent to CH2Cl2 and increasing the catalyst loading to 10 mol % (entry 9) allowed for syn-3a to be obtained with considerably enhanced regio-and chemoselectivity and isolated in 69% yield, 1:12.5 dr, and 97:3 er. [a] Reactions run under standard conditions shown; isolated yields and er of the major diastereomer. [b] Regiomeric ratio (rr) is the ratio of 6,3-addition to 6,5-addition and was determined by 500 MHz 1 H or 162 or 202 MHz 31 P NMR spectroscopy of the unpurified mixture; dr, listed as anti:syn, was determined by 500 MHz 1 H or 162 or 202 MHz 31 P NMR spectroscopy of the unpurified mixture. [c] Isolated product contains 9% syn-3b and 7% 4b. [d] 3.0 equiv 1. [e] 2.0 equiv 1. [f] Conversion of imine 2f, 3:2 3f/4f:5f. [g] Conversion of imine 2g, 1:1.3 3g/4g:5g. [h] Isolated product contains 10% syn-3j and 10% 4j. [i] Conversion of imine 2k; 4k is the major product (see Figure 2). [j] Isolated product contains 7% anti-3l and 19% 4l. [k] Isolated product contains 9% syn-3m and 20% (Z)-3m. [l] Isolated product contains 19% anti-3m and 19% (Z)-3m. [m] Isolated product contains 12% anti-3n. nd = not determined. A number of aryl aldimines of varying substitution patterns may thus be coupled with azatriene 1 to deliver either antior syn-diamines (Table 2). Diamines with a variety of arene functional groups, such as methoxy (3b), halide (3c-d, 3i, 3k), trifluoromethyl (3e), ester (3f), nitrile (3g), and alkyl (3j) were prepared. Additionally, several heterocyclic aldimines were investigated and are tolerated by the copper-based catalysts, including pyridine (3l), pyrrole (3m), pyrazole (3n), indole (3o), and thiophene (3p). Yields range from 33% to 91% for the major diastereomer of any isolated product, demonstrating the broad potential of the method to prepare both vicinal diamine diastereomers with a diverse chemical landscape. In general, reactions we explored with Ph-BPE deliver anti-diamines 3 in >20:1 dr and 98:2 er. Contrastingly, stereoselectivity for syn-diamine formation with t-Bu-BDPP is considerably more variable, showing a wide range of both dr (1:3 to 1:>20) and er (86.5:13.5 to 97:3). Still, couplings favor syn-diamines over the anti isomers and with good enantioselectivity (7:1 syn:anti and 94:6 er for the syn). Regioselectivity for the allylic diamine is also greater with Ph-BPE as the supporting ligand (15:1 rr in most cases) and more variable with t-Bu-BDPP (3:1 to >20:1 rr), which is one factor in the higher yields obtained for the anti diastereomer. Chemoselectivity for reductive coupling versus imine reduction is tied to imine electronics with both catalysts: more electron-rich imines deliver a higher proportion of C-C bond formation. The copper complex derived from t-Bu-BDPP was more greatly influenced in this regard. For example, p-chloro syn-3d is obtained in 53% yield but p-CF3 syn-3e in just 33% yield despite the reactions having similar regio-and diastereoselectivity. Intriguingly, reaction of 2-iminopyrrole 2k with either catalyst affords an appreciable quantity of the (Z)-olefin isomer (ca. 2-3:1 E:Z) although only (E)-alkenes are obtained in all other cases. From this initial data set, several differences in trends in reaction metrics from transformations involving Ph-BPE (L1) and t-Bu-BDPP (L7) are notable. Whereas more electron-rich aldimines lead to greater diastereoselectivity when L7 is employed (compare syn-3b-g, ranging from 1:4.5 to 1:13 dr), the reaction of p-methoxy imine 2b in the presence of L1 leads to only 7.5:1 dr. In contrast, anti-3c-e are generated in >20:1 dr. Likewise, regioselectivity (3:4) is greatest for reaction of 2b versus other imines with L7 and poorest with L1. Aryl aldimines bearing ortho substituents (2j-k) lead to perfect regio-and diastereoselectivity for syn-3j-k with L7. At the same time, this ortho substitution engenders the lowest enantioselectivity observed for syn-diamines with L7 (91:9 er for syn-3j and 86.5:13.5 er for syn-3k). With L1, however, anti-3j, with its orthomethyl group, is obtained in only 6:1 dr and 2.5:1 rr. ortho-Bromo anti-3k is the minor isomer from the reductive coupling (1:5.5 3k:4k), is formed in only 6:1 dr, and was not isolated. 2-Azatrienes bearing alkyl substituents at the 6-position (6) enable diamines (7) with longer chain olefin substituents to be obtained (Table 3). With the greater chemoselectivity for crosscoupling shown by Cu-Ph-BPE in azatriene couplings, anti-7a-h are isolated in good yields (51-89%) even with electronically neutral imine 2a. This contrasts with transformations with substituted azadienes, which required electron-rich imines to avoid reduction. Both diastereoand enantioselectivity are excellent (12:1 to >20:1 dr and 95:5 to 99:1 er), but in most cases regioselectivity is more modest than with terminal azatriene 1 (7:1 to 12:1 rr for anti-7a-g). Triamine anti-7h, however, is formed as a single regioisomer. The Cu-t-Bu-BDPP catalyst is more prone to imine reduction, and with the greater sterics of substituted azatrienes 6, more electron-rich imines are required to achieve appreciable yields of syn-diamines (Table 3). Within these confines, a number of azatriene-imine combinations afford syn-diamines in good yields (39-76% for 7i-l). Diastereo-and regioselectivity are good (1:7 to 1:>20 dr and 9.5:1 to >20:1 rr) and enantioselectivity remains high (93.5:6.5 to 97:3 er). Scheme 2. Azatriene 1 Additions to an Aliphatic Aldimine and a Ketimine with Cu-Ph-BPE Ph-BPE also permits azatriene couplings with an aliphatic aldimine and a ketimine (Scheme 4). Diamine anti-9 is formed with 9.5:1 dr and 88:12 er from aldimine 8 and azatriene 1; the product was isolated as an 8:1 mixture of E/Z isomers. Ketimine 10 undergoes a highly diastereoselective addition, forming anti-11 in 20:1 dr, although regio-(6:1 rr) and enantioselectivity (85:15 er) are moderate. Intriguingly, the allylation reaction leads to only 2.5:1 E/Z selectivity for the olefin within 11. Cu-t-Bu-BDPP is ineffective in these couplings, generating a complex mixture. For preparative scale diamine synthesis, we employed lower catalyst loadings and higher reaction concentrations (Scheme 5). Excellent yields of the two diamine diastereomers are thereby obtained within a few hours. For instance, anti-3a was generated in 86% yield with just 1.2 mol % Ph-BPE. Similarly, 2a was converted to syn-3a (61% yield) in the presence of just 3.3 mol % of the Cu-t-Bu-BDPP catalyst. Regio-and stereoselectivity are largely unaffected by the scale up and modified conditions. Scheme 3. Larger Scale Diamine Synthesis ## Mechanistic Studies. In order to gain a better understanding of factors governing the stereochemical outcome of the reductive couplings with the two optimal catalysts, we carried out a number of additional experiments. Having qualitatively observed a relationship between aryl aldimine electronics and the diastereoselectivity of diamine formation, we first initiated a more detailed study to determine if there were a true correlation and, if so, its magnitude. The results are shown as Hammett plots in Figure 1. Each ligand shows a linear dependence for the reaction diastereoselectivity upon the imine's electronic character although this tie is greater for Ph-BPE (L1). For both ligands, the ratio of the normally observed major diastereomer to the minor isomer increases as the imine becomes more electron-deficient. With Ph-BPE, the selectivity morphs from a reaction that slightly favors the syn-diamine with a p-NMe2 group (1:1.2 dr) to a highly anti-selective process (66:1 dr) with the p-CF3 imine ( = 1.4, R 2 = 0.98, Figure 1A). For t-Bu-BDPP (L7), however, the p-NMe2substituted imine still leads to a fairly syn-selective reaction (1:7 dr) but the diastereoselectivity increases to a maximum of just 1:13 with a p-fluoro group ( = 0.30, R 2 = 0.99, Figure 1B). For each ligand, there is a break in the plot where diastereoselectivity then decreases as the imine becomes even more electron-poor. The break is indicative of a change in the diastereodeterming step in the reactions. For Ph-BPE, the erosion does not significantly impact the synthetic utility, with the p-nitro imine delivering the corresponding diamine in 22:1 dr ( = -0.63, R 2 = 0.97, Figure 1A); with t-Bu-BDPP, the p-cyano syn-diamine 3g is modestly favored (1:4.5 dr,  = -0.73, R 2 = 0.98, Figure 1B). It should be noted that product regioselectivity shows a poor correlation with imine electronics. ## Table 4. Comparison of (E)-and (Z)-Azatrienes [a] [a] Reaction with 0.1 mmol imine 2a. Entries 1 and 3 run under the conditions of Table 1, entry 1; entries 2 and 4 run under the conditions of Table 1, entry 9. [b,c] See Table 1. We next investigated how stereochemistry of the azatriene might play a role in the chemo-, regio-, and stereoselectivity of the imine couplings (Table 4). Under their respective optimized conditions, the copper catalysts bearing L1 or L7 show little difference in regio-(3a:4a) or chemoselectivity (3a/4a:5a) for the addition of either (E)-1 or (Z)-1 to imine 2a (compare entry 1 with 3 and entry 2 with 4). The same major enantiomer of anti-3a is formed with L1 regardless of azatriene geometry (>99:1 er, entries 1 and 3). Likewise, the L7-derived catalyst leads to 97:3 er in favor of the same major enantiomer of syn-3a beginning with either azatriene stereoisomer (entries 2 and 4). Diastereoselectivity is largely unaffected. We also measured the er of the minor diastereomer of the reactions. Somewhat surprisingly we discovered that it is formed with poor enantioselectivity in each case. Additionally, we stopped the reactions of both (E)-and (Z)-1 after 30 seconds with the Cu-Ph-BPE catalyst. There was approximately 60% conversion to anti-3a but none of the recovered azatriene had undergone stereochemical inversion in either case, suggesting CuH insertion is irreversible. To examine the azatriene aryl groups' influence upon product distribution and stereoselectivity, we prepared o-tolyl containing 12 and carried out reductive coupling with imine 2a (Table 5). In both cases, 6,5-addition product 14 is favored over 1,2-diamine 13, significantly so with t-Bu-BDPP (1:9.5 13:14, entry 2). Diamine 13 is obtained in low dr and 14 with modest selectivity. [a] Reaction with 0.1 mmol imine 2a. Entry 1 run under the conditions of Table 1, entry 1 and entry 2 under the conditions of Table 1, entry 9. [b] Determined by 500 MHz 1 H NMR spectroscopy of the unpurified mixture. We were able to obtain an X-ray crystal structure of the major stereoisomer of 4k (Figure 2), which is the major product of azatriene (1) reductive coupling with the o-bromo imine (Table 2). The observed stereochemistry indicates that the allyl-copper that leads to 4 has copper bound to the same face as that which leads to 3 and that imine facial selectivity is the same in both instances. The stereoconvergence of the (E)-and (Z)-azatriene isomers with each catalyst might be explained by several mechanistic possibilities, while the diastereodivergence observed for the two catalysts suggests a mechanistic dichotomy in the C-C bond-forming step. Furthermore, the profound diamine diastereoselectivity dependence on the imine electronics observed with the Ph-BPE-derived catalyst is significantly different from our prior azadiene additions to N-phosphinoyl imines with the same catalyst, where the anti-diamine was obtained with >20:1 dr in all cases. We propose that although both azatriene isomers 1 may undergo migratory insertion to the CuH species derived from either ligand with olefin facial selectivity, that is irrelevant as all possible stereoisomers of allyl-copper I can equilibrate through (E,E)-III via intermediates II (Scheme 4, left). These equilibria are likely faster than the addition of any species to the imine (Curtin-Hammett conditions) and, with the allyl-copper formation irreversible, provides the most likely explanation for the data in Table 2. The mechanism for C-C bond formation with each catalyst is less certain. In both instances, we propose a closed transition state, and our working hypothesis is shown in Scheme 4 (right). With Ph-BPE (L1), we suggest that reaction takes place through O-coordination of the imine [28c] (IV) but with t-Bu-BDPP (L7) via coordination of the imine's nitrogen atom (V). Therefore, the stereochemical outcome with L1 can be explained by -addition of (S,E)-II to the imine's Re face (IV), whereas the L7-promoted reaction takes place by -addition of (R,E)-I to the same face of the imine (V). From the phosphine ligands we have examined for this transformation, it is clear that Ph-BPE is an outlier in favoring the anti-diamine to any degree. The product stereoisomer observed is the same as in our previous Cu-Ph-BPE-catalyzed azadiene couplings with this class of imines, which deliver -alkyl diamines, suggesting a similar addition mode; however, in the earlier chemistry, there was no dr dependence on imine electronics. These data indicate a mechanistic pathway towards syn-diamines available to Cu-L1 with azatrienes but not azadienes, likely a -addition mode via N-coordination of the imine (i.e., V). The significant, positive  observed at lower  values in the Hammett plot (Figure 1A) implies that C-C bond formation is the diastereodetermining step, with addition through IV becoming more stabilized compared to the alternative as the imine becomes more electrophilic. At higher p-values, the negative  is consistent with imine coordination becoming diastereodetermining. Therefore, the most electrophilic imines become less discriminating in their coordination with and subsequent addition to the myriad allyl-copper species available. The t-Bu-BDPP reactions display a similar electronic trend although the break in the plot occurs with electron-neutral imines (Figure 1B). Furthermore, although the right-hand half of the plot has a comparable negative  value to the Ph-BPE reactions, the correlation at small  values shows a significantly smaller positive . It may be that the anti diamines formed with t-Bu-BDPP also arise through intermediate IV although several possibilities exist. For example, the path to the anti diamine may not involve O-coordination of the imine but rather a different -addition mode, such as from (S,Z)-I, to an N-coordinated imine. It should be noted that since the minor diastereomer of 3 with each ligand is racemic, the stereodetermining step for the minor three stereoisomers in the coupling have similar free energies. Further evidence in support of these two addition models can be found in the imine coupling of azadiene 15 with the Cu-t-Bu-BDPP catalyst (Scheme 5). Under our previously established conditions for this transformation with Ph-BPE, anti-diamine 16 is obtained as the major isomer (5:1 anti:syn), similar selectivity to what we observe with DCyPE (3:1 anti:syn). Thus, without the possibility of N-coordination of the imine, the major pathway funnels the azaallyl-copper species through an O-coordination/-addition mode. The majority of couplings lead to products that exclusively contain an (E)-alkene; however pyrrolo imine 2m (Table 2), alkyl aldimine 9 (Scheme 2), ketimine 11 (Scheme 2), and the p-NMe2 and p-NHPh aryl aldimines utilized in the Hammett study (Figure 1) all afford measureable quantities of the (Z)-isomer. Although the reason for alkene stereochemical erosion is unclear, the phenomenon appears to be tied to imine electrophilicity as these five partners are among the least electrophilic we examined. The shift in regioselectivity with di(o-tolyl)azatriene 12 (Table 5) towards 6,5-addition product 14 with both catalysts and the poor diastereoselectivity observed for 1,2-diamine 13 implies a disruption in the allyl-copper equilibria due to added steric hindrance in II and III (Scheme 4) compared to azatriene 1. The stereochemistry of amine 4k (Figure 2), obtained with Ph-BPE, can be explained either by -addition of (S,E)-II to the imine (versus -addition IV) or by an selective addition of (S,E)-I. The high selectivity for 14 with t-Bu-BDPP is somewhat puzzling as hindered ortho-substituted N-phosphinoyl imines lead to syn-diamines 3j-k (Table 2) as the exclusive products (reaction through V). It may be that (R,E)-I is less accessible when employing 12 (versus 1) because irreversible CuH insertion to the azatriene initially occurs to furnish (S,E)-I. ## CONCLUSION We have developed the first examples of Cu-catalyzed diastereodivergent and enantioselective reductive coupling reactions. Through the use of a new umpolung reagent, 2-azatrienes, we have successfully prepared both synand anti-diamines through addition to N-phosphinoylimines. The synthesis of the syn-isomers was enabled by the bis(phosphine) t-Bu-BDPP, the first use of this ligand in CuH processes. Ongoing work is dedicated to uncovering more details of the mechanism of this reaction and to the development of other transformations with 2-azatrienes. ## Corresponding Authors *xxshao@hznu.edu.cn *steven.malcolmson@duke.edu ORCID: 0000-0003-3229-0949

chemsum

{"title": "A Diastereodivergent and Enantioselective Approach to syn-and anti-Diamines: Development of 2-Azatrienes for Cu-Catalyzed Reductive Couplings with Imines that Furnish Allylic Amines", "journal": "ChemRxiv"}

dopamine-grafted_heparin_as_an_additive_to_the_commercialized_carboxymethyl_cellulose/styrene-butadi

## Abstract: Graphite is used commercially as the active material in lithium ion batteries, frequently as part of a graphite/SiO x composite. Graphite is used in conjunction with SiO x to overcome the limited energy density of graphite, and to lessen the adverse effects of volume expansion of Si. However, electrodes based on graphite/SiO x composites can be made with only 3-5 wt % SiO x because of the increased failure of electrodes with higher SiO x contents. Here, we developed a new polymer binder, by combining dopamine-grafted heparin with the commercial binder carboxymethyl cellulose (CMC)/ styrene butadiene rubber (SBR), in order to more effectively hold the SiO x particles together and prevent disintegration of the electrode during charging and discharging. The crosslinking using acidbase interactions between heparin and CMC and the ion-conducting sulfonate group in heparin, together with the strong adhesion properties of dopamine, yielded better physical properties for the dopamine-heparin-containing CMC/SBR-based electrodes than for the commercial CMC/SBR-based electrodes, and hence yielded excellent cell performance with a retention of 73.5% of the original capacity, a Coulombic efficiency of 99.7% at 150 cycles, and a high capacity of 200 mAh g −1 even at 20 C. Furthermore, a full cell test using the proposed electrode material showed stable cell performance with 89% retention at the 150 th cycle.Lithium-ion secondary batteries are being intensively studied due to their potential uses as large-scale energy storage devices in applications such as electric vehicles as well as energy storage systems 1-4 . However, currently available lithium-ion secondary batteries use graphite as the anode material, and as a result do not have sufficient energy density to efficiently power such high-energy applications.Much effort is hence being devoted to increasing the energy density levels of lithium-ion batteries. To achieve this, materials based on Li alloys, such as lithium-tin and lithium-silicon, have attracted much recent attention as alternate anode materials due to their high capacities [5][6][7][8][9] . For example, the silicon electrode has a high theoretical capacity, more than ten times greater than that of the currently used anode material for lithium-ion batteries (i.e., graphite), which has a theoretical capacity of only about 372 mAh g −1 . However, silicon has inherent problems due to volume expansion and shrinkage of about 400% during the charge-discharge process of Si + 4.4Li ⇆ Li 4.4 Si 5,6 . When such changes in volume are repeated, mechanical stresses build up that damage the anode, induce its components (e.g., conducting agents and binders) to separate from one another, and cause loss of electrical pathways by isolating Si particles having insulation properties, leading to rapid deterioration of the battery system . Thus, many researchers have examined modified forms of silicon, for example, nano-sized particles, in order to reduce the stress caused by the volume expansion of Si 13,14 . The preparation of these nano-architectured silicon anode materials, however, inevitably increases the production costs. Silicon oxides (SiO x ), on the other hand, are relatively inexpensive to produce and have been commonly used as composite materials with Si to lessen the adverse effects of the volume expansion of Si. In these composites, either crystalline or amorphous Si cores are uniformly dispersed in a SiO 2 matrix, and the stable lithium-silicates formed after lithiation of this material have been shown to reduce the magnitude of the changes in the volume of Si during cycling 12, . Indeed, in currently used and commercialized active materials based on graphite, replacing some of the graphite with SiO x has been found to be the most practical way of overcoming the limited energy density of the graphite. The SiO x content of the active material, however, has been limited to at most 3-5 wt.% due to the rapid increase in electrode failure with increasing SiO x content as a result of the above-mentioned side reactions of Si. Aside from the problems directly related to the silicon itself, developing a strong polymer binder that can endure large volume changes of Si and prevent electrodes from rupturing has also been widely pursued recently. Polymers such as poly(acrylic acid) (PAA) that consist of many carboxylic acid functional groups 18,19 and polysaccharides such as alginates 20,21 , Xanthan gum 22 and Pullulan 23 have been shown-due to their excellent physical properties resulting from their strong interactions with OH groups on silicon particles-to be of use for increasing the lifespan of Si anodes. These polymers may thus be good candidates as polymeric binders for Si as well as other high-capacity anode materials such as Sn and Ge, and may even be a viable substitute for carboxymethyl cellulose (CMC)/styrene butadiene rubber (SBR), the binder currently used in commercial graphite anodes. In addition, cross-linkages via covalent bonds or even noncovalent ones such as hydrogen bonds have been exploited to further improve the physical properties of these polymer binders, yielding markedly improved cell performance. We report herein a new polymer binder based on dopamine-functionalized heparin/CMC/SBR, in which dopamine-grafted heparin was used as an additive in the commercially available CMC/SBR binder, and show that it improves the performance of SiO x /graphite composite-based anodes. Heparin is a highly sulfonated, biocompatible, water-soluble, and naturally derived polysaccharide. It has various functional groups, which have contributed to its effectiveness and hence wide use as an anticoagulant and inhibitor of both angiogenesis and tumor growth 32 . Although heparin is difficult to apply directly as a binder to Si anodes due to its poor physical properties when used alone 33 , the many functional groups of heparin allow it to interact with CMC. These functional groups include amines, which can reversibly interact with the acid groups in CMC through hydrogen bonding, and sulfonates, which can further contribute to Li + conduction. In addition, dopamine, inspired by its role as an adhesive in mussel foot protein, was further incorporated onto heparin due to dopamine's well-known excellent adhesion to various surfaces including Si anodes 34 . Therefore, we produced dopamine-functionalized heparin and added it to the CMC/SBR binder in order to (1) mechanically strengthen this binder by crosslinking, (2) improve the adhesion property of the corresponding binder to Si and hence suppress the volume expansion of Si, and (3) endow the crosslinked binder system with the ability to conduct ions. ## Results Electrochemical performance and failure mode analyses. The addition of dopamine to heparin was accomplished via the N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC) coupling reaction, as shown in Fig. 1. To prevent the oxidation of dopamine, the pH was maintained at 6.0 using a phosphate buffer solution (PBS) while the reaction proceeded. A structural analysis of the dopamine-functionalized heparin (dopamine-heparin, 1) was carried out by performing UV-Vis and FT-IR spectroscopy investigations. The UV-Vis spectrum of 1 (Fig. 2a) showed a characteristic absorption peak at a wavelength of about 280 nm, confirming that the catechol moiety of dopamine was successfully grafted onto heparin 32,34 . In addition, the absence of any additional peaks at wavelengths greater than 300 nm indicated that no undesired oxidation of dopamine occurred 32 . The amount of dopamine incorporated was measured to be 11 wt.% of the total weight of dopamine-heparin 1, as determined by carrying out a UV-Vis spectroscopic quantitative analysis using dopamine standard solutions (see Supplementary Fig. S1) 32 . Further analyses of 1 using FT-IR spectroscopic investigations confirmed the structure (Fig. 2b): the C = O stretching peak of heparin at 1620 cm −1 shifted to 1640 cm −1 after conjugation of dopamine due to the formation of amide bonding. In addition, a peak corresponding to the secondary amide NH bending appeared at 1560 cm −1 32,35 , indicating that an amide bond successfully formed between the dopamine and heparin. Furthermore, a new peak corresponding to the aromatic OH stretching of the catechol moiety appeared at 1376 cm −1 . Preparation of electrodes and analyses of the mechanical properties. Electrodes were fabricated using a composite of SiO x and graphite as the active material, Super-P as the conducting agent, and CMC/SBR, heparin/CMC/SBR or dopamine-heparin/CMC/SBR as the polymer binder in a ratio of active material: conducting agent: binder of 80:10:10 by weight. The weight ratio of CMC to SBR was 1:2, and 10 wt.% of heparin or 1 relative to CMC was added to CMC/SBR for the fabrication of the heparin/CMC/SBR-or dopamine-heparin/CMC/ SBR-based electrode, respectively. The results were compared with those obtained using the electrode containing only CMC/SBR as the polymer binder. The binding affinities for the three electrodes (CMC/SBR, heparin/CMC/SBR and dopamine-heparin/CMC/ SBR) were then investigated by carrying out 180-degree peel tests (Fig. 3a). In the case of the SiO x /graphite electrode composed of the pristine CMC/SBR polymer binder, an average binding affinity of 0.757 N was obtained, while this value increased to 0.907 N when heparin was added to CMC/SBR (heparin/CMC/SBR electrode). A dramatic increase in the binding affinity was further obtained for the electrode with the CMC/SBR binder containing dopamine-heparin (dopamine-heparin/CMC/SBR), with a measured value of 1.287 N. These results suggested that the addition of heparin to CMC/SBR mechanically strengthened the electrode due to the formation of physical interactions between CMC and heparin, and that the durability of the electrode in the presence of mechanical stress was further increased by the strong adhesion resulting from the inclusion of dopamine in the dopamine-grafted heparin (dopamine-heparin, 1) 27,34 . We further used an optical microscope to examine the surface of the tape detached from the electrode after the peel test to determine whether the addition of dopamine-heparin (to form a dopamine-heparin/CMC/SBR binder) reduced the amount of Si exfoliated from the electrode. As shown in Fig. 3b, the amount of the electrode material detached from the current collector was inversely related to the adhesive force generated by the binder in Fig. 3a 28 . The tapes pulled off the pristine CMC/SBR-based electrode were stained with a lot of slurry and transmitted little light, indicating that much of the Si material in these cases became detached from the electrode and transferred to the tape. In contrast, the tapes pulled off the heparin/CMC/SBR-based electrode were stained with only a very small amount of slurry, and were observed to transmit a considerable amount of light, indicative of the better ability of heparin/CMC/SBR than of pristine CMC/SBR to prevent material from being detached from the electrode and consistent with the stronger adhesion force of the heparin/CMC/SBR binder based on the peel test. And this trend continued with the inclusion of dopamine: tapes pulled off the dopamine-heparin/ CMC/SBR-based electrode, showed the least amount of slurry and transmitted the most light, consistent with the dopamine-heparin/CMC/SBR binder having displayed an adhesion force greater than those of the other binders. ## Electrochemical properties. After investigating the mechanical properties of the electrodes containing the heparin/CMC/SBR and dopamine-heparin/CMC/SBR binders, we then assessed the electrochemical performances of their corresponding cells. We performed the electrochemical evaluations at a low rate (specifically with a discharge rate of 0.2 C and charge rate of 0.5 C, where 1 C equals 450 mA g −1 ), and compared the results with the cell performance of the pristine CMC/SBR-based electrodes (see Fig. 4). The voltage profiles of the prepared cells during the 1 st cycle showed similar trends irrespective of the binder (Fig. 4a), suggesting that the addition of heparin or dopamine-heparin (1) did not cause any particular side reactions. In addition, the voltage profiles of both the heparin-and dopamine-heparin-based electrodes showed lower resistance during lithiation than did the electrode prepared from CMC/SBR. The enhanced physical properties of the heparin and dopamine-heparin binder systems have been suggested to result in less polarization due to the reduction of the contact resistance of the electrode 36 . Having the highest mechanical strength, the dopamine-heparin-based electrode indeed showed the lowest resistance. In addition, the specific discharge capacities of the cells prepared using the three binders were also determined in order to assess the effect of the binder on cyclability (Fig. 4b). The specific capacity of the CMC/SBR-based electrode started to deteriorate rapidly after 50 cycles, and at 100 cycles showed a value of 286 mAh g −1 and a retention of 62.3% of the original specific capacity. After about 100 cycles, this electrode showed particularly poor cycling stability. This poor stability was attributed to the CMC/SBR binder, with its poor adhesion properties, no longer being able to accommodate the repeated changes in the volume of the silicon electrode. The cell performances of the electrode made of the heparin/CMC/SBR binder and that made with the dopamine-heparin/ CMC/SBR binder, however, were much better than the cell performance of the electrode made with the pristine CMC/SBR binder. The superior performance of the former two cells was attributed to their enhanced physical properties. A specific capacity of 380 mAh g −1 , representing a retention of 81.3% of the original capacity, was observed at 100 cycles, and a specific capacity of 268 mAh g −1 and retention of 57.3%, were observed at 150 cycles for the electrode including heparin/CMC/SBR. The dopamine-heparin/CMC/SBR-based binder system showed the best cell performance. Although the specific capacity and retention values of the dopamine-heparin/CMC/ SBR-based binder system at 100 cycles, with values of 378 mAh g −1 and 81.3%, respectively, were similar to those of the heparin/CMC/SBR-based system, a much greater reversible capacity of 343 mAh g −1 and retention of 73.5% were obtained at 150 cycles for the dopamine-heparin/CMC/SBR-based system than for the heparin/CMC/ SBR-based system. Since the three cells were subjected to the same conditions, the superior specific capacity and percent retention of original specific capacity up to the 150 cycles for the cell having the dopamine-heparin/ CMC/SBR binder were therefore attributed mainly to the superior adhesion properties of this binder material. The dopamine-heparin-CMC/SBR electrode also displayed excellent Coulombic efficiency (Fig. 4c). It displayed an initial Coulombic efficiency (after the formation cycles) of 97.4%, a value of 98.9-99.5% at the 20 th cycle, and very high value of 99.7% at the 150 th cycle. These values were higher than those of the CMC/SBR-and heparin/CMC/SBR-based electrodes. We also performed rate capability tests for the three types of electrode (Fig. 4d). Their 0.1 C-capacities were similar to those in the formation stage of the cycling performance experiment. The CMC/SBR-based electrode (having the worst mechanical properties) showed a much larger decay of specific capacity as the C rate was increased; the performance of this electrode indicated that it would be difficult to use above 15 C. The heparin-based electrode, in contrast, exhibited a stable capacity up to 3 C. Although a constant capacity loss occurred when the rate was increased to above 3 C, the cell showed stable operation up to 20 C. The dopamine-heparin electrode showed stable performance up to 5 C, and showed a reasonably good cell performance of about 200 mAh g −1 even at 20 C. We attributed the enhanced rate properties of the heparin and dopamine-heparin electrodes to the improved physical properties of these binders, which helped to prevent the conducting agents and active materials from detaching from the electrodes even at high C rates. In addition, the functional groups in the heparin structure were thought to further provide a lithium transfer pathway capable of conducting Li + ions 28, . Electrochemical impedance spectra (EIS) were also acquired for the electrodes with and without the dopamine-heparin, and the overall resistance on the electrode surface was found to be smaller when the dopamine-heparin was included (Fig. 5). Since there were no other factors besides the addition of dopamine-heparin, this result was thought to be due to the formation of a solid electrolyte interphase (SEI) layer. Surface analyses using SEM and XPS further confirmed these results (see below). Finally, a full cell, having a LiNi 0.6 Co 0.2 Mn 0.2 O 2 positive electrode and the composite negative electrode containing the dopamine-heparin/CMC/SBR binder, was tested to evaluate the binder in a practical battery application (Fig. 6). To ensure reliable performance, one formation cycle was performed after the fabrication of the cell, followed by charging and discharging at 0.5 C between 3.0 and 4.1 V. Even though the initial high specific charging capacity of the full cell (Fig. S2), the obtained capacities at subsequent cycles suddenly fall because of the initial irreversible reaction of the negative electrode limiting the number of reversible lithium ions and the high kinetic resistance from the high amount of the active material causing high polarization. On the other side, an excellent cell performance, with retention of 92% of the initial performance at the 100 th cycle and of 89% of the initial performance at the 150 th cycle, was obtained, strongly suggesting that the proposed binder system has potential for use in practical Si-based lithium-ion batteries. In addition, the voltage profiles of the full cell (Fig. S2) showed 66.5% Coulombic efficiency at the pre-cycling step, which can be attributed to irreversible SiO x conversion and SEI formation during pre-cycling. After the pre-cycling, however, the system required only two cycles to show >99% Coulombic efficiency, which was maintained for more than 100 cycles (Fig. 6). ## Morphological analysis. The CMC/SBR-based and dopamine-heparin/CMC/SBR-based electrodes were visualized before and after cycling using scanning electron microscopy (SEM), as shown in Fig. 7, in order to determine the effects of the binder material on the morphology of the electrode surface. The two electrodes showed similar surface morphologies immediately after fabrication, but quite different morphologies after 150 cycles, at which point the surface of the pristine CMC/SBR electrode was relatively thickly and unevenly covered by an SEI layer, while the surface of the dopamine-heparin/CMC/SBR-based electrode maintained a relatively clean, uniform and porous appearance. Energy dispersive spectrometry (EDS) analyses of the prepared electrodes (based on analyses of mapped images of oxygen atoms, which originated from SiO x and the polymer binder, and of silicon atoms, which originated from SiO x ) showed that the materials comprising the electrodes were well dispersed for both electrodes irrespective of the addition of the dopamine-heparin to the CMC/SBR (Fig. S3). However, the EDS results showed that the portion of the weight of the CMC/SBR electrode due to oxygen was about 22% greater after cycling than before cycling, whereas the portion of the weight of the dopamine-heparin-based electrode due to oxygen was only about 20% greater after cycling (Fig. S4). Since changes in oxygen content as a result of cycling have been suggested to originate from the formation of the SEI layer and/or the decomposition of electrolytes, these results might indicate that less electrolyte decomposed when dopamine-heparin was present. We also carried out XPS analyses of the surfaces of the CMC/SBR and dopamine-heparin/CMC/SBR electrodes. For the CMC/SBR electrode, the peak intensity corresponding to C 1s was much lower after 150 cycles than before cycling, whereas that for O 1s significantly increased, confirming the formation of a thick SEI layer caused by the side reactions of the carbonate-based electrolytes (Fig. S5). In contrast, in the case of dopamine-heparin/ CMC/SBR electrode, the changes in the peak intensities of the C 1s and O 1s peaks were not large even after 150 cycles, suggesting that there was relatively little SEI formation 38,42 . The XPS results were found to be in good agreement with those of the SEM analyses. ## Discussion We developed a new polymer binder system based on dopamine-grafted heparin (dopamine-heparin) as an additive to the CMC/SBR binder for SiO x /graphite composite electrodes in order to (1) produce physical crosslinking using acid-base interactions between heparin and CMC, (2) increase the ion conductivity using the sulfonate groups of heparin, and (3) strengthen the adhesion of the binder to the electrode due to the inclusion of dopamine. While the addition of heparin to the CMC/SBR binder (to convert commercialized CMC/SBR-based electrodes to the heparin/CMC/SBR-based electrodes) improved the physical properties of the electrodes and hence the overall performance of the corresponding cells, even better physical properties were achieved when dopamine-heparin was added (to produce dopamine-heparin/CMC/SBR electrodes). Also, while previously described electrodes could be made with only 3-5 wt.% SiO x because of the increased failure of these electrodes with increasing SiO x content due to side reactions of Si, we were able to achieve excellent cell performance for the dopamine-heparin/CMC/SBR-based electrode with SiO x /graphite composite active material containing up to 7 wt.% SiO x . Specifically, the resulting dopamine-heparin/CMC/SBR-based electrode displayed a specific capacity of 343 mAh g −1 and retention of 73.5% of the original capacity, and Coulombic efficiency of 99.7% after 150 cycles, together with a high capacity of 200 mAh g −1 even at 20 C. Furthermore, a stable cell performance with 89% retention at the 150 th cycle was achieved in a full cell test using this electrode. Overall, the inclusion of dopamine-grafted heparin not only enhanced the physical properties of the corresponding dopamine-heparin/ CMC/SBR-based electrode but also its ionic conductivity, and hence demonstrated its potential for use in practical Si-based lithium-ion batteries. ## Methods Materials. Heparin sodium salt, N-hydroxysuccinimide (NHS), and dopamine hydrochloride were purchased from Sigma-Aldrich. N-(3-Dimethylaminopropyl)-N'-ethylcarbodiimide hydrochloride (EDC), sodium dihydrogen phosphate and sodium hydrogen phosphate were purchased from Alfa-Aesar. Dimethyl carbonate and electrolyte (1 M LiPF 6 in ethylene carbonate/ethylmethyl carbonate (EC:EMC = 3:7 v/v) with 10 wt.% fluoroethylene carbonate (FEC)) were purchased from PanaxEtec. Silicon monoxide (SiO x ) was obtained from Osaka Titanium Technologies. Carboxymethyl cellulose (CMC) was purchased from Nippon Paper Chemicals. Styrenebutadiene rubber (SBR) was purchased from ZEON. Synthesis of dopamine-grafted heparin 1. Heparin (1.0 g) was added to an aqueous phosphate buffer solution (PBS, 100 mL, pH = 6.0) in a round-bottom flask, and this mixture was stirred until a homogeneous solution was obtained. Thereafter, EDC (1.63 g, 8.5 mmol) and NHS (0.98 g, 8.5 mmol), dissolved in a PBS solution (pH = 6.0), were subsequently added to the heparin solution. Finally, to this mixture was added an aqueous solution of dopamine (0.5 g, 2.6 mmol) in PBS (pH = 6.0). The resulting mixture was stirred for 9 h at r.t. After completion of the reaction, the dopamine-grafted heparin was purified from the reaction mixture by subjecting this mixture to dialysis twice using a cellulose membrane (MWCO = 12 kDa, Sigma-Aldrich) in distilled water for 6 h, and the product was collected by freeze-drying. The relative amount of dopamine in the dopamine-heparin 1 was calculated by performing quantitative analysis using UV-Vis spectroscopy (Fig. S1) 32 . Standard dopamine solutions were prepared for a calibration curve, and then the molar absorptivity was obtained from this calibration curve slope. Finally, the absorbance of the dopamine-heparin solution was measured, and the dopamine content was determined using the following Equation (1): where A is the absorbance, Ɛ is the molar absorptivity, b is the path length of the sample, and c is the concentration of the sample. Fabrication of the SiO x /Graphite composite electrode. The composite electrode composition was 80:10:10 wt.% of active materials: conducting agent:polymer binder, with 7 wt.% of the active materials being SiOx and the remainder being graphite. The conducting agent is Super-P. The binder compositions tested were a 1:2 weight ratio of CMC:SBR and a 0.1:0.9:2 weight ratio of heparin(-dopamine):CMC:SBR. To make a SiO x / graphite composite electrode, we first mixed all of the solid materials, i.e., the graphite and SiO x active materials, the conducting agent and CMC binder powder in a mortar based on the designed composition. This mixture of solids was combined with enough distilled water and a solution of SBR to make a slurry with the desired viscosity, which was coated onto a Cu foil. The as-formed electrode produced in this manner was dried in a convection oven at 80 °C for 30 minutes. The prepared electrode was made to have a mass loading of 5 mg cm −2 . The prepared electrodes were dried under vacuum at 120 °C for 6 h to eliminate residual water. The theoretical capacity of the SiO x /graphite composite electrode was determined to be 450 mAh g −1 by weight ratio of active materials, whereas the capacities of graphite and SiO x have been determined to be 372 mAh g −1 and 1500 mAh g −1 , respectively. Characterization and Measurements. FT-IR spectra were recorded on a PerkinElmer Spectrum Two ATR spectrometer. UV-vis spectra were recorded on a PerkinElmer Lambda 365 spectrometer. Scanning Electron Microscopy (SEM) was performed using a JEOL JSM-7000F instrument. To see the surface morphology, the cycled Si electrodes were washed with dimethyl carbonate before being imaged. For evaluating the mechanical properties of the electrodes with various polymer binders, the prepared electrodes were each cut into rectangular shapes (1.2 cm × 3.0 cm) and attached to 12-mm-wide 3 M tape. The peel strength of each tested electrode specimen was then recorded with a universal testing machine (UTM, Shimadzu EZ-L) by pulling the tape at a constant displacement rate of 30 mm min −1 . ## Electrochemical performances and failure mode analyses. To evaluate the electrochemical performances of the prepared electrodes, 2032-type coin cells were assembled in an Ar-filled glove box, using multi-layered porous PE/PP/PE as the separator, a lithium disc as the counter electrode, and 1 M LiPF 6 in ethylene carbonate/ethylmethyl carbonate (EC:EMC = 3:7 v/v) with 10 wt.% fluoroethylene carbonate (FEC) (PanaxEtec) as an electrolyte. Galvanostatic discharge−charge cycling was then performed using a CPS-Lab battery cycler (Basytec) at 25 °C in a temperature-controlled chamber. All of the cells were repeatedly discharged to 0.005 V vs Li/Li + and charged to 2.0 V vs Li/Li + at a constant C-rate of C/20 lithiation (discharge)-C/10 delithiation (charge) in the first cycle and C/10 lithiation-C/10 delithiation in the next two cycles for pre-cycling, and then at C/5 lithiation-C/2 in the following cycles. Rate capability tests were carried out by repeating discharging to 0.005 V vs Li/Li + and charging to 2.0 V vs Li/Li + . The pre-cycling was performed before the cyclability test, and the current for lithiation was fixed at C/20, and the currents for delithiation were varied from 1/2 to 20 C (C/2, 1 C, 2 C, 3 C, 5 C, 10 C, and 20 C). For postmortem analyses, the cycled cells were disassembled in an Ar-filled glove box and recovered electrodes were washed with dimethyl carbonate to eliminate residual electrolyte. Then, the surface morphologies of cycled electrodes were investigated by performing scanning electron microscopy (SEM, JEOL JSM-7800F). The EIS was measured at a frequency range of 10 mHz-100 kHz with an AC amplitude of 10 mV at 0.01 V. Full cell measurement. The full-cell was assembled using an N/P ratio of 1.1. The same SiO x /Graphite composite electrode was used for the negative electrode and fabricated 94:3:3 ratio of active materials:super-P:binder. The composite cathode electrode being fabricated with a 90:5:5 ratio of LiNi 0.6 Co 0.2 Mn 0.2 O 2 :super-P:PVdF were used. The full cells were repeatedly charged to 4.1 V and discharged to 3.0 V at a constant C-rate of C/10 for both the 1 st charge and discharge as a pre-cycling step and at C/5 of lithiation and C/2 of delithiation in the subsequent cycles. And 1 C was determined to be 170 mA per gram of the positive electrode material.

chemsum

{"title": "Dopamine-grafted heparin as an additive to the commercialized carboxymethyl cellulose/styrene-butadiene rubber binder for practical use of SiOx/graphite composite anode", "journal": "Scientific Reports - Nature"}

discovery_of_sars-cov-2_m^pro_peptide_inhibitors_from_modelling_substrate_and_ligand_bindi

## Abstract: The main protease (M pro ) of SARS-CoV-2 is central to viral maturation and is a promising drug target, but little is known about structural aspects of how it binds to its 11 natural cleavage sites. We used biophysical and crystallographic data and an array of biomolecular simulation techniques, including automated docking, molecular dynamics (MD) and interactive MD in virtual reality, QM/MM, and linearscaling DFT, to investigate the molecular features underlying recognition of the natural M pro substrates.We extensively analysed the subsite interactions of modelled 11-residue cleavage site peptides, crystallographic ligands, and docked COVID Moonshot-designed covalent inhibitors. Our modelling studies reveal remarkable consistency in the hydrogen bonding patterns of the natural M pro substrates, particularly on the N-terminal side of the scissile bond. They highlight the critical role of interactions beyond the immediate active site in recognition and catalysis, in particular plasticity at the S2 site.Building on our initial M pro -substrate models, we used predictive saturation variation scanning (PreSaVS) to design peptides with improved affinity. Non-denaturing mass spectrometry and other biophysical analyses confirm these new and effective 'peptibitors' inhibit M pro competitively. Our combined results provide new insights and highlight opportunities for the development of M pro inhibitors as anti-COVID-19 drugs. ## Introduction Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the etiological agent of coronavirus disease 2019 (COVID-19) that caused the World Health Organization to declare a global pandemic in March 2020. At the time of writing, >210 million COVID-19 cases and >4.4 million deaths have been reported worldwide. 1 A key step in maturation of SARS-CoV-2, a single-stranded positive-sense RNA virus, is hydrolysis of its polyproteins pp1a and pp1ab. Most of the cleavage events-at 11 sites-are performed by the SARS-CoV-2 main protease (M pro ; 3 chymotrypsin-like or 3CL proteinase, 3C-like protease, 3CL pro ; or non-structural protein 5, Nsp5). M pro is a nucleophilic cysteine protease, which in solution is predominantly homodimeric. Each protomer consists of three domains, and the active site contains a cysteine-histidine catalytic dyad, Cys-145 and His-41, located near the dimer interface. 2 SARS-CoV-2 M pro is 96% identical to the M pro of SARS-CoV, which causes SARS. 3 Dimerisation of M pro is proposed as a prerequisite for catalysis: the N-terminus of one protomer contributes part of the active site of the other. 4 Indeed, the monomeric form of SARS-CoV M pro is reported to be inactive. 5 Evidence from non-denaturing mass spectrometry (MS)-based assays indicates that M pro monomers are not only inactive (at least with tested substrates), but do not bind 11-mer substrates with high affinity. 6 SARS-CoV-2 M pro and SARS-CoV M pro have similar substrate specifcities, both recognizing the motif: [P4:Small] [P3:X] [P2:Leu/Phe/Val/Met] [P1:Gln]Y[P1 0 :Gly/Ala/Ser/Asn], "Small" denoting Ala, Val, Pro or Thr; "X" any residue; and "Y" the scissile amide (Fig. 1). 7,8 In part, because such sequences are not known to be recognised by any human protease, M pro represents an attractive drug target. 4 Although no clinically approved M pro drugs are available, small molecule inhibitors and peptidomimetics have been designed to inhibit SARS-CoV M pro and, more recently, SARS-CoV-2 M pro . 11,12 Indeed, a covalent M pro inhibitor from Pfzer has recently entered clinical trials. 13,14 Multiple crystallographic and computational modelling studies concerning the M pro mechanism and inhibition are available; the CORD-19 database 25 documents many such studies. It is proposed that during M pro catalysis, His-41 deprotonates the Cys-145 thiol, which reacts with the carbonyl of the scissile amide to give an acyl-enzyme intermediate. This intermediate is stabilised by a hydrogen bond network that holds the scissile amide carbonyl in an 'oxyanion hole'. The Cterminal part of the product likely leaves the active site at this stage. The acyl-enzyme intermediate is subsequently hydrolysed with loss of the N-terminal product regenerating active M pro . Computational and mechanistic studies on SARS-CoV M pro26-29 and SARS-CoV-2 M pro30-32 suggest that in the resting state His-41 and Cys-145 are likely neutral and that the protonation states of nearby histidines (e.g. His-163, 164, and 172) affect the structure of the catalytic machinery-although it has been suggested in SARS-CoV M pro that the protonation state of the catalytic dyad may change in the presence of an inhibitor or substrate. 33 A different picture has been obtained from neutron crystallographic studies, which indicate that an ion pair form of the dyad is favoured at pH 6.6. 34 While neutron crystallography, in principle, enables the direct determination of hydrogen atom positions, questions remain about how pH and the presence of active site-bound ligands influence the precise-and likely dynamic-protonation state(s) of the dyad. Important questions remain regarding M pro catalysis, including to what extent the active site protonation state, solvent accessibility, induced ft, and substrate sequence influence activity. The lack of this knowledge makes it difficult to carry out effective computational studies on M pro catalysis and inhibition. Fig. 1 Substrates processed by SARS-CoV-2 M pro . (a) The 11 SARS-CoV-2 M pro cleavage sites, and the corresponding 11-residue peptides, s01-s11; positively/negatively charged residues are blue/red, respectively; histidine is purple; residues with polar sidechains are green; and cysteine is yellow. (b) Comparison between the 11 substrate sequences (generated by WebLogo) 9 highlighting the completely conserved Gln at P1 and the highly conserved Leu at P2. (c) View of an energy minimised model, built using apo M pro (PDB: 6yb7, light grey surface), 10 of M pro complexed with s05 (dark grey sticks); subsites S4-S4 0 are labelled. The oxyanion hole formed by the M pro backbone NHs of Gly-143, Ser-144 and Cys-145 is cyan. (d) The reaction catalysed by M pro exemplified by s01. Substrate residues important in recognition (see main text) are highlighted. With the aim of helping to combat COVID-19, in April 2020 we embarked on a collaborative effort involving weekly virtual meetings, initially to investigate the relationship between M pro substrate selectivity and activity. We employed an array of classical molecular mechanics (MM) and quantum mechanical (QM) techniques, including non-covalent and covalent automated docking, molecular dynamics (MD) simulations, density functional theory (DFT), combined quantum mechanics/ molecular mechanics (QM/MM) modelling, and interactive MD in virtual reality (iMD-VR). Our results provide consensus atomic-level insights into the interactions of M pro with 11residue peptides derived from the 11 natural cleavage sites (named "s01" to "s11", in order of occurrence in the viral polyprotein, Fig. 1a). The identifcation of key interactions between M pro and its substrates, together with analysis of fragment/ inhibitor structures, 35,36 led to the design of peptides proposed to bind more tightly than the natural substrates, several of which inhibit M pro . The results are freely available via GitHub (https://github.com/gmm/SARS-CoV-2-Modelling). ## Protonation state of the catalytic dyad The protonation state of the dyad following substrate binding was studied with s01-bound (Fig. 1a) M pro37 using QM/MM umbrella sampling simulations at the DFTB3/MM and uB97X-D/6-31G(d)/MM levels of theory (Section S1.1 †). The protonation states of nearby histidines were also evaluated (Table S2.1 and Fig. S2.5, S2.6 †). His-41 was treated as Nd-protonated in the neutral state of the dyad, as reported by Pavlova et al. to be preferred for both uncomplexed and N3 inhibitor-bound M pro based on MD studies. 30 The protonation state of the dyad-neighbouring His-163, which interacts with Tyr-161, Phe-140 and the substrate P1 Gln sidechain, was also studied. Three His-163 protonation states were considered: (i) Nd-protonated, neutral ("HID"); (ii) N3-protonated, neutral ("HIE"); and (iii) Nd and N3-protonated, positively charged ("HIP"). 38 For all three His-163 protonation states, the forward trajectories (neutral, N to ion pair, IP) showed the anionic Cys-145 form to be stabilised solely by interaction with positively charged His-41 (Fig. 2a and S2.2A-C †), and the thiolate in an unreactive conformation for nucleophilic attack. This suggests that such a zwitterionic state is transient, with concerted proton transfer and simultaneous nucleophilic attack of the thiolate onto the scissile amide carbon being more likely than a stepwise mechanism. 15 By contrast, the backwards PT trajectories (from IP to N) showed stabilisation of the Cys-145 thiolate by His-41 and the P1 Gln backbone N-H, in the case of HID-163 and HIP-163 (Fig. S2.2E, F †). For HIE-163, additional stabilisation comes from interactions with the backbone N-H and the hydroxyl of the P1 0 Ser, and an additional water that diffuses into the active site (Fig. 2b and S2.2D †). The zwitterionic state with HID-163 was less stable than the neutral state and the zwitterionic states with HIE-163 and HIP-163 (Fig. S2.1C †). This was due to perturbation of the interaction network with Tyr-161 and Phe-140, suggesting that a Ndprotonated His-163 is unlikely. Double protonation of His-163 results in a loss of both interactions in the forwards and backwards PT trajectories. Despite both HIP-163 and HIE-163 giving similar PT free energy profles, the loss of these interactions suggests HIP-163 is unfavourable for productive catalysis. These QM/MM results therefore suggest that an N3-protonated neutral His-163 is most likely. Along with conserving interactions with Tyr-161 and Phe-140, an N3-protonated His-163 also formed a hydrogen bond with the P1 Gln side chain (Fig. 2a and b), an interaction not observed in PT trajectories with HID-163 and HIP-163. Considering that DFTB3 overestimates the proton affinity of methylimidazole, it is expected that this method will overstabilise the zwitterionic state relative to the neutral state. 39 To account for this, the backwards PT reaction with a Ndprotonated His-163 was modelled at the uB97X-D/6-31G(d)/MM level of theory. This showed the zwitterionic state was 24.3 kJ mol 1 above the neutral state, an increase of 26.4 kJ mol 1 compared to DFTB3/MM (Fig. S2.3 †). Applying the free energy difference between uB97X-D/6-31G(d) and DFTB3 at each reaction coordinate value as a correction to the combined QM/MM free energy profle in the case of HIE-163, the neutral catalytic dyad is preferred, with the ion pair being 28.5 kJ mol 1 higher in energy than the neutral state (Fig. 2c). Similar results were obtained with a different QM approach (Fig. S2.4 and associated ESI Movie †). ## Models of SARS-CoV-2 M pro -substrate peptide complexes To understand substrate specifcity and to assess their relative binding affinities, we constructed 11 models of SARS-CoV-2 M pro complexed with its cleavage site-derived substrates 40 as 11-amino acid peptides, from P6 to P5 0 (Section S1.2; Fig. 1a and S2.7 †). We refer to these peptides as 'substrates' as their hydrolytic sites are all cleaved by M pro (vide infra). The substrates were modelled in crystallographic chain A of the M pro dimer (PDB: 6yb7) 10 with a neutral dyad; unless otherwise stated, all M pro residue numbers and names in the following discussions refer to chain A. Initial Three of the 11 cleavage site-derived peptides (s01, s02, and s05) were also modelled with interactive MD using virtual reality (iMD-VR), as an alternative to comparative modelling and traditional MD. iMD-VR provides an immersive 3D environment for users to interact with physically rigorous MD simulations. The three substrates were chosen because s01 and s02 have the highest relative efficiencies (of SARS-CoV M pro ) of all substrates; while s05 has the second-lowest catalytic efficiency but the same P2 and P1 residues as s01. 44 Throughout both the explicit-solvent MD and implicitsolvent iMD-VR simulations, all the substrates remained tightly bound in the active site (Fig. S2.8-S2.11 †). Substrate backbone stability was maintained especially in the central region, with only the N-and C-terminal residues showing substantial flexibility. Local sidechain fluctuations were present, notably at the solvent-facing P3 residue (Fig. 1c and S2.9 †). C-terminal P 0 -residues consistently fluctuated more than the N-terminal P-side residues, likely in part because of fewer protein-substrate hydrogen bonds on the P 0 side (vide infra). 2.2.1 Conserved hydrogen bond interactions. Crystallographic studies on SARS-CoV M pro revealed the importance of hydrogen bonds in binding of substrate s01, 37 which has the same cleavage site sequence in SARS-CoV-2 (Fig. 1a). To investigate whether this was true for SARS-CoV-2 M pro complexed with its 11 substrates, we analysed the hydrogen bond (HB) prevalence at each subsite in both explicit-solvent MD and implicit-solvent iMD-VR simulations. Twelve HBs were consistently identifed (Fig. 3) and their distance and angular distributions analysed (Fig. S2.12 †). In both explicit-solvent MD and iMD-VR simulations, all 11 substrates are primarily held in place by four consistently formed backbone-backbone HBs: Glu-166 at S3 (2 & 3) and Thr-26 at S2 0 (10 & 11; Fig. 3). Backbone HBs 1 and 12 further from the cleavage site show greater variation between the MD and iMD-VR studies. Although P1 Gln is conserved in all 11 M pro cleavage sites (Fig. 1a and b), HBs 5-9 formed in the S1 site are observed less often than 2, 3, 10 and 11, but outnumber other sites. In both types of simulations, HB 8 from the Cys-145 backbone amide is consistently formed, suggesting that this could play a fundamental role in catalysis. HB 8 forms part of the oxyanion hole, which stabilises the tetrahedral intermediate formed upon nucleophilic attack by Cys-145 Sg on the scissile amide. M pro 's exquisite specifcity for Gln at P1 is likely due to formation of HB 6 with His-163, and to a lesser extent HB 7, along with the narrowness of the S1 pocket accommodating the Gln sidechain in an extended conformation. 2.2.2 Hydrophobicity analysis. Hydrophilicity analysis shows that the S1 subsite is substantially hydrophilic in all 11 substrate-M pro complexes, while S2 is consistently hydrophobic (Fig. 3d). In accord with the MD-HB analysis (Fig. 3b), the conservation of subsite interactions decreases further from the cleavage site. Although S3 and S2 0 show a slight bias towards hydrophilic interactions, none of the other subsites show a consistent pattern. 2.2.3 Other non-covalent interactions. Beyond HBs, several other interaction types are conserved across the substrates; these were identifed by running Arpeggio 45 on snapshots extracted from the explicit-solvent MD simulations (Fig. 4). Six of the eight most common P1 interactions are present in most (i.e., $9/11) substrates. This includes the previously described HBs between the P1 backbone oxygen and the backbone NHs of Cys-145 (HB 8) and Gly-143 (HB 9) that constitute the oxyanion hole, as well as HB 6 (His-163) and HB 7 (Phe-140) that stabilise the P1 Gln sidechain. In addition, interactions with Ser-144, His-163, His-164, Glu-166 and, to a lesser extent, Phe-140 were prevalent at P1. None of the other subsites show this level of consistency in residue-level contacts, although some interactions such as hydrophobic contacts with Met-49 and Met-165 were always present at P2. Furthermore, important stabilising backbone HBs (HBs 2-3 between Glu-166 and P3; and HBs 10-11 between Thr-26 and P2 0 ) were conserved in all substrates. Finally, P 0 interactions are less common than those on the P side (Fig. 4). The same trend was found when docking s01, s02, and s05 using iMD-VR, where P 0 residues tended to be more flexible than P-side residues. We further analysed the energetic contributions of each M pro residue using the Molecular Mechanics-Generalised Born Surface Area (MM-GBSA) method (Section S2.3.1 †), which highlighted hotspot residues that were also recognised by Arpeggio as conserved contacts. 2.2.4 Density functional theory analysis of the interaction network. We performed linear-scaling DFT (BigDFT 50 ) calculations using representative snapshots extracted from explicitly solvated MD trajectories of M pro complexed with s01, s02, and s05. By automatically decomposing large molecular systems into coarse-grained subsets of atoms (or 'fragments') in an unbiased manner, 51 quantities like inter-fragment bond order and interaction strengths, E cont , can be easily calculated (Fig. 5a, S2.18-S2.21 and Section S2.4 †). This analysis supports the essential roles of Glu-166 and Thr-26, with interactions observed in all three peptides s01, s02, and s05, consistent with the HB analysis described earlier (Fig. 3). Gln-189 consistently hydrogen-bonds with P2 (HB-4) in s01 and s05, but rarely in s02. This weakening of HB-4 in P2 may be due the greater bulk of Phe in s02 (Fig. 5b). Conserved contacts are present in the three substrates between Cys-145 and both P1 and P1 0 residues. Interactions between His-41 and P2/P1 0 are observed for s01 and s05, and showing QM interaction energies between 22 selected residues of M pro and s01, s02 and s05. (b) QM interaction networks where node colour indicates interaction strength, from dark blue (strongest) through green to yellow (weakest). Square nodes denote substrate, while circular ones denote M pro . The thickness and colour of the edges show the fragment bond order between residues, a unitless measure associated with bond strength and analogous to bond order; black is strongest, orange is weakest. 51 Interaction energies and bond orders were computed using BigDFT and ensemble-averaged results of MD snapshots. between Glu-166 and P1/P3 (and, to some extent, to P4) for all three substrates. This analysis singles out the character of s02, which is dominated by the bulky character of its P2 Phe. Substitutions at P2 may have a substantial effect on the interaction network close to the catalytic site. While the P side exhibits an interconnected character especially from P1 to P4 (Fig. 5b), the network on the P 0 side has a more linear character, once again indicating that hotspot residues responsible for binding are present on the P side. Distributions of E cont are shown in Fig. S2.20. † The following trends emerge from our studies on M pro in complex with models of its 11 substrates: (i) binding stability is partly conferred by a series of HBs from P4 to P4 0 , in particular between the backbones of M pro Glu-166 and Thr-26 and substrate positions P3 and P2 0 respectively, as well as HBs involving the conserved P1 Gln sidechain; (ii) substrate residues N-terminal of the cleavage site (P side) form more, and more consistent, contact interactions with M pro compared to the P 0 side, with interactions at Met-49, Gly-143, Ser-144, Cys-145, His-163, His-164, Met-165 and Glu-166 being most conserved. We conclude that the S1 and S2 pockets are prime targets for active site substrate-competing inhibitor design due to their well-defned hydrophilic character, large energy contributions to substrate binding, and vital conserved hydrogen bonds in S1 for substrate recognition. In all cases, P1 Gln recognition is mainly driven by interactions with Gly-143, Ser-144, Cys-145 (oxyanion hole) and His-163 and Phe-140. The S2 subsite, however, is highly flexible, especially at Thr-45, Ser-46 and Met-49. Although P2 is conserved in terms of hydrophobicity (Leu, Phe, Val), the S2 pocket is highly flexible and can adapt to accommodate functional groups of varying sizes, including aliphatic and aromatic groups. The outer regions of the active site (S3-S6 and S2 0 -S5 0 ) vary in flexibility, echoing our MD simulations. ## Monitoring of substrate sequence hydrolysis by mass spectrometry To rank the SARS-CoV-2 M pro preferences for hydrolysis of the 11 cleavage sites, we monitored turnover of 11-mer peptides by solid-phase extraction (SPE) coupled to mass spectrometry (MS). Interestingly, after the N-terminal autocleavage site s01, s11 was the next preferred substrate for catalysis (Fig. S2.22 †). Peptides s06, s02, and s10 were hydrolysed less efficiently than s11. Slow turnover was observed for s07 and s09. Evidence for low turnover of s05 was obtained after prolonged incubation with M pro (9.56%) (Fig. S2.23 †). Under our standard conditions, no evidence for cleavage was observed for s03, s04, and s08. We then examined turnover under non-denaturing MS conditions using ammonium acetate buffer (Fig. 7a). Peptides s01, s06, s08, s10 and s11 evidenced fast turnover. The level of substrate ion depletion was >70% after 1 min and >90% after 6 min incubation. Peptides s02, s04 and s09 showed substrate ion depletion from 35 to 45% after 1 min incubation, >70% depletion after 6 min, and >90% depletion after 12 min. Peptides s03, s05 and s07 demonstrated slow turnover that was below 50% after 12 min incubation. 53 and a reference uncomplexed structure (PDB 6yb7). 10 Each M pro subsite is colour-coded. In the protein region of the mass spectra, complexes of the M pro dimer and the cleavage products were observed after 1 min of incubation for the fast-turnover substrates s01, s06, s08, s10 and s11, and also s02, s04 and s09 (Fig. 7b1-b4). For the slowturnover substrates s03, s05 and s07, only M pro complexes with intact substrates were observed after 1 min incubation. For longer incubation times, complexes between M pro and the products from these substrates emerged and increased in abundance (Fig. 7b5 and b6). The rank order of the substrates in part depends on the MS method used, likely due to the differences in the buffers and concentrations used: i.e., non-denaturing MS used ammonium acetate buffer and an M pro concentration of 5 mM, which is higher than the 0.15 mM used in denaturing MS assays. Higher concentrations of both enzyme and substrate in the nondenaturing MS experiments explain the faster substrate turnover than seen with denaturing MS, especially as the concentration of catalytically active M pro dimer would be higher at higher enzyme concentrations. 6,54 Regardless of the MS method used, a clear trend is observed in the catalytic turnover of the cleavage site-derived peptides. The rank order of substrate preference under denaturing MS conditions was s01 > s11 > s06 > s02 > s10 > s07 > s09 > s05 (Fig. S2.22 and S2.23 †). Under non-denaturing conditions (Fig. 7) turnover was: fast (s01, s11, s06, s10, and s08), medium (s04, s02, and s09), and slow (s05, s03, and s07). Substrates s01, s11 and s06 turned over fastest; while s07, s05 and s03 were slow as measured by both methods. This is in broad agreement with the reversed phase high performance liquid chromatography analysis of substrate turnover by SARS-CoV M pro , where s01 and s02 display fast turnover; s10, s11 and s06 manifest medium turnover; and the rest (s09, s08, s04, s03, s05, s07) show slow turnover. 44 Both of our MS studies on SARS-CoV-2 M pro indicate that s02 consistently displayed slower turnover than s11. Previous reports on SARS-CoV M pro have shown evidence for cooperativity between subsites during substrate binding, in particular during autocleavage of the M pro C-terminal site (s02), where the Phe at P2 induces formation of the S3 0 subsite to accommodate the P3 0 Phe residue. 55 SARS-CoV-2 M pro substrate s02 has a Phe at P2, but not at P3 0 (Fig. 1a). The absence of a Phe at the P3 0 position may in part explain the reduced activity of SARS-CoV-2 M pro for s02 relative to s01, compared to the same pair in SARS-CoV M pro . 44 The observed turnover of all 11 SARS-CoV-2 cleavage-sitederived peptides by M pro is consistent with our atomistic models, where the peptides remain bound in the active site during MD simulations and where the scissile amide carbonyl remains well-positioned in the oxyanion hole (e.g., HB 8 in Fig. 3) for reaction initiation. The stability of the M pro -peptide interactions involving the S2 and S1 subsites, as well as backbone-backbone HBs 2, 3, 10 and 11, could explain the observation using non-denaturing MS of complexes of M pro with products-because of slow product dissociation. Nevertheless, we envisage that the order of substrate turnover rates is likely determined by various factors, including peptide conformations, the influence of the P2 and P1 0 residues on the catalytic dyad (as highlighted by the BigDFT analysis), entropic effects, and rates of product dissociation, all of which prompt ongoing experimental and computational investigations. ## In silico mutational analysis of substrate peptides enables peptide inhibitor design Building on insights gained from our binding studies of SARS-CoV-2 M pro and the 11 SARS-CoV-2 polypeptide substrate sequences, we designed peptides that could bind more tightly than the native substrates. We quantifed the per-residue energetic contributions of these sequences to the overall binding in the M pro active site and proposed substitutions that would increase affinity. We hypothesised these peptides would: (a) behave as competitive inhibitors, and (b) provide counterpoints for comparison with natural substrates, shedding light on requirements for M pro binding and, perhaps, turnover. ## In silico alanine scanning and predictive saturation variation scanning We used the interactive web application BAlaS to perform Computational Alanine-Scanning mutagenesis (CAS) using BudeAlaScan 56 and the BUDE_SM algorithm 57 for Predictive Saturation Variation Scanning (PreSaVS). 58 Both are built on the docking algorithm BUDE, 59 which uses a semi-empirical free energy force-feld to calculate binding energies. 60 To identify key binding interactions of the natural substrate peptides to M pro , the 11 substrate:M pro complexes were frst subjected to CAS using BAlaS. By sequentially substituting for alanine, the energetic contribution of each substrate residue to the overall interaction energy between the singly mutated peptide and M pro is calculated using: where DG wt is the interaction energy between the peptide and M pro , and DG Ala is the interaction energy for the peptide with a single alanine mutation at a given position. The more positive the value for each residue, the greater the contribution from that substrate residue to binding. This method was used later to evaluate potential inhibitor peptides. Having identifed residues contributing most to the binding energy of the natural M pro substrates, each of the sequences was subjected to PreSaVS using the BUDE_SM algorithm. This sequentially substitutes each substrate residue with a range of residues (D, E, F, H, I, K, L, M, N, Q, R, S, T, V, W and Y). BUDE_SM calculates the DDG ¼ DG wt DG mut for the binding interaction of each, entire, singly mutated peptide with M pro . Substitutions predicted to improve binding over wildtype sequences have a positive DDG. Fig. 8 shows an example of the BUDE_SM PreSaVS results for all the P2 substitutions for the 11 substrate peptides (normally Leu, Phe, or Val in the 11 substrates). The most positive results suggest that Phe, Trp and Tyr favour increased predicted affinity at the P2 position (Fig. 8b). However, although Tyr generally increased the predicted binding affinity (DDG sum ¼ 68.8 kJ mol 1 ), it was not considered for substitution at P2 due to its negative effect at this position in s11 (scoring 18.9, Fig. 8a). Candidate residues for each position, from P6 to P5 0 , were shortlisted similarly based on those with the best total, and the fewest unfavourable, scores. In addition to the computed DDG values, we considered the propensity of each residue to promote an extended conformation. All bound substrates are largely extended, so entropic penalties may be avoided if inherently extended conformations could be favoured in the designed peptide. Thus, the best bforming (and therefore least a-forming) residues from the frst triage were selected (Fig. 9a). 61 We also considered solubility. This was achieved by limiting the number of hydrophobic residues in each designed peptide and ensuring a net positive charge (except p14, which was neutral). ## Designed peptide sequences Employing the criteria described above, fve new peptides, p12-p16, were designed (Fig. 9b). Comparison of the computed DDG values for s01-s11 (Fig. 9c) and p12-p16 (Fig. 9d) reveals that substitutions at the P sites provide only occasional, moderate improvements to binding energy over the corresponding substrate P sites, with the notable exception of P2, which can accommodate Trp, Phe or Lys. These results agree with the HB analysis, which predicts that the sidechains of residues that are on the N-terminal side of the cleavage site (P sites) contribute more to binding than C-terminal, P 0 sites. The most striking difference between substrates and designed peptides is in this P 0 region, where the predicted binding energy contributions for the designed peptides exceed those of the substrates, an advantage that is distributed over most of the designed P 0 positions. The fnal step in design was to assess the relative binding affinities of the substrates and designed peptides. Hence the summed DDGs (Fig. 9e) provide a proxy for the binding energies (BAlaS) 62 for the substrates and designed peptides with M pro . The substrate:M pro complexes are stabilised by an average of 46.5 kJ mol 1 , whereas the designed-peptide:M pro complexes are predicted to have, in some cases, double the interaction stability of the substrates, with an average of 96.0 kJ mol 1 . The full analysis is in the ESI fle SI_BAlaS_BUDE_SM_12-04-2021.xlsx. † ## Synthesis and analysis of designed peptides To test the designed sequences, p12, p13, p15 and p16 were synthesised with a carboxyl-amide C-terminus by solid phase synthesis. Their M pro inhibitory activity was determined by dose-response analysis (Table 1) using SPE MS, monitoring both substrate s01 (1191.68 Da) depletion and N-terminally cleaved product (617 Da) formation. Ebselen which reacts multiple times with M pro63 was used as a standard (IC 50 ¼ 0.14 AE 0.04 mM; Fig. 10). All four designed peptides manifested similar potency with IC 50 values ranging from 3.11 mM to 5.36 mM (Table 1 and Fig. S3.1 †). Strikingly, despite the presence of Gln at P1 in all the designed peptides assayed, no evidence for hydrolysis was observed by SPE MS. This observation was supported by LCMS of the peptides incubated overnight with M pro (Fig. S3.2 †). We probed the inhibition mode of the designed peptides by monitoring changes in IC 50 while varying the substrate concentration (2 mM, 10 mM, 20 mM and 40 mM TSAVLQYSGFRK-NH 2 s01; K m $ 14.4 mM). 63 The results indicated a linear dependency between substrate concentration and IC 50 values (Fig. 10a-d). This was not observed with a control 15-mer peptide or ebselen (Fig. 10e and f). Analysis of the data by the procedure of Wei et al. 64 implies competitive inhibition (Fig. S3.3 and Tables S3.1, S3.2 †). By contrast, the same analysis for ebselen did not support competitive inhibition, consistent with MS studies showing it has a complex mode of inhibition. 63 Three of the synthesised peptides-p12, p13, and p15-have a Trp at P2 (Fig. 9b) while the other, p16, has a Lys at P2. The 11 M pro substrates all have hydrophobic residues (Leu, Val or Phe) at P2 (Fig. 1a). To investigate if the nature of the hydrophobic P2 residue, or the hydrophilic nature of the Gln at P1, alters the interaction of the peptide and hence its reactivity at the active site, we synthesised p13-WP2L, s01-LP2W, and s01-QP1W. There was no evidence for cleavage of p13-WP2L or s01-QP1W. However, s01-LP2W underwent partial cleavage (12.6 AE 4.5)% after overnight incubation. These results suggest that the presence of a Trp at P2 hinders catalytically productive binding, at least with these peptides, and that other residues (including the P1 0 and P2 0 residues) play roles in orienting the substrates for cleavage (vide infra). We then used non-denaturing protein MS to study enzymesubstrate/product/inhibitor complexes simultaneously with turnover. Complexes between M pro dimer and p12 and p13 were observed, together with the uncomplexed M pro dimer in the protein region of the mass spectra. No binding was observed for p15 and p16, due to relatively high noise in that m/z region. None of the designed peptides were cleaved by M pro , as recorded in the peptide region. As a control, s01 was S3.2 †). See Experimental Section S1.8 † for assay details. added to the protein/inhibitor mixtures; for all the inhibitors, turnover of s01 was observed after 3 min incubation. Depletion of s01 was 95%, 91%, 70% and 78% in the presence of p12, p13, p15 and p16, respectively, with an 8-fold excess of inhibitor over M pro , versus >98% depletion for the M pro /s01 mixture without the inhibitor. In the protein region of the mass spectra, complexes between M pro dimers and the s01cleavage products were observed in the presence of p13, but the abundance of these complexes was lower than the abundance of M pro /p13 complexes (Fig. 11). These results validate the above-described evidence that the peptide inhibitors both bind and competitively inhibit M pro . 3.4 Understanding the basis of SARS-CoV-2 M pro inhibition by the designed peptides 3.4.1 Modelling of the designed peptides. Modelling of p12 and p13 shows that both bind stably at the active site during MD simulation (Fig. S3.4-S3.8 †). Like the natural substrates, key HBs form with Glu-166, Thr-26, Thr-24, and the oxyanion holecontributing Cys-145 (Fig. S3.9-S3.11 †). However, HBs involving the P1-Gln sidechain of p12 and p13 showed greater variability. The favourability of the P2 Trp mutation, as predicted by the BAlaS scores, prompted us to investigate its binding. In line with the plasticity observed at S2, a variety of conformations are observed during MD simulations at this position, showing varying degrees of immersion in S2 (Fig. 12 and S3.12, S3.13 †). Similar results were obtained using iMD-VR (Fig. 12). Analysis of the conformations of the most populated cluster from MD using Arpeggio-generated hydrophilicity maps (Fig. S3.14 †) reveals that the P2 Trp is more deeply buried within S2 than the native P2 residues in the natural substrates, forming more than double the number of hydrophobic contacts in the cases of p12 and p13. Some conformations involve the indole ring p-p-stacking, or hydrogen-bonding via its indole N-H, with the catalytic His-41 sidechain, forming an extended HB network (Fig. 12). It is possible that these interactions may hamper the ability of His-41 to deprotonate Cys-145 at the start of peptide hydrolysis, which could be tested using QM/MM Fig. 13 QM contact interaction graph for p13 and M pro . Interactions are computed using ensemble-averaged results of MD snapshots with the BigDFT code. 50 calculations. Interestingly, DFT-based interaction analysis reveals that one of the slowest turnover substrates, s05 (Fig. 5), and inhibitor p13 (Fig. 13), share similar short-range interaction networks. 3.4.2 Comparative peptide docking. To investigate the ability of the M pro subsites to recognise residues in the designed sequences, AutoDock CrankPep (ADCP) was used (Table S3.3 †). 65 A trial was performed by redocking s01 into the H41A SARS-CoV M pro structure originally complexed with s01 (PDB entry 2q6g). 37 ADCP successfully positioned the peptide mostly correctly in its top solution, with the C a positions from P5 to P1 0 deviating by less than 1 (Table S3.4 and Fig. S3.15 †). Deviations increased up to 16 at P5 0 as the peptide coiled up in the P 0 positions, but this is deemed acceptable since the S 0 subsites are less well defned, as discussed earlier. Following the promising redocking results with ADCP, s01-s11, p12, p13, p15, and p16 were docked with an M pro structure originally complexed with the N3 inhibitor (PDB entry 7bqy; 1.7 resolution). 2 Substrate-docked structures were found to have the P4 and P2 residues correctly positioned in their corresponding S4 and S2 pockets (Table S3.5 †). From P1 to P5 0 the poses were more variable, with some peptide backbones turning through S1 rather than continuing an extended conformation, likely due to the less well-defned S 0 subsites (Fig. S3.16 †). For the designed peptides, by contrast, docking appeared less successful (except p16), with none of the top 10 solutions positioning the peptide in the manner observed in our MD simulations (Fig. S3.17 †). The S2 pocket in 7bqy binds the Leu sidechain of N3 and is probably too shallow to accommodate the larger Trp sidechain, given the assumption of a rigid receptor in ADCP docking. Hence, the four designed peptides were also docked to the C145A M pro structure in complex with the s02 cleaved product (PDB entry 7joy; 2 resolution), 66 which has a deeper S2 pocket that binds the P2 Phe sidechain in s02. Interestingly, for both p12 and p16, the top docked solution matched our design more closely (Table S3.6 and Fig. S3.18 †). Docking of p13 and p15 was challenging, possibly due to the difficulty of recognising a larger Leu (p13) or Ile (p15) residue in the S4 pocket, which originally accommodated a Val sidechain. This highlights the ability of the M pro active site to adapt when binding to different substrates or inhibitors. ## Summarydesigned peptides We used in silico Predictive Saturation Variation Scanning to design peptides that were shown in vitro to inhibit M pro competitively. Structures of p12 and p13 generated by both iMD-VR docking and comparative modelling behaved similarly, in terms of HB formation and peptide backbone RMSD and RMSF, when performing MD. These studies highlight how the S2 subsite can adapt its size and interaction network via induced ft to accommodate different substrate or inhibitor P2 residues. Notably, while these models suggested similarly stable binding modes as seen with the natural substrates, turnover of the inhibitor peptides by M pro was not detected. This may relate to the more favourable predicted binding affinity of the designed peptide-M pro complexes, both in terms of higher overall interaction energies, and greater contribution of the P 0 residues than in the natural substrates. Our MD simulations suggest it is also possible that the larger P2 residue prevents the catalytically vital His-41 from adopting a reactive conformation (Fig. 12). ## M pro -ligand interaction analysis Having elucidated how M pro recognises its substrates and our designed peptide inhibitors, we hypothesised that this might be reflected in the extensive small-molecule inhibitor work on M pro and could, in turn, be exploited for the design of novel smallmolecule inhibitors and peptidomimetics. We explored whether ligands sharing the same contacts as the natural substrates could lead to better inhibitory activity. We analysed all 91 X-ray structures of small molecule fragments complexed 3), and between the His-163 N3 and the heterocyclic nitrogen of the fragment (HB 6, Fig. 3). (e) Overlay of the P4-P1 0 -truncated structure of peptide inhibitor p13 (grey) from an MD snapshot and cluster 5 binder x0678 (pink), with the x0678 co-crystal structure (white surface). with M pro obtained by high-throughput crystallographic screening at Diamond's XChem facility, 35 as well as the dataset of 798 designed inhibitors and 245 crystal structures obtained from the COVID Moonshot project. 36 We analysed them by investigating their protein-ligand interaction patterns. ## Interaction analysis of XChem fragments We separated fragments into non-active-site binders (25 fragments) and active-site binding/likely-substrate competing molecules (66 fragments; Fig. S4.1 †). A fngerprint bit-vector was constructed for every active-site binding fragment, with each bit denoting the presence or absence of a given interaction with M pro residues, and used for clustering fragments by their interaction fngerprint Tanimoto similarity, 67 with 1 corresponding to identical contacts, and 0 to no shared contacts (Fig. 14, S4.2-S4.5 and Table S4.1 †). All the fragments and ligands in clusters 1 and 2 (except x0397, x0978 and x0981) are covalently bound to Cys-145. As a result, a highly conserved binding mode is observed for the carbonyl-containing covalent warheads (e.g., chloroacetamides), where the carbonyl oxygen binds into the oxyanion hole between residues Gly-143 and Cys-145, mimicking substrate HBs 8 and 9 (Fig. 3). Cluster 5 stands out as the only major cluster with fragments that bind deeply into S1, one of the main conserved contacts identifed in all substrates. Cluster 5 shows a distinct binding motif primarily driven by: (i) hydrogen bonding between a carbonyl oxygen on the fragment and the Glu-166 backbone NH-group; and (ii) a strong polar interaction between His-163 and the fragment. Notably, the protonation of the imidazole of His-163 appears to depend on the fragment. Overall, the primary functionality that facilitates interaction with His-163 is the nitrogen-containing heterocycle present in almost all ligands in cluster 5 (Fig. 15); the exception is x0967, which forms the His-163 HB via its phenol oxygen. Such heterocycles are well suited to replace the substrate P1 Gln sidechain by mimicking its HB donor/acceptor abilities. In addition, most cluster 5 binders also extend into the hydrophobic S2 pocket, although there is no clear preference in functional group at S2. This agrees with our plasticity analysis, which shows that S2 can accommodate a large variety of functional groups (Fig. S4.4 †). As seen in the overlap of peptide inhibitor p13 and cluster 5 representative x0678 (Fig. 14e), the binding modes of both inhibitors in the S1 and S2 subsites are very similar, with both forming HBs to His-163 (HB-6) and Glu-166 (HBs 2 & 3) and binding deep in the S2 pocket. In addition, all cluster 5 ligands (Fig. 15) contain an amide or urea linker between the P1 and P2 binding groups making them interesting building blocks for the development of peptidomimetics. The interactions between the fragments, substrates, and peptide inhibitors with M pro were analysed by employing linear scaling DFT. Using short-range (E cont ) DFT interactions with M pro as a "descriptor" for clustering, a cluster containing both the substrates and the cluster 5 compounds To test whether cluster 5 inhibitors are promising building blocks for optimization, we identifed all assayed and crystallized cluster 5 binders in the COVID Moonshot project database 36 as of 11th Jan 2021 and analysed them using Arpeggio. Compounds were deemed cluster 5 inhibitors if they shared at least 70% of the contacts identifed in fragment cluster 5. We observed that cluster 5 inhibitors have a signifcantly higher proportion of "strong" binders, classifed as IC 50 < 99 mM (85% of cluster 5 compounds), unlike the rest of the Moonshot project database (67%). Closer analysis can be found in Section S4.2 and Fig. S4.7. † In summary, based on Arpeggio and BigDFT contact analysis and reported assay data, cluster 5 binders are promising building blocks for substrate-competing inhibitor design. ## Covalent docking of COVID Moonshot compounds To accommodate induced ft and create high quality poses of covalent inhibitors for future optimisation, we selected 540 covalently-reacting compounds from 10 001 Moonshotdesigned compounds and docked them using AutoDock4 68 into the M pro structure of the corresponding covalent "inspiration" fragments. 36 We generated an interaction Tanimoto distance matrix as described earlier, and analysed the ability of the procedure to recapitulate the binding mode of the parent fragment. The normalized shape and pharmacophoric overlap (SuCOS 69 ) of the lowest energy pose of the highest populated cluster for each Moonshot compound was compared with the inspiration covalent XChem fragment (Fig. S4.8 †). When controlling for the smallest maximum common substructure (MCS) that encompasses the covalent warhead and one additional atom in the compound, 379 designs remain, 132 (34.8%) of which adopted the binding mode of the inspiration fragment. Given the high similarity between the fragments and docked designs, it is likely that these binding modes are more representative of the actual binding mode. A summary of the work-flow is shown in Fig. 16. In summary, our covalent docking method is more likely to identify the correct binding mode when substantial overlap exists between the inspiration fragment and designed compound beyond the covalent warhead (Section S4.3 and Fig. S4.8, S4.9 †). This generated 132 high quality docked poses which serve as inspiration for future inhibitor design and were used in our proposals for compound derivatisation in Section 4.3. All poses of the 540 docking runs are available at https:// github.com/gmm/SARS-CoV-2-Modelling. ## Implications for future inhibitor design We compared the interactions of the cluster 5 binders with those in the substrates, peptide inhibitors, and XChem fragments. Interestingly, unlike the peptides, almost none of the cluster 5 binders interact with the oxyanion hole. The only cluster 5 compounds where this contact is made are a series of covalent inhibitors, none of which showed promising potency (Fig. S4.11 †). An exhaustive search of Moonshot structures showed that at the time of the analysis, no non-covalent inhibitor has ever been tested that includes both the typical cluster 5 binding mode while also being able to interact with the oxyanion hole. We compared the structures of the top 10 compounds in cluster 5 (part of the dataset analysed in Section 4.1) to the docked structures of covalent Moonshot designs (Section 4.2). Two compounds-FOC-CAS-e3a94da8-1 and MIH-UNI-e573136b-3-were selected based on their high normalized SuCOS overlap with their inspiration fragments, strongly suggesting that their docked binding modes reflect the actual poses. 70 Both compounds bind into the oxyanion hole as well as into S1 and S2, providing a clear opportunity for extension of the cluster 5 binders (Fig. S4.12 †). Most cluster 5 binders place the aromatic heterocycle into the S1 site and the carbonyl oxygen of the amide linker bonds to Glu-166 (Fig. 17). The position of this amide nitrogen overlays perfectly with the ring amine present in the docked compound FOC-CAS-e3a94da8-1. Thus, extension of cluster 5 binders into the oxyanion hole could be achieved by adding a substituent at the amide nitrogen. A promising candidate for extension is x10789, which makes a HB with the backbone oxygen of Glu-166 (Fig. 17a) and mimics the non-prime side binding mode of peptide inhibitor p13 (Fig. 14e), even binding into the S4 site via its b-lactam ring (Fig. 17a). Additional expansion into the oxyanion hole and S1 0 through the amide linker could yield a powerful peptidomimetic inhibitor, combining protein interactions observed for the substrates, peptide inhibitors and small molecule fragments. When comparing interactions exhibited by cluster 5 binders (Glu-166, His-163) or covalent fragments (Gly-143, Cys-145) with the contacts present in the docked structure of the recently published Phase 1 clinical trial candidate PF-07321332 by Pfzer (Fig. 17b), 13,14 a nearly identical interaction pattern to the cluster 5 binding motif is observed. However, note that for reacted PF-07321332, AutoDock4 was unable to place the negatively charged azanide nitrogen in the oxyanion hole, which is the expected position given its similarity to related warheads previously docked (Fig. S4.9 †). ## Conclusions A wealth of crystal structures of SARS-CoV-2 M pro is available, including hundreds with ligands. There is thus the question of how best to use this static information to help develop M pro inhibitors optimised in terms of efficacy and safety for COVID-19 treatment. The dimeric nature of M pro , coupled with its multiple substrates, makes it challenging to understand the structural and dynamic features underpinning selectivity and catalysis, as is the case with many other proteases. Such an understanding is, of course, not essential to develop medicines, as shown by work with other viral proteases. However, it may help improve the quality of such medicines and the efficiency with which they are developed. It will also lay the foundation for tackling anti-COVID-19 drug resistance-a challenge we will likely encounter as experience with the HIV global pandemic implies. The scale of global efforts on M pro makes this system an excellent model for collaborative efforts linking experimental biophysics, modelling, and drug development (Fig. 18). The results of our combined computational studies, employing classical molecular mechanics and quantum mechanical techniques, ranging from automated docking and MD simulations to linear-scaling DFT, QM/MM, and iMD-VR, provide consistent insights into key binding and mechanistic questions. One such question concerns the protonation state of the 'catalytic' His-41/Cys-145 dyad, an important consideration in the rates of reaction of covalently linked M pro inhibitors which ultimately relates to their selectivity and potency. Our results indicate that a neutral catalytic dyad is thermodynamically preferred in M pro complexed with an unreacted substrate, justifying the neutral state for MD simulations. A more reactive thiolate anion may be deleterious to the virus, as it will be susceptible to reaction with electrophiles. Importantly, analysis of the active site suggests that the precise mechanism of proton transfer in the His-41/Cys-145 dyad involves dynamic interactions with other residues, including His-163, His-164, Asp-187, and a water hydrogen bonded to the latter two residues and His-41. Proton transfer may be considered a relatively simple part of the overall catalytic cycle-these results thus highlight how M pro catalysis is likely a property of (at least) the entire active site region, with a future challenge being to understand motions during substrate binding, covalent reaction, and product release. The models we have developed of M pro in complex with its 11 natural substrates provided a basis for analysis of key interactions involved in substrate recognition and for comparison with (potential) inhibitor binding modes. Notably, the P 0 (Cterminal) side of substrates appears to be much less tightly bound than the P (N-terminal) side, where there is remarkable consistency in the hydrogen bonding patterns across the substrates. This difference may in part reflect the need for the P 0 side to leave (at least from the immediate active site region) after acyl-enzyme complex formation and prior to acyl-enzyme hydrolysis. The tighter binding of the N-terminal P-side residues suggests these are likely more important in substrate recognition by M pro . This is also reflected in potent inhibitors, such as N3 and peptidomimetic ketoamides, 2,4 which predominantly bind in these non-prime S subsites. The development of S-site-binding inhibitors may also reflect the nature of the substrates used in screens leading to them, which typically comprise an S-site binding peptide with a C-terminal group enabling fluorescence-based measurement. Our results imply that there is considerable scope for developing inhibitors exploiting the S 0 subsites, or both S and S 0 subsites, though relatively more effort may be required to obtain tight binders compared to targeting the S subsites. Consistent with prior studies, our work highlights the critical role of the completely conserved P1 Gln residue in productive substrate binding and analogously in inhibitor binding. However, the nature of the P2/S2 interaction is also important in catalysis. In the natural substrates (Fig. 1), the P2 position is Leu in 9 of the 11 substrates, Phe in s02 (which displays medium turnover efficiency), and Val in s03 (which is a poor substrate). Our results show that the S2 subsite plays a critical role in recognition and inhibition. S2 is highly plastic (Fig. 6 and S4.4 †) and can accommodate a range of different sidechains, including larger groups, though not necessarily in a productive manner. The observation that substrates with a P2 Leu vary in efficiency reveals that interactions beyond those involving P1 and P2 are important, reinforcing the notion that (likely dynamic) interactions beyond the immediate active site are important in determining selectivity both in terms of binding and rates of reaction of enzyme-substrate complexes. Notably, the results of computational alanine scanning mutagenesis followed by design, aimed at identifying peptides that would bind more tightly than the natural substrates, led to the fnding that substitution of a Trp at P2 ablates hydrolysis creating an inhibitor. The observations with peptide inhibitors of M pro have precedence in studies with other nucleophilic proteases, including the serine protease elastase, showing that substrate substitutions away from the scissile P1/P 0 residues can cause inhibition. 72,73 There is thus scope for the extensive development of tight binding peptidic and peptidomimetic M pro inhibitors for use in inhibition and mechanistic/ biophysical studies, with the Trp at P2 of the peptide inhibitors being a good point for SAR exploration, potentially by (i) replacement of the indole hydrogen with suitable alkyl or aryl substituents; (ii) introduction of substituents with different stereoelectronic properties at C-2 or C-5 of the indole ring; or (iii) cyclization by the insertion of a methylene group linking position 2 of the indole ring to the a-nitrogen of Trp itself. 74 Finally, the combined analysis of interactions involved in substrate binding and extensive structural information on inhibitor/fragment binding to M pro enabled us to identify a cluster of inhibitors whose interactions relate to those conserved in substrate binding (e.g., involving the Glu-166 backbone, His-163 sidechain, and/or the oxyanion hole formed by the Cys-145 and Gly-143 backbones). Building out from these 'privileged' interactions (Fig. 17) might be a useful path for inhibitor discovery. Indeed, an M pro inhibitor now in clinical trials 13,14 exploits the same 'privileged' interactions that we identifed. We hope the methods and results that have emerged from our collaborative efforts will help accelerate the development of drugs for treatment of viral infections, and particularly COVID-19. ## Methods A detailed description of the experimental and computational methods employed in this work is provided in the ESI. †

chemsum

{"title": "Discovery of SARS-CoV-2 M^pro peptide inhibitors from modelling substrate and ligand binding", "journal": "Royal Society of Chemistry (RSC)"}

europium_luminescence:_electronic_densities_and_superdelocalizabilities_for_a_unique_adjustment_of_t

## Abstract: We advance the concept that the charge factors of the simple overlap model and the polarizabilities of Judd-Ofelt theory for the luminescence of europium complexes can be effectively and uniquely modeled by perturbation theory on the semiempirical electronic wave function of the complex. With only three adjustable constants, we introduce expressions that relate: (i) the charge factors to electronic densities, and (ii) the polarizabilities to superdelocalizabilities that we derived specifically for this purpose. The three constants are then adjusted iteratively until the calculated intensity parameters, corresponding to the 5 D 0 → 7 F 2 and 5 D 0 → 7 F 4 transitions, converge to the experimentally determined ones. This adjustment yields a single unique set of only three constants per complex and semiempirical model used. From these constants, we then define a binary outcome acceptance attribute for the adjustment, and show that when the adjustment is acceptable, the predicted geometry is, in average, closer to the experimental one. An important consequence is that the terms of the intensity parameters related to dynamic coupling and electric dipole mechanisms will be unique. Hence, the important energy transfer rates will also be unique, leading to a single predicted intensity parameter for the 5 D 0 → 7 F 6 transition.The theoretical foundation for the first lanthanide luminescence models began to burgeon in the late ′ 20 s. Through the Point Charge Electrostatic Model, PCEM, Bethe estimated the magnitude of the crystalline electric field on the energy levels of the 4d and 4f orbitals 1 . In 1937, Van Vleck assigned the narrow spectral lines observed for the lanthanide ions to 4f transitions. Further, in this same article, Van Vleck addressed the nature of these electronic transitions and classified them as governed by electric dipole, magnetic dipole and electric quadrupole mechanisms 2 . Furthermore, eight years later, by semi-quantitative calculations, Broer and coauthors demonstrated that the electric dipole mechanism was sufficient to explain the observed experimental intensities 3 .These were the works that inspired and gave support to Judd-Ofelt theory 4,5 . Indeed, in 1962, Judd 4 and Ofelt 5 published, in an independent manner, their studies on the transitions between the electronic energy levels in the 4f sub-shell of lanthanide ions. In their articles, they both formulated essentially the same theory that quantitatively explains the radiative optical transitions in the lanthanide ions, in which they used Racah algebra to arrive at expressions for the oscillator strengths related to the forced electric dipole terms within 4f n configurations 4,5 . The calculation of intensity parameters through the Judd-Ofelt theory depends on the contributions of two important terms that represent two mechanisms, each: (i) the forced electric dipole mechanism, and (ii) the dynamic coupling mechanism. Calculation of the forced electric dipole mechanism depends on the odd parity crystal field parameters. Until the early 80 s, these parameters were obtained from PCEM. However, at that time, reports in the literature suggested that the even-parity terms obtained by PCEM do not correspond to the observed crystal field splitting in lanthanide ions . To overcome part of these discrepancies, in 1982, Malta introduced the Simple Overlap Model-SOM 9 . The SOM model assumes two postulates: (i) the 4f energy potential is generated by charges, uniformly distributed in a small region located around the midpoints of the lanthanide-ligand chemical bonds, and (ii) the total charge in each region is equal to -geρ, where g is a charge factor, e is the fundamental electric charge, and ρ is a parameter proportional to the magnitude of the total overlap between the lanthanide ion and the ligand atoms. While PCEM only treats the metal-ligand atom bonds as a purely electrostatic phenomenon, SOM introduces a correction to the crystal field parameters of PCEM in order to confer to it a degree of covalency through a charge factor g. However, the SOM article 9 did not provide equations for its calculation. The calculation of intensity parameters through the Judd-Ofelt theory further depends on equations describing the dynamic coupling mechanism, which in turn depends on structural aspects (coordination geometry), and is thus sensitive to the chemical environment around the lanthanide ion through polarizabilities, α i , of the i directly coordinated atoms of the ligands 10 . So far, there are no expressions that allow the calculation of these polarizabilities. More recently, Malta and co-workers introduced the concept of the overlap polarizability of a chemical bond and proposed an ordinal scale of covalence for lanthanide complexes 11 . They also proposed an equation for calculating this new overlap polarizability 11 . However, this overlap polarizability is only a part of the polarizability itself, i.e. this equation does not fully quantify the polarizabilities α i . Hence, to this day, the charge factors of the SOM model and the polarizabilities of the Judd-Ofelt theory do not have any mathematical expressions to allow them to be evaluated. In 1994, we introduced the Sparkle Model to carry out semiempirical molecular orbital calculation of lanthanide complexes at the AM1 level 12,13 , making it also possible to calculate UV-Vis absorption spectra from the Sparkle Model geometry via INDO/S 14 . The model was improved in 2004 15 with the addition of Gaussian functions to the core-core repulsion and proved useful for the design of luminescent complexes 16,17 . Subsequently, robust statistical methodologies were incorporated into the parameterization procedure, and the model has been since parameterized for a variety of existing and widely distributed semiempirical models, such as AM1 18 , PM3 19,20 , PM6 21 , PM7 22 , and RM1 23 . We designed the Sparkle Models to predict mainly geometries-the most computing time intensive part of lanthanide complex computational chemistry. Indeed, once one has a fully optimized geometry, more advanced single point calculations on the complexes can then be carried out with more workability. The variety of Sparkle Model implementations is important because ligands in the complexes vary, and different semiempirical models tackle particular bonding situations differently: some more accurately than others. Therefore, having a palette of Sparkle Models to choose from, adds a strong value to the experimentalist. All are fully available in the MOPAC software 29 . Recently, the RM1 model for lanthanides has been introduced 30 . In this model, the europium atom is now represented in the semiempirical calculation as an atom with a core depicting [Xe4f 6 ]; while assigning to its semiempirical valence shell, three electrons and the following set of semiempirical atomic orbitals: 5d 6s 6p. The RM1 model for lanthanides so defined, does extend the accuracy of the previous Sparkle Models to types of coordinating bonds other than Eu-O and Eu-N; the most common ones for Eu being Eu-C, Eu-S, Eu-Cl, and Eu-Br. Both the Sparkle Model and the RM1 model for the lanthanides are quantum chemical models, which generate electronic wave functions, and therefore yield a wealth of information. However, it is noteworthy that, up to now, the electronic wave functions of these models have not been directly used in the context of lanthanide luminescence. Indeed, the last 20 years were fraught with publications about the development and application of theoretical methods to study the luminescent properties of lanthanide compounds, especially for systems containing europium ions. However, even now, less than 3% of published studies involving lanthanide ions make use of theoretical tools 31 . In 2013, in order to better disseminate these theoretical tools, our group published an article showing, systematically, the theoretical study of a simple system of europium 32 , and then released our new luminescence software package, LUMPAC 31 . This is the first and only software specifically designed for the study of luminescence properties of systems containing lanthanide ions. This first version, which is available via our homepage (www.lumpac.pro.br), was designed to be efficient and user-friendly. In the short time that it has been available, it is already in use by many experimental groups worldwide. So far, in the first version of LUMPAC 31 and other articles , the charge factors g i and polarizabilities α i are frequently adjusted in order to reproduce the experimental intensity parameters Ω 2 exp and Ω 4 exp . During the adjustment procedure, the calculated intensity parameters (Ω λ calc ) from the optimized geometry, obtained from one of our Sparkle Models, are compared with the experimental intensity parameters (Ω λ exp ). In this article, we advance the concept that the charges, g i, and polarizabilities, α i , for europium complexes, within SOM and Judd-Ofelt theory, can be effective and uniquely modeled by energy variations resulting from perturbations on the semiempirical electronic wave function of the complex. In our conceptualization, the charges will be determined from first order perturbation theory, and the polarizabilities from second order perturbation theory. ## Results Uniqueness of g i and α i . First, we carried out a series of tests to determine the uniqueness of the adjusted set of parameters g i and α i for some representative complexes. We found out numerically that the number of degrees of freedom is actually smaller than the theoretical maximum of 2N c , where N c is the coordination number of the complex, due to restrictions that result from the need to accommodate the geometry and the values of Ω 2 exp and Ω 4 exp in g i and α i . Nevertheless, the residual number of degrees of freedom is still quite large. Indeed, there is an enormous space of solutions for this problem, with drastically different values of g i and α i leading exactly to the same experimental values of both Ω 2 exp and Ω 4 exp . That was an unsettling finding because the different g i and α i imply in different predicted Ω calc 6 values, on which the radiative emission rate depends. Further, the contribution to the intensity parameters from coupling dynamics (Ω λ dc ), which depends on α i , and from electric dipole (Ω λ ed ), which depend on g i , will also vary and be non-unique for any geometry and any two given values of Ω 2 exp and Ω 4 exp . Furthermore, Ω λ ed , is used to predict the energy transfer rates via the multipolar mechanism, which will, in turn, be non-unique, depending on the g i and α i values chosen from the space of solutions to the problem of finding a set of g i and α i consistent with the two values Ω 2 exp and Ω 4 exp . So, in this article, we introduce a way of determining the sets of g i and α i in a unique manner for any given complex, from which the geometry and the values of Ω 2 exp and Ω 4 exp are known. Determining g i and α i uniquely from semiempirical calculations. In order to model the effect by the metal ion on the directly coordinated atoms of the ligands, we use first and second order perturbations on the semiempirical wavefunction 40 , as fully described in the Supplementary Information which accompanies this article. Accordingly, in this article, we postulate that the charge factors g i of the SOM model 41 are equal to the following expression obtained from first order perturbation theory: where Q will be a single parameter to be applied to all zero differential overlap, ZDO, electronic densities, q i , of all directly coordinated atoms i. The expression for the ZDO electronic density at any atom μ of the complex, q μ is ∑∑ where i' runs over all occupied molecular orbitals of the complex, p runs over all atomic orbitals of atom μ , and µ ′ c pi is the corresponding linear coefficient. Likewise, we further postulate that the polarizabilities α i of Judd-Ofelt theory 4,5 are given by: obtained from second order perturbation theory, with constants D and C being the same for all directly coordinated atoms i of a given complex, and SE σ is the electrophilic superdelocalizability of any atom σ of the complex, originally introduced by Simas 40 , where i' runs over all occupied molecular orbitals of the complex, p and q run over all atomic orbitals of atom μ , and σ ′ c pi and σ ′ c qi are the corresponding linear coefficients. Our electrophilic superdelocalizabilty is therefore a generalization to an all valence electron method of the corresponding superdelocalizability of Fukui 42 . In this sense, our electrophilic superdelocalizability 40 is unique and differs from the one in the article by Lewis 43 and also from the one in the article by Brown and Simas 44 because these do not take into account the cross-products of the atomic orbitals for each molecular orbital. And it also differs even more from the delocalizability of Schüürmann 45,46 , D E (i), because not only, as Lewis 43 and as Simas and Brown 44 , he does not take into account the cross-products of the atomic orbitals for each molecular orbital, but also because, instead, he uses a different denominator (see Eq. S56 of the Supplementary Information). As we carried out research for this article, we also tried to use the superdelocalizability of Lewis 43 and of Simas and Brown 44 , and also the delocalizability of Schüürmann in Eq. ( 58). However, they all did not produce good fittings. Therefore, we stayed with our superdelocalizability as defined by Eq. ( 4) above. Complete derivations of first order and second order perturbations on the semiempirical wavefunction of the lanthanide complex, leading to the ZDO electronic densities, to the electrophilic superdelocalizabilities Eq. ( 4), and to Eqs. ( 1) and ( 3), are fully presented in the Supplementary Information. Parameters Q, D, and C are then adjusted for each complex in order to reproduce the various experimentally obtained Ω λ exp with λ = 2, 4. And in the process of finding the optimal Q, D, and C parameters, we use the nonlinear optimization technique generalized simulated annealing 47 . ## LUMPAC implementation. Both the charge factors from SOM 41 and the polarizabilities from Judd-Ofelt theory 4,5 introduced by Jørgensen and Judd 10 , to be used in Eqs (S10,S14), must be positive. Since the charge factors, as advanced in the present article, are being calculated from Eq. ( 1) as a product between the always-positive ZDO electronic densities at the directly coordinated atoms, and a multiplicative constant Q, then this constant Q must always be positive as well. Likewise, as the electrophilic superdelocalizability is always a negative quantity, and the polarizability must be a positive quantity. Then, from Eq. ( 3), SE i •D+ C > 0 and therefore, the following inequality must be always obeyed: After a large number of simulations, we noted that the D and C parameters optimized to reproduce the experimental intensity parameters were almost always positive values. Therefore, we restricted the acceptable values of D and C to positive ones. Further justification of that can be arrived at, by comparing the homomorphic equations Eq. ( 3) and Eq. (S54) of the Supplementary Information, when it becomes clear that constant D can be interpreted as being (δβ στ ) 2 of Eq. (S54), necessarily a positive value. Since SE i is negative because of the occupied orbital energies in its denominator, then the product SE i •D is a negative quantity. As such, the polarizability can only be turned positive by a positive C. Moreover, we also noted that, very frequently, the optimized values of D and C obeyed approximately the following rule: D ≈ 2C, the vast majority with values of D lying in the interval 1 < D < 2.5C. For the cases where this ratio fell outside this range, the D/C ratio found by the non-linear optimization techniques tended to be invariably too small, close to zero. In these cases, the adjustment procedure has not usually been successful. Indeed, when D ≈ 0 ≪ C, then the effect of the superdelocalizabilities on the polarizabilities is being zeroed and the polarizabilities of all atoms become essentially similar and approximately equal to C. In these cases, we regard such fittings to be devoid of physical meaning and discard them. So, there must be some truth in the fact that acceptable fittings always display a D/C ratio ≈ 2 au −1 , which adds to the strength of the methodology we are introducing in this article. We avail ourselves of this fact and define here a binary outcome acceptance attribute for the adjustment, that is: we consider the adjustment acceptable whenever D/C > 1 au −1 , and unacceptable whenever D/C ≤ 1 au −1 . Accordingly, as starting guesses for the parameters in the non-linear optimization of Eq. ( 5), we then choose values subjected to the conditions Q > 0, D > 0, C> − SE i •D, and D ≈ 2C. Ω 6 exp is rarely observed from emission spectra because it is always displaced towards longer wavelengths and it is also very weak. Thus, in the process of obtaining the fit, we chose to minimize the quadratic errors in Ω 2 exp and Ω 4 exp , while simultaneously minimizing Ω 6 exp according to Eq. ( 5), below. ## Discussion We decided to test the methodology advanced in this article on all europium complexes whose crystallographic structures could be obtained from the Cambridge Crystallographic Database , and whose values of Ω 2 exp and Ω 4 exp have been published. Thirteen very different complexes obeying this criterion were found, and are listed in Table 1. To exemplify how the new methodology functions, consider the crystallographic structure of the complex of CSD code GIPCAK, Eu(BTFA) 3 (4,4-BPY)(EtOH), shown in Fig. 1, where BTFA stands for 4,4,4-trifluoro-1-phenyl-2,4-butanedione, and 4,4-BPY for 4,4′ -bipyridine. We then carried out a single point Sparkle/RM1 calculation in order to obtain the ZDO electronic densities and electrophilic superdelocalizabilities at the directly coordinating atoms of the ligands to be used in the fitting procedure. We also carried out single point RM1 model for Eu(III) calculations to obtain the ZDO electronic densities and electrophilic superdelocalizabilities, so that we can now compare the electronic properties results, at the crystallographic geometry, from a Sparkle Model with those from the RM1 model for Eu(III), which has valence orbitals at the europium ion center. Results are presented in Table 2, which shows the values for Q, D, and C for both Sparkle/RM1 and RM1 model for europium calculations, together with values, at the directly coordinated atoms of the ligands, of ZDO electron densities, electrophilic superdelocalizabilities, and the corresponding charge factors g and polarizability α values. Note that the ratio D/C is 2.13 au −1 for the Sparkle/RM1 case and 2.43 au −1 for the RM1 for Eu(III) case. Observe that the present fitting naturally groups the polarizabilities of the directly coordinated atoms in same ligand groups. That is, the oxygen atoms from one of the BTFAs, BTFA3, have similar Sparkle/ RM1 polarizabilities of 6.78 3 and 6.87 3 ; the oxygen atoms from BTFA2 also have similar polarizabilities of 1.30 3 and 1.23 3 ; and for polarizabilities of the oxygen atoms of BTFA1 the values are 0.183 3 and 0.0517 3 . Further, the polarizability of the oxygen from the coordinated ethanol is 4.59 3 , and of the nitrogen atom of 4,4-BPY the value is 1.89 3 , both being intermediary values. This grouping of polarizabilities is in line with what had been the practice until now, and implemented in version 1.0 of LUMPAC 31 . Note that this same grouping also naturally occurs in the RM1 model for Eu(III) for the electronic properties of this same complex (see Table 2). While before, and also in LUMPAC, that had to be done by hand, here the groupings naturally emerge from the quantum chemical calculations. Table 3 presents the Q, D, and C parameters for similar fittings for all 13 complexes considered, with the electronic densities and superdelocalizabilities computed by single point (using the 1SCF keyword of MOPAC) Sparkle/AM1, whereas Table 4 shows corresponding results computed by single point RM1 model for Eu(III). Noticeably, a general trend is followed in both Tables 3 and 4 for the quantities Q, D, and C, with all fittings having resulted being acceptably good, except for complex QAMLIB where the errors in both Ω calc 2 and Ω calc 4 are larger. The same happens when we examine the fittings for the RM1 model for Eu(III), also at the crystallographic geometries. For both models, the error in Ω calc 2 for complex QAMLEX is mildly acceptable, but the error in Ω calc 4 is not. However, the impact of Ω calc 4 in A rad is much smaller, which makes this situation slightly less ## Sparkle/RM1 RM1 model for Eu(III) Ligand Atom , or from an intrinsic inadequacy of the whole model. That is open to investigation. Nevertheless, for all other complexes the obtained fittings were very good. We then decided to verify what would happen if the crystallographic geometries were not available. To simulate this situation, we then optimized the geometry of the complexes by both RM1 model for Eu(III) and Sparkle/RM1, and carried out the fittings. Results are presented in Tables 5 and 6. In Table 5, five of the geometries clearly seemed to have been incorrectly predicted and the model could not properly carry out the fitting. For all these five cases (complexes of CSD codes GIPCAK, QAMLIB, VENLEH, VENLIL, and YETTUN) the ratio D/C was found to be close to zero after the fitting attempt. One other borderline case, where the ratio D/C was found to be 1.00 au . a Cambridge Crystallographic Data Centre deposited CSD entry. function as a compass in pointing to the acceptability of the fit. We are simply presenting these fits in Table 5 to illustrate the cases when we needed to reject the fits as devoid of physical meaning. The cells in the lines corresponding to these fits have been painted gray, so as not to be confused with the other acceptable ones. The fact that the fits are acceptable when we use crystallographic geometries, and sometimes are not when we use Sparkle Model geometries, suggests that the adjustments seem to fail when the predicted geometries are not sufficiently accurate. This indicates that the choice of the semiempirical model to carry out the geometry optimization is a crucial step in this process. Since the RM1 model for Eu(III) is a more accurate model in terms of obtaining geometries, we expect that the fits will be more successful in this case. And that is corroborated by the results in Table 6, which shows fits from both geometric and electronic properties from RM1 model for Eu(III) calculations. This time, only one fit needed to be rejected due to the fact that the ratio D/C was close to zero: the fit for complex DEVHOC. Accordingly, results presented in Tables 5 and 6 do reinforce the fact that excellent geometries are an important requirement for the fitting to be successful. The robustness of the fitting can be strengthened by the relative similarity and stability of parameters Q, D, and C across all tables. Indeed, for example, for complex RATKUU, the values of Q in Tables 3 to 6 are: 0.0546 au −1 , 0.0550 au −1 , 0.0275 au −1 , and 0.193 au −1 . The corresponding values of D are: 60.2 au −1 . 3 , 46.7 au −1 . 3 , 47.6 au −1 . 3 , and 47.5 au −1 . 3 . And the corresponding values for C are: 27.6 3 , 20.0 3 , 24.4 3 , and 22.1 3 . In all cases studied, with either crystallographic or theoretically optimized geometries, only a single minimum could be found in the fit given the constraints imposed on the problem: Q > 0, D > 0, C > − SE i •D, and D ≈ 2C, a result consistent with the uniqueness of the fits being introduced here. Such uniqueness makes possible eventual future interpretations of the meanings of the quantities Q, D, and C. In the Supplementary Information, we present tables with results for both single point calculations at the experimental geometries, as well as for fully optimized geometries, for all the other Sparkle Models: Sparkle/AM1, Sparkle/PM3, Sparkle/PM6, and Sparkle/PM7. We now have enough data to test the hypothesis that a geometry, closer to the crystallographic one, will tend to produce more acceptable fittings as measured by the binary outcome acceptance attribute for the adjustment represented by D/C, where D/C > 1, for acceptable fittings and D/C ≤ 1 for unacceptable ones. For all complexes, we measured the difference between the theoretically predicted coordination polyhedron (Eu(III) included) and the crystallographic one, by means of their minimized root-mean-square deviation, RMSD, corrected for the number of atoms. The minimization was performed on the polyhedra by translation and rotation, via the Kabsch algorithm 51 employing a freely available Python script (http:// github.com/charnley/rmsd). Table 7 shows all RMSD values for all complexes considered, computed between the crystallographic coordination polyhedra and the theoretically predicted ones, for all semiempirical models taken into consideration. When the theoretical intensity parameter adjustment resulted unacceptable, we painted the respective cell gray. As a result, we have 27 gray cells and 51 other ones. The mean RMSD of the gray cells is 0.525 and the mean RMSD of the 51 other cells is 0.367 . This indicates that, indeed, when the error in the geometry of the theoretical coordination polyhedron is larger, the theoretical intensity parameter adjustment tends to result unacceptable. We now turn to quantify the statistical significance of this statement by verifying whether the mean RMSD of the gray cells, 0.525 , is truly larger than the mean RMSD of the other cells, 0.367 . The t-statistic for both sets of data is 3.089, which, for 26 degrees of freedom, does indicate that the mean for the gray cells is indeed larger than the mean for the regular cells within a 99.8% confidence level. This result does reinforce the fact that the choice of the semiempirical model to carry out the geometry optimization is truly a crucial step in this process, because one needs an accurate geometry for an acceptable adjustment of the theoretical intensity parameters. Moreover, if one does not possess the crystallographic geometry and computes the geometries of the complex by two theoretical methods, and one of them leads to an acceptable adjustment, and the other does not, it is likely that the geometry of the one, which yielded the acceptable adjustment, will be closer to the crystallographic geometry. ## Conclusions In this article, we advanced a procedure for fitting, in a unique manner, the theoretical intensity parameters Ω λ calc to reproduce the experimentally obtained Ω λ exp from either crystallographic geometries, or from geometries obtained from Sparkle Model or RM1 calculations on complexes. Thus, we now have a procedure, which is seemingly a robust one, and that leads to a unique set of g and α necessary for the prediction of the intensity parameters. The relative stability of the Q, D, and C parameters for the same complex when the semiempirical method employed is varied, as can be seen from Tables 3 to 6, as well as from the tables in the Supplementary Material, further strengthens the uniqueness of the adjustment being advanced in this article. In addition, in the absence of crystallographic geometries, these can be obtained from either one of the Sparkle Models , or from the more accurate RM1 model for lanthanides 30,52 . The model contains a built in quality control index, the ratio D/C, which suggests that, in general, something seems not to be correct with the geometry vis a vis the intensity parameters when the value of this ratio is close to zero. Besides, for the adjustment to occur in a perfect manner, a requirement of the whole luminescence model seems to be that the geometry and intensity parameter values must be consistent with each other. In the absence of crystallographic geometries, one can try to optimize the complex with different Sparkle Models or with the RM1 model for lanthanides until a good fit is obtained. That is because each Sparkle Model underlying semiempirical method, AM1, PM3, PM6, PM7, or RM1, treats every type of ligand differently, and one method may be better for some ligand characteristics than others. As an additional evidence of consistency of our model, we showed that semiempirical methods that lead to an acceptable theoretical intensity parameter adjustment, also tend to produce more accurate geometries, likely closer to the true crystallographic one. The uniqueness of the adjustment has a number of very good consequences for luminescence research, since, as mentioned before, a unique set of parameters Q, D, and C will lead to a single predicted Ω calc 6 value. Further, the contribution to the intensity parameters from dynamic coupling, (Ω λ dc ), which depends on α i , and from electric dipole (Ω λ ed ), which depends on g i , will also be unique for any given geometry and any two given values of Ω 2 exp and Ω 4 exp . Furthermore, Ω λ ed , which is used to predict the energy transfer rates via the multipolar mechanism, will be also unique. ## Methods All Sparkle calculations were carried out using MOPAC2012 29 , and all RM1 model for europium calculations were carried out by a modified version of the same software. Calculations were done either at the crystallographic geometry, or by fully optimizing the geometry at the particular level of theory, when great care was taken to ensure that no imaginary vibrational frequencies were present. A modified version of LUMPAC was then coded to implement the new methodology being advanced in this article, and will be made available as a new version at http://www.lumpac.pro.br. This modified version was used to obtain the results presented in Tables 2 to 7 and Tables S1 to S16 of the Supplementary Information. This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

chemsum

{"title": "Europium Luminescence: Electronic Densities and Superdelocalizabilities for a Unique Adjustment of Theoretical Intensity Parameters", "journal": "Scientific Reports - Nature"}

pyro-synthesis_of_a_high_rate_nano-li3v2(po4)3/c_cathode_with_mixed_morphology_for_advanced_li-ion_b

## Abstract: A monoclinic Li 3 V 2 (PO 4 ) 3 /C (LVP/C) cathode for lithium battery applications was synthesized by a polyol-assisted pyro-synthesis. The polyol in the present synthesis acts not only as a solvent, reducing agent and a carbon source but also as a low-cost fuel that facilitates a combustion process combined with the release of ultrahigh exothermic energy useful for nucleation process. Subsequent annealing of the amorphous particles at 8006C for 5 h is sufficient to produce highly crystalline LVP/C nanoparticles. A combined analysis of X-ray diffraction (XRD) and neutron powder diffraction (NPD) patterns was used to determine the unit cell parameters of the prepared LVP/C. Electron microscopic studies revealed rod-type particles of length ranging from nanometer to micrometers dispersed among spherical particles with average particle-sizes in the range of 20-30 nm. When tested for Li-insertion properties in the potential windows of 3-4.3 and 3-4.8 V, the LVP/C cathode demonstrated initial discharge capacities of 131 and 196 mAh/g (,100% theoretical capacities) at 0.15 and 0.1 C current densities respectively with impressive capacity retentions for 50 cycles. Interestingly, the LVP/C cathode delivered average specific capacities of 125 and 90 mAh/g at current densities of 9.6 C and 15 C respectively within the lower potential window. ## R echargeable lithium ion batteries (LIBs) are considered as one of the most promising power sources for electric vehicle (EV) and hybrid electric vehicle (HEV) applications 1 . The present state of affairs in the emerging market of Li-ion batteries indicate that any Li-battery cathode requires to satisfy the demands of fast charging capability, high energy density, long shelf-life and safety issues. Bearing this in mind, intense efforts are on to identify phosphate-based cathodes due to the strongly covalent (PO 4 ) 32 units that provide greater structural stability than commercial LiCoO 2 hosts even under deep charging conditions and elevated temperatures . Among the known phosphates, monoclinic Li 3 V 2 (PO 4 ) 3 (LVP) is a promising LIB cathode due to its high operating voltage (,4 V) and ability to deliver a maximum theoretical capacity of 197 mAh g 21 as 3Li 1 ions are extracted/inserted per unit in the voltage range 3.0-4.8 V. However, monoclinic LVP suffers from low electrical conductivity (,2.4 3 10 28 S cm 21 ) and hence the complete utilization of its theoretical capacity coupled with high rate performances becomes intricate. A majority of the strategies to overcome this obstacle is focused on producing particles with electrically conductive coatings or composites 5, and reducing particle sizes 14,15 . Given the fact that reversible insertion of all three Li atoms that occupy distinct crystallographic positions in monoclinic LVP phase is feasible under slow or even fast charge/discharge rates, literatures on carbon coatings and developing composites of micro-sized LVP are available 13, . Nevertheless, only a few studies on the solid state and sol-gel syntheses of nanostructured LVP are reported 5,11,15 . One of the difficulties to prepare nano-sized LVP is probably due to the high crystallization energies which may not only lead to unavoidable particle growth at elevated temperatures but also make carbon coating or composite formation complicated . Meanwhile, polyol approaches are advantageous for the preparation of electrode materials at considerably moderate temperatures. Moreover, poly alcohols or polyols such as ethylene glycol and di/tri/tetra ethylene glycol can play multi-roles of a solvent, a reducing agent and a carbon source . Further, the importance of combustion based approaches to synthesize LVP electrodes have also been highlighted recently 26,27 . Hence, it remains essential to develop synthetic routes wherein particle crystallization and limiting particle growth are facilitated simultaneously in order to arrive at optimized carbon coated/composite LVP nanoparticles with high power capabilities. Herein, we report a pyro-synthesis performed in a polyol medium that acts as a solvent, a conductive carbon source and a reducing agent to maintain lower vanadium oxidation states of V (III) and facilitate the formation of LVP. Further, the polyol acts a low-cost fuel, which on combustion yields carbonized structures that may aid in carbon coating and tend to restrict excessive particle growth. Our previous report on a rhombohedra Na 3 V 2 (PO 4 ) 3 /C cathode produced by this pyro-technique revealed the presence of micro and nano-sized particles. The prepared NVP/C cathode clearly demonstrated the possibility of reversible insertion of 4 Na 1 ions within the potential window of 3.8 -1.5 V vs Na/Na 122 . Hence, the pyro-synthesis was employed in the present study to produce an LVP/C cathode and assess its insertion capabilities within different potential windows at various charge/discharge rates. The asprepared carbon-saturated Li 3 V 2 (PO 4 ) 3 possesses amorphous characteristics and require post-heat treatment for complete particle crystallization. However, the as-prepared amorphous LVP/C requires apparently shorter sintering reaction times of approximately 5 h to produce carbon-coated particles that does not display excessive growth. Moreover, the saturated carbon on LVP particles appears to prevent the particle growth even at high sintering temperatures of ,800uC. Although particle growth, to a certain extent, is inevitable at elevated temperatures, the morphology of the prepared sample indicates the presence of spherical and rod-shaped particles with surface carbon coatings. The conductive coating may not only tend to prevent excessive particle growth at high temperatures but also enhance the electrical conductivity of the prepared LVP. In addition, the existence of a carbon network that may improve the electrical connectivity between the particles is also revealed. All these factors appear to significantly influence the overall electrochemical performance of the prepared LVP/C cathode that demonstrated decent reversible insertion of 2/3 Li 1 ions even under high current densities. ## Results and discussion The polyol assisted pyro-synthesis to obtain Li 3 V 2 (PO 4 ) 3 /C is a straightforward approach and is clearly illustrated in Figure 1. The first stage is represented by the complete dissolution of the starting precursors in the polyol solution. The subsequent stage involves the ignition of the homogenous solution with a torch which ultimately results in the rapid precipitation of nanoparticles. In fact, the polyol, ethylene glycol acts as a low-cost fuel that induces a flame and thereby releases ultrahigh exothermic energy which is usefully exploited for precursor decomposition and subsequent nucleation of nanoparticles. The entire reaction proceeds as a continuous process and is completed in just a few seconds. In addition, the presence of phosphoric acid (H 3 PO 4 ) accelerates the carbonization of the polyol to yield carbocations and carbon-carbon double bonds at high temperatures 28 . The resulting carbonized structures may act as physical barriers to prevent particle growth at elevated reaction temperatures. Further, the generation of high energy coupled with the short reaction time facilitates rapid nucleation and tend to suppress grain growth. The hydrocarbon containing polyol serve as a carbon source and the reducing environment offered by the polyol ensures that vanadium is maintained at the lower V(III) state. The as-prepared sample obtained after the pyro-synthetic reaction in Figure 1 corresponds to the carbon-saturated amorphous Li 3 V 2 (PO 4 ) 3 particles. The post heat-treatment at 800uC in argon atmosphere for a short sintering time of 5 h is sufficient to produce highly crystalline particles. Although the elevated heat treatment tends to make particle growth inevitable, the carbonized structures formed from polyol combustion may prevent excessive particle growth. Further, the presence of a carbon network that may act as electrical conduits between the particles may tend to improve the electrochemical properties at even high charge/discharge rates. The crystal structure of LVP/C was elucidated by a combined analysis of the Rietveld refinement performed on the X-ray and neutron diffraction data of the prepared sample. In general, the common problem of detecting lighter atoms (Li) in the presence of heavy transition atoms (V) with many electrons using X-ray maybe overcome by using neutron diffraction which can produce scattered rays of proportional intensities from all atoms irrespective of being light or heavy. Therefore, a combined Rietveld refinement analysis on the X-ray and neutron diffraction data was performed to estimate the structural parameters of LVP and the results are summarized in Table 1 and Table S1 (see Supporting Information). The fitted patterns obtained from the X-ray and neutron diffraction studies are shown in Figures 2a and b respectively. The space group for Li 3 V 2 (PO 4 ) 3 was determined as P2 1/n and, as anticipated, all the constituent atoms including the three lithium atoms, in the monoclinic unit cell are distributed over distinct co-ordinates at the crystallographic 4e site (Fig. 2c). The refined lattice parameter values are slightly lower and the lattice volume of the monoclinic unit cell is significantly lower in the present Li 3 V 2 (PO 4 ) 3 /C than those reported 5,11,20, . In general, the reported lattice volume for LVP/ C is relatively higher than that for pure LVP. However, in the present study, the estimated lattice volume of 886.9 A ˚3 for the monoclinic unit cell is the lowest so far among the reported values for LVP and LVP/C samples. According to Liu et. al, the estimated lattice-volume may depend on the presence of carbon or graphene in the sample 30 . The fractional coordinates of the elements are in close agreement with those reported for monoclinic Li 3 V 2 (PO 4 ) 3 . Moreover, it is well known that the atomic positions of Li are critical to influence the electrochemical performance of Li 3 V 2 (PO 4 ) 3 30 . The three distinct Li sites in the monoclinic lattice at room temperature consist of one four-fold co-ordination (Li1) and two five-fold co-ordination (Li2 and Li3) sites. The refined isotropic thermal data indicate that the values for the Li atoms situated at two locations (Li1 and Li2) are far higher than that of the third lithium site (Li3) (see Table S1, Supporting Information). This observation clearly indicates that the Li1 and Li2 sites predominantly contribute to ionic conduction. Although the Li2 and Li3 sites in the present Li 3 V 2 (PO 4 ) 3 are swapped in comparison to that studied by Patoux et. al, the result is contrary to the finding by the latter that the five-fold co-ordination sites (Li2, Li3) are mainly involved in ionic conduction 20 . The bond lengths and bond angles specific to Li sites were calculated using VESTA software and are listed in Table S2 (see Supporting Information). The Li-O bond lengths at the tetrahedral Li1 site vary between 1.87 and 2.04 A ˚. On the other hand, the longest Li-O bond lengths at the pseudo-tetrahedral Li2 and Li3 sites are calculated to be 2.61 and 2.33 A ˚respectively. The V-O bond distances for the two slightly distorted VO 6 octahedral units representing two distinctive vanadium sites V1 and V2 are estimated to lie in the ranges of 1.82-2.17 A ˚and 1.88-2.12 A ˚respectively (Table S2). The bond-valence sum values for the two octahedral sites of vanadium, namely V1 and V2, were determined to be 3.2 and 3.5 respectively. The calculated values thus indicate that the oxidation states of vanadium are close to 13. The average bond lengths of the PO 4 tetrahedral units vary between 1.54 and 1.61 A ˚(Table S2, Supporting Information). Furthermore, the low values of reliability (x 2 , 2) and weighted (R p , R wp , 5%) factors indicate the consistency of the combined refinement analysis. The ICP analysis also confirmed the molar stoichiometric composition of the prepared Li 3 V 2 (PO 4 ) 3 (Table S3, Supporting Information). Electron microscopic images of the prepared LVP/C were recorded to obtain further information on the particle size/morphology and carbon existence. Figure 3a shows the low-magnification field emission scanning electron microscopy (FE-SEM) image of LVP/C that appears to reveal rod-type particles dispersed among a major portion of apparently smaller particles with spherical morphologies. The rod-shaped particles are observed to range from a few hundred nanometers to micrometers (Figure 3a). This observation implies that particle growth remains unavoidable due to sintering of the as-prepared sample at elevated temperatures. Further, the average particle-sizes of the spherical particles are observed to be in the range of 20-30 nm. In addition, the high magnification SEM image in Figure 3a inset appears to reveal some aggregation between particles. It is highly probable that the short sintering time (,5 h) may be insufficient to cause excessive particle growth even at elevated temperatures (,800uC). The as-prepared LVP which shows amorphous characteristics or consists of loosely bound functional groups at the atomic scale may undergo phase transition to highly crystalline nanoparticles even under very short sintering times. Furthermore, the carbonized structures formed from polyol combustion may not only contribute to limiting particle growth but also facilitate the maintenance of vanadium in the lower V (III) oxidation state. The Transmission electron microscopy (TEM) images of LVP/C under low and high magnification are shown in Figure 3(b-f). The contrast observed in the image of Figure 3b clearly appears to distinguish between the carbon-coated particles and the carbon network present in the sample. The relatively darker locations appear to indicate the presence of carboncoated LVP particles whereas the brighter image seems to display a network-type structure of probably carbon. Fig. 3c shows the tip of a single rod-type particle and the corresponding high magnification image is presented in Figure 3d. The highly magnified picture reveals that the lattice fringe value of 3.03 A ˚is distinguishable in addition to identifying a carbon-coating of 7 nm thickness along the particle boundary. The Fourier transform (FFT) image of the corresponding area is also shown in the inset of Figure 3d. On the other hand, Figure 3e shows a LVP particle with spherical-type morphology. The corresponding high magnification image, displayed in Figure 3f, directly visualizes the lattice fringe in the spherical particle. Further, the FFT image of the portion under study is shown in Figure 3f inset and confirms the crystalline characteristics of the prepared LVP. It is possible that the mixed morphologies of spherical and rod-type particles may promote intimate particle connectivity and thereby influence electrical properties and hence electrochemical performance 35 . The Energy Dispersive X-ray (EDX) elemental mapping studies were performed to further confirm the presence of carbon in the sample. Figure 4 shows the dark field (DF) STEM image and the corresponding elemental mapping images of C, V and P. The images reveal that carbon is well-dispersed among the sample and is similar to that procured for the counterpart elements of V and P. The observation, therefore, appears to indicate the presence of surface carbon coatings and indirectly suggests the presence of a carbonnetwork as well among the LVP particles. The elemental analysis confirmed that the carbon content on the prepared LVP/C was 6.87%. The estimated value appears to be sufficient enough to facilitate the carbon layer and network formation 12,36,37 . The decent dispersion of the respective elements in the sample implies that no phase separations in the nanometer scale occurred during the formation of LVP/C particles. Further, the Brunauer-Emmett-Teller (BET) analysis aided in estimating the surface area of the sample to be 124.15 m 2 g 21 . It is anticipated that the combined factors of carbon painting on LVP and the presence of a carbon-network facilitates the enhancement of electrical conductivity in the prepared phosphate cathode. Figure 5 presents the electrochemical performances of the LVP/C electrode vs Li 0 /Li 1 . The LVP/C electrode was cycled under two potential ranges of 3-4.3 and 3-4.8 V at a current density of 0.1 mA cm 22 . Figure 5a The discharge plateaus that occur at two potentials namely, 3.64 and 3.55 V correspond to the re-insertion of the second Li into the monoclinic host. On completion of the discharge cycle, the complete reduction from V (IV) to V (III) state in the monoclinic host is ensured. The delivered initial specific discharge capacity is 131 mAh/g, which corresponds to almost 100% theoretical capacity (,132 mAh/g) utilization. The slightly higher value of first charge capacity (,143 mAh/g) may be considered as a common problem of initial electrochemical activation of active materials. Average discharge capacities of about 131 mAh/g (Fig. 5a inset) are consistently maintained with almost 100% Coulombic efficiencies at a current density of 0.15 C (10 mA/g) for 50 electrochemical cycles. Hence, the 100% utilization of theoretical capacity by the prepared LVP/C cathode corresponds to the successful reversible insertion of 2 Li per formula. Further, the differential capacity (dQ/dV) plot displayed in Figure 5c clearly reveals that Li-insertion is reversible within the potential window of 3-4.3 V. The electrochemical response of the prepared LVP/C cathode under a voltage window of 3.0 -4.8 V is displayed in Figure 5b. The initial voltage curves reveal that a discharge capacity of 196 mAh/g, which corresponds to almost 100% of theoretical capacity (,197 mAh/g), is achieved under a current density of 0.1 C (20 mA/g). The apparently higher discharge capacity delivered by the LVP/C cathode within the wider potential window corresponds to the insertion/extraction of 3Li per Li 3 V 2 (PO 4 ) 3 . Precisely, the sloping charge profile in the potential range of 4.5-4.8 V may be assigned to the kinetically difficult extraction process of the third Li and the formation of V 2 (PO 4 ) 3 . However, the disorder caused by the vanadium charge ordering in this fully de-lithiated phase contributes to the solid solution phase behaviour observed during the corresponding discharge cycle 29 . The differential capacity (dQ/ dV) plots in Figure 5d reveal a broad but not so intense peak around On the other hand, the broad nature of the peak around 4.0 V during discharge may indicate the formation of a Li x V 2 (PO 4 ) 3 solid solution, as revealed by the sloping discharge profile of the test cell cycled between 3-4.8 V. Although Coulombic efficiencies are maintained in proximities to 100% under extended cycling (Figure 5b inset), the cycle performance of the LVP/C cathode indicates that initial discharge capacity is decreased to 85% after 50 cycles. To procure more information on the capacity fading in the present LVP/C cathode, insitu XRD measurements were performed on a specially designed half cell (see details in Methods section). The spectro-electrochemical cell was put under galvanostatic test within the potential of 3-4.3 V for the initial cycle. The galvanostatic measurement during the second cycle was performed within the wider potential domain of 3-4.8 V. The electrochemical curves obtained for the initial cycle (3-4.3 V) and the second cycle (3-4.8 V) are provided in Figures 6a and 6b respectively. The vertical bars that indicate the specific locations at which the XRD scans were initiated are numbered sequentially. The initial ten XRD scans were initiated at progressive state of charges (SOCs) during initial charging until 4.3 V whereas the following eleven scans (from 11 to 21) were recorded at different state of discharges (SODs) during subsequent discharging until 3 V. The in-situ XRD patterns procured at different SOCs and SODs during first charging and discharging are plotted against the scan numbers in Figures 6c and 6e respectively. On the other hand, the scan numbers 22 to 31 and 32 to 42 correspond to the different SOCs and SODs, respectively, in the second cycle (Figure 6b). The in-situ XRD patterns recorded at various SOCs and SODs of the second cycle are presented against their scan numbers in Figures 6c and 6e respectively. The major diffraction lines of the LVP have been marked in Figure 6(c-f) in order to identify the changes in the peak positions during electrochemical cycling. On comparing the XRD patterns obtained before and after the first electrochemical cycling (scan numbers 1 and 21) within the potential region of 3-4.3 V, no significant variations in the diffraction peak positions are observed. A similar trend in the positions of the major diffraction peaks is observed before and after the completion of the second cycling within the potential domain of 3-4.8 V. This finding indeed supports the conclusion of a recent investigation that the structural integrity of the LVP/C cathode is maintained in spite of the two-phase reaction mechanism and the large volume changes 38 . Therefore, the present results clearly suggest that the capacity fading observed when the LVP/C cathode is cycled to cut-off potentials of 4.8 V may be related to the kinetic problems or the high resistance offered towards the deintercalation of the final Li 1 ion [Li 3-x V 2 (PO 4 ) 3 , (x $ 2)] 29 . Precisely, the fully de-lithiated phase of V 2 (PO 4 ) 3 that exhibits some disorder-ing undergoes a lithium-site/electron ordering or a disorder-to-order transition during the first lithium insertion that ultimately increases the unit cell size of LiV 2 (PO 4 ) 3 . The small increase in the unit cell dimensions of LiV 2 (PO 4 ) 3 formed during discharging affects the kinetics of the electrochemical reaction and leads to the detrimental effect on the cycle performance of the LVP/C cathode when cycled until high cut-off potentials of 4.8 V. In order to assess the rate capabilities, the C-rate performances of the prepared LVP/C cathode under the two different potential windows corresponding to the reversible intercalation of 2 and 3 Li ions per formula are displayed in Figure 7(a-d). Namely, a discharge capacity in the proximity of 130 mAh/g corresponding to 2 Li per formula is realized even at a relatively high current density of 640 mA/g (4.8 C). At a current density of 1280 mA/g (9.6 C), an average discharge capacity of 125 mAh/g corresponding to 95% theoretical value is maintained. It is worth noting that a remarkable average capacity of about 90 mAh/g, corresponding to 68% of theoretical capacity is achieved at rates as high as a current density of 2000 mA/g (15 C) (Figure 7 a,b). The first set of current rates applied to the test cell beginning from 0.05 C until 15 C completes after the 27 th cycle. From the 28 th cycle, the same set of current densities beginning from 0.05 C is repeated sequentially. It is clearly seen that as the applied current rate is progressively increased from 0.05 to 0.3 C, average capacities of 132 mAh/g that correspond to 100% theoretical capacity utilization is still delivered by the LVP/C cathode (Figure 7b). In fact, reports on LVP/C cathodes with different morphologies demonstrating capacities in proximities to 100 mAh/g at current densities beyond 10 C rates are available 13,15,23 . An LVP/ Graphene nanocomposite prepared by a sol-gel method demonstrated consistent capacities of 118 and 109 mAh/g at 5 and 20 C rates respectively 11 . The specific capacities registered by the present LVP/C cathode are indeed comparable to the reported values. Despite the capacity loss experienced when cycled until 4.8 V, the pyro-LVP/C cathode registers average discharge capacities of 137 mAh/g at 1280 mA/g (6.4 C) and 77 mAh/g at 2000 mA/g (10 C), as shown in Figures 7(c, d). As the initial set of current densities progressing from 0.05 C to 10 C are repeated sequentially from the 28 th cycle, the LVP/C cathode still delivers discharge capacities of 140 and 134 mAh/g at 640 (3.2 C) and 1280 mA/g (6.4 C) current rates respectively. A hydrothermally prepared Sc-doped LVP (Li 3 V 1.85 Sc 0.15 (PO 4 ) 3 /C) cathode was reported to demonstrate capacities of 80 mAh/g at 5 C rates within a potential window of 3-4.8 V 37 . Other reported hydrothermal routes to one dimensional nanostructures and micro-flakes of LVP yielded capacities of 101 and 164 mAh/g at 10 and 5 C rates respectively 39 . The registered capacities in the present study are comparable to those reported for other monoclinic-Li 3 V 2 (PO 4 ) 3 cathodes cycled until high cut-off potentials of 4.8 V 11,17,18 . The impressive rate capabilities of the prepared LVP/C may be explained on the basis of several reasons. Firstly, the polyol combustion provides spontaneous ultrahigh energies that facilitate the nucleation and growth of amorphous particles. During post-heat treatment, the amorphous LVP undergo phase transition to highly crystalline LVP nanoparticles. The carbonized structures that formed as a result of the polyol combustion may serve as barriers and thus tend to inhibit particle growth. Secondly, the short time durations of post-heat treatment ensures the presence of carbon in the final product. The polyol carbonization also enables carbon painting on particle surfaces which can in turn act as act as electrical conduits during electrochemical reaction. Thirdly, the presence of a carbon-network appears to favor better particleconnectivity and hence improve the electronic conductivities and contribute to high rate performance. Moreover, the presence of rod-type particles in addition to nano-scaled particles may offer better inter-particle connectivity and thereby contribute to enhanced electrochemical performances 35 . The surface morphology of the asprepared sample (not shown) clearly indicates the presence of just spherically shaped particles with almost similar particle-sizes in the annealed sample. However, the major difference lies in the fact that annealing produces random growth of rod-type crystals, as presented in Figure 3. Furthermore, the preparation process adopted in the present study is a cost-effective and comparatively faster process than conventional techniques and may lead to the possibilities for large-scale development of cathode materials for high power lithium ion batteries 24 . In summary, a polyol assisted pyro-synthetic reaction was used to obtain a monoclinic LVP/C cathode with mixed morphologies for lithium batteries. A combined refinement was performed on the XRD and neutron diffraction data to determine the lattice parameter values. The LVP/C cathode revealed morphologies of carbon-coated spherical/rod-type particles by using SEM and TEM studies. The prepared LVP/C cathode demonstrated initial discharge capacities of 131 and 196 Ah/g, equivalent to 100% theoretical capacities of 132 and 197 mAh/g within the potential windows of 3-4.3 and 3-4.8 V that correspond to 2 and 3 Li extractions at current densities of 0.15 and 0.1 C respectively. More importantly, remarkable average specific capacities of 125 and 90 mAh/g at current densities of 1280 and 2000 mA/g respectively were delivered by the LVP cathode within the smaller potential range and may suit high power applications. The impressive rate capabilities of the prepared LVP/C may be attributed to the effective surface carbon-coating on the LVP particles. In addition, the presence of a carbon network also appears to contribute to the impressive electrochemical performance. Moreover, the mixed morphology of nanorods and spherical particles may enhance electrochemical performance. Furthermore, the rapid pyro-synthesis adopted in the present study may offer opportunities not only to realize phosphate-based electrodes for battery applications but also may provide solutions to scale-up the production of prospective energy materials. ## Methods Material synthesis. Li 3 V 2 (PO 4 ) 3 powders were obtained by the pyro-synthetic approach using Lithium acetate (C 2 H 3 LiO 2 , $99% -Aldrich), vanadium acetylacetonate (C 15 H 21 O 6 V, 97% -Aldrich), and phosphoric acid (H 3 PO 4 , $85% -Daejung) as starting materials. Initially, the starting precursors were dissolved in 80 mL of ethylene glycol (C 2 H 6 O 2 , $99% -Daejung) in the molar ratio 1.55151.5 (Li5V5P) at room temperature. After obtaining a homogenous solution, the final solution was uniformly poured onto a hot-plate maintained at 200uC. The polyol precursor solution was ignited with a torch to induce a self-extinguishable combustion process. Subsequently, the as-prepared powder was annealed at 800uC for 5 h under argon atmosphere to obtain the LVP/C powders with high crystallinity. Neutron and X-ray diffraction studies. Neutron powder diffraction (NPD) data were collected at room temperature for the prepared sample using a high-resolution powder diffractometer (HRPD) at the Hanaro Center of Korea Atomic Energy Research Institute. The neutron diffraction studies were performed over scattering angles ranging from 15u to 160u using 1.8367 A ˚neutrons, with 4 h of collection time per pattern. 5 g of the sample was contained in a vanadium can. Diffraction data were co-refined with X-ray diffraction (XRD) data obtained from a MPD X-ray diffractometer with Cu Ka radiation (l 5 1.5406 A ˚) operating at 40 kV and 30 mA within the scanning angle, 2h, ranging between 20 and 100u in steps of 0.02u, using the General Structure Analysis System (GSAS) 40 . Electron microscopy (HR-TEM and FE-SEM) analyses. The particle morphologies and sizes were determined by field emission-scanning electron microscopy (FE-SEM) using an S-4700 model from HITACHI. The High-resolution transmission electron microscopy (HR-TEM) and transmission electron microscopy (TEM) images were recorded using an FEI Tecnai F20 at a 200 kV accelerating voltage. Molar stoichiometric and surface area studies. The stoichiometric molar composition of Li 3 V 2 (PO 4 ) 3 /C was analyzed by inductively coupled plasma atomic emission spectrometer (ICP-AES) using OPIMA 4300 DV from Perkin Elmer. The sample surface area was measured by the Brunauer Emmett and Teller (BET) method using a Micromeritics ASAP2010 (Norcross, GA, USA) instrument. Electrochemical characterization. The electrochemical properties of the Li 3 V 2 (PO 4 ) 3 particles were evaluated using lithium metal as the reference electrode. The cathode was fabricated by mixing the active material with 23 wt% carbon (16 wt% carbon black plus the , 7 wt% carbon already present in the active material) and polytetrafluoroethylene (PTFE) (7 wt%) was used as a binder. Usually, a loading of 3.5 mg cm 22 as the active material was used. This mixture was pressed onto a stainless steel mesh and dried under vacuum at 150uC for 10 h. The cell consisted of a cathode and lithium metal anode separated by glass fiber. The electrolyte used was a propylene carbonate (EC) containing 1 M LiPF 6 . In-situ XRD measurements. The in-situ XRD measurements were performed at beamline 1D KIST-PAL, Pohang Accelerator Laboratory (PAL) using a MAR345image plate detector operating at2.5 GeVwith a maximum storage current of 200 mA. The X-ray beam was focused by a toroidal mirror and monochromatized to 12.4016 keV (0.9997 A ˚) by a doublebounce Si (111) monochromator. The Si(111) monochromator and a Si(111) analyzer crystal were used to provide a high-resolution configuration in reciprocal space. The patterns were recorded based on the wavelength value of 0.999 A ˚. However, the XRD patterns displayed in the present study were plotted after re-calculation of 2h values based on the conventional Cu Ka radiation (l 5 1.5414 A ˚). During the preparation of the in-situ cell, the electrode active material mixed with carbon black and PTFE binder (in the ratio mentioned earlier) was cast on stainless steel mesh and assembled in a spectro-electrochemical cell. The cell was cycled to a fully charged/discharged state by a portable potentiostat at constant rate of 0.1 mA/cm 2 . Kapton tape was applied on the apertures of the outer cases of the test cell.

chemsum

{"title": "Pyro-synthesis of a high rate nano-Li3V2(PO4)3/C cathode with mixed morphology for advanced Li-ion batteries", "journal": "Scientific Reports - Nature"}

tunable_enzyme_responses_in_amphiphilic_nanoassemblies_through_alterations_in_the_unimer–aggregate_e

## Abstract: Developing design rules that offer tailorability in materials' response to enzymes is of great importance, as such materials are of interest in a variety of biomedical applications including sensing, diagnostics and drug delivery. Using an amphiphilic oligomeric platform, we show that the degree of polymerization and hydrophilic-lipophilic balance variations can be utilized to alter the unimer-aggregate equilibrium, which in turn offers robust tunability of the host-guest properties of the amphiphilic nanoassemblies.We found that oligomeric assemblies with higher degree of polymerization are less sensitive to enzymatic degradation and release the guest molecules at a slower rate. Similarly, increasing the hydrophilicity makes these assemblies more sensitive to enzymes. These trends can be understood by correlating these changes to predictable modifications in the dynamics of the unimer-aggregate equilibrium, which affects the substrate availability for enzymes. These findings provide insights into rationally tuning the response of enzyme-sensitive supramolecular assemblies. ## Introduction Enzymes, as one of the most essential class of macromolecules in living organisms, are known to catalyse more than 5000 biochemical reactions efficiently and serve a variety of functions in biological processes. 1 Therefore, dysregulation of enzymatic activities has been associated with many human pathologies. In this context, introducing enzymes as stimuli to trigger specifc responses in artifcial supramolecular assemblies has been of interest, as they have potential application in areas such as activity profle based biological imaging and drug delivery. A promising design strategy that leads to such materials involves covalent incorporation of substrate functionalities in self-assembling molecules, such as amphiphilic macromolecules, where the specifc catalytic action of an enzyme covalently modifes the substrate moiety. If it were to be designed such that the product of this enzymatic reaction exhibits distinctly different self-assembly features, compared to the substrate, then there exists a unique opportunity for programmable changes in the nanostructures and their host-guest properties. Many supramolecular systems including polymeric nanoparticles, hydrogels, silica nanoparticles and gold nanoparticles have displayed adaptive behaviours toward enzymes. Tunability of the kinetics of the enzymatic response still remains a challenge, as it is mainly influenced by two factors: accessibility of an enzyme to the substrate moiety and the degree of difference in the host-guest properties between the reactant and product assemblies. In the case of amphiphilic assemblies, our group and others have shown that enzymatic activation usually occurs in the unimeric state, where the substrate is more accessible to the enzyme than in the assembled micellar form. 28,29 Following these fndings, we have been interested in investigating how the reaction kinetics and the ensuing changes in the host-guest characteristics would be affected by tuning the unimer-aggregate equilibrium to alter the assemblies' accessibility to the enzyme. Moreover, we were interested in identifying as to how structural changes in host assemblies, induced by an enzyme, would affect the rate of disassembly and kinetics of guest molecule release. We envisaged that oligomeric amphiphiles would be an ideal choice to address this question, because: (i) these molecules have critical aggregation concentrations (CACs) that are quite low and compare very well with those of amphiphilic polymers; (ii) despite the fact that they do exhibit low CACs, unlike polymers, these are amenable to a well-defned structure-property relationship study as the degree of oligomerization can be precise. Here we report a new modular design of oligomeric amphiphiles with which a precise control over the degree of polymerization (DP) and functional group placement in the scaffolds can be achieved (Fig. 1). These oligomers are expected to self-assemble in the aqueous phase and host hydrophobic guests in their interiors. By varying the DP and hydrophilic moieties of host molecules, we explore the molecular features that underlie the kinetics of enzymatic response in these supramolecular assemblies. ## Results and discussion Since enzymatic activation usually occurs in the unimeric state, where the substrate is more accessible to the enzyme than in the assembled micellar form, we envisaged that shifting the equilibrium between the unimer and the assembled state would provide an opportunity to alter the enzymatic reaction rate. Degree of polymerization is one of the key factors that can alter this equilibrium and thus change the accessibility of an enzyme to its substrate. To test this possibility, it is critical that all the designed amphiphiles possess the same hydrophiliclipophilic balance (HLB). For this purpose, a series of oligomeric amphiphiles from dimer (2-EG5) to pentamer (5-EG5), have been synthesized (Scheme 1). To further evaluate the effects of DP on the enzymatic response, a polymer, P-EG5, with 14 repeating units was also synthesized. In these amphiphiles, penta-ethylene glycol (EG5) monomethyl ether moieties are installed as the hydrophilic functionality, while alkylated coumarin moieties are used as the hydrophobic units. Both these units are attached to the meta-positions of a benzoyl building block, which are then attached to well-defned oligoamines to generate amphiphiles with different degrees of oligomerization. In all these systems, the coumarin moiety is chosen as the covalently-appended model guest molecule. In order to release this guest molecule in the presence of an enzyme, we use an acetal-ester linkage to connect the coumarin to the oligomer. The esterase-induced cleavage of the carboxylate moiety would create a hemi-acetal coumarin, which is hydrolytically unstable. This hemi-acetal therefore rapidly hydrolyzes further to generate a highly fluorescent, 4-methylumbelliferone. In addition to releasing this covalently attached molecule, this transformation also replaces the aryl moiety on the hydrophobic side of these amphiphiles with a carboxylic acid moiety. This results in a signifcant change in the HLB of the amphiphile. Note that this series of amphiphiles share all the common structural features including the backbone, and hydrophobic and hydrophilic functionalities; the only variation within this series of amphiphiles is DP. Therefore, this investigation allows us to inquire about the impact of DP upon selfassembly and enzyme induced disassembly events. In addition to DP, the HLB of oligomers is another factor that impacts the unimer-aggregate equilibrium. To test this possibility, with the same oligomer series as above, we simply increased the length of the oligoethyleneglycol chain length from fve to eight units. Thus, we synthesized four more oligomers 2-EG8, 3-EG8, 4-EG8, and 5-EG8 (Scheme 1). We hypothesized that the increase in hydrophilicity upon going from penta-ethylene glycol monomethyl ether (EG5) to octaethylene glycol (EG8) monomethyl ether would increase the dynamics of the unimer-aggregate equilibrium, which will then increase the availability of the substrate moiety for the enzymes. In this study, we also test this hypothesis. The amphiphilic oligomers were designed in such a way that they can be synthesized in a modular fashion, providing a facile way to vary the number of repeating units and functional group placement. The synthetic routes to the target oligomers are exemplifed by the synthesis of trimer 3-EG5 in Scheme 2 (see the ESI † for detailed procedures and characterization). The 3,5disubstituted-benzoyl chloride molecule 1a was reacted with N,N 00 -dimethyl diethylenetriamine under basic conditions to generate the substituted oligoamine scaffold 1b. This molecule now contains the pentaethyleneglycol hydrophilic unit and the alkyne moiety to anchor the hydrophobic unit. The hydrophobic and fluorogenic enzyme substrate was then attached to all three repeat units of the oligomer using the Huisgen 1,3-dipolar cycloaddition reaction, the so-called "click" chemistry, 16 to yield the desired oligomer 3-EG5 (Scheme 2). We frst investigated whether these oligomeric amphiphiles would form aggregates in the aqueous phase, since they contain both hydrophobic and hydrophilic moieties. If self-assembly occurs, the interior of these assemblies would have the capability to non-covalently encapsulate hydrophobic molecules. To test this, the oligomers were directly dissolved in phosphate buffer and non-covalent incorporation of a solvatochromic dye, Nile Red, within these assemblies was attempted. We found that at lower concentrations of oligomers, the emission intensity of Nile Red was quite low. However, once the concentration of the oligomers reached a certain point, a rather sharp increase in emission intensity was observed. This onset point is taken to be the onset of hydrophobicity-driven aggregation, which is estimated to be the critical aggregation concentration (CAC) of these oligomers. As shown in Table 1, with DP increasing from 1 to 13, the CAC values of these oligomers vary from 75 mM to 0.58 mM (Fig. S1 †). In general, oligomers with higher DP tend to aggregate at lower concentrations, despite the fact that the HLBs of all these oligomers are identical. At the same DP, the systems with longer ethylene glycol chains as the hydrophilic moiety exhibited higher CAC values. The solution phase sizes of these nanoassemblies were then measured by dynamic light scattering (DLS) at a concentration above their CACs. We observed an average hydrodynamic diameter ranging from 100 to 300 nm for these assemblies (Table 1). The spherical morphology and size of these assemblies were further ascertained using transmission electron microscopy (TEM), as shown in Fig. S2. † Note that we hypothesized that if the HLB of the oligomers was kept constant, then oligomeric amphiphiles with higher DP would be hydrolyzed by the enzyme at a slower rate than their counterparts with lower DP. To test this, we frst measured the enzymatic cleavage rates of all oligomers. Since the enzymatic reaction releases the fluorescent byproduct 4-methylumbelliferone, we were able to monitor the cleavage rates spectroscopically. For an accurate comparison, it is necessary that all these oligomer solutions are not only prepared at concentrations above their respective CACs but also contain the same concentration of the substrate functionalities, regardless Scheme 2 Synthesis route to oligomers exemplified using 3-EG5 scheme. of their DP. To meet these two criteria, we prepared oligomer solutions that contain 200 mM enzyme substrates (based on coumarin), i.e. 100 mM dimer, 66.7 mM trimer, 50 mM tetramer, and 40 mM pentamer, and then treated with 60 nM esterase. As shown in Fig. 2, a clear trend of the enzymatic reaction rate was observed for these oligomers with EG5 as the hydrophilic moiety; amphiphile 2-EG5 exhibited the fastest enzymatic rate over 48 hours, systematically followed by 3-EG5, 4-EG5 and 5-EG5. Moreover, when the same concentrations of the enzyme and the substrate were used in the case of the 14-mer P-EG5, the molecular weight of which is comparable to that of polymers, little hydrolysis was observed from the emission spectra. These results are consistent with our hypothesis that a higher DP would result in a slower enzymatic reaction rate, which in turn provides a convenient handle to tune reaction rates of enzymes and the resultant release of the covalently bound molecules. When the same experiments were performed with the second series of oligomers (the EG8 series) that contain longer ethylene glycol chains as the hydrophilic group, a similar trend indeed observed, i.e. the hydrolysis rate decreases for oligomers with higher DP. These results again confrmed our hypothesis that amphiphiles with higher DP are less accessible to enzymes and thus more stable compared with oligomers with lower DP. Meanwhile, comparison of the two series of oligomers also allows us to evaluate the HLB effects on the enzymatic hydrolysis rates of these oligomers. Note that the basis for our hypothesis that a higher degree of oligomerization would cause a slower reaction rate is that the dynamics of the unimeraggregate equilibrium would be slower at higher DP. The results above support this hypothesis. If this were true, then it should also follow that if the hydrophilicity of these oligomers changes, the dynamics of the unimer-aggregate equilibrium would also be affected, which would in turn alter the sensitivity of these oligomers to enzymes. To test this idea, we compared the hydrolysis rates of EG5 oligomers and EG8 oligomers under the same experimental conditions. Interestingly, we observed that the cleavage rates of the covalently attached molecules from 2-EG5 and 2-EG8 were very similar (Fig. 2c). However, when the DP increases to trimeric or higher, n-EG8 oligomers with longer ethylene glycol chains indeed consistently exhibited faster cleavage, compared to their corresponding n-EG5 oligomers with shorter ethyleneglycol chains (Fig. 2d-f). These results suggest that lowering the hydrophilicity of oligomers will make them more stable in the presence of esterase. This is reasonably expected, because an increase in hydrophilicity is expected to increase the dynamics in the unimer-aggregate equilibrium, which facilitates the enzyme's access to its substrate functionalities. We attribute the lack of signifcant difference between 2-EG5 and 2-EG8 assemblies to the fact that these low order oligomers are already sufficiently dynamic, such that there is no signifcant advantage to increasing the hydrophilicity of the oligomeric amphiphiles from EG5 to EG8. Next, we were interested in evaluating the effect of the enzyme-induced changes in the HLB of the amphiphiles upon their host characteristics for hydrophobic guest molecules. We were especially interested in identifying whether this anticipated molecule release event will follow a DP-and hydrophilicity-dependent trend observed in the covalent modifcation of the amphiphile. To test this, we encapsulated a hydrophobic fluorophore, 1,1 0 -dioctadecyl-3,3,3 0 ,3 0 -tetramethylindocarbocyanine (DiI), into these assemblies. The DiIencapsulated oligomeric assemblies were treated with esterase and the molecule release was assessed by fluorescence change. A change in fluorescence is anticipated in this case because the DiI molecule is insoluble in aqueous solutions and therefore precipitates out of solution, upon release from the hydrophobic pockets of these amphiphilic assemblies. As with the experiments above, the concentrations of esterase and the substrate functionalities in the oligomers were maintained for appropriate comparison of the relative rates of molecule release. Indeed, we found that the guest release depends on the DP of oligomers at constant HLB of the molecule, i.e. within the same oligomer series (EG5 or EG8 series). That is, assemblies from higher order oligomers exhibit the ability to more stably encapsulate the guest molecules and respond to the enzyme much more slowly compared to the lower order oligomers (Fig. 3). Also, assemblies with longer ethylene glycol chains can release guest molecules much faster in the same time range (Fig. 3 and S3 †). These results show that a precise control over the release kinetics of non-covalently encapsulated guest molecules can also be achieved by tuning the molecular structures. Comparison of data for the covalent molecule release based on the enzymatic cleavage of the substrate functionalities and the release of the non-covalently bound guest molecules revealed that the latter process lags behind the former process. The potential reason behind this difference is that the enzymatic cleavage of the covalently attached guest molecules happens frst, which is followed by the loss in capability of the amphiphilic assemblies to hold the guest molecules to cause molecule release. In this scenario, the intermediate states of the aggregated assemblies generated by the enzymatic reaction (e.g. only one of the coumarin moieties cleaved in a pentameric amphiphile) also can bind to guest molecules, but their relative ability to act as a host might be lower. This process in conjunction with the need for a critical concentration of DiI to cause its precipitation likely manifests itself as the lag in the non-covalent guest release, relative to the covalent modifcation of the oligomers by the enzyme. Since the enzymatic cleavage of hydrophobic groups seems to be the primary reason for assemblies to lose their stability and capability to hold guest molecules, it is likely that this enzyme reaction induces morphological changes of the aggregated assemblies. To test this possibility, we monitored the temporal evolution of the size of these assemblies by DLS. We found that the size of assemblies changes immediately after esterase was introduced into these systems. As shown in Fig. 4, both 2-EG5 and 3-EG5 completely disassembled in the presence of the enzyme: the size of assembly 2-EG5 sharply decreased from 240 nm to 20 nm, while assembly 3-EG5 formed an 35 nm assembly from an initial size of 220 nm in 48 hours; this size change was also confrmed by TEM images which showed clear spherical structures initially but few visible aggregates after 48 hours of enzymatic reactions (as shown in Fig. S4 †). However, the size of oligomers 4-EG5 and 5-EG5 remained relatively unchanged over the same timeframe. Furthermore, a similar trend of assembly size change was observed for n-EG8 oligomers with longer EG chains under the same experimental conditions. While 4-EG8 and 5-EG8 were rather more stable in the presence of the enzyme, both 2-EG8 and 3-EG8 completely disassembled in the presence of the enzyme at a faster rate compared with the corresponding n-EG5 oligomers, respectively. These results suggested that enzymatic cleavage can induce the disassembly process. Also, both DP and Fig. 3 Non-covalent guest (DiI) release from nanoassemblies. HLB variations of oligomeric amphiphiles can alter the disassembly kinetics, which correlate well with the guest release profles of both covalently bound and non-covalently bound hydrophobic molecules. ## Conclusions In summary, two series of oligomeric amphiphiles were prepared to evaluate the possibility of tuning enzyme-induced changes in their self-assembly properties and host-guest characteristics. We have shown that: (i) when the degree of oligomerization increases in the amphiphiles, the enzymatic reaction rate decreases. This offers a straightforward opportunity to tune the release kinetics of covalently-appended guest molecules. (ii) This reaction kinetics can also be tuned by varying the hydrophilic-lipophilic balance of the selfassembling substrate molecule itself, where an increase in hydrophilicity accelerates the molecule release rates. (iii) In the assemblies where the enzyme-induced alteration in the HLB occurs at a reasonable rate, i.e. in lower order oligomers, a signifcant change in size and morphology of the assemblies was also observed. (iv) Non-covalently bound guest molecules can also be released from these amphiphilic assemblies in response to the enzyme-induced alteration in the HLB, the trends of which closely follow those observed in the release of the covalently bound guest molecules. The trends in the enzymatic reaction rates and the change in the host-guest characteristics can be understood by correlating structural variations to the change in the dynamics of the unimer-aggregate equilibrium. Factors that lead to faster unimer-aggregate equilibrium dynamics lead to faster enzymatic response. Overall, this study provides two simple and straightforward approaches for altering enzyme-induced changes in amphiphilic assemblies, which in turn offer tunability of the release kinetics of covalently and non-covalently bound guest molecules from these assemblies. The fndings presented here could provide a basis for designing enzyme responsive materials with controlled release capabilities for biomedical applications.

chemsum

{"title": "Tunable enzyme responses in amphiphilic nanoassemblies through alterations in the unimer\u2013aggregate equilibrium", "journal": "Royal Society of Chemistry (RSC)"}

effective_carbon_number_and_inter-class_retention_time_conversion_enhances_lipid_identifications_in_

## Abstract: Chromatography is often used as a method for reducing sample complexity prior to analysis by mass spectrometry, the use of retention time (RT) is becoming increasingly popular to add valuable supporting information in lipid identification.The RT of lipids with the same headgroup in reverse-phase separation can be predicted using the effective carbon number (ECN) model. This model describes the effect of acyl chain length and degree of saturation on lipid RT, which increases predictably with acyl chain length and degree of saturation. Furthermore, we have found a robust correlation in the chromatographic separation of lipids with different headgroups that share the same fatty acid motive. By measuring a small number of lipids from each subclass it is possible to build a model that allows for the prediction of the RT of one lipid subclass based on another.Here, we utilise ECN modelling and inter-class retention time conversion (IC-RTC) to build a glycerophospholipid RT library with 481 entries based on 136 MS/MS characterised lipid RTs from NIST SRM-1950 plasma and lipid standards.The library was tested on a patient cohort undergoing coronary artery bypass grafting surgery (n=37). A total of 129 unique circulating glycerophospholipids were identified, of which, 57 (4 PC, 24 PE, 4 PG, 15 PI, 10 PS) were detected with IC-RTC, thereby demonstrating the utility of this technique for the identification of lipid species not found in commercial standards. ## Introduction Lipids are a complex class of molecules that play a pivotal role in the structure and function of living cells, including membrane structure, energy storage, inflammation, protein folding and aspects of second messenger signaling [reviewed in (1)(2)(3)]. The well-established role of circulating cholesterol contained in low-density lipoproteins (LDL) has been essential to developing our current understanding of atherosclerotic cardiovascular disease and continues to be used as a surrogate measurement for cardiovascular disease risk in clinical practice (4,5). Improvements in mass spectrometry (MS) technology and data analysis methods have caused an influx of clinical cardiovascular lipidomics research (6). Among other findings, lipidomics has identified lipoprotein subclass-specific moieties which may contribute to aggregation or inflammation (7), identified total-and LDL-cholesterol independent changes in the circulating lipidome that are correlated with atherosclerotic cardiovascular disease (8)(9)(10)(11)(12), and has been shown to improve recurrent event prediction in patients with clinical cardiovascular disease (13). Traditional analysis of the lipidome often relies on the generation of fragmentation information from nominated masses to identify compounds with MS techniques. This is rigorous and allows for thorough compound characterization; however, targeted acquisition methods (e.g., multiple reaction monitoring (MRM)) require the selection of target lipids before analysis, and any compounds not specified are not measured. Conversely, the identification of compounds from untargeted data, such as MS-only, time of flight (ToF) data, is difficult without MS/MS data for verification. Therefore, most MS approaches must consider the advantages and disadvantages between targeted and untargeted methodologies before acquisition. In recent years, major advances have improved the sensitivity and selectivity of MS based methods for the analysis of lipids. Alongside this, utilizing the chromatographic dimension of separation allows for the prediction of retention times (RT) of compounds based on their physicochemical properties, which has been explored extensively in metabolomics and lipidomics previously (14)(15)(16)(17)(18)(19)(20)(21). Reverse-phase separation allows for the prediction of RT by using the effective carbon number (ECN) model, which describes the tendency of species to elute earlier with fewer carbons in the acyl chain and as the degree of unsaturation increases (22)(23)(24) Despite these advances, there is a great degree of heterogeneity between measurement techniques, sample preparation, and use of internal standards in MS-based lipidomics protocols (25,26). Therefore, appropriate method validation must be performed to determine the reliability and accuracy of measurements. The most widely applicable methods for verification involve the use of stable isotope-labelled internal standards for quantification, and standard reference materials to check measurements against established consensus. Here we describe an approach that leverages hydrophobic interaction (HILIC) chromatography, which is selective for lipid headgroup rather than acyl chain, to identify abundant phosphatidylcholine (PC) species in the National Institute of Standards (NIST) Standard Reference Material-1950(SRM 1950) pooled blood plasma. Abundant species were analyzed with Quadrupole-ToF (QToF)-MS/MS with reverse phase chromatography to build a retention time library. We then utilize the predictable RT behavior of lipids in reverse-phase liquid chromatography to predict the RT of lysophosphatidylglycerol (LPG), lysophosphatidylethanolamine (LPE), lysophosphatidylcholine (LPC), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), phosphatidylinositol (PI) and phosphatidylserine (PS) species from the gradient conditions of a previously published method (19) to create a glycerophospholipid (GPL) RT library. The use of interclass-retention time conversion (IC-RTC), demonstrated for the first time here, also allows for resolution of acyl chain composition for several lipid species without using MS/MS, where they are separated in the chromatographic dimension. Additionally, this method preserves the data from unannotated compounds detected by ToF-MS, allowing for retrospective identification of new compounds. By combining ECN and IC-RTC, it is possible to generate large, accurate lipid databases from a much smaller set of thoroughly characterized measurements. This can be used to expand the number of lipid species measured from a sample cohort post acquisition. To determine the efficacy of IC-RTC in a clinical context, we applied this technique to detect circulating lipid species that could not be detected in the NIST SRM-1950 plasma sample. ## Chemicals EquiSPLASH and UltimateSPLASH ONE lipidomics standards were purchased from Avanti Polar Lipids (Alabaster, AL). Formic acid, ammonium formate, 2-propanol, acetonitrile, and water were purchased from Honeywell (Charlotte, NC). All reagents used were LC-MS grade. A stock solution of EquiSPLASH solution was prepared in a 1:173 v/v dilution in 1:1 v/v isopropyl alcohol: acetonitrile for lipid extraction (Extraction Buffer). ## Plasma Samples NIST SRM-1950 blood plasma standard was purchased from Merck (Darmstadt, Germany). Human blood plasma was collected from patients undergoing coronary artery bypass grafting (CABG) surgery at the Royal Adelaide Hospital (n = 37). Clinical data were gathered from patient records and case notes at the time of consent. Specimen and data collection were in accordance with approved human ethics (CALHN ethics number R20180206). All patients were fasted for 8 h prior to sample collection. Patient characteristics were as follows: the mean age was 65.7 ± 9.5 years, 73% of participants were male, and 54% of patients had diabetes. Clinical samples were assigned an internal ID and injected in the following sequence: six reagent blank injections to bed in the column, two injections of extracted NIST SRM 1950 plasma sample (one spiked with internal standard, one without), one injection of EquiSPLASH internal standard spiked extraction buffer, followed by the clinical samples in randomized order. NIST SRM 1950 plasma was used as a quality control (QC) sample and injected after every 10 samples, immediately followed by a reagent blank. ## Lipid Extraction Plasma samples were removed from -80°C storage and were thawed at 4°C overnight prior to extraction. 174 µL of internal standard spiked extraction buffer was added to 6 µL of blood plasma. Samples were vortexed for 3 s and centrifuged for 15 min ca 16,000 × g and the supernatant aliquoted into QuanRecovery plates (Waters, Milford, MA). NIST SRM-1950 spiked with EquiSPLASH aliquots were used for library building. For the identification of retention characteristics of other GPL sub-classes, 1 μL of UltimateSPLASH ONE was spiked into 173 μL of extraction buffer, which was used to extract an aliquot of NIST plasma. For characterization of the effects of double bond position and stereochemistry on RT, three separate 1:500 dilutions of 1 mg/mL Avanti PC 18:1(9E)/18:1(9E) (850376), PC 18:1(9Z)/18:1(9Z) (850375), and PC 18:1(6Z)/18:1(6Z) (850374) standards (in methanol) were spiked into EquiSPLASH spiked extraction buffer. ## HILIC-MS Candidate Discovery Samples were mass analyzed on a Xevo G2-XS Q-ToF (Waters, Milford, MA) coupled to an Acquity I-Class-FTN UPLC system (Waters, Milford, MA) in negative ion mode. Lipids were separated with a 100 mm Waters Acquity BEH HILIC column (2.1 mm ID, 1.7 µm particle size). Sample chamber temperature was set to 8 °C and the column heater was set to 50 °C. 10 μL of extracted NIST SRM 1950 plasma was injected on column. Mobile Phase A was a 10 mM ammonium acetate solution in 95:5 acetonitrile:water v/v. Mobile Phase B was a 10 mM ammonium acetate solution made up in 50:50 acetonitrile:water v/v. Gradient conditions are described in Table 1. Samples were mass analyzed in negative ion MS E acquisition mode, where collision voltage was switched between elevated and low energy every 0.1 s and the resultant data written to two different functions. Capillary voltage was set to 3.0 kV. Collision voltage was set to 6 V in the low CE channel and set to 30 V in the elevated CE channel. Prior to data acquisition the mass spectrometer was calibrated between 50-2000 Da with sodium formate, with an RMS residual mass error of 0.4 ppm. Mass calibration was maintained using leucine enkephalin lockspray which was sampled for 0.2 s every 30 s. Purge solution was composed of 1:1:1:1 water:acetonitrile:isopropyl alcohol:methanol. Seal wash solution was 10% methanol in water. Needle wash solution was 50% acetonitrile in water. Table 1: LC gradient for HILIC method. ## RP-LC-MS & MS/MS Conditions Lipid extracts were separated using a 100 mm Waters Acquity Premier CSH C18 column (2.1 mm ID, 1.7 µm particle size) heated to 55 °C. Mobile phase A was prepared by combining 1200 mL of acetonitrile, 800 mL of water, 2 mL of formic acid, and 1260.3 mg of ammonium formate. Mobile phase B was prepared by combining 1800 mL of isopropyl alcohol with 200 mL of acetonitrile, 2 mL of formic acid and 1260.3 mg of aqueous ammonium formate (dissolved in 2 mL of water). Sample chamber temperature was set to 8 °C. Conditions for the LC gradient are specified in Table 2. ## Selection of Candidate Ions for Spectral Library generation Mass spectral peaks were extracted from a RT range of 2.9 to 3.3 minutes from the HILIC experiment. This chromatographic region contains the PC 18:1(d7)/15:0 standard and several other common PC candidate masses (Figure 1). An MS/MS-candidate list was generated using the criteria that: the nominal m/z of the 12 C isotope was even; the peak height was over 1x10 3 counts and that the Kendrick mass defect fell between .4 and .8 Da. Putative PC were identified in descending precursor ion intensity order based on intensities in the NIST SRM-1950 sample. The criteria set for identification are listed in Table 3. Other GPL species were selected for MS/MS analysis based on the NIST community consensus paper (20). LPC species were identified based on exact mass and the neutral loss of the acetate ion in the elevated collision energy channel of the MS E data. ## Raw Data Preprocessing An overview of data processing methods is given in Figure 2. Raw feature lists were extracted using MZMine2 version 2.52 (26). A mass list was generated from spectra using the selection criteria that peaks had to have peak heights greater than 500 counts and fall in the 100 to 1000 m/z range. Chromatograms were built from mass lists using the ADAP chromatogram builder module, where a minimum peak height of 1000 counts was required for detection, and a minimum of 5 points with height greater than 500 counts were required for inclusion. An m/z tolerance of ±0.015 Da was set for inclusion of points to the same peak. Extracted peaks were smoothed with an 11-point smooth and deconvoluted with the wavelets (ADAP) chromatogram deconvolution module (27). A signal-to-noise (S/N) threshold of 10 was set with intensity window S/N estimation. A minimum feature height of 1000 counts was selected with a coefficient/area threshold minimum of 50. Peak duration range was set between 0.05 min and 0.50 min. RT wavelet range was set from 0.00 -0.15. ## RT Calibration & Lipid Identification Logic for Clinical Samples Retention calibration was performed individually on the two linear sections of the solvent gradient corresponding to the elution of lyso-and diacyl lipid species respectively. The RTs of the lysolipid-region of chromatograms were calibrated with a flat calibration using stable isotope labelled LPC and LPE internal standards (0.5 min -2.0 min). The diacyl-region RTs were similarly calibrated using labelled PC, PG, PI, PE, and PS internal standards (2.1 min -12.1 min). Library matching criteria for RT correction standards were: m/z within ±10 mDa of theoretical value and chromatographic RT within 7.5% of the library value. RT correction factor was calculated as the relative change in RT for the internal standards between all standard peaks. A linear regression of time vs change in RT was calculated to identify a drift coefficient, which was used to standardize RTs to those in the library in the diacyl-region of the gradient. This correction was applied on a per-sample basis to account for any potential inter-and intra-batch RT variation. Potential sources of variability include: chromatographic column equilibration, variable rates of solvent evaporation from mobile phase reservoirs and measurement inaccuracies in the preparation of new batches of mobile phases. There are two potential situations that can occur when identifying species with the same sum-composition. In the case of lipids that are baseline-separated chromatographically in the acquired MS/MS data peaks, were annotated as distinct measurements, e.g. PC 36:4 can be split into PC 18:2_18:2 and PC 18:1_18:3, with RTs of 5.00 min and 5.15 min respectively. In the case where multiple lipids co-elute within 3 s of other species with the same sum-composition, an aggregate measurement was recorded and assigned only to the total carbon and double bond level. For example, PC 32:0 with a single chromatographic peak at 7.12 min cannot be further resolved with MS-only data, although from MS/MS data is known to be a combination of PC 16:0_16:0 or PC 18:0_14:0. All lipids were normalized to the peak area of the internal standard in the same subclass. From RT-aligned data, lipids were identified based on the RT and identification criteria specified in Table 3. Mass tolerance was set to ±15 mDa, and corrected RT tolerance was set to ±3 s. Lipid identifications were binned to the nearest .7 min to simplify complex regions of the diacyl chromatography and aggregate sn-1 and sn-2 lyso-lipids. ## Headgroup Adduct Candidate Masses for ## Untargeted Data Processing Detected peaks from samples were normalized to the peak area of the PC 33:1 (d7) internal standard. Normalized peak areas were sorted in descending order and iteratively identified with ± 0.015 m/z and ± 0.05 min tolerances. ## Missing Value Handling To account for missing values in data, an approach was adopted from Wei et al, which uses a hybrid of quantile regression imputation of left-censored data (QRILC) and random forest imputation and to handle values which were below LOD and missing at random, respectively (28). After compound identification, raw peak areas of identified lipids were normalized to the PC 33:1(d7) internal standard, to allow comparison with the untargeted data processing. The complete dataset was combined, and missing values were handled based on their minimum recorded intensity (see Figure 2). For features that had a minimum intensity below 5% of the peak area of the internal standard and were detected in more than 70% of samples, missing values were assumed to be left-censored missing data (i.e. below LOD), where missing data points were imputed with QRILC, using the impute.QRILC function from the imputeLCMD R package (29). For features detected in more than 70% of samples and minimum intensity higher than 5% of the internal standard, missing values were attributed as missing at random or missing completely at random, where random forest imputation was used to calculate missing values using the MissForest package in R (30) Fully processed data was mean-centered and scaled using the scale function in base R. ## Data Analysis and Visualization PCA plots were generated using the base prcomp function in R and visualized using the ggplot2 package (31). ## Lipid Identifications A total of 74 PC species were identified using MS/MS from 105 candidate masses (Table 4). Of these, were resolvable to the acyl-chain composition level (i.e., no evidence of other fatty acids with the same elution profile). An additional 6 structural isomers were identified that separated chromatographically. For non-PC subclasses, 14 PE and 5 PI species were observed, along with a total of 22 LPC, 14 LPE, and LPG species. ## Linear Elution of PC species in RP-LC-MS The elution of PC species was observed to be predictable with a second order polynomial trendline, as previously reported (17,18). Here, we observed an increase in RT as the carbon number increased, while keeping the number of double bonds consistent (Figure 3). The inverse relationship was also true when plotting RTs against double bonds, while keeping the carbon number consistent. Low abundance lipids were resolved by utilizing this relationship to provide narrowed RT windows for identification (Figure 3). Furthermore, we also analyzed three PC standards that only differed in their double bond position and/or stereochemistry (Figure 4). We found that PCs containing the 18:1(9Z) were baseline separated from the 18:1(6Z) and 18:1(9E) species, which co-elute in reverse-phase LC-MS/MS. ## Prediction of Non-PC GPLs in Reverse-Phase Chromatography with IC-RTC The RT of different GPL subclasses relative to PCs with the same acyl chain composition was plotted to study the effect of differing headgroups on RT. The resulting trendline was linear with an R-squared value of 1.0 for all observed GPL subclasses (Figure 5). This relationship was used to predict RTs of GPL subclasses from PC identifications (Table 4). To predict the RT of a non-PC GPL, the equations listed in total of 1,513 features were ECN and exact-mass matched in clinical samples. Only species which were quantifiable to their internal standard and predicted using IC-RTC were included. ## Clinical Sample Results A sum-total of 98,821 features were detected in 37 clinical samples and a NIST QC, a mean of 2,352 features were detected per sample. A total of 129 lipid species were identified, where 77 were present in >70% of samples (details in Supplementary Table 1). A total of 1,198 features were detected in >70% of samples. To highlight the heterogeneity of the lipidome, PCA was used to visualize identified lipid data (Figure 6, A) and unannotated feature data (Figure 6, B). A total of 57 lipids unique to clinical samples were detected based on ECN and interclass retention prediction, primarily belonging to PI and PE subclasses (Table 5). When stratifying samples between clinical diagnoses of recent myocardial infarction or stable angina, six unidentified compounds were significantly different between groups (p < 0.05, unpaired t-test); however, these features were not assessed further due to low abundance and high variation. ## Discussion The data shown here illustrates the utility of RT-based characterization of lipid species from a matrix containing a large amount of unknown lipid species. The prediction of RT based on lipid physicochemical properties has been previously described in an intra-subclass context but has not yet been demonstrated on an inter-subclass basis to our knowledge. Our small clinical dataset also highlights the heterogeneity of the lipidome in human patient samples and the necessity of additional identification mechanisms besides exactmass matching in clinical workflows, where IC-RTC and ECN matching reduces false-positive annotations. ## RT Prediction of Lipid Species with ECN The ECN model has been previously described as an additional mechanism to assess the validity of results in reverse-phase chromatography (22)(23)(24). By utilizing the relationship between RT and carbon chain length, it was possible to predict the RT of less abundant species to provide additional identifications (Figure 7). A source of RT variability, potentially impacting RT prediction, is the double bond positioning and stereochemistry of unsaturated fatty acids. This variability is illustrated in Figure 4, where the 6Z, 9Z and 9E variants of PC (18:1/18:1) elute with slightly different measured peak tops. This effect has been reported in the context of PC, PG and TAG species (22,32). In mammalian systems, there is limited diversity in abundant unsaturated fatty acids, these are commonly 16:1, 18:1, 18:2, with either 7Z/E or 9Z/E double bond positioning and stereochemistry (33). This diversity may be reflected in Figure 7, where multiple peaks are observed for lipids with the same MS/MS fragmentation e.g. 18:1_20:4 and 18:1_22:6, although this cannot be conclusively proven with the equipment used. The utilization of post-ionisation structural elucidation techniques such as ozonolysis may be able to resolve this further (34). By measuring only one fatty acid and mapping the effect of differing combinations on the other acyl chain (Figure 7 A), the accuracy of the ECN RT prediction is significantly improved (Figure 7 B). Provided that there is enough biological diversity in samples to detect all combinations of species, a combination of aggregated and single fatty-acid prediction is useful to improve lipidome coverage. RT error i.e. the difference between predicted and observed measurements, is described in Supplementary Table 1. ## Inter-Class Retention Time Conversion of GPLs To our knowledge, the mechanism which allows the conversion of RT between lipids of different subclasses has not been described previously. By utilizing the conversion method described here, it is possible to leverage the high abundance of PC species in blood plasma for the discovery of less abundant lipids from other GPL subclasses. While non-PC lipids are generally much lower in abundance, it is valuable to improve lipidome coverage. In all assessed lipids, the mean error in RTs obtained through IC-RTC was less than 1.5 s for compounds manually verified with MS/MS, and between 0 s and 5 s for compounds processed with the automated retention correction. As shown in Table 5, the abundance of predicted non-PC GPLs is considerably lower than that of other subclasses, as previously described in the context of the NIST SRM-1950 blood plasma standard (20). An additional strength of the method described here is that it can assist in deconvolution of species from MS1 data; for instance, the neutral loss of the acetate ion of PC produces species which are isobaric with PE species with two fewer carbons in the aggregate acyl chain. This fragmentation also occurs in-source, which would complicate identification of compounds from MS-Only data if IC-RTC was not used. However, by using IC-RTC as an isolating mechanism for isobaric peaks, overlaps in RT and mass are restricted to lipid species with the same aggregate carbon chain composition in the diacyl region of chromatography (Figure 8). For broader lipidome analysis in non-plasma research contexts, IC-RTC streamlines identification, as theoretical RTs for lipid species have already been recorded for an additional 347 GPL species could not be routinely detected in blood plasma here. Despite the high accuracy of the IC-RTC, it is crucial that significant findings involving the detection of predicted species are assessed with MS/MS to confirm headgroup fragmentation and acyl chain composition before any assumptions about the underlying biology are made, particularly in cohorts with low statistical power. ## Clinical Data Given the heterogeneity of human blood plasma, the primary use of clinical samples was to detect lipids not reported in the NIST 1950 standard using the ECN and test IC-RTC on a pilot cohort for future studies. We were able to detect a total of 57 candidate precursors that follow the two methods with mass error below 15 mDa that were not identified previously with MS/MS, primarily belonging to PI and PE subclasses (Table 5). The consensus measurements of other GPLs are several times lower than their PC counterparts, therefore their rare detection in clinical samples is not surprising but allows for extra coverage of unexpected lipid species from ToF data (20). The sample preparation protocol took approximately one hour to complete for 37 samples, because of its simplicity it is highly reproducible from sample to sample. The maximum sample throughput using this method is 40 samples per hour, making data acquisition time the bottleneck at 22.5 min per sample. In positive ion mode the triacylglycerol, sphingomyelin and ceramide species elute in the 12.1 min to 16.0 min range of the reverse-phase gradient (data not shown); however, this step could be omitted if GPLs are the only analytes of interest. Using the sample preparation method all internal standards were observed with 100% success and a 0% false positive rate (Supplementary Figure 1). Experimental reproducibility was high, as illustrated by the low scatter in the dimensionality reduction of targeted and untargeted data of the NIST QC sample (Figure 6). Eight significantly altered compounds (p < 0.05, students t-test) were detected in the unannotated data when stratifying between by clinical diagnosis of myocardial infarction vs stable angina. These results were excluded from any subsequent analysis due to having maximum abundance below 1% of the PC internal standard in all cases. When considering the normalized peak areas of all identified lipids, a significant difference in the distribution of lipids containing 5, 6 and 7 double bonds was noted when comparing clinical samples vs technical replicates of the NIST SRM-1950 plasma (Supplementary Figure 2). Due to the high innate biological variance and expected differences in the targeted patient sampling of this study (patients undergoing cardiothoracic surgery for vascular complications) and the participants who donated plasma for the NIST standard, differences were expected when looking at simplified measures of the lipidome. ## Utility of ECN and IC-RTC for Lipidomics Method Development ECN provides an additional degree of structural identification confidence that cannot otherwise be attained without fragment ion information. Furthermore, with the exceptionally high accuracy of IC-RTC, maintaining theoretical RTs of all lipid species is recommended to increase lipidome coverage. This advantage extends particularly well to instruments utilizing MS-only ToF or similar technologies, where no additional fragmentation data is generated. ## Conclusion The use of ECN to confirm lipid RTs provides an additional degree of confidence to compound identification in untargeted lipidomics, particularly when dealing with complex matrices, such as human blood plasma. Furthermore, the use of one lipid subclass (in this instance, PCs) as the basis for the prediction of RT for other subclasses with IC-RTC demonstrates a substantial advantage of reverse-phase chromatography. Nominal RTs for other GPL species can be mined from untargeted lipidomics data, with a mechanism more rigorous than exact mass alone. The combination of ECN and IC-RTC accelerates compound library building, where several theoretical RTs are generated from only one observed measurement. For use with a clinical cohort, this method was used to identify a substantial amount of previously undetected lipid species, present at low abundance. For large scale clinical datasets using untargeted methodology, the methods discussed here allow for the more rigorous identification of lipids from reversephase LCMS data.

chemsum

{"title": "Effective Carbon Number and Inter-Class Retention Time Conversion Enhances Lipid Identifications in Untargeted Clinical Lipidomics", "journal": "ChemRxiv"}

low_carbon_strategies_for_sustainable_bio-alkane_gas_production_and_renewable_energy

## Abstract: Propane and butane are the main constituents of liquefied petroleum gas and are used extensively for transport and domestic use. They are clean burning fuels, suitable for the development of low carbon footprint fuel and energy policies. Here, we present blueprints for the production of bio-alkane gas (propane and butane) through the conversion of waste volatile fatty acids by bacterial culture. We show that bio-propane and bio-butane can be produced photo-catalytically by bioengineered strains of E. coli and Halomonas (in non-sterile seawater) using fatty acids derived from biomass or industrial waste, and by Synechocystis (using carbon dioxide as feedstock). Scaled production using available infrastructure is calculated to be economically feasible using Halomonas. These fuel generation routes could be deployed rapidly, in both advanced and developing countries, and contribute to energy security to meet global carbon management targets and clean air directives. Broader contextThere is an urgent need to develop sustainable and renewable biofuels to address the depletion of fossil fuels and the consequences of their combustion on climate change. Commercially viable bio-LPG production (propane and butane blends) would answer both concerns by reducing the demand on petroleum and natural gas usage, and improving air quality by utilising a cleaner-burning fuel. A secondary global concern is the disposal and/or recycling of organic waste, enabling sustainable energy capture and utilisation, and improvement in the environment and living conditions. Both of these concerns can be met by generating biologically-sourced alkane gases through cultivation of engineered microbial hosts fed on waste carbon sources. The microbial 'chassis' could be engineered to utilise specific waste types (e.g. biodiesel waste or salted milk whey), and low cost bioprocess 'hubs' could be localised at existing waste generating industries. This would increase the recycling of industrial waste, thereby reducing the industries carbon footprint, improving waste management strategies and generating further income. ## Introduction The race to develop economically viable microbial biofuels is a consequence of the pressing need to reduce carbon emissions, improve air quality and implement renewable and sustainable fuel strategies. 1 Current over reliance on fossil fuels has led to concerns over energy security and climate change. This has driven new policies to restrict greenhouse gas emissions, increase the recycling of waste biomaterials and accelerate the delivery of the bioeconomy. 2 Effective sustainable biofuels strategies would comprise scalable production of transportable and clean-burning 3fuels derived from a robust microbial host, cultivated on renewable waste biomass or industrial waste streams, with minimal downstream processing, and (limited) use of fresh water. Embedding production techniques within existing infrastructures for waste processing and fuel distribution would minimise expenditure. Tailoring to specific waste streams would support local economies, waste management, energy self-sufficiency, and carbon reduction in both advanced and developing countries. Propane is an ideal biofuel. This simple hydrocarbon gas is a highly efficient, clean-burning fuel requiring little energy to store in a liquefied state. 2 It is currently obtained from natural gas and petroleum refining. Propane is the third most widely used transportation fuel (20 million tons per annum globally), with existing infrastructure and global markets well established. It is also used for domestic heating and cooking, non-greenhouse gas refrigerants and aerosol propellants. 3 Its 'drop-in' nature boosts the calorific value of current methane/biogas supplies, with lower energy requirements for liquefaction and storage. The only existing alternative production method is the Neste ´process, an energy intensive, catalytic chemical conversion of biodiesel waste (glycerol) reliant on natural gas derived hydrogen. 4 No natural biosynthetic routes to propane are known. Engineered biological pathways to propane have been developed based on decarbonylation of butyraldehyde incorporating natural or engineered variants of the enzyme aldehyde deformylating oxygenase (ADO). The low turnover number of ADO (B3-5 h 1 ), however, limits implementation of these pathways in scaled bio-propane production. 5,6,8 Here we describe blueprints for the scaled and economic production of bio-alkane gas (propane and butane, or 'Bio-LPG') using engineered forms of a recently discovered, blue lightdependent, fatty acid photodecarboxylase (FAP) that catalyzes decarboxylation of fatty acids to n-alkanes or n-alkenes (Fig. 1). 10,11 We have taken a systems engineering approach to convert waste VFAs to bio-alkane gas in live bacterial cultures. The strategies we describe could enable environment-friendly in situ gas generation (e.g. in rural and/or arid communities), dependent on the availability of abundant waste resources, and implemented with CO 2 capture. These low carbon strategies could provide economic, sustainable, secure and clean alternatives to extant petrochemical LPG supplies. ## Materials, services and equipment All chemicals and solvents were of analytical grade or better. Gene sequencing and oligonucleotide synthesis were performed by Eurofins MWG (Ebersberg, Germany). All oligonucleotide sequences can be found in Tables S2-S4 (ESI †). Gene synthesis was performed by Geneart (Thermo Fisher), with codonoptimization for E. coli or Synechocystis. The mounted highpower blue LEDs and LED drivers were from Thorlabs (Ely, U.K.), with spectra centered at 455 nm (bandwidth (FWHM) 18 nm, 1020 mW typical output) and 470 nm (FWHM 25 nm, 710 mW typical output). The custom-built LED blue light array had area of 396 cm 2 of relatively consistent light intensity and a fixed average culture-to-LED distance of 8 cm (Fig. S1; ESI †). Light 'intensity' was measured with a Li-Cor light meter with a Quantum sensor in mmol photons m 2 s 1 (or mE), with background light value subtracted. The photobioreactor was a thermostatic flat panel FMT 150 (500 mL; Photon Systems Instruments, Czech Republic) with integral culture monitoring (OD 680/720 nm), pH and feeding control and an LED blue light panel (465 nm; maximum PPFD = 1648 mE photons). E. coli strain BL21(DE3) was modified by chromosomal deletion of two aldehyde reductase genes yqhD and yjgB (BL21(DE3)DyqhD/DyjgB/Kan R ; GenBank: ACT44688.1 and AAA97166.1, respectively) as described previously. 5 The kanamycin selection gene was removed using the Flp-mediated excision Transformed E. coli cultures (three biological replicates) were grown at 37 1C in LB medium containing kanamycin (30 mg mL 1 ) to a density of OD 600 B 0.6-0.8. CvFAP expression was induced with IPTG (0.1 mM) and cultures were supplemented with 10 mM butyric acid. Triplicate aliquots (1 mL) of each culture were sealed into 5 mL glass vials and incubated at 30 1C for 16-18 h at 200 rpm, illuminated with a blue LED panel. Headspace gas was analysed for hydrocarbon content using a Micro GC. Data were normalized by dividing the propane titres (mg L 1 culture) by the relative protein concentration compared to the wild type (WT) enzyme (Fig. S2; ESI †). Error bars represent one standard deviation for triplicate biological repeats (n = 3). Inset: Structure of the palmitic acid binding region of CvFAP (PDB: 5NCC) shown as a cartoon with secondary structure colouring. Models of butyrate and palmitate in the active site of (c) wild-type and (d) G462V variant of CvFAP. The position of palmitate in the wild-type enzyme is crystallographically determined (PDB: 5NCC). The positions of the remaining ligands were determined by Autodock Vina, and mutagenesis to G462V was simulated using SwissPDBViewer 4.10. The protein is shown as a cartoon with secondary structure coloring, with selected residues shown as sticks. FAD, palmitate and butyrate are shown as atom-colored sticks with yellow, green and blue carbons, respectively. In panels c and d, the dashed line shows a hydrogen bond between palmitate and the wild-type enzyme, while the dotted lines indicate the distance between the C4 carbon of palmitate and the Ca atom of residue 462. All crystal structure images were generated in Pymol. ## Paper Energy & Environmental Science methodology (BL21(DE3)DyqhD/DyjgB). 12 Synechocystis sp. PCC 6803 was modified by chromosomal deletion of the acyl-ACP synthetase (Daas) gene as described previously. 13,14 Halomonas strains TD01 15 and TQ10, and modified pSEVA plasmids have been described previously. 16 Halomonas strain TQ10-MmP1 is a modified version of the TQ10 strain, which had been cured of a recombinant plasmid. ## Gene synthesis, sub cloning and mutagenesis The following N-terminally truncated (DN) FAP enzymes were synthesized (Table S1 Each gene was sub cloned into pETM11 with a N-His 6 -tag for rapid protein purification. The gene encoding thioesterase Tes4 from Bacteroides fragilis (UniProt: P0ADA1) was obtained from plasmid pET-TPC4, as described previously. 5 For the valine to propane pathway, leucine 2-oxoglutarate transaminase from E. coli (ilvE; P0AB80); 3-hydroxypropionaldehyde dehydrogenase from E. coli (Hpad; P23883) and branched-chain keto acid decarboxylase from Lactococcus lactis (KdcA; Q6QBS4) were synthesised and sub-cloned into pET21b (C-His 6 -tag), pETM11 (N-His 6 -tag) and pET28b (N-His 6 -tag), respectively. Variant CvFAP G462V was generated by site-directed mutagenesis of the wild-type using the QuikChange whole plasmid synthesis protocol (Stratagene) with CloneAmp HiFi PCR premix (Clontech). Additional variants (e.g. G462N/W/L/C/I/F/ A/H/Y and those at neighbouring positions; see Fig. 1) were generated using the Q5 and QuikChange site directed mutagenesis kits (New England Biolabs and Novagen, respectively). PCR products were analysed by agarose gel electrophoresis, followed by gel purification (NucleoSpin Gel), or purified using the PCR clean-up kit (Macherey-Nagel). Constructs were transformed into E. coli strain NEB5a (New England Biolabs) for plasmid recircularization and production. The presence of the mutations was confirmed by DNA sequencing followed by transformation into E. coli strains BL21(DE3) and BL21(DE3)DyqhD/ DyjgB 5 for functional expression studies. ## Molecular modelling Substrates palmitic and butyric acid were docked into chain A of the crystal structure of the palmitic acid bound CvFAP structure 5NCC using Autodock vina. 17 Non-polar hydrogen assignment was performed using AutoDock Tools 1.5.6. A cubic search volume with 15 sides was defined with the coordinates of C6 of palmitic acid as the centre, and an exhaustiveness of fifty. Twenty conformations were analysed and the lowestenergy conformation with the substrate in the correct orientation (carboxylate pointing towards the FAD) was selected. Mutations were performed in SwissPDBViewer 4.10, 18 using the exhaustive search function to identify the best rotamer for the mutated residue. ## Multi-enzyme construct generation N-His 6 -CvFAP G462V was sub-cloned into plasmids pET21b and pBbA1c 19 by PCR-mediated In-Fusion cloning. Plasmids were transformed into E. coli strain NEB5a, BL21(DE3) and BL21(DE3)DyqhD/DyjgB 5 for functional expression studies. The multi-gene valine to propane construct was assembled with CvFAP variant G462I in pBbE1k with a single pTrc promoter ( pTrc-ilvE-Hpad-KcdA-CvFAP G462I ) by overlap extension PCR, with vector linearisation and insert(s) amplifications performed by PCR. ## Halomonas construct generation CvFAP wt , CvFAP G462V and CvFAP G462I coding sequences were amplified from pETM11 (lacking His 6 -tag) and inserted (NcoI-XhoI) into Halomonas-compatible plasmid pHal2 downstream of the MmP1 IPTG-inducible phage T7-like RNA polymerase promoter. This promoter is composed of an optimized MmP1-lacO-RiboJ-RBS sequence 16,20 and the CvFAP translation initiation site (bold) comprises part of an NdeI restriction site (underlined): Promoter induction only occurs in Halomonas strain TQ10-MmP1, which contains the cognate chromosome-integrated MmP1 phage RNA polymerase gene. 16,20 pHal2 is derived from pSEVA441 21 and contains the pRO1600 broad host range replication origin; the pRP4 origin of conjugative transfer (oriT); and genes conferring spectinomycin/streptomycin and kanamycin resistance, for selection in Halomonas TQ10-MmP1 and in the E. coli conjugative donor strain S17-1, 22 respectively. pHal2-CvFAP variants were introduced into Halomonas TQ10-MmP1 by conjugation as follows. Kanamycin-resistant transformed colonies of donor E. coli S17-7 were mixed with TQ10-MmP1 on YTS agar plates (yeast extract 5 g L 1 , tryptone 10 g L 1 , NaCl 30 g L 1 , agar 15 g L 1 ), incubated overnight at 37 1C, then streaked onto YTN6 agar (5 g L 1 yeast extract; 10 g L 1 tryptone, 60 g L 1 NaCl, pH 9, 15 g L 1 agar) containing spectinomycin (50 mg mL 1 ) to select for Halomonas transconjugants. Plasmid content of the transconjugants was confirmed by DNA isolation, restriction mapping and sequencing. Chromosomal insertion of the CvFAP G462I gene with the MmP1 promoter (pHal2-derived) and valine to propane pathways (IPTG-inducible and pPorin 69) were performed using a novel suicide vector (pSH) protocol based on previously published methods. 23,24 The insertion plasmids contained the biocatalytic and chloramphenicol resistance (Cam R ) genes surrounded by homology arms, an I-SceI restriction site and a colE1 ori that is not compatible with replication in Halomonas. This plasmid was co-conjugated into Halomonas TQ10-MmP1 with a second spectinomycin-resistant plasmid (pSceI) expressing the gene for the restriction enzyme I-SceI. Expression of I-SceI enabled the linearization of pSH plasmids, facilitating chromosomal integration. 23,24 The sites for integration were chosen based on the intergenic regions in Halomonas showing prior high recombinant protein expression. 23 as the pSH plasmid is not replicated. Integration was confirmed by colony PCR, genomic sequencing and in vivo propane production after pSceI plasmid curing. 23,24 Synechocystis construct generation Two versions of the non-His 6 -tagged C. variabilis CvFAP G462V gene with identical amino acid sequences were constructed in Synechocystis sp. PCC 6803 (fap G462V_Ecoli and fap G462V_cyano ), differing by applying codon optimisation for E. coli and Synechocystis, respectively. For fap G462V_cyano , plasmid pIY505 (pJET-'FAP') 14 variant G462V was generated by site-directed mutagenesis using the QuikChange whole plasmid synthesis protocol. To construct pJET-fap G462V_Ecoli , the non-His 6 -tagged gene in pETM11 was amplified by PCR and cloned into the blunt-ended pJET1.2 plasmid. The gene encoding Tes4 was amplified from construct pET-TPC4 14 and ligated into blunt pJET1.2 plasmid. To clone the mutated fap genes and/or tes4 genes into the erythromycin resistant RSF1010 plasmid, the Biopart Assembly Standard for Idempotent Cloning (BASIC) method was used as described previously. 13,14,25 Gene expression was controlled using either the cobalt-inducible Pcoa or constitutive Ptrc (no lacI) promoters. Prefix and suffix linkers used to create the plasmids are listed in Tables S3 and S4 (ESI †). The following constructs were generated: (i) pIY894: Ptrcfap G462V_cyano ; (ii) pIY918: Ptrc-tes4, fap G462V_Ecoli ; (iii) pIY906: Pcoa-tes4, fap G462V_cyano ; and (iv) pIY845: Pcoa-tes4. Plasmid assembly was validated by DNA sequencing. Plasmids were transformed into the E. coli helper/cargo strain (100 mL; E. coli HB101 strain carrying the pRL623 and RSF1010 plasmids), conjugal strain (E. coli ED8654 strain carrying pRL443 plasmid) 26 and Synechocystis sp. PCC 6803 lacking acyl-ACP synthetase (encoded by slr1609; Daas strain; OD 730 B 1) using the tri-parental conjugation method described previously. 13,14 Each strain had been pre-treated by washing with LB and BG11-Co medium for E. coli and Synechocystis, respectively, to remove antibiotics. The mixture was incubated for 2 h (30 1C, 60 mE), then spread onto BG11 agar plates without antibiotic, and incubated for 2 d (30 1C, 60 mE). Cells were scraped from the agar plate, resuspended in 500 mL of BG11-Co medium, and transferred onto a new agar plate containing 20 mg mL 1 erythromycin. Cells were allowed to grow for one week until colonies appeared. ## Protein expression and lysate production Wild type CvFAP-pETM11 homologues in E. coli BL21(DE3) were cultured in LB Broth Miller (500 mL; Formedium) containing 30 mg mL 1 kanamycin at 37 1C with 180 rpm shaking until OD 600nm = 0.2. The temperature was maintained at 25 1C until OD 600nm = 0.6-0.8. Recombinant protein production was induced with 50 mM IPTG, and maintained at 17 1C overnight. Cells were harvested by centrifugation (8950 g, 4 1C, 10 min), and analysed for protein content using 12% SDS-PAGE gels (Mini-PROTEAN TGX Stain-Free Precast Gels, Bio-Rad). Protein gels were imaged using a BioRad Gel Doc EZ Imager and the relative protein band intensity was determined using the BioRad ImageLab software. Cell pellets were resuspended in lysis buffer (1.2-1.7 mL g 1 pellet; 50 mM Tris pH 8 containing 300 mM NaCl, 10 mM imidazole, 10% glycerol, 0.25 mg mL 1 lysozyme, 10 mg mL 1 DNase I and 1 protease inhibitors) and sonicated for 20 minutes (20 s on, 60 s off; 30% amplitude). Cell-free lysate was prepared by centrifugation at 48 000 g for 30 minutes at 4 1C. Lysate samples were analysed for recombinant protein expression by SDS PAGE. ## Hydrocarbon production In vitro propane production reactions (1 mL) of FAP homologues were composed of cell-free lysate and butyric acid (0.4 mM) in sealed 4 mL vials. Reactions were incubated at 30 1C for 24 h at 180 rpm under illumination (blue LED; 455 nm). In vivo propane production of CvFAP WT and variants in E. coli was performed by the following general protocol: Cultures (20-100 mL) in LB medium containing kanamycin (30 mg mL 1 ; pETM11) or ampicillin (50 mg mL 1 ; pET21b) were incubated for 4-6 h (OD 600 B 1) at 37 1C and 180 rpm, followed by induction with IPTG (100 mM) and butyric acid supplementation (1-1000 mM; pH 6.8). Triplicate aliquots (1-5 mL) each of 3 biological replicate cultures were sealed into vials (4-20 mL) and incubated at 30 1C for 16-18 h at 200 rpm, illuminated continuously with an LED (455 nm or 470 nm). Comparative in vivo studies with 10 mM butyric, isobutyric, valeric, 2-methylbutyric and isovaleric acids were performed as above, with culture induction at OD 600 of 0.6-0.8. For all in vitro and in vivo alkane gas production studies, the headspace gas was analysed for propane content using a Micro GC. Data is expressed as mg hydrocarbon production per litre of fermenting culture. Propane production in Halomonas was performed by a modified E. coli protocol as follows: Cultures were grown in phosphate buffered (50 mM K 2 HPO 4 pH 6.6) YTN6 medium containing spectinomycin (50 mg mL 1 ) for 5 h at 37 1C and 180 rpm, followed by IPTG induction at OD B 1.6. The remainder of the in vivo propane production process was performed as above, with butyric acid concentrations of 10-80 mM. For studies with the valine pathway, amino acids (up to 30 mg L 1 ) were added after induction in place of VFAs. Autolysed brewery yeast extract (waste amino acid source) was produced by culturing waste brewery barley grains from a North of England supplier in YPD medium (10 g L 1 yeast extract, 20 g L 1 peptone and 20 g L 1 glucose), followed by autolysis (2 h at 50 1C) and autoclaving. Waste milk medium was composed of milk whey containing 60 g L 1 NaCl and pH adjusted to 9.0. Propane production in Synechocystis was performed in BG11 medium 13,14 using a modified protocol as follows: Starter cultures in BG11 medium were incubated at 30 1C under 30 mE white LED until OD 720nm reached 1.0 (B4 days). Replicate culture aliquots (2 mL) were harvested by centrifugation and re-suspended in 1 mL BG11 medium supplemented with sodium bicarbonate (150 mM), cobalt(II) nitrate hexahydrate (100 mM; for Pcoa cultures only), 50 mg mL 1 kanamycin, and 20 mg mL 1 erythromycin at 30 1C AE butyric acid (10 mM). ## Paper Energy & Environmental Science Cultures were sealed within 4 mL gas tight vials and incubated at 30 1C for 24-48 h under blue light (average 63 mE). ## Halomonas cultivation Cultures were grown in phosphate buffered YTN6 medium containing spectinomycin (50 mg mL 1 ) for 5 h at 37 1C and 180 rpm. Recombinant protein expression was induced with IPTG (0.1 mM; OD 600 B 1.6), and cultures were supplemented with butyric acid (0-100 mM, buffered at pH 6.6). Triplicate aliquots (1 mL) of cultures were sealed into 4 mL glass vials and incubated at 30 1C for 16-18 h at 200 rpm, illuminated with a blue LED panel. For studies with the Synechocystis extract, cultures (1 mL) were incubated post induction with lysed Synechocystis extract in place of butyric acid. Headspace gas was analysed for hydrocarbon content using a Micro GC. Photobioreactor cultivation was performed with high salt glycerol medium at pH 6.8 (5 g L 1 yeast extract, 1 g L 1 glycerol, 60 g L 1 NaCl, 50 mg mL 1 spectinomycin and 0.5 mL L 1 antifoam; 400 mL) in batch mode, pre-equilibrated at 30 1C with 60-100% stirring output. An overnight starter culture (10 mL) of Halomonas TQ10-MmP1 containing pHal2-CvFAP G462V was added, and the culture was maintained at 30 1C with an airflow rate of 1.21 L min 1 , automated pH maintenance, culture optical density monitoring and ambient room lighting until mid-log phase (4-5 hours). Protein induction by IPTG (0.1 mM) was followed by sodium butyrate addition (60-80 mM pH B 6.8) with continual blue light exposure (1656 mE) for B48 h. Propane production was monitored at 15 min intervals by automated headspace sampling using a Micro GC, while aqueous butyrate and glycerol depletion were quantified by HPLC. ## Synechocystis cultivation The photobioreactor (400 mL) was set up in batch mode with starter culture diluted 3 : 1 in fresh BG11 + medium (BG11 pH 8.0 13,14 containing TES buffer and 1 g L 1 sodium thiosulphate) with 150 mM NaHCO 3 . The culture was maintained at 30 1C with maximal stirring, airflow of 1.21 L min 1 , illumination by a white LED (30 mE), automated pH maintenance (1 M acetic acid in 2 BG11 + ) and optical density monitoring (680 nm and 720 nm). After reaching OD 720nm of B0.5, cobalt(II) nitrate hexahydrate (100 mM) was added as required, the warm white illumination was increased to 60 mE and the integral actinic blue LED light panel provided 500-750 mE blue light. The culture was maintained at 30 1C for 18-48 hours, fed and not fed respectively. Manual headspace sampling for propane content was performed by Micro GC, and butyrate depletion was quantified by HPLC. ## Analytical techniques Propane levels were determined by headspace injection using an Agilent 490 Micro GC, containing an Al 2 O 3 /KCl column, a thermal conductivity detector (TCD) and a heated injector (110 1C; 100 ms injection) using helium as the carrier gas (10.2 psi). During continuous monitoring mode, fermenter exhaust gases were dried by passage through an ice-cooled condenser prior to entering the Micro GC. Compounds were separated isothermally (100 1C) over 120 s under static pressure conditions, with a sampling frequency of 100 Hz. Aqueous culture metabolites (glycerol and butyric acid) were analysed by HPLC using an Agilent 1260 Infinity HPLC with a 1260 ALS autosampler, TCC SL column heater and a 1260 refractive index detector (RID). Cell-free culture supernatant samples (10 mL injection) were analysed isocratically on an Agilent Hi-Plex H column (300 7.7 mm; 5 mM H 2 SO 4 ) at 60 1C with a flow rate of 0.7 mL min 1 for 40 minutes. Analyte concentrations determined by Micro GC or HPLC were calculated by comparing the peak areas to a standard curve generated under the same running conditions. ## Techno-economic analysis (TEA) In this analysis, a number of assumptions were made to provide projected economics and establish benchmarks to assess the state of technology based on current research performance. Design basis and costing data for non-fermentation unit operations were obtained from earlier studies and publications. Rigorous structural and parametric optimisation, heat integration and site analysis were not included at this stage. The main tasks utilized were: (a) Conceptual design of a pilot-scale continuous process as the base case. Each reactor has a 1 m 3 working volume, with an inside battery limit (ISBL) plant (fermentation and propane purification) cost of a process Bd500 000. ## Biocatalyst selection and redesign We surmised that FAP could be engineered to increase the decarboxylation of butyric acid (and other short chain volatile fatty acids; VFAs) 11 to form propane (and other hydrocarbon gases) to enable their production at scale (Fig. 1a). FAP from Chlorella variabilis NC64A has a reported reaction quantum yield of greater than 80% and it catalyzes a 1-step bioconversion of waste VFAs into alkanes. 10 However, it has a reaction specificity that is strongly in favor of long chain fatty acids (C14-C18). 10,11,31 We screened a range of previously identified potential FAP homologues 10 for propane production with butyric acid. Direct kinetic comparison of each purified homologue was not possible as protein instability (aggregation) and flavin loss occurred to varying degrees during protein purification of each enzyme. Consequently, comparative propane production in live cells or cell-free lysates was the main approach taken in this study, with FAP concentrations estimated by quantification of the protein band density from SDS-PAGE (Fig. S2; ESI †). The presence of contaminating flavin and other chromophore (e.g. heme) -containing enzymes in cell lysates prevented active enzyme concentration determination by spectral analysis. The most suitable FAP enzymes for hydrocarbon gas production were identified as the Chlorella variabilis NC64A (CvFAP) and Chlamydomonas reinhardtii (CrFAP) homologues 10,11,31 using biotransformation assays of cell-free extracts (Table S5; ESI †). 10 We selected CvFAP as the target biocatalyst, given the availability of a three-dimensional crystal structure. 10 Using structure-based engineering we identified regions in the natural substrate binding channel of CvFAP that when targeted by mutagenesis may be able to confer increased activity of the enzyme with short chain VFAs. We made a collection of twentyeight CvFAP variants, targeting residues G462, G455, Y466, V453, T484 and A457 for substitution (Fig. 1b). A key substrate channel residue G462 was identified, and substituted for ten other residues (Val, Asn, Trp, Leu, Cys, Ile, Phe, Ala, His and Tyr; Fig. S2 and Table S6; ESI †). The side chains of these residues are in close proximity to bound palmitate in the co-crystal structure of CvFAP WT , and variants were designed to disfavor long chain fatty acid binding (Fig. 1b inset). 10 Propane production in E. coli cultures exposed to blue light was measured and normalized according to each variant's relative expression level (Fig. S1 and S2; ESI †). The most promising variants were found to be G462V and G462I (Fig. 1b; Table S6; ESI †), while variants at other positions produced less propane than CvFAP WT (Fig. 1b). Some variants appeared to show higher propane yields than G462V (e.g. G462A/F/C), but inconsistent expression levels of these variants within biological repeats made it difficult to accurately quantify this. Molecular docking simulations of CvFAP WT and variants G462V and G462I were performed using Autodock Vina 17 to investigate the effects of the amino acid substitutions. This analysis predicted a 30-40-fold weaker binding of palmitate for the G462V/I variants (decrease in binding affinity of B2 kcal mol 1 ; Table S7; ESI †), with a small increase in binding of butyrate (B30% tighter binding; Fig. 1c and d). This is consistent with the observation that purified CvFAP G462I exhibited an increase in propane production compared to wild-type (9-fold after 1.5 hours; Fig. S3; ESI †). Conversely, CvFAP WT showed a 2.6-fold higher rate of pentadecane production from palmitic acid after 45 minutes than CvFAP G462I . Given the high variability in activity detected with some variants using cell lysates (Table S6; ESI †), more detailed activity assays were performed with purified CvFAP wild-type and G462I variant (normalised against FAD content; Fig. S3; ESI †). CvFAP G462I showed increased propane production (6.48 AE 1.04 mM propane) compared to wild-type enzyme (0.76 AE 0.31 mM propane), confirming the trend seen with the cell lysate screening data. We also determined that purified wild-type and variant CvFAP G462I are susceptible to rapid photoinactivation (Fig. S3; ESI †), so process scale production will require regeneration of the biocatalyst through continuous replenishment with live cell culture. This is possible by continuous biomass cultivation -an approach we have adopted in developing the design of bio-LPG production platforms described below. Next, we investigated the effect of butyric acid concentration on propane production using live E. coli cultures. As butyric acid lowers the pH of LB medium (Fig. S4a; ESI †), we performed studies with the CvFAP G462I variant with pH control (buffered at pH 6.8; Table S8, ESI †). This showed propane production was maximal at around 50 mM butyric acid (Fig. S4b; ESI †). Under these conditions, the molar ratio of propane production to butyrate consumption was 1.9 to 1. Propane titres were also affected by the plasmid backbone used, and the position of fused purification tags to CvFAP. We observed a 6.4-fold increase in propane production using CvFAP G462V in plasmid pET21b compared to pETM11 (48.31 AE 2.66 mg L 1 culture). These levels are higher than maximal propane production observed previously with ADO (32 mg L 1 culture) using an E. coli strain upregulated for butyric acid production. 5 This highlights the need to explore multiple plasmid backbones, and the effect of location and size of protein tags on gas production in vivo. ## Tuneable bio-LPG blends Photodecarboxylation of other short chain fatty acids (butyric, isobutyric, valeric, 2-methylbutyric and isovaleric acid) was investigated in vivo with wild-type and four CvFAP variants (G462V/A/I/F). The major gas produced, whether propane, butane or isobutane, was dependent on which VFA was supplied (Fig. 2a). Additional gases produced arose from naturally arising VFAs in cell extracts. Variant G462I showed the highest levels of gas production, especially with the branched chain substrates isovaleric and 2-methylbutyric acids (5-8-fold higher than with G462V; Table S9; ESI †). With CvFAP G462I , propane and butane production from butyric and valeric acid were less than 2-fold higher than with CvFAP G462V . Variants G462V and G462A generated similar levels of propane and butane, but G462A showed a greater variation in hydrocarbon titre (Fig. 2a; Table S9; ESI †). These data consolidate the finding that residue 462 is important in conferring activity with short chain carboxylic acids. The photodecarboxylation products of these VFAs can be used to make biological LPG-blends. As CvFAP G462V generated similar titres of propane and butane, this was chosen to explore the production of tuneable bio-LPG blends of varied composition. The most common gases found in LPG blends are propane and n-butane. Blends may also contain isobutane, ethane, ethylene, propylene, butylene and isobutylene. The exact composition of LPG is country-specific, and can be varied between seasons as required, for example, in order to assure proper vaporisation in winter. 32 In the UK, LPG is 100% propane, while in Italy the propane : butane ratio varies from 90 : 10 to 20 : 80 (Fig. 2b). As CvFAP G462V can generate both propane and butane at similar titres, the possibility of producing country-specific bio-LPG blends by varying the ratio of externally supplied butyric: valeric acids was investigated. The butyric : valeric acid ratios fed to live cultures were closely correlated with propane : butane ratios measured in the culture headspace (Fig. 2b; Table S10; ESI †). ## Paper Energy & Environmental Science This indicates the ease with which tuneable bio-LPG blends can be generated. Manipulation of the externally supplied VFA feed ratios, or modulation of VFA concentrations generated in vivo, could therefore offer simple routes to generate LPG blends in scaled production. ## Engineered metabolic pathways for bio-alkane gas production Upregulation of cellular VFA biosynthesis is an alternative means of biosynthesising alkane gas with engineered CvFAP biocatalysts. Ideally, the chemical precursor for a VFA biosynthetic pathway should be a major component of available waste feedstocks. Amino acids derived from protein-rich waste products are simple, cheap and readily available carbon sources. They are prevalent in salted milk whey from cheese manufacture, 33 brewery waste yeast 34 and proteinaceous food waste. 35 A pathway was constructed from valine to propane beginning with the deamination of valine to a-ketovalerate, catalysed by leucine 2-oxoglutarate transaminase (ilvE) from E. coli 36 (Fig. 3a). Irreversible decarboxylation of a-ketovalerate leads to isobutyraldehyde catalysed by branched-chain keto acid decarboxylase (KdcA) from Lactococcus lactis. 37 Isobutyric acid is then formed by the action of 3-hydroxypropionaldehyde dehydrogenase (Hpad) from E. coli, 38 which is subsequently decarboxylated by CvFAP G462I to form propane (Fig. 3a). An ADO/ferredoxin-dependent decarbonylation of isobutyraldehyde to propane 9 can provide a further 'dark' pathway to operate alongside the light-dependent pathway should this be required. This pathway was engineered in E. coli (Fig. 3b inset) and cultures were supplemented with valine (0-30 mg L 1 ). In each case, propane, isobutane and butane were detected in the headspace. Propane levels increased on feeding valine by up to 17-fold higher with 30 mg L 1 valine (109.7 AE 6.3 mg L 1 propane; Table S11; ESI †). (Iso)butane titres decreased with increased valine supplementation. These observations are likely due to an increase in (valine-derived) butyrate levels, leading to Effect of (a) CvFAP-pETM11 variant and (b) butyric:valeric acid blends with CvFAP G462V -pBbA1c on hydrocarbon production. Cultures (20 mL; 3 biological replicates) were grown in LB medium containing kanamycin (50 mg mL 1 ) at 37 1C to OD 600 B0.6-0.8. Recombinant protein expression was induced with IPTG (0.1 mM) followed by culture supplementation with fatty acid substrates (10 mM) after 1 h at 30 1C. Triplicate aliquots (1 mL) from each culture were sealed into 4 mL glass vials and incubated at 30 1C for 16-18 h at 200 rpm, illuminated with a blue LED panel. Headspace gas was analysed for hydrocarbon content using a Micro GC. Culture conditions and gas measurement were performed as described in the legend of Fig. 2, except the cultures were supplemented with valine (0-30 mg mL 1 ) 1 h after IPTG-induction, instead of VFAs. The numerical data for panel b can be found in Table S11 (ESI †). Error bars represent one standard deviation for triplicate biological repeats (n = 3). Differences in ratios between the difference gases likely reflects competitive binding of the three amino acids for a common CvFAP binding site. ## Energy & Environmental Science Paper the favouring of propane production over other alkane gases, similar to that seen in Fig. 2b. Endogenous valine and/or butyrate levels are likely high because propane yields were about three-fold higher than the externally fed valine concentration (Fig. 3b). The propane titres observed using this pathway are comparable to those obtained by external feeding of butyrate to CvFAP G462V alone (Table S8; ESI †). To operate with this pathway, waste feedstocks would need to be protein rich (e.g. food waste). Amino acid content of thirty-nine samples of vegetal and dairy product food waste from EU industrial agro-food systems has been shown to have a valine content that varies from 6.4-29.4 mg g 1 waste, 35 and also can act as general carbon sources. Therefore, these abundant wastes are attractive feedstocks for alkane gas production. A microbial chassis for scaled bio-alkane gas production E. coli was chosen to demonstrate laboratory scale production of bio-alkane gases using CvFAP variants as proof-of-concept. It may not be suitable, however, for scaled production. Microbial fuels and chemicals production is a costly process, and places high demands on both capital and operational expenditures. Typically, steel-based bioreactors with complex monitoring systems are used, with high running costs (e.g. energy-intensive aeration, mixing and downstream processing), production rates and titres. Sterilisation is required to minimise microbial contamination, and growth under aseptic conditions is necessary. There are also environmental concerns over waste processing and disposal, and production methods use large quantities of clean water. These multi-faceted issues can increase production costs. At the outset, by selecting Halomonas st. TD01 as a production host, we tackled many of these issues. 39 Halomonas grows at high salinity (e.g. 20% w/v NaCl) and at pH values as high as 12. Continuous cultures have been grown for over three years in industrial-scale vessels for the biomanufacture of polyhydroxyalkanoates at greater than 1000 tonnes scale, with no decline in growth potential. 15 Seawater, waste-water and recycled water can be used without sterilisation, conserving fresh water and reducing energy expenditure. Bioreactors can be constructed using low cost materials (e.g. plastics, ceramics and cement). Scaled production of polyhydroxyalkanoates using Halomonas is at a 65% cost saving compared to E. coli. 21 This suggests that distributed bio-LPG biomanufacture could be more profitable using Halomonas. We constructed a Halomonas-compatible plasmid pHal2 containing CvFAP using a broad host range pSEVA-derived plasmid 40 with an IPTG-inducible promoter (Fig. 4a inset). 16,41 Halophilic in vivo alkane gas production with both CvFAP G462V and CvFAP G462I variants was performed with butyric acid (Fig. 4a), with the highest titres shown for CvFAP G462V (157.1 AE 17.1 mg L 1 culture) with 80 mM butyrate in buffered medium (Fig. 4b and Table S12; ESI †). Variant CvFAP G462I produced more propane than CvFAP G462V , and both variants produced several-fold more propane than the wild-type (Fig. 4a). These titres are about ninefold greater than those reported for E. coli containing engineered ADO-dependent pathways, and five-fold greater than E. coli containing CvFAP G462V and fed with butyrate. 5,6,8 Propane production progressively increased with light 'intensity' (up to about 2000 mmol s 1 m 2 ; Fig. 4b inset) but declined at higher light intensities, most likely due to increased photoinactivation of CvFAP (Fig. S3; ESI †), or light-dependent effects on cell viability. Halomonas therefore proves to be well able to support the production of propane, showing titres comparable to valine supplemented E. coli cultures expressing the KdcA-dependent pathway (Fig. 3). For scaled-up production, engineered strains require waste biomass feedstock mixed with seawater and recycled water grown aerobically at high salinity with minerals, vitamins and VFAs. Wastewater streams (with salt supplementation) are suitable for production at inland sites. Autolysed spent brewery Cultures were grown in phosphate buffered YTN6 medium containing spectinomycin (50 mg mL 1 ) for 5 h at 37 1C and 180 rpm. Recombinant protein expression was induced with IPTG (0.1 mM; OD 600 B 1.6), and cultures were supplemented with butyric acid (0-100 mM, buffered at pH 6.6). Triplicate aliquots (1 mL) of cultures were sealed into 4 mL glass vials and incubated at 30 1C for 16-18 h at 200 rpm, illuminated with a blue LED panel. Headspace gas was analysed for hydrocarbon content using a Micro GC. Reactions were performed as biological repeats (technical repeats for the inset). Error bars represent one standard deviation for triplicate biological repeats (n = 3). Technical repeats were also performed in triplicate. ## Paper Energy & Environmental Science yeast or similar (e.g. hydrolysed 'spent' Halomonas cells) can be used to provide essential vitamins. Biodiesel waste (60-70% glycerine) is a cost-effective carbon source for bacterial growth. 42 VFAs can be sourced readily from anaerobic digestion (AD) (e.g. 50 g L 1 butyrate from fed-batch fermentation of brown algae; 43 36 g L 1 with kitchen waste 44 ). Bio-LPG production using waste feedstocks was investigated at laboratory scale in a flatbed photobioreactor. Non-sterile fermentations were performed using 'clean' (laboratory grade reagents) and 'crude' (seawater and waste glycerin) media. Seawater and biodiesel waste impurities affected Halomonas growth and propane production to only a minor extent (Fig. 5). Maximal propane production occurred 4-6 hours after induction, with an overall yield of B90 mg g 1 cells (over 2 days). A steady decline in propane production after the early peak rate was observed (Fig. 5b; Fig. S5; ESI †), attributed to plasmid instability and/or loss 45 or possibly CvFAP inactivation. In addition, the continuous exposure of Halomonas (or microorganisms in general) to blue light may impact on cell viability, compromising further replenishment of the CvFAP catalyst in vivo. In spite of this, we have shown that waste feedstocks and seawater are capable of supporting bio-alkane gas production. ## Stable strains for bio-LPG production Genome-integrated alkane gas producing Halomonas strains are required for scaled production to eliminate the need for antibiotic-mediated plasmid maintenance. We integrated an IPTG-inducible CvFAP G462I into the genome of Halomonas and tested its bio-alkane gas producing ability in a photobioreactor for 48 to up to B100 h. 16 Cumulative propane yields were less than half those obtained with plasmid borne pHal2-CvFAP G462V (Fig. 5c), in spite of the G462I variant displaying higher activity than CvFAP G462V (Fig. 4a). This lower propane titre is not surprising as only one copy of CvFAP G462I was integrated into the genome. However, these titres compare favourably to plasmid-based expression (Fig. 5c). Propane production rates were maximal around 60 h (Fig. S6; ESI †). Beyond 60 h, cell viability of Halomonas was shown to decline, as evidenced by a reduction in viable cells (colonies) detected during plate counting assays. This is likely due to prolonged high intensity light exposure. Continuous culture replenishment of the photobioreactor with fresh 'dark-grown' culture would mitigate against this loss of cell viability. Next, we eliminated the need for IPTG induction of CvFAP G462V using a modified constitutive pPorin-like promoter. 46 Constitutively-expressed CvFAP G462V cultures showed elevated (2.7-fold) propane yields (237 mg g 1 cells) compared with induced cultures (Fig. 5c; Fig. S6; ESI †), with a concomitant reduction in cell density attributed to slower cell growth. We also integrated a 'valine-to-propane' pathway (constitutive and inducible versions) into the Halomonas chromosome and demonstrated bio-alkane gas production. Higher overall alkane titres were obtained with the integrated IPTG-inducible pathway (1.19 AE 0.01 mg L 1 isobutane; Fig. 5d) compared to the constitutively expressed strain. Extensive industrial waste amino acid sources can be found in the dairy (salted milk whey) and brewery industries (autolysed yeast). ## (c) Cumulative propane production by Halomonas expressing plasmid-borne or chromosomally integrated CvFAP variants. Cultures (400 mL) were grown in high salt glycerol medium at pH 6.8 containing 50 mg mL 1 spectinomycin (plasmid borne cultures only) and 0.5 mL L 1 antifoam. Conditions were maintained at 30 1C with 65-100% stirring, an airflow rate of 1.21 L min 1 in the dark until mid-log phase (4-5 hours). Recombinant protein expression was induced with IPTG (0.1 mM) where required, followed by the addition of sodium butyrate (60 mM pH B 6.8) and blue light exposure (1656 mE). Cultures were maintained for 48-110 h and propane production was monitored at 15 to 20 minute intervals by automated headspace sampling using a Micro GC. In panel c, data for inducible and constitutively expressed plasmid-borne expression systems are indicated by grey and orange circles, respectively, while the chromosomally integrated expression system is indicated by green circles. R = lac repressor; Spec R = spectinomycin resistance gene. (d) Production of alkane gases in Halomonas using the valine-dependent pathway. The IPTG-inducible and constitutive promoters were pTrc and pPorin 69, respectively (Table S2; ESI †). General culture conditions (non-sterile) and gas measurements were performed in YTN6 media, as described in the legend of Fig. 4, containing autolysed brewery yeast without VFA addition. Milk whey medium (pH 9) was composed of cheesemaking residual salt whey from a North of England supplier supplemented with 60 g L 1 NaCl. Error bars represent one standard deviation for biological repeats (n = 3). ## Energy & Environmental Science Paper We compared the bio-alkane gas production of the plasmidborne KcdA-dependent pathway using these waste supplies and found in all cases, bio-alkane gases are produced (Fig. 5d). The titres and compositions of the produced gas were dependent on feedstocks, which reflects on the relative amino acid and/or overall nutritional compositions of each feedstock. ## Bio-LPG from carbon dioxide An ideal energy strategy would directly utilise CO 2 as the carbon source for the production of biofuels. A microbial carbon capture solution could take advantage of the photosynthetic ability of cyanobacteria to fix CO 2 into organic carbon. Synechcocystis PCC 6803 is an ideal host because it grows rapidly and is genetically tractable. It is tolerant to abiotic stress and growth requirements are well understood. 47,48 Conversion of CO 2 into medium chain-length fatty acids 13 and long chain hydrocarbons 14 has been described. We previously bioengineered Synechcocystis PCC 6803 by incorporating E. coli thioesterase A, which catalyses the conversion of fatty acyl-ACP to free fatty acids. We also knocked out the native fatty acyl ACP synthase gene (Daas) to minimise the reverse reaction (Fig. 6a). 14 These changes increased the availability of free fatty acid precursors in vivo, enabling hydrocarbon biosynthesis direct from CO 2 , instead of via an external carbon source. 14 In this work, we incorporated the gene for CvFAP G462V into wild-type Synechocystis and a Daas gene knockout strain in the presence or absence of Tes4 (a butyryl-ACP thioesterase from Bacteroides fragilis) 9 under the control of a cobalt-inducible promoter (Pcoa) or a constitutive promoter (Ptrc) (Fig. 6a and Fig. S7; ESI †). CvFAP G462V was chosen for the experiments in Synechocystis due to the reproducibility of its high expression levels (Fig. S2; ESI †). Under batch culture conditions, neither the wild-type nor the Tes4/Daas strains produced detectable propane. Only low levels of propane (11-14 mg per L culture per d) were produced when using strains carrying the CvFAP G462V gene that were supplied externally with butyrate (Fig. S7; ESI †). Encouraged by these findings, we then cultivated the Synechocystis Tes4/Daas strain carrying CvFAP G462V in the photobioreactor under photosynthetic conditions (see Experimental section) with supplementary blue light exposure. This strain showed moderate propane production (11.1 AE 2.4 mg propane per L culture per day), which is equivalent to B12.2 AE 2.6 mg propane per g cells per day (Fig. 6b and Fig. S7; ESI †). To the best of our knowledge, this is the first demonstration of the direct conversion of CO 2 into propane. A factor to consider when using Synechocystis for propane production is the photobleaching of photosynthetic pigments in the presence of high intensity blue light. 49 This fixes a practical upper limit on blue light intensity to maximise CvFAP activity whilst minimizing the extent of photobleaching (Fig. S8 and S3, respectively; ESI †). Photobleaching was apparent when light exposure was maintained between 500 to 800 mE, but not during prolonged exposure (140 h) between 300 to 500 mE (Fig. S8; ESI †). A complementary strategy is to feed bio-alkane-producing Halomonas with osmotically lysed Synechocystis; the latter acting both as a carbon (growth) and butyrate feedstocks. 50 This bypasses photobleaching effects, but retains the ability to capture CO 2 . Synechocystis can also be degraded by AD to generate VFAs for Halomonas gas production. To test the feasibility of this approach, we fed batch cultures (1 mL aliquots) of Halomonas expressing CvFAP with lysed Synechocystis as supplementary carbon and butyrate sources, and produced propane gas (25.3 AE 5.8 mg L 1 culture; Fig. 6b). These titres were enhanced compared to control cultures reliant on only endogenous butyrate alone (YTN6 medium; 0.9 AE 0.1 mg L 1 culture). Use of Synechocystis as a feedstock could enable Halomonas to produce bio-LPG from industrial or atmospheric CO 2 rather than being reliant on waste organic matter. The International Energy Agency has estimated that carbon capture and storage/ utilisation (CCS) could potentially contribute to a 19% reduction in CO 2 emissions by 2050 using existing technologies. 51,52 Fig. 6 Propane production employing natural photosynthetic CO 2 capture. (a) Scheme for engineering Synechocystis to enable propane production by CO 2 fixation. (b) Cumulative propane production of Synechocystis Daas expressing CvFAP G462V and Tes4 (strain pIY819). The photobioreactor (400 mL) was set up in batch mode with starter culture diluted 3 : 1 in fresh BG11 + medium (BG11 pH 8.0 13,14 containing TES buffer and 1 g L 1 sodium thiosulphate) in the presence of 150 mM NaHCO 3 . Both pH control and CO 2 supply were maintained using 1 M NaHCO 3 in 2 BG11 + . The culture was maintained at 30 1C with maximal stirring with an airflow rate of 1.21 L min 1 , illumination of warm white light (30 mE), automated pH maintenance (1 M acetic acid in 2 BG11 + ) and optical density monitoring (680 nm). After reaching an optical density of B0.3 (B20 h), cobalt(II) nitrate hexahydrate (150 mM) was added, warm white illumination was increased to 60 mE and the integral actinic blue LED light panel was activated to provide 500-750 mE blue light (460-480 nm). The culture was maintained at 30 1C for 18-48 hours, with manual headspace sampling to quantify propane by Micro GC. A coupled Synechocystis-Halomonas process could further enhance management of industrial CO 2 emissions as well as generate bio-LPG to meet energy demands. ## Vision and economics of bio-LPG production Multiple designs of scaled bio-LPG production hubs have been envisaged, which differ in waste feedstock supply, bioreactor design (B1000-10 000 L scale) and light requirements. A prototype design based on Halomonas cultivation, could be located in a coastal region with on-site seawater, an anaerobic digester (AD) for VFA production and optionally a cyanobacteria photobioreactor for CO 2 fixation and VFA supply (Table S11; ESI †). The AD plant would be tuned to generate a specific VFA blend by modulating the waste composition (e.g. oil and salt concentrations), microbial consortium and running conditions (e.g. temperature). The site could also contain multiple photobioreactors for bio-LPG production (Fig. 7) fed from a dark bioreactor for Halomonas propagation prior to bio-alkane gas production. These photobioreactors could be classical flat-bed photobioreactors, or even low-cost pressurised polyethylene bags with external illumination. 53 Propane could be harvested using gas-scrubbing methodologies, 54 linked to existing desiccant moisture removal and liquefaction technology. On-site generated bio-LPG could be used to feed adjacent heavy industry, or be transported using local distribution infrastructures. For on-site usage of the alkane gas blends, the exact gas composition (e.g. propane vs. butane) does not need to conform to local government requirements if it is not sold under the 'LPG' label. This option allows the usage of local organic waste that may not generate a specific ratio of VFAs, or if the waste composition is likely to vary considerably. Also, small distributed plants can utilise local power generated by solar, wind or tidal technology to power LED illumination, considerably reducing operating costs. The selection of suitable sites for bio-alkane gas production will be dictated by the need to use locally supplied waste feedstocks and seawater. The current global price for non-AD butyrate is around d2-3 per kg, which is not cost-effective for bio-alkane gas production. Local food waste could be used as both a carbon source and amino acid supply (KdcA pathway utilization) in place of AD-generated VFA blends. For example, the UK generates approximately 7 million tonnes of household food waste annually, of which 0.6-0.7 million tonnes can be collected by local authorities and treated through waste management systems. 55 Based on laboratory scale studies, food waste has a great potential as a VFA fermentation substrate due to high VFA yield (up to 0.43 gVFA per g substrate). 30 The feasibility of this approach was investigated by performing a preliminary Techno-Economic Analysis (TEA) and carbon footprint analysis, limited to the process itself, of a prototype design for Halomonas bio-alkane gas production (Note S1, including Fig. S9-S11 and Tables S13-S26; ESI †). This analysis is intended to understand the gap between the early-stage research and commercial realisation, and to illustrate potential strategies to bring further process improvements. The TEA contains a high degree of uncertainty because the process is still at a low technology readiness level. Nevertheless, the analysis results can be used to highlight bottlenecks and hotspots which have significant impact on the economic or environmental aspect of the process, so that future research and process design can be directed in the most effective direction. The base case was modelled on a conceptual design of a pilot-scale continuous process with one 1 m 3 bioreactor for biomass synthesis, two 1 m 3 photobioreactors in series for propane synthesis, and one 1.72 m 3 anaerobic digester to generate butyrate feed. Ten further TEA cases were formulated by introducing additional measures and strategies to strengthen the economic potential of the process (Note S1; ESI †). Examples of cost reducing strategies include implementing non-sterile fermentation, valorization of sidestreams to produce additional valuable chemicals, optimization of cell productivity and process scale-up. The design basis, process specifications and engineering as well as financial assumptions are summarized in Note S1 (ESI †). Based on this information, the TEA model generated estimates for plant performance, production costs, minimum propane selling prices (MPSP) and CO 2 emissions for all the cases. In comparison to chemical routes, CvFAP catalyst 'poisoning' by photoinactivation is overcome by continuous replenishment from a 'dark' bioreactor (biomass production in the dark). Reduction in energy costs associated with light supply (CvFAP photoactivation; cyanobacterial growth) is central to production costs savings. This could be managed by using solar energy or wind farm electricity. Substituting blue LEDs for concentrated wavelength-filtered sunlight (e.g. 425-475 nm) could reduce the energy burden and associated costs. The daily blue light intensity in the Northern United Kingdom 56 averages around 29.0 W m 2 , while the required photobioreactor intensity is up to 424 W m 2 , dependent on microbial chassis. A 15-fold solar concentration is required to generate the required blue light intensities which could be met using existing solar concentration technologies, similar to the Australian National University parabolic trough or the Entech Incorporated Fresnel lens (each achieving 30-fold concentration). 57 To allow diurnal propane production, light supplementation via LEDs could be supplied outside daylight hours. Alternatively, Halomonas could be engineered to include an alternative non-light requiring ADO-dependent pathway from valine (Fig. 3a), enabling propane production during the dark phase. A cost-effective solution to bio-alkane production requires a significant reduction in illumination costs. This was modelled by the utilisation of natural sunlight with solar concentrators, cleaner wind power and the localisation of the bioreactors within developing countries with lower operating costs (Case 6, 10 and 11; Tables S16, S23-S24; ESI †) In addition, process scalingup and the generation of secondary revenue are necessary to increase the cost-effectiveness of the process (Case 5 & 9). This could include the conversion of waste Halomonas and Synechocystis biomass to fertiliser, further processing through AD plants as a source of VFAs, or desiccated to produce fish food at larger scale. As about 25% of the carbon content of supplied butyrate would be lost as CO 2 , the energy ratio of propane production (i.e. energy output/energy input) at maximal productivity is estimated to be 3.18. (Case 11; Tables S17, S23-S24; ESI †). The projected propane yields were estimated at 358 tonnes per year for a pilot production system scaled-up by 10-fold, with projected combined revenue of d3.1 M (primary and secondary products; Table S24; ESI †). This is based on developing a Halomonas strain with multi-copy insertion of CvFAP G462I to ensure stable propane titres similar to plasmid-borne systems. The TEA study also predicted a 300-fold decrease in propane production cost at scale (US $626.80 kg 1 to US $1.89 kg 1 ) in comparison to traditional, unoptimised biotechnological approaches. The UK LPG market is ca. 82 970 barrels per day, 55 equivalent to B0.086 tonnes of propane per barrel or 2 540 000 tonnes per year. If the future transport sector were to use ca. 10% of the current market for bio-LPG as a drop-in fuel this would require around 710 of the said pilot process operating at maximal productivity and consume 225 tonne of crude glycerol per year. After four decades of technology development, ethanol derived from starch or cellulosic biomass is currently the dominant biofuel for liquid transportation and power generation. Bioethanol can now be made economically and at large scales sufficient to contribute to a nation's fuel market; this is not yet feasible for other newly emerged biofuels e.g. butanol (non-ABE derived), biofene (farnesene), and bisabolene. Ethanol biorefineries have the capacity to utilise biomass feedstock, transforming most of the components into valuable products, which are integrated readily with existing industrial infrastructures. These features are desirable also for bio-alkane gas production processes, if they are to meet the ambitious goal of utilising propane/alkane blends as drop-in-fuels. Also, fuelling vehicles with Bio-LPG is one way to diversify the availability of clean fuels and to increase energy security so that economies are not over reliant on provision of ethanol. Bio-LPG has its own niche in the transportation fuel sector -for example, it is suitable for high-mileage vehicles by offering improved engine life and lower maintenance costs. As with ethanol production, Bio-LPG production will require scaled technology development and optimisation, and more detailed TEA evaluation as the technologies mature. The above example illustrates how an integrated biorefinery strategy could in principle be used to supply local energy requirements and generate income, while recycling industrial CO 2 and food waste. A second strategy utilizing the multi-step pathway from amino acids could be employed, using food, brewery or dairy waste to supply the necessary amino acids (in place of AD-generated butyrate). There are likely other configurations around these examples that could likewise be deployed at scale, enabling bio-LPG manufacture from waste biomass and atmospheric/industrial CO 2 , and provide renewable energy solutions for localised economies around the globe. That said, further exploration of the TEAs will be required coupled to further rounds of microbial strain optimization, as individual bio-production formats are scaled at higher technology readiness levels. ## Conclusions The development of any synthetic biological solution for chemicals production into a commercially viable process requires the consideration of both (bio)catalytic process optimisation and an understanding of the techno-economic challenges of developing scaled bioprocesses. We tackled both of these challenges when developing a series of biocatalytic solutions to alkane gas production. We investigated (i) single vs. multi-catalytic pathways from abundant waste feedstocks, (ii) multiple chassis screening and development, (iii) lab-scale production in vivo and (iv) finally proposed designs for multiple scaled bio-LPG production 'hubs', utilising local waste materials and taking advantage of the available infrastructure. This combinatorial approach is key in any commercial development as it focuses the research and development beyond simple proof-of-principle demonstration, and directs progression towards practical solutions to process/ economic bottlenecks. We have shown that sustainable and renewable solutions to highly efficient and clean-burning bio-alkane fuels (tuneable ## Paper Energy & Environmental Science bio-LPG) is possible. Further optimisation of the microbial chassis to achieve stable high titres, coupled to improvements in the bio-LPG bioprocess hub design will drive the concept towards the realisation of a commercially-viable process. Localisation of these hubs close to existing waste-generating heavy industry in both advanced and developing countries will assist with waste management, reduce the carbon footprint, and increase energy security. This could positively contribute towards global carbon management targets and clean air directives. ## Conflicts of interest A patent application (PCT/EP2019/060013) entitled 'Hydrocarbon production' is pending in relation to the production of hydrocarbon gases in engineered microbial strains. P. R. J. is a board member, and M. S. and N. S. S. are founding directors of C3 Biotechnologies Ltd.

chemsum

{"title": "Low carbon strategies for sustainable bio-alkane gas production and renewable energy", "journal": "Royal Society of Chemistry (RSC)"}

deciphering_the_mechanism_of_o₂_reduction_with_electronically_tunable_non-heme_iron_enzym

## Abstract: A homologous series of electronically tuned 2,2 0 ,2 00 -nitrilotris(N-arylacetamide) pre-ligands (H 3 L R ) were prepared (R ¼ NO 2 , CN, CF 3 , F, Cl, Br, Et, Me, H, OMe, NMe 2 ) and some of their corresponding Fe and Zn species synthesized. The iron complexes react rapidly with O 2 , the final products of which are diferric mu-oxo bridged species. The crystal structure of the oxidized product obtained from DMA solutions contain a structural motif found in some diiron proteins. The mechanism of iron mediated O 2 reduction was explored to the extent that allowed us to construct an empirically consistent rate law. A Hammett plot was constructed that enabled insightful information into the rate-determining step and hence allows for a differentiation between two kinetically equivalent O 2 reduction mechanisms. ## Introduction Molecular oxygen (O 2 ) dependent iron oxygenases are important in a variety of life processes such as respiration and drug metabolism. Therefore, a fundamental grasp of the elementary steps involved is of great signifcance. However, the diverse 1,2 primary and secondary coordination sphere of the enzyme active sites that cause different selectivity 3,4 and observed reactive intermediates 5,6 make a general understanding of the mechanism a complicated matter. The initial step in a mechanism involving O 2 produces a formally Fe(III)-superoxide species via an inner or outer sphere electron transfer mechanism. These two limiting cases are difficult to distinguish. 7,8 Studies that might enable differentiation by testing specifc hypotheses in enzymatic O 2 activation require systematic variations of a metalloprotein active site. However, a major challenge to this approach is the inherent difficulty associated with changes to an active site by means of site-directed mutagenesis, not to mention loss of activity that may result from such alterations. The systematic electronic and steric tuning of synthetic enzyme models offers a potential solution to this dilemma. One of the most powerful techniques used to investigate mechanism that takes advantage of systematic changes is the linear free energy relationship in the form of a Hammett plot. 12 While linear free energy relationships have been used to great success in understanding O 2 activation, 10,13 the specifc use of the Hammett plot in inorganic and organometallic reactions is not as common and, to our knowledge, only a few reports have demonstrated the utility of the Hammett plot in O 2 reduction by synthetic non-heme iron complexes. 14,15 The frst step in O 2 activation at non-heme centers, namely the two limiting cases of inner vs. outer sphere reduction of O 2 , has not been thoroughly addressed when compared to heme analogues that have been extensively studied. 7, In fact, the discussion about O 2 binding and reduction in non-heme centers is predominantly described as an inner sphere process. 9,19 While an inner sphere reduction to form Fe III -superoxo species is reasonable and probably true in many cases, the alternative outer sphere description is equally plausible. To this end, we report a systematically varied series of Narylacetamide ligands that contain remote substituents for electronic tuning of metal-ligand bonding for the purpose of using a Hammett plot to gain insight into the rate-limiting step of O 2 reduction. Herein we report the synthesis and characterization of these new ligands in addition to the biologically relevant Fe and Zn metal complexes. Finally, the iron complexes react with molecular oxygen and the mechanism of this reaction was deciphered with the aid of a Hammett plot. To our knowledge, this study serves as the frst kinetic analysis that specifcally attempts to address the outer vs. inner sphere hypothesis in non-heme iron enzyme model complexes. ## Synthesis of ligands The new ligands described in this report are based after the trisacetamide ligands frst used by Borovik and coworkers. 20 They prepared a variety of aliphatic and aryl acetamide ligands and used them in coordination chemistry studies with frst-row transition metal complexes including O 2 activation and stabilization of unusual electronic 24,25 and coordination environments. 20 We adopted the synthetic strategy for the known 2,2 0 ,2 00 -nitrilotris(N-(3,5-dimethylphenyl)acetamide) compound (H 3 L dmp ) 25 to prepare the new ligands, which involves heating a solution of nitrilotriacetic acid in pyridine and triphenylphosphite with the appropriate aniline (Scheme 1). Eleven ligands H 3 L R (R ¼ NO 2 , CN, CF 3 , F, Cl, Br, Et, Me, H, OMe, NMe 2 ) were thus obtained in good yield and high purity. A plot of the 1 H-NMR acetamide NH resonance vs. the Hammett parameters reveals a linear correlation (Fig. S1 †) confrming electronic communication between the substituents and the arylacetamide nitrogen atom 26,27 that will serve as the donor to a transition metal ion. ## Metal complex synthesis and characterization The ligands can be deprotonated in dimethylacetamide (DMA) solvent with three equiv. of KH to afford the respective ligand salt. These are then treated with M(OAc S1 †). When M ¼ Fe, the acetato ligand adopts a bidentate orientation for R ¼ NO 2 (1 NO 2 ) rather than the monodentate mode in the R ¼ H molecules (1 H ), possibly indicating a more electron defcient metal. The acetate ligation in the respective Zn complexes (R ¼ H, NO 2 ) is similar to the iron complexes except that the long Zn-O in [ZnL NO2 (OAc)] is about 0.4 longer. The solution state-structure of the two Zn complexes (R ¼ H and NO 2 ) was probed by 1 H-NMR spectroscopy. With the exception of a broadened peak for the acetato ligand, which might indicate fluxional ligation or exchange, the peaks are sharp and reveal a C 3 symmetric coordinationgeometry on the NMR time scale (Fig. 2 and S2 †). Hence, we assume that the solution-state structure of the Fe(II) ions is somewhat similar to the Zn complexes. 28,29 The iron salt [Me 4 -N] 2 [Fe(II)L NO2 (OAc)] (1 NO2 ) was characterized by Mössbauer spectroscopy in the solid state and has parameters consistent with an S ¼ 2 species. This is in agreement with the room temperature solution NMR Evans' method magnetic moment of m eff ¼ 4.91 m B . ## Bulk oxidation of [FeL NO 2 (OAc)] 2À with O 2 The iron complexes react rapidly with molecular oxygen (pure O 2 or in air) forming a red compound (2 R ). Preparative scale reactions performed using O 2 and 1 NO 2 in DMA or MeCN resulted in good isolated yield ($88%) of [Me , respectively (Scheme 2). In contrast to previous studies using O 2 on similar platforms, 23 the fnal products isolated herein are Scheme 1 Synthesis of H 3 L R and metal complexes. dimeric m-oxido complexes rather than mononuclear complexes with terminal hydroxido ligands. The lack of steric protection is the probable cause for this difference since the bulky aliphatic-acetamide ligand L iPr stabilizes the terminal hydroxido K[FeL iPr (OH)]. 23,25 Another major difference between 2a/b and monomeric ferric hydroxido complexes with similar ligands is that one of the ligand arms in 2a/b has an altered binding mode, having undergone tautomerization. As such, each iron center in 2a/b contains one anionic oxygen donor from one of the acetimidate moieties and two anionic nitrogen donors binding in the usual fashion from the other two acetamidate arms. The assignment of the charges on the donor groups is supported by the number of counter ions in the unit cell and the substantial differences in the C-N bond lengths (Table 1). For example, the {K[L dmp FeNO]} 7 complex has C-N bond distances of 1.34 , which is comparable to 1 NO2 and 1 H and consistent with the acetamidate ligation. Similarly, two of the C-N bonds in each of the crystallographically related halves of the 2a and 2b molecules are 1.34 and 1.35 , respectively, further indicating acetamidate ligation. The remaining C-N bond distance in 2a and 2b is shorter, 1.30 , and supports the assignment of acetimidate ligation. A key difference between the 2a and 2b is the coordination number at the iron centers (Fig. 3). 2b attains a fve coordinate geometry with loss of local three-fold symmetry in the primary coordination sphere due to the tautomerization and hence an asymmetrical binding mode. In contrast, the iron centers in 2a are six-coordinate due to the additional ligation of the bridging acetate ligand. The synthetic procedure for 2a and 2b differ only in the solvent used (DMA and MeCN, respectively). Thus, the formation of two similar dimeric m-oxido complexes is likely caused by the greater extent to which DMA can stabilize the trianionic 2a during crystallization. Interestingly, the structure of 2a resembles carboxylate-oxobridged diiron enzymes that are important in a number of biological transformations that use molecular oxygen. 30,31 It is well established that the bridging ligands strongly influence the magnetic properties of these active sites and influence chemistry. Similarly here, the binding of acetate to the diferric core appears to influence the magnetic properties of the complex. For example, 2b has a magnetic moment of 2.26 m B (DMSO, room temperature) that is similar to other m-oxido differic complexes and 2a has a higher magnetic moment of 3.01 m B . 32 However, little can be said about these differences because the solution speciation of 2a appears to be complicated. For instance, the UV-vis spectra of 2a and 2b are essentially identical in DMA and suggest that the binding is minimal in solution. In fact, treatment of a solution of 2b in DMA with 0.10, 1.0, 10 and 30 equivalents of [Me 4 N][OAc] causes a shift in the UV-vis spectrum to lower energy to occur with no isosbestic point implicating multiple binding modes or complicated equilibria (Fig. S6 †). The formation of the mu-oxo species 2 NO2 likely forms from the condensation of O 2 derived {Fe(III)OH} n (n ¼ 1 or 2) species. 33 To test this premise, we quantifed the water produced in the reaction between 1 NO 2 and O 2 in bulk oxidations using 19 F-NMR spectroscopy and the water sensitive reagent iodosobenzene difluoride (PhIF 2 ). 34 Specifcally, the volatiles from a solution of freshly prepared 2b were transferred to a clean, dry flask via trap-to-trap vacuum distillation on a high-vacuum line. The distillate was transferred into a glove box and treated with freshly prepared PhIF 2 in solvent dried with newly activated alumina and the solution contents were analyzed by 19 F-NMR spectroscopy. Any water in solution reacts with PhIF 2 to form the [FHF] anion, which can be quantifed using BF 4 internal standard. It was found that about 0.5 equivalent H 2 O formed per molecule of 1 NO2 used (three runs, 54%, 37%, and 30% yield H 2 O based on iron). Hence, it is reasonable to assume that "Fe(III)OH" moieties form in the reaction, likely through a C-H bond cleavage reaction. To further test this hypothesis, we included 10 equiv. of dihydroanthracene (DHA) in a bulk oxidation reaction, but we did not observe anthracene as a product. A likely reactive intermediate in the oxidation of 1 R is a superoxo species with an accessible active site that is exposed to free solvent; for such an intermediate, DHA may not be able to compete kinetically with solvent molecules in a bimolecular reaction. Recently, it has been shown that enzymatic and synthetic iron-superoxo species are competent for such transformations, 6,19, but other intermediate species (e.g., oxo) are also possible. ## Mechanism of iron mediated O 2 reduction The kinetics of the reactions of [FeL H (OAc)] 2 (1 H ), [FeL Cl (OAc)] 2 (1 Cl ), and 1 NO 2 with O 2 in DMA were investigated with UV-vis spectroscopy. The reactions with O 2 and 1 R in DMA were carried out in Schlenk UV-vis cuvettes that were degassed and equilibrated at 20 C prior to exposure to 0.75 atm of pure, dry O 2 . The dissolution of O 2 initially causes complication in the kinetic analysis and has been described before as prohibitive to mechanistic studies. 38 However, we have conducted a mass transfer analysis that accounts for this complication and is further enabled by the fact that O 2 saturation occurs early enough that we can determine frst order rate constants (see ESI †). A representative plot of spectra obtained by treatment of 1 H in DMA with O 2 is shown in Fig. 4. In the case of 1 NO2 , the UVvis spectrum of the fnal species (designated 2 NO2 ) in low concentration experiments is almost identical to 2b with added acetate in solution (Fig. S6 †). Specifcally, the l max of the fnal product is shifted by 10 nm from 2b and has a lower extinction coefficient. 32 Considering the complicated equilibrium between 2b and [Me 4 N][OAc], we propose that the fnal product generated in UV-vis cuvettes is an isomer of 2a that converts into 2a upon crystallization at higher concentrations. To avoid complications from incomplete knowledge about the speciation of 2 R , we performed our kinetic analysis by following the consumption of 1 R by method of extent of reaction. For the three complexes 1 H , 1 Cl , and 1 NO 2 it was determined using log-log plots and flooding methods that the reaction is frst order in iron and has a complicated dependence on acetate and O 2 (Fig. S8-S11 †). Taken together, we propose the following mechanism (Scheme 3): (1) reversible acetate dissociation is followed by (2) a rate limiting O 2 binding step (step 2-III) or outer sphere electron transfer (step 2-I and 2-II); (3-4) reduction of O 2 is then followed by several fast steps to form an iron(III) compound 2 R . A steady-state approximation of the proposed mechanism with the mono anionic [Fe(II)L R ] serving as the intermediate gives a single term rate law of the following form (see ESI † for derivation): This rate law simplifes further to k obs [FeL R (OAc)] (eqn S1-S5 †). Following consumption of 1 R as a function of time provides frst order plots with a k obs ¼ 0.017 s 1 AE 0.004 for R ¼ H (Table S2 †) that is effectively independent of [FeL R (OAc)]. To further test the rate law, we kept iron concentration constant and varied the concentration of O 2 in the presence of additional acetate (20, 30, and 40 equiv. [Me 4 N][OAc]) and plotted 1/k obs against 1/[O 2 ] (Fig. S10 †). The plots with different acetate concentration each furnish a horizontal region with a y intercept The values of k 2 , K eq , and k 1 are 0.39 M 1 s 1 , 0.08, and 0.27 M 1 s 1 , respectively, were obtained through solving a system of equations (eqn S11-S14 †). These values are an estimate that is accurate to the order or magnitude presented due to the tolerance set in the MATLAB code we used to solve the system of equations. Taking advantage of the fact that the rate law can be approximated by k 1 [FeL R (OAc)] in the absence of free acetate (k 1 z k obs ), the rate constant was measured over the temperature range from 10 to 70 C. Unfortunately, the Eyring plot (see Table S3 †) contains a large degree of scatter because the rate has a negligible dependence on temperature; k obs has a value of 0.023 s 1 AE 0.004 from the range of 10 to 70 C (Table S3 †). There also appears to be an inflection point near 10 C, but the large scatter makes this Eyring plot difficult to interpret and possibly not informative outside the context of other similar studies. Iron centers that bind O 2 often have small enthalpy of activation, reflecting the fact that O 2 is a poor ligand. 38,39 Our lack of a clear relationship between temperature and rate may also reflect a small entropic contribution. This is only speculative however and there might be other factors such as competing pathways with relatively similar barriers. For instance, Busch observed complicated parabolic dependence of the rate constant with temperature for O 2 with myoglobin, hemoglobin, and cyclidene complexes. 40 Busch's interpretation of the temperature dependence relied on competing inner and outer sphere O 2 reduction pathways in addition to competitive ligand bindingall of which are possible in the system studied here. These two possible O 2 reduction pathways, one involving rate limiting inner sphere O 2 binding and reduction (step 2-III in Scheme 3) and the other rate limiting outer sphere electron transfer followed by rapid superoxide coordination (step 2-I and 2-II, respectively, in Scheme 3), provide the same rate law and are difficult to distinguish. Herein lies the advantage of the Hammett plot to decipher reaction mechanisms. The frst order rate constants for fve of the 1 R complexes were plotted against the Hammett parameter (s) and from this plot a negative slope was obtained (Fig. 5 and S7 †). An even better ft was obtained when we used the Swain-Lupton correlation that takes into account both inductive and resonance effects. 41 The negative slope in these plots indicates positive charge build up in the transition state and is expected for an outer sphere electron transfer event. An alternative interpretation is positive charge buildup arises from loss of acetate ligand. However, the ratedetermining step is not acetate loss and so we surmise that the data are most consistent with a rate-determining outer sphere electron transfer event. Scheme 3 Proposed O 2 reduction mechanism with FeL R . To put our work in context, Sun and coworkers have studied the O 2 reduction dioxygenase model reaction with a sixcoordinate non-heme iron complex. 15 Assuming that O 2 reduction is rate limiting in their reaction, the negative slope in their Hammett plot also indicates that an outer sphere mechanism is operative. This is expected for a six-coordinate iron species. However, Que and coworkers reported a Hammett plot with a positive slope indicating a nucleophilic mechanism for O 2 reduction (inner sphere). 14 It should be noted that our investigation and Sun's were conducted in DMA and DMF, respectively, whereas Que's investigation was carried out in MeCN. We also briefly investigated the O 2 reduction in MeCN and, similarly to Que, constructed a Hammett plot with a positive slope (Fig. S7c †). This positive slope in MeCN suggests an inner sphere mechanism. Hence, the frst step in O 2 reduction mechanisms appears to have signifcant solvent dependence. ## Conclusions In summary, we have synthesized eleven new ligands and coordinated them to a variety of frst-row transition metals including biologically relevant iron and zinc. The iron compounds react with O 2 , and we determined the identity of the oxidized iron products for the R ¼ NO 2 variant. These products (2 NO2 ) are oxido bridged diferric complexes with unusual acetimidate binding modes resulting from one of the ligand arms of L NO 2 having undergone tautomerization. This binding mode has not been observed prior to our work for these acetamidate ligand platforms. Furthermore, the iron centres in 2a are bridged by an acetato ligand and are six-coordinate. The trisacetamidate ligand platform usually enforces three-fold symmetry that results in the formation of trigonal bipyramidal M III ions. Hence, the observation of the new sixcoordinate binding mode in 2a serves as precedent for hexa coordination of intermediates that might form in water or O 2 activation reactions. The differic molecules 2a and 2b also appear to exhibit a fluxional binding of acetate that, in a future study, may provide insight into how acetate binds to differic protein active sites. The mechanism of the formation of 2 R was determined to proceed through a rate limiting reduction of O 2 with a rate constant of k 2 z 0.4 M 1 s 1 . This reduction process was determined to follow acetate dissociation from 1 R with an equilibrium constant of 0.08. The dependence of the rate on temperature was minimal, so the Eyring plot was of little value, suggesting that both enthalpy and entropy of activation are close to zero consistent with other O 2 binding activation parameters. The use of the Hammett plot revealed a negative slope that is consistent with an outer sphere reduction of O 2 in DMA. The nature of the O 2 reduction step in non-heme iron metalloprotein active sites is a fundamental elementary step in O 2 activation mechanisms. Thus, this mechanistic investigationmade possible by a series of electronically tuned ligand-metal complexesserves as an important step in answering questions regarding O 2 activation with non-heme iron centres. Namely, what is the nature of the frst step in O 2 binding in irreversible O 2 reduction mechanisms? Is the frst elementary step that involves O 2 an outer sphere reduction of O 2 , or is it a binding event that is inner sphere electron transfer in nature? The question has been explored extensively for heme centres, but the situation is rather unclear for non-heme metalloenzyme O 2 dependent active sites. Our study serves as the frst kinetic analysis that specifcally attempts to address the outer vs. inner sphere hypothesis in non-heme iron enzyme model complexes. The data indicates that outer sphere reduction is the frst step in DMA, but solvent and probably counterion play a role in changing the mechanism and require further exploration of this challenging problem.

chemsum

{"title": "Deciphering the mechanism of O₂ reduction with electronically tunable non-heme iron enzyme model complexes", "journal": "Royal Society of Chemistry (RSC)"}

rapid_synthesis_of_monodisperse_au_nanospheres_through_a_laser_irradiation_-induced_shape_conversion

## Abstract: We develop a facile and effective strategy to prepare monodispersed Au spherical nanoparticles by two steps. Large-scale monocrystalline Au nanooctahedra with uniform size were synthesized by a polyol-route and subsequently Au nanoparticles were transformed from octahedron to spherical shape in a liquid under ambient atmosphere by non-focused laser irradiation in very short time. High monodipersed, ultra-smooth gold nanospheres can be obtained by simply optimizing the laser fluence and irradiation time. Photothermal melting-evaporation model was employed to get a better understanding of the morphology transformation for the system of nanosecond pulsed-laser excitation. These Au nanoparticles were fabricated into periodic monolayer arrays by self-assembly utilizing their high monodispersity and perfect spherical shape. Importantly, such Au nanospheres arrays demonstrated very good SERS enhancement related to their periodic structure due to existence of many SERS hot spots between neighboring Au nanospheres caused by the electromagnetic coupling in an array. These gold nanospheres and their self-assembled arrays possess distinct physical and chemical properties. It will make them as an excellent and promising candidate for applying in sensing and spectroscopic enhancement, catalysis, energy, and biology. Gold nanoparticles have attracted much interest because of their tunable Surface Plasmon Resonance (SPR) peaks, leading to important applications in the fields of chemistry, biology and materials sciences [1][2][3][4][5][6][7] . As is well known, the SPR properties of metallic nanoparticles (NPs) can be calculated by solving Maxwell's equations 8 in terms of theory, and exact solutions to Maxwell's equations are known only for spheres, concentric spherical shells, spheroids, and infinite cylinders 9 . Specifically, Mie theory is the exact analytical solution of Maxwell's equations for a nanoparticle with defined shape 10 . However, it is difficult to synthesize monodispersed noble metallic nanoparticles in a wide range of sizes and controlled shapes on a large scale, and their observed optical absorption property can not obey Mie theory prediction accurately.In recent years, gold spherical nanoparticles have attracted much attention in fundamental research 11,12 , such as Fano resonance 13,14 . Since such spherical nanostructures will give precise SPR spectra and can exclude the undesirable SPR signals originating from the particle corner, facets and size distribution in contrast with the arbitrary shape 10,15 . Hence, a gold perfectly spherical nanoparticle is an ideal model to verify the optical property predicted by Mie theory under light irradiation.More recently, focus has also turned to cellular and medical applications for such gold nanoparticles 2,16 . Besides the effects of size and surface functional group of gold nanoparticles, a particle shape is another important influence for delivering it into the cell 17 . For instance, Chithrani et al. showed that gold nanoparticles with spherical shapes took shorter wrapping time to wrap the entire bulk in comparison to the nanorod because of the decrease in the surface area 18,19 . Therefore it is urgent to obtain the homogeneous gold nanospheres without any facets. However, the crystalline gold nanoparticles (NPs) prepared in a solution phase always tend to grow anisotropic and accompany with a high tendency to form distinct facets naturally 20,21 , driven by the surface free energy minimization 22 . To date, there have many demonstrations focusing on the synthesis of Au nanospheres, such as a citrate reduction method 23,24 , Brust2Schiffrin method 25,26 , seeding growth method 27,28 . However, most of the nanospheres reported in these methods are not in a truly spherical shape. Usually, they are multiply twinned particles with more or less rounded profile and with smaller facets on the surface 9 . In other words, these nanoparticles obtained by classical growth methods should be called quasi-spheres. It still keeps a challenge to obtain the ultrasmooth gold nanospheres with high monodispersity. To address this requirement, some investigations have been tried to produce real-nanospheres. These methods can be roughly divided into wet chemical methods and novel physical methods for auxiliary. In wet chemical methods, Lee et al. 22 developed a strategy to prepare ultra-smooth, highly spherical monocrystalline gold particles by using the growth in solution and subsequently chemical etching method. Undeniably, these spherical gold crystals are smoother than the one synthesized by conventional chemical methods. But the etching process only removes selectively the grain edges and boundary which have higher surface free energy. Scrutinizing these particles, one can find that the leaving intact particles still have smaller facets. In physical methods, a laser irradiation has been used to modify varieties of noble metal nanostructure to spherical morphology. Koshizaki's group developed a non-focused laser irradiation method to produce spherical sub-micrometer particles of various materials in nanocolloids. In their method, ultra-smooth spherical sub-micrometer particles have been obtained, but display an uncontrollable size distribution, resulting in polydispersity. For contrast, Werner, D. et al. 33 reported that gold nanospheres with controllable size distribution can be fabricated by tuning the adscititious pressure, laser intensity and excitation wavelength in laser irradiation. They emphasized that the application of an external high pressure have suppressed the formation of the bubble, under this case, the size of gold nanospheres can be controlled only by further changing the laser fluence. The application of adscititious pressure makes manipulation in the preparation process complex. Herein, we develop a facile and effective strategy to achieve monodispersed Au spherical nanoparticles. Firstly, large-scale monocrystalline Au octahedral nanoparticle with uniform size are synthesized by a straightforward polyol-route. The size dimensions of Au octahedra can be manipulated from tens to hundreds of nanometers. Secondly, non-focused laser irradiation technology is applied to transform Au particles from octahedron to spherical shape in a liquid under ambient atmosphere in short time. Ultra-smooth gold nanospheres with high monodispersity can be obtained by simply optimizing the laser fluence and irradiation time. Moreover, the morphology transformation of Au octahedral nanoparticle under the nanosecond laser irradiation strongly depends on a photothermal melting-evaporation process. Further, these Au nanoparticles can be fabricated into ordered arrays by self-assembly technique due to their high monodispersity and perfect spherical shape. Importantly, such Au nanospheres array shows a significantly SERS performance associated with their periodic structure due to existence of many SERS hot spots between neighboring Au nanospheres in an array. These gold nanospheres and their self-assembled arrays possess distinct physical and chemical properties that will make them as an excellent and promising candidate for applying in sensing and spectroscopic enhancement, catalysis, energy, and biology . ## Results Pristine Au nanoparticle aqueous colloidal solutions obtained by polyol process contain a large amount of Au uniform, highly mono-dispersed octahedral nanoparticles with average edge lengths of 72 (63.1) nm. It should be mentioned that this edge length is slightly smaller than real one, because the measured value is a horizontal projection of edge length and the vertex of the octahedral nanoparticles is faded out (see Fig. S1 in support information). These nanoparticles were placed under a non-focused nanosecond pulsed laser with wavelength of 532 nm for irradiation using gentle laser fluence at 3.84 mJ cm 22 for 60 seconds. Under such optimized experimental condition, a large amount of spherical nanoparticles were produced, as shown in Figure 1. SEM images in Figure 1a and 1b show that these Au NPs exhibit uniform spherical shapes. Corresponding TEM image shown in Figure 1c indicates these Au nanoparticles demonstrate perfect spherical shapes. An inset in Figure 1d is a feature TEM image of one Au NP, reflecting ultrasmooth surface of particle. Figure 1d is selected area electron diffraction (SAED) pattern of a Au NP in its inset, indicating a face-center cubic (FCC) crystal structure 21 . Additionally, it also reveals that the irradiated Au spherical nanospheres are single crystalline. The average particle size is about 75 (62.6) nm (Figure 1e), showing an excellent monodispersity with very narrow size distribution. All of the above results confirm that the final Au nanoparticles, obtained by non-focused laser irradiation, are ultra-smooth and perfect spherical shape with a highly single crystalline. Figure 2 shows the measured absorption spectra of the Au spherical nanoparticle obtained by laser irradiation and pristine Au nanooctahedra dispersed in water. The optical spectra of Au nanospheres and pristine Au nanooctahedra dispersed in water display absorption peaks centered at 542 nm and 578 nm respectively, which originates from the surface plasmon resonances (SPR) 37 . The SPR peak of Au nanoparticle presents a blue-shift after changing their shapes from octahedral to sphere after laser irradiation, meanwhile, the color of colloid solutions changes from light wine reddish to light brick reddish as shown in the inset of Figure 2. The change of SPRs peaks depends extremely on the nanoparticle structures, because the structures with sharp corners induce the inhomogeneous distribution of surface electronic cloud that yielding additional charge separation and different multipolar moments, as reported by Noguez and Katherine 15,38 . When compared with the octahedral structure, the spherical structure of the same size has a blue-shifted peak. It can be attributed to the reduced charge separation in these structures 39 . Once the charge separation decreased, an increasing restoring force is achieved for the dipole oscillation, resulting in a larger frequency and shorter wavelength. Plasmonic materials, while irradiating with light, can generate plasmon-mediated evanescent fields near their surfaces, which has great potential applications in nanomedicine, nano-optics and plasmonic solar cells and so on. Au nanocrystals with controlled shapes, sizes and surfaces, are ideal building blocks for fabrication of plasmonic materials because of their unique and useful optical phenomena, especially for the spherical geometry. Our results reflect that Au uniform octahedral nanoparticles can be prepared by a straightforward polyol-route, and a further laser irradiation is facile, efficient approach to change them into monodispersed Au perfect spherical nanoparticles. Due to their uniform spherical shapes, these Au nanospheres can be self-assembled into hexagonal close-packed monolayer more easily and this monolayer can be further transferred onto a desired substrate. As shown in Figure 3a and 3b, a large-scaled, close-packed periodic array is obtained by a Langmuir-Blodgett technique. The interparticle gap distances of this array are in the range of sub-10 nm, as shown in Fig. S2. Such Au nanosphere array has important applications in areas of surface-enhanced Raman scattering (SERS) 43 , catalysis 44 and plasmonic crystals 45 , and would render the creation of novel optical materials relying on their tunable plasmonic response 46 . SERS properties of this Au nanosphere arrays on silicon substrates were also investigated. Figure 4a shows the Raman spectra of R6G molecules (10 26 M) on the Au nanosphere array (Curve i) and the normal Au film (Curve ii) for comparison. Obviously, the Au nanosphere array exhibits significantly higher SERS performance than that of the simply sputtered Au film. Qualitatively, the Raman counts at 615 cm 21 for the Au nanosphere array is about 30 times higher than the normal Au film. When the R6G solution decreased into low concentration (such as 10 211 M), the SERS signal is still distinct within 1 s in integral time and 5 mW in laser power, as shown in Figure 4b. ## Discussion The morphological change from nanooctahedra to nanosphere of Au nanoparticle is mainly affected by laser fluence and irradiation time. Laser fluence. Figure 5 shows the FE-SEM and TEM images of Au NPs obtained by laser irradiation (l 5 532 nm, 20 Hz) at varies of laser fluence for 240 s. The original octahedral nanoparticles display distinct edges and corners with no obvious defects on their surface (Figure 5a and 5b). At a low fluence, for instance, 1.76 mJ cm 22 , it is found that octahedral Au NPs can be rarely transformed into spherical shape (Figure 5c and 5d). While increasing the laser fluence to 2.87 mJ cm 22 , more than half number of octahera are transformed into spherical shape (Figure 5e and 5f). When the fluence is 3.84 mJ cm 22 , a completely spherical shape transformation is achieved (Figure 5g and 5h). If the fluence is too high, a size reduction accompanied with smaller fragments of Au NPs will be witnessed because of the photo-thermal evaporation process 47 . For instance, at the high fluence of 5.50 mJ cm 22 , spherical nanoparticles accompanied with small gold cluster are clearly confirmed in the SEM and TEM images (Figure 5i and 5j). These results indicate that the effect of laser fluence is remarkable during laser irradiation and an increase in the laser fluence results in a distinct morphology change from octahedral to spherical at a suitable range of fluence. This can be attributed to the laserinduced surface melting or entire bulk melting of the origin Au NPs after absorbing suitable energy from laser pulses 48 . Besides the shape change during laser irradiation, a fusing process of the Au NPs can be found. This could be the results of the laser-induced agglomeration of two fully melting Au NPs in colloid solution in a little probability, as marked in Figure 5h. In this case, a small quantity of larger spherical Au particles will be generated from this fusion procedure. Occasionally the intermediate fusing state during fusion can be remained since the quick quenching of the fusing particles in the flowing solution, because this colloid solution keeps vigorously stirring during laser irradiation. Figure 6 shows the corresponding absorption spectra of Au NPs in DI water irradiated for 240 s at different laser fluences. It indicates that there is a remarkable blue shift in their localized surface plasmon resonances (LSPR) peak with increase of laser fluence due to shape changes of Au nanoparticle (Figure 6a). When Au nanooctahedra is irradiated by laser at 1.76 mJ cm 22 for 240 s, the LSPR peak shifts from 578 to 565 nm, which can be attributed to the slight melting of the Au octahedra corner, where the LSPR effect is extremely sensitive 49 . While increasing fluence to 2.87 mJ cm 22 , the absorption peak containing a shoulder peak appears. This curve can be divided into two peaks, centered at 540 and 570 nm, which correspond to the LSPR peaks of Au spherical and quasi-octahedral structures, respectively (Figure 6b). When laser fluence increased to 3.84 mJ cm 22 , the LSPR peak is shifted into 542 nm because of the complete formation of Au nanospheres. While increasing into a high fluence of 5.50 mJ cm 22 , the LSPR peak further shifts to 538 nm due to the size reduction of Au nanospheres. In short, a gentle laser irradiation at 3.84 mJ cm 22 can induce completely shape changes from octahedra to sphere and even size reduction, but it is not feasible to obtain a highly monodispersed spherical nanoparticles only by controlling the laser fluence. According to the above results, there might have two possible routes to tackle this dilemma to prepare monodispersed and perfectly spherical nanoparticles. One is to prolong the irradiation time at lower fluence of 2.87 mJ cm 22 , the other is to shorten the irradiation time at higher fluence of 3.84 mJ cm 22 . Nonetheless, considering the pulse-to-pulse laser fluctuations, the lower laser fluence (2.87 mJ cm 22 ) may not be enough to melt the bulk Au NPs completely. Previous studies have pointed out that the high surface-areato-volume ratio of Au NPs led to a lower surface melting temperature compared to the bulk melting temperature 50 . Thus, the irradiation of low laser fluence produce relatively low temperature and might only induce the surface melting of the Au NPs where the inner core of Au NPs still remains solid state. As shown in Figure 5f, the formation of the partial spherical Au NPs could be the result of the fluctuation of laser pulse, which substantially leads to the particle temperature fluctuating above and below the surface melting temperature. Additionally, prolonging the irradiation time will raise the fusion probability among the melting Au NPs. Therefore, it is possible to decrease the irradiation time at 3.84 mJ cm 22 to prepare monodispersed, Au perfectly spherical nanoparticles. And it will be investigated as following part. Irradiation time. At fluence of 3.84 mJ cm 22 for 240 s, although most of Au nanoparticles are spherical shapes, a few dimer and larger Au nanoparticle turns up during laser irradiation due to fusing of two melted particles, as marked in Figure 5h. These dimer and larger nanoparticles are caused by redundant time of laser irradiation and further increasing the fusion possibility. Figure 7 shows the FE-SEM and TEM images of Au NPs obtained by laser irradiation for varies of decreasing irradiation time at 3.84 mJ cm 22 . When increasing the irradiation time gradually, an increasing proportion of spherical Au NPs transformed from Au octahedral nanoparticles was observed, as shown in Figure 7(a-f). Figure 7e and 7f indicate that the expected perfect-spherical Au NPs with highly monodispersity are obtained for laser-irradiation for 60 s. Further prolonging the irradiation time up to 120 s will result in a fusing procedure of the spherical NPs to form Au dimer particles (as marked in Figure 7h). This observation correlated well with our speculation that the fusing of Au NPs is due to the laser irradiation overtime. We also studied the corresponding absorption spectra of Au NPs at different irradiation time as shown in Figure 8. Figure 8a demonstrates an evolution of blue shift of the absorption spectra varying from 20 to 60 s. When increasing the irradiation time, one can observe that a LSPR peak at 578 nm originated from Au nanooctahedra gradually disappears and an absorption peak at 542 nm caused by Au spherical nanoparticles appears, which corresponds well with the results of the FE-SEM and TEM images. Although further prolonging the irradiation time leads to appearance of a few Au dimer and larger nanoparticles (less than 10% of the total number of Au nanoparticles), the peak position is changed not so much, as depicted in Figure 8b. Therefore, we can conclude that the ultraspherical Au NPs with highly monodispersity can be achieved only in optimal conditions, that are, gentle laser fluence (3.84 mJ cm 22 ), proper irradiation time (60 s). Besides above influence factors, a vigorous stir during laser irradiation also affects the formation of high-quality Au nanospheres. It also should be pointed out that the concentration of the colloid solution before laser irradiation should be diluted into a low constant by DI water, because high concentration tends to increase the possibility of collision and fusion among nanoparticles during laser irradiation. Further, to confirm the validity of this method, a similar optimization process was also practiced with the smaller but uniform size of Au nanoocthedra. The ultra-spherical and monodisperse Au NPs with 50 nm in diameter can be obtained by laser irradiation at 3.97 mJ cm 22 in laser fluence and 90 s in irradiation times (Fig. S3). The formation mechanism of Au spherical nanoparticles. In previous studies , the photothermal evaporation and Coulomb explosion mechanisms are two major models for interpreting the pulse laser-induced size reduction of Au NPs . Recently, Werner et. al 58 have demonstrated a clear classification of the photothermal evaporation and coulomb explosion mechanisms based on the twotemperature model (TTM) for simulation. The two-temperature in this model represents for Lattice temperature (T L ) and electron temperature (T e ). They concluded that the photothermal meltingevaporation model is more suitable for the system of nanosecond pulsed-laser excitation. Because, for the gold system, the lattice's melting and boiling temperature (1336 and 3150 K respectively 59 ) is easily reached, while leaving the electron temperature (above 7300 K 58 ) insufficiently for the coulomb explosion to take place. Accordingly, the photothermal melting-evaporation model is suitable for our work due to application of nanosecond pulsed-laser irradiation and it will lead to a deeper understanding of the mechanism in the morphology transformation of Au NPs. Herein, based on the previous mechanisms presented by Werner et al. 50,58 and Pyatenko et al. 60 , we simplify this model to understand the formation mechanism of our work and give as follow. A laser energy, E abs , absorbed from an individual pulse of laser beam with pulse energy, E 0 (t), is equal to 60 E abs ~E0 (t) In this equation, beam cross section is S 0 , J(t) is laser fluence dependent on pulse to pulse time t, J(t) 5 E 0 (t)/S 0 , and s l abs is the particle absorption cross section, which is strongly dependent on laser wavelength and refractive index of surrounding medium n m . As described in TTM models, the laser energy absorbed by the metal NP is eventually transformed into heat, leading to a rise of the temperatures of T e , T L and T s (surrounding medium). The differential heat equation is given as follow 58 , Here, (m e , m L , m s ) and (c e , c L , c s ) are the mass and specific heat capacity of the electron, lattice and surrounding medium, respectively; k is the thermal conductivity, D represents Laplace operator, g denotes the coefficient of the electron-phonon coupling, E abs is the absorbed laser energy and F represents the interface energy transfer between the NP and liquid. For simplifications, the first terms on the right side of equation (2, 3) are neglected (for K e (T e , T L )?DT e and K L (T L )?DT L < 0) 58 . Additionally, for nanosecond laser excitation, the time evolution of T e and T L is in quasi-equilibrium (T L < T e ) during the excitation period 58 , and the mass of electron could be negligible compared with the one of lattice. For convenience, the heat loss caused by the radiation cooling and heat diffusion to the surrounding water (F) is also negligible during particle heating, because the typical times needed for particle cooling/solidification range from 10 25 to 10 26 s 60 , which are much longer than the nanosecond laser (t p 5 10 28 s). On the other hand, the lattice heat will not be accumulated from one laser pulse to the next, for the time required for the cooling/solidification of Au NP is much shorter than 50 ms (20 Hz). According to the above approximation, the equations (2-4) will be generalized into a simplified form during the particle heating process. Integrated with the equation ( 1), the general relationship between the nanoparticle temperatures (T) and laser fluences (J) is given in equation ( 5). Where, the subscript j defines as a mathematic set of the particle thermodynamic states, e.g. heating, melting and boiling to evaporation. H is the relative enthalpy per unit mass. Herein, the laser fluence (J) for each laser pulse can be considered as a constant. It is thought that a high absorption cross section s l abs is achieved at 532 nm exciting because of the SPR property of Au NPs. When adopting the laser beam with wavelength of 532 nm, s l abs is also a constant value since the octahedral Au NPs employed in our research are uniform and highly monodispersed. With the help of equation ( 5), it is easy to understand the mechanism of the photothermal melting-evaporation of shape changes from Au nanooctahedron to nanosphere under nanosecond pulsed-laser irradiation. Figure 9 shows a schematic to illustrate the morphological evolution of the octahedral Au NPs under nanosecond pulsed-laser irradiation. When laser fluence J p is smaller than the bulk melting threshold J m , only particle heating is achieved or even a slight melting sharp corner is taken place. Or perhaps a surface melting of particles will be achieved due to their high surface-area-tovolume ratio 61 . While at the condition of J m , J P , J evp , a bulk melting occurs and the particle turns into ultra-spherical liquid phase. Once the heat dissipation to the surrounding medium, particles solidify into ultra-spherical phase for keeping. However, when the condition of J P . J evp is met, the particle reaches the boiling temperature and starts to evaporate. The laser-induced size reduction takes place, and the slightly smaller and much smaller Au NPs are observed as a result of the layer-by-layer surface evaporation. A large scale ultra-spherical and highly monodispersed Au NPs can be prepared via this two-step strategy at optimal conditions. However, a small quantity of spherical Au NPs will agglomerate and fuse into a larger spherical NPs for the overtime laser irradiation, as evidenced from the TEM image of Figure 5h and 7h. Therefore, the interaction among the spherical Au NPs should be taken into consideration during laser irradiation. According to Takeshi et al.'s previous research 62 , laser irradiation using a non-focused laser beam will bring about the removal and decomposition of stabilizing molecules on the surface of the Au NPs, and that causes the agglomeration of NPs dynamically, then induces the fusion of the agglomerated Au NPs. In our work, a small quantity of the Au nanospheres obtained by laser irradiation also displays a tendency to agglomerate into dimer dynamically. While irradiated by the next laser pulse, the dimer turns to a bulk melting state and dynamically fusing into a larger spherical particle, as illustrated in Figure 9b. This aggolomeration of Au Nanospheres can be explained for the partial removal and decomposition of PDDA molecules on the surface of Au NPs during laser irradiation. In that case, prolonging the irradiation time will set up a vicious circle of the fusion process. In such vicious circle, larger gold nanospheres are gathering increase and lead to a poor monodispersity. Therefore, we confirm that an appropriate irradiation time is essential to maintain the monodisperity of Au nanospheres while the phase transformation takes place. SERS active substrate of Au periodic nanosphere array. Finite-Difference Time-Domain (FDTD) simulation was employed to fully understand local electric field distribution of Au nanosphere array under light excitation, as shown in Figure 10. As we all know, two mechanism models, electromagnetic (EM) enhancement and chemical enhancement (CE), are the predominant roles in explaining the SERS phenomenon 63 . From the EM point of view, when the frequency of incident light is resonant with the SPR of the Au nanoparticle, a redistribution of the local EM field is induced by the coherent action between their dipolar filed and the exciting electric field. This process leads to a special 'hot spot' position, in which the EM field around the NP is greatly enhanced. The molecule near or adsorbed at the hot spot is excited and results in much enhanced Raman-scattered signals. What's more, gaps (in sub-10 nm range) between adjacent Au nanoparticles provided further enhanced local EM fields because of the electromagnetic coupling at the junctions between neighboring nanoparticles 43 . This interparticle electromagnetic coupling causes the amplification of the polarization of the plasmons, thus generating large enhancement SERS signal from molecules in illuminated area, known as 'hot spots'. Herein, the periodic NP arrays results from selfassembly provide maximum surface density of well-defined ''hot spots'' upon optical excitation, exhibit enormous near-field enhancements exploitable for large SERS enhancements. In conclusion, a facile and effective strategy is developed to synthesize monodispersed Au spherical nanoparticles by two steps, preparation of large-scale monocrystalline Au nanooctahedra with uniform size by a polyol-route and subsequently rapid synthesis of monodispersed Au nanospheres through a laser irradiation-induced shape conversion of Au nanooctahedra in a liquid under ambient atmosphere. High monodispersed, ultra-smooth gold nanospheres can be obtained by simply optimizing the laser fluence and irradiation time. We discovered that there have two major parameters that affect the formation quality of Au nanospheres: laser fluence and irradiation time. Mechanism investigation suggests that photothermal melting-evaporation model is suitable for the system of nanosecond pulsed-laser excitation to understand the complex process in our work. These Au nanoparticles could be fabricated into periodic monolayer arrays by self-assembly utilizing their high monodispersity and perfect spherical shape. These Au nanospheres arrays demonstrated very good SERS enhancement related to their periodic structure due to existence of many SERS hot spots between neigh-boring Au nanospheres in an array and hence can be used as an excellent and promising candidate for sensing and spectroscopic enhancement, catalysis, energy, and biology. ## Methods Uniform Au octahedral nanocrystals were prepared by a facile polyol route according to our previous work 37 . Briefly, a given amount of chloroauric acid (HAuCl 4 , Aldrich) as gold source, poly(diallyldimethylammonium) chloride (PDDA, M w 5 100,000 , 200,000, 20 wt % in H 2 O, Aldrich) as surfactant, and HCl solution (Aldrich) were introduced into an ethylene glycol solution in a glass vial under vigorous stir. The final concentration of AuCl 4 2 ions, PDDA, and HCl in the initial gold precursor was 0.5 mM, 25 mM, and 5 mM, respectively. Finally, the mixture solution was heated at 195uC for 30 min in an oil bath without stir. The final color of the solution appears to light wine reddish, reflecting the formation of Au nanocrystals. The colloidal Au octahedral nanoparticle diluted with deionized (DI) water (153 volume ratios, 4 mL in total) was added into a clean weighing bottle (25 mm 3 40 mm) under magnetic blender. A non-focused Nd: YAG laser operated at 20 Hz with a wavelength of 532 nm and pulse duration of 10 ns vertically irradiated into the Au colloidal solution in weighing bottle. The diameter of laser beam irradiated on the solution was about 5 mm. The solution was continuously stirred by magnetic rotor during irradiation. With increase of irradiation time, the color of colloidal Au nanoparticle was gradually changed from light wine reddish to light brick reddish. The final product was collected by centrifugation at 12, 000 rpm for 45 min and washed repeatedly with pure DI water for further characterization and optical property study. These Au nanospheres obtained by laser irradiation were modified with excess amounts of 30 mM dodecanethiol in ethanol solution as the capping agent for further self-assembly 64 and then centrifuged, dispersed into chloroform. The Au monolayer nanosphere arrays were formed by dropping the mixed solution onto the water surface by self-assembling process until the chloroform was fully evaporated. This Au monolayer was gently picked up by using a clear silicon substrate, dried with flowing N 2 gas, and then placed into a ultra-violet ozone cleaning systems (UVOCS) for 1 h to remove the surface coating materials. For Raman spectral examination, these pretreated samples were soaked into Rhodamine 6G (R6G) with different concentrations for 30 min, rinsed with DI water, and dried with flowing N 2 gas for further SERS characterization. The morphology of the final product was characterized by a field emission scanning electron microscopy (FE-SEM, FEI Sirion 200). The samples for TEM examination were prepared by dropping the Au nanoparticle colloidal solution on copper grids with thin carbon coating and then drying by evaporation in air at room temperature, selected area electron diffraction (SAED) studies were performed on a JEM-200CX operated at 200 kV. For optical measurement, the products were dispersed in water and then optical absorption spectra were recorded with a spectrophotometer (Cary 500) in the wavelength range of 200-800 nm, using an optical quartz cell with a 10 mm path-length. The Raman spectra were recorded on a macroscopic confocal Raman spectroscopy (LAICA DM 2500) by using a laser beam 532 nm in wavelength for excitation at room temperature. The integral time and accumulations were 1 s and 3 times for all samples respectively.

chemsum

{"title": "Rapid Synthesis of Monodisperse Au Nanospheres through a Laser Irradiation -Induced Shape Conversion, Self-Assembly and Their Electromagnetic Coupling SERS Enhancement", "journal": "Scientific Reports - Nature"}

helical_frontier_orbitals_of_conjugated_linear_molecules

## Abstract: Compounds containing allenes, cumulenes and oligoynes (polyalkynes) have attracted attention for both their conformation and reactivity. Whilst the textbook molecular orbital description explains the general electronic and molecular structure of the cumulenes, there are anomalies in both the crystal structures and cycloaddition products involving oligoynes and allenes; the understanding of these molecules is incomplete. Through a computational study we elucidate that the frontier orbitals of the allene and oligoyne families are extended helices. These orbitals are the linear analogue to the Möbius aromatic systems, which also display non-linear p interactions. The axial chirality found in allenes and oligoynes is intimately related to the topology of the frontier orbitals, and has implications for predictions of cycloaddition pathways, structure stability and spectroscopy. ## Introduction Allenes, cumulenes and oligoynes (polyalkynes) are intriguing structural motifs; they partake in cycloadditions, act as Michael acceptors and are redox active. There has been a recent resurgence in studies relating to these molecules, as their inherent chirality, reactivity and unusual physical properties are useful in organic synthesis and as linking units in metalorganic frameworks (MOFs). Allene, 1, is the most fundamental cumulated alkene; three carbons are bound linearly with double bonds. Through a basic molecular orbital construction of the p-system, the terminating motifs obtain an orthogonal conformation, Fig. 1. Allene may be axially chiral; the most simple example, 1,3-dimethyl allene, has two stereoisomers, (M) and (P). The next smallest cumulated alkene, cumulene (2), is planar by the same orbital description proposed for allene. The four carbons result in a planar molecule, achiral unless the terminating protons are substituted with a point chiral motif (e.g. ()-menthol). A recent publication by Tykwinski and co-workers explored the structures of 2 and its extended variants, all of which had an even number of linear carbons and were achiral. 8 To the best of our knowledge, the most relevant studies of the next odd cumulene, containing fve carbons, are related to interstellar diradicals, and other quasistable analogues. There is no evidence to suggest that the odd cumulated alkenes are less stable than their even relatives, but rather that the previous synthetic routes do not yield odd catenated chains. There is scope for the synthesis of these molecules through some form of modifed metathesis, but this has thus far not been explored. The term 'cumulene' is not only the conventional molecular name of 2, but is also used to refer to similar extended alkenes with four or more carbon atoms. This does not distinguish between the odd (1-like) and even (2-like) systems, or the molecule itself. Cumulenes like those described by Tywinkski and co-workers are even numbered and are thus planar (D 2h ) and pose no possibility for axial chirality. The odd cumulated alkenes are terminated in orthogonally-orientated functional motifs (D 2d ), and therefore may be axial chiral. In this work we refer to odd numbered cumulated alkenes as the family of 'allenes' and the even cumulated alkenes referred to as the family of 'cumulenes'. Fig. 1 The archetype allene, 1, and cumulene, 2, with their simplified p bonding shown on the right. Even cumulated alkenes may be made from the hydrolysis of oligoynes similar to that of 3. Related linear systems include the simple conjugated oligoyne, acetylene dicarboxylate, 4, which is presumed planar, as is the aromatic terminated oligoyne, 5. The cumulene family may be made by the two-electron oxidation of a cumulated alkyne precursor. Indeed the precursor alkynes, like that of 3, share a similar electronic structure to their two-electron oxidised cumulene derivatives. Curiously, there have been a signifcant number of studies on extended conjugated oligoynes (ECOs), and their conformation; 4 and 5 are archetypes. It has been assumed that the orientation of the ECO family is planar like the cumulene family, and some computational investigations have fxed this planarity through enforced D 2h symmetry. 9,19 Some studies do correctly suggest that ECOs demonstrate non-planar geometries: the terminating functional motifs rotate out of plane if they partake in the p conjugation of the alkyne. 20,21 This geometry is similar to that observed in the allene family, with one fundamental difference: ECOs have even numbers of carbon in the chain. It is certainly less obvious why ECOs are structurally analogous to allenes. There are fleeting examples of such isolated structures; acetylene dicarboxylic acid has been observed in the orthogonal orientation in its crystal hydrate, seemingly stabilised through hydrogen bonding. 22 More recently, orthogonallyorientated oligoynes have been observed in MOFs, primarily as linking units. 23,24 IRMOF-0 (ref. 25) encompasses the 'twisted' acetylene dicarboxylate, 4, and a similar MOF isolated by Burrows and co-workers features an orthogonal aromatic carboxylate terminated ECO. 26 To understand the chemical similarities of the allenes and ECOs, and the differences between cumulenes and allenes, we must frst understand the electronic structure of these molecules. Through quantum-chemical calculations we report on an unusual electronic characteristic in the allene and ECO families: extended helical molecular orbitals. The topology of these MOs has implications for predictions of cycloaddition products. The directed electronic structure is the linear analogue of the electronically chiral cyclic systems (canted p interactions) like that observed in Möbius aromatics. 27 Furthermore, it is these extended electronic helices which result in the observed orthogonality of the allene and ECO systems. The cumulene family displays no helical nature, further emphasising the difference between even and odd cumulated alkenes. ## Electronic differences between the 'odd' allenes and 'even' cumulenes The textbook MO construction of the p interactions between orthogonal atomic p-orbitals is sufficient to describe the electronic structure of the cumulenes. It is, however, inadequate in the description of the allenes. Quantum-chemical computations of these molecules confrm that the frontier molecular orbitals (FMOs) are comprised of p-orbitals for both systems, but perhaps not what one would expect. These are drawn schematically in Fig. 2 and formally in Fig. 4. The allene family have extended FMOs with helical topology. Through extended non-linear p interactions, allenes exhibit extended conjugation similar to that observed in Möbius aromatics, Fig. 2a, which are cyclic conjugated systems that feature the electronic topology of a Möbius ring. 27 The cumulenes show the expected p-orbital orthogonality with a centred node in the HOMO, Fig. 2b. The HOMOa and b of the achiral allenes are a degenerate pair, comprised of the right and left-handed helices orientated 90 from each other. The rotation and polarisation of the p-orbitals of the central carbon atom results in this interesting electronic structure, with visibly helical molecular orbitals, Fig. 2a. The achiral allenes feature both the right and lefthanded bias through electronic degeneracy. The splitting of this degeneracy may be achieved chemically by forming an axially chiral allene, thus differentiating the helical MOs. Helical molecular orbitals can be observed in the cumulene family if the terminal motifs are non-planar. However, the equilibrium geometry of cumulenes is planar, and thus there is no potential for axial chirality. Even chemical substitution, e.g. methylation, does not induce non-planarity of the terminal motifs; therefore, there is no observed helical selectivity in organic reactions or in computational models. The FMOs of the axially chiral (P)-1,3-dimethyl allene are shown in Fig. 4b; the degeneracy is removed upon substitutional methylation. The resulting MOs are helical; the HOMO of (P)-1,3-dimethyl allene is the left-handed helix. The HOMO1 is found 0.002 eV below the HOMO, and is righthanded. Considering the other enantiomer, (M)-1,3-dimethyl allene, the HOMO is inverted to a right-handed helix. Contrast these helices to the molecular orbitals of a point chiral molecule: the chirality can be described by a non-linear asymmetric dipole about the stereogenic centre, like the direction and magnitude of the hands on a clock reading 01:45, Fig. 3a. A helix is the projection of this dipole through a third dimension, Fig. These linear molecules are unusual because they maintain both chiral identities in their molecular orbitals. The helical nature of the allene orbitals is best observed, but not limited to, the FMOs. For axially chiral allenes, like those shown in Fig. 4b, it is important to note that the 'handedness' of the HOMO and LUMO are the same (and opposite to that of the HOMO1 and LUMO+1). Helicity is a familiar structural topology in organic chemistry; 29-33 a familiar example is the helicene family. Logically, the HOMO is composed of the highest energy occupied linear combination of atom-centred orbitals and is intimately related to the molecular geometry. For conventionally chiral systems there is no HOMOn that reflects the opposite enantiomer. Considering the helicenes, there is structurally only one electronic enantiomer present. In linear molecules, like the allenes, both identities are portrayed, further lamenting the unique behaviour of the helical FMOs explored herein. There is experimental precedence for the existence of helical molecular orbitals. Consider the familiar [2 + 2] cycloaddition described in Fig. 5. In a general sense, the occupied orbitals of 7 interact with the unoccupied orbitals of 8 forming a cyclobutane. The criteria for such reactions vary. Some require irradiation to promote the reaction of the frst excited state of the electron acceptor (8), others are thermally favourable. There are many examples of allenes partaking in similar reactions, best summarised in the recent review by Alcaide and coworkers. 7 There are accounts of allenes partaking in [3 + 2] cycloadditions (acting as the 3 component) through either in the presence of a nucleophilic organocatalyst 40,41 or a metal catalyst. 42,43 In such instances, an enolate is formed, subsequently racemising the allene. It is rather the [even + even] cycloadditions that demonstrate the application of such helical FMOs. The allene cycloaddition reactions obey the Woodward-Hoffmann rules, but the general consensus is that these cycloadditions are not concerted. 7 In [even + even] cycloadditions involving chiral allenes, there are many reports of enantioselectivity in the formation of radical intermediates. Two intriguing examples are the thermal [2 + 2] of 6 and methyl propiolate, 47 and a more complex [2 + 2] intramolecular photocycloaddition. 48 In both instances single molecule calculations affirm that the HOMOs are 'left-handed'. Previously, high selectivity in cycloadditions involving chiral allenes was rationalised through steric preferences of the reagents. 47 This is almost certainly not the case; whilst the interacting geometry is of thermodynamic importance, the enantioselectivity is the product of orbital symmetry conservation (i.e. the symmetry of the HOMO of the allene in relation to the LUMO of 8 is allowed). Fig. 6a describes a proposed symmetry-allowed in-phase orbital reorientation, which is intimately related to the 'handedness' of the allene HOMO. This is in contrast to the symmetry-forbidden p-orbital reorientation shown in Fig. 6b, where the extended conjugation is broken rapidly. Therefore, the application of the helical molecular orbital description provides insight into the predictions of cycloaddition reactions and accurately describes the observed high enantioselectivity. Indeed, the products of the cycloaddition reflect the orbital symmetry of the precursors, allowing the heuristic prediction of the electronic structure of the allene precursor. This may become more important once large odd cumulated alkenes are synthesised and used in similar reactions. ## Extended conjugated oligoynes Unlike the allenes and cumulenes, which are orthogonal and planar, respectively, the geometries of ECOs exhibit more variation. 49 The simplest ECO, acetylene dicarboxylate (4), and the ester derivative, are generally considered to be planar, resulting in a non-degenerate HOMO/HOMO1 due to the inequivalent p x and p y interactions. The generalised family of ECOs are shown in Fig. 7. The ECO equilibrium geometry can be found by rotation of the conjugated terminating groups (carboxylate torsion angle; CTA). The potential energy surface has been constructed for the archetype ECO systems, Fig. 8. The acetylene dicarboxylate (4) orbital images accompanying the PES illustrate the HOMOs at the local maximum (CTA ¼ 0 , D 2h symmetry), twist transition (CTA ¼ 45 , D 2 symmetry) and the global minimum ($90 , D 2 , near D 2d , symmetry). The global minimum was found through both an unconstrained optimisation of 4, 11 and 12 (depicted as shaded points in Fig. 8), and the fxed CTA calculations used to construct the PES. Very good agreement between the calculated and experimental bond lengths is observed, with a difference of the order 10 2 . All of the molecules described in Fig. 7, and their structural analogues, have non-planar global minima. A near orthogonal conformation (CTA ¼ 87.28 /92.72 ) is energetically favourable for the monoalkyne system, 4, with a 10.33 kJ mol 1 energy stabilisation over the planar variation. A similar trend is observed for the larger oligoynes, 11 (CTA ¼ 88.76 /91.24 ) and 12 (CTA ¼ 89.88 /90.12 ), with 4.99 and 2.96 kJ mol 1 , respectively. It is notable that both the planar and orthogonal conformations are saddle points (both D 2h and D 2d prevent the formation of the helices). As catenation is increased, the barrier of rotation is reduced and the global minimum converges to the orthogonal conformation, CTA ¼ 90 . The origins of the unfavourable planar geometry are the combination of repulsive interactions between the terminating functional groups (visualised in the PES, CTA ¼ 0 ) and the constructive orbital overlap, similar to the Möbius systems, upon twisting (Fig. 9). Thus for different reasons, extended conjugated oligoynes obtain a similar geometry to that of the allene family, with one important difference: all extended conjugated oligoynes are made up of even numbered carbon chains. The structural twist is observed in all of the ECO series; selected examples are shown in Fig. 9. The electronic result is the formation of a degenerate extended helical p system similar Fig. 7 Extended conjugated oligoynes and their variable terminating motifs. In the case where R is excluded the systems are considered as the di-anion. to that of allene. In addition to the occupied electronic orbitals, most virtual orbitals are also helical in nature. The lowestunoccupied molecular orbital is helical and degenerate. These helices are more robust than the HOMO orbitals with respect to chemical substitution; we have not discovered a oligoyne terminating in any conjugated motif that removes the extended helical nature of the orbitals. This fnding may become important when exploring optical excitations, and have interesting implications for cycloaddition reactions. Similar to the cumulenes, an approach for creating inequivalent right and left-handed helices in oligoynes is the installment of a point chiral motif, Fig. 7d, f and g. 50 The fundamental problem with a chiral ester functionality is that the alkoxy oxygen is no longer formally part of the conjugated system. Effective removal of the helical p-system can be achieved by attaching electron-withdrawing groups (e.g. -Cl, shown schematically in Fig. 7e and formally in Fig. 9) or an organic molecule (e.g. -Me, Fig. 7a-d) to the terminal oxygen atoms. The reduction in conjugation is a product of the re-hybridisation of oxygen, removing p overlap with the alkyne chain. Extended electronic helices observed in 5, and relevant derivatives, are affected to a lesser extent by substituents on the ring. Depending on the substituent, strongly electron withdrawing groups, like the carboxylic acids, cause the helical MOs to appear in the HOMO5 to HOMO8 region. An alternative approach is the substitution of the terminal carbons for chiral sulfoxides or oxazolidinones. We have proposed the R-phenylsulfnyl substituent as a conjugated alternative to carboxylate termini (as shown by 13, Fig. 9d). It is not necessary for the 'dangling' substituents to be conjugated to the oligoyne. In this instance, the conjugation in the sulfoxide is extended. By the addition of the point chiral motif, the oligoyne HOMOs are split, again reflecting the chirality of the system. In this case, the separation of the HOMO (right-handed helix) and HOMO1 (left-handed helix) is $0.02 eV. One promising example is the oxazolidinone described in Fig. 9e. The nitrogen acts similar to the sulfur in this instance, conjugating to the alkyne motifs and, more importantly, providing a bias for the helices. The specifc oxazolidinone, (R)-4-isopropyloxazolidin-2one (14), is both synthetically plausible, and, like the sulfoxide, can be isolated in a solid (similar to presence of 4 in MOFs). Single-molecule calculations of 14 result in a 0.42 eV energy separation of the HOMO and HOMO1. The chiral oxazolidinone and sulfoxide demonstrate a separation of degenerate orbitals in the ECO family. The attachment of these conjugated oligoynes amplifes the effect of point chirality, and should be detectable spectroscopically. With respect to reactivity, the oligoynes may act as both the electron accepting and donating motifs. The cycloaddition discussion for allenes is applicable to oligoynes; however, the isolation of a single helix remains an area that has yet to be thoroughly explored. In principle, the kinetic product of the chiral oligoyne [2 + 2] cycloaddition reaction should reflect a similar outcome to the allenes; the helical conformation is thermodynamically favoured. There should be chiral selectivity in cycloadditions involving the oxazolidinone (the HOMO is helical and separated by 0.42 eV from the inverse HOMO1). There are ECOs that do not exhibit extended helices. A contemporary example is porphyrin systems linked by diynes, 54 where there are two interactions that disfavour the formation of the orthogonal structure and thus helical orbitals. Firstly, the tetrahedral Zn ions interact with the porphyrin, altering the hybridisation of the alkyne/porphyrin orbitals. Secondly, the extended conjugation is terminated with unconjugated silyl motifs. As demonstrated here, unconjugated motifs remove the extended helicity. The planar conformation is still conventionally conjugated in the equilibrium geometry; the HOMO and HOMO1 consist of nodes not observed in the extended helices of orthogonal ECOs. ## Conclusion The description of both the orbital structures of the allenes and extended conjugated oligoynes, and the reactivity of these motifs is incomplete. Herein we have described the distinct electronic structure of the allene and extended conjugated oligoyne families; both the allenes and the ECOs display extended helical frontier orbitals. Whilst these extended helices are degenerate in the unsubstituted systems, chemically activated chirality (distinct right and left-handed helices) may be preferentially performed through intelligent design. Our results justify the observed structural orthogonality in allenes and ECOs. The orthogonality of the ECOs is the result of the conformation being in a potential energy minimum, a stabilisation due to the formation of extended helices. The allenes are inherently structurally orthogonal, but display similar helical orbitals. This helical nature provides a rigorous explanation for enantioselectivity in cycloadditions involving chiral allenes. The helicity found in the ECOs is predicted to be observable spectroscopically and in enantioselective cycloadditions. Furthermore, these structures are ones of importance with applications as both ligands and guest molecules in metalorganic frameworks. Their rigidity and potential optical response makes them interesting ligand candidates. Importantly, we have also demonstrated a fundamental difference between the cumulenes and allenes. The cumulene family (even numbered carbons) are planar and do not display helical MOs, whilst the allene family (odd numbered carbons) are orthogonally terminated and display electrohelicity. The distinct behaviour further laments that molecular orbitals are not always what they appear; indeed, an ending with a twist. Since our original submission, an investigation of the mechanical properties of extended conjugated systems has been reported, 55 which identifes a periodic torsional response consistent with the canted p interactions that we observe. ## Computational details All quantum chemical calculations were performed using the GAMESS-US package. 56 The atomic geometries were optimised within the Kohn-Sham density functional theory (DFT) 57 construct. The total energy and forces were calculated with the hybrid meta-generalised gradient approximation functional, TPSSh, 58 using a triple zeta basis set, 6-311+G(3d). 59 The potential energy surfaces were obtained by relaxing the internal coordinates as a function of the dihedral angle between the terminal groups. A range of semi-local and non-local exchangecorrelation functionals (e.g. PBE, 60 PBE0, 61 M06 (ref. 62)) and a wave function based method (MP2) 63 were found to give similar predictions. To ensure quantitative results, the electronic structure and internal energies were also calculated for select systems at the coupled cluster level of theory including single, double and perturbative triple excitations, CCSD(T), 64,65 with the same basis set. The potential energy surfaces and molecular orbitals show very good agreement with the TPSSh results, with the exception that the barriers for rotation for the oligoynes are slightly reduced (+2 kJ mol 1 for 4), which is consistent with a more rigorous description of electron correlation. 66 Visualisations of the structures and orbitals were made using the codes VESTA and GABEDIT. 67,68

chemsum

{"title": "Helical frontier orbitals of conjugated linear molecules", "journal": "Royal Society of Chemistry (RSC)"}

adsorption_isotherm_of_methylene_blue_dye_on_groundnut_shell_and_sorghum_husk_kyauta_sunday,_erugo_i

## Abstract: Dye removal using low cost adsorbent is a suitable method for textle wastewater treatment. The aim of this research work is applicaton of groundnut shell and sorghum husk in powdered form as low cost adsorbent for methylene blue and Congo red dye removal from water at laboratory scale. Batch studies were carried out to determine the adsorpton equilibrium of methylene blue and Congo red dye on the two diferent adsorbent and also to verify the reported data using Langmuir, Freundlich and Temkin isotherm. The study revealed that for the adsorpton of methylene blue on sorghum husk, Langmuir Isotherm fied well the experimental data with R 2 value of 0.9710, while for the adsorpton of methylene blue on groundnut shell. The Temkin binding energy (b T ) are positve for all the experiment, indicatng that the process is endothermic. The present work will help to carry out studies in packed column and later scale it up for industrial applicaton. ## Background The released voluminous amount of toxic Substances in to the water system has afected the human beings as well as aquatc animals. Dyes are used extensively in various industries such as textless rubbers plastcss printngs leathers cosmetcss etc.s and also in producton of colored products. Dyes afect the penetraton of sunlight into the water bodies and thus interfere with the growth of bacteria and hinder photosynthesis in aquatc plant. It poses a serious threat to mankind and water qualitys thereby is a mater of vital concern. It causes acute and chronic efects on exposure to human skins such as allergics dermattss skin irritatons cancers mutaton etc. It is very difcult to separate dye from water by using Using conventonal techniques ( K. Upadihyays 2019). Among the various treatment methodss adsorpton processs chemical coagulaton focculaton degradaton process etc. have been explored to remediate these dyes in the waste water (S. Khatris 2016). Among the various techniques available for its remediaton adsorpton technique has been proved to be most efectve process. Adsorpton is preferred over other processes due to possible regeneraton and recovery of the sorbet. Adsorpton is an unit operaton process which refers to atachment of molecules of liquid or gas on the surface solid substance. It is based on the fact that some solids preferentally adsorb other solute from the soluton onto their surfaces. Dyes are partcularly removed using various adsorbents. Many such adsorbents have been explored for its removal. Howevers in this research work we considered adsorpton process using groundnut shell and sorghum husk ## PROBLEM STATEMENT The world is becoming more industrialized therefore the amount of toxic material (e. g dye) released in to the environment is becoming highs some agricultural by-product can be used for water purifcaton so as to comply with the environmental regulaton. ## AIM AND OBJECTIVE The aim of the research work is to study the adsorpton Isotherm for methylene blue dye on groundnut shell and sorghum husk. ## The objectves of the research work are: 1. To determine the adsorpton isotherm that best fts the batch adsorpton data. 2. To determine values of Langmuirs freunlish and Temkin constants which defned the adsorpton process. ## SCOPE. The research is limited to the adsorpton of methylene blue dye using groundnut shell and sorghum husks and the Isotherm model used are Langmuirs Freundlich and Temkin Isotherm. ## Justification; 1. Massive amount of groundnut shell and sorghum husk are readily available to be used as potential material for adsorption instead of being disposed and causing environmental pollution. ## Adsorpton Isotherms Isotherms give an equilibrium relatonship between the amounts of adsorbate adsorbed on the adsorbent surface and its concentraton in the soluton at a constant temperature. Numerous adsorpton models are available in the literature to ft the experimental adsorpton data. In this studys the data were fted using Langmuir Freundlich and Temkin models. Each of the these models makes use of a parameter q e (i.e. adsorpton capacity per unit mass of the adsorbent at equilibrium) in mg/g The linear form of Langmuir expression: Where C e is the equilibrium concentration of dye solution (mg/L), q e is the equilibrium capacity of dye on the adsorbent (mg/g) q o is the monolayer adsorption capacity of the adsorbent (mg /g), b is the Langmuir adsorption constant (L/mg) and is related to the free energy of adsorption . The essential characteristic of Langmuir Isotherm is expressed in terms of dimentionless constant separation factor or equilibrium parameter RL, is define as: Where Co is the initial adsorbate concentration (mg/g) and b is the Langmuir constant (L/mg) (kimet et al, 2014). In the context, Lower R L value reflect the adsorption is more favorable. In deeper explanation, RL value indicate the adsorption nature is either unfavorable (R L >1), Linear (R L =1), favorable (0<R L >1), or irreversible (RL=0) (Kimel et al, 2014). ## Freudnch Isotherms A freunlich Isotherm is a mathematical equilibrium between a fluid and a solid material. The freunlich expression is a mathematical equetion between an imperical expression representing the Isotherm variation of Liquid or gas on a solid material. The slope ranges between 0 and 1 is a measure of adsorption intensity or surface heterogeneity, becoming more heterogeneous as its value gets closer to zero. Whereas, a value below unity implies chemisorption process where 1/n above one is an indicative of cooperative adsorption. It's linearized and non-linear. The linear form can be written as: Ln Where, k f and n (dimensionless constants) are the Freundlich adsorption isotherm constants, which indicates the capacity and intensity of the adsorption, respectively. ## Temkin Isotherm Model Temkin Isotherm assumes that the adsorption heat of all molecules decreases linearly with with the increase in the coverage on the surface of the adsorbent. The general form of Temkin equation is giving below The linear form equation is written as The curve is obtain by plotting q e against ln C e , with A T as Temkin isotherm equilibrium binding constant (L/g), b T Temkin isotherm constant R, is universal gas constant (8.314 j/mol K) and T, is temperature (J. Piccin et al, 2011) ## Groundnut Shell Groundnut botanically known as arachis hypogeal belongs to legumiminose family. A complete seed of groundnut is called the pod and the outer layer is called the shell. Brian George et al. investigate the groundnut shell fibers was found to be 38mm and 0.2mm diameter. Average tenacity of groundnut shell fiber is of 1.06g/den. Also, average strain of the fibers was found to be 7. ## PREPARATION OF THE ADSORBENT Adsorbents (groundnut shell and sorghum husk) were collected from Sabon kauras Bauchi local government areas Bauchi state. The agricultural by product were washed and dried at low temperature (<105˚C) for 48 hours to remove moisture content. After the drying processs material were ground to fne powder and sieved through 600 µm size. The adsorbents is then used for the research work. ## Procedure for Adsorpton Isotherm Studies 50mg/L were prepared. 4.0g of groundnut shell of the prepared sample was taken in to a conical fasks containing the solutons with a known inital concentraton and the soluton was maintained at a constant PH and strred untl equilibrium concentraton is reached. The equilibrium concentraton is recorded and used for the Isotherm study. The same procedure was used to carry out the same experiment using 3.6g of sorghum husk and methylene blue dye soluton. ## RESULTS AND DISCUSSION This secton presents the results and discussions of the research. The following data were used: Volume of the soluton used = 50ml ## Methylene Blue Adsorpton on Sorghum Husk The Langmuirs Freundlich and Temkin Isotherm study for Methylene Adsorpton on Sorghum husk was carried out. The Langmuirs Freunlich and Temkin plot were given in fgure 4.1s 4.2 and 4.3 respectvely. The value of q e is calculated using equaton (2.1) and the values needed for the calculaton of Isotherms were recorded in From the Langmuirs Freunlich and Temkin isotherm plot of methylene blue removal on sorghum husks R 2 values of Langmuir constant was 0.99700 which is higher than R 2 value of freundlich and Temkins this means that Langmuir data fted well than freundlich (0.983) and Temkin (0.9758) data in the adsorpton process and That shows that the process is a Monolayer adsorpton process. ## Methylene blue adsorpton on Groundnut Shell The Langmuirs freundlich and Temkin Isotherm for methylene blue adsorpton on Groundnuts shell was carried out. The Langmuirs Freundlich and Temkin plots are given in fg 4.4s 4.5 and 4.6 respectvely. The value of q e is calculated using equaton (2.1) and the Values needed for the calculaton of Isotherms were recorded in Table A2. From the above Langmuirs Freunlich and Temkin isotherm plot of methylene blue removal on groundnut shells R 2 value of freundlich constant is 0.960 which is higher than R 2 value of freundlich and Tebkins this means that Langmuir data fted well than Langmuir (0.9436) and Temkin (0.944) data in the adsorpton process. That show that the Adsorpton process is multlayer adsorpton process. The Temkin binding energy (b T ) are positve for all the experiments indicatng that the process is endothermic. ## SUMMARY The batch adsorpton experiment on adsorpton of methylene blue and Congo Red on groundnut shell and sorghum husk was carried out. The experiment data needed for the diferent isotherm were generateds adsorpton isotherm such as Langmuirs Freundlich and Temkin models were used to ft the experiment data. Correlaton coefcient was used to see how well the models best fts the data. ## CONCLUSIONS For the adsorpton of methylene blue on sorghum husks Langmuir Isotherm is more

chemsum

{"title": "ADSORPTION ISOTHERM OF METHYLENE BLUE DYE ON GROUNDNUT SHELL AND SORGHUM HUSK KYAUTA SUNDAY, ERUGO IMMACULATE AMARACHI", "journal": "ChemRxiv"}

dynamic_role_of_the_correlation_effect_revealed_in_the_exceptionally_slow_autodetachment_rates_of_th

## Abstract: Real-time autodetachment dynamics of the loosely-bound excess electron from the vibrational Feshbach resonances of the dipole-bound states (DBS) of 4-bromophonoxide (4-BrPhO -) and 4-chlorophenoxide (4-ClPhO -) anions have been thoroughly investigated. The state-specific autodetachment rate measurements obtained by the picosecond time-resolved pump-probe method on the cryogenically cooled anions, exhibit the exceptionally long lifetime (τ) of ~ 2.5 ± 0.6 ns (as the upper bound) for the 11' 1 vibrational mode of the 4-BrPhO -DBS. Strong mode-dependency in the wide dynamic range has also been found, giving τ ~ 5.3 ps for the 10' 1 mode, for instance. Though it is nontrivial to get the state-specific rates for the 4-ClPhO -DBS, the average autodetachment lifetime of the 19' 1 20' 1 /11' 1 mode has been estimated to be ~ 548 ± 108 ps.Observation of these exceptionally slow autodetachment rates of vibrational Feshbach resonances strongly indicates that the 'correlation effect' may play a significant role in the DBS photodetachment dynamics. The Fermi's golden rule has been invoked so that the correlation effect is taken into account in the form of the interaction between the charge and the induced dipole where the latter is given by the polarizable counterparts of the electron-rich halogenated compound and the diffuse non-valence electron. This report suggests that one may measure, from the real-time autodetachment dynamics, the extent of the correlation effect contribution to the stabilization and/or dynamics of the excess non-valence electron among many different types of the long-range interactions of the DBS. Since firstly conceived by Fermi and Teller, 1 the dipole-bound state (DBS) of the anion, where the excess electron is loosely bound to the neutral core by the long-range monopole-dipole interaction, has been intensively investigated both experimentally and theoretically for many recent decades. The DBS is known to play an important role as the doorway state to the stable valence anion formation. Namely, as the slow electron approaches the neutral molecule or radical, the incoming electron is captured in the form of the Feshbach resonances by the long-range attractive interaction potential, and it is followed by the subsequent coupling and/or relaxation into the more stable anion species. 14,15 Detailed pictures of the whole processes of the electron-capturing, coupling, and relaxation are thus quite essential for the thorough understanding of the anion chemistry as well as the entry/exit dynamics of the redox reactions. The DBS has been found to be ubiquitous and identified in a number of chemical and biological systems to date. Notably, thanks to the combined techniques of the laser and cryogenically cooled ion-trap, 16 the understanding of the DBS has recently been enormously advanced in terms of the precise information of energetics, vibrational structures, and state-specific autodetachment (or relaxation) dynamics. 5,17 Regarding dynamics, however, the state-specific autodetachment rate of the DBS was directly measured only recently for the phenoxide (PhO -) anion, 18 allowing for the stringent comparison of the experiment with the theoretical prediction. The Fermi's golden rule has been found to be extremely useful to explain the strongly mode-dependent autodetachment rates although the prediction of the absolute values seems to be still quite challenging. 18 The lower threshold of the dipole moment for holding the excess electron was firstly proposed to be 1.625 D, 19 although it has been refined repeatedly after the correction of the Born-Oppenheimer approximation, 20,21 for instance. More practically, however, it seems to be widely accepted now, as a rule of thumb, that the DBS may exist when the dipole moment of the neutral core exceeds 2.5 D. 22,23 Although the long-range interaction of the dipole moment of the neutral core with the excess non-valence electron has been considered to be the most critical factor in the electron binding/unbinding dynamics, many theoretical works have suggested that the (especially dispersive) electron correlation effect should be largely responsible for the excess electron binding to the neutral core in the DBS. 4, The significant contribution of the electron correlation effect to the binding energy of DBS has been theoretically demonstrated from the quantum-mechanical calculations using the 2 nd -order Møller-Plesset perturbation (MP2), the coupled cluster singles and doubles (CCSD) theory, or the quantum Monte Carlo method. In the same context, it is notable that the critical value of the dipole-moment for the existence of the π-type DBS is still in dispute. Although the extent of the correlation effect is highly anticipated to be strongly dependent on individual chemical systems, the importance of the correlation effect in the DBS seems to be well received in the scientific community. Apparently, however, it is nontrivial to experimentally identify the correlation effect in terms of the static and/or dynamic role in the DBS. Even though the electron binding energy of the DBS is often expected to be proportional to the dipole-moment magnitude of the neutral core, it does not necessarily mean that the electron binding of the DBS is governed by the dipole moment only, as many different factors related to the long-range interaction potential could be cancelled out or added up depending on the chemical details. 34 For example, the smaller (or lager) binding energy does not necessarily mean the smaller (or larger) contribution of the correlation effect, and vice versa. In this aspect, we have here found that the correlation effect may be reflected in the dynamic property of the DBS rather than in the static binding energy. For example, two different DBS chemical systems of the similar binding energies could be quite different in terms of the extent of the correlation-effect. Herein, we argue that the autodetachment dynamics could reveal the nature of the electron binding in terms of the dynamic role of the correlation effect in the electron binding/unbinding dynamics of the DBS. We have investigated the picosecond (ps) timeresolved autodetachment dynamics of the DBS vibrational Feshbach resonances prepared by the one-photon photoexcitation of the cryogenically-cooled 4-bromophenoxide (4-BrPhO -) and 4chlorophenoxide (4-ClPhO -) anions. Exceptionally slow autodetachment dynamics observed in some vibrational Feshbach resonances of these anions have been analyzed from the new perspective that the correlation effect may play a significant role in the autodetachment dynamics. Photodetachment spectra of the cryogenically-cooled 4-BrPhOand 4-ClPhOanions taken by monitoring the total photoelectron signal as a function of the pump laser wavelength are shown in Figure 1. In both spectra, the stepwise increases of the photoelectron signal represent the electron-affinity (EA) thresholds. The rather sharp peaks are attributed to the vibrational Feshbach resonances of the DBS whereas the diffusive structureless background signal originates from the direct photodetachment of the anion. The overall structures of the photodetachment spectra are more or less identical to those obtained by the nanosecond (ns) laser pulse reported by the Wang group 22 except the broad bandwidths of the vibrational bands due to the intrinsic property of the picosecond (ps) laser pulse (ΔE ~ 20 cm -1 , Δt ~ 1.7 ps). The binding energy of the DBS has been precisely estimated to be 24 or 11 cm -1 for 4-BrPhOor 4-ClPhO -, respectively. 22 The zero-point energy (ZPE) level of the DBS is hardly identified, as the ps laser bandwidth is comparable to the electron binding energy for both anions. Instead, the 11' 1 /20' 1 30' 1 (Peak-I) and 11' 2 /10' 1 (Peak-II) bands could be well identified for 4-BrPhO -, whereas the 19' 1 20' 1 /11' 1 (Peak-III) The pump laser wavelength is tuned at the particular DBS vibrational band while the spatially overlapped non-resonant probe laser pulse (791 nm) is given at different delay times. At the zero delay-time, the DBS is most efficiently depopulated to give a spike 18 with the pump-probe crosscorrelation width of ~2.88 ps. With the increase of the pump-probe delay, the transient signal shows the apparent recovery (which is equivalent to the decay in the transients shown in Figure 2) due to the autodetachment process, giving the lifetime of the DBS Feshbach resonance from the exponential fit to the experiment. The Peak-I transient of 4-BrPhOshows the biexponential behavior with two distinct lifetimes, Figure 2. The faster decaying component gives the lifetime (τ) of ~ 13.5 ± 7.0 ps whereas the lifetime of the slow-decaying component is found to be extremely long, giving τ ~ 2.54 ± 0.60 ns (as the upper bound) with the relative amplitude ratio of 0.43 to 0.57, respectively. As the fundamental 11' 1 and 20' 1 30' 1 combinational modes are expected to be co-excited within the ps pump laser spectral window, two distinct lifetimes are ascribed to the autodetachment of two different vibrational modes. For the appropriate matches between the individual vibrational modes and their associated lifetimes, the velocity-map electron image taken at the pump wavelength of the DBS resonance is compared with that taken at the adjacent non-resonant (allowing just the direct photodetachment) pump wavelength. The image of the former minus that of the latter then gives the nature of the DBS band as the propensity rule of Δv = -1 is rather strictly obeyed. Accordingly, in the photoelectron spectrum taken from the peak-I, the -ν11 peak is the consequence from the autodetachment of the 11' 1 mode whereas the -ν20 or -ν30 peak results from that of the 20' 1 30' 1 combination band via the wobbling motion associated with the ν20 or ν30 mode, respectively. The relative contribution of the 11' 1 or 20' 1 30' 1 band is then, from the comparison of integrated photoelectron peak areas, estimated to be 0.64 or 0.36, respectively (Supporting Information). From the fact that the 11' 1 mode is more responsible for the peak-I, therefore, the exceptionally long lifetime of 2.54 ns should be due to the autodetachment from the 11' 1 mode whereas the faster decaying component (τ ~ 13.5 ps) is attributed to the autodetachment from the 20' 1 30' 1 combinational mode. While the autodetachment lifetime of 13.5 ps for the 20' 1 30' 1 mode sounds reasonable in terms of the order of magnitudes, the lifetime of 2.54 ns for the 11' 1 mode seems to be extraordinarily long for the vibrational autodetachment process. It should be noted though that the estimated lifetime of 2.54 ns has the somewhat large uncertainty due to the narrow temporal window (0 -1.8 ns) of the present experimental condition. Nonetheless, it is quite remarkable that the autodetachment rate of the 11' 1 mode of the 4-BrPhO -DBS is ~ 75 (or less) times slower compared to that of the 11' 1 mode of the PhOof which the lifetime has been measured to be ~ 33.5 ps. 18 Considering that the infrared intensity of the ν11 mode of the 4-BrPhO is only two times weaker than the ν11 mode of the PhO (vide infra), the retardation of the DBS autodetachment of the former compared to the latter by nearly two orders of magnitudes is quite exceptional. The similar analysis for the 11' 2 /10' 1 band (Peak-II) of 4-BrPhOhave been carried out (Supporting Information), giving τ ~ 82 or 5.3 ps for the 11' 2 or 10' 1 mode, respectively. The autodetachment rate of the 11' 2 mode of 4-BrPhO -(τ ~ 82 ps) is also quite slow compared to that of the 11' 2 mode of PhO -(τ ~ 12 ps), 18 supporting the experimental finding of the extremely slow autodetachment rate for the 11' 1 mode of 4-BrPhO -. In order to explain the experiment, we have invoked the Fermi's golden rule which has been widely used for the autodetachment rate. 𝑘𝑘 = 2π ℏ ��𝜙𝜙 𝑓𝑓 �𝑊𝑊�𝜙𝜙 𝑖𝑖 �� 2 𝜌𝜌(𝐾𝐾𝐾𝐾 𝑒𝑒 ) Eq.1 Eq.2 Here, 𝜙𝜙 𝑖𝑖 and 𝜙𝜙 𝑓𝑓 are the initial and final total wavefunctions, respectively, whereas 𝑣𝑣 𝑖𝑖 (𝑒𝑒 𝑖𝑖 ) or 𝑣𝑣 𝑓𝑓 (𝑒𝑒 𝑓𝑓 ) is the initial or final vibrational (electronic) wavefunction, respectively. ρ is the density of states which is the function of the electron kinetic energy (𝐾𝐾𝐾𝐾 𝑒𝑒 ). U is the charge-dipole interaction potential for the excess electron whereas Q is the normal mode coordinate associated with the particular vibrational mode. When the electron binding potential is confined to the interaction between the charge and permanent-dipole moment (µ0), F(Q) is proportional to the magnitude of the derivative of µ0 with respect to Q, 𝜕𝜕𝜇𝜇 0 𝜕𝜕𝜕𝜕 . As the infrared (IR) intensity is proportional to it is approximately regarded as the quantitative measure for the relative autodetachment rate of the corresponding vibrational mode. 35,39 Actually, the mode-dependent behavior of the autodetachment rate of the PhO -DBS could be quite successfully explained by the relative IR intensities of the individual vibrational modes as well as the Franck-Condon derivative factor for the overtone band. 18 In this regard, the nearly two-orders of magnitudes increase of the lifetime of the 11' 1 mode of 4-BrPhOcompared to that of the 11' 1 PhOmode cannot be explained by the simple application of the above conventional Fermi's golden rule, especially as the IR intensity of the former is only two times weaker than that of the latter (vide supra). It should be emphasized that the autodetachment rate is little influenced by the amount of the electron-binding energy. Rather, the loosely-bound electron is shaken off by the dynamic change of the interaction potential induced by the vibrational wobbling motion. 40 In this regard, one may invoke the aforementioned electron correlation effect into the autodetachment dynamics for the explanation of the large discrepancy of the experiment from the conventional Fermi's golden rule, especially as the electron-rich halogen atomic moiety is expected to be strongly correlated with the non-valence electron at the positive end of the dipole. Instead of the quantummechanical Hamiltonian, we have brought a simple physical model where the interaction potential in the Fermi's golden rule is modified to include the interaction between the charge and the (newly-added) induced dipole moment (𝜇𝜇 ⃑ 𝑖𝑖𝑖𝑖𝑖𝑖 ). The effective dipole moment (𝜇𝜇 ⃑ 𝑒𝑒𝑓𝑓𝑓𝑓 ) is then the sum of the permanent and the induced dipole-moments. The induced dipole moment can be expressed by the relation of 𝜇𝜇 ⃑ 𝑖𝑖𝑖𝑖𝑖𝑖 = 𝛼𝛼̈𝐾𝐾 �⃑ , where 𝛼𝛼̈ is the polarizability tensor of the neutral core whereas 𝐾𝐾 �⃑ is the local electric field given by the excess non-valence electron. This may belong to the same context with a recent report by the Wang group that the excess dipole-bound electron may play a significant role as the source of the intramolecular E-field. 41 The simple physical model proposed here, though the quantum-mechanical correlation effect would be much more sophisticated, may be at least conceptually consistent with the correlation effect as far as the electron-radical interaction is concerned, as the electrons of the neutral core are allowed to be dynamically correlated with the excess non-valence electron by the mediation of the polarizability and the local electric field. F(Q) could be then re-written as follows. Here, 𝐾𝐾 �⃑ is assumed to be independent of Q. At the equilibrium position, as 𝐾𝐾 �⃑ heads from the neutral core to the dipole-bound electron, the neutral core is polarized so that the oxygen moiety is negatively charged whereas the opposite-positioned bromine moiety should be positively charged according to 𝜇𝜇 ⃑ 𝑖𝑖𝑖𝑖𝑖𝑖 = 𝛼𝛼̈𝐾𝐾 �⃑ . The resultant induced-dipole, therefore, points the same direction as the permanent dipole (Figure 3a). And yet, the autodetachment process is not determined by the static property of the dipole. Rather, it is governed by the dynamic interplay between the instant changes of the permanent-and induced-dipoles with respect to the particular vibrational normal mode (Q). Accordingly, the directions of 𝜕𝜕𝜕𝜕 is positive (or negative) while the slope of 𝜕𝜕𝛼𝛼∂𝜕 𝜕𝜕 𝐾𝐾 �⃑ is negative (or positive), then the magnitude of the vector sum diminishes to give the decrease of the autodetachment rate. In this case, the autodetachment should be retarded due to the correlation effect. It should be noted that, when the dipole-bound electron is regarded as a point-charge lying on the molecule-fixed z-axis (Figures 1 and 3), all the in-plane vibrational modes of PhOor 4-BrPhOend up with the instant changes of dipole moments along the z-axis. In order to testify the physical model, we have calculated the for the ν11 mode of 4-BrPhO. Apparently, the latter is the consequence from the reduction of the permanent dipole moment of 4-BrPhO with the positive displacement of the ν11 mode whereas the polarizability along the z-axis instantly increases by the same displacement (Figure 3). Substitution of the electronegative Br atom on the para position should be responsible for the opposite behavior of the dipole-moment change with ν11, compared to that of PhO. Therefore, it is most likely that the correlation effect embodied in the charge-induced dipole interaction should impede the autodetachment of the 11' 1 mode of 4-BrPhOwhereas it expedites that of the 11' 1 mode of PhO -. Though the quantitative comparison is nontrivial, it gives the rational explanation why the autodetachment rate could be exceptionally slow for the 11' 1 mode of 4-BrPhO -. Notably, the magnitude of 𝜕𝜕α 𝜕𝜕𝜕𝜕 could be larger for 4-BrPhOcompared to that of PhObecause of the lager polarizability of the former than the latter although the more sophisticated calculation is highly desirable (Supporting Information). The fast autodetachment rate (τ ~ 13.5 ps) observed for the 20' 1 30' 1 mode of 4-BrPhOis probably due to the cooperation effect of the combination mode in the wobbling motion as demonstrated previously for PhO -(Supporting Information). 18 Regarding the 11' 2 overtone mode of 4-BrPhO -, its lifetime of 82 ps is much longer than the lifetime of 12 ps measured for the 11' 2 mode of PhO -. According to the derivative Franck-Condon factor in Eq. ( 2), the autodetachment rate of the overtone mode is anticipated to be ~ 4 times faster than that of the fundamental mode. 18 In that sense, if the lifetime of the 11' 1 mode of 4-BrPhOis taken to be 2.54 ns, the lifetime of ~ 600 ps is expected for the 11' 2 mode. In the same context, if the lifetime of 82 ps is taken for the 11' 2 mode, the autodetachment lifetime of the 11' 1 mode is expected to be ~ 330 ps, which is already quite long for the vibrational autodetachment lifetime. Therefore, although the lifetime measurement of 2.54 ns is subject to the further refinement, it seems to be quite certain that the autodetachment rate of the 11' 1 mode of 4-BrPhOis exceptionally slow. The fast autodetachment rate of the 10' 1 mode of 4-BrPhOwith τ ~ 5.3 ps is mainly attributed to the much stronger IR intensity of the ν10 mode which is ~ 30 times larger than that of ν11, although the Fermi's golden rule could not give the quantitative explanation of the experiment (Supporting Information). It is interesting to note that the magnitude of 𝜕𝜕α 𝜕𝜕𝜕𝜕 is much smaller than that of 𝜕𝜕μ 0 𝜕𝜕𝜕𝜕 for the 10' 1 mode of 4-BrPhO -, Figure 3, suggesting that the dynamics of the corresponding mode is little influenced by the correlation effect. Similar analysis has also been carried out for 4-ClPhO -. For the Peak-III in Figure 1, the 19' 1 20' 1 /11' 1 DBS band undergoes the autodetachment process via -ν19 (or -ν20) from the 19' 1 20' 1 combination mode whereas the autodetachment from the 11' 1 mode is responsible for the -ν11 peak. The relative ratio of the former to the latter in the Peak-III is estimated to be 0.86:0.14. The peak-IV of 4-ClPhO -(Figure 1) is assigned to the 11' 1 19' 1 20' 1 /11' 2 band. In the photoelectron spectrum, the photoelectron peak populated by the autodetachment via the ν20 or ν19 mode is found to be quite small. Similar to the case of Peak-III, the relative contribution of the 11' 1 19' 1 20' 1 and 11' 2 mode to the Peak-IV is estimated to be 0.86:0.14. Unfortunately, however, it turns out to be nontrivial to extract two different lifetimes from the transient of Peak-III or Peak-IV by the biexponential fit, mainly due to the relatively poor S/N ratio, Figure 2. Instead, the single exponential fit to the experiment give the averaged autodetachment lifetime of ~ 548 ± 108 or 50.0 ± 8.5 ps for the Peak-III or Peak-IV, respectively. Overall, the autodetachment rate of the 4-ClPhO -DBS is also estimated to be quite slow compared to that of PhO -, indicating that the correlation effect on the electron-binding dynamics could also be quite significant in 4-ClPhO -. It should be emphasized that the autodetachment dynamics is expected to be strongly dependent on the individual chemical systems, and thus the theoretical analyses for individual chemical systems should be carried out case by case. Conversion) for the pump laser pulse, while the other half of the fundamental was used as the probe laser pulse. Time-resolved photoelectron images were obtained by scanning the delay times between the pump and probe pulses using a couple of retro-reflectors on the optical delay stage (DDS220, Thorlabs). Photoelectron images were reconstructed by the BASEX 43 or polar onion peeling (POP) programs. 44

chemsum

{"title": "Dynamic Role of the Correlation Effect Revealed in the Exceptionally Slow Autodetachment Rates of the Vibrational Feshbach Resonances in the Dipole-Bound State", "journal": "ChemRxiv"}

protective_effect_of_starch-stabilized_selenium_nanoparticles_against_melamine-induced_hepato-renal_

## Abstract: Melamine and its analogues are illegally added to raise the apparent protein content in foods. The elevated concentrations of these compounds cause adverse effects in humans and animals. In this contribution, the protective effects of the synthesized starch-stabilized selenium nanoparticles (Se-NPs@starch) on melamine-induced hepato-renal toxicity have been systematically investigated.The Se-NPs@starch were characterized by X-ray photoelectron spectroscopy (XPS) analysis, energy dispersive spectroscopy (EDS) mapping analysis, TEM, and FT-IR. Starch plays a crucial role in the stabilization and dispersion of Se NPs, as noticed from the TEM and EDS investigations.Furthermore, the atomic ratio of Se distribution over the starch surface is approximately 1.67%.The current study was conducted on four groups of adult male rats, and the oral daily treatments for 28 days were as follows: group I served as control, group II received Se-NPs@starch, group III was exposed to melamine, while group IV was treated with melamine and Se-NPs@starch. The results reveal a significant alteration in the histoarchitecture of both hepatic and renal tissues induced by melamine. Furthermore, elevated liver and kidney function markers, high malondialdehyde, and increased expression levels of apoptosis-related genes besides a reduction in GSH and expression levels of antioxidant genes were observed in the melamine-exposed group.Interestingly, the administration of the Se-NPs@starch resulted in remarkable protection of rats against melamine-induced toxicity through increasing the antioxidant capacity and inhibiting oxidative damage. Collectively, this study provides affordable starch-stabilized Se-NPs with potent biological activity, making them auspicious candidates for prospective biomedical applications. ## Introduction Melamine is frequently used in different products as furniture, laminates, plastics, food utensils, coatings, glues, commercial filters, and dining ware . Melamine intake at very low concentrations is harmless for animals because it is hardly metabolized by animals and rapidly eliminated via urine (more than 90%) . Furthermore, World Health Organization (WHO) and U.S. Food & Drug Administration (USFDA) determined the safe level of melamine concentration in human food materials to be 2.5 mg/kg . However, melamine and its analogues, such as nitrogen-rich triazine compounds, are illegally added to various foodstuffs to falsify the protein contents, resulting in fatal adverse effects in children and pets . Recently, exposed dairy products containing melamine have raised concerns regarding its toxicity . For example, nephrotoxicity was noticed among the children who ingested milk-infant formula contaminated with melamine . Unfortunately, this outbreak was accompanied by urinary stones in both infants and children. As a consequence, 294,000 of them were influenced with more than 50,000 hospitalized and 6 died due to acute renal failure . Besides, renal failure and death were reported in domesticated dogs and cats after exposure to pet food contaminated with melamine and its analogs . Melamine can accumulate in the brain, liver, spleen, and bladder as an effect of oral administration over time . Elevated concentrations of melamine in these organs induce adverse effects on the growth of fetuses and neonates . Likewise, ingestion of melamine leads to sperm abnormalities and DNA damage , in addition to detrimental impacts on the male reproductive system . Testicular damage is observed to be associated with oxidative stress . Oxidative stress is a pathophysiological phenomenon that emerges from the imbalance between the formation of reactive oxygen species (ROS) and the anti-oxidative efficiency of the enzymatic and nonenzymatic antioxidants . Excessive ROS can cause alterations in the cellular macromolecules including lipids, proteins, and DNA, followed by a disruption of the cell wall, cellular enzymes inactivation, and eventually cell death . On that front, substantial efforts have been devoted to inhibiting melamine toxicity . Metal nanoparticles are of utmost importance in medicine and pharmacology due to their fascinating properties such as chemical stability, non-toxicity, and biocompatibility . Among them, selenium (Se) is a fundamental trace element required for the regular physiological function of growing animals and is obtained from plants. Besides, Se is very important to human health because of its potent pro-oxidant, and antioxidant effects, anti-inflammatory, and immunity-boosting capabilities . Remarkably, many Se compounds are reported to have significant anti-cancer activity and chemopreventive properties . More outstandingly, Se NPs show an antiviral effect against the current pandemic of SARS-CoV-2 (coronavirus disease 2019), which has been confirmed using Ebselen . However, there is a very narrow boundary between the acceptable concentrations of Se intake and its toxicity . Generally, the pro-oxidant and antioxidant effects, bioavailability, and toxicity of Se are strongly affected by its chemical form, particle size, and concentration. These properties are crucial to determine the interaction of Se with biological entities . For example, the toxicity of Se is remarkably inhibited in the nano-size zerovalent form. Moreover, Se-NPs with particles size of 5-200 nm can efficiently scavenge free radicals such as superoxide anion and 1,1-diphenyl-2-picrylhydrazyl (DPPH) . While selenomethionine is a very safe natural source of Se with high bioavailability, lower toxicity 5 accompanied with a similar ability to increase selenoenzyme levels were detected for Se-NPs of average particle size of 36 nm . Zhang et al. showed that the toxicity of Se-NPs in mice is seven times lower than that of sodium selenite and three times lower than that of organic Se compounds . Additionally, nanoparticle applications in the fisheries and livestock world revealed that Se-NPs enhance the efficacy of growth, digestion, immunomodulation, and reproduction as well as increase the productivity of stress-ridden fish and livestock . Despite the fascinating properties of NPs, they suffer from aggregation in suspensions due to their high surface energy, consequently, their activity sharply declines as a function of time . Hence, to control the dispersion and size of Se-NPs, natural polysaccharides like chitosan , and cellulose were used as natural stabilizers and size controlling agents. One should emphasize that the shelf-life storage of the stabilized metal nanoparticles via the immobilization approach is extremely long and can exceed 5 years . In continuation of our previous work on Se-NPs and their interesting aspects , starch-stabilized Se-NPs (Se-NPs@starch) were synthesized and characterized by XPS, FT-IR, and TEM. The potential activity of the synthesized starch-stabilized Se-NPs was evaluated for inhibiting the toxicity of hepatic and renal tissues induced by melamine ingestion. The experiments were conducted on four groups of male albino rats including group I (control), group II received starchstabilized Se-NPs, group III received melamine, whereas group IV received both melamine and starch-stabilized Se-NPs. Statistical analysis of the results is also provided. ## Chemicals and reagents Melamine (C3H6N6, LOBA Chemie, India) was obtained from El-Mekkawy Company, Cairo. Selenium (Sigma-Aldrich, CAS 7782-49-2 034-001-00-2), potato starch (CAS Number: 9005-25-8), ascorbic acid (CAS Number: 50-81-7), and nitric acid (HNO3, CAS Number 7697-37-2) were of high-purity grade from Sigma-Aldrich. ## Selenous acid preparation 0.050 M of selenous acid (H2SnO3) was prepared by dissolving 0.987 g of Se metal in concentrated HNO3 before heating till dryness and dissolved in 250 mL distilled water . ## Synthesis of selenium nanoparticles impregnated on starch (Se-NPs@starch) Se-NPs were synthesized by a versatile and green procedure using starch as a stabilizer, and ascorbic acid as a reducing agent that reduces Se (IV) to Se. Se-NPs are stabilized by starch according to the literature but with replacing cellulose with starch . In detail, the selenous acid/starch aqueous solution was prepared as follows: 50.0 g potato starch was boiled in 100 mL water and then mixed with 250 mL of 0.05 M selenous acid before completing the total volume to 800 mL. Afterward, 100 mL of 0.20 M ascorbic acid solution was added dropwise into the selenous acid/starch solution to start the reduction reaction. After the addition of ascorbic acid, the solution changed from transparent to red, indicating the formation of Se-NPs. The obtained crimson red solution was concentrated to less than 100 mL using a rotary evaporator at 80 °C. A crimson red gel is formed, which is dried at 80 °C in an oven to produce shiny crimson red particles (Se-NPs@starch). The synthesized Se-NPs@starch was characterized by X-ray photoelectron spectroscopy (XPS), Fourier-transform infrared (FTIR), and Transmission Electron Microscope (TEM). ## Instrumentation To investigate the surface chemical composition and oxidation state of Se, X-ray photoelectron spectroscopy (XPS) analysis was conducted using a UHV Multiprobe system (Scienta Omicron, Germany) with a monochromatic X-ray source (Al Kα) and an electron analyzer (Argus CU) with 0.60 eV energy resolution. Fourier-transform infrared (FTIR) spectra were recorded on a Nicolet iS10, Thermo-Fisher Scientific, USA, using a KBr pellet. The TEM images of Se-NPs@starch were obtained by a TEM of the Se-NPs@starch using a JEOL model 1200EX electron microscope at an operating voltage of 120 kV. ## Experimental animals and ethical approval Adult male albino rats (n= 40, average B.W= 200±20 g) were kept in plastic cages under the standard hygienic conditions (60% relative humidity, 24 ± 3°C room temperature, and 12: 12-h light: dark cycle with ad libitum access to food and drinking water). Rats were treated humanely according to NIH guidelines, and the experimental procedure was accepted by the Institutional Animal Care and Use Committee (IACUC) of the Faculty of Veterinary Medicine, Cairo University (Approval number Vet CU28042021291). ## Experimental design Rats were acclimated for one week, then they were assigned to 4 equal groups of ten each (n=5 rats/cage); group I (untreated control group) only received distilled water, Group II: Se-NPs@starch treated group (2 mg Se-NPs/kg) as previously mentioned by Rezvanfar et al. , group III: melamine-treated group (300 mg/kg) as previously reported by An et al. , and group IV: Se-NPs@starch co-treated group (received melamine (300 mg/kg) plus Se-NPs (2 mg/kg)). All treatments were received daily by stomach tube for 28 days. ## Sample collection and preparation After 28 days of treatment, serum samples were obtained from rats for measuring the function of both liver and kidney. Then, rats were euthanized by cervical dislocation for collection of the liver and kidney samples. Some specimens of these organs were stored at -80 °C for estimation of oxidative stress biomarkers in their tissue homogenates and quantitative real-time PCR (Rt-PCR) analysis for some antioxidant-and some apoptotic-related genes, while other specimens were histopathologically and immunohistochemically investigated after fixation in 10% neutralbuffered formalin (NBF) solution. ## Determination of Liver Function Markers The activities of both serum aspartate-and alanine-aminotransferase (AST & ALT) were assayed using reagent kits (Spectrum Diagnostics Co., Egypt) following the provided instructions. 9 ## Determination of kidney Function Markers The levels of serum creatinine and urea were assayed using reagent kits (Spectrum Diagnostics Co.) following the provided instructions. ## Hepatic and renal oxidative stress biomarkers A Teflon tissue homogenizer was used to homogenize both liver and kidney samples in 10% (W/V) ice-cold 0.1 M phosphate buffer saline (PBS) (pH 7.4). Crude tissue homogenate was centrifuged at 15,000 rpm for 15 min at 4 ο C and subsequently used for determination of reduced glutathione (GSH) content as previously mentioned by Ellman , malondialdehyde (MDA) content according to Ohkawa et al. , in addition to total protein concentration as previously reported by Bradford . ## qRT-PCR analysis for Nrf-2, GPX, c-Myc, and CASP 3 genes The relative hepatic and renal Nrf-2, GPX, c-Myc, and CASP 3 mRNA abundance was determined by qRT-PCR analysis using GAPDH as a housekeeping gene. Total RNA was extracted from approximately 100 mg liver and kidney tissues using the total RNA Extraction Kit (Vivantis, Malaysia). By using M-MuLV Reverse Transcriptase (NEB#M0253). RT-PCR was carried out after confirming the concentration and purity of RNA. By fluorescence-based real-time detection method with a fluorescent SYBR Green dye (Thermo Scientific, Cat. No. K0221), a quantitative assessment of cDNA amplification for each gene was performed. The primer sequence used for RT-PCR analysis is shown in Table 1. The real-time PCR conditions were carried out as follow: 95 ο C for 5 min (initial denaturation) and then 40 cycles at 95 ο C for 15 s, 60 ο C for 30 s, and 72 ο C for 30 s in each experiment. Negative controls that were free of the template were included. 10 Each qRT-PCR was performed with three biological replicates, and each biological replicate was assessed three times. Using the comparative 2 −ΔΔCT method, the relative transcription levels were calculated . Table 1. Primer sequence used for qRT-PCR. ## Gene symbol Gene description Accession number Primer Sequence Both liver and kidney specimens collected from the four groups were immediately fixed for 48-72 h in 10% NBF, embedded in increasing grades of alcohol. Then, tissue specimens were cleared with xylene and fixed in paraffin. 3-4 μm paraffin sections were processed for staining with hematoxylin and eosin (H&E) for histopathological investigation as described by Bancroft and Gamble . All the noticeable histopathological alterations in the liver and kidneys were graded using a classical semiquantitative scoring system to assess the degree of lesion severity between different groups (n= 5 slides representing 5 rats per group) as follows: (0) none (normal histology) while (1), ( 2), ( 3) and ( 4) indicate mild <25%, moderate 25%-50%, severe 50%-75%, and extensive severe >75% tissue damage, respectively . The main histopathological parameters used for liver injury were congestion, vacuolar degeneration, and inflammation while, those for renal damage were tubular degeneration and necrosis, glomerular atrophy, congestion, and hyalinization. ## Immunohistochemical analysis Immunohistochemistry was carried out on deparaffinized hepatic and renal sections (4-μm thick) for detection of apoptosis according to the manufacturer's protocol. Sections were immersed in 0.3% hydrogen peroxide (H2O2) in phosphate-buffered saline (PBS) for 20 min for deactivation of endogenous peroxidase, incubated with rabbit polyclonal anti-caspase-3 antibodies (active/cleaved) (100-56113, Novus biologicals) 1:100 for 1hr then washed out. After that, tissue sections were incubated for 20 min with secondary antibody Horse Radish peroxidase (HRP) Envision Kit (DAKO), washed out, and incubated for 10-15 min with diaminobenzidine (DAB). Sections were washed, counterstained in hematoxylin, dehydrated in alcohol, cleared with xylene, and covered with a coverslip for microscopic examination. Caspase 3 -stained liver and kidney sections were assessed using Leica Quin 500 software (Leica Microsystems, Switzerland) for morphological analysis. Caspase 3 immunostaining was quantified as area percentage in different slides (n=5 fields/group) at magnification power x400. The areas displaying brown color immunoreaction (positive) were selected for estimation. Mean values and standard error mean (SEM) of each specimen were obtained and statistically analyzed. ## Statistical analysis Descriptive statistics were presented as mean ± SEM. The obtained results were subjected to oneway analysis of variance (ANOVA) using SPSS version 25.0 software (IBM, USA) followed by Tukey post hoc test. Data were considered significantly different at P-value less than 0.05. ## Characterization XPS investigations were performed to identify the surface chemical composition of Se-NPs@starch and the valence state of Se. Fig. 1a shows the XPS spectra (survey) of the Se-NPs@starch sample. From the survey spectra, the peaks corresponding to Se, O, and C are observed. The high-resolution spectrum of Se implies the peaks of Se 3d5/2 and Se 3d2/3 at approximately 55.6 and 56.6 eV, respectively confirming the presence of elemental selenium (Se 0 ), as shown in Fig. 1b. Furthermore, the absence of Se 2p, 2s, and 1s peaks indicates that Se (IV) was entirely reduced during the synthesis of the Se-NPs@starch . Besides Se 0 , the sample contained organic material on the surface as revealed by the presence of C, and O (Fig. 1c). The C 1s peak in the XPS spectrum of starch has been deconvoluted into three peaks at 284.80, 286.4, and 288.7 eV, as presented in Fig. 1c . ## Liver and kidney functions Hepatic damage was estimated by the determination of ALT and AST activities. Renal damage was estimated by the determination of serum urea and creatinine levels. Fig. 4B indicates that melamine significantly increased the serum urea level from 14.84 to 22.48 mg/dL and creatinine level from 0.75 to 1.50 mg/dL compared to the control group. Co-administration of Se-NPs@starch did not significantly decrease the serum urea level to 18.96 mg/dL and significantly decreases the creatinine level to 0.84 mg /dL in contrast to the melamine-intoxicated group. ## Hepatic and renal oxidative stress biomarkers For studying the redox state of the cell, MDA (LPO biomarker) and some antioxidant machinery were evaluated. ## Hepatic and renal MDA Content Based on the results presented in Fig. 5, melamine markedly (p˂ 0.05) increased the hepatic and renal MDA level from 1.53 to 5.22 μΜ/mg and from 1.08 to 3.50 μΜ/mg protein, respectively, compared to the control group. Co-administration with Se-NPs@starch significantly declined both liver and kidney MDA contents to 2.30 μΜ/mg protein and 2.10 μΜ/mg protein, respectively, compared to the melamine-intoxicated group. ## Antioxidant Machinery Antioxidant Machinery was assessed by the determination of GSH content and mRNA relative expression for Nrf-2 and GPx genes. ## Hepatic and renal GSH content Fig. 6 shows that melamine considerably depresses the hepatic and renal GSH content from 3.83 to 2.47 μΜ mg -1 and from 3.62 to 2.31 μΜ mg -1 protein, respectively, compared to the control group. Co-administration with Se-NPs@starch noticeably increased both liver and kidney GSH contents to 3.01μΜ/mg protein and 3.03 μΜ/mg protein, respectively compared to the melamineintoxicated group. For the melamine-intoxicated group, the Nrf-2 gene was substantially downregulated to 0.28-fold and 0.16-fold compared to the control one in both liver and kidney tissues, respectively. While coadministration with Se-NPs@starch modulated the gene expression level to 0.71-fold in the liver and 0.40-fold in the kidney. Such alteration in the gene expression was significant when compared to the melamine-intoxicated group (Fig. 7). Furthermore, after exposure to melamine, the GPX gene showed a considerable downregulation in both hepatic and renal tissues compared to the control one. Se-NPs@starch co-administration extensively elevated the expression level of the GPX gene from 0.2 to 0.7-fold and from 0.2 to 0.82-fold for the liver and kidney tissues, respectively compared to the melamine-intoxicated group as shown in Fig. 7. ## Histopathological investigations The microscopical examination of the hepatic sections obtained from the control group revealed a normal histological architecture of the liver that appeared with normal radiating cords of hepatocytes, normal central vein, and blood sinusoids (Fig. 9a). The Se-NPs@starch exposed group exhibited normal hepatocytes that appeared polyhedral with normal vesicular and central nuclei as well as arranged in hepatic cords that were radiating from the central vein. The hepatic cords were separated by normal blood sinusoids. In addition, the portal triad showed the normal histological structure of the bile duct and branches of the portal vein and hepatic artery (Fig. 9b and 9c). The hepatic tissues of the melamine-exposed rats had severe damage that was observed in the form of distortion of the hepatic cord's arrangement, edema, severe congestion, and dilatation of the central vein. Blood sinusoids appeared severely dilated with clearly observable activated Von Kupffer cells. Furthermore, hepatocytes displayed severe ballooning and remarkable cytoplasmic vacuolization as well as different stages of degeneration and apoptosis. The dark brown deposition was also noticed inside the cytoplasm of hepatocytes. The portal area was disrupted, and the lining epithelium of the bile duct revealed distortion, in addition to inflammatory cells infiltration and severe congestion of the portal vein, as shown in Fig. 9d-j. On the other hand, the melamine-exposed group co-treated with Se-NPs@starch exhibited a partial recovery of the hepatic damage. The central vein appeared less congested with normal dilatation, while the blood sinusoids appeared with normal architecture. Some hepatocytes showed degeneration and ballooning with cytoplasmic vacuolization, but these alterations in the hepatic architecture were less in severity (Fig. 9k, l) in comparison to the melamine-exposed group. The microscopical examination of the renal tissues for the control group and Se-NPs@starch exposed group revealed a normal histological structure of the renal cortex; the renal corpuscle contained normal glomerular tuft with plentiful capsular space enveloped by intact Bowman's capsule in addition to intact tubular epithelial cells of both proximal and distal convoluted tubules (PCT & DCT) as illustrated in Fig. 10a and 10b, respectively. While the renal tissue sections from the melamine-exposed group exhibited various histopathological alterations in the renal corpuscles, glomeruli, PCT, DCT, and renal vasculature (Fig. 10c-g). Some renal corpuscles revealed widening of Bowman's capsule (Fig. 10c-e) and widening of capsular space (Fig. 10c, d), while others showed distortion and shrinkage of Bowman's capsule (Fig. 10f, g) with a narrowed capsular space (Fig. 10e). Moreover, some renal corpuscles revealed loss of glomeruli with a decreased number (hypocellularity) of mesangial cells (Fig. 10c), while others showed a complete loss of cells (Fig. 10d). Peri-glomerular edema was also plentiful (Fig. 10d). Congestion of glomerular capillary tuft (Fig. 10c-g) and diminished glomeruli with a prominent nuclear condensation were also detectable (Fig. 10f, g). In addition, degeneration, and necrosis of PCT and DCT with desquamated epithelial cells were observed (Fig. 10d-g). Also, some parts of DCT were lined by squamous cells and possess vacuolization and desquamation of other lining cells with pyknosis (Fig. 10d). Severe interstitial (Fig. 10c) and vascular (Fig. 10g) congestion in addition to hyalinization (Fig. 10f, g) were noticed. On the other hand, the melamine-exposed group co-treated with Se-NPs@starch demonstrated a remarkable improvement in the histological architecture. The renal cortex was nearly normal with less degeneration, congestion, and hyalinization (Fig. 10h). Moreover, the sections of renal medulla obtained from the control group and Se-NPs@starch exposed group revealed a normal histological structure of collecting tubules as observed in Fig 10i &10j, respectively, while vacuolar degeneration of collecting tubules, hyalinization, and congestion of interstitial blood capillaries were prominent in those obtained from the melamineexposed group (Fig. 10k). In contrast, a marked improvement of renal medulla with less degenerated collecting tubules was observed in the melamine-exposed group co-treated with Se-NPs@starch (Fig. 10l). and hyalinization (yellow chevron) are observed. h: The melamine-exposed rats co-treated with Se-NPs@starch show restoration of most renal corpuscles, glomeruli (G), and renal tubules (yellow arrow), while some renal tubules still show degeneration (red arrow). i: Renal medulla of the control group reveals the normal structure of collecting tubules. j: The Se-NPs@starch exposed group exhibits nearly normal collecting tubules of the medulla. k: The melamine-exposed group shows a prominent vacuolar degeneration of collecting tubules (yellow arrow), hyalinization (yellow chevron), and congestion of interstitial blood capillaries (yellow star). l: The melamineexposed group co-treated with Se-NPs@starch demonstrates a partial recovery of renal medulla with less degenerated collecting tubules (yellow chevron). Also, some collecting tubules appear with squamous lining cells (yellow arrow). Scoring criteria for the observed histopathological changes in the hepatic and renal architectures of different groups are presented in Fig. 11. The highest score in all histopathological parameters was noticed in the melamine-exposed group, while the SeNPs co-treated group revealed a substantial reduction in the elevated histopathological score induced by melamine. ## Immunohistochemical finding According to our immunohistochemical findings in Fig. 12, liver sections obtained from both control and Se-NPs@starch exposed rats exhibited a negative reaction to caspase 3, while a strong immuno-expression was noticed in the melamine-exposed one. However, the liver sections obtained from the group exposed to melamine and co-treated with Se-NPs@starch indicated a mild expression of caspase 3 (Fig. 12). 28 Furthermore, the renal tissues of both control and Se-NPs@starch exposed groups displayed a mild positive immunoreactivity to caspase 3. In contrast, the melamine-exposed group revealed a strong positive immune reaction. However, a moderate immunoreactivity was noticed in the melamineintoxicated group co-treated with Se-NPs@starch (Fig. 13). 29 Moderate expression in the melamine-exposed group co-treated with Se-NPs@starch. According to the data analysis, the melamine-intoxicated group demonstrated a remarkable elevation (p˂ 0.05) in the area% covered by caspase 3 in both hepatic and renal tissues compared to the control and Se-NPs@starch exposed group. On the other hand, the concurrent administration of Se-NPs@starch significantly (p˂ 0.05) reduced the caspase 3 area% induced by melamine in both hepatic and renal tissues by 87.92% and 65.9%, respectively, compared to the melamineintoxicated group (Fig. 14). ## Discussion The kidney is the major target organ in melamine toxicity, but there are some other tissues such as the liver, muscles, colon, and spleen that were investigated for melamine toxicity . This study has been focused on testing the ability of oral administration of Se-NPs@starch to attenuate melamine-induced liver and kidney dysfunction. Therefore, the serum AST and ALT activities were assessed as the principal hallmarks of liver toxicity . According to the present investigation, melamine induces severe hepatic damage as represented by markedly elevated serum functions of AST and ALT which further evidenced by severe histopathological alternations in the normal structure of hepatic tissues as shown in Fig. 9 (d-j) that indicated severe damage of the cellular membrane followed by leakage of hepatic enzymes into the bloodstream . These results are in good agreement with those reported recently by Abd-Elhakim et al. . However, treatment with Se-NPs@starch in group IV markedly decreased the serum activities of both ALT and AST, in addition to improving liver functions, which are highly indicative of the hepatoprotective effect of Se-NPs@starch on melamine-induced hepatic impairment. These findings are consistent with Bai et al. and Sohrabi et al. . Additionally, the ameliorative effects of Se-NPs@starch are also confirmed by the less hepatocellular damage observed in the histopathological examination of liver tissues as shown in Fig. 9 (k & l). These results are in good agreement with those reported by Amin et al., Bai et al.,and Hamza et al. . These outcomes may be attributed to maintaining the hepatocyte's integrity or regeneration of damaged hepatocytes . Furthermore, urea is considered the initial acute renal biomarker that elevates by any type of renal injury. While creatinine only elevates when most of the kidney function is lost, therefore, it is considered the most reliable renal indicator . Consequently, any significant elevation in the serum urea and creatinine levels can indicate kidney damage. Our results reveal that melamine-intoxicated rats exhibited a significant elevation in the serum creatinine and urea levels, indicating melamine-induced renal glomerular impairment. These findings are consistent with those of Al-Seeni et al. and Abd-Elhakim et al. . Moreover, the renal damage induced by melamine is indicated histopathologically by severe deterioration of renal architecture as shown in Fig. 10 (c-g & k) which was also observed by Lee et al. , Abd-Elhakim et al. , Lee et al. , Yasui et al. and Peerakietkhajorn et al. . In contrast, the concomitant administration of Se-NPs@starch with melamine considerably decreases the serum urea and creatinine levels. Moreover, administration of Se-NPs@starch substantially restores the melamineinduced histological alterations to the kidney tissues as observed in Fig. 10 (h & l) supporting the nephroprotective property of Se-NPs@starch against melamine toxicity. These results agree with those obtained by others . The nephroprotective efficiency role of the Se-NPs@starch 32 could be due to its potent action in scavenging free radicals , thus protect the cellular structure from oxidative damage . Based on the obtained findings, oxidative stress and apoptosis are involved in the pathophysiology of the hepato-renal toxicity induced by melamine. Here, the levels of lipid peroxidation marker and GSH, in addition to mRNA expression level for GPx and Nrf2 were assessed to evaluate the level of oxidative stress. Compared to the control, the MDA level in both tissue homogenates was strongly elevated, while the content of GSH and the mRNA expression level for GPx was sharply reduced in the melamine-intoxicated rats. Therefore, our results suggest that melamine elevates the susceptibility of tissue to oxidative damage via increasing oxidative stress and reduction of endogenous antioxidant capacity in both tissues. Our findings coincide with the previous results stated by Al-Seeni et al. and Abd-Elhakim et al. . A marked elevation of MDA concentration reflected the severity of cell damage resulted from the oxidation of unsaturated fatty acids. Moreover, a significant depletion in antioxidant enzymes (GPx and GSH) reflects the impaired antioxidant defense mechanism to counteract the elevated levels of free radicals . The molecular mechanism underlying the hepato-renal damage induced by melamine is not entirely clear. According to our results, melamine administration induced hepato-renal oxidative damage as substantiated by marked downregulation for Nrf2 mRNA gene expression. As known, Nrf2 is a key transcription factor that has a critical contributor role to protect cells from damage induced by inflammation and oxidative stress. Translocation of Nrf2, from the cytoplasm to the nucleus, was necessary for its regulation of antioxidant/detoxification enzyme expression . As observed in our study, melamine downregulates the Nrf2 mRNA level. This is an indicator of an oxidative stress response. On the other hand, Se-NPs@starch co-administered group (group IV) significantly reduced the level of MDA, stabilized nonenzymatic antioxidant GSH level, and up-regulated antioxidant enzyme (GPx) mRNA expression level in the liver and kidneys. Thus, these findings suggest an ameliorating effect of Se-NPs@starch against melamine-induced excessive production of ROS. Se-NPs@starch administration results in antioxidative and nephron-protective effects as was reported by Khater et al. . Furthermore, the antioxidant activity of Se-NPs@starch was reported by Bai et al. and Sheiha et al. for liver tissue and these findings support our results. Moreover, selenium can combat oxidative stress, consequently leading to the cellular redox balance, due to its incorporation as selenocysteine into GPx and thioredoxin reductase . Se-NPs@starch lead to an increase in the activities of both GPx and glutathione S-transferase, resulting in less oxidative stress . In addition, Se-NPs@starch detoxify hydroperoxidase and lipid peroxides that accumulate in the cytoplasm and mitochondria. Moreover, Se improves the antioxidant enzyme capacity, thereby leading to cellular protection from oxidative damage . Khalaf et al. and Rashad et al. verified the antioxidant effect of Se-NPs@starch. Furthermore, most recent studies link the potent antioxidant effects of Se with the activation of the Nrf2 factor. Se can upregulate the transcription of Nrf2 against metal intoxication , as observed in our result. The elevated levels of oxidative stress may hinder the pathway of Nrf2 and thereby enhancing the deleterious effects induced by toxic agents . Zhang et al. reported that selenium triggered the Nrf2 activation and thereby upregulating the transcription of many genes like glutathione S-transferase . Concerning cell death, a series of physiological symptoms is initiated as a result of several biochemical lesions induced by the rapid reaction of free radicals with cellular elements that eventually lead to apoptosis. Therefore, our study strongly recommends using Se-NPs@starch as a potent antioxidant to diminish the apoptosis induced by melamine intoxication in both hepatic and renal tissues. In 34 the current study, intoxication with melamine significantly induces overexpression of the mRNA expression level of both apoptotic genes; caspase-3 and c-Myc. This observation is in harmony with that of Hsieh et al. . This is confirmed by a significant strong positive caspase 3 immunoreactivities of hepato-renal tissues in melamine-intoxicated rats as presented in Fig. 12-14. Activation of caspase 3 can be triggered either by extrinsic or intrinsic factors inducing mitochondrial stress, and it plays important role in cell apoptosis . Activation of caspase 3 is an important hallmark of DNA fragmentation and nuclear condensation in apoptotic cells . c-Myc is a potent transcription factor that regulates the proliferation, growth, and differentiation of cells, whereas deregulation of c-Myc as in the situation of cellular stress induces apoptosis and suggested that Bax and caspase activation are involved in c-Myc induced apoptosis. The hepato-renal cell death induced by melamine may be attributed to the excessive release of ROS after mitochondrial dysfunction. Yiu et al. reported that melamine activates Ca2-sensing receptors which in turn causes a sustained Ca2 entry in the cell . Cell death may be mediated either by ROS generation together with elevation of Ca ions inducing a caspase-mediated apoptotic pathway or through activation of the mitochondrial proapoptotic (Bax-1/ Bcl-2) pathway and ROSmediated cytotoxicity . Therefore, the potent anti-apoptotic activity of affordable Se-NPs@starch has been commonly postulated as one of the vital mechanisms underlying its beneficial bioactive properties. This activity was previously reported . Many studies have shown the potential ameliorative effect of Se-NPs@starch against oxidative stress, nuclear damage, and cell death that may be due to the ability of Se-NPs@starch to enhance the antioxidant defense mechanism, scavenge free radicals efficiently , upregulate Nrf2 and heme oxygenase-1, as well as impede the inflammatory response and apoptotic cascade . ## Conclusions A versatile and green method for the fabrication of Se-NPs that are effectively stabilized by starch is described. The starch-stabilized Se-NPs were characterized by XPS, FT-IR, EDS elemental mapping, and TEM investigations. Se-NPs are distributed homogeneously over the starch surface and have a spherical shape with a particle size ranging from 20 to 140 nm. The protective efficacy of starch stabilized Se-NPs on the melamine-induced hepato-renal toxicity has been evaluated. Melamine ingestion can result in a range of toxicological effects on the liver and kidneys. The intoxication with melamine is characterized by an elevation of ALT, AST, serum urea, and creatinine, besides higher MDA levels, an increase in the expression level of the apoptosis-related gene, a reduction in the GSH, and a decrease in the expression level of antioxidant genes. These findings were associated with a strong positive immune expression of caspase-3 as well as severe distortion and alteration in the hepato-renal tissues. Fascinatingly, the daily oral administration of starch-stabilized Se-NPs (2 mg Se-NPs@starch/kg/day) for 28 days significantly reduces the apoptotic effect produced by melamine intoxication (300 mg melamine/kg/day) in the liver and kidney tissues of adult rats. Moreover, Se-NPs@starch can substantially improve the liver and kidney function parameters, alleviate the oxidative stress, apoptosis, and histopathological injuries exerted by melamine. This study highlights the detrimental effects of melamine ingestion on the liver and kidneys and provides Se-NPs stabilized by starch as an affordable and effective material to inhibit these toxic effects.

chemsum

{"title": "Protective effect of starch-stabilized selenium nanoparticles against melamine-induced hepato-renal toxicity in male albino rats", "journal": "ChemRxiv"}

research_on_multi-effect_evaporation_salt_prediction_based_on_feature_extraction

## Abstract: In the multi-effect evaporation salt making process, the smooth operation of the salt making process is crucial. As the salt production process continues, many unstable factors will cause the salt production process not to proceed smoothly. These factors can be discovered in advance by predicting the salt production data, thus, it is of great significance to predict the multi-effect evaporation salt production data. In the process of multi-effect evaporation and salt production, the multiple saltmaking devices make the influence between the parameters closer, and the influence of a single parameter on itself is sometimes ductile. Therefore, the data of multi-effect evaporation and salt production have the characteristics of high dimensions, high complexity and temporal information. If the historical salt production data is used for data prediction directly, the prediction model will take a long time and the prediction effect is not good. Thus, how to predict the multi-effect evaporation salt production data is the main research problem of this paper. In view of the above problems, according to the characteristics of multi-effect evaporation salt production data, this paper analyzes and improves the self encoder for feature extraction of multi effect-evaporation salt production data, so as to solve the problem of high dimensions and high complexity of salt production data. On this basis, combined with the time-series information contained in the salt production data, a multi-effect evaporation salt production data prediction model is proposed based on long-term and short-term memory cycle neural network to solve the prediction problem of time-series salt production data. Experiments show that the prediction model can predict and prevent the problems in salt production line in advance. It has a certain theoretical research value and application value in the intelligent production process and production line optimization of salt chemical industry.The salt chemical industry is an important part of the chemical industry and an important source of the economy. With the continuous development of the chemical industry, the salt chemical industry is faced with the challenge of developing from a highly labor-intensive and high-energy production method to a low-energy and highefficiency direction. In the face of this challenge, how to optimize the salt production process is the problem faced by the whole salt chemical industry. At present, the research on neural network, artificial intelligence, and big data is in-depth. How to apply them to the optimization of salt production process and get a better solution is particularly important. In recent years, the construction of industrial information infrastructure has been gradually accelerated, and the chemical production process has gradually completed the information coverage. The integration of emerging computer technology and industrial development provides the basis for the application of intelligent technology in the industrial field. The salt chemical industry is no exception. Most salt-making factories have supporting distributed control systems. DCS(Distributed Control System) uses computer technology to centrally control and manage the production process, and this system will support a background database to record a large amount of historical production data. These historical production data are of great value for the solution of optimizing the salt production process through neural network technology. Salt chemical companies mostly use mature multi-effect evaporation and combined heat and power technology to produce salt. Although the salt production technology is relatively mature, after years of production, the problems that occur in the production process will gradually accumulate.In the process of salt production, although the process of salt production is monitored by the DCS, when the staff observe the problem data, the problems in the production line have already occurred. For example, the multi-effect evaporator will have the problem of scaling at the salt discharge foot during the salt making process. The main reason is that the washing water is introduced into the salt discharge foot of the evaporation tank, and the solution at the salt discharge foot has a more violent cooling process, especially the 1-3 effect evaporation tank. Therefore, it is particularly important to predict the changing trends of the key parameters. In the process of multi-effect evaporation, the production process is continuous, and there is a potential connection and influence between different production parameters. Therefore, similar scaling problems can be obtained by analyzing the historical production data and the potential correlation between the production parameters, thus predicting the trend of key parameters in salt production, adjusting the salt production parameters in advance, and preventing problems. Therefore, it is of great practical and economic significance to establish a parameter prediction model based on the historical salt production data and predict the key salt production parameters through neural network technology. In the actual multi-effect evaporation salt production process, the salt production line will generate a large number of production data with high dimension and nonlinearity in real time. Generally speaking, if the production data with high dimension and large nonlinearity is directly used to train the prediction model, it will not only make the prediction model training efficiency low, but also make the prediction accuracy relatively low. Therefore, how to process the original salt production data and train the prediction model becomes the key point of the prediction problem. Compared with low-dimensional data, high-dimensional data contain more complex information and data analysis is more difficult. Using the high-dimensional data to train the prediction model directly will lead to the problems of long training time and poor prediction accuracy. Feature extraction is the key technology to deal with high-dimensional data. A series of transformations are carried out to map high-dimensional data into low-dimensional subspace, so as to obtain the low-dimensional feature representation of the original data. Using the extracted feature data of the original data to train the prediction model will greatly reduce the training time of the prediction model and improve the performance of the prediction model. The goal of various learning algorithms is to complete the prediction of the data, thus data prediction is the key problem that the learning algorithm needs to solve. The principle of data prediction is to learn the fixed pattern of the original data through the learning model, thereby predicting the data, thus data prediction is the ultimate goal of the entire learning process. There are many algorithms for data prediction in the computer field, such as recurrent neural networks. RNN (Recurrent neural network) is a type of neural network with memory function and is suitable for learning time-series data. RNN is the same as the general neural network, which is composed of explicit layer neurons, hidden layer neurons, and output layer. The difference is that hidden layer neurons are not only affected by the input neurons, but also affected by the current state of the hidden layer neurons themselves. The subsequent state, so RNN has the memory ability, which has an advantage for the learning task of time-series data. Power system load forecasting, especially the short-term power load forecasting of individual users, plays an important role in future grid planning and operation. In addition to large-scale centralized residential power supply, the power load of a single user involves high volatility and uncertainty, thus the prediction of power load is quite challenging 1 . Kong et al. proposed a framework based on recurrent neural networks to solve the above problems and tested it on public data. The proposed method outperformed other algorithms in the short-term load forecasting of individual users. The time sequence of power load is highly nonlinear, thus it is very difficult to accurately predict the power load. Zheng et al. found that short-term power load prediction can predict the future short-term load, and proposed a prediction framework based on long-short-term memory recurrent neural network. This framework can make use of the long-term dependence in the electrical load time series for accurate prediction 2 . Wang et al. studied the prediction problem of helping information spread on graphs through representation learning, especially the probability of inactive nodes activated at the next time point in the cascade of information dissemination 3 . The author believed that the deep learning method is successful in diffusion prediction, but it is not enough to explore the cascade structure, thus the author proposed that cascade is not only a sequence of nodes sorted by activation time, it has a richer potential structure, indicating the diffusion process on the data graph. The author introduced a new data model, the diffusion topology, to fully describe the cascade structure. The author found through research that using existing neural networks to model diffusion topology is a difficult task. Therefore, a novel topological recurrent neural network based on recurrent neural network was proposed, and it showed promising performance on multiple real data sets. For the personalized recommendation problem, Donkers et al. used recurrent neural networks to model the sequence data and generate effective personalized recommendations 4 . Sherstinsky et al. aimed at the current problem of insufficient RNN training formulas 5 , drawing on signal processing theory, drawing standardized RNN formulas from differential equations, and proposing and proving a precise statement that produced RNN expansion techniques, which helped to understand RNN more clearly. For the problem of traffic speed prediction, Lv et al. proposed a model based on recurrent neural network to achieve more accurate traffic speed prediction 6 . The author learns time-series patterns by integrating RNN and convolutional neural network models to adapt to the traffic dynamics in the surrounding area. For the problem of how to extract useful information from protein sequences, Liu et al. used RNN for protein function classification 7 . The multi-effect evaporation salt production process is a continuous production process, thus the salt production data have the characteristics of time series. Time-series data are a type of data with a certain order in the time series 8 , that is, the data change continuously according to the time axis, such as weather and air quality data, video data, stock data and vehicle flow data 9 , etc. They are widely used in many fields 10 . The salt production data generated by the multi-effect evaporation salt production process is a kind of time-series data, thus this paper studies the time-series data prediction algorithm. In this paper, the principle and construction of the feature extraction model based on neural network are studied, and the feature extraction model is used to extract the feature of the multi-effect evaporation salt production data, then the feature extracted from the original production data is used to train the prediction model, and then the key data of the multi-effect evaporation salt production are predicted. ## Results Multi-effect evaporation salt production data. This article mainly conducts prediction research on multi-effect evaporation salt production data. The salt production data used are all derived from real multi-effect evaporation salt production data. The salt production data are collected by DCS and stored in the background database, with one record per minute Data, the data contains 1935 parameter values. Among them, 589 parameters are numeric parameters, and the remaining parameters are non-numeric parameters. In this article, only numeric parameters are used. In terms of the collection of experimental data, a total of 80,000 consecutive time periods of data were obtained as experimental data sets, each of which contained 589 numerical parameters. Because this article is a prediction study on salt production data, some key parameters in the salt production line are selected as the prediction targets, namely the solid-liquid ratio, vapor pressure (kPa) and the vapor pressure (kPa) in the four evaporation tanks EV11 to 14(i.e., the actual ID of evaporation tank in the factory) in the salt production line. Salt leg flow ( m 3 /h ), where the attributes of key parameters are shown in Table 1. Table 1 shows the maximum, minimum, median, mean, variance and standard deviation of the parameters. According to the variance and standard deviation, it can be found that the fluctuations of the parameters of EV14 solid-liquid ratio, EV11 steam pressure and EV11 to 14 salt leg flow rate are very sharp, thus the degree of nonlinearity is extremely high, which are difficult parameters to predict. The variance and standard deviation of EV13 steam pressure and EV14 steam pressure are small, and the data change is relatively stable, thus they belong to better predicted parameters. Generally speaking, the experiments in this paper include more difficulty to predict parameters. Because the unit of different parameters of the original salt-making data is different, that is, the dimensions are different, it cannot be directly used for model training. Therefore, this paper preprocesses the experimental data, that is, the dimensions of the experimental data set are converted to be the same through a standardized method. The standardization method used in this article is MinMaxScaler standardization, as shown in formula (1). Among them, min is the minimum value of the parameter, and max is the maximum value of the parameter. In this way, all parameters in the experimental data can be mapped to the scope between 0 and 1, so that the standardized experimental data have the same dimension. After standardization, this paper divides the experimental data. In summary, there are a total of 80,000 experimental data, each of which contains 589 parameter values. In this paper, the first 70,000 data are used as the training set for the training of the feature extraction models and prediction models, and thereafter 10,000 data are used as the test set used for the evaluation of the model prediction performance. ## Experiment preparation. The experiment is divided into 3 parts in total: training of the feature extraction model based on deep confidence network, training of the AE+LSTM(i.e., Autoencoder combine with Long short-term memory) model and prediction analysis of the AE+LSTM model. The Autoencoder is composed of two modules: an encoder and a decoder. In this paper, the encoder is composed of one explicit layer and 2 hidden layers, and the decoder is composed of 2 hidden layers. In this paper, the number of neurons in the explicit layer of the encoder is set to 589, the number of neurons in the first hidden layer and the second hidden layer of the encoder is set to 200 and 50, and the number of neurons in the first hidden layer and the second layer of the decoder is set to 200 and 589. The initial values of the edge weights in the Autoencoder network are randomly generated by a uniform distribution function, and the bias on each neuron is initialized as 0. The setting of the LSTM model has been introduced in the third. Vol:.( 1234567890 Predictive analysis. After the AE+LSTM model training is completed, this article uses 10,000 pieces of test data to evaluate its prediction effect. In this paper, the solid-liquid ratio, steam pressure and salt leg flow rate of the EV11 to 14 evaporation tanks are used as the prediction targets. PCA+LSTM(i.e., principal components analysis combine with Long short-term memory) and AE+SVM(i.e., Autoencoder combine with Support Vector Machine) models are compared for prediction, as shown in Fig. 3. The PCA+LSTM model considers the time-series information of the data, but because its feature extraction model is PCA, the extracted features are fixed. The AE+SVM model uses AE as the feature extraction model, thus the extracted features are continuously learned, but the SVM model does not consider the temporal information of the data. Figure 3a uses solid-liquid ratio as the target for prediction. From EV11 to EV14, the AE+LSTM model achieve the minimum prediction error. However, AE+SVM model present the worst performance. the prediction error of PCA+LSTM is higher than AE+LSTM. The solid-liquid ratio is a parameter with large fluctuations in the multi-effect evaporation salt production process, we can conclude in this experiment that considering the temporal information will get a better effect and the performance of AE+LSTM is better than PCA+LSTM. Figure 3b is based on the prediction of steam pressure. The AE+LSTM model has the best effect and the smallest prediction error, followed by PCA+LSTM. The changes in the vapor pressure of EV13 and 14 evaporation tanks are generally stable, but the performance of AE+SVM is very bad. AE+LSTM and PCA+LSTM model all consider the time information of data, we conclude that the time information of data is very importance to the data of multi-effect evaporation salt production process. Figure 3c is based on the prediction of the salt leg flow rate. The AE+LSTM model has the best effect and the smallest prediction error. The salt leg flow rate is a parameter with large fluctuations. AE+LSTM not only considers the changes in data characteristics, but also considers the time series information of the data. Therefore, the AE+LSTM proposed in this paper can better predict the multi-effect evaporation salt production data. With the continuous growth of data dimensions, "dimensional disaster" has gradually become a concern for researchers in the computer field, because too high data dimensions will make the performance of various learning models poor. Therefore, in the face of the challenges brought by highdimensional data, feature extraction techniques emerged at the historic moment 11 . Feature extraction is mainly accomplished by transforming the feature space 12 , which is an important technique for representing high-dimensional data, and is a necessary preprocessing step for large-scale industrial data 13,14 . It is used in pattern recognition, data mining, and computer vision. All fields have applications 15 . In detail, feature extraction is to find the low-dimensional feature subspace of the data through mathematical methods 16 , and map the high-dimensional data into its low-dimensional subspace, so that the original data can be well represented in the low-dimensional subspace and distribution 17,18 . Different feature extraction methods have different performances 19,20 . Typical feature extraction methods include principal component analysis (PCA). Principal components analysis (PCA) is a typical feature extraction algorithm, widely used in the field of data compression and data analysis. PCA projects the high-dimensional features of the original data into a low-dimensional subspace in a linear or non-linear manner according to mathematical theory, and this lowdimensional subspace retains the original data space distribution as much as possible. In the projection process, PCA can identify the direction vectors called principal components from the data. In these direction vectors, the data in the original data set have the largest change in data value, that is, the data in the original data set. The changes are mainly reflected in these main direction vectors, thus the original data can be well represented using these direction vectors. PCA has good theoretical properties and attracts a large number of researchers. The classic PCA algorithm is widely used in the field of feature extraction. Not only that, many improved algorithms based on PCA have been proposed to better solve the corresponding problems. In order to be able to extract more useful features, Yi et al. proposed a new PCA algorithm, namely the Joint Sparse Principal Component Analysis (JSPCA) algorithm 21 . The JSPCA algorithm relaxes the orthogonal constraints of the transformation matrix so that more features can be freely combined to represent the data in low dimensions. JSPCA imposes a joint sparse constraint on the objective function, that is, a paradigm constraint on the loss term and the regular term, which improves the algorithm's greatness. The author analyzes the theory of the algorithm and gives a simple and effective optimization plan. Experiments show that this algorithm can better extract useful features in the data set. By improving the PCA algorithm, Fan et al. proposed a learning framework based on multiple similarity metric subspaces, namely an improved principal component analysis (MPCA) algorithm 22 . MPCA calculates three similarity matrices according to the similarity measurement method: interactive information matrix, angle information matrix, and Gaussian kernel similarity matrix. The author uses the feature vector of the similarity matrix to generate a new subspace, that is, the similarity subspace, and finally uses the feature selection method to generate a new complete similarity subspace, so as to realize the feature extraction of the data. In the process of data analysis, outliers are a problem that cannot be ignored. Thus Rahmani et al. proposed a simple, powerful, and robust principal component analysis algorithm 23 . In the article, the author proposes that as long as there are enough data points in the low-dimensional subspace, then these points have a strong correlation. In contrast, outliers usually do not exist in low-dimensional structures, thus outliers are unlikely to have a strong similarity to a large number of data points. As a result, outliers can be distinguished. This algorithm calculates the data correlation by normalizing the Gram matrix of the data, and then recovers the subspace through a small number of data points. Because its calculation process only involves matrix multiplication, this method is faster than the classic PCA calculation, and it can still perform well in the presence of abnormal points in the data. LSTM neural network prediction model. Long short-term memory (LSTM) model is a kind of recurrent neural network. It is proposed to solve the long-term dependence of RNN. Due to its unique design, it is suitable for predicting events with long-time intervals in time series. LSTM has the characteristics of recurrent neural network, as shown in Fig. 4. LSTM is composed of input layer, hidden layer and output layer. The state of the hidden layer neuron module at the current time point will affect the state of the module at a later time point. Time-series data will affect each other, that is, there is a causal relationship between the data at the current time point and the data at the www.nature.com/scientificreports/ previous time point and after the time point. The recurrent neural network can learn the association between the time-series data through the structure shown in Fig. 1. Therefore, the effect of the prediction of time-series data is outstanding. However, the traditional recurrent neural network can only remember the correlation between the data in a short period of time. To solve this problem, LSTM, because of its special structural design, can Learn through association, thus it has a good performance in the prediction of time series data. LSTM model is a special improvement based on RNN, which can be used to learn the long-term dependence of the data 24 . Illumination and photovoltaic data are typical time-series data. In order to accurately predict the photovoltaic power in the smart grid, Abdel-Nasser et al. used LSTM to predict the output power of the photovoltaic system, thereby providing an important guarantee for the safe operation of photovoltaics 25 . Qing et al. proposed a novel illumination forecasting scheme to predict solar irradiance through LSTM 26 . Video data is a commonly used time-series data in computer vision. In the study of picture subtitles, automatic description of natural language based on video content has attracted widespread attention. Gao et al. proposed a novel LSTM-based framework that can convert video into natural sentences. The framework integrates the attention mechanism with LSTM to capture the salient structure of the video and explore the correlation between multi-modal representations to generate sentences with rich semantic content 27 . In the 3D skeleton sequence, human motion recognition has attracted the attention of many researchers. LSTM has advantages in modeling dependencies. Liu et al. proposed an LSTM framework, that is, global context awareness attention LSTM, used for bone motion recognition. The algorithm can selectively focus on information-rich joints, and further improve the attention ability in each frame by using the global context storage unit 28 . Weather data is also a typical time-series data. Huang and others have developed a framework based on convolutional neural networks and LSTM to solve the problem of air pollutant index prediction, and use historical data to predict the air pollution index 29 . Because of its structural design, LSTM has very good performance in time-series data learning. When learning time-series data, LSTM learns the current and past states of neurons to generate potential characteristics of time series data, thus it can be used for the prediction of salt production data in the multi-effect evaporation salt production process. Autoencoder. Autoencoder (AE) can learn artificial neural networks from the data through unsupervised learning. The AE is composed of an encoder module and a decoder module. The encoder maps the data to the feature subspace, and the decoder is responsible for reconstructing the data through the features. Chorowski et al. applied self-coding neural networks to speech waveforms 30 , and used Autoencoder neural networks to extract meaningful potential speech features in an unsupervised manner. The goal of this algorithm is to learn the features of higher-order semantic content from the signal. Because the learning behavior of the self-encoder model depends on the representation of potential constraints, the author applies three variants of the self-encoder and different constraints so that the model can learn the features that meet the needs. Zeng et al. proposed a hybrid model combining a stacked self-encoder and Mel frequency cepstrum coefficients 31 , which extracts key information through the self-encoder to improve model performance. The features involved in image processing are high-dimensional. For systems like facial expression recognition, selecting the most important features is a very critical task. Usman et al. studied the performance of deep self-encoders in feature extraction 32 , performing facial expression recognition on multiple hidden layers. Compared with other feature selection and size reduction techniques, feature performance extracted from stacked self-encoders behaves better. Condition monitoring is one of the main tasks in the industrial process. Mechanical parts such as motors, gears and bearings are the main components of the industrial process. Any failure in them may cause the entire process to stop completely, resulting in serious losses. Therefore, it is critical to predict defects before they occur, but most methods are based on the processing of raw sensor data, which is complicated and inefficient. The latest development of feature extraction methods based on self-encoders provides methods for them, but they are mainly limited to the field of image and audio processing. Based on self-encoders and online sequential learning networks, Roy et al. developed an automatic feature extraction method for online status monitoring 33 . Experiments show that the method performs well. Effective condition monitoring can improve the reliability and safety of equipment. Feature extraction determines the performance of the monitoring model to diagnose faults. Maurya et al. proposed a feature extraction technology based on the fusion of low-order features and high-order features 34 to detect machines Failures and potential anomalies. The author uses signal processing techniques to extract low-order data features, and uses deep neural networks based on stacked Autoencoders to extract high-order data features. The acoustic data set collected by the air compressor verifies the effectiveness of the proposed method. On the feature extraction problem, how to map the data into the appropriate subspace is the key to the problem, and the dimensions of the original data also affect the performance of the feature extraction algorithm. High-dimensional data will greatly reduce the performance of the classic feature extraction algorithm, making it impossible to extract more important features in the data. The feature extraction algorithm based on neural network, because of its structural design, can be reflected in the feature extraction of high-dimensional data with very good performance. The Autoencoder is composed of an input layer, a hidden layer, and an output layer, and its structure is shown in Fig. 5, where the hidden layer is divided into an encoder and a decoder according to functions. Autoencoder and deep-confidence network have different training methods. Deep-confidence network trains RBM layer by layer, while Autoencoder trains the entire network. In the training process, the data enter the network from the input layer of the autoencoder, and the features of the data are learned by the encoder, and the learned features are sent to the decoder for decoding, that is, the original data are reconstructed by the learned features. In this process, the function of the hidden layer in the center of the Autoencoder is to obtain the characteristics of the data. The Autoencoder adjusts the entire network through reconstruction errors, and this process is repeated until the reconstruction error values tend to converge. The trained Autoencoder can be used for feature extraction, In traditional Autoencoder, the activation functions used by hidden layer neurons are mostly sigmoid activation functions. This type of activation function maps the data to a low-dimensional subspace through a nonlinear transformation. The purpose of the entire mapping is to make the data features as separable as possible. In the study of multi-effect evaporative salt production data, the extracted features need to be as separable as possible, but the individual features conform to the normal distribution as much as possible, which is conducive to the prediction of the prediction model later. Therefore, this paper makes some improvements on the autoencoder model. ## Methods This article mainly studies how to use the Autoencoder for feature extraction of multi-effect evaporation salt production data. From the analysis of the aforementioned Autoencoder model, it is known that the Autoencoder is composed of an encoder and a decoder. The activation function does not make the entire model well learn the characteristics of multi-effect evaporation salt production data. Therefore, this paper proposes a feature extraction algorithm based on Autoencoder for the feature extraction of multi-effect evaporation salt production data as shown in Table 2. Algorithm 1 is a feature extraction algorithm based on the Autoencoder. This algorithm has improved the multi-effect evaporation salt production data, mainly to make changes in the activation function of the encoder. The activation function uses the sigmoid function, as shown in formula ( 2), but the hidden layer that plays the coding function uses the normal distribution function as the activation function, as shown in formula (3). www.nature.com/scientificreports/ In Algorithm 1, the input of the algorithm is the salt-making data, a list of the number of neurons included in each hidden layer of the autoencoder h_list , the learning rate l and the number of training times max_epoch . Step (1) normalize the salt production data to obtain the standardized data v, and obtain the data dimension and the number of hidden layers h_num through the len function. Steps (2) to (5) are the initialization of each hidden layer of the Autoencoder. The init_layer_couple function initializes the neurons of the adjacent layer. Lc is the subnetwork composed of the display layer and the first hidden layer, lc .b is the hidden layer bias of the subnetwork, lc .W is the weight matrix of the self-network. Steps (6) to (8) are the training process of the Autoencoder. The specific training process is shown in Algorithm 2. Algorithm 2 is the training process of the Autoencoder, as shown in Table 2. The input of the algorithm is data v, the subnet list lc of the Autoencoder, the learning rate l, and the number of subnets h_num . Steps (2) to (6) are the forward propagation process of the encoder. Here, we set the activation function of the neuron at the end of the encoder to be a normal distribution function, and the activation function of the remaining hidden layers is the ReLU function. Steps (7) to (8) are used to calculate the gradient of the network, and Steps (9) to ( 12) are used to update the parameters in the network. The algorithm finally returns the updated lc list. Steps (9) to (13) of Algorithm 1 are the process of acquiring the data features of the Autoencoder. Through the above steps, the Autoencoder has been completely trained. Finally, the trained Autoencoder coding layer is used for data feature extraction. Finally, the algorithm returns the extracted data features. In this paper, Algorithm 1 and LSTM are combined to form the AE+LSTM prediction algorithm to predict the multi-effect evaporation salt production data. Algorithm 2 is the training process of the Autoencoder, as shown in Table 3. The input of the algorithm is data v, the subnet list lc of the Autoencoder, the learning rate l, and the number of subnets h_num . Steps (2) to ( 6) are the forward propagation process of the encoder. Here, we set the activation function of the neuron at the end of the encoder to be a normal distribution function, and the activation function of the remaining hidden layers is the ReLU function. Steps (7) to (8) are used to calculate the gradient of the network, and Steps (9) to ( 12) are used to update the parameters in the network. The algorithm finally returns the updated lc list. Steps (9) to (13) of Algorithm 1 are the process of acquiring the data features of the Autoencoder. Through the above steps, the Autoencoder has been completely trained. Finally, the trained Autoencoder coding layer is used for data feature extraction. Finally, the algorithm returns the extracted data features. In this paper, Algorithm 1 and LSTM are combined to form the AE+LSTM prediction algorithm to predict the multi-effect evaporation salt production data. After that, the features are input into the LSTM for training and prediction. LSTM has a special design for long-term memory on its hidden layer neurons, so that the entire network can learn long-term associations between data. Its hidden layer neuron structure is shown in Fig. 6. In Fig. 6, the hidden layer neuron of LSTM has three kinds of inputs, namely the input of the hidden layer neuron at time t, the output of the hidden layer neuron h t−1 at time t − 1 and the state parameter of the hidden layer neuron Cell t−1 at time t − 1 . The LSTM hidden layer neuron is mainly composed of four modules. In module 1, the function σ of the function is to selectively forget some information in the hidden layer neuron h t−1 output at time t − 1 and the input data v t at time t. The output of module 1 is shown in formula (4). www.nature.com/scientificreports/ In module 2, the role of this module is to update information, and selectively update the information in the hidden layer neuron output h t−1 at time t − 1 and the input data v t at time t. The output of module 2 is shown in formulas ( 5) and ( 6). The function of module 3 is to update the state of the hidden layer neurons at the current moment according to the outputs of module 1 and module 2. The update formula is shown in formula (7). After the update operation of module 3, the state of the hidden layer neurons is updated. Module 4 extracts the features of the hidden layer neuron output h t−1 at time t − 1 and the input data v t at time t through the current state of the neuron, and outputs the current state hidden layer, as shown in formulas (8) and (9). Through the special design of the above process, LSTM can remember and forget the features in the data, and iterate the state of the hidden layer neurons at different times to learn the long-term association of the data. Therefore, LSTM can perform learning tasks on time-series data, especially the data that have long-term dependence in the data, and obtain better learning results. This article considers the time series of multi-effect evaporation salt production, and in the multi-effect evaporation salt production process, the parameter state at the current time point will affect the parameter state at the future time point, and the effect of the reaction in the multi-effect evaporation tank will continue very long. Long time, that is, there is a long-term dependence between data. Therefore, this paper uses LSTM as the prediction model to predict the multi-effect evaporation salt production data. Received: 1 July 2020; Accepted: 12 October 2020 (4)

chemsum

{"title": "Research on multi-effect evaporation salt prediction based on feature extraction", "journal": "Scientific Reports - Nature"}

integrating_single_ni_sites_into_biomimetic_networks_of_covalent_organic_frameworks_for_selective_ph

## Abstract: Selective photoreduction of CO 2 into a given product is a great challenge but desirable. Inspired by natural photosynthesis occurring in hierarchical networks over non-precious molecular metal catalysts, we demonstrate an integration of single Ni sites into the hexagonal pores of polyimide covalent organic frameworks (PI-COFs) for selective photoreduction of CO 2 to CO. The single Ni sites in the hexagonal pores of the COFs serve as active sites for CO 2 activation and conversion, while the PI-COFs not only act as a photosensitizer to generate charge carriers but also exert a promoting effect on the selectivity.The optimized PI-COF with a triazine ring exhibits excellent activity and selectivity. A possible intra-and inter-molecular charge-transfer mechanism was proposed, in which the photogenerated electrons in PI-COFs are efficiently separated from the central ring to the diimide linkage, and then transferred to the single Ni active sites, as evidenced by theoretical calculations. ## Introduction Selective photoconversion of CO 2 to a target industrial product is an intriguing approach to simultaneously enrich solar energy and utilize CO 2 . 1,2 However, efficient photoreduction of CO 2 with high selectivity, particularly in aqueous solution, is a considerable challenge because of the multi-electron reaction process and the competing H 2 evolution in the CO 2 reduction reaction. 3,4 So far, many molecular metal complex based photocatalytic systems have been developed to selectively reduce CO 2 into solar fuels with high efficiency. Molecular complexes as precursors of single active sites with tailorable and versatile coordination possess maximum efficiency of catalytic sites in the reaction. However, most of the systems suffer from insufficiently stable and expensive photosensitizers to achieve high performance. 9 The utilization of semiconductors would be a promising alternative approach for photocatalysis, 10,11 as they always possess higher photostability. In practical applications, the photocatalytic performance of semiconductorbased catalytic systems is still limited because of their exterior surface catalytic mechanism, always leading to a limited utilization of photogenerated charges. Taking inspiration from nature, where photocatalysis for converting solar energy into chemical energy occurs in the hierarchical networks in plants' leaves with non-precious metal catalysts, 15,16 the combination of single metal sites and a hierarchical porous semiconductor may offer an applicable approach towards the development of photocatalytic systems for selective conversion of CO 2 . 17 In this context, it is desirable to explore novel semiconductors with intrinsic hierarchical porosity to accommodate single active sites and maximize the transfer of photogenerated charges to the active sites. The development of covalent organic frameworks (COFs) 18 provides a promising platform for photocatalysis. The periodic and permanent porosity endow COFs with a naturemimicking architecture, while the diverse compositions and synthetic approaches allow COFs with tunable microstructures and optical and electronic structures. The highly conjugated structure in-plane in COFs can ensure the mobility of photoinduced charges. Moreover, COFs with tunable porosity can accommodate guest molecules for target applications. In terms of catalysis, the microenvironment in the cavity of a heteroatom-rich COF may impose complicated effects on the active sites and its catalytic performance. 40 Compared to metalorganic frameworks, organic polymers and inorganic networks, COFs possess a metal-free skeleton and periodic porosity to form biomimetic microenvironments as in plants' leaves for the accommodation of metal molecular catalysts. Profting from the unique characteristics of COFs, it is reasonable to expect the integration of single metal sites in photoactive COFs for photoreduction of CO 2 . Herein, we report an integration of single Ni sites in the biomimetic channels in polyimide covalent organic frameworks (PI-COFs) for selective photoreduction of CO 2 to CO (Fig. 1). The excellent catalytic performance mainly arises from the synergistic effects of single Ni sites and the PI-COFs, in which the engineered PI-COFs via adjusting the building units not only act as hosts for accommodating single Ni sites but also are responsible for the generation and separation of charge carriers. ## Results and discussion Three polyimide covalent organic frameworks (PI-COFs) were synthesized by coupling pyromellitic dianhydride (PMDA) with tris(4-aminophenyl)amine (TAPA), 1,3,5-tris(4-aminophenyl) benzene (TAPB) and 1,3,5-tris(4-aminophenyl)triazine (TAPT), respectively, and are denoted as PI-COF-1, PI-COF-2 (ref. 41) and PI-COF-TT, 42 respectively. The formation of the COFs was assessed by Fourier transform infrared (FT-IR), solid-state 13 C NMR spectroscopy and powder X-ray diffraction (PXRD). FT-IR spectra of all three COFs showed strong peaks at 1720-1725 cm 1 (Fig. S1-S3 †), confrming the formation of fvemembered imide rings. In the solid-state 13 C NMR spectroscopy, PI-COFs showed the characteristic signal for the carbonyl carbon of the imide ring at 165 ppm (Fig. S4 †). 43 The overlapping peaks from 116.5 to 144.7 ppm were attributed to phenyl carbons. The PXRD patterns of PI-COFs show prominent peaks, indicating their crystalline nature. The experimental profles of PI-COFs match well with their simulated PXRD patterns indicating a serrated stacking bnn net with adjacent sheets slipping by 1/4 of the unit cell distances (Fig. 2a-c). 41 All PI-COFs exhibit similar geometries, whereas the linkages in PI-COF-2 and PI-COF-1 twist rather than remaining planar as in PI-COF-TT, and the phenyl group is tilted by 48.44 and 41.37 with respect to the diimide plane, respectively. The N 2 adsorption/ desorption measurements show that all as-synthesized PI-COFs have a pore size in the range of 1.5 to 3.5 nm calculated using the nonlocal density functional theory (NLDFT), which is in agreement with the pore sizes predicted from the theoretical crystal structures. The Brunauer-Emmett-Teller (BET) surface areas of PI-COF-1, PI-COF-2 and PI-COF-TT are calculated to be 475, 1175 and 825 m 2 g 1 , respectively (Fig. S5-S7 †). All three PI-COFs exhibit good CO 2 uptake behavior due to the polarity of the polymer surface 44 (Fig. S8-S10 †), and PI-COF-TT shows the highest isosteric heat (Q st ) (29.76 kJ mol 1 ) for CO 2 adsorption at low coverage (Fig. S11 †). This may be attributed to the dipolequadrupole interactions between CO 2 molecules and the imide groups and triazine rings in the framework. 44,45 Scanning electron microscopy (SEM) images show that PI-COF-1 and PI-COF-2 exhibit a "sphere-like" morphology (Fig. S12 †), while PI-COF-TT is a cross-linked network (Fig. S13 †). Crystalline domains were identifed in PI-COF-TT in the transmission electron microscopy (TEM) images (Fig. 2d). The presence of regular lines with a spacing of $3.0 nm in TEM images was consistent with the interatomic distances of the plane inferred from PXRD and computational models. TEM images of PI-COF-1 and PI-COF-2 showed that the dense nanospheres were comprised of small nanoparticles (Fig. S14 †). PI-COFs are stable in common organic solvents, water and acidic aqueous solutions (Fig. S15 and S16 †). Thermogravimetric analysis (TGA) showed that PI-COFs can be stable up to 450 C in argon (Fig. S17 †), and no obvious changes of crystal intensity were observed after heating to 300 C (Fig. S18 †). UV-vis diffuse reflectance spectra (DRS) show that all three PI-COFs absorb light in the ultraviolet and parts of the visible region (Fig. 2e), suggesting the optical band gaps of 1.74, 2.09 and 2.49 eV for PI-COF-1, PI-COF-2 and PI-COF-TT (Fig. S19 †), respectively. The intrinsic absorption band edge varies with the central ring of the COFs. From adjusting triazine, phenyl and tertiary amine units as the central ring, the electron-rich property of the central ring increased, so a remarkable red-shifted absorption of PI-COFs was observed. The band structures of the as-synthesized PI-COFs were obtained from the combination of optical absorption spectra and Mott-Schottky plots (Fig. S20-S22 †), and are shown in Fig. 2f. Obviously, the changes of the electronic and steric properties of the central ring result in a progressively enlarged band gap and the corresponding conduction band (CB) and valence band (VB) positions. It is noteworthy that the reduction potentials of the CB electrons in PI-COFs are much more negative than the reduction potentials of CO 2 to various hydrocarbon fuels, making them suitable for reduction of CO 2 . Although Re-modifed COFs have been previously employed in photocatalytic reduction of CO 2 , 46,47 it is desirable to develop non-precious metal-based catalytic systems. 48 The [Ni(bpy) 3 ] 2+ complex has been shown to reduce CO 2 actively in electrocatalysis 49 or in homogeneous photocatalysis. 50 The combination of photoactive COFs and [Ni(bpy) 3 ] 2+ as photocatalytic systems may offer a viable approach for conversion of CO 2 . The photocatalytic activity of PI-COFs toward reduction of CO 2 in aqueous solution was studied by using self-assembled [Ni(bpy) 3 ] 2+ (bpy is 2,2 0 -bipyridyl) as a precursor of active sites. The [Ni(bpy) 3 ] 2+ complex can be easily formed upon addition of a bpy ligand and Ni(II) salt as evidenced by UV-vis absorption spectra. A remarkable red-shift of bpy in UV-vis absorption spectra indicates the coordination of bpy with Ni ions (Fig. 3a). The encapsulation of molecular Ni complexes into the COFs (denoted as Ni@PI-COF-TT) was confrmed by aberrationcorrected high-angle annular dark-feld scanning transmission electron microscopy (HAADF-STEM), energy dispersive X-ray (EDX) mapping, X-ray photoelectron spectroscopy (XPS), PXRD and N 2 sorption measurements. HAADF-STEM images clearly show the crystalline domain in Ni@PI-COF-TT (Fig. 3b and c). The fast-Fourier transform (FFT) image confrms the crystalline nature of Ni@PI-COF-TT. Evidently, the bright spots corresponding to the Ni single sites are distributed in the hexagonal pores of Ni@PI-COF-TT (Fig. 3d), and not on the external surface of the stacking layers. EDX elemental mapping and TEM images reveal that Ni ions are uniformly dispersed in the framework (Fig. 3e and S23 †). XPS investigations with Ar-ion etching were conducted on Ni@PI-COF-TT. The peak at 856.0 eV was assigned to Ni 2+ (Fig. 3f). After Ar + etching, the intensity of the characteristic peaks of Ni 2+ relatively increased, further confrming the encapsulation of the molecular Ni complex in the channels of Ni@PI-COF-TT. A new peak at 854.1 eV is attributed to the reduction of Ni 2+ to Ni 0 , which was caused by Ar ion bombardment. Compared to PI-COF-TT, the PXRD peak of Ni@PI-COF-TT at 2.7 corresponding to the (110) facet weakened, due to the disorder induced by the Ni complexes inside the channels of Ni@PI-COF-TT (Fig. S24 †). The PXRD peak intensity can be recovered after the removal of [Ni(bpy) 3 ] 2+ from the COF by washing with water. The BET surface area of Ni@PI-COF-TT is obviously reduced yet can be recovered after [Ni(bpy) 3 ] 2+ removal (Fig. S25 †). The reduction of porosity in Ni@PI-COF-TT is mainly attributed to the occupancy of molecular Ni complexes in the channels. In the photocatalytic CO 2 reduction reaction, under the optimized reaction conditions, PI-COF-TT can generate 1933 mmol g 1 CO with a 93% selectivity over H 2 production in a 4 h reaction (Fig. 4a), while PI-COF-1 and PI-COF-2 showed relatively low catalytic activities (Fig. 4b). It is noteworthy that the activity of PI-COF-TT is lower than that of the previously reported Ni-TpBpy-based catalytic system, mainly as a result of the absence of precious metal sensitizers. Nevertheless, the catalytic performance of PI-COF-TT is among the best compared to other noble-metal-free catalytic systems that have been reported so far (Table S1 †). Control experiments show that the photocatalytic reduction of CO 2 results from the coexistence of PI-COF-TT and [Ni(bpy) 3 ] 2+ (Fig. 4c). Cycling experiments indicate the catalytic and structural stability of PI-COF-TT in the photocatalysis (Fig. 4d and S26 †). To confrm the origin of the as-formed CO, 13 CO 2 labelling experiments were performed. A major signal at a mass/charge ratio of 29 on the spectrum corresponding to 13 CO appears, confrming that the generated CO comes from the reduction of CO 2 (Fig. 4e). There were no detectable hydrocarbon products such as HCOOH and CH 3 OH when analyzing the reaction solvent through 1 H NMR and highperformance liquid chromatography (HPLC) (Fig. S27 †). The trend of CO production matches well with the optical absorption spectrum of PI-COF-TT (Fig. 4f), suggesting that the CO 2 reduction is indeed induced by PI-COF-TT. The quantum efficiency (AQE) of PI-COF-TT was estimated to be 0.55% at 380 nm. Acetonitrile was found to be a favourable reaction solvent for CO 2 reduction with a relatively high activity and selectivity (Fig. 5a), mainly because of its appropriate coordination ability to Ni, which can not only contribute to the stabilization of the Ni active sites, but also retain the accessibility to the Ni center. Other metal complexes were tested under similar reaction conditions (Fig. 5b). The results show that the Co complex could also reduce CO 2 to produce CO efficiently but with a low selectivity for CO compared to H 2 . H 2 O in the reaction could signifcantly influence the catalytic activity and product selectivity. Upon increasing the amount of H 2 O in the reaction, the production of both CO and H 2 decreases (Fig. 5c), probably due to the intrinsic hydrophobicity of PI-COFs. With the increment of the amount of Ni 2+ in the reaction the reduction production gradually increased, further confrming that the Ni single sites act as catalytically active sites (Fig. 5d). The photoelectrochemical properties of the PI-COFs were investigated by electrochemical impedance spectroscopy (EIS) and transient photocurrent measurements. Nafon solution was used as an additive to PI-COF powder to form the active layer, 51 which can form a homogeneous ink with the COF and further help to attach onto the surface of electrodes. Control experiments showed negligible effects of Nafon on the current density of electrodes (Fig. S28 †). As shown in Fig. 6a, Nyquist curves show that all three PI-COFs exhibit two semicircles. The small semicircles in the high-frequency region correspond to the charge transfer resistance, while the large ones in the lowfrequency region are related to diffusion resistance. The radius of the semicircles in the high frequency region of PI-COF-TT is smaller than that of other two PI-COFs, suggesting that the triazine ring in the PI-COFs obviously improves the rate of charge transfer. 52 Linear potential sweep measurements of PI-COF-TT showed a higher current density of 10.35 mA cm 2 at 0.1 V vs. RHE compared to PI-COF-1 and PI-COF-2 (Fig. S29 †). The current density is directly correlated with the increase in the catalytic performance. In addition, PI-COF-TT exhibits higher photocurrent intensity than PI-COF-1 and PI-COF-2 (Fig. 6b), which probably contributes to the improved photocatalytic activity. 53 To show the morphology of the active photosensitizer layer, SEM of the PI-COF electrodes was performed. Top view SEM images revealed that the FTO glass electrodes can be well covered by PI-COFs (Fig. S30 †). All PI-COF deposits show interstitial voids and textural porosity which may also contribute to their catalytic performance. 54,55 Fig. 4 An interesting photochromic phenomenon was observed in the photoreduction of CO 2 over PI-COFs (Fig. S31 †). When the catalytic system was irradiated under a CO 2 atmosphere, as shown in Fig. 7a and b, the color of PI-COF-TT changed from the original yellow to green to yellow to orange. After the introduction of air into the reaction, the color of PI-COF-TT changed from orange back to the original yellow. The color change of PI-COF-TT takes place in the presence of triethanolamine (TEOA) under light irradiation (Table S2 †). The electron paramagnetic resonance (EPR) spectrum of the suspension of PI-COF-TT and TEOA in acetonitrile shows a prominent increment of EPR signals after light illumination (Fig. 7c and d), indicating charge generation upon photoexcitation, 56 and leading to the photochromic phenomenon. In comparison with PI-COF-TT, the EPR signal intensity of Ni@PI-COF-TT decreased predominantly under identical conditions, revealing the electron transfer from PI-COF-TT to [Ni(bpy) 3 ] 2+ . The solid-state photoluminescence (PL) emission spectra of PI-COFs show negligible peaks around 400-700 nm compared to the strong peak of PMDA at about 430 nm (Fig. S32-S34 †). The weak emission of PI-COFs mainly results from the intramolecular charge transfer from the central rings acting as the electron donor to pyromellitic diimide units acting as the electron acceptor in PI-COFs. 31,57 Cyclic voltammetry (CV) and optical tests of pyromellitic diimide, N,N 0bis(phenyl)pyromellitimide, triphenylamine, triphenylbenzene and triphenyltriazine as model compounds of building blocks were performed to reveal their relative energy levels (Fig. S35 †). The LUMO levels of N,N 0 -bis(phenyl)pyromellitimide are very close to those of triphenylamine, triphenylbenzene and triphenyltriazine, which means that the connection of pyromellitic diimide and triphenylamine, triphenylbenzene or triphenyltriazine affording an extended conjugation framework could presumably contribute to the electron transfer. When a physical mixture of pyromellitic diimide and triphenyltriazine was employed in the catalytic reaction, no reduction products such as CO, CH 4 and H 2 were observed (Fig. S36 †), which further reveals the vital role of the formation of the conjugated structure in PI-COFs for charge transfer. Density functional theory (DFT) calculations show that the CB and VB wave functions of PI-COFs are separately localized on the diimide units and the central rings, respectively (Fig. 7e). Due to their narrow band gaps, the PI-COFs can be excited to form electron-hole pairs by light irradiation. The photogenerated electrons transfer from the central rings to the diimide units. The planar pconjugated structure in PI-COF-TT brings a superior separation of photogenerated electrons and holes compared to PI-COF-1 and PI-COF-2, 58 thus contributing to the enhanced catalytic efficiency. It is noteworthy that an obvious induction period accompanies the photochromic phenomenon. Thus, DFT calculations and control experiments were carried out to further explore the reaction mechanism. The in situ formed molecular Ni complexes preferred to be adsorbed to the diimide unit and not to the triazine ring, as evidenced by DFT calculations (Fig. 8a). However, the distance between the Ni complexes and framework complexes with a relatively large molecular volume cannot diffuse effectively into the pores of PI-COFs, leading to the inefficient catalytic activity. The recovered PI-COF-TT does not generate the CO product in the additional reaction run in the absence of Ni complexes, further confrming that Ni active sites are free in the channel and only physically adsorbed on the pore surfaces. 59 Additionally, when a bpy ligand with large substituents was used, the production of CO and H 2 was not detected under similar reaction conditions (Fig. 8d). This could be due to the steric hindrance of the substituent on the bpy ligand, leading to inefficient contact between the Ni center and COF wall. These noncovalent interactions are always considered to be conducive for retaining their intrinsic properties as well as the sufficient interfacial mass transport. 48,60 Thus, the induction period could be mainly ascribed to the collisional electron transfer from the COF to molecular Ni complexes to form molecular Ni active sites, 48 In order to understand the selectivity trends between the CO and H 2 products, the adsorption energies of CO 2 and H 2 O onto [Ni(bpy) 2 ] 0 with and without PI-COF-TT were calculated (Fig. S41 and Table S3 †). As shown in Fig. 9a, the adsorption energy of CO 2 on [Ni(bpy) 2 ] 0 in the presence of PI-COF-TT was signifcantly lower than that of H 2 O, implying the stronger affinity of molecular [Ni(bpy) 2 ] 0 toward CO 2 , which would facilitate the formation of the key intermediate Ni-CO 2 adducts and afford selectivity for the CO 2 reduction product rather than the H 2 product. It is noteworthy that the adsorption energy of CO 2 on Ni active sites was obviously reduced by PI-COF-TT through the hydrogen bonding interactions between the hydrogen atom in the PI unit and the activated CO 2 molecule (Fig. 9a, inset). Accordingly, the selective photoreduction of CO 2 was greatly dependent on the selective adsorption and activation of CO 2 on the metal active sites and the special reaction microenvironment. The synergistic photocatalytic system containing PI-COFs and single Ni sites facilitates the selective activation of CO 2 and inhibits the competitive H 2 evolution, leading to the enhanced catalytic activity and selectivity. Based on the above results, a possible mechanism for the selective photoreduction of CO 2 over PI-COF-TT with molecular Ni complexes was proposed (Fig. 9b). Light illumination on PI-COF-TT generates electron-hole pairs. The ## Conclusions In summary, we demonstrated a design of an integration of PI-COFs with Ni single sites for selective photoreduction of CO 2 . The electronic properties of PI-COFs can be facilely tuned for photocatalysis. The photogenerated electron-hole pairs in PI-COFs under irradiation can be efficiently separated through an intra-and inter-molecular charge-transfer mechanism to drive the reduction of CO 2 . The excellent catalytic performance of the COF-based catalytic system mainly arises from the synergistic effects of the photoactive PI-COF and single Ni sites, in which the PI-COF with a nature-mimicking architecture serves as the functionalized host for single Ni sites to promote the reaction. This work presents an integration of Ni single sites in PI-COFs for CO 2 reduction and a deep understanding of the electron-transfer mechanism in COFs. Based on the diversity of COFs and molecular metal complexes, we believe, inspired by our work, there would be more functionalized COFs developed for robust and efficient catalytic systems for sustainable energy conversion. ## Experimental Synthesis of PI-COF-TT TAPT (35.1 mg, 0.10 mmol) and PMDA (32.7 mg, 0.15 mmol) were placed in a mixed solution of mesitylene/NMP/ isoquinoline (0.5 mL/0.5 mL/0.05 mL). The tube was flash frozen at 77 K (liquid N 2 bath) and degassed by pump-thaw three times. The tube was sealed and heated at 200 C for 5 days, giving a yellow precipitate. The precipitate was purifed by Soxhlet extraction using tetrahydrofuran overnight, and fnally dried under vacuum at 80 C to give PI-COF-TT (yield 66%). ## Photocatalytic reduction of CO 2 A mixed solution of acetonitrile, H 2 O and triethanolamine (TEOA) (3 : 1 : 1, 5 mL) containing PI-COFs (10 mg), Ni(ClO 4 ) 2 -$6H 2 O (2 mg, 5.5 mmol) and 2,2 0 -bipyridyl (15 mg, 0.1 mmol) was purged with CO 2 for 15 min. The solution was then irradiated under UV-Vis light (300 W Xe lamp, PLS-SEX 300/300UV, 780 mW cm 2 ) at 313 K. After each reaction time, the generated gas in the headspace of the reaction vessel was sampled with a gastight syringe and determined by gas chromatography (Agilent 7890B) with both TCD and FID detectors. H 2 was detected using a TCD detector. CO was converted to CH 4 in a methanation reactor and then analyzed using an FID detector. The apparent quantum efficiency (AQE) at 380 nm was calculated by the following equation: AQE ¼ (2 amount of CO molecules evolved in 1 h/number of incident photons in 1 h) 100%.

chemsum

{"title": "Integrating single Ni sites into biomimetic networks of covalent organic frameworks for selective photoreduction of CO₂", "journal": "Royal Society of Chemistry (RSC)"}

tandem_processes_promoted_by_a_hydrogen_shift_in_6-arylfulvenes_bearing_acetalic_units_at_<i>ortho</

## Abstract: 6-Phenylfulvenes bearing (1,3-dioxolan or dioxan)-2-yl substituents at ortho position convert into mixtures of 4-and 9-(hydroxy)alkoxy-substituted benz[f]indenes as result of cascade processes initiated by a thermally activated hydrogen shift. Structurally related fulvenes with non-cyclic acetalic units afforded mixtures of 4-and 9-alkoxybenz[f]indenes under similar thermal conditions. Mechanistic paths promoted by an initial [1,4]-, [1,5]-, [1,7]-or [1,9]-H shift are conceivable for explaining these conversions. Deuterium labelling experiments exclude the [1,4]-hydride shift as the first step. A computational study scrutinized the reaction channels of these tandem conversions starting by [1,5] -, [1,7]-and [1,9]-H shifts, revealing that this first step is the rate-determining one and that the [1,9]-H shift is the one with the lowest energy barrier.260 Scheme 2: Preparation of benz[f]indenes 5 and 6. Reagents and conditions: i) cyclopentadiene, pyrrolidine, anhydrous methanol, rt, 10 h; ii) DMSO, microwave, 120 °C, 120 W, ## Introduction Fulvenes (also known as pentafulvenes), a unique class of trienes, have intrigued chemists for decades due to their theoretical interest and synthetic applications . In this latter sense, fulvenes can be involved in multiple modes of cyclization processes such as [4 + 2] , [6 + 2] , and [6 + 3] [ cycloaddition reactions resulting in the construction of diverse fused ring systems. Other classical pericyclic processes that may potentially occur in fulvene fragments (electrocyclic and ene reactions, sigmatropic rearrangements and shifts) have received less attention, most probably with the only exception of the Claisen rearrangement . Notably, thermally promoted H-shifts remain, to the best of our knowledge, completely unexplored in fulvene frameworks . A part of our recent research focused on showing the special ability of cyclic acetalic functions (1,3-dioxolanes, thiolanes, oxathiolanes, dioxanes, dithianes, oxathianes) for promoting the migration of its acetalic H atom in a hydride-like manner. As result, we have disclosed a variety of tandem processes initiated by -and -hydride shifts from the acetalic carbon atom toward electrophilic molecular fragments . Thus, we have reported that ortho- (1,3-dioxolan-2-yl)benzylidenemalonates 1 undergo tandem hydride shift/cyclization sequences leading to the corresponding indan-1-one-2,2-dicarboxylates 2. Remarkably, the first step of these processes consists of an uncommon -hydride shift of the acetalic H atom following the activation of the benzylidenemalonate fragment by scandium(III) triflate as the catalyst (Scheme 1) . Scheme 1: Lewis acid-catalyzed -H transfer/1,5-electrocyclization tandem processes of benzylidenemalonates 1 leading to indan-1-ones 2. Because fulvenes are widely used as the direct precursors of cyclopentadienyl anions following the addition of nucleophiles, including the hydride anion, to its exocyclic sp 2 carbon atom , we wondered whether acetalic functions could be employed as internal H donors in intramolecular hydride-like shifts, analogous to that highlighted in Scheme 1, toward fulvene frameworks. With this goal in mind we designed the unknown acetalfulvenes 3 (Scheme 2) as potential candidates for assaying the -hydride shift of its acetalic H atom toward the exocyclic C4 carbon atom of the fulvene fragment (note the numbering in the Scheme 2). At this point it is worth noting that other possible H migrations were not ruled out at the outset of this investigation. For example, C5 and C7 may be the respective termini of sigmatropic -H or -H shifts, whereas C6 could be also prone to participate in a less common -H shift . The variety of potential intramolecular H migrations in these reactive species might well justify by itself the research here disclosed. ## Experimental study The starting acetal-fulvenes 3 were prepared by the condensation of substituted 2-(1,3-dioxolan-2-yl)benzaldehydes 4 with cyclopentadiene following a well-established synthetic methodology . With the aim of promoting the desired hydride transfer by thermal activation, we first heated the parent acetalfulvene 3a under a variety of reaction conditions (benzene 110 °C sealed tube; toluene 120 °C sealed tube; DMF 120 °C) but unfortunately without success. Only when a DMSO solution of 3a was heated at 120 °C for 7 h the acetal-fulvene converted into a complex mixture from which we were able to isolate the benz[f]indenes 5a and 6a, in a relative 2:1 ratio and a poor global yield (34%). We next tested the same and similar processes in a microwave apparatus. As presumed, conversions of a series of acetal-fulvenes 3a-f under 120 W microwave irradiation at 120 °C in DMSO required much shorter reaction times (20-40 min) and led in all cases to the isolation of the respective benz[f]indenes 5 and 6, in a 2:1 ratio (Scheme 2). The overall yield of the isomeric mixtures 5 + 6 did not improve significantly with respect to the conventional thermal conditions previously used with 3a, ranging from medium to low as depicted in Table 1. Despite our chromatographic (column, thin-layer) efforts, additional pure products other than 5 and 6 could not be isolated from the complex final reaction mixtures. The structural determination of the reaction products, the 9-and 4-hydroxyalkoxy regioisomers 5 and 6, was easily accomplished by using the habitual analytical and spectroscopic techniques, whereas the distinction between each two regioisomers is basically supported by 1 H NMR NOE difference experiments. For the major isomers 5 enhancements of the signals due to the H-C2 and CH 2 -OAr protons were observed when the methylenic protons were irradiated. In contrast, similar irradia- A number of additional mechanistic alternatives arise by considering that the experimentally observed transformations of 3 are initiated by other H shifts alternative to the initially expected one. Thus, a -H shift from the acetalic carbon of 3a to the C5 carbon atom of the cyclopentadiene ring, would lead to the transient ortho-quinodimethane structure 9a, which might transform into two similar intermediates, 10a and 11a, by a sequence of consecutive -H shifts around the cyclopentadiene ring. Next, intermediates 10a and 11a would undergo 6π-electrocyclic ring closures (6π-ERC) to the respective dihydrobenzindenes 12a and 13a. Finally, these two species would experiment the acetalic ring opening by a formal β-elimination proccess leading to the respective final products 5a and 6a (Scheme 4). This mechanistic scheme is further complicated when considering that ortho-quinodimethane intermediates 10a and 11a could also result from the respective -H and -H shifts Scheme 4: Alternative mechanistic paths for the conversion 3a → 5a + 6a initiated by -, -or -H shifts. occurring in the starting acetal-fulvene 3a (Scheme 4). Thus, the different mechanistic paths represented in Scheme 4 for explaining the conversion 3a → 5a + 6a share several common steps, essentially differing in their first hydrogen shift, -, -or -H. Obviously, any attempt to discern which one is the actual reaction path (if only one!) among the range of potential mechanistic alternatives for these transformations seems a huge task. Seeking for additional experimental data in order to approach such objective, we reasoned that the mechanism initiated by a -hydride shift, as summarized in Scheme 3, could be differentiated from those starting by -, -or -H shifts, represented in Scheme 4, by deuterium labelling experiments. Thus, if the conversion of the deuterated acetal-fulvene 14, in which deuterium replaces the proton at the acetalic carbon of 3a, was actually initiated by a -deuteride shift, the transformation of the dihydrobenzo[f]indenic species 15, which should form in first instance, would yield the final benz[f]indenes via intermediates 16 and 17. These are labelled with deuterium at C4 of the major product 18 and at C9 of the minor one 19 as well as probably at additional positions of the fused five-membered ring (Scheme 5). With this in mind we prepared the monodeuterated acetalfulvene 14 (see Supporting Information File 1) and submitted it to the habitual reaction conditions. As result, a mixture of the monodeuterated benz[f]indenes 18 and 19 was obtained, again in a relative 2:1 ratio (Scheme 6). 1 H NMR analyses showed that only protons, not deuteriums, were linked to the C4 atom of 18 and to C9 of 19. Instead, one deuterium atom is found at the methylene group of each regioisomer. Moreover, we carried out a second labelling experiment by using now as starting material the monodeuterated acetalfulvene 20 bearing the deuterium atom at C4. This species converted into a 2:1 mixture of the monodeuterated regioisomers 21 and 22 as the only isolated reaction products (Scheme 6). In both compounds, the deuteration percentage at their respective C4 and C9 positions was determined, by 1 H NMR analyses, to be higher than 98%. This latter result shows that the deuterium atom attached at C4 in the original acetal-fulvene does not migrate in the course of the reaction. In combination, these two labelling experiments are conclusive for discarding an initial -deuteride (or hydride) shift as the first step of the previously discussed conversions since, in such a case, we should have found deuterium linked to C4/C9 of the benzindenes 18/19 resulting from the first labelling experiment. Next we tried to understand why the regioisomeric benz[f]indenes were, in all our reactions, produced in a ratio close to 2:1. It is well known that 1H-indenes are prone to undergo isomerization by H or group migrations at its cyclopentadiene ring . Consequently, we postulated that such an isomeric ratio would correspond to the thermodynamic equilibrium between both isomers, 5 and 6, established by two consecutive -H shifts around its five-membered ring. To test this hypothesis, we heated a 4:1 mixture of isomeric 5c and 6c in deuterated DMSO solution at 120 °C for 24 h. In this way we could verify by 1 H NMR analyses of reaction aliquots that the initial isomeric ratio remained constant over time and heating. Interestingly, the initial 4:1 ratio changed to 2:1 at the end of a related experiment carried out by stirring a DMSO solution of the same isomeric mixture in the presence of a catalytic amount of triethylamine at room temperature for 2 h (Scheme 7). This result seems to indicate that the experimentally recurrent 2:1 proportion between the regioisomeric products 5 and 6, reached by equilibration in the latter experiment, should be due to the presence of adventitious minor amounts of basic species either in the DMSO solutions of the experiments or in the course of the processing of the crude reaction mixtures and the purification steps. Besides, we also explored the thermally induced transformations of related fulvenes bearing non-cyclic acetalic units (Scheme 8 and Table 2). To this end, benzaldehydes 23 were transformed into fulvenes 24 by the usual procedure, whereas its microwave heating (DMSO, 120 °C, 120 W) yielded a mixture of the benz[f]indenes 25 and 26 in the habitual 2:1 ratio. These conversions most probably occur, in mechanistic terms, similarly to those of the acetal-fulvenes 3 (Scheme 4), although in the present cases with the formal β-elimination of a methanol or ethanol molecule. These results show that non-cyclic acetalic units are as effective as the cyclic ones on achieving the conversion of acetalfulvenes into the corresponding benz[f]indenes under microwave irradiation. ## Computational study With the aim of scrutinizing the putative reaction paths leading from the fulvene 3a to the isomeric benz[f]indenes 5a and 6a we have carried out a computational study at the B3LYP/6-31+G** theoretical level. Scheme 9 shows the diversity of the computed reaction paths leading from reactants to products. The geometries of the located transition structures associated to the first mechanistic step of each path, the H shift, are shown in Figure 1 . We anticipated three general reaction channels, paths A-C (see Scheme 9). In fact, these pathways only differ in the first step. Path A starts by a -H shift, path B by a -H shift and path C by a -H shift. This overall mechanistic scheme is somewhat complicated due to the number of steps of each reaction path and by the fact that some stationary points belong to more than one of these three pathways. Obviously we also envisaged a fourth mechanistic alternative, path D, just that initially conceived starting by the -hydride shift of the acetalic H atom to the exocyclic C4 carbon atom of the fulvene unit. All our efforts aimed to locate its corresponding transition structure were unsuccessful. Nevertheless, this latter mechanistic alternative was discarded by the isotopic labelling experiments commented above. In the following paragraphs we intend to discuss in a simplified way the results of our calculations on the potential surface of the transformations summarized in Scheme 9. The first step of path A consists of a -H shift from the acetalic carbon atom to C5 (see the numbering in Scheme 9). We located the transition structure TS1-A, connecting the fulvene 3a with the ortho-quinodimethane intermediate 9a. As expected TS1-A shows the typical geometry of a suprafacial hydrogen shift (see Figure 1), the computed energy barrier associated to this step being fairly high, 47.5 kcal•mol −1 . Intermediate 9a could then experiment two alternative -H shifts of its H-C5 proton, migrating either to C6 or to C9, its two vicinal carbon atoms at the cyclopentadiene ring. For the -H shift to C9 we located the transition structure TS2, 19.6 kcal•mol −1 above in energy than 9a, connecting it with its isomer 10a. For the alternative -H shift to C6 we located the transition structure TS3, 19.0 kcal•mol −1 above in energy than 9a, leading to the isomeric structure 10a' which is in fact a rotamer of 10a. These two energy barriers are reasonably low and should be easily surmountable under the experimental reaction conditions. Additionally, we were able to locate a transition structure, TS4, connecting 10a and 10a' by rotation around the C4-C5 single bond. The computed barrier for the conversion of 10 into 10a' via TS4 is only 7.0 kcal•mol −1 , whereas the one for the reverse transformation is 7.9 kcal•mol −1 . Accordingly, equilibration between 10a and 10a' is predicted to occur rapidly by C4-C5 bond rotation rather than by two consecutive -H shifts via the isomeric intermediate 9a (see Scheme 9). Intermediate 10a' can also convert into a third ortho-quinodimethane isomer 11a via the transition structure TS5 by another -H shift from C6 to its vicinal C7 carbon atom at the cyclopentadiene ring. The computed energy barrier for this step is 27.6 kcal•mol −1 , significantly higher than those corresponding to the similar -H shifts via TS2 and TS3 commented above (19.6 and 19.0 kcal•mol −1 ). The lower barriers of these two latter transition structures are attributable to its more extended conjugation in comparison with the partially crossconjugated TS5. A series of reaction steps starting from intermediates 10a and 11a can lead respectively to the final benzindenes 5a and 6a. Thus, intermediate 11a undergoes a disrotatory 6π-electrocyclic ring closure via the transition structure TS6 to give the tricyclic species 13a. By an analogous electrocyclic ring closure through TS7, compound 10a is converted into the isomeric spirotricycle 12a. The computed energy barriers for these processes are relatively small, 13.9 and 17.0 kcal•mol −1 respectively. Again the differences in the extent of the electronic conjugation in these electrocyclization transition states can give account of the relative stabilities of TS6 and TS7. Two transition structures, TS8 and TS9 were located for the respective transformations of 13a and 12a into the final benzin-denes 5a and 6a, involving each one the opening of the acetalic ring with simultaneous transfer of an hydrogen to one of the oxygen atoms (in other words, a concerted β-elimination along a C-C single bond), with the concomitant aromatization of the central ring. The computed energy barriers for these concerted β-eliminations are high, 41.7 and 40.8 kcal•mol −1 respectively . Concerning the alternative reaction paths B and C, we have located essentially the same stationary points that in path A with the sole difference of the respective first mechanistic steps. Path B starts with a -H sigmatropic rearrangement through TS1-B leading to intermediate 10a', which then transforms via the mechanistic paths commented above. The geometry of TS1-B (see Figure 1) is in accordance with a suprafacial transfer of the H atom between the acetalic carbon and C6. The calculated energy barrier associated to this step is 41.4 kcal•mol −1 , 6.1 kcal•mol −1 lower in energy than the initial -H shift of path A. This difference could be rationalized attending to the geometries of both transition structures TS1-A and TS1-B, more specifically to the distance between the two carbon atom termini of the H migration, shorter in TS1-B (2.63 ) than in TS1-A (2.67 ). Therefore, TS1-B is earlier than TS1-A. The geometry of TS1-B also accounts for its greater conjugation as the spatial positioning of the cyclopentadiene ring allows its orbital overlapping with the rest of the π system. Moreover, TS1-B is less sterically congested and also less distorted than TS1-A (see the bond distances and bond angles displayed in Figure 1). For the first step of path C we have located a transition structure, TS1-C, connecting fulvene 3a with the intermediate 11a by a -H sigmatropic shift (see Scheme 9). The computed energy barrier is very high, 64.3 kcal•mol −1 . This large value is probably due to the heptatrienic fragment not being able of adopting the helical all s-cis conformation, optimal for an antarafacial -H shift, as result of the conformational restrictions imposed by the cyclopentadiene ring (Figure 1). As a consequence, the distance between the two carbon atom termini of the H migration is considerably long (2.88 ), thus accounting for the high computed energy barrier. To summarize so far, by comparing the energy barriers associated to the three alternative H shifts, this study predicts that path B is the one involving the lowest energy barrier and, in accordance, the calculations predict that the transformation of fulvene 3a into the benzindenes 5a and 6a should take place via an initial -H shift. Moreover, we also considered that 5a and 6a could equilibrate by two consecutive -H shifts occurring at the five-mem-bered ring. By exploring the potential energy surface associated to these transformations we were able to locate transition structures TS10 and TS11 connecting 5a and 6a through the intermediate 27a (see Scheme 9). The computed energy barriers for the conversions 5a → 27a, and 6a → 27a are fairly high, 43.2 and 42.6 kcal•mol −1 , respectively, as expected on going from a fully aromatic central ring to an orthoquinoid structure, whereas those calculated for the reverse conversions are considerably lower (13.3 kcal•mol −1 in both cases). By analysing the overall picture showing the different mechanistic paths connecting 3a with the final benzindenes 5a and 6a we can extract the following conclusions: 1) On going from 3a to the two tricyclic intermediates 12a and 13a, the rate determining reaction step is predicted to be the first one, i.e. the initial hydrogen migration, and this study predicts that a -H shift is less costly in terms of energy than a -H or -H one. The more extended conjugation, the lower steric hindrance and the shorter C-C distance between the two carbon atoms termini of the H migration in TS1-B, the transition structure of the -H shift, can account for its lower energy when compared with those of the alternative two other H shifts. Nevertheless, the higher electronic conjugation in TS1-B, in comparison with those of TS1-A and TS1-C, could be also decisive in accounting for the differences in the respective energy barriers. 2) Concerning the two key polyenes 10a and 11a, precursors of the tricycles 12a and 13a, respectively, the computed energy barriers of this study show that i) 10a most probably forms by an easy rotational isomerization of 10a', instead of the alternative path involving the -H shift from 9a; and ii) 11a will form mainly from 10a' rather than directly from 3a. That is, 10a and 11a should form via intermediate 10a' resulting from the -H shift, which then transforms into 11a by a -H shift (barrier of 27.6 kcal•mol −1 ) or equilibrates to its rotamer 10a (barrier of 7.0 kcal•mol −1 ). 3) The two 6π-electrocyclic ring closures converting respectively 10a and 11a into 12a and 13a involve low energy barriers, that corresponding to the conversion of 11a into 13a being lower than that of 10a into 12a (13.9 and 17.0 kcal•mol −1 , respectively). 4) The overall processes 3a → 5a and 3a → 6a are exothermic by 31.3 and 30.7 kcal•mol −1 , respectively. The interconversion between 5a and 6a is predicted to take place via two consecutive -H shifts at the pentagonal ring through transient intermediate 27a. As a final point, we have also explored the potential energy surface associated to these conversions by considering the effect of the solvent used in the experimental study, DMSO. The computed energy barriers in the gas phase and in DMSO are depicted in Table 3. In general, the values of the energy barriers do not vary noticeably in DMSO when compared with those in gas phase. Only ΔE 8 and ΔE 9 are appreciably lower in DMSO with respect to those in gas phase by 3.9 and 3.5 kcal•mol −1 , respectively (Table 3). Consequently, according to these calculations, the rate determining step in the transformations 3a → 5a and 3a → 6a in DMSO should be the first one, i.e. the -H shift, with an energy barrier slightly lower than that calculated in the gas phase. In summary, this computational study shows that the conversion of fulvene 3a into the benzindenes 5a and 6a could take place by a variety of alternative reaction paths according to a complicated mechanistic scheme. By analysing in detail the energy barriers computed for each mechanistic step, the energetically preferred path starts with a -H sigmatropic rearrangement of the acetalic hydrogen atom leading to an orthoquinodimethane intermediate, further transforming into the isomeric final products by two alternative reaction channels. These two latter pathways may involve up to three consecutive steps such as -H shifts, 6π-electrocyclic ring closures, C-C rotations and formal β-eliminations. The interconversion between the isomeric benzindenes 5a and 6a could also occur by means of two consecutive -H shifts through an unstable benzisoindene intermediate. ## Conclusion The ability of benzofulvenes bearing 1,3-dioxolane or -dioxane units in ortho position for undergoing cascade processes initiated by an H shift step has been tested. Such acetal-fulvenes, under thermal activation, transformed into mixtures of the corresponding 4-and 9-(hydroxy)alkoxy-substituted benz[f]indenes in a 1:2 ratio. Analogous fulvenes bearing noncyclic dialkoxymethyl units when submitted to similar thermal conditions also afforded 1:2 mixtures of the respective 4 and 9-alkoxybenz[f]indenes. Such 1:2 ratio has been interpreted as the one corresponding to the thermodynamic equilibrium established between both isomers. Mechanistic paths initiated by an initial -, -, -or -H shift are conceivable for explaining these cascade transformations leading to benz[f]indenes. The results of deuterium labelling experiments excluded a -hydride shift as the initial step. The reaction of the unsubstituted 1,3-dioxolane-fulvene has been computationally studied by DFT methods. The results of this study revealed that the first mechanistic step, the H shift, is the rate-determining one and that, among the alternative -, -or -H migrations, the energy barrier of the -H shift is the lowest one, a fact that is rationalised attending to some key structural and electronic characteristics of the respective transition states. The calculations have also shown that the tandem conversions of the starting fulvenes into benz[f]indenes are exergonics, the 9-substituted regioisomer being the thermodynamically-controlled major product, in accordance with the experimental results.

chemsum

{"title": "Tandem processes promoted by a hydrogen shift in 6-arylfulvenes bearing acetalic units at <i>ortho</i> position: a combined experimental and computational study", "journal": "Beilstein"}

highly_stable_low_redox_potential_quinone_for_aqueous_flow_batteries

## Abstract: Aqueous organic redox flow batteries are promising candidates for large-scale energy storage.However, the design of stable and inexpensive electrolytes is challenging. Here, we report a highly stable, low redox potential, and potentially inexpensive negolyte species, sodium 3,3',3'',3'''-((9,10anthraquinone-2,6-diyl)bis(azanetriyl))tetrakis(propane-1-sulfonate) (2,6-N-TSAQ), which is synthesized in a single step from inexpensive precursors. Pairing 2,6-N-TSAQ with potassium ferrocyanide at pH 14 yielded a battery with the highest open-circuit voltage, 1.14 V, of any anthraquinone-based cell with a capacity fade rate <10%/yr. When 2,6-N-TSAQ was cycled at neutral pH, it exhibited two orders of magnitude higher capacity fade rate. The great difference in anthraquinone cycling stability at different pH is interpreted in terms of the thermodynamics of the anthrone formation reaction. This work shows the great potential of organic synthetic chemistry for the development of viable flow battery electrolytes and demonstrates the remarkable performance improvements achievable with an understanding of decomposition mechanisms. ## Introduction Safe and economical energy storage technologies are indispensable for the deep penetration of intermittent renewable energies such as photovoltaic and wind electricity. Aqueous redox flow batteries are promising candidates for large-scale energy storage compared to other storage devices such as pumped-hydro, flywheel, and lithium-ion batteries, owing to the highly modular configuration, long cycle life, and good safety features. 2,3 Aqueous vanadium redox flow batteries (VRFBs) have been successfully established by many manufacturers, due to their long cycling life and high-power density. 4 However, cost reductions in VRFBs are anticipated to be difficult due to the abundance of vanadium and its fluctuating price. Consequently, aqueous organic redox flow batteries (AORFBs) are attracting tremendous research interest, as the redox active materials comprising earth abundant elements are potentially inexpensive. 3, Additionally, the physical and electrochemical properties of redox organics, such as aqueous solubility, molecular size, molecular net charge, redox potential, and chemical stability could be tailored for improved performance via molecular functionalization. 8,9 One drawback of many reported AORFBs, however, is their fast capacity fade because redox organics are susceptible to degradation reactions such as nucleophilic substitution, disproportionation, and tautomerization. 3 To date, various redox-active organics based on quinone, viologen, phenazine, alloxazine, 36 ferrocene 23,24,37,38 and nitroxide radical derivatives 22,28,32,39 have been reported for AORFBs. Most of them, however, exhibit high capacity fade rates of 0.1%-1%/day, 3 which is unsuitable for practical application. Recently, anthraquinone derivatives such as 2,6-DBEAQ, 2,6-DPPEAQ, DPivOHAQ and DBAQ have demonstrated very good long-term stability. 11,16,19 However, their widespread application is hindered by the high synthetic cost due to sophisticated synthesis or expensive precursors involved. Additionally, there is an apparent trade-off between anthraquinone cycling stability and the redox potential. 40 The highly stable anthraquinones such as 2,6-DBEAQ, DPivOHAQ, and DBAQ have a redox potential more positive than -0.52 V vs. standard hydrogen electrode (SHE) at pH 12 and above; anthraquinones with a more negative redox potential exhibited less stable cycling performance. 8,41 For a negolyte molecule, however, a low redox potential is desired to achieve high cell voltage. Therefore, developing inexpensive and stable anthraquinone negolytes with a low redox potential remains crucial for the practical implementation of AORFBs. Here, we report a potentially inexpensive and low redox-potential anthraquinone negolyte with outstanding cycling stability. The anthraquinone sodium 3,3',3'',3'''-((9,10-anthraquinone-2,6diyl)bis(azanetriyl))tetrakis(propane-1-sulfonate) ( negolytes and highlight the great potential of organic synthesis towards inexpensive and stable electrolytes for grid-scale energy storage application. Given the higher water solubility, lower redox potential, and possible lower synthetic cost, 2,6-N-TSAQ was selected for further electrochemical study. The advantage of 2,6-N-TSAQ over some other low redox potential anthraquinones 8,41 is that it has four negative charges on the solubilizing groups, leading to a high intermolecular Coulomb repulsion and a low collision rate. ## Results and Discussion According to Marcus theory, 43,44 these properties could decrease the reaction rate of the disproportionation (known to cause capacity decay in anthraquinone negolyte). 19,40 The large Coulomb repulsion and bulky functionalization also decrease the molecular permeability across cation exchange membranes, increasing the cell lifetime. The permeability of 2,6-N-TSAQ through sodium-exchanged Nafion NR212 was measured in a two-compartment diffusion cell. Due to a very low crossover rate, we estimate a maximum permeability of 3 x 10 -14 cm 2 /s (Figures S6 and S7), which is even lower than that reported for the tetra-anionic anthraquinone derivative 2,6-DPPEAQ. 16 The Pourbaix diagram of 2,6-N-TSAQ, shown in Figure 3a, indicates the molecule undergoes a two-proton/two-electron process below pH 10, a one-proton/two-electron process over pH 10-12, and a pH-independent two-electron process at pH > 12 with a redox potential around -0.63 V vs. SHE. The corresponding CV profiles at various pH are shown in Figure S8. It should be noted that the pH is the local pH of anthraquinone molecules. For an unbuffered case, e.g., 1 M NaCl, the formal potential of 2,6-N-TSAQ is -0.62 V, which is close to the redox potential at high pH; such a phenomenon was also observed in other anthraquinones when a pH buffer was not used. 16,17 Based on the Pourbaix diagram, the pKa1 and pKa2 of reduced and protonated 2,6-N-TSAQ are estimated to have values around 10 and 12, respectively, which are slightly larger than those of anthraquinone negolytes with more positive redox potentials. 9,11,16,19 We attribute this to the strong electron donating effect of lone pair electrons on nitrogen atoms decreasing the pKa of hydroxy groups of the 9,10-dihydroxyanthracene (reduced state of anthraquinone). was charged/discharged at 40 mA/cm 2 between 0.6 V and 1.4 V with a potential hold until the current dropped to 2 mA/cm 2 to get the full capacity. The OCV increased from 0.8 to 1.31 V as the SOC increased from ~0% to ∼100% (Figure 3a). The 0.2 V increase of OCV from 0 to ~1% SOC, 0.09 V increase from ~1% to ~10%, 0.08 V increase from 10% to 90% SOC, and 0.14 V from 90% to final OCV, indicates the utilization of 2,6-N-TSAQ is more than 99% under the operating conditions according to the Nernst equation. The peak galvanic power density at 10% SOC was 0.15 W cm -2 and increased to 0.18 W cm -2 at 90% SOC (Figure 3d). The power density is mainly limited by the high-frequency ASR, which is dominated by the membrane resistance (Figure 3c) with a value around 1.6 Ω•cm 2 . Therefore, the power density is expected to be improved with a lower-resistance membrane. A long-term cycling test of the 0.1 M 2,6-N-TSAQ/ferrocyanide flow battery at pH 14 was performed with the same cell. The cell was cycled at 40 mA cm -2 with potential holds at 1.4 V for charging and 0.6 V for discharging until the current density dropped to 2 mA cm -2 . The initial volumetric discharge capacity was 4.76 Ah/L, corresponding to a capacity utilization of 88.9% of the theoretical value. However, the OCV at different SOCs in Figure 3c 0.1 0.01 8.314 J/mol/K 298.15 K (ln ln ) 6.18 kJ/mol 0.9 0.99 × × − = assuming the variation of anthrone concentration is negligible. Likewise, when the SOC of anthraquinone negolyte increases from 90% to 99.9%, the Gibbs free energy change for anthrone formation becomes more negative by approximately 11.93 kJ/mol. Therefore, to suppress anthrone formation in practical deployment, charging to a high SOC should be avoided, i.e., it is desired not to conduct a potential hold at the end of the charging half-cycle. A simple galvanostatic cycling protocol, however, cannot be used in research to evaluate very low capacity fade rates. 3,45 . For future research on anthraquinone negolytes species, a potential hold after, say, every 30 cycles of galvanostatic cycling might be advisable. In summary, we synthesized three sulfonated anthraquinone derivatives, carbon-linked, nitrogen-linked, and oxygen-linked. The nitrogen-linked anthraquinone (2,6-N-TSAQ) showed a much lower redox potential than the others due to the strongest electron donating effect of lone pair electrons on nitrogen atoms. Because it is synthesized from inexpensive precursor with a one-step N-alkylation method, the mass production cost could be low. Despite the Coulomb repulsion afforded by its four negatively charged sulfonate groups, the cycling performance of 2,6-diaminoanthraquione (97%), 1,3-propanesultone (98%), sodium hydride (60% in mineral oil), anhydrous dimethyl sulfoxide, anhydrous N,N-Dimethylformamide, potassium carbonate, and palladium(II) acetate (98%) were purchased from Sigma-Aldrich. 2,6dihydroxyanthraquinone (98%) was purchased from AK scientific. Sodium allylsulfonate (94%) was purchased from Ambeed, inc. Ltd. Hydrogen was purchased from Airgas. The materials were directly used without further purification. ## Synthesis of 2,6-N-TSAQ 3 g of 2,6-diaminoanthraquinone (12.59 mmol) was added to 50 mL anhydrous dimethyl sulfoxide. Then 2.1 g sodium hydride (60%, 52.46 mmol) was added to the solution under vigorous stirring. After 15 minutes, 6.41 g 1,3-propanesultone (98%, 52.46 mmol) was added to the above mixture. The solution was stirred at room temperature for overnight. Afterward, ethyl acetate was added to the solution and collect the red solid. The crude product was further washed with ethyl acetate to remove any mineral oil. Yield:9.7 g (95%). The 1 H NMR spectrum of 2,6-N-TSAQ is shown in Figure S1. Afterwards, ethyl acetate was added to the solution to collect the yellow solid. The crude product was washed with ethyl acetate to remove any mineral oil. Yield: 6.4 g (97%). The 1 H NMR spectrum of 2,6-O-DPSAQ is shown in Figure S3. Evaporate the solution in vacuum to collect the solid. Yield: 1.51 g (70%). The 1 H NMR spectrum of 2,6-DPSAQ is shown in Figure S4. This membrane was sealed between a donating compartment containing 10 mL of 0.1 M 2,6-N-TSAQ in 1 M NaOH and a receiving compartment containing 10 mL of 0.17 M Na2SO4 in 1 M NaOH. The electrolyte in the receiving compartment was designed to minimize osmotic pressure gradients influencing permeability: a freezing point osmometer (Advanced Instruments Inc., Model 3300) confirmed the osmolarity difference between the compartments was only 0.018 Osm. Both compartments were stirred continuously using magnetic stir bars. Three identical H-cells stirred for 13 days, and 2 mL aliquots were periodically removed from the receiving side to measure absorbance spectra, which were then replaced with fresh receiving solution. Due to the exceptionally low crossover rate, the spectrophotometer was unable to detect the peaks characteristic of 2,6-N-TSAQ, so an upper limit was assigned based on the highest absorbance value observed at 455 nm during the experiment (Fig S8). Using the derivation of Fick's Law reported previously, 1 2,6-N-TSAQ cannot exceed 3 × 10 -14 cm 2 /s under these conditions.

chemsum

{"title": "Highly Stable Low Redox Potential Quinone for Aqueous Flow Batteries", "journal": "ChemRxiv"}

intermolecular_dearomative_[4+2]_cycloaddition_of_naphthalenes_via_visible-light_energy-transfer-cat

## Abstract: Dearomative cycloaddition reaction serves as a blueprint for creating three-dimensional molecular topology from flat-aromatic compounds. However, severe reactivity and selectivity issues make this process challenging. Herein, we describe visible-light energy-transfer catalysis for the intermolecular dearomative [4+2] cycloaddition reaction of feed-stock naphthalene molecules with vinyl benzenes. Tolerating a wide range of functional groups, a variety of structurally diverse 2acyl naphthalenes and styrenes could easily be converted to a diverse range of bicyclo[2.2.2]octa-2,5-diene scaffolds in high yields and selectivities. The late-stage modification of pharmaceutical agents further demonstrated the broad potentiality of this methodology. The efficacy of the introduced methods was further highlighted by the post-synthetic diversification of the products. Furthermore, photoluminescence, electrochemical, kinetic, and control experiments support the energy transfer catalysis Constructing three-dimensional (3D) molecular scaffolds from two-dimensional (2D) molecules is highly challenging yet significantly impacts organic synthesis and drug discovery programs. [1][2][3][4] The cycloaddition reactions that convergently and predictably join multiple molecular fragments have been recognized as a powerful tool for this purpose (Figure 1). [5][6][7][8] Notably, in [4+2] cycloaddition reaction, two new σ-bonds, and one π-bond are formed in a 3D six-membered ring topology from two simple unsaturated reaction components, diene, and dienophile (Figure 1a). 9, 10 In fact, this thermally allowed process has been a fundamental reaction type demonstrating its molecular complexity generating power for many years. [11][12][13] In this context, polycyclic aromatic hydrocarbon such as naphthalene also contains alternating double bonds. Besides, they are abundant and inexpensive feed-stock chemicals (present at about 10% in typical coal tar). 14 However, these 2D molecules displayed limited application in 3D complexity generating cycloadditions reactions due to severe challenges associated with breaking the increased stabilization conferred by aromaticity (resonance energy = 80.3 kcal mol -1 ) and selectivity (Figure 1b,c). 15,16 A typical thermal dearomative [4+2] cycloaddition with naphthalenes required harsh reaction conditions (high temperature up to 210 o C, pressure up to 10 3 atm), 17 specially designed reaction conditions, [18][19][20] or reactive dienophile 21,22 (hereafter called arenophile) to overcome the high kinetic barrier (Figure 1c). 23 22, 24 However, since the free energy is often positive for such a reaction, the reverse reaction is thermodynamically preferred resulting in lower product yields (Figure 1d, blue curve). 19, 25 Photochemistry provides alternative strategies for achieving challenging and unusual chemical transformations in this context. 23,26,27 However, since most organic molecules are incapable of absorbing visible light efficiently, direct high-energy ultraviolet (UV) light irradiation is required. Indeed, the UV-light mediated dearomative [4+2] cycloaddition with naphthalenes is known. 28,29 However, their utility in organic synthesis is minimal due to the requirement of specific arenophiles, meager product yields, and unpredictable side reactions conferred by UV light (Figure 1c). Eliminating UV irradiation should ideally broaden the synthetic applicability of this process with enriched structural diversity. The recent renaissance of visible-light photocatalysis provides a new space for dearomative [4+2] cycloaddition reaction via sensitization induced energy transfer (EnT) catalysis 30, 31 or direct visible-light excitation of the dienophile in some cases. 23 Conceptually, the EnT process can selectively excite a ground state of a polycyclic hydrocarbon by using an appropriate photosensitizer to a higher triplet state (naphthalenes exhibit ETs of 54-60 kcal mol -1 ), 32 lowering the kinetic barriers significantly compared to thermal processes (Figure 1d, black curve). 33 Furthermore, the milder reaction conditions and substantially higher ET of the dearomatized product prevent the reverse reaction resulting in higher product yields.Recently, Glorius and coworkers demonstrated dearomative [4+2] cycloaddition of pyridines (intramolecular) 30 and bicyclic azaarenes (intermolecular, Figure 1e) 33 via visible-light EnT catalysis. A stoichiometric Brønsted acid additive was shown to play a vital role in the latter reaction to increase the reactivity of quinolines' triplet state toward olefins. 31 You and coworkers reported intramolecular dearomative cycloaddition of indole tethered naphthalenes (Figure 1f). 34 The dearomative intramolecular [2+2] cycloaddition of 1-naphthol derivatives via visible-light EnT catalysis was recently developed by Glorius. 35 The intramolecularity prepaid the entropic requirements for the last two reactions. Besides, various groups demonstrated the application of the EnT process in diverse chemical transformations. However, to the best of our knowledge, intermolecular dearomative [4+2] cycloaddition reactions of naphthalenes with unactivated alkenes have not been documented yet. Herein, we report our initial results on visible-light EnT mediated dearomative [4+2] cycloaddition of naphthalenes with styrenes to access bicyclo[2.2.2]octa-2,5-diene scaffolds (Figure 1g). We commenced the visible-light mediated dearomative [4+2] cycloaddition reaction using 2-acetyl naphthalene 1 and 1fluoro-4-vinyl benzene 2 as the model substrates (Table 1, Tables S1-S6). Pleasingly, the irradiation of blue light-emitting diodes (λmax = 427 nm) in the presence of commercially available photosensitizer Ir[(dFCF3ppy)2dtbbpy]PF6 (PC1, 1 mol%) in acetonitrile solution of 1 and 2 (in stoichiometric ratio) at room temperature resulted in the formation of the desired [4+2] cycloadduct 3 in 98% yield in 2:1 diastereomeric ratio, favoring endo (entry 1). Notably, the regioisomeric product 4 resulting from the syn attack of the alkene to the 6-membered aromatic ring was not observed by 1 H nuclear magnetic resonance analysis of the crude reaction mixture. The ortho and meta cycloaddition products (Figure 1b) were also undetected in a measurable amount. We then followed the kinetics of the reaction via 1 H NMR spectroscopy (Figure S18). The rate of formation of 3 was found to be almost equal to the disappearance of 1 (Figure S19). This supports the direct formation of [4+2] cycloadduct 3 without the intermediacy of an [2+2] adduct as previously been observed for the intramolecular reaction. 34 Product 3 was computed to be thermodynamically uphill (Figure S21). However, it can be purified by flash chromatography on silica gel and was found to be stable when heated at 60 o C for 24 h. The structure of endo-3 was confirmed by single-crystal X-ray crystallography. 63 Control experiments confirm that the light source and photosensitizer were necessary for this [4+2] cycloaddition reaction, and their absence led to no product formation (entries 2-3). As anticipated, the reaction does not provide any product formation when heated at 150 o C for 24 h (entry 4). We then stride forward to gain more insights into the EnT mediated dearomative [4+2] cycloaddition reaction (Figure 2). The cyclic voltammetry analysis suggested that an electron transfer from the excited PC1 (E1/2Ir(III)*/Ir(IV) = -0.89 V vs SCE) to either 1 or 2 (reduction potentials E1/2 = -1.81 V and -1.42 V vs SCE, respectively) is not thermodynamically feasible (Figure 2a). Similarly, an oxidative quenching of the excited PC1 (E1/2Ir(III)*/Ir(II) = 1.21 V vs SCE) can also be ruled out. However, the Stern-Volmer analysis shows that 2-acetyl naphthalene 1 efficiently quenches the luminescence emission of the excited photosensitizer (Stern-Volmer quenching constant kq = 1.6 ×10 4 M -1 s -1 ) (Figure 2b, c). In comparison, vinyl benzene 2 quenched the excited PC1 at a lower rate (figure 2c). On the other hand, we have computed the S0-T1 gap for 1 and 2 using B3LYP/6-311+G(2d,p) level of density functional theory (Figure S20). The calculated triplet energies ET = 55.1 kcal mol -1 for 1 and 57.9 kcal mol -1 for 2 are in good agreement with the previously reported values. 58 Accordingly, the triplet energy transfer from the photoexcited PC1 (ET = 60.1 kcal mol -1 ) 39 to substrate 1 is more likely. The reaction was also found to be amenable to some triplet-energy sensitizers, albeit at lesser efficiencies (Table 1, entries 5-9). However, a direct correlation between the ETs of the catalysts and product yields could not be drawn. 43 To gain more insight into the interaction of other photosensitizers with the substrates 1 and 2 and their efficiencies in catalyzing the EnT mediated cycloaddition reaction, we have performed the luminescence quenching experiments with Ir(ppy)3 (PC3) and 4CzIPN (PC5). Both the catalysts give a moderate to low yield of 3. We have found both 1 and 2 do not interact with the excited PC3 at an appreciable rate (Figure 2d). The luminescence intensity of excited PC5 is moderately quenched by increasing the concentration of 1 (kq = 0.30 ×10 4 M -1 s -1 ) (Figure 2e). However, 2 does not communicate with the excited PC5. In addition, it was confirmed that an electron donor-acceptor complex between 1 and 2 does not occur, as evidenced by the lack of a change in the UV/vis spectrum (Figure 2f). The absorption spectrum of 2 possesses small residual absorption at 370 nm (molar absorptivity ε = 48.1 M -1 cm -1 at 370 nm). However, direct irradiation using 370 nm Kessil LEDs (emission spectral window of ∼360 to ~410 nm) only gave a 35% yield of 3 (d.r. = 2:1) (Figure 2g). These results additionally suggest that the photochemical dearomative cycloaddition reaction of 1 is more efficiently triggered upon energy transfer from the excited PC1. Additionally, complete inhibition of the dearomative [4+2] cycloaddition reaction under oxygen (a well-known ET quencher) further supported the triplet energy transfer process (figure 2h). Based on the observations outlined above and the literature report, it is proposed that the photosensitizer upon visible-light excitation produces its long-lived triplet excited state. The naphthalene 1 is then activated via EnT to afford triplet intermediate 1 (T1). The latter is then engaged in [4+2] cycloaddition with styrene 2 to yield the cycloadduct 3 via radical capture, intersystem crossing (ISC), and bond formation (Figure 2i). The orbital coefficients of the relevant frontier molecular orbitals were examined to understand the origin of regioselectivity (see page S67-S68). The preferential formation of 3 can approximately be conceptualized considering the maximum interaction between the largest orbital components in the triplet intermediate 1 (T1) and the lowest unoccupied molecular orbital of 2. However, a detailed computational elaboration is necessary. Finally, we explored the generality, and synthetic utility of the EnT mediated dearomative [4+2] cycloaddition reaction (Figure 3). First, the scalability of our protocol was tested for a 1.0 mmol scale reaction between 1 and 2 that provided 0.25 g (84% yield) of 3 with the same diastereoselectivity. Then, the scope of the arenophile was investigated. Halogenated (F, Cl, Br) styrenes were noticed to be excellent coupling partners, yielding bicyclo[2.2.2]octa-2,5-diene scaffolds 3-8 in high yields and selectivities. The electronic nature of the substituents does not perturb the reaction. Vinyl benzenes with both electronwithdrawing (CF3) and electron-donating substituents (OMe, OAc, Me, SMe) at different positions of the aryl ring participated in this reaction in equal efficiencies. The adducts 9-16 were obtained in 25-80% yields with a similar diastereomeric ratio. Interestingly, vinyl naphthalene could also be utilized as an arenophile, and the adduct 17 was isolated in 39% yield. It indicated selective energy transfer to 1. Notably, vinyl biphenyls containing diverse electronic substituents, heteroarenes, and 4-vinyl pyridines were compatible with this reaction, and the dearomative [4+2] cycloadducts 18-20 were obtained in moderate to good yields. However, αand β-methyl styrene fails to deliver the product in synthetically useful yields under these conditions. Pleasingly, the reaction was compatible with styrene derivatives of drug molecules, including gemfibrozil, clofibric acid, fenbufen, ketoprofen, and oleic acid. The adducts 21-25 were isolated in 64-80% yields and moderate diastereoselectivities. It highlighted the functional group compatibility and synthetic utility of the process for late-stage diversification of these molecules. The scope of naphthalene derivatives was then tested. Acyl naphthalenes with diverse alkyl chains participate in this EnT mediated [4+2] cycloaddition reaction, delivering the adducts 26-28 in high yields and selectivities. Naphthalenes containing cyclic alkyl groups, including cyclobutene 29 and cyclopropane 30, were also viable reaction partners. A CC-bond cleavage in strained cyclopropane rings was not observed. In comparison, Brown recently discovered EnT mediated CC-bond cleavage in bicyclo[1.1.0]butyl naphthyl ketone for [2π + 2σ] cycloaddition reactions with styrenes. 64 Severe ring strain in the caged molecule can be reasoned responsible. N-Methyl imidazole bearing keto naphthalene performed well in this reaction 31. The structures of both the endo-and exo-diastereomers were assigned by single-crystal X-ray crystallography. 63 Interestingly, 2carboxyalkyl naphthalenes also smoothly undergo visible-light mediated dearomative cycloaddition reaction providing high 80-85% yields of the cycloadducts 32-34, albeit in a moderate diastereomeric ratio. The reaction was also compatible with the naphthalene derivatives of (-)-menthol 35, (-)-borneol 36, and cholesterol 37. It again highlights the utility of the visible-lightdriven dearomatization reaction. To further demonstrate the synthetic application of the [4+2] dearomative cycloaddition reaction, we have performed the derivatization of the cycloadduct endo-3 to obtain complex 3D molecular architecture. The palladium-catalyzed reduction of the α,β-unsaturated double bond of endo-3 in the presence of diphenyl silane gave the saturated product 38 in 52% yield in >20:1 d.r. NOE spectroscopy confirmed the stereochemical assignment. Furthermore, a Michael addition of N-Boc-L-cysteine methyl ester delivered the adduct 39 in 56% yield and 2:1 d.r. Notably, the Corey-Chaykovsky cyclopropanation of endo-3 exclusively offered the cyclopropane 40 in 63% yield, >99:1 d.r. Similarly, the epoxidation of endo-3 in the presence of H2O2/NaOH delivered the epoxide 41 in 70% isolated yield as a single diastereomer. In conclusion, a new strategy for intermolecular dearomative [4+2] cycloaddition reaction of abundant aromatic compound naphthalene was developed. The reaction operates via visible-light energy transfer utilizing a commercially available sensitizer. Diverse bicyclo[2.2.2]octa-2,5-diene scaffolds were synthesized in high yields. The reaction can be scaled-up and is amicable for late-stage modification of various bioactive molecules. The ease of post-synthetic diversifications adds to the utility of the introduced method. Photoluminescence, electrochemical, kinetic, and control experiments helped understand the features of this visible-energy transfer catalysis. We anticipate that the protocol will further illuminate the EnT catalysis for designing complex molecular scaffolds. Further research exploring the reaction scope and mechanisms is currently underway in this laboratory.

chemsum

{"title": "Intermolecular Dearomative [4+2] Cycloaddition of Naphthalenes via Visible-Light Energy-Transfer-Catalysis", "journal": "ChemRxiv"}

a_hyphenated_preconcentrator-infrared-hollow-waveguide_sensor_system_for_n2o_sensing

## Abstract: Following the Kyoto protocol, all signatory countries must provide an annual inventory of greenhousegas emission including N 2 O. This fact associated with the wide variety of sources for N 2 O emissions requires appropriate sensor technologies facilitating in-situ monitoring, compact dimensions, ease of operation, and sufficient sensitivity for addressing such emission scenarios. In this contribution, we therefore describe an innovative portable mid-infrared chemical sensor system for quantifying gaseous N 2 O via coupling a substrate-integrated hollow waveguide (iHWG) simultaneously serving as highly miniaturized mid-infrared photon conduit and gas cell to a custom-made preconcentrator. N 2 O was collected onto a solid sorbent material packed into the preconcentrator unit, and then released via thermal desorption into the iHWG-MIR sensor utilizing a compact Fourier transform infrared (FTIR) spectrometer for molecularly selective spectroscopic detection with a limit of detection (LOD) at 5 ppbv. Highlighting the device flexibility in terms of sampling time, flow-rate, and iHWG design facilitates tailoring the developed preconcentrator-iHWG device towards a wide variety of application scenarios ranging from soil and aquatic emission monitoring and drone-or unmanned aerial vehicle (UAV)mounted monitoring systems to clinical/medical analysis scenarios.Nitrogen was first identified as an essential nutrient for all plants and humans in the middle of the 19 th century 1,2 . Since the remarkable invention of the Haber-Bosch process in the beginning of the last century, an unlimited supply of Nr has been deposited worldwide, thus enabling an abrupt growth of the human population due the increase in food availability 3 .Approximately 100 years after the introduction of the Haber-Bosch process, a study led by Rockstrom and colleagues 4 proposed a safe operating space for mankind, is associates the earth as an eco-system with anthropogenic procedures and their thresholds. If certain boundaries are crossed, unacceptable environmental changes may be initiated putting the rather delicate environmental balance at risk. Among the procedures that have already extensively exceeded the safety threshold is the nitrogen cycle. As mentioned, large amounts of reactive nitrogen are annually applied into the soil during food production as a major component of fertilizers. Recent studies have shown that less then 40% of Nr is used by plants, thus resulting in an excess of reactive nitrogen spreading into various environmental compartments 3 . Most nitrogen-based fertilizers use soluble nitrates, urea or ammonium. Consequently, lixiviation of soluble Nr to aquatic systems such as rivers, lakes, estuaries, and aquifers has been described of major environmental concern 5 . This fact results in thousands of tons of Nr annually moved to atmosphere as volatile forms (NH 3 , NOx and N 2 O), which contributes to extend the range of this pollution to regional and even global scale 6,7 .According to the nitrogen cycle, both bacterial nitrification and denitrification pathways are responsible for the emission of nitrous oxide into the atmosphere 8 . Since the introduction of the Haber-Bosch process, the background levels of N 2 O in atmosphere have increased by almost 20% from 270 to 330 ppb 8 . Nitrous oxide is considered a major greenhouse gas next to carbon dioxide and methane with an estimated contribution of approx. 6% to global warming, despite the fact that N 2 O is present at much lower concentration than CO 2 . The global warming potential (GWP) of N 2 O is approximately three hundred times larger than CO 2 . Besides, in the stratosphere N 2 O is oxidized to NO and NO 2 , which are both among the constituents responsible for the destruction of the ozone layer. Hence, next to its role as environmental contaminant, nitrous oxide is considered among the most relevant ozone-depleting gas during in the 20 th century 9 . According to the Kyoto protocol, all signatory nations must provide an annual greenhouse gas inventory, which should include N 2 O 10 . However, establishing the N 2 O budget is non-trivial due to (i) the low concentration of N 2 O in the atmosphere (i.e., in the ppbv range) when compared to CO 2 , and (ii) the spatial and temporal variations in the episodes of N 2 O emission from soil and aquatic systems 9 . In addition, a wide range of factors affects the rate of emission including but not limited to pH, temperature, moisture, soil composition, and microflora 5 . Currently, extensive research has been focused on fundamentally understanding the processes involved in N 2 O emissions from soil and aquatic systems, and have triggered substantial interest in the to the development of analytical strategies and methods providing reliable and accurate data . Latent emissions provided by aquatic systems and soil reactions are believed to be the dominating mechanisms for N 2 O release into the atmosphere. For characterizing such scenarios, the development of sensor systems capable of in-situ in-field determination with suitable sensitivity and reliability are an essential demand. Conventionally, gas chromatography with electron capture detector (ECD) is applied during laboratory studies on collected environmental samples providing high sensitivity and accuracy 9 . However, extended sampling and measurement times, and the dimensions and operational requirements of the instruments render them less feasible for in-field scenarios, and certainly do not allow for potential usage as drone-or unmanned aerial vehicle (UAV)-mounted sensing systems. Among the optical sensors, the mid-infrared spectral region (MIR; 2.5-25 μm) has been widely utilized for N 2 O detection 11,16,17 utilizing IR-absorption phenomena 18 . To achieve the desired sensitivity, usually multi-pass gas cell providing extended absorption path lengths are applied. Again, despite many studies demonstrating suitable sensitivity the involved cost of instrumentation and optically sophisticated gas cells along with bulky system dimensions to date limit extended field applications of such MIR analyzer systems. As a viable alternative to complex gas cells, hollow waveguides (HWG) serving as a light pipes made of dielectric materials or metals with a coaxial channel enabling radiation propagation by reflection at the inside wall have emerged 19 . If gaseous samples are injected into the hollow core, the HWG may simultaneously act as a miniaturized gas cell providing several meters of optical path length yet demanding only minute gas sample volumes (i.e., several milliliters). Simultaneously, HWGs serve as an optical waveguide propagating radiation at rather low attenuation losses. Consequently, HWGs facilitate the development of miniaturized MIR gas sensing systems providing extended and well-defined optical path lengths with high optical throughput, while maintaining fast response times (<1 min) due to the small transient gas sample volume. The utility and adaptability of HWGs coupled to FTIR spectrometers operating in the MIR spectral range for various gas sensing applications has extensively been described in literature . Substrate-integrated hollow waveguide (iHWGs) constitute a new generation of hollow waveguides pioneered by the research team of Mizaikoff and collaborators 23 with exceptionally compact substrate dimensions (e.g., made from aluminum), an adaptable (i.e., designable) optical path lengths via the integration of meandered hollow waveguide channels at virtually any desired geometry into an otherwise planar substrate. Furthermore, a minute volume of gas sample is required for analysis (i.e., few hundreds of microliters), and the short transient residence time of these sample volumes within the active transducer region facilitate real-time monitoring rendering iHWGs tailorable to a wide range of application scenarios . Combining the iHWG technology with a compact preconcentrator device, we demonstrate in the present study the first field-deployable, compact MIR N 2 O gas sensor operating at the demanded concentration levels. Nitrous oxide was collected from gaseous samples using molecular sieve 5 A during 10 min at a gas sample flow rate of 200 mL min −1 corresponding to a total sampled volume of 2 L, and was then thermally desorbed at 270 °C for 1 min into a gas volume of 5 milliliters, thus facilitating an enrichment effect by a factor of 400. For quantitative analysis, the pronounced fundamental v 3 MIR band of the N 2 O molecule occurring at 2222-2177 cm −1 range was evaluated enabling a limit of detection of 5 ppbv. ## Results Optimization of the preconcentration parameters. Several parameters were evaluated and optimized in order to maximize the analytical signal intensity during the preconcentration step 28 . N 2 O molecules were trapped using molecular sieve 5 A as sorbent, and were then thermally desorbed into nitrogen with the desorption flow carrying the molecules into the iHWG. Sampling of 5 ppmv of N 2 O at 200 mL min −1 for 10 min was used during the optimization. The desorption flow was evaluated in the range from 3 to 20 mL min −1 (see Figure S1) yielding an optimized gas flow of 5 mL min −1 used during all further experiments. The desorption temperature and time were optimized at 270 °C for 1 min yielding reliable and reproducible N 2 O-MIR signals (Figures S2 and S3). The improvement factor resulting from the introduction of the preconcentration step was calculated as approx. 500 by comparison of the integrated peak area associated with the N 2 O signal at 2222-2177 cm −1 of standard mixtures at 500 ppm directly injected to the detection system, and 2.5 ppm subjected to the preconcentration procedure. Figure S4 shows the spectrum at each condition. Analytical figures-of-merit. For quantification, a calibration function was established based on the evaluation of the peak area with integration boundaries from 2222 to 2177 cm −1 vs. the nitrous oxide concentration. For each concentration, the mean value of three independent replicate measurements was calculated. The resulting goodness of the fit of the linear calibration function was determined at r 2 > 0.99 in the concentration range of 0.1 to 2.5 ppmv of nitrous oxide following A = 0.813 [N 2 O] + 0.0118. The calculated limit of detection (LOD) was considered at three-times the standard deviation of the blank signal, and was determined at 5 ppbv. The obtained analytical figures-of-merit are summarized in Table 1. Evaluation of spiked and real-world samples. In order to evaluate the utility of the proposed methodology for analyzing real-world atmospheric samples, Tedlar ® gas sampling bags were (i) filled with untreated atmospheric air collected outdoors and quantitatively spiked with N 2 O obtaining a final concentration of 0.5 and 2.5 ppmv for calibration purposes, and (ii) filled only with untreated atmospheric air for analyzing the real-world background concentration based on the established calibration. Thereafter, these samples were inserted into the preconcentration system by suction using a pocket pump set at 200 mL min −1 for 10 min, and were then analyzed as previously described for the calibration samples. The obtained concentrations of the spiked samples derived from the calibration function were 0.42 ± 0.09 and 2.39 ± 0.11 ppmv resulting in recovery rates of 84% and 96%. The N 2 O concentration was then determined in untreated atmospheric samples was 306 ± 1 ppbv. This value is in excellent agreement with the anticipated atmospheric background concentration of N 2 O (i.e., on average 330 ppbv). ## Discussion Nitrous oxide has a pronounced absorption band within the mid-infrared spectral range associated with the fundamental ν 3 vibration of the N 2 O molecule located at 2270-2160 cm −1 (Fig. 1) 29 . However, a spectral interference caused by broadening of the CO 2 band has been identified during real-world sample evaluation. This interference Table 1. Summary of analytical figures-of-merit obtained for the iHWG-FTIR nitrous oxide gas sensing system. was reduced by limiting the integration boundaries associated with the N 2 O signal at 2222-2177 cm −1 . Figure 1 exemplarily shows the spectrum of N 2 O, CO 2 , and the real-world sample highlighting the applied signal integration boundaries. The relation between the absorption signal and desorption time is demonstrated using nitrogen as desorption gas at a flow rate of 5 mL min −1 . As evident in Fig. 2, no spectrum is recorded via the iHWG in the beginning of the desorption procedure, which is attributed to the dead volume between the iHWG-FTIR sensor and the preconcentrator. Approximately 150 s after sample desorption, the IR signal characteristic for N 2 O abruptly increased reaching a maximum value after approx. 175 s, which was then continuously decreasing again to zero (i.e., at approx. 200 s). The preconcentrator-iHWG system was then purged with nitrogen for 200 s to ensure complete removal of N 2 O from the sorbent, thereby avoiding any memory effects. A gas phase preconcentration-based analytical method may be also evaluated by its efficiency in trapping molecules passing through the solid sorbent. Fundamentally, the trapping efficiency increases with the sample flow rate up to a maximum value, and then decreases. For the current configuration, the sample flow rate was evaluated in the range of 50-200 mL min −1 . A gas sample flow of 200 mL min −1 for 10 min was determined providing an optimum signal; however, the flow rate may readily be adapted to other measurement scenarios and sampling requirements. In summary, in the present study N 2 O molecules were trapped at room temperature using molecular sieve 5 A, and were then thermally desorbed at 270 °C into the iHWG-FTIR detection system using nitrogen at a flow rate of 5 mL min −1 . For quantitative purposes, integration boundaries from 2222-2177 cm −1 were utilized to avoid spectral interference by CO 2 . Using a sampling flow rate of 200 mL min −1 for 10 min enabled a limit of detection of 5 ppbv. The sensitivity of the method may readily be improved by increasing the sampling flow rate and the analysis time, and by increasing the absorption path length provided by the iHWG. Furthermore, alternative (i.e., more efficient) sorbent materials may further improve the achieved preconcentration effect. Smartly optimizing these parameters, it is immediately evident that the developed iHWG MIR sensor system provides a generic transducer platform that may be flexibly and modularly adapted to a wide range of monitoring scenarios. Besides, replacing the FTIR spectrometer with a suitable tunable quantum cascade laser (tQCL) light source should result in a significant reduction of the physical dimensions of the developed IR sensing device. ## Methods Materials and FTIR spectrometer. Nitrous oxide and nitrogen were acquired from MTI Industriegase AG (Neu-Ulm, Germany). The sample preconcentration tube was custom-made from a quartz tube in U-shape with an inner diameter of 3.5 mm. The effective packaging length containing molecular sieve 5 A (Sigma-Aldrich, St Louis, USA) -a crystalline metal aluminosilicate frequently applied as solid sorbent for gaseous compounds 30 was 40 mm. Thermal desorption was performed using a heating wire wrapped around the tube and connected to a voltage/current controller (Basetech, BT-305, Germany). All measurements were performed using a compact portable FTIR spectrometer (Alpha OEM, Bruker Optics Inc, Ettlingen, Germany) equipped with a liquid nitrogen cooled mercury-cadmium-telluride (MCT) detector (FTIR-22.1.00, Infrared Associates, Stuart/FL, USA). A straight-line iHWG made from polished aluminum providing an integrated hollow waveguide channel (i.e., absorption path length) of 15 cm at device dimensions of 150 × 46 × 17 mm (length x width x depth) was coupled to the spectrometer using gold-coated off-axis parabolic mirrors with a focal length of 1′′ (Thorlabs, Dachau, Germany). IR spectra were recorded in the range of 4000-650 cm −1 at a spectral resolution of 4 cm −1 ; 10 spectra were averaged per measurement. The OPUS 7.2 software package (Bruker Optics Inc, Ettlingen, Germany) was used for data acquisition and evaluation. ## Measurement protocol. The experimental setup is schematically shown in Fig. 3. Polytetrafluoroethylene tubes with Luer-lock ® adapters were used to connect the system components. Three-way gas valves were adjusted for switching different sample pathways during a measurement routine. The calibration protocol was executed in four steps (a-d) as follows: (a) Standard gas mixtures: Nitrous oxide and nitrogen cylinders were used to prepare standard gas mixtures. Suitable gas flow rates of N 2 O and N 2 were mixed and delivered to the sensing system at a flow rate of 200 mL min −1 . The sample preparation was carried out for a period of 3 min to obtain homogenous solutions. Then, valves V 1 and V 4 were set to enable the passage of the sample through the preconcentration tube. (b) Sampling: Samples were introduced into the preconcentration tube containing the sorbent at a gas flow rate of 200 mL min −1 for 10 min (i.e., resulting in sampling 2 L of each sample). For the standard gas mixtures, the flow was regulated using the gas mixing system. For real-world samples, Tedlar ® gas sampling bags containing air samples were attached to the preconcentration tube, and the gas sample was transported through the system by suction using a pocket pump (Model 222-3, SKC, Dorset, United Kingdon) with a calibrated gas flow rate of 200 mL min −1 . (c) Cleaning: After sampling, the valves V 1 and V 4 were set to close the preconcentration pathway. Nitrogen at a flow rate of 200 mL min −1 was used to regenerate the system for 3 min. Thereafter, a background IR spectrum was collected averaging 100 scans. (d) Thermal desorption/Data acquisition: After cleaning, the preconcentration tube was heated (i.e., via the resistive heating wire controlled by a temperature-calibrated voltage/current controller @ 5.3 A and 18.5 V) for 1 min. For the elution procedure, nitrogen was inserted into the preconcentration tube via V 1 and V 4 at a flow rate of 5 mL min −1 , thereby carrying the desorbed N 2 O molecules to the iHWG-FTIR sensing device through valve V 5 .

chemsum

{"title": "A Hyphenated Preconcentrator-Infrared-Hollow-Waveguide Sensor System for N2O Sensing", "journal": "Scientific Reports - Nature"}

the_onset_of_faba_bean_farming_in_the_southern_levant

## Abstract: Even though the faba bean (Vicia faba L.) is among the most ubiquitously cultivated crops, very little is known about its origins. Here, we report discoveries of charred faba beans from three adjacent Neolithic sites in the lower Galilee region, in the southern Levant, that offer new insights into the early history of this species. Biometric measurements, radiocarbon dating and stable carbon isotope analyses of the archaeological remains, supported by experiments on modern material, date the earliest farming of this crop to ~10,200 cal BP. The large quantity of faba beans found in these adjacent sites indicates intensive production of faba beans in the region that can only have been achieved by planting non-dormant seeds. Selection of mutant-non-dormant stock suggests that the domestication of the crop occurred as early as the 11 th millennium cal BP. Plant domestication| Vicia faba L.| Pre-Pottery Neolithic B| radiocarbon dating| Δ 13 C analysis. The shift from foraging to food production, which in many cases marked the transition from a hunter-gatherer to a sedentary lifestyle, brought about substantial changes in the history of humankind and its relation to the ecosystem 1 . The process of plant domestication, based on the selection of phenotypes with characteristics that are more suitable for agriculture, systematically reduced the range of ecotypes available for selection and improvement and led to the loss of 'wild-type' traits that ensure the survival of species in their natural habitat. Unraveling the early history of the faba bean would highlight a turning point in human civilization, by documenting the transition from food collection to food production, and could provide new insights that could help improve the crop for the future. We use stable carbon isotope analysis, radiocarbon dating and archaebotany to deepen our understanding of the nature of the selection process that accompanies domestication of the faba bean. The faba bean is a major crop in many countries, including China, Ethiopia and Egypt, and it is widely grown for human consumption throughout the Mediterranean region and in parts of Latin America 2 . Worldwide it is the third most important feed grain legume after soybean (Glycine max) and pea (Pisum sativum) and the faba bean is the most efficient N-fixing legume used to reduce the emission of the N 2 O greenhouse gas 3 . Despite its importance, little is known about its wild progenitor and the area of its origin. Cubero 4 distinguished between four varieties within the domesticated species of faba bean (major, equina, minor and paucijuga), and all are interfertile. He suggested the Near East as the core area where Vicia faba L. originated, and advocated that V. faba var. major and equine derived from the var. minor. Studies of the genetic profile support Cubero's hypothesis 5 . To date, the most abundant findings of ancient faba bean (V. faba L.) are from the Pre-Pottery Neolithic B (PPNB) site of Yiftah' el, in the Southern Levant, and they were radiocarbon dated to the 10 th millennium cal. BP 6 . While faba bean are plentiful at Yiftah' el, very few faba beans were recovered in other PPN sites in the Levant (Fig. 1). This situation changed when relatively large amounts of faba beans were found in Early PPNB contexts at the site of Ahihud (n = 6205), and the Middle PPNB sites of Nahal Zippori 3 (n = 132) and Yiftah' el (n = 1069) (Fig. 2) (Supplementary information) . These discoveries were the motivation for the present study on the origins of farming the faba bean. The large numbers of specimens found and the geographical proximity of the 3 sites of Ahihud, Nahal Zippori 3 and Yiftah' el (Fig. 1b), that therefore share the same environmental conditions, offered a unique opportunity to study the beginning of cultivation of this legume and its process of domestication in the southern Levant. To date the study of domestication of legumes has been based on visible changes of the plant, namely an increase of seed size and a reduction of the natural dispersal mechanism of the seeds (i.e. dehiscent pods). Since pods are rarely found in archaeological contexts, seed size has been considered the best trait to identify domesticated legumes 10 . Here we test the assumption that size is the main change that occurs at the early stage of domestication by measuring biometrical parameters and biochemical properties of the archaeological seeds. Experiments were carried out on modern faba bean to estimate the variability of size due to charring. We used standard size analysis to measure the biometric traits of charred faba beans from the five contexts in Ahihud, Nahal Zippori 3 and Yiftah' el. High-precision radiocarbon dating was used to assess the absolute chronology of seeds from each context, while stable carbon isotopes (Δ 13 C) were measured to infer information on the water input of the ancient seeds during their life cycle and test the correlation between water status and size. ## Results Charring experiments on modern material. The first question we address is the effect of charring on faba beans. The experiment was carried out on 120 modern seeds charred at various temperatures and 46 , Abu Hureyra 47 , Dja' de 48 Tell el Kerkh 49 , Iraq ed-Dubb 50 , Nahal Zehora 51 , Horvat Galil 52 , Yiftah' el 35 (and this work), Nahal Zippori 3 53 for different times (Supplementary Table S1). We discovered that modern seeds explode above 200 °C, and therefore the ancient seeds must have been charred at ~200 °C or below. At 250 °C 33% of seeds explode, and at 300 °C all of them explode. The length decreased by ~17%, for both short and long periods of charring (R 2 : 0,93 for 4 h and R 2 : 0,93 for 12 h) (Fig. 3a). Breadth and thickness were also measured but they were found subjected to substantial changes when burned longer than 4 h so they were considered less relevant than length for biometric analysis. Thus charring results in a homogeneous reduction of length, while breadth and thickness change in an unpredictable way. The charring experiment also showed that the stable carbon isotopic composition (Δ 13 C) is retained in seeds after charring at 200 °C × 4 h or at 200 °C × 12 h (R 2 : 0,77; R 2 : 0,84;) (Fig. 3b) (Supplementary Table S2). We thus conclude that length and stable carbon isotopic composition can provide reliable information on the archaeological samples. S3) The samples from Ahihud (Early PPNB) were significantly longer (~20%) than those from Nahal Zippori 3 and Yiftah' el (Middle PPNB) (p < 0,001) (Supplementary Table S4). The relative chronology, based on the associated lithics found in the sites, attributes Ahihud to the Early PPNB, while the other two sites are from the Middle PPNB 7,9 . Radiocarbon dates of faba beans from the three sites are consistent with this relative chronology. A simple Bayesian sequence of 14 C dates shows that Ahihud dates to 10,235-10,125 cal BP and falls at the end of the Early PPNB (Fig. 4). These dates are older than the ones from Nahal Zippori 3 and Yiftah' el, that both dated to 10,160-9,890 cal BP, which fall in the beginning of the Middle PPNB 8 (Supplementary Table S5). When the sizes of these samples are plotted on an absolute chronological scale, it is clear that the longer faba beans from Ahihud are significantly older than the shorter ones from the other sites (R 2 : 0,98) (Fig. 5a,b). ## Size analysis of archaeological material and radiocarbon dating. Δ 13 C analyses of archaeological faba beans. In order to understand the influence of edaphic conditions on the bean size, we analyzed the stable carbon isotope ratio (Δ 13 C) of 95 archaeological faba beans. The Δ 13 C was previously shown to be a reliable parameter to ascertain the amount of water received by the faba bean during its growth 14 . The samples from Ahihud (Early PPNB) had higher values of Δ 13 C when compared with those of Nahal Zippori 3 and Yiftah' el (Middle PPNB) (p < 0,001). A positive correlation was found between the average length and the average Δ 13 C of each group (R 2 : 0,72) and this shows that the size mainly depends on water availability at the site of growth (Supplementary Table S6) (Fig. 6a,b). This result is in agreement with the hypothesis that the water received during the period of growth accounts for the size of the beans. ## Discussion Domestication of legumes. The distinction between cultivation and domestication of the most common edible plants is controversial. Some favor the opinion that a regime of tilling, sowing and reaping (cultivation) acts as a selective force on wild plants, selecting for mutations adapted to the new environmental conditions 15,16 . Others claim that mutations that are favorable for agriculture (domestication) must have been selected before the crop could be successfully cultivated 17,18 . The mutations typically include reduced seed dormancy, the loss of dispersal mechanisms (indehiscent pods) 18 , reduced seed coat thickness 19 , and increased seed size 10 . The first two mutations do not leave visible traces on the legumes, but wherever large stocks of legumes are found, plants with domestic-traits (non dormant/non dehiscent) must have been used 16 . Furthermore, experiments conducted on wild modern pea, chickpea and lentil prove that neither harvesting of wild stands or cultivation of wild legumes results in profitable yields 17,18,20 . The thickness of seed coat should be thinner and smoother in domesticated stocks to facilitate water penetration and germination 19 . The validity of this mutation as a trait of domestication remains controversial, because no great differences were found between the domesticated species of lentil, bitter vetch, grass pea and their wild relatives 21,22 . Increase in seed size is considered to be one of the major domestication-traits in crops, but analyses conducted on archaeological legumes such as pea, lentil and cowpea show that the increase in seed size does not occur at the early stage of domestication but rather later as a result of crop improvements 23,24 . Cultivation versus domestication of the faba bean. Scholars that invoke a protracted process of cultivation as an unconscious cause of domestication stress three mechanisms that should increase the seed yield: the seedling depth 15 , ploughing 25 and selection of larger seeds for seeding 26 . It holds that seeds buried deeper by human planting develop larger reserves because they emerge from a greater depth 27 . In the case of the faba bean, greater burial depth does not increase seed size, while maximum yield is obtained when the seeds are buried no deeper than 8 cm. Faba bean, as well as chickpea and lentil, has an hypogeal germination, meaning that its cotyledon remains where the seed is sown while only the shoot emerges from the soil surface; as a consequence, the seed must be buried close to the surface to sprout above the ground . A negative correlation was also found between tillage and yields 31 . Agronomists agree that tillage does not help to maximize the faba bean yield once the crop is farmed under dry climatic conditions. The relatively shallow root system of faba bean relies on the water accumulated within the first 30-40 cm from the soil surface; therefore soils that have a more stable structure prevent water from percolating to greater soil depths. Another advantage of compacted, non-tilled soils is that residues of crops remain on the surface and this prevent excessive water evaporation, which is a major constraint for plants growing in dry and semi-dry environments 32 . The selection of larger seeds for seedling is commonly considered a profitable way to obtain plants that have bigger seeds and produce higher yields 33 . Thus a relation between seed yield and seed size anticipates that seed size should have a positive influence on seed yield. Nonetheless such a correlation has not been recorded for the faba bean. On average, medium and small size grains of faba bean produce maximum seed yield compared to larger seeds 29,34 . The explanation for such a contradictory result is that plants originating from smaller seeds produce a greater number of pods, which are, on average, longer that those produced by plants originating from large seeds. Plants originating from small faba beans sprout, flower, form pods and mature faster, leading to the highest harvest index values. Small-sized grains produce good yield over a range of seasonal conditions, while large seeds are more sensitive to adverse seasonal conditions such as drought and low temperature 29 . The agronomic studies show that common practices associated to cultivation do not lead to higher yields. Therefore it is unlikely that cultivation acted as a selective force in the process of domestication of the crop. Other mechanisms must have been adopted in order to transform the faba bean into a crop. Seed dormancy (and pods dehiscence) is a typical trait of the wild progenitors of the major legumes (i.e. Lens culinaris ssp. orientalis, Pisum humile, Cicer reticulatum, Vigna radiata subsp. sublobata and Phaseoulus vulgaris), so it is safe to assume that unknown progenitor of faba bean would also have some mechanism to delay the germination. The problem of dormancy in wild legumes can be overcome by soaking the seeds in water (or abrade the seed coat, increase the temperature etc) to encourage the germination of the legumes; thus, repeated use of free-geminating seeds would saturate the natural gene bank with non-dormant seeds in a relatively short time 15 . Another option includes the inadvertent selection of domestication traits by means of cultivation of wild stands of plants; this could also have resulted in the widespread introduction of non-dormant seeds in the natural gene pool, but only after 5-6 cycles of (unprofitable) cultivation 18 . Regardless of whether the selection of the original seeds had been 'conscious' , through the selection of non dormant seeds straight from the wild stands, or 'unconscious' , by inducing phenotopic changes in the local gene pool, the loss of seed-dormancy is the only circumstances that enables the production of legumes in large quantities. The presence of a large quantity of seeds in Ahihud and Yiftah' el reinforce the idea that non-dormant (and non-dehiscent) stocks were used to ensure reliable harvesting, resulting in the build-up of characteristic domestication traits. The size of the archaeological faba bean cannot be used as a domestication trait, because the Δ 13 C shows that size depends on the amount of water received by the plant. We therefore rely on other parameters to assess domestication. The presence of storage facilities found in Ahihud and Yiftah' el , as well as the presence of other crops in great quantities, such as the 7,2 kg of lentils found in Yiftah' el 35 , are consistent with the notion that the surplus was kept for seeding, to ensure the continuity of legume production and the sustainability of the settlers. It is worth noting that by the Early PPNB, cereals were already domesticated 36 . Within this context, the domestication of the faba bean represents an important additional step that settlers took to reduce the risk of famine. ## Conclusions The present study provides new insights about the geographic area where the faba bean originated. The large stocks of seeds found in the coeval adjacent sites of Ahihud, Nahal Zippori 3 and Yiftah' el show that the faba bean was already domesticated in the 11 th millennium cal BP in the lower Galilee, Israel. Agronomic studies of the modern faba bean, show that cultivation alone does not improve the yield. Therefore only the selection of domestication traits, such as free-germination, could have maximized the agricultural output. The analysis of the Δ 13 C of the faba beans, which allows to identify the amount of water received by the plants during the period of growing, shows that the seeds from Ahihud were farmed under moister conditions compared to the seeds found in Yiftah' el. In the light of this evidence, we conclude that the intensive farming of the faba bean began in the 11 th millennium in the lower Galilee in a period of greater water availability and continued into the 10 th millennium, due to the ability of the local farmers to select seeds able to germinate under dryer conditions. ## Methods Charring experiments on modern material. Several studies have sought to replicate the condition under which archaeobotanical remains have been charred. Scholars found out that well preserved and undistorted archaeobotanical grain are more likely to have been heated at lower temperature (between 150 °C and 300 °C) in reducing conditions . As results, we decided to conduct our experimental charring in reducing conditions at variable temperatures for different periods of time. For the experiment, 120 small faba beans (Vicia faba var. minor) were used because of the similarity in size and shape of this variety to the archaeological specimens. Length, breath and thickness of the seeds were measured using a Leica Image System Analysis (LAS V3.8) attached to a binocular microscope (Leica M80). Five groups, of 24 seeds each, were charred at 200 °C × 4 h, 200 °C × 12 h, 250 °C × 2 h, 250 °C × 4 h, 300 °C × 0.5 h in anoxic atmosphere in a muffle oven. Each seed was wrapped in aluminum foil. Measurements of length, breath and thickness were taken after the charring only on those seeds that were still whole (Supplementary Table S1). Additional 64 seeds were used to check the variation of Δ 13 C as result of charring. Previous studies had already showed that Δ 13 C of seeds did not change as effect of charring 12,41 , but none of these studies have ever tested the effect on faba beans. We therefore decided to test the variation of Δ 13 C on 32 seeds kept in reducing condition in a muffle oven at a temperature of 200 °C × 4 h, and on another 32 that were burned in the same condition but for a longer period of time (12 h) (Supplementary Table S2). Size analysis of archaeological material. 469 seeds were selected among the remains collected in the archaeological contexts: the two storage pits found in Ahihud (L450_E14; L398_D13), the plaster floor in Nahal Zippori 3 (L273_C11), and the storage contexts from Yiftah' el (L5073_G18; L715_F40). The length, breadth and thickness were measured using a Leica Image System Analysis (LAS V3.8) attached to a binocular microscope (Leica M80) (Supplementary Table 3). Since experiment on modern material showed that breadth and thickness vary in unpredictable ways as result of charring, length was the only parameter used to compare the size of seeds coming from the different contexts. A Z-Test at two row was used to prove that the average length of the populations were statistically different. P value < 0,001 was considered the threshold to accept this hypothesis (Supplementary Table S4). Radiocarbon dating of archaeological material. The samples were pre-treated, graphitized and measured by Accelerator Mass Spectrometry at D-REAMS Radiocarbon Laboratory of the Weizmann Institute of Science, Israel. The legumes (~30 mg of material) were cleaned using Acid-Base-Acid treatment as in Yitzhaq et al. 42 . The samples prepared for dating were combusted to CO 2 in quartz tubes containing about 200 mg of copper oxide (Merck) and heated to 900 °C for 200 min. The CO 2 was divided into 3 aliquots and each was reduced to graphite using cobalt (Fluka) (about 1 mg) as a catalyst and hydrogen at 700 °C for 20 hr. The 14 C ages were calibrated to calendar years BP using the IntCal13 atmospheric curve 13 using the software OxCal v 4.2.3 54 . Bayesian modeling was used to build the chronological sequence between the three sites according to the material culture remains. The legumes found in Ahihud are considered to be within the EPPNB and therefore are the earliest, while the material from Nahal Zippori 3 and Yiftha' el, both MPPNB, are younger. The dates of Ahihud start the model. Archaeologically, the storage places from Ahihud were not synchronous, therefore the 14 C dates RTK 6866, 6875 and 6868 are set as a sequence within themselves. Finally, dates from Nahal Zippori 3 and Yiftha' el (RTK 6864,6865, 6892 and 6991) were added to the model in a phase, since there were no chronological differences among these two MPPNB sites (Supplementary Table S5). Stable carbon isotopes analysis of archaeological material. 95 archaeological seeds were selected among the specimens found in the three sites. The seeds were purified from all contaminants using the same treatment reported for 14 C samples. Analysis of the organic carbon by the dry combustion method was performed using an Elemental Analyzer (1112 Flash EA, Thermo-Finnigan) interfaced with an Isotope Ratio Mass Spectrometer (EA-IRMS Delta V Plus, Thermo Scientific), under the following conditions: the oxidation reactor was filled with chromic oxide over silvered cobaltous-cobaltic oxide and was maintained at 1020 °C. The reduction reactor filled with reduced copper wire was maintained at 650 °C. The chromatographic column for the gas separation was maintained at 50 °C. Helium carrier gas flow was 100 ml/min. To account for changes in δ 13 C of atmospheric CO 2 (δ 13 C air ) during the two hundred year span between the sites, the Δ 13 C of the plant was calculated using the δ 13 C air and the carbon isotope ratio of the plant (δ 13 C plant ) as described by Farquhar et al. 43 Where the δ 13 C air was inferred by interpolating a range of data from ice-core records covering the whole Holocene (http://ftp.cmdl.noaa.gov/ccg/; http://web.udl.es/usuaris/x3845331/AIRCO2_LOESS. xls). The Δ 13 C was calculated for all the archaeological faba beans, using these updated estimates of past δ 13 C air ## 43 . The average value of Δ 13 C of each group as average value of 19 measurements. The higher is the value of the discriminant Δ 13 C the higher is the water input received by the faba bean during the period of growth. (Supplementary Table S6). A Z-Test at two row was used to prove that the average length of the populations were statistically different. P value < 0,001 was considered the threshold to accept this hypothesis (Supplementary Table S7).

chemsum

{"title": "The onset of faba bean farming in the Southern Levant", "journal": "Scientific Reports - Nature"}

on_the_formation_of_nanobubbles_in_vycor_porous_glass_during_the_desorption_of_halogenated_hydrocarb

## Abstract: Vycor porous glass has long served as a model mesoporous material. During the physical adsorption of halogenated hydrocarbon vapours, such as dibromomethane, the adsorption isotherm exhibits an hysteresis loop; a gradual ascent is observed at higher pressures during adsorption, and a sharp drop is observed at lower pressures during desorption. For fully wetting fluids, an early hypothesis attributed the hysteresis to mechanistic differences between capillary condensation (adsorption) and evaporation (desorption) processes occurring in the wide bodies and narrow necks, respectively, of 'ink-bottle' pores. This was later recognized as oversimplified when the role of network percolation was included. For the first time, we present in-situ small angle x-ray scattering measurements on the hysteresis effect which indicate nanobubble formation during desorption, and support an extended picture of network percolation. The desorption pattern can indeed result from network percolation; but this can sometimes be initiated by a local cavitation process without pore blocking, which is preceded by the temporary, heterogeneous formation of nanobubbles involving a change in wetting states. The capacity of the system to sustain such metastable states is governed by the steepness of the desorption boundary.For more than half a century, Vycor porous glass 1 has been used as a model mesoporous material; according to Brunauer's classification 2 , it exhibits a type IV adsorption isotherm with an H2 hysteresis loop (see Fig. 1, inset). The early 'knee' at low relative pressure is taken to indicate the formation of an adsorbed monolayer of adsorbate molecules. The first models considered that there were two types of pores present, each with a size distribution. The first type were V-shaped, and these filled and emptied reversibly. The second type, known as 'ink-bottle pores' , had a narrow neck and a relatively wide interior. According to the Kelvin equation 3 , the vapour pressure above the concave meniscus of a wetting liquid decreases with curvature, which is inversely proportional to the meniscus and pore radius. Thus, as relative vapour pressure was increased to one, the gradually increasing steepness of the adsorption branch was taken to reflect the combined effects of monolayer adsorption and gradual capillary condensation in the wide pore interiors with a large distribution of sizes. But, as relative vapour pressure subsequently decreased during the reverse process, the delayed but sharp drop in the desorption branch of the adsorption isotherm was taken to indicate the evaporation from the wide pore bodies via the narrow necks, the latter with a relatively narrow size distribution. Later on, it was recognised that such a description was oversimplified, and the role of network effects was taken into account. This phenomenological paradigm was in accord with IUPAC recommendations 4 , and the Vycor porous glass has provided a classic example for its experimental demonstration. However, by revisiting our small angle x-ray 5 and small angle neutron 6 scattering data (SAXS and SANS), we have concluded that the strong increase in scattered intensity at the commencement of the desorption process may be attributed to the temporary formation of myriad nanobubbles inside the porous glass (for more details of this process, see Fig. 1 and the description in the next section). The extent to which these nanobubbles influence the desorption process and the manner in which they form are both discussed in this report. It was Ross and co-workers 6 (of whom KLS is also an author of this report) who first observed this upturn. They interpreted the increase in the scattered intensity as the result of a spaghetti-like percolation cluster induced by mass fractals. However, percolation by its own can not fully explain this upturn, especially at the beginning of the hysteresis area; a driving mechanism is also needed. Moreover, the concept of nanobubbles had not been established or verified at that time. It is only in recent years that an intense research effort has been devoted to the study of nanobubbles, and especially to their formation and stability, since they appear to last for days or even months. This paradoxical behaviour contradicts the classical view of, for example, the air-water interface, for which the high Laplace pressure inside small bubbles should cause them to dissolve instantly in favour of larger ones (the phenomenon of Ostwald ripening). While detailed research on pore networks giving rise to type H2 hysteresis loops continues today with novel mesoporous materials , the effect of various factors on adsorption hysteresis remains an open question. There are two main mechanisms of desorption in such networks: a) when evaporation of the capillary condensate from the pore body occurs after emptying of its neck, the mechanism is known as pore blocking and b) when the pore body empties first, while the pore neck remains filled, the mechanism is known as cavitation. In the first mechanism, the onset of evaporation is associated with a percolation threshold where a continuous path of open pores to the external surface is formed. In the second mechanism, the growth of gas bubbles in the condensed fluid is involved. Naturally, the size of the pore necks is taken to be the factor that determines which mechanism will prevail. When the neck size is small, but not small enough such that the negative capillary pressure will expand the condensed liquid beyond its limiting tensile strength, desorption will obey the pore blocking/percolation mechanism. On the other hand, when the neck size is small enough, the negative capillary pressure exceeds the limiting tensile strength of the liquid and cavitation will succeed. In addition, other factors such as the surface rugosity of the pore walls may also play a role in the evaporation mechanism, leading to alternative scenarios as extensions of these two cases. For instance, although the formation of nanobubbles of sub-critical size does not lead to a cavitation instability, such nanobubbles may nevertheless assist in the percolation transition. Rosinberg et al. 22 provided a comprehensive theoretical description of hysteresis during the capillary condensation of gases in mesoporous disordered materials. They suggested that a percolation-dominated draining process does not require the introduction from the outset of a pore-blocking mechanism that limits the accessibility of the filled pores to the outer surface of the material. Woo et al. 23 have also studied the desorption mechanism of fluids by Monte Carlo simulation on a matrix configuration representative of Vycor porous glass. They concluded that cavitation via nucleation of bubbles inside the pores plays a role in the desorption process. They further suggested the existence of a percolation transition which required neither a pore-blocking mechanism nor cavitation. By using SANS, Hoinkis and Kuhn 24 examined in situ the sorption of nitrogen at 78 °K on a mesoporous silica glass having a rough internal surface. During desorption, a strong signal at low values of the scattering vector Q was also observed. They interpreted this result in terms of ramified vapour-filled void clusters, and they further speculated that these clusters may originate from a percolation process; this process could occur with or without heterogeneous nucleation or cavitation and the self-similar growth of bubbles. Bonnet et al. 25 studied the collective effects which occur during adsorption-desorption in Vycor porous glass by light scattering. They concluded that, as temperature increases, a crossover from percolation to cavitation is evident for the evaporation process. In a review article, Monson 26 discussed the hysteresis for fluids in mesoporous materials. For disordered pore networks like those in Vycor glass, evaporation from different regions depends upon their spatial location. He suggested that pore blocking and cavitation are key components of the desorption mechanism. Further relevant work includes a noteworthy review by Landers et al. 27 on the characterization of porous materials and one by Thommes and Cychosz on the same topic 28 . ## Results Figure 1 shows the SAXS measurements from which the formation of nanobubbles is inferred. The spectrum of dry/empty Vycor (i.e. at p/p o = 0) is characterized by the peak at Q = 0.025 −1 (curve 0); here, Q = 4π sinθ /λ and 2θ is the scattering angle. On the other hand, during the desorption process, the spectrum of Vycor at (p/p o ) des = 0.54 (curve 1), which is at the onset of the steep part of the desorption branch (see inset), is characterized at low Q by an increase in the scattered intensity to well above the spinodal peak (compare with curve-0). During adsorption, an adsorptive film is deposited on the pore walls, and eventually all pores are filled with capillary condensate. Since dibromomethane (CH 2 Br 2 ) contrast matches the silica matrix, the scattered intensity constantly decreases as p/p o increases. During desorption, however, curve 1 shows that the scattered intensity at low Q, just before the pores empty, increases sharply. This is true, in spite of the fact that, at this relative pressure, the isotherm in the inset indicates that only 3% of the adsorbate has evaporated. The preceding situation is similar to that of the so-called 'opacity point' . In much earlier work, a silica gel-water system was found to assume a turbid appearance at a point close to the beginning of the steep part of the desorption isotherm; this was termed by Zsigmondy 29 as the opacity point. A similar situation was also observed for Vycor porous glass 30 . The glass, which is transparent when saturated, acquires an intense whitish turbidity when a small amount of liquid is removed by evaporation. Haynes and McCaffery 31 have examined the turbidity in Vycor porous glass with light scattering. They attributed the phenomenon to a non-uniform distribution of full and empty pores large enough to act as Mie scatterers, sustained by hysteresis effects. Based on molecular dynamics calculations of the condensation process within Vycor porous glass, a density redistribution of the adsorbate within the hysteresis region was also concluded by Valiullin et al. 32 As already mentioned, the observed sharp increase in scattering intensity has been confirmed by previous results we have obtained for SAXS and SANS. Furthermore, this outcome is far from universal. Results for other adsorbing and desorbing systems show no such peak; see Figures S1 and S2 in the supplementary information. Such systems are unable to generate the necessary tensile strength in the liquid adsorbate required to create nanobubbles, even by heterogeneous nucleation. As a consequence for these adsorbents, the large sudden jump in scattering intensity upon desorption (as seen in Figure 1) is absent, nor is there evidence of bulk cavitation at the lower knee. To study the metastabilities fixed in Vycor by hysteresis effects, we conducted a scanning of the desorption isotherm to some depth within the hysteresis area, in conjunction with SAXS. Figure 2 shows the results. From point A on the desorption boundary, an adsorption/desorption scanning cycle is performed (ABCA), and from point A΄ an adsorption-only scanning is performed (A΄C΄…). Points A and A΄, which are at different relative pressures, differ between each other by an adsorbed amount which is roughly equal to that between points C and C΄, the latter points being at equal relative pressure. An equivalent situation to this may now be described, as follows. A fluctuation from an initial state A can lead to transient states, e.g. A΄ and B, in adjacent pore regions. After equilibrium is re-established, B moves down to C and A΄ enters the hysteresis loop to C΄, where the two states are in equilibrium at the same p/p o . This will result in a redistribution of the capillary condensate within the system, which is clearly illustrated at the lower inset of Fig. 2. In both coloured areas, the sum of negative and positive Δ I(Q) is equal to zero. It is the steepness of the desorption boundary which defines the capacity of the system to maintain such metastable distributions; in a sense, the steeper the boundary curve, the larger the saturation differences that can be sustained. In order to gain a better understanding of our results, we draw a picture of a single pore in Vycor. However, it should be noted that this is only a restrictive case for illustrative purposes; the real pore system in Vycor porous glass is far more complicated, and in some cases the descriptions provided have proven controversial. Based on a simulated 2-D TEM image reported by Kim and Glinka 33 with the aid of small-angle scattering data, we have drawn Fig. 3 to summarize a number of average-size estimates of various pore features of Vycor porous glass. The pore walls are sinusoidal in profile and define a pore body formed from two cavities with necks at each end. In the middle, where the sinusoidal walls converge, the pore body is narrowed but not as much as in the necks. The length of the pore body is roughly equal to the Bragg spacing, d. This latter is related to the SAXS scattering vector Q corresponding to the characteristic peak of the Vycor porous glass spectrum, via the expression d = 2π /Q; it ranges between 250 and 285 . The average pore size is about 70 . Furthermore, the pore walls of Vycor porous glass are rough, with a fractal dimension of about 2.3. This roughness has an upper cut-off limit of about 15 . CH 2 Br 2 and Nitrogen BET areas are found to be 80 and 135 m 2 /g, respectively. Other details for the glass may be found elsewhere 34 . Adsorption isotherms for CH 2 Br 2 and Nitrogen are presented in the supplementary material. ## Discussion In a previous study 5 , it was found that the roughness of the internal surface of the Vycor porous glass does not rely entirely on its micro-porosity. During the leaching process, a hydrogel layer is deposited on the pore walls. Following drying, this soft hydrogel is converted to an asymmetric xerogel layer which includes cavities, bridges, and bumps conferring a roughness to the surface in a similar manner to that of e.g. a woven textile fibre. Under these circumstances, the adsorbate molecules experience the pore surface as though it consisted of a porous textile of fibres, and so interact with the adsorbent surface according to a Cassie-Baxter 35 type wetting process (Fig. 4a). However, during desorption, the negative capillary pressure associated with the smaller pores that control the entrance to the pore body result in the exertion of a tensile force on the condensed liquid. Under this force, the adsorbate molecules may find room in the underlying xerogel; nanobubbles are formed and accommodated by the space thus freed. The adsorbate now interacts with the Vycor surface according to a Wenzel 36 type wetting process (Fig. 4b). An ideal chemical model for this wetting transition is presented in Fig. 5. During the synthesis of amorphous Vycor porous glass, important phase separation and phase equilibria effects can take place 37 . A vertical and horizontal Si polymerisation with the chemical post-synthesis treatment will result in a typical pore system for Vycor porous glass, where geometrically different broken siloxane chains and silanol groups are expected at the pore surface, thus explaining its roughness (Fig. 5a). When halogenated hydrocarbons (such as CH 2 Br 2 ) are adsorbed at a moderate temperature, e.g. 293 K, on the Vycor surface, the surface siloxane chains will bend towards the surface because of the presence of significant repulsive forces between, on the one hand, the siloxane bonds and the silanol groups (the latter having a basic nature) and, on the other, the electronegative character of the bromine group (Fig. 5b,c). This leads to adsorbed molecules of CH 2 Br 2 lying on top of bended siloxane chains (the Cassie-Baxter model). During desorption, with a rearrangement of the surface siloxane chains due to rotational, vibrational and migrational effects, nanobubbles can be formed (the Wenzel model). A similar molecular rearrangement was observed during the adsorption and desorption of alkylamines on clays, which depended on the length of the alkyl chain 38 . On the other hand, such behaviour, which is typical for Vycor porous glass, is very unlikely 39 to arise in the semi-crystalline mesoporous materials studied by e.g. Voort et al. 40 and Ravikovitch et al. 41 , where their model is based entirely on the physical mechanisms of adsorption and desorption for nitrogen, argon and krypton at 77 °K and 87 °K. Furthermore, the adsorption/desorption temperature, together with the surface roughness, the concentration and flexibility of the siloxane chains and the chemical properties of the adsorbate molecules are all very important factors in the behaviour of the desorption process in Vycor porous glass, as described in the proposed alternative model. We now analyse the energy barrier to nanobubble formation, as follows. The negative pressure or tension (τ ) of the capillary condensed liquid at a relative vapour pressure of p/p o is given by 42,43 : where V L is the molar volume, and T is the isotherm temperature. In the present case, (p/p o ) des = 0.54 and thus τ = − 21.6 MPa. Since p o for CH 2 Br 2 at 293 K is equal to 4.65 kPa, the correction of Eq.1 for the vapour pressure is negligible. When a nanobubble is formed by homogeneous nucleation, the total energy Δ E hom is the sum of the surface free energy required to form a nanobubble of radius R nb and the work of nanobubble formation (equivalent to the lowering of free energy): where σ is the surface tension (for CH 2 Br 2 σ = 40.2 mN/m). This energy reaches a maximum value (Δ Ε hom ) max = 16π σ 3 /3τ 2 at the Kelvin radius r k = 2σ /|τ | at a given p/p o ; above this radius, the bubble formation leads to a lowering of free energy and is thus spontaneous. At p/p o = 0.54, from (1), r k = 37 and then (Δ Ε hom ) max = 2.34 × 10 −18 J or 579 k B T, where k B is the Boltzmann constant and T is the isotherm temperature. According to nucleation theory the rate of bubble formation is proportional to exp(− Δ Ε max /k B T). We follow the procedure outlined by Grosman and Ortega 43 to estimate Δ E max from adsorption isotherm data. Based on the amount adsorbed as measured from the isotherm, and the geometry of a single pore (see Fig 3), we find the energy barrier for nucleation to be Δ E max = 1.56 × 10 −19 J = 39 k B T. According to the expression (Δ Ε hom ) max = 16π σ 3 /3τ 2 , this energy value corresponds to a negative pressure of |− 84| MPa, which is around four times higher than the actual value of |τ |. Therefore, at p/p o = 0.54, bulk cavitation (via homogeneous nucleation) is unlikely. However, bubbles can also be produced by heterogeneous nucleation; that is to say, they are formed on the pore walls and particularly rough ones with reduced bubble surface free energy, rather than in the bulk fluid. In this case, the negative pressure of − 21.6 MPa which arises at p/p o = 0.54, and where the steep upturn in the scattering intensity at low Q is observed, may qualify for such a local cavitation event. To this end, let us consider the energy required to form a fraction of an interfacial nanobubble (INB) on the Vycor surface, and relate it to the energy involved in passing one mole of condensed liquid from Cassie-Baxter to Wenzel wetting states. The maximum value of (Δ E het ) max corresponding to the heterogeneous nucleation of an INB of critical size R c , is given by 44 and ω is the contact angle (taken on the same side as the liquid). For (Δ Ε het ) max = Δ Ε max and ω = 133 ο , an INB of lateral size α /2 = 27 and R c = r k = 37 is concluded. Fig. 6 shows the results for both types of nucleation. The free energy barrier (Δ G cw ) in moving from a Cassie-Baxter to a Wenzel wetting state is highly dependent on the height of the surface pillars and the liquid contact angle 45 . For the adsorption of CH 2 Br 2 on Vycor porous glass, an autophobic behaviour requiring the use of a finite angle of contact was previously suggested 46 . For pillar heights less than a critical value (about 13.5 ), the Wenzel wetting state prevails. For pillars higher than this critical height, the Cassie-Baxter state is metastable. Coexistence of Wenzel and Cassie-Baxter states is thus possible, depending on the local characteristics of the pore wall roughness. In the case of Vycor porous glass, where the characteristic height of roughness features is about 15 , is Δ G cw not more than 1 k B T. It is noted that the strength of a hydrogen bond with halogenated groups is about 160 J/mol. In any event, Δ G cw is much less than the energy of heterogeneous formation of an INB. The effect of pore geometry on the conditions for cavitation has also been discussed by Ravikovitch and Neimark 47 . They have concluded that the cavitation pressure in spherical pores is higher than that for cylindrical pores. That is, the lower closure point of the hysteresis loop depends not only on the adsorbate and the temperature but also on the pore geometry. Although this is an important conclusion, it is interesting to note that this is not a property of the solid. It is the liquid which has the intrinsic property of taking on the shape of the vessel that contains it; there is no physicochemical interaction, in this reasoning, between bulk liquid and solid. To show this subtle difference, let us provide the following example. Everett 48 introduced a descriptor for the geometry of the pores in a solid using a numerical factor γ , such that: where V p is the pore volume, A is the solid surface area and r p is the mean pore size. This would mean that for non-intersecting cylindrical capillaries of uniform size which are open at both ends γ = 1, whereas for closed spherical pores γ = 2/3; i.e. for equal pore volumes, A cyl < A sph . However, for the liquid column which is accommodated within this cylindrical pore (assuming flat menisci), the surface area will always be greater than that for the spherical blob, as is expected; i.e. for equal volumes the former will always be less 'bulky' than the latter. Since homogeneous nucleation takes place within the volume of the bulk liquid, it may be concluded that (p cav /p o ) cyl < (p cav /p o ) sph . During heterogeneous nucleation, the opposite situation arises. In this case, there is an interaction between the solid/liquid interface, which is readily inferred from the energy required to form an INB: where A lg and A gs are respectively the areas of the liquid/gas and gas/solid interfaces, A eff is the effective area defined by dA eff = ∫C lg δ V lg and C lg is the curvature of the liquid/gas interface. Note that the term in parenthesis is the Gauss equation 49 , and by assuming that the contact angle is independent of the volume, Eq.4 can be transformed to: Now, the CH 2 Br 2 /Vycor system has a surface-to-volume ratio of about 4 × 10 8 m −1 whereas the N 2 / Vycor system has a value of 5 × 10 8 m −1 . The liquid-like adsorbed film on the pore walls, which protects the interior of the capillaries from surface contaminants and irregularities that otherwise may serve as nucleation sites 19 , is rather shallow and at a much higher temperature in the case of CH 2 Br 2 compared to that of N 2 . The surface roughness will increase this autophobicity, and thus will increase the probability for a local cavitation event at higher p/p o . However, in heterogeneous nucleation a local cavitation event by its own is not critical if it cannot propagate within the pore network. When pores are unconnected or loosely connected, heterogeneous nucleation in a small fraction of the pores may not have an effect on the macroscopic properties of the medium; the event will be confined by the pore boundaries. However, in Vycor, there are about 3 × 10 17 pores/g which are fully interconnected. Therefore, a local cavitation event in one of the pores may spread to some extent to neighboring pores, thus making a noticeable difference in e.g. the scattering properties of the medium. At very low values of Q, the scattering is generally determined by large entities (ones with a length scale larger than 1,000 ). We explain the large rise in scattered intensity at low Q as follows. At early stages of the desorption process, nanobubbles with sizes of the order of about 50-60 result in an heterogeneous cavitation event which occurs locally, rather than globally, within the porous network. This localized cavitation event spreads towards adjacent network portions and, from there, develops into a vapour cluster by coalescence; this is large enough to give the strong upturn in the scattering spectrum. Percolation without the need for a pore blocking mechanism may thus develop. However, although initially the pore blocking mechanism is not actively involved in this process, it still plays an important role in the desorption process. This is to govern the spatial extent of cavitation events by defining the steepness of the desorption boundary and consequently the capacity of the system to lock them into metastable equilibria. Furthermore, hysteretic behaviour may arise as a consequence of surface interactions and can be explained without additional assumptions about the pore structure or on the detailed shapes of the liquid menisci 50 . Figure 7 illustrates a schematic for the progressive desorption process within the porous system. Further work with adsorbents of similar surface nature but different pore size is underway. ## Methods In this study, we present in-situ measurements on the adsorption of dibromomethane (CH 2 Br 2 ) onto Vycor porous glass using small angle x-ray scattering. Dibromomethane is able to 'contrast match' with amorphous silica; in this way, when a set of glass pores is filled with condensed CH 2 Br 2 liquid, they will cease to act as an X-ray scatterer and only the remaining empty pores will produce a measurable scattering intensity. It should be noted, however, that the sample cell which facilitates this adsorption process in conjunction with SAXS measurements may introduce an error in the temperature (held at 293 K), and consequently in the relative pressure, of the order of ± 0.2 K and ± 0.01, respectively. Small angle x-ray scattering measurements were performed on a JJ X-ray system (Denmark) equipped with a Rigaku Helium-3 detector and a Cu (λ = 1.54098 ) rotating anode operated at 40 kV and 40 mA. The sample-to-detector distance and the beam centre were precisely determined by calibration with the Ag-behenate standard (d001 = 58.38 ). Scattering data were corrected for dark current and empty tube scattering. The Q-range is varied approximately from 0.004 to 0.11 −1 . Nitrogen adsorption measurements at 77 K were performed using an Autosorb-1 static volumetric system (Quantachrome Instruments). Dibromomethane adsorption-desorption isotherms were conducted gravimetrically at 293 K by means of an Intelligent Gravimetric Analyser (IGA, Hiden Isochema). In both adsorption experiments the samples were outgassed overnight at 473 K under high vacuum. Although further details on the experimental procedure have been published elsewhere 5,6 , this is a novel type of experiment and a first time to our knowledge of scanning the hysteresis loop in conjunction with SAXS. Since the properties of the glass may vary between samples from different lots 34 , it is worth noting that our new and our old data, obtained at different time periods and places and with different Vycor samples, chemicals, and instruments, all reflect the same result; that is, an intensity increase at very low Q, well above the spinodal peak.

chemsum

{"title": "On the Formation of Nanobubbles in Vycor Porous Glass during the Desorption of Halogenated Hydrocarbons", "journal": "Scientific Reports - Nature"}

radical_tropolone_biosynthesis

## Abstract: Non-heme iron (NHI) enzymes perform a variety of oxidative rearrangements to advance simple building blocks toward complex molecular scaffolds within secondary metabolite pathways. Many of these transformations occur with selectivity that is unprecedented in small molecule catalysis, spurring an interest in the enzymatic processes which lead to a particular rearrangement. In-depth investigations of NHI mechanisms examine the source of this selectivity and can offer inspiration for the development of novel synthetic transformations. However, the mechanistic details of many NHI-catalyzed rearrangements remain underexplored, hindering full characterization of the chemistry accessible to this functionally diverse class of enzymes. For NHI-catalyzed rearrangements which have been investigated, mechanistic proposals often describe one-electron processes, followed by single electron oxidation from the substrate to the iron(III)-hydroxyl active site species. Here, we examine the ring expansion mechanism employed in fungal tropolone biosynthesis. TropC, an α-ketoglutaratedependent NHI dioxygenase, catalyzes a ring expansion in the biosynthesis of tropolone natural product stipitatic acid through an under-studied mechanism. Investigation of both polar and radical mechanistic proposals suggests tropolones are constructed through a radical ring expansion. This biosynthetic route to tropolones is supported by X-ray crystal structure data combined with molecular dynamics simulations, alanine-scanning of active site residues, assessed reactivity of putative biosynthetic intermediates, and quantum mechanical (QM) calculations. These studies support a radical ring expansion in fungal tropolone biosynthesis. ## Introduction Enzymes catalyze challenging transformations with exquisite control over chemo-, site-and stereoselectivity, directing the synthesis of structurally-complex products from bioavailable starting materials. These transformations often take advantage of the three-dimensional architecture of the enzyme active site to guide reactive intermediates toward formation of the desired product. 1 Biocatalytic control over the fate of reactive intermediates is a common mechanistic architype in a variety of enzymes such as cyclases, which leverage active site geometries to induce selective cyclization reactions, polyketide synthases that deliver specific products from large, structurally-complex intermediates, Diels-Alderases, which choreograph the relative position of two reactive components to form a stereo-enriched product, and nonheme iron (NHI) dioxygenases, an enzyme class known to catalyze numerous skeletal rearrangements through radical or polar mechanisms. In many cases, reactions of identical intermediates generated outside of the enzyme active site proceed in an uncontrolled fashion, leading to racemic products or fundamentally divergent reactivity. In contrast to the selectivity demonstrated by enzyme-mediated transformations, small molecule-enabled transformations often require careful substrate design to achieve a desired rearrangement, adding synthetic steps and reducing the efficiency and sustainability of the resulting synthesis. Radical-based rearrangements are particularly impacted by selectivity challenges, as competing radical pathways often complicate anticipated reaction outcomes. As a result, selectivity can be achieved only through optimization of substrate design and reaction conditions to favor the desired reaction pathway. In comparison, biocatalysts operate with precise control over the reaction outcome, enabling highly selective and direct synthesis of natural product scaffolds. Therefore, thorough examination of the mechanistic details of enzyme-catalyzed rearrangements is critically important to the development of novel, selective approaches to complex molecule synthesis. Non-heme iron (NHI) enzymes perform an array of oxidative transformations and rearrangements in secondary metabolism. 4, Members of the α-ketoglutarate-dependent family of NHI dioxygenases couple the oxidative decarboxylation of α-ketoglutarate (α-KG) to the activation of molecular oxygen, generating an iron(IV)-oxo species (17, Fig. 1D). Through this common mechanism of oxygen activation, NHI enzymes catalyze a variety of selective transformations that are often initiated by hydrogen atom abstraction, followed by a variety of processes including hydroxylation, desaturation, halogenation, endoperoxidation, epimerization, ring expansion and ring contraction, among others. 4, Understanding how the fate of a radical intermediate is controlled by each catalyst has motivated structural, spectroscopic and computational studies of NHI enzymes. In-depth interrogation of NHI-catalyzed reaction mechanisms has revealed critical enzyme-substrate interactions that dictate the reaction outcome in several cases. These studies provide mechanistic detail for several NHI-catalyzed rearrangements, including ring contractions and ring expansions. Ring-modifying transformations often proceed through an initial C-H atom abstraction, followed by radical rearrangement and termination by single electron oxidation or hydroxylation to yield product. The NHI active site architecture dictates the reaction outcome by exerting fine control over events downstream from hydrogen atom abstraction, enabling highly selective transformations. For example, AusE and PrhA are highly related NHI enzymes (78% sequence ID) that catalyze divergent transformations on a common substrate to generate preaustinoid A3 (1) and berkeleydione (2), respectively. Structural and computational studies have shown that minor differences in their active site architecture play a critical role in determining the outcome of the initial enzyme-catalyzed desaturation, ultimately leading to the formation of two distinct natural products through a divergent radical rearrangement process. 7,9 Similar mechanistic studies found that the NHI-catalyzed biosynthesis of cycloclavine (3) and deacetoxy-cephalosporin C (4) proceed through a one-electron reaction pathway, suggesting a common mechanistic architype for ring rearrangements catalyzed by NHI enzymes. 18 Natural products often possess medium-sized rings (7-11 atoms) which can be difficult to access using small molecule techniques. One such class of synthetically-challenging targets are tropolones, a structurally-diverse group of bioactive metabolites with an aromatic cycloheptatriene core structure containing an α-hydroxyketone moiety (see Figure 1C, 11). Tropolones have been synthesized through a variety of approaches, typically involving ring expansion. 22 These transformations include classic two-electron rearrangements such as the Büchner reaction, 23 the de Mayo fragmentation 24 and [5+2] cycloadditions. 22 In addition, radical-based approaches have been successful in synthesizing tropolones from ortho-dearomatized catechols (5) through a selective radical ring expansion mechanism (Figure 1B). This approach required pre-functionalization of the arene to ensure that radical initiation occurred at the methyl halide ipso to the site of dearomatization, leading to the desired rearrangement (Figure 1B). While these methods have been successful in enabling ring expansion, they require arduous synthetic efforts to achieve the desired substitution pattern of the tropolone natural product, preventing facile access to the target compound. 22 In contrast, Nature efficiently assembles complex tropolones using available biosynthetic machinery. In fungi, the identification of the stipitatic acid (12) biosynthetic gene cluster in T. stipitatus provided a blueprint for the assembly of these aromatic seven-membered rings in Nature (Figure 1C). 21 Fungal tropolone biosynthesis, in the case of stipitatic acid, commences with polyketide synthase production of 3methylorcinaldehyde ( 9) and subsequent flavin-dependent monooxygenase-mediated oxidative dearomatization. 21,27 It has been established that the resulting dienone (10) undergoes an oxidative ring expansion catalyzed by an α-KG-dependent NHI enzyme, TropC, to afford stipitaldehyde (11). Downstream enzymatic modifications of stipitaldehyde (11) generate stipitatic acid (12). 21 Since the identification and characterization of enzymes involved in stipitatic acid biosynthesis, several other tropolone natural products have been shown to proceed through the same oxidative dearomatization/ring expansion cascade to generate the tropolone core. Motivated to understand the chemical steps of this powerful transformation, we initiated our studies on the chemistry and mechanism of the TropC-catalyzed ring expansion. In particular, we sought to investigate two transformations which TropC has been shown to catalyze (1) ring expansion to generate stipitaldehyde ( 11) and ( 2) a fragmentation reaction which produces trihydroxybenzaldehyde (18), a known shunt product in fungal tropolone biosynthesis (Figure 1D). As the fate of a radical intermediate in NHI dioxygenase-catalyzed transformations is critical to the observed reactivity of the enzyme, we envisioned that TropC-catalyzed ring expansion could occur through several possible mechanistic pathways. Cox and coworkers initially proposed that TropC performs a polar ring expansion to produce stipitaldehyde (11, Figure 1D, Path 1). Their proposal commences with TropC-mediated C-H atom abstraction on dienone 10 to generate radical species 14. Subsequent rebound hydroxylation would afford diol 13, which is proposed to undergo a semi-pinacol rearrangement to form stipitaldehyde (11). We anticipated that this rearrangement could be assisted by residues in the TropC active site through activation of the alcohol leaving group in diol intermediate 13. The proposed ring expansion mechanism is also consistent with the observed formation of a shunt product in this biosynthetic pathway, 18, which is anticipated to arise through loss of formaldehyde and concomitant rearomatization to produce trihydroxybenzaldehyde 18 (Figure 1D, Path 3). 21 Based on the radical reaction pathways that have been proposed for other ring modifying enzymes, we envisioned an alternative mechanism to arrive at stipitaldehyde in which radical 14 directly undergoes ring expansion, followed by radical termination via a single electron oxidation involving the iron(III)-hydroxyl species (Figure 1D, Path 2). Under this model, the product generated by the enzyme is determined by the radical termination process which occurs in the active site. If rebound hydroxylation predominates, then trihydroxybenzaldehyde (18) will be the major product formed. If radical rearrangement is preferred under the reaction conditions, then stipitaldehyde (11) formation will dominate. To decipher the ring expansion mechanism used by TropC, we aimed to investigate the active site architecture of the enzyme, as well as the reactivity of proposed ring expansion intermediate 13. ## Results and discussion To determine the nature of the TropC-catalyzed ring expansion, we began by performing a computational analysis of the proposed pathways (Figure 1D, Path 1-3). Through these calculations, we aimed to compare the relative barriers of proposed ring expansion pathways to assess the feasibility of a one-or two-electron ring expansion mechanism. Each mechanistic pathway was evaluated using a small molecule NHI mimetic complex (see Supplementary Figure S32). We manually performed tautomerization of reaction pathway products to model stipitaldehyde (11) and the trihydroxybenzaldehyde shunt product 18 as protonated, neutral structures, reflecting the enzymatic reaction conditions for TropC-catalyzed ring expansion (pH 7.0). Water molecules were used as proton shuttles to mimic the protonation and deprotonation events that would be facilitated by residues in the enzyme active site. We began our computational investigations using the radical intermediate 14 (Figure 2), which is the divergence point for each of the pathways along the potential energy surface (PES). We first explored the radical rearrangement (Path 2), which occurs with a barrier of 12.7 kcal/mol (16). A subsequent H-atom abstraction by an iron(III)-hydroxyl species to produce the tautomer of stipitaldehyde (11) was calculated to be barrierless. This calculation was supported by literature precedent which demonstrated that tropolones could be synthesized from ortho-dearomatized radicals (Figure 1B). 22,26 We then explored the rebound hydroxylation and semi-pinacol rearrangement (Path 1). As expected, rebound hydroxylation to produce diol 13 was found to be a barrierless process in our active site model. 12, In an enzyme active site, we anticipate that substrate movement and alignment would be restricted by local residues, resulting in thermodynamic barriers to this process. 12, The barrier for the proposed semi-pinacol rearrangement was observed to be excessively high at 48.8 kcal/mol (21), indicating that this transformation is not likely to occur. These barriers are in agreement with computational analysis by Siegbahn and coworkers examining ring expansion mechanisms of ortho-dearomatized phenols, indicating that a semi-pinacol type ring expansion is unlikely to occur in these systems. 32 In comparison, fragmentation to afford an aromatic product from diol 13 (Path 3) to produce formaldehyde and trihydroxybenzaldehyde 18 was found to be a low barrier process at 18.2 kcal/mol (23). These calculations suggest that diol 13 favorably undergoes fragmentation and rearomatization (Path 3), rather than the proposed semi-pinacol ring expansion (Path 1). The calculated low barrier for Path 3 is supported by experimental observation of trihydroxybenzaldehyde formation during the TropC-catalyzed reaction. Taken together, the QM simulations support our alternative mechanistic proposal in which the radical rearrangement is a kinetically and thermodynamically accessible route towards tropolone formation (Path 2). Next, we sought experimental evidence to discriminate between the two reaction pathways under consideration for TropC-catalyzed ring expansion. We aimed to further interrogate the proposed ring expansion mechanisms by performing a detailed mutational analysis of TropC in order to determine if specific active site residues are responsible for catalyzing the ring expansion. To gain structural information to guide this work, we obtained TropC crystals using the sitting drop vapor diffusion method (see supplemental information for details). Following data collection, we found that TropC crystallized in the P3121 space group with two molecules of TropC in the asymmetric unit (PDB ID: 6XJJ). The crystal structure of unliganded TropC was solved at a resolution of 2.7 with the molecular replacement method using the structure of thymine-7-hydroxylase (T7H) of Neurospora crassa as a search model (Figure 3). 33 Our analysis of the structural data indicated that TropC adopts a similar architecture as other α-KG-dependent NHI dioxygenases 34 . The three-dimensional structure of TropC consists of double-stranded β-helix (DSBH or jelly-roll) fold at the core of the protein 34 . The DSBH core of TropC is comprised of ten anti-parallel β-strands, which form two β-sheets, called major and minor β-sheets. Major β-sheets of TropC include β1-5, β8 and β10, and the minor β-sheet consists of β6, β7 and β9 (Supplementary Figure S29). The exterior of the major β-sheets of the DSBH fold is surrounded by α-helices (α1-3, α5 and α7) to form a compact globular structure. Structures similar to TropC in the Protein Data Bank (PDB) were explored by using the DALI server (Supplementary Table S2), 35 revealing the highly similar structures of isopenicillin N synthase from Pseudomonas aeruginosa PAO1 (PaIPNS, PDB ID: 6JYV, Z-score: 32.2) 36 and thymine-7-hydroxylase (T7H, PDB ID: 5C3Q, Z-score: 31.8) of Neurospora crassa. 36 An overall structural comparison of TropC with PaIPNS and T7H suggested a putative substrate binding site for TropC (Supplementary Figure S30). Overlaying TropC with T7H revealed several shared residues which define the substrate binding site in the T7H structure (Supplementary Figure S31). 33 Specifically, residues F284 and F213 align well with two conserved aromatic residues in T7H, F292 and Y217, respectively. 33 F292 and Y217 have been demonstrated as critical for substrate binding and alignment in T7H, with F292 providing π-π stacking interactions that align thymine in the enzyme active site for productive catalysis. 33 The conservation of these aromatic residues in homologs of T7H and their alignment with F284 and F213 in TropC suggests that these residues may play an analogous role in substrate alignment. 33 In addition, bound metal ion was observed in the structure of TropC, which was refined to be Fe(III). As has been observed with the active site of other α-KG-dependent NHI dioxygenases, our structure demonstrates that the conserved HxD/E…H metal binding residues are involved in interaction with Fe(II) 36 . Specifically, our data showed the coordination of residues H210, D212 and H269 (located at the loop between β4-5 and β9) to the active site Fe(III) atom (Figure 3). Crystallography experiments yielded vital information on structural features of TropC and enabled further investigation of mechanistic considerations by computational and mutagenic analysis. Toward this goal, we anticipated that an enzyme-substrate complex would provide critical information for determining the active site residues that are involved in ring expansion catalysis. To construct a computational substrate-bound model, we modeled the missing electron density from the C-terminus and used an overlaid structure of T7H to establish the coordinates of critical enzyme cofactors, such as α-KG, ferrous iron and substrate (see Supplemental Information for model construction and simulation details). To prepare the model for substrate binding studies, we performed a molecular mechanics (MM) minimization over 5000 steps. The system was then prepared for a combined quantum mechanics and molecular dynamics (QM/MD) simulation. The system was subjected to three phases of MD simulations in which the system was heated, equilibrated, and sampled for a total of 12 ns to generate an enzyme-substrate-cosubstrate complex with the appropriate geometry and alignment for C-H atom abstraction (see Supplemental Information for details). Our analysis of the substrate-bound TropC model revealed that substrate 10 was flanked by numerous active site residues that could potentially be involved in substrate binding or catalysis (Figure 4C). Residues from the TropC substrate pocket were identified and selected for alanine screening to determine if specific amino acids are critical for the ring expansion process. We anticipated that a TropC variant, without the required residue for a polar, semi-pinacol-type ring expansion, would be unable to generate stipitaldehyde, resulting in a change to the ratio of observed enzyme products toward formation of the shunt product, 18 (Figure 4A). We carried out mutagenesis of twelve active site residues, generating the corresponding alanine variants (Figure 4B and Supplementary Figure S2) as well as variants with isosteric and isoelectronic residues at the same position. These residues were chosen using the substrate-bound TropC model as a guide and were selected for their proximity to the substrate as well as their ability to participate in the proposed reaction pathways. Many of the resulting variants were soluble, but catalytically inactive, suggesting that the structural changes in some variants prevented productive catalysis. Catalytically active variants exhibited reduced enzymatic conversion to products, but the product profiles were largely unaltered, and stipitaldehyde (11) remained the major product under the reaction conditions, suggesting that most of the substrate-flanking residues were not involved in catalysis. However, the F284A variant uniquely demonstrated a shift in the product profile to produce nearly equimolar amounts of stipitaldehyde (11) and trihydroxybenzaldehyde (18). Analysis of the substrate-bound TropC model suggests that this residue could be important for positioning of the substrate in the active site. This proposed role is analogous to the function of residue F292 in T7H, which has been shown to be critical for binding and alignment of substrate for productive catalysis. The binding mode of substrate 10 in the TropC MD simulation demonstrates that F284 and F213 flank the substrate in the active site, potentially guiding the proper alignment for C-H atom abstraction. This proposal is likewise analogous to the reported role of F292 and Y217 in T7H, suggesting a similar function in aligning the TropC substrate to achieve catalysis. 33 To further investigate this hypothesis, we generated the isosteric variant TropC F284Y which reconstituted the ring expansion activity of the enzyme, providing further evidence that F284 is a critical residue for determining the product mixtures generated by TropC. We aimed to further analyze this reaction by generating a TropC F213A variant but were unable to produce soluble protein with this construct. Despite this challenge, the data generated through alanine scanning of the TropC active site suggested that substrate alignment plays a role in determining the product profile of the enzyme. Perturbations of this alignment often result in a change to the products generated by the enzyme, providing clues about the mechanism of the transformation. 7,9 In particular, substrate alignment relative to the iron(III)-hydroxyl species is key in discriminating whether rebound hydroxylation occurs over other NHI-catalyzed reactivity, such as halogenation. 37 In the context of our TropC model, these observations support the proposed mechanistic pathway for radical ring expansion (Path 1) in which the fate of the radical (rearrangement versus rebound hydroxylation) determines which product is generated by the enzyme. In addition, mutagenic changes to polar residues did not produce a corresponding shift in the product profile to the production of trihydroxybenzaldehyde 18, suggesting that local acid or base catalysis does not drive a ring expansion mechanism (Figure 1D, Path 1). We aimed to further interrogate the proposed semi-pinacol rearrangement (Figure 1D, Path 1) by evaluating the reactivity of the proposed diol intermediate 13. To generate the target diol in the absence of TropC, a biocatalytic approach was employed using CitB and TropB. 21,27, We aimed to directly synthesize diol 13 through oxidative dearomatization of benzylic alcohol 25 using TropB. Notably, trihydroxybenzaldehyde 18 was detected in these reactions, but diol intermediate 13 was not observed, suggesting that the fragmentation reaction described in Path 3 (Figure 1D) occurs spontaneously under the reaction conditions. To further probe whether diol 13 is a ring expansion intermediate in TropC catalysis, we again performed a TropB-catalyzed oxidative dearomatization of benzylic alcohol 25 and included TropC in this reaction. We envisioned that diol 13, when generated in situ, could enter the active site of TropC and undergo a ring expansion reaction as proposed in reaction Path 2 (Figure 1D). In this reaction, exclusive formation of trihydroxybenzaldehyde ( 18) was observed with no stipitaldehyde (11) detected, suggesting that fragmentation and rearomatization of diol 13 is the predominant mode of reactivity observed for this putative intermediate (Figure 4D). In support of this finding, Riess and coworkers noted this same fragmentation and rearomatization reactivity in their attempts to perform ring expansion reactions on ortho-dearomatized phenols to produce tropolones (Supplementary Figure S29). 41 These observations also agree with the computational modelling which suggests that fragmentation and rearomatization is a low-barrier process (18.2 kcal/mol). Taken together, these data indicate that diol intermediate 13 is unlikely to be the correct ring expansion intermediate in the TropC-catalyzed reaction, further suggesting that a radical mechanism (Path 2) leads to tropolone formation in this NHI system. ## Conclusions These experimental observations and QM calculations suggest that TropC-catalyzed ring expansion occurs through a radical-based process (Path 2), rather than the previously proposed rebound hydroxylation/semi-pinacol sequence (Path 1). This mechanistic proposal is supported by structural characterization, modelling and mutagenic analysis of the TropC active site, demonstrating that modification of a residue involved in substrate positioning altered the products generated by the enzyme. These data suggest that substrate positioning in TropC determines which radical termination step predominates: rebound hydroxylation or radical rearrangement followed by single electron transfer. Furthermore, we have demonstrated that rebound hydroxylation generates an intermediate (13) which does not undergo ring expansion, but rather a fragmentation and rearomatization process to produce trihydroxybenzaldehyde 18. The observed reactivity of intermediate 13 illustrates the thermodynamic favorability for rearomatization over ring expansion in ortho-dearomatized phenols, as has been documented in previous studies. 32,41 This revised proposal of radical-based ring expansion is also supported by literature precedent, demonstrating that tropolones can be synthesized from ortho-dearomatized radicals. These findings provide strong evidence of a radical-based mechanism for fungal tropolone biosynthesis, representing a major shift in the current understanding of the role of NHI enzymes in this process. In addition, our observations provide new insight into the mechanistic processes harnessed by NHI enzymes for the selective synthesis of complex molecules. We anticipate that computational studies that consider active site geometries could provide additional insight into these transformations and the observed behavior of TropC variants. Furthermore, we envision that the molecular details of NHI-catalyzed ring expansion can be leveraged to address current challenges in the synthesis of tropolone natural product scaffolds.

chemsum

{"title": "Radical tropolone biosynthesis", "journal": "ChemRxiv"}

the_presence_of_microplastics_in_commercial_salts_from_different_countries

## Abstract: The occurrence of microplastics (MPs) in saltwater bodies is relatively well studied, but nothing is known about their presence in most of the commercial salts that are widely consumed by humans across the globe. Here, we extracted MP-like particles larger than 149 μm from 17 salt brands originating from 8 different countries followed by the identification of their polymer composition using micro-Raman spectroscopy. Microplastics were absent in one brand while others contained between 1 to 10 MPs/Kg of salt. Out of the 72 extracted particles, 41.6% were plastic polymers, 23.6% were pigments, 5.50% were amorphous carbon, and 29.1% remained unidentified. The particle size (mean ± SD) was 515 ± 171 μm. The most common plastic polymers were polypropylene (40.0%) and polyethylene (33.3%). Fragments were the primary form of MPs (63.8%) followed by filaments (25.6%) and films (10.6%). According to our results, the low level of anthropogenic particles intake from the salts (maximum 37 particles per individual per annum) warrants negligible health impacts. However, to better understand the health risks associated with salt consumption, further development in extraction protocols are needed to isolate anthropogenic particles smaller than 149 μm. Since their mass production in the 1950s, global plastic production has been increasing, which exceeded 322 million tons in 2015 1 . Mismanaged plastic waste could find their ways to oceans 2 . Meanwhile, continuous fragmentation of large plastic objects has resulted in the accumulation of smaller particles called microplastics (MPs, sized between 1 and 1000 μ m 3,4 ). Also, MPs may be directly introduced to the aquatic environments through their primary sources (e.g., synthetic sandblasting media, cosmetic formulations, textiles) . The widespread distribution of MPs in aquatic bodies is well documented, such as in the Celtic sea 8 , Laurentian Great Lakes 9 , Persian Gulf 10 , and in sub-tropical gyres 11 . Accordingly, it is expected that products originating from the contaminated water bodies are also loaded with MPs. Several studies have shown the presence of MPs in seafood products like clams 12 and fish 13 . Therefore, the consumption of seafood products could be a significant route of exposure to MPs in humans. For instance, top European shellfish consumers are expected to ingest up to 11,000 plastic particles per annum 14 . Microplastics might be of health concern since they have been shown to carry hazardous chemicals 15,16 and microorganisms 17 . Despite the relatively well-documented occurrence of MPs in seafood products, little is known about MP loads in abiotic saltwater products, which are expected to inevitably contain contaminants from the water. Sodium is an essential element for the human body to maintain homeostasis and is mainly consumed as common salt (sodium chloride, NaCl) 18 . Commercial sea and lake salts are mainly produced through a crystallization process as a result of seawater evaporation or naturally occurring brine under the combined effects of sunlight heat and wind 19,20 . There are concerns over the potential transfer of water contaminants into sea salt after the crystallization and concentration process 20 . In the only available study on abiotic products, Yang et al. 21 showed the presence of up to 681 MPs/kg in salts originating in China. However, the chemical composition of individual MP particles was not determined, with MP identification based instead on grouping the particles according to their morphological features and analyzing representative samples through Fourier Transform Infrared Spectroscopy (FT-IR). Nevertheless, visual sorting is not the most reliable method to identify MPs 22,23 as this could significantly overestimate the concentration of anthropogenic materials. No report is available, however, on MP loads in salt samples from other regions of the world that are potentially consumed by around 6 billion people, excluding billions of other organisms such as cattle that need to consume salt on a regular basis 24 . In this study, we further investigated the presence of MPs by extracting particles from 17 different brands of salt originating from 8 countries over 4 continents. Density separation and visual identification were employed to initially isolate MP-like particles. Finally, all particles were analyzed by micro-Raman spectroscopy for their chemical composition. ## Results The presence of a large quantity of white sediment in the lake salt from Malaysia (Table 1) blocked the 149 μ m-pore size filter membrane. These particles were identified as calcium carbonate (CaCO 3 ) using Raman spectroscopy. Therefore, this sample was excluded from MP analysis. A total of 72 MP-like particles were isolated from 16 salt brands. The average particle size (mean ± SD) was 515 ± 171 μ m. The size of the smallest particle was 160 μ m and the largest sized 980 μ m. Figure 1 presents a histogram of the number of the particles sorted by size. As depicted by Fig. 2a, 30 particles (41.6%) were confirmed as plastic polymers, 17 particles (23.6%) were pigments, 21 particles (29.1%) were not identified, and 4 particles (5.50%) were non-plastic items (i.e., amorphous carbon). The major plastic polymers were PP at 40.0% of the total plastic polymers followed by PE (33.3%), polyethylene terephthalate (PET; 6.66%), polyisoprene/polystyrene (6.66%), polyacrylonitrile (10.0%), and polyamide-6 (nylon-6, NY6; 3.33%) (Fig. 2b). Particles identified as pigments were phthalocyanine (82.3%), chrome yellow (5.88%), hostasol green (5.88%), and hostaperm blue (5.88%) (Fig. 2c). The abundance of MPs per salt sample ranged from 0 per kg in the salt sample # France-F (i.e. Country of origin: France, brand F) to 10 in the salt sample # Portugal-N. Figure 3a and b are the stacked bar chart of the number of plastic polymer and pigment particles isolated from each salt brands, respectively. 1. Detailed information on the salt samples analyzed in this study. 1 Polyethylene. 2 Polyethylene terephthalate. 3 Polypropylene. With regards to the particle morphology, the predominant type were fragments (63.8%) followed by filaments (25.6%), and films (10.6%) (Fig. 4). No sphere beads were isolated from the salt samples. Figure 5 illustrates the microscopic images of some of the isolated particles. Supplementary Information Fig. 1a-e present the microscopic images and spectra from some of the isolated MPs along with the spectra of reference materials. ## Discussion In this study, we developed a simple and cost-effective protocol to isolate MPs from salt samples. The presence of insoluble particles quickly blocked the filter papers with pore sizes of 2.7, 8, and 22 μ m. Initially, we hypothesized that the blockage was due to the presence of organic materials. However, lack of changes on the filtration after KOH digestion showed the presence of digestion-resistant organic or inorganic materials. Therefore, dilutions with deionized water, followed by filtration through a membrane with a larger pore size (149 μ m), was the simplest and quickest method to isolate MPs from the salt samples. Our recent study demonstrated a high efficcincy of 4.4 M NaI solution to separate (recovery rate > 95%) plastic polymers from high-density particles like shell fragments from bivalves and sand grains 3 . To minimize the chance of overestimating or underestimating MP prevalence in the salt samples, in addition to NaI extraction, we implemented microscopical examination and Raman spectroscopy. Despite processing the samples through NaI extraction and visual identification, 5.50% of the particles were amorphous carbon (Supplementary Information Figure 1e), which underscores the necessity of using spectroscopic techniques to identify the chemical composition of the isolated particles. Micro-Raman spectroscopy is a highly specific technique used to identify the composition of biological, mineral or polymer samples. It offers a number of advantages such as analysis of microscopic particles while being non-invasive towards the samples 23,25 . In addition, Raman measurements do not depend on the transmission of light through the particle, which consequently allows an accurate analysis of thicker or pigmented particles 23 . Some of the earlier studies solely relied on the morphological characteristics of MPs, like shape and color to identify MPs in environmental samples 26,27 while others partially confirmed the particle composition through a random selection of the isolated particles 21,28 . Although observation is an indispensable part of polymer identification, it cannot be employed as a stand-alone technique for particle characterization because it is unlikely to be sufficient to identify the polymer type through morphological features. Consistent with the findings of this study, fragments and filaments have been reported as the main form of MPs 29,30 . The absence of microbeads in the salt samples may indicate their low prevalence in aquatic environments. Polypropylene (PP) and polyethylene (PE) were the most abundant plastic polymers in the salt samples (40.0% and 33.3%, respectively) which is consistent with reports on their wide distribution in the marine environment 22 . The presence of these polymers in salt samples could be due to the low density of PP (0.90-0.91 g/cm 3 ) and PE (0.91-0.96 g/cm 3 ) allowing these to float on the water surface and be readily directed into saltpans. In addition, their low density may facilitates their spread by becoming airborne. Polyisoprene/polystyrene (styrene-isoprene-styrene block copolymer) were the other detected synthetic polymers in the salt samples. These polymers are used when elasticity and easy processing is required, such as in adhesives and sealants 31 . A few isolated particles from the salt samples had a similar composition to their packaging. This might indicate degradation of the packaging materials leading to the contamination of the salt product. Nevertheless, this hypothesis was rejected since all the fragments or films were highly corroded, indicating their long-term presence in the environment. Almost one fourth of the isolated particles were identified as pigments (phthalocyanine, chromate yellow, and hostaperm blue) because the strong Raman signal of these pigments hindered the identification of plastic polymers. Phthalocyanine is a synthetic pigment and is extensively used in the plastics industry 32 and was the main pigment isolated from the salt samples (Fig. 2c). Hostaperm blue falls under the copper phthalocyanine chemical class and is an industrial dye that is mainly used in the plastics industry 32 . Victoria blue is commonly used as a coloring agent in polyacrylic fibers (Supplementary Information Figure 1a) 30 that are mainly introduced to the marine environment through the washing of clothes after passing through sewage treatment plants 34 . Meanwhile, lead chromate (yellow) pigment is a toxic compound that has extensive applications in paints and plastic industries owing to its excellent light-fastness and low cost 33,35 . Earlier studies have attributed exposure to lead chromate pigment with incidents of bronchial carcinoma 36 , cerebrovascular disease 37 , and nephritis 38 in humans. However, the occurrence of only one particle of lead chromate pigment per Kg of salt # South Africa-Q, poses a negligible threat to the health of consumers. Initially, we hypothesized that the pigment particles might be paint particles. However, since none of the extracted particles shared similar mechanical properties as paint particles like brittleness 39 , we suggest the absence of paint particle in the salt samples. Other than plastics, pigments are widely used in other materials like textile, rubber, and fiberglass 40,41 . Van Cauwenberghe et al. suspected that the particles identified as copper phthalocyanine, polychloro copper phthalocyanine, and permanent red in deep-sea sediments 42 , as well as copper phthalocyanines and haematite in bivalves 14 to be plastic materials. Similarly, in this study we could confirm that the pigmented particles had an anthropogenic origin but could not ensure they were MPs. In the present study, a significant portion (29.1%) of the particles was not identified by Raman spectroscopy. Photo-degradation and weathering are the two major factors suggested as the causes for variation in the spectroscopic spectra of polymers such as PVC 23 . Moreover, the presence of additives could alter the polymer spectra and hinder comparisons with the reference library 23,42 . Another reason for having unidentified samples is the lack of a comprehensive spectra library to identify mixed samples 23 . Lake salt from Malaysia was excluded from MP analysis because this contained a large volume of sediments, which were later identified as CaCO 3 . Calcareous sedimentation is a common process occurring in lakes mainly due to the assimilation of carbon dioxide by photosynthesizing plants and/or seasonal temperature effects on the solubility of carbon dioxide and calcite 43 . It should be noted that the other lake salt (# Iran-I) did not contain calcareous sedimentation, which shows the variation in calcium contents among different lake salt brands. In 2010, the global daily sodium consumption was 3.95 g/day (equivalent to 9.88-10.2 g salt/day 44 ) corresponding to 3.6 to 3.7 Kg salt per annum. The number of anthropogenic particles (MPs and pigments) detected in the salt samples ranged from 0 (sample # France-F) to 10 (sample # Portugal-N) MPs/Kg. Based on this data, humans could ingest a maximum of 37 plastic particles annually. This should take into account that this maximum value is based on the assumption that sea salt is the sole source of sodium intake. Other sources of sodium supply are food additives like monosodium glutamate and preservatives. Therefore, in real scenarios, the maximum MPs intake is probably even less than 37 particles. Microplastics may cause adverse effects to organisms through causing micro injuries (mainly in the case of fragments) 4 or the release of pollutants that had been sorbed during their prolonged incubation in the water. In the case of the latter, MPs have shown the ability to sorb persistent organic pollutants (POPs) 45 and subsequently desorb them under simulated gut conditions 46 . However, recent studies have argued the lower importance of MPs as a vector for translocation of POPs to aquatic biota as compared to other routes like food and water 15,16 . Despite a potentially high concentration of POPs and other contaminants in MPs, the combination of their small particle size and low prevalence indicate that the consumption of sea salt does not appear to be a major route for the contaminant transfer into the human body as compared to other sources like water and food. Due to technical limitations, however, we could only quantify the level of particles larger than 149 μ m. The prevalence of smaller particles in the salt samples, however, might be higher than the larger ones. Smaller sizes could facilitate their translocation into other organs and, therefore, cause a higher degree of toxicity. For example, in a study by Lu et al. 47 , 20 μ m polystyrene (PS) microspheres accumulated in the gills and gut of zebrafish (Danio rerio), while 5 μ m microbeads were incorporated into the gills and gut as well as liver. Further advances in isolation techniques are needed to quantify smaller MP particles before making a more accurate justification on the health impacts of sea salt consumption. However, it should be taken into account that salt is not the only edible item that has been shown to contain MPs. These have been previously detected in clams 12 , mussels 14 , fish 13 and unexpectedly in honey 48 as well as beer 49 . Therefore, the long-term consumption of various products containing MPs might become a concern. Due to their low density and slow degradation, plastics are becoming the chief cross-border contaminant that often travels far from their original source. Hence, MPs found in the salt samples of one country could have been produced by another country thousands of miles away. A potential solution to this global dilemma requires a strong commitment from all the countries to make a substantial improvement in plastic disposal and recycling. ## Conclusions The results of this study did not show a significant load of MPs larger than 149 μ m in salts originating from 8 different countries and, therefore, negligible health risks associated with the consumption of salts. The increasing trend of plastic use and disposal 50 , however, might lead to the gradual accumulation of MPs in the oceans and lakes and, therefore, in products from the aquatic environments. This should necessitate the regular quantification and characterization of MPs in various sea products. ## Methods Materials. A total of 17 brands of salt from Australia, France, Iran, Japan, Malaysia, New Zealand, Portugal, and South Africa were purchased from a Malaysian market. NaCl (analytical grade) was purchased from Merck (Darmstadt, Germany), and sodium iodide (NaI), potassium hydroxide (KOH), and ethanol 95% were supplied by R&M Chemicals (UK). Solutions of NaI (4.4 M) and KOH (10% w/v) were prepared by dissolving the powder/pellet in ultrapure deionized water. GF/D microfiber filter paper (pore size 2.7 μ m), and filter membrane No. 540 and 541 filter membranes (hardened ashless, pore size 8 and 22 μ m, respectively) were purchased from Whatman. The 149 μ m filter membranes were supplied by Spectrum Laboratories (USA). High-density polyethylene (HDPE), low-density polyethylene (LDPE), PP, PS, PET, polyvinyl chloride (PVC), NY6, and nylon-66 (nylon-66, NY66) virgin plastic fragments were supplied by Toxemerge Pty Ltd (Australia). About 10% of particles were sized below (D10) 40 μ m, 50% below (D50) 140 μ m, and 90% (D90) below 310 μ m. ## Extraction method development. In an attempt to employ a solvent to dissolve the salt particles or efficiently digest the organic materials, we compared digestion using a KOH solution 3 with appropriate dilutions with deionized water. A 20 g sample of sea salt from sample # Australia-A was placed into a 250 mL DURAN glass bottle (Schott, Germany) sealed with a premium cap and pouring ring (Schott, Germany), and then 200 mL KOH solution or deionized water were added (1:10 w/v). The bottle was manually shaken until the full dissolution of salt. These solutions were then incubated at 40 °C for 48 h and were vacuum filtered through 2.7, 8, or 22 μ m membranes. Results showed that both solvents failed to pass through the 2.7, 8 or 22 μ m filter membranes owing to the presence insoluble or digestion resistant materials. Therefore, dilution with deionized water was used to dissolve salts due to its cost effectiveness, accessibility, and inertness to the plastic polymers. Next, another experiment was run to find a filter membrane with the most suitable pore size. ## Filtration optimization. Upon poor filtration of the salt solution (dissolved in water or digested with KOH solution) through a 22 μ m filter membrane, a series of sieves were individually used to identify the best pore size for filtration. Briefly, a solution of sea salt was prepared in deionized water (1:10 w/v) and added into different 8-inch testing sieves-full height stainless steel frame and wire (pores sizes 38, 45, 53, 63, 90, 106, 125, or 150 μ m; Daigger Co., Vernon Hills, IL, USA) and mechanically shook on an orbital shaker at 200 rpm for 2 h. The salt solution could only fully pass through the 150 μ m pore size sieve. Therefore, disposable filter membranes with the nearest pore size to 150 μ m were chosen (Spectrum Laboratories, USA; Pores size: 149 μ m) for assessing MP loads in the salt samples. ## Method validation. To validate the dilution method, 5 g of NaCl (analytical grade), 0.05 g of crushed shell from Asian green mussel (Perna viridis), 0.05 g of sand grains, and 0.1 g of HDPE, LDPE, PP, PS, PET, PVC, NY6 or NY66 were weighed on a scale (0.1 mg precision) and then added into a 100 mL laboratory bottle in triplicate. The bottles were filled with 50 mL deionized water and shook manually until the complete dissolution of salt. After that, they were vacuum filtered through 8 μ m-pore size filter membrane using a vacuum pump (Gast vacuum pump, DOA-P504-BN, USA) connected to a filter funnel manifold (Pall Corporation, USA). To separate the high-density particles (shell fragments and sand grains), the filter membrane was soaked in 10-15 mL of NaI (4.4 M, density: 1.5 g/cm 3 ) in a 50 mL glass bottle and sonicated at 50 Hz for 5 min followed by agitation on an orbital shaker (200 rpm) for 5 min. Finally, the solution was centrifuged at 500 × g for 1 min, and the supernatant containing MPs was vacuum filtered through another 8 μ m filter membrane. To ensure the full isolation of MPs, this stage was repeated again. Finally, the filter membrane was placed into a clean glass Petri dish and dried at 50 °C in an oven for 5 h. No significant change was noticed in the weight of the filter membrane belonging to the procedural blank after filtration (Student's t-test, p > 0.05) showing the full dissolution of NaCl. The recovery rates of all polymers were > 95%. Therefore, dissolution in deionized water followed by density separation using NaI was used to isolate MPs from the commercial salt samples (Fig. 6). Extraction and characterization of MPs in the commercial salt samples. Extraction. One package of sea salt (200-400 g) was mixed with 2-4 L of deionized water in a 2 or 5 L laboratory bottle and then vacuum filtered through a filter membrane (pore size 149 μ m) to collect the insoluble materials. Between 3 and 5 salt packages (replicate) per brand were used to make a final weight of 1 kg. The filter membrane was placed into a 50 mL laboratory bottle and was subjected to NaI treatment as described earlier. The pellet was re-suspended in 10-15 mL of NaI, sonicated, agitated, and centrifuged to ensure the complete separation of MPs trapped within the insoluble materials. ## Characterization. Visual selection of the extracted particles. The filter membranes were inspected using a Motic SMZ-140 stereomicroscope (Motic, China). Visual inspection was performed and MP-like particles were sampled based on their morphological characteristics like color and shape. However, low-density particles such as carbon-based materials, plant tissues, and remains of invertebrate exoskeletons could not be excluded through density separation. The remains of these invertebrates were mostly yellow, light brown, or black in color, and possessed a soft texture and an even surface with smooth edges. The plant-based materials were dark brown or black. Extra care was taken when sampling the brown and black particles due to their similarity in appearance to non-plastic items. Microplastic-like particles were divided into foam (lightweight particles with spongy texture), fragments (jagged and irregular shape particles which often have an uneven surface), fibers/filaments (thin, straight and often cylindrical particles), films (thin plane of flimsy particles), or beads (rounded particles) 51 . The selected particles were photographed using a camera apparatus (AxioCam, ERc 5S, Germany). To measure the size of particles, digital images were examined using ImageJ software. Because of the irregularity of most of the isolated particles, sizes of the largest cross-section was measured. Raman spectroscopy. Particles were analyzed over a range of 150 to 3000 cm −1 using a Raman spectrometer (Horiba LabRam HR Evolution) equipped with a Single Mode Open Beam Laser Diode (Innovative Photonic Solutions) operating at a wavelength of 785 nm coupled with a charge-coupled device detector (Horiba Synapse). The experimental conditions were adapted as much as possible to limit fluorescence and increase the spectral quality of the measured particles. Another laser line (514 nm) was tested but due to the high level of fluorescence in the visible range, the laser 785 nm was selected, even though some fluorescence was still observed using this wavelength. The acquisition time and number of accumulations were adjusted for each scan such that the detector did not saturate and that the signal to noise was sufficient for performing a library search. A high numerical aperture objective (100X with NA 0.90) was used to increase the signal collection. The confocal hole was partially closed to limit the collection of the fluorescent background. The laser power was set low (below 3 mW) to avoid burning or damaging the material. The low power of the laser did not permit complete efficient bleaching; however, since the acquisition time and the number of accumulations were increased whenever necessary, some photobleaching occurred while accumulating. Before the library search, to reduce noise and enhance the spectrum quality without losing subtle spectral information, each spectrum passed through a baseline correction and denoising procedure (Labspec 6, Horiba Scientific). Pre-processed spectra were then evaluated and compared to the following spectral libraries: Raman polymers and monomers from Bio-Rad Sadtler and Raman Forensic from Horiba using the KnowItAll software from Bio-Rad. The Correlation algorithm (KnowItAll, Bio-Rad) was used to evaluate each query spectrum to the spectra of the databases. The Hit Quality Index (HQI) was used to rank the results of a spectral search. The HQI, which was scaled between 0 and 1000, indicates how well each spectrum from the database matches the test spectrum. The HQIs > 700 were accepted as evidence of a reliable match between the unknown and the reference spectrum. Furthermore, to investigate the possibility of MP leakage from the packing material to the salts, the chemical composition of the plastic packaging was also determined. ## Contamination prevention. To avoid contamination, cotton lab coat and nitrile gloves were worn at all times. All liquids (deionized water, ethanol) were filtered through 2.7 μ m glass fiber filter membranes. The glassware were washed once with dishwashing liquid, then with deionized water, and finally with ethanol. The work This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

chemsum

{"title": "The presence of microplastics in commercial salts from different countries", "journal": "Scientific Reports - Nature"}

seawater_electrolysis_for_hydrogen_production:_a_solution_looking_for_a_problem?

## Abstract: As the price of renewable electricity continues to plummet, hydrogen (H2) production via water electrolysis is gaining momentum globally as a route to decarbonize our energy systems. The requirement of high purity water for electrolysis as well as the widespread availability of seawater have led significant research efforts in developing direct seawater electrolysis technology for H2 production. In this Perspective, we critically assess the broad-brush arguments on the research and development (R&D) needs for direct seawater electrolysis from energy, cost and environmental aspects. We focus in particular on a process consisting of sea water reverse osmosis (SWRO) coupled to proton exchange membrane (PEM) electrolysis. Our analysis reveals there are limited economic and environmental incentives of pursuing R&D on today's nascent direct seawater electrolysis technology. As commercial water electrolysis requires significant amount of energy compared to SWRO, the capital and operating costs of SWRO are found to be negligible. This leads to an insignificant increase in levelized cost of H2 (<0.1 $/kg H2) and CO2 emissions (<0.1%) from a SWRO-PEM coupled process. Our analysis poses the questions: what is the future promise of direct seawater electrolysis? With an urgent need to decarbonize our energy systems, should we consider realigning our research investments? We conclude with a forward-looking perspective on future R&D priorities in desalination and electrolysis technologies. Hydrogen is back in fashion as a route to decarbonize our energy systems. Globally the hydrogen market is expected to grow by 47% from 142 billion USD in 2019 to 209 billion USD in 2027. 1 Since hydrogen is an energy carrier and not an energy source, it can be made dirty or clean. Today, over 95% of the 70 million tons of hydrogen produced annually comes from steam methane reforming (SMR), releasing 830 million tons of CO2 every year. 2,3 While blue hydrogen routes coupling SMR to carbon capture and storage (CCS) technologies are being tested at scale 4 , green or sustainable hydrogen made from water electrolysis and powered by low-carbon energy is crucial to attain climate neutrality. 2, As the price of renewable electricity continues to plummet, sustainable hydrogen production via water electrolysis is gaining momentum globally. 8 ## Water Electrolysis Technologies Today, the two main electrolyzer technologies that exist commercially are the alkaline electrolysis and proton exchange membrane (PEM) systems. Alkaline electrolysis is a mature and commercial technology, used since the 1920s, for hydrogen production in the fertilizer and chlorine industries. 2 Several alkaline electrolyzers with a capacity of up to 165 MW were built in the last century, although almost all of them were decommissioned when natural gas and SMR for hydrogen production took off in the 1970s. 2 Alkaline electrolyzers are characterized by lower capital costs compared to PEM systems due to the avoidance of precious catalysts. 9,10 While alkaline electrolysis systems operate at high efficiency (~55-70% LHV), low current density (< 0.45 A/cm 2 ) and low operating pressures (< 30 bar) negatively impact system size and hydrogen production costs. 11 Also, dynamic operation (frequent start-ups and varying power input) is limited (25-100% of nominal load) for alkaline electrolyzers, and can negatively affect system efficiency and gas purity. 12 On the other hand, PEM water electrolysis was pioneered by Grubb in the early fifties and General Electric Co. led development in 1960's to overcome the drawbacks of alkaline electrolysis. 2 The PEM systems run on pure water as an electrolyte solution, and so avoid the recovery and recycling of the corrosive potassium hydroxide electrolyte that is necessary in alkaline electrolyzers. Today, industries are inclined towards PEM system due to its compact design, high system efficiency (~52-69% LHV) at high current density (> 1-2 A/cm 2 ), fast response, dynamic operation (0-160% of the nominal load), low temperatures (20-80 °C) and the ability to produce ultrapure hydrogen at elevated pressure (30-80 bar). 2,9,12,13 PEM has seen drastic reduction in electrolyzer stack costs over the last few years and is expected to be the dominant technology for sustainable hydrogen production by 2030. 2,8,9 ## Direct Seawater Electrolysis One of the requirements of PEM water electrolysis is the need of highly pure water feeds with a minimum requirement of American Society for Testing and Materials (ASTM) Type II deionized (DI) water (resistivity > 1 MΩ-cm) while ASTM Type I DI water (> 10 MΩ-cm) is preferred. 14 ASTM defines Type II water, as required in commercial electrolyzers, as having a resistivity of > 1 MΩ-cm, sodium, and chloride content < 5µg/L and < 50 ppb of total organic carbon (TOC). 15 Alkaline electrolyzers are less stringent on water quality as compared to PEM, but still needs high purity water to achieve long-term stability. Such high purity water as required by water electrolysis systems is produced through a combination of either reverse osmosis (RO), multi-stage flash distillation (MSF), electrodialysis, multiple effect distillation (MED) to desalinate water, and commonly an additional technology such as ion exchange or electrodeionization (EDI). 16,17 The additional capital and operating cost associated with water purification has been the common argument that has spurred research activities into direct electrocatalysis of seawater for H2 production, with the rationale that seawater represents ~96.5% of earth's water resources. 18,19 A technology for direct seawater splitting could potentially be used in coastal arid zones that have limited access to freshwater yet plenty of access to seawater and renewable electricity from solar, wind and geothermal. Over the last few decades, significant research efforts have gone into direct seawater electrolysis (Figure 1). In the last decade, the field has seen 700+ scientific publications, and 340+ patent applications, which translates into millions of dollars of research funding allocated globally. Seawater electrolysis could be done to either produce chlorine via chloride oxidation or oxygen via water oxidation. Although chlorine is a valuable industrial chemical, the quantities produced for the growing hydrogen market would far exceed global demand for Cl2. 22 Therefore, one of the major challenges has been the development of active and stable anode catalysts for selective oxygen evolution over chlorine. 18,19 The competing chlorine evolution reaction (CER) is thermodynamically unfavorable compared to the oxygen evolution reaction (OER) (~480 mV higher in alkaline media), but it is a two-electron reaction, in contrast with OER which involves four electrons. This difference in the numbers of electrons involved makes OER kinetically unfavorable. 18,19,23 While some progress has been made on the development of selective catalysts for OER from seawater, reaching industrially relevant current densities (> 300 mA/cm 2 ) has been a major challenge. 18,19 Even though carbonate and borate ions are present in seawater, their average concentration is too low to sustain high current densities. Majority of reports at industrially relevant current densities use seawater with a borate buffer or additives such as KOH. 18,19 Furthermore, because seawater is essentially a non-buffered electrolyte, which causes a change in pH near the electrode surface during electrolysis (as high as 5-9 pH units), leading to salt precipitation as well as catalyst and electrode degradation. 24 Other challenges include the presence of other ions, bacteria, microbes, as well as small particulates, which limit the long-term stability of catalysts and membranes. 23 Despite the resources and efforts that have gone into developing this technology, direct seawater splitting remains in its infancy and distant from commercialization. ## Desalination Technologies Desalination -particularly via seawater reverse osmosis (SWRO) has seen tremendous technology advancements. Over the years, with improvement of membrane technology, more efficient energy recovery devices, and process optimization of reverse osmosis (RO) systems, have resulted in lowering the energy requirements, capital (CAPEX) and operating costs (OPEX) associated with the technology. In the last few decades, the energy requirement of SWRO desalination plants has decreased from ~9-10 kWh/m 3 to < 3 kWh/m 3 currently. 25,26 This has led to a decrease in levelized cost of SWRO desalinated water from > 2.2 $/m 3 to < 0.6 $/m 3 27 and resulted in a 6.5-fold increase in global desalination capacity (Figure 2). As of 2020, total production capacity reached > 100 million m 3 /day, ~70% of which is based on RO. The increase in production capacity is expected to follow the same trend in the next decades, as per planned and under construction plants. This raises the questions: what is the future promise of direct seawater splitting as compared to SWRO coupled with commercial water electrolysis for widespread implementation ? Further, with an urgent need to decarbonize our energy systems, should we consider realigning research priorities to disrupt the current fossil-fuel based carbon economy? Figure 1. (a) Annual number of publications extracted from https://www.dimensions.ai/, when a search for the topic "seawater splitting" was performed. (b) Annual number of patent applications found in Patsnap database (https://www.patsnap.com/), when a search for the topic i.e., "seawater" and "electrolysis" was performed in the title, abstract or claims. ## Sea Water Reverse Osmosis Coupled with Water Electrolysis In this viewpoint, we sought to address these questions by presenting a case study of a PEM water electrolysis system for 50 tons/day H2 production capability coupled to a SWRO plant for its water feed (Figure 3). The process is powered by the grid, which sources electricity from both fossil and renewable sources. We analyze both the economic and environmental feasibility of using SWRO water for PEM water electrolysis. As shown in Figure 3, PEM electrolysis plant consists of the electrolyzer stacks and the mechanical and electrical balance of plant (BoP) components. 34 The electrical BoP consists of the AC to DC rectifier for converting grid electricity while the mechanical BoP consists of other auxiliary components such as pumps, heat exchangers, temperature swing adsorption (TSA) subsystem and most importantly a deionizer (DI) system. 34 The design of SWRO and PEM systems are adapted from references 34,35 . The SWRO plant contains the RO unit which uses a membrane barrier and pumping energy to separate salts from saline water. Using high-pressure pumps, water is forced through semipermeable membranes that have a dense separation layer (thin film composite membrane) allowing the passage of pure water molecules while rejecting dissolved salts and other impurities. 36 In addition, in order to control RO membrane (bio)fouling and scaling, the SWRO system necessitates physical (e.g., dual media, sediment and carbon filters or low-pressure membranes) and chemical (e.g., coagulant polymer, antiscalant, acid, chlorination/dechlorination) pretreatment steps with variable complexity depending on raw feed water quality. 37 A combination of these filters provides a broad spectrum of reduction. The carbon filters remove volatile organic compounds (VOCs), chlorine (not tolerated by polyamide RO membranes) and other contaminants that give water a bad taste or odor, sediment filter removes dirt, colloidal matter and debris while the RO membranes remove >99.8% of total dissolved solids (TDS). 38 The SWRO-PEM coupled system could be located near coastal regions with intense solar irradiation and wind patterns to produce renewable electricity using photovoltaics or wind turbines or even offshore structures if hydrogen supply for shipping for example was desired.. With ample access to seawater such regions are already equipped with large desalination plants, as shown in Figure 4, making it feasible to couple RO technology to PEM water electrolyzers. Not surprising, such locations have also been identified as potential locations for implementation of direct seawater electrolysis if and when the technology can be commercialized in future. PEM electrolysis plants typically need ~ 10 kg water to produce 1 kg H2, that is total water requirement of 500 m 3 /day of SWRO water for 50 tons/day H2 PEM plant. 2 The breakdown of the daily energy required by a coupled SWRO-PEM process is shown in Figure 5(a), highlighting the low energy (0.1% of total energy) required by SWRO. This is a direct result of the energy intensive water electrolysis process, with ~55.44 kWh energy (including BoP) needed to electrolyze 10 kg water versus only 0.03 kWh to desalinate same amount of water. 25,26,34 A breakdown of the CAPEX associated with building a SWRO-PEM plant is shown in Figure 5(b). The most comprehensive cost analysis on PEM electrolyzer systems was recently published by the U.S. S1. 34 For a 50 tons/day H2 plant, the total uninstalled capital costs are ~ 460 $/kW, with approx. 26% costs associated with BoP. At the same time, capital costs for a SWRO plant are dependent on technology, location, environmental regulations and most importantly the plant size. 27 For example, a medium size 10 million gallons per day (37,800 m 3 /day) SWRO plant would cost ~ 80 million $, whereas a smaller plant with 0.5 million gallons per day (1890 m 3 /day) capacity would cost ~7 million $. 40 For our case study, we estimated a direct capital cost of ~1.86 million $ for 500 m 3 /day SWRO plant. In contrary to the broad-brushed argument by many in literature, this analysis reveals that the CAPEX of the SWRO plant contributed only ~3% of total direct CAPEX required for the coupled process (Figure 5(b)). A breakdown of the OPEX for the coupled SWRO-PEM process is shown in Figure 5(c). The OPEX of PEM systems are dominated by electricity costs due to the energy intensive electrolysis process with other contributions from O&M and stack replacement costs. On the other hand, the typical OPEX for SWRO plants comprises power consumption, membrane replacement, waste stream disposal, chemicals, labor, and O&M cost. 33,41 Assuming an electricity cost of 0.05 $/kWh, the OPEX of SWRO plant represent a small fraction (~0.2%) of the total OPEX for coupled process and is dominated by electricity costs to run the PEM electrolyzer (~95%). To this end, we calculated the levelized cost of H2, which is ~3.81 $/kg without considering SWRO, which marginally increases to ~3.83 $/kg when cost of SWRO water is accounted for (Figure 5(d)). The analysis reveals that the use of SWRO water does not add any significant cost to the H2 produced, due to the low energy, CAPEX and OPEX for SWRO as compared to PEM electrolysis. We further analyzed the carbon footprint of a SWRO facility coupled with PEM electrolysis for H2 production. We ignored any emissions associated with construction and decommissioning as these contributions are minimal when compared to the operating phase of the plant. 42 We first calculated the CO2 emissions to produce a kilogram of H2 from the SWRO-PEM electrolysis process using the average emission intensities of various energy sources, as shown in Figure 6(a). 43 One obvious observation is that H2 produced via water electrolysis with purely fossil fuel (coal, oil, natural gas) based electricity would end up producing more CO2 than that of presentday SMR process (8-12 kg of CO2/kg H2). 44 Secondly and more importantly the contribution of SWRO to CO2 emissions is negligibly small when compared to PEM water electrolysis, irrespective of the energy source (Inset of Figure 6(a)). ## Department of Energy (DOE) with the parameters summarized in Table We also calculated CO2 emissions for the more practical scenario where the SWRO-PEM plant gets its required power from the electricity grid (Figure 3). Figure 6 (b) shows the CO2 emissions from a coupled SWRO-PEM process based on average emission intensity from electricity generation in different jurisdictions. 45 The analysis indicates the SWRO-PEM process for large scale H2 production is environmentally compelling only in countries with a carbon intensity of electricity < 0.18 kg CO2e/kWh. Today, such low carbon footprint from electricity generation is only possible in countries having significant fraction of their electrical energy from renewables, such as Canada, Sweden, and Iceland. 46 In countries like China and the United states which are currently the biggest CO2 emitters in the world, such low carbon intensity would be an ambitious target to achieve in the next couple of decades unless there is a major shift in energy policies and production methods. For truly green H2 production, one could consider the example of Iceland where 100% renewable electricity on grid emits only ~0.48 kg-CO2e/kg H2. 47 ## Conclusions and Outlook In summary, our analysis shows there are limited economic and environmental incentives for pursuing research and development on today's nascent direct seawater splitting technology as opposed to simply coupling industrially mature SWRO with water electrolysis routes for sustainable H2 production in the foreseeable future. With fast growing multiple challenges in energy, water, environment, food, and health affecting modern society, we will likely be better off prioritizing R&D investment in technologies that have the greatest chance for widespread deployment in near future, including coupled SWRO and PEM systems. This seems to us a more practical and immediately deployable route than large-scale investments in developing catalysts and systems for direct electrolysis of seawater with all its attendant uncertainties. Despite fast development with great promise for future, PEM electrolysis routes to hydrogen production remains expensive for widespread roll out. Therefore, further investment in R&D efforts from academia and industry for developing low cost and energy efficiency electrocatalysts is vital for future market growth. With 1.2 billion people around the globe living in areas of physical water scarcity, there are opportunities to further develop energy efficient and economically compelling desalination technology. Worldwide, desalination is considered an immediate solution to the problem of water scarcity and quality that will worsen with continued population growth and more prolonged droughts linked to climate change. 48 Using desalinated water for electrolysis has an added advantage of being able to treat water from a wide variety of sources, such as brackish groundwater, surface water, seawater, and domestic and industrial wastewater. 28 To make it more affordable and accessible, research efforts should be directed towards improving desalination processes, devising more effective and durable membranes, for example, to produce more water per unit of energy. 48 There are environmental issues to tackle as well such as the disposal or processing of the concentrated brine, which in addition to being extremely salty also contains treatment chemicals. 51 In excessive concentrations, they have the potential to negatively affect the marine environment. There are efforts to eliminate wastewater discharge via zero liquid discharge (ZLD) approaches and exploring the potential of high-pressure reverse osmosis (HPRO), among other technologies, to efficiently desalinate hypersaline brines. 49 Furthermore, in some settings, these brines may be considered as a resource for high value minerals and energy recovery instead of being a waste with discharge constraints. Therefore, we pose these questions: Should we consider realigning our R&D priorities? Is direct seawater splitting a solution looking for a problem that has already been solved?

chemsum

{"title": "Seawater Electrolysis for Hydrogen Production: A Solution Looking for a Problem?", "journal": "ChemRxiv"}

stereoregular_functionalized_polysaccharides_<i>via</i>_cationic_ring-opening_polymerization_of_biom

## Abstract: We report the facile synthesis and characterization of 1,6-a linked functional stereoregular polysaccharides from biomass-derived levoglucosan via cationic ring-opening polymerization (cROP). Levoglucosan is a bicyclic acetal with rich hydroxyl functionality, which can be synthetically modified to install a variety of pendant groups for tailored properties. We have employed biocompatible and recyclable metal triflate catalystsscandium and bismuth triflatefor green cROP of levoglucosan derivatives, even at very low catalyst loadings of 0.5 mol%. Combined experimental and computational studies provided key kinetic, thermodynamic, and mechanistic insights into the cROP of these derivatives with metal triflates.Computational studies reveal that ring-opening of levoglucosan derivatives is preferred at the 1,6 anhydro linkage and cROP proceeds in a regio-and stereo-specific manner to form 1,6-a glycosidic linkages. DFT calculations also show that biocompatible metal triflates efficiently coordinate with levoglucosan derivatives as compared to the highly toxic PF 5 used previously. Post-polymerization modification of levoglucosan-based polysaccharides is readily performed via UV-initiated thiol-ene click reactions. The reported levoglucosan based polymers exhibit good thermal stability (T d > 250 C) and a wide glass transition temperature (T g ) window (<À150 C to 32 C) that is accessible with thioglycerol and lauryl mercaptan pendant groups. This work demonstrates the utility of levoglucosan as a renewably-derived scaffold, enabling facile access to tailored polysaccharides that could be important in many applications ranging from sustainable materials to biologically active polymers. ## Introduction Polymers are indispensable in modern life and the global economy; however, more than 90% of current synthetic polymers are derived from fossil fuels and 87% of these produced polymers contribute to plastic waste, usually after a single use. Lignocellulosic biomass is one of the most promising renewable feedstocks for sustainable polymers due to its worldwide abundance and availability. The major component of lignocellulosic biomass is cellulose, and thermochemical conversion processes-such as fast pyrolysis-provide an efficient route to convert biomass cellulose into renewable chemicals. The major product of cellulose fast pyrolysis is levoglucosan (Fig. 1A), which can be obtained with yields as high as 80%. Not surprisingly, levoglucosan has been identifed as a building block for high value-added chemicals, pharmaceuticals, and surfactants. As a result, there has been increasing interest in the commercial production of levoglucosan from biomass. Techno-economic analyses indicate that levoglucosan could be produced by thermochemical conversion at a low cost of $1.33 per kg to $3.0 per kg, which is competitive with many petroleum-derived monomers. 12, Structurally, levoglucosan is an attractive feedstock for the synthesis of sustainable polymers. This anhydrosugar contains a bicyclic acetal linkage (Fig. 1A, highlighted in orange) that is amenable to cationic ring-opening polymerization (cROP). 18 Levoglucosan also offers rich functionality due to its three hydroxyl groups, which can be synthetically modifed prior to cROP to install a variety of pendant groups for tailored properties. Furthermore, the incorporation of the rigid glucopyranose ring in the polymer backbone via cROP can increase the thermal stability of levoglucosan based polysaccharides. 19 Stereoregular polysaccharides are also of interest in biological and biomedical applications such as drug delivery and blood anti-coagulation due to their biocompatibility, multiple functionalities, and stereochemistry dependent properties. In addition, stereoselective cROP of levoglucosan to obtain polysaccharides with 1,6-a-glycosidic linkages is attractive from a sustainability standpoint since these linkages have demonstrated susceptibility to enzymatic and acidic hydrolysis. Despite these advantages, levoglucosan has received considerably less attention in the sustainable polymers feld as compared to levoglucosenone, another anhydrosugar than can be obtained in small quantities from cellulose pyrolysis. 13 The cROP of protected levoglucosan derivatives was frst reported more than 50 years ago, however the synthetic routes employed had some major drawbacks. 38 These derivatives are generally synthesized from levoglucosan by reaction with alkyl/ aryl bromide in the presence of sodium hydride in dimethylformamide, which is hazardous due to its thermal instability. Benzyl and allyl derivatives of levoglucosan have been two of the most widely studied monomers with cROP proceeding at low temperatures in the presence of PF 5 or TMSOTf. The resulting polymers were de-benzylated/deallylated for use as polysaccharide mimetics and HPLC stationary phases, respectively. 38,44,47 These polymerization conditions can afford high molecular weight polymers (M n of 50 kDa), however reaction times were generally long (80 h) and initiators (PF 5 & TMSOTf) that are both highly toxic and difficultto-handle were used. 38,48 Additionally, in these previous studies stereoselective cROP of levoglucosan required rigorous conditions such as high vacuum and very low temperature (60 C for the tribenzyl monomer, 0 C for the triallyl monomer). 38,41,44,45 Overall, identifcation of alternative low toxicity catalysts for cROP of levoglucosan under mild conditions has remained a challenge, thereby hindering their large scale and sustainable synthesis. This has potentially limited the use of levoglucosan in the sustainable polymers feld as compared to other biomassderived sugars such as levoglucosenone. Despite the initial investigations in cROP of levoglucosan, the material properties of the resulting polysaccharides have not been characterized and computational insight into the thermodynamics and mechanism of levoglucosan cROP has not been provided. Additionally, very limited work has been done to utilize the rich hydroxyl functionality of levoglucosan for the synthesis of functional polysaccharides. Fu et al. synthesized amphiphilic polysaccharides by ring-opening copolymerization of tripropargyl levoglucosan, followed by azide-alkyne cycloaddition modifcation of the copolymer. 49 However to the best of our knowledge this is the only example of reactive levoglucosanbased materials, providing ample space for the development and characterization of functional levoglucosan polymers. Taken together, these factors have led us to explore methods to improve levoglucosan polymerization conditions while preserving reactive functionalities for polymer modifcation. Herein, we report a synthetic platform for stereoregular 1,6a linked levoglucosan-based polysaccharides with different pendant functional groups. An array of catalysts was screened for the cROP of levoglucosan-based benzyl (Bn) 2 and allyl (All) 3 functional monomers (Fig. 1A) to identify green alternatives. Scandium and bismuth triflate-which are biocompatible and recyclable catalysts-were identifed as promising candidates and used for further studies to understand the kinetics, thermodynamics, and mechanism of ring-opening for 2 and 3. Postpolymerization modifcation of poly(3) was performed with UVmediated thiol-ene click chemistry to afford functional polymers with a range of solubility and thermal properties. This work paves the way for future development of this renewable platform, enabling facile access to tailored polysaccharides for sustainable polymers to biomaterials applications. ## Monomer synthesis Monomer 2 was purchased from a commercial supplier and monomer 3 was synthesized on a multi-gram scale using a modifed method for the etherifcation of starch (75% yield, ESI Sections 2.1 & 9.1 †). 50 This synthetic route is safe with easy to handle reagents, employing NaOH as a mild base in dimethyl sulfoxide. We also explored a "greener" synthetic pathway for 3 using NaOH, a phase-transfer catalyst, and water as the reaction medium; 51 the monomer was successfully prepared with this method, albeit in lower yield (36% yield; ESI Section 2.2 †). ## Screening of cROP conditions Combined molecular mechanics and density functional theory (DFT) calculations were performed to determine the ring-strain free energies for isodesmic reaction of 2 and 3 with dimethyl ether. Unsurprisingly, both the monomers have comparable ring-strain values, and for both 2 and 3 ring-opening is energetically favored at the 1,6-anhydro linkage over the 1,5-linkage (Fig. 1B), supporting polymerization of 2 and 3. To identify less toxic catalysts to promote cROP of 2 and 3, and to understand the effect of solvent and catalyst loading on cROP of 2 and 3, a library of polymerization experiments was performed. We initially screened a range of metal and organic catalysts due to their commercial availability, low toxicity, and ability to ring open cyclic ethers for the polymerization of 2 (Fig. S3 and Table S1 †). 48, Cationic initiators BF 3 OEt 2 and MeOTf that are conventionally employed for cROP of cyclic acetals were also screened in addition to the array of catalysts for direct comparison. 57,58 All screening reactions took place at room temperature for 72 h in dichloromethane (initial monomer concentration [M] 0 ¼ 1.0 mol L 1 ). Successful catalysts were then used to screen cROP of 3; two metal triflates [Sc(OTf) 3 and Bi(OTf) 3 ] were identifed for cROP of both 2 and 3 (Table 1, entries 1, 2, 7, 8). Both metal triflates provided comparable conversion to the BF 3 OEt 2 and MeOTf controls (Table S2 †) and provide the additional beneft of being recyclable (via a simple aqueous extraction) while not releasing corrosive byproducts such as triflic acid commonly released by alkyl triflates. 48 However, these metal triflates have limited solubility in dichloromethane and hence further studies were performed to understand solvent effects. We investigated the addition of varying quantities of acetonitrile (MeCN), which is a better solvent for the catalyst, in the polymerization mixture. To our surprise, no polymerization occurred when >10% MeCN by volume was present in the reaction medium. Moreover, monomer conversion generally decreased with increasing amount of MeCN from 0-10% (Tables S3 and S4 †) and we hypothesize this is due to the better solvation of the ionic active species by MeCN, as reported in the literature. 59,60 We then studied the effect of varying catalyst loading on cROP of 2 and 3 and found that polymerization of both monomers could be conducted at M(OTf) 3 loadings as low as 0.5 mol% (Table 1, entries 4, 6, 10). Notably, the pendant groups do not seem to drastically impact conversion, as similar values were observed for both the monomers likely due to their comparable ring-strain (Fig. 1B). With the identifed metal tri-flates, M w values up to 18.6 kDa for poly(2) (DP ¼ 31) and 12.5 kDa for poly(3) (DP ¼ 24) could be achieved with moderate dispersities (Table 1, entries 5 and 12). These moderate molecular weights are most likely caused by intermolecular chain transfer and back-biting reactions which are common features in cROP of cyclic acetals. 57 Chain transfer reactions are favored in the cROP of cyclic acetals due to the higher basicity of oxygen atoms in the polymer chain as compared to the monomer. 57 Moreover, it is hypothesized that metal triflate mediated cROP of levoglucosan derivatives follows a catalytic approach as reported in the literature. 54 In a catalytic approach, one metal triflate molecule catalytically produces a large number of polymer molecules, thereby leading to shorter polymer chains. 54 Since the metal triflate mediated cROP of levoglucosan follows a catalytic approach, calculating a desired or target molecular weight for a given set of polymerization conditions is difficult. ## Thermodynamics of polymerization DFT calculations were employed to provide insights into the mechanism of levoglucosan cROP catalyzed by Bi(OTf) 3 , Sc(OTf) 3 , and PF 5 (as a control for comparison to previous studies), as well as to understand catalyst efficiency. Various mechanistic pathways were considered (higher energy pathways are provided in the ESI Section 11 †), and the lower energy pathway is depicted in Fig. 2A. First, monomer 2 or 3 is activated by the catalyst to form a thermodynamically stable complex (I1). When PF 5 is utilized as the catalyst, activation of 2 or 3 is unfavorable relative to the metal triflates due to poor coordination of PF 5 . This suggests that PF 5 exhibits less efficient initiation, leading to the high M n observed in previous studies. 44,46 After the frst step, I1 undergoes ring opening of 1,6anhydro linkage to generate the carbenium intermediate I2 through a TS structure TS1, and the activation free energies are similar for monomer 2 or 3 with their respective catalysts, Bi(OTf) 3 and Sc(OTf) 3 (Fig. 2A, B and S45 †). Next, the nucleophilic addition of another molecule of monomer to the electrophilic carbon of intermediate I2 generates intermediate I3 via a pseudoaxial approach through TS2. In this step, the DG ‡ values for nucleophilic addition of 2 and 3 are similar in all cases (Fig. 2B). Interestingly, it is evident that computed energetics of the nucleophilic addition of monomer to I2 to form the alternative intermediate structure I4 (Fig. 2A) indicate it as energetically disfavored (possibly due to the steric hindrance between the incoming monomer and catalyst), thereby leading to stereospecifc cROP of 2 and 3 (Fig. S46, ESI Section 11.3 †). Further, Fig. 2C shows that the monomer addition at C6 of 2 and 3 is energetically disfavored compared to the anomeric carbon. This indicates that cROP proceeds in a regiospecifc manner with attack on the anomeric carbon to form 1,6-glycosidic linkages. Finally, computational mechanistic insights also provide some understanding of the moderate molecular weights observed experimentally. Specifcally, higher energy barrier pathways may contribute to slower propagation leading to shorter polymer chains (Fig. S48 †). Finally, the increase in DG ‡ and the increase in charge separation between the active ionic species with increased chain length may hinder propagation to high molecular weight polymer. Polymer stereoregularity 1 H NMR analysis of poly(2) and poly(3) compared to the respective monomers indicates that the polymers possess 1,6-a glycosidic stereoregularity as the anomeric proton resonances (in the b-confguration) appeared at 4.94 ppm (Fig. 3B and D). 41 Additionally 13 C NMR of poly(2) and poly(3) also showed the a confguration as the anomeric carbon resonances appeared at 97.8 ppm and 97.4 ppm, respectively (Fig. S35 and S37 †). 41,49 The stereoregularity of the polymers was also confrmed via optical rotation measurements. The optical rotation values for poly(2) and poly(3) were measured to be +91.5 cm 3 dm 1 g 1 and +84 cm 3 dm 1 g 1 respectively, with the optical rotation values for 2 and 3 being 31.5 cm 3 dm 1 g 1 and 45 cm 3 dm 1 g 1 , respectively. Overall, these results indicate that the cROP of 2 and 3 with metal triflates yields highly stereoregular levoglucosan polymers with 1,6-a glycosidic linkages. Notably, this work demonstrates that highly stereoregular levoglucosan polymers can be synthesized under mild conditions without the need for an energy intensive process involving high vacuum and low temperatures, as reported previously. 1 H NMR analysis also indicates that the alkene functionality in poly( 3) is intact during cROP with the vinyl proton resonances at 5.91 ppm, 5.27 ppm and 5.14 ppm. The successful synthesis of poly(3) while preserving multiple pendant allyl groups will further allow for click-chemistry modifcations. ## Kinetics of polymerization Apart from understanding the ring-opening mechanism, we were also interested in investigating the cROP kinetics for each monomer to determine optimum reaction time (ESI Section 5 † for details). We observed a short induction period of $20 min during cROP of 2 (Table S7 and Fig. S5 †), which reaches an equilibrium conversion of 64% in #24 h (Fig. 4A). Conversely, cROP of 3 has a much longer induction period of $4 h (Table S8 †) and reaches an equilibrium conversion of 83% in #72 h (Fig. 4B). Remarkably, the induction period for cROP of 3, as evidenced by 1 H NMR spectroscopy, was accompanied with a drastic color change in the solution throughout the 4 h period (Fig. S10 and S11 †). We hypothesize that this long cROP induction period of 3 is due to the non-productive coordination of Bi(OTf) 3 with the allylic ether oxygens and the glucopyranose ring oxygen. This was supported by DFT energetics calculations depicting that non-productive coordination of Bi with the allylic ether oxygens and the glucopyranose ring oxygen is favored at multiple locations (Fig. 4C). Furthermore, while comparing 2 and 3, we fnd that the relative free energies of binding the M(OTf) 3 to each of these respective oxygen atoms vary slightly (Fig. 4C and S44 †). However, the free energy trends and magnitudes are similar across both 2 and 3. Along with tracking conversion over time, the molar mass and dispersity of the growing polymer chains was also monitored throughout this kinetic study. For poly(2), the M n increased up to 40% conversion (Fig. S18 †), and for poly(3) the M n increased up to 60% conversion (Fig. S19 †). Lastly, narrow dispersities were maintained for poly(2) throughout the reaction duration ($1.2 to 1.4, Fig. S20 †), whereas for poly(3) dispersity was higher in the initial stages and gradually decreased to a stable value ($1.5, Fig. S21 †). ## Thiol-ene post polymerization modication The allylic pendant groups enable facile post-polymerization modifcation of poly(3), which we envisioned could serve as a stereo-and regio-regular scaffold for rapid, UV-initiated thiol-ene click reactions for further tailoring polymer properties. 61 Such a renewably-derived and stereoregular scaffold is attractive for many applications such as sustainable polymers and biologically active polymers. Thioglycerol and lauryl mercaptan were chosen as model thiols as they are structurally similar to renewable glycerol and lauryl alcohol. 7,62,63 Additionally, the contrast in hydrophilicity/ hydrophobicity of these thiols was expected to provide starkly different properties after modifcation (Fig. 5A). Realtime Fourier-transform infrared (RT-FTIR) spectroscopy was used to study the kinetics of thiol-ene reactions with poly(3). showed an increase in M n as compared to poly(3), but some of this increase may be due to fractionation as a result of puri-fcation steps (Table S9 †). The solubility of synthesized polymers was tested in a range of solvents (Table S10 †). As expected, poly(4) and poly(5) display starkly different solubility properties, with poly(4) being the only water-soluble polymer in the synthesized library. ## Thermal properties Lastly, the thermal properties of the synthesized polysaccharides were examined via thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC) to understand thermal stability and thermal transitions respectively. Dextran, a natural and commercial polysaccharide, was used as a control because it also contains 1,6-a-glycosidic linkages and glucopyranose rings. Except for poly(3), each of the homopolymers demonstrated excellent thermal stability with a T d,10% > 300 C (Fig. 5E), consistent with the thermal stability of dextran and trimethylated levoglucosan polymer (T d,10% ¼ 347 C). 19 Comparatively, poly(3) had a slightly lower T d,10% (250 C), which could be attributed to the fragmentation of pendant allylic ether groups at this temperature, consistent with literature. 64 In general, levoglucosan-based polysaccharides possess excellent thermal stability even at moderate molecular weights, likely due to the rigid glucopyranose ring in the backbone. The glass transition temperature (T g ) of poly(2), poly(3), and poly(4) was observed to be 32 C, 5 C, and 14 C, respectively (Fig. S24-S26 †); no T g was observed for poly(5) down to 150 C, which is consistent with other lauryl-pendant polymers. 62,65 The T g values of poly(2), poly(3), and poly( 5) are predictably lower than that of the control dextran (199 C), potentially due to the strong hydrogen bonding between unsubstituted dextran chains. Surprisingly, poly(4) exhibited a sub-zero T g despite the presence of pendant hydroxyl groups, potentially due to the added flexibility of the aliphatic methylene units. Additionally, the T g for poly(2) is lower than that of trimethylated levoglucosan polymer (T g $300 C). 19 It is hypothesized that this result can be explained based on the steric bulk of pendant groups. As compared to methyl groups, benzyl groups possess higher steric bulk, leading to higher free volume between adjacent polymer chains. A higher free volume will allow polymer chains to slide past each other more easily, thereby resulting in a lower T g . Overall, DSC analysis demonstrates that the T g of levoglucosanbased polysaccharides can be tailored based on the identity of the pendant groups, with a remarkable T g window of >180 C accessible with the few pendant groups evaluated in this study. DSC analysis also revealed an interesting double melting peak for poly(5) as shown in Fig. 5F (T The crystallinity of poly( 5) is most likely due to lauryl side chain crystallization, also observed in other polymeric systems with lauryl side chains. 62,65 However, the double melting phenomenon could be caused by multiple factors. One explanation is the lamellar thickness model, which attributes the double melting behavior to the presence of lamellae with two different thicknesses. Other potential explanations include the crystallization of poly(5) chains with varying degrees of side chain functionalization and some extent of polysaccharide backbone crystallization. It is also worth highlighting that even at moderate molecular weights, poly(3) offers very high functionality for modifcation (for M n ¼ 6.7 kDa the number of allyl groups ¼ 72), enabling the facile synthesis of highly functional materials. The high thermal stability, wide range of accessible T g values, and the potential crystallization phenomena of these derivatives demonstrate that poly(3) is an excellent renewable scaffold for post-polymerization reactions to tailor properties to desired applications in a variety of applications. Furthermore, the levoglucosan platform provides easy access to fully functionalized dextran derivatives with the ability to install the desired pendant group both pre-and post-polymerization. ## Conclusions In conclusion, we have synthesized functional stereoregular 1,6a linked polysaccharides with tunable thermal properties from cellulose-derived levoglucosan via cROP and postpolymerization thiol-ene click reactions. Through systematic screening experiments, we have identifed green and recyclable metal triflate catalysts for cROP of levoglucosan derivatives under mild conditions. We have also provided kinetic, thermodynamic, and mechanistic insights into cROP of these derivatives. Computational studies reveal that initial ringopening of levoglucosan derivatives is energetically favored at the 1,6 anhydro linkage and subsequently nucleophilic addition of monomer to the carbenium intermediate leads to 1,6a glycosidic linkages in a regio-and stereo-specifc manner. The allyl-functional polymer has been identifed as an excellent scaffold for synthesis of highly functional materials via rapid, UV-meditated thiol-ene modifcations. Levoglucosan-based polysaccharides generally demonstrate excellent thermal stability and a vast T g window of >180 C is accessible with the few pendant groups highlighted in this study. We believe that levoglucosan exhibits great potential as a renewable feedstock for the development of next generation sustainable and biocompatible polymers with tailored properties.

chemsum

{"title": "Stereoregular functionalized polysaccharides <i>via</i> cationic ring-opening polymerization of biomass-derived levoglucosan", "journal": "Royal Society of Chemistry (RSC)"}

exploring_a_new_ligand_binding_site_of_g_protein-coupled_receptors

## Abstract: Identifying a target ligand binding site is an important step for structure-based rational drug design as shown here for G protein-coupled receptors (GPCRs), which are among the most popular drug targets. We applied long-time scale molecular dynamics simulations, coupled with mutagenesis studies, to two prototypical GPCRs, the M3 and M4 muscarinic acetylcholine receptors. Our results indicate that unlike synthetic antagonists, which bind to the classic orthosteric site, the endogenous agonist acetylcholine is able to diffuse into a much deeper binding pocket. We also discovered that the most recently resolved crystal structure of the LTB4 receptor comprised a bound inverse agonist, which extended its benzamidine moiety to the same binding pocket discovered in this work. Analysis on all resolved GPCR crystal structures indicated that this new pocket could exist in most receptors. Our findings provide new opportunities for GPCR drug discovery. ## Introduction Many essential cellular processes, including cell regulation, signal transduction, and the immune response, are mediated by specifc protein-ligand interactions. The identifcation and characterization of specifc protein binding sites for particular ligands are crucial for the understanding of the functions of both endogenous ligands and synthetic drug molecules. Thus, the detection and characterization of ligand binding sites are important steps towards protein function identifcation and drug discovery. A newly revealed ligand binding site will provide a new opportunity for a drug target, to design new classes of compounds based on new chemical environments. 8 Besides traditional biochemical methods such as systematic mutagenesis experiments, structural biology or NMR, 9 computational methods offer alternative powerful and efficient approaches for ligand binding site detection. 7,10 In this work, we applied molecular dynamics simulations and computer modeling, combined with functional mutagenesis experiments, to explore new ligand binding sites of GPCRs. GPCRs typically detect extracellular signals (photons, odorants, hormones, and neurotransmitters) 11,12 on the cell surface and undergo multiple conformational changes that enable the binding and activation of intracellular proteins (e.g. G-proteins, arrestins and kinases). More than a third of modern therapeutic compounds are estimated to target GPCRs. 7 Thus, understanding the ligand binding process and detecting a new ligand binding site for GPCRs in molecular details are of great importance in revealing GPCR-mediated signaling and improving GPCR-targeted drug design. Recent crystal structures of GPCRs show that different small molecules can bind to different regions of receptors including (1) the traditional orthosteric ligand binding site in the vicinity of the highly conserved (W 6.48 ), 17 (2) the allosteric ligand binding site next to extracellular loop 2 (ECL2), 18,19 (3) allosteric ligand binding sites between transmembrane (TM) helices TM2 and TM3, 18,19 (4) allosteric ligand binding sites between TM3 and TM4, 20 (5) the intracellular G protein binding region 21 and (6) the outer surface of the receptor in the middle of the TM area 22 (Fig. S1 †). Computational methods have become important tools for understanding the structural and dynamical function of GPCRs. 7,23 Molecular dynamics (MD) simulations and computational modeling can be used to address specifc questions about the dynamic properties of the modelled system, which are difficult to illuminate in experiments. For example, with MD simulations, it is possible to sample the process of ligand binding to a specifc GPCR. Many atomic details, including molecular switches, binding site expansions or domain movements, can be efficiently revealed through MD simulations. Here we explore a new ligand binding site for two GPCRs, muscarinic M3 and M4 receptors, which specifcally activate Gaq and Gai proteins, respectively (Table 1). 31,32 Crystal structures have been solved for M3 and M4 receptors containing antagonists in orthosteric binding sites, whereas no agonistbound structures are available. A new binding site for muscarinic acetylcholine receptors could provide new opportunities for the understanding of ligand binding and activation processes. We therefore simulated the entire binding processes of the endogenous agonist acetylcholine (ACh) to both M3 and M4 receptors. Surprisingly, the resulting position of ACh bound to the particular receptors revealed an additional new binding site next to the highly conserved Asp (D 2.50 ). 33 We also observed that ligand binding leads to the expansion of the binding site. We further inspected over 200 ligand bound GPCR crystal structures and discovered that most ligands were located in traditional orthosteric sites. However, the recently resolved crystal structure of the leukotriene B4 (LTB4) receptor comprised a bound bitopic ligand which extended from this typical orthosteric site to the new ligand binding site found in the present work. This observation strongly suggests that our newly discovered binding site might also exist in other GPCRs. Among these >200 structures, we systematically inspected 39 unique family A GPCRs and found that most of them indeed contain this potential binding site for binding small ligands or extending larger ones from the main orthosteric site. Our fndings provide a new opportunity for GPCR drug discovery. ## Biological testing Split luciferase biosensor cAMP assay for measuring activation of Gi protein. Promega's split luciferase based GloSensor cAMP biosensor technology was used to determine GPCR mediated cAMP production. M4 receptor DNA (4 mg) and Glo-Sensor cAMP DNA (4 mg, Promega) were co-transfected into HEK293 T cells using Lipofectamine 2000 (Life technologies). After at least 24 h, the cells were seeded in 384 well white clear bottom plates (Greiner) with DMEM (Life technologies) supplemented with 1% dialyzed FBS at a density of 15-20 000 cells in 20 ml medium per well and incubated at 37 C in 5% CO 2 for at least 6 h before analysis. Wells were loaded for 20 min at 37 C with 20 ml of 2 mg ml 1 Luciferin D sodium salt prepared in HBSS at pH 7.4. All the following steps were carried out at room temperature. To measure agonist activity of the M4 receptor, 10 ml of acetylcholine solutions ranging from 0 to 30 000 mM was added to the cells 15 min before addition of 10 ml of isoproterenol at a fnal concentration of 200 nM, followed by counting of the plate for chemiluminescence using EnVision (Perkin Elmer) after 15 min. Chemiluminescence intensity was plotted as a function of ACh concentration and normalized to percent acetylcholine with 100% for E max acetylcholine cAMP inhibition and 0% for the isoproterenol-stimulated cAMP baseline. Data were analyzed using log (ACh) vs. response in GraphPad Prism. Ca 2+ Mobilization assay for measuring activation of Gq protein. To measure ACh-induced G protein coupling to the M3 receptor and the subsequent increase of intracellular calcium ion concentration, HEK293T cells were seeded in a 10 cm dish and incubated overnight. After 24 h, the cells were transfected with 4 mg M3 receptor plasmid and 20 ml Lipofectamine 2000 (Life technologies). 6 h later, the cells were seeded in 384 well plates at a density of 15 000 cells per well in DMEM containing 1% dialyzed FBS and incubated overnight. On the assay day, the cells were incubated (20 ml per well) for 1 h at 37 C with Fluo-4 Direct dye (Invitrogen) reconstituted in FLIPR buffer (1 HBSS, 2.5 mM probenecid, and 20 mM HEPES, pH 7.4). After the dye was loaded, the cells were placed in a FLIPR TETRA fluorescence imaging plate reader (Molecular Devices); acetylcholine dilutions were prepared at 3 fnal concentration in FLIPR buffer and aliquoted into 384 well plates. The fluidics module and plate reader of the FLIPR TETRA were programmed to read baseline fluorescence for 10 s (1 read per s), and then 10 ml of drug/ well was added to read for 6 min (1 read per s). Fluorescence in each well was normalized to the average of the frst 10 reads (i.e., baseline fluorescence). The maximal fluorescence intensity increase was measured 60 s after acetylcholine addition. Data were analyzed using GraphPad Prism. ## Loop modelling and structural preparations Loop flling and refnements. The resolved crystal structures of M3 and M4 receptors comprised the engineered receptors and inserted proteins in the intracellular loop ICL2 to facilitate crystallisation. Before starting MD simulations, we removed the corresponding inserted proteins from the M3 and M4 crystal structures and used the loop refnement protocol in Modeller 34 V9.10 to reconstruct and refne the ICL2 region. A total of 20 000 loops were generated for each receptor, and the conformation with the lowest DOPE (Discrete Optimized Protein Energy) score was chosen for receptor construction. Repaired models were submitted to Rosetta V3.4 for loop refnement with kinematic loop modeling methods. 35 Kinematic closure (KIC) is an analytic calculation inspired by robotics techniques for rapidly determining possible conformations of linked objects subject to constraints. In the Rosetta KIC implementation, 2N -6 backbone torsions of an N-residue peptide segment (called non-pivot torsions) were set to values randomly drawn from the Ramachandran space of each residue type, and the remaining 6 phi/ psi torsions (called pivot torsions) were solved analytically by KIC. Protein structure preparations. All protein models were prepared using Schrodinger suite software under the OPLS_2005 force feld. 36 Hydrogen atoms were added to the repaired crystal structures at physiological pH (7.4) with the PROPKA 37 tool to optimize the hydrogen bond network provided by the Protein preparation tool in Schrodinger. Constrained energy minimizations were carried out on the full-atomic models, allowing the maxium RMSD for heavy atoms of 0.4 . Ligand structure preparations. All ligand structures were obtained from the PubChem 38 online database. The LigPrep module in Schrodinger 2015 suite software was introduced for geometric optimization by using the OPLS_2005 force feld. The ionization state of ligands was calculated with the Epik 39 tool employing Hammett and Taft methods in combination with ionization and tautomerization tools. 39 Molecular simulations Molecular dynamics simulations. Membrane systems were built with the membrane building tool g_membed 40 in Gromacs with the receptor crystal structure pre-aligned in the OPM (Orientations of Proteins in Membranes) database. 41 Preequilibrated 120 POPC lipids coupled with 92 000 TIP3P water molecules in a box of $ 68 68 96 were used to build the protein/membrane/water system. We modeled the protein, lipids, water and ions using the CHARMM36 force feld. 42,43 The ionization states of both protein and the ligand were assigned properly according to the results from Schrodinger software. Ligands were assigned to the CHARMM CgenFF force feld. 44 The ligand geometry was submitted to the GAUSSIAN 09 program 45 for optimization at the Hartree-Fock 6-31G* level when generating force feld parameters. The system was gradually heated from 0 K to 310 K followed by a 1 ns initial equilibration at constant volume with the temperature set at 310 K. Both the ligand molecule and protein backbone were restrained by a force constant of 10 kcal mol 1 2 during this step. Next, an additional 40 ns restrained equilibration was performed at constant pressure and temperature (NPT ensemble; 310 K, 1 bar), and the force constant was tapped off gradually from 10 to 0 kcal mol 1 . The backbones of the proteins and the heavy atoms of the ligands were restrained during the equilibration steps. All bond lengths to hydrogen atoms were constrained with M-SHAKE. The van der Waals interactions were included using the switching function in the range of 10-12 . Long-range electrostatic interactions were computed using the Particle Mesh Ewald (PME) summation scheme. All MD simulations were done using Gromacs. 46 The simulation parameter fles were obtained from the CHARMM-GUI website. 47 The MD simulation results were analyzed using Gromacs 46 and VMD. 48 The solvent accessible surface area was calculated using Gromacs. Figures were prepared using PyMOL and Inkscape. 49 ## Metadynamics simulations Free-energy profles of the systems were calculated using welltempered metadynamics in Gromacs 46 V5.1.4 with Plumed 50 V2.2.1 patches. Metadynamics adds a history-dependent potential V(s, t) to accelerate sampling of the specifc collective variables (CVs) s (s 1 , s 2 , ., s m ). 51 V(s,t) is usually constructed as the sum of multiple Gaussians centered along the trajectory of the collective variables (eqn (1)). During the simulation, another Gaussian potential, whose location is dictated by the current values of the collective variables, is periodically added to V(s,t). 51 In our simulations, distances between the quaternary N atom of ACh and a side chain oxygen of D 2.50 were assigned as the collective variables s 1 , while the width of Gaussians, s, was set as 0.05. The time interval, s, was 0.09 ps. Well-tempered metadynamics involves adjusting the height, w j , in a manner that depended on V(s,t) where the initial height of Gaussians w was 0.05 kcal mol 1 , the simulation temperature was 310 K, and the sampling temperature DT was 298 K. The convergence of our simulations was judged by using the free energy difference between states A and B at 10 ns intervals. Once the results stopped changing over time, the simulation was considered as converged. 51 Each metadynamics simulation was performed for 100 ns, and the results were analyzed upon convergence. ## Analysis Interaction fngerprint calculations. The IFP was performed using PLIP software. 52 PLIP detects frequent non-covalent protein-ligand interactions including hydrogen bonds, hydrophobic contacts, pi-stacking, pi-cation interactions, salt bridges, water bridges and halogen bonds. 52 We used 200 frames from the fnal 20 ns MD simulations for the IFP analysis. A plot was generated for the combination of all kinds of interactions. Parameters used for IFP calculations were kept as default. New ligand binding site predictions and calculations. Inspections for the 39 unique crystal structures of the family A GPCRs were performed using ConCavity, 53 a widely used program for ligand site prediction. Both the grid size and resolutions were set to 0.1 . The volume of the newly discovered binding site for each receptor was calculated using the POVME 54 program. All the other analyses have been done using Gromacs and VMD. Figures were prepared using PyMOL and Inkscape. ## Results and discussion The entrance pathway of M3-ACh We frst present the analysis of eight 3 ms-long MD trajectories, each of which starts with either the antagonist tiotropium (TTP)-bound receptor or the ACh bound at the extracellular vestibule of the receptor (Fig. 1). The crystal structures of both inactive M3 (PDB: 4DAJ) 31 and M4 (PDB: 5DSG) 32 receptors were used directly for long time scale MD simulations. In order to probe the binding mode of ACh in the crystal structure of the M3 receptor, 32 we frst placed ACh at the entrance of the orthosteric binding pocket. Next, we submitted this M3-ACh structure to all-atom MD simulations. At the beginning of the simulations (Fig. 1a and b, Movie S1 †), ACh quickly enters (during 0.1-0.2 ms) the orthosteric site (zone II). Then, ACh is frst confned to an aromatic cage characterized by residues Y148 3.33 -Y506 6.51 -W503 6.48 -Y529 7.39 . The volume of this pocket increases substantially from the initial 86 AE 2 3 to 121 AE 3 3 at the end of MD simulations. We superimposed a typical conformation of ACh (at $0.1 ms snapshot) with the crystal structure of M3-TTP (4DAJ) 31 and found that the conformation of ACh is identical to that of TTP (Fig. S2 †). This supports the hypothesis that the binding site of ACh in the two mAChRs might be the same as observed in the corresponding orthosteric structures (zone II in Fig. 1). 32 However, during the 0.5-0.6 ms simulation period, ACh drifts into a much deeper new binding site (zone III) next to the highly conserved residue D113 2.50 (Fig. 1a and b). Consequently, the volume of zone III expands from the initial 57 AE 1 3 to 148 AE 2 3 at the end of the MD simulations. Interestingly, the side chain of the highly conserved W503 6.48 undergoes a corresponding flip, leaving space for the entry of the ACh molecule (Movie S1 and Fig. S3 †). Such molecular switches have been observed both computationally 55 and experimentally 56 for adenosine receptors. In contrast, in the antagonist bound M3-TTP system, the antagonist TTP remains stably bound in the orthosteric site (zone II) (Fig. S4 †) throughout the entire simulations. The volume of the orthosteric site (zone II) and that of the new binding site (zone III) is 162 AE 2 3 and 56 AE 1 3 , respectively, in this M3-TTP system. We also investigated the crystal structures of M2 receptors, bound to both an agonist (PDB: 4MQS) 57 and to an antagonist (PDB: 3UON). 58 The highly conserved W400 6.48 also underwent molecular switches (Fig. S5 †) comparable with that found for the M3 receptor in the present work. ## The binding mode of M3-ACh To validate the above-mentioned fndings, we frst employed a protein-ligand interaction fngerprint (IFP) analysis for M3-ACh (Fig. 2a) followed by mutagenesis experiments (Fig. 2c and S6 †) for M3 receptors. IFP was performed based on the snapshots from the fnal 20 ns MD simulations. IFP identifed that ACh frequently interacts with fve residues including A112 2.57 , I116 2.53 , A150 3.35 , S154 3.39 and W503 6.48 . Thus, we next mutated I116 2.53 , S154 3.39 and W503 6.48 individually into Ala residues. The signaling assays of the different M3 receptor mutants showed noticeably decreased activities (Table 2). Specifcally, the ACh induced activation of Gq by the wild-type (WT) M3 shows an EC 50 of 0.5 mM, whereas receptor mutants I116 2.53 A, S154 3.39 A and W503 6.48 A show an EC 50 of 1.8 mM, 1.7 mM and 4.0 mM, respectively. The entrance pathway of ACh into the M4 receptor Furthermore, we extended our investigations to M4 receptors and performed similar MD simulations as before for the M3 receptor. We found that ACh also diffuses into a much deeper binding site of the M4 receptor next to the highly conserved D78 2.50 , however at a different time scale. It frst spends a longer period (up to 0.25 ms) in the extracellular vestibule (zone I) before reaching the orthosteric site (zone II) at 0.3-0.4 ms (Movie S2, † Fig. 1b, c, f and g). After stabilizing for a considerable time in this region, ACh continues moving from the orthosteric site (zone II) to the new binding site (zone III) at 1.9-2.1 ms, where it remains frmly bound until the end of the MD simulations. The volumes of zone II and zone III in M4-ACh also change significantly during the MD simulations. The initial volume of zone II of 92 AE 1 3 expands to 140 AE 2 3 in M4-ACh at the end of MD simulations. The corresponding volume changes in zone III are from 46 AE 1 3 to 139 AE 2 3 . These results are in complete agreement with our previous studies that agonist binding leads to an expansion of the extracellular orthosteric regions. 19 The binding mode of M4-ACh As described for the M3 receptor, we also applied both IFP analysis and mutagenesis experiments to validate our fndings for M4-ACh. IFP indicated that there are more residues involved in M4-ACh than in M3-ACh interactions, including I81 2.53 , S116 3.36 , S119 3.39 , V120 3.40 , F409 6.44 , W413 6.48 and N445 7.45 . We then employed mutagenesis to all these residues individually which led to dramatic effects on the M4 activity in each case (Fig. 2d and Table 2). Specifcally, W413 6.48 A led to a completely inactive M4 mutant towards ACh; the activities of mutants I81 2.53 A and S116 3.36 A also decreased 100 times compared to that of the WT M4 receptor. In contrast, the mutations of F409 6.44 A, S119 3.39 A and N445 7.45 A show smaller effects with 4-14 fold decreased activities (Fig. S6 † and Table 2), while the mutation V120 3.40 A increased the EC 50 by 16 fold. ## The binding mode differences between M3-ACh and M4-ACh Moreover, we found that the fnal binding positions of ACh in M3 and M4 are different (Fig. 1b and 2a) at the end of MD simulations. In M3-ACh, ACh forms (i) hydrophobic interactions with A112 2.57 , I116 2.53 , A150 3.35 , S154 3.39 and W503 6.48 , and (ii) ionic interactions between its quaternary nitrogen and highly conserved residue D113 2.50 . The flexible acetyl tail sits in the void space between TM2 and TM3, next to A150 3.35 . In contrast, ACh flipped by 180 degrees in M4-ACh compared to the binding confguration in M3-ACh: the ester head group comes into contact with V120 3.40 , F409 6.44 , W413 6.48 and N445 7.45 , whereas the quaternary nitrogen interacts with I81 2.53 , V115 3.35 , S116 3.36 and S119 3.39 . Such variations probably stem from the differences in position 3.35: A150 3.35 in M3 and V115 3.35 in M4. All other residues in the ligand binding sites of M3 and M4 are exactly the same. The smaller A150 3.35 in M3 creates more space which can accommodate the ester group but it is still too small to ft the bigger quaternary nitrogen (Fig. 2a). When ACh approaches this region of M3, the ester group can be stabilized between TM2 and TM3. The V115 3.35 in M4 takes up more space than A150 3.35 in M3. The hydrophobic sidechain of V115 3.35 favors the hydrophobic interactions with -CH 3 groups from the quaternary nitrogen. The positively charged quaternary nitrogen could be further stabilized in this area by the highly conserved D 2.50 through ionic interactions (Fig. 2b) in both M3 and M4. ## The kinetic differences in binding M3-ACh and M4-ACh Noticeably, the time scales for ACh entering the new ligand sites of M3 (0.5 ms) and M4 (2.0 ms) are quite different. To understand these observations, we introduced well-tempered metadynamics MD. MD simulations are employed to explore the free energy profle of the entire binding processes of ACh for each receptor (Fig. S7 †). We defned the distance between the quaternary nitrogen of ACh and the carboxyl group of D 2.50 as the collective variable (CV). The results of our metadynamics MD simulations are in complete agreement with our unbiased long-time scale MD simulations, revealing three distinct energy states: the transient locations of ACh in the extracellular vestibule zone I, the orthosteric site zone II, and the fnal stable new binding site zone III (Fig. 1a and c). Interestingly, we found that the free energy barrier for the movement between zone I and zone II is very low ($0.1-0.5 kcal mol 1 ) facilitating the diffusion of ACh from the extracellular vestibule towards the orthosteric site. However, the free energy barrier between zone II and zone III is around 1.6 AE 0.1 kcal mol 1 for M3-ACh and 3.8 AE 0.2 kcal mol 1 for M4-ACh. These large differences in the energy barriers explain why the diffusion of ACh from zone II to zone III takes longer in the M4 than in the M3 receptor. This is probably due to the fact that in position 3.35 in TM3, M4 has a larger residue (Val) than M3 (Ala) which creates a higher steric barrier for ACh to enter the deep binding pocket. ## New putative binding sites in crystal structures To validate whether our newly discovered ligand binding pocket is also applicable to other GPCRs, we analyzed more than 200 crystal structures of family A GPCRs comprising bound ligands (Fig. 3a and S8 †) from GPCRDB. 59 In the recently deposited crystal structure of the leukotriene B4 (LTB4) receptor (PDB: 5X33), 60 the benzamidine moiety of the inverse agonist BIIL260 occupies the same new ligand binding site which we have discovered for ACh in the M3 and M4 receptors in this work, contacting D66 2.50 through an ion-lock interaction (Fig. 3b). Moreover, the highly conserved W236 6.48 in the LTB4 receptor was also found undergoing molecular switches induced by the binding of BIIL260 (ref. 60) (Fig. S5b †). Interestingly, the mass center vector of BIIL260 was found at a unique position compared to those of all other ligands resolved in the crystal structures: it directly pointed to the highly conserved D 2.50 , whereas the mass center vectors of all other ligands are located far away from D 2.50 . The normalized electrostatic potential surface area (Fig. 3d and e) reveals positively charged surfaces for BIL260 which facilitate the ion lock interaction with D 2.50 , very similar to that observed for ACh (Fig. 2a and b). In contrast, all other ligands resolved in GPCR crystal structures have either hydrophobic or negatively charged surfaces in the same region, which is very unfavorable to interact with the negatively charged D 2.50 . This is probably the reason why all the other ligands do not bind to this newly discovered pocket. We propose that a well-designed ligand with both a proper mass center vector and a positively charged head group might be able to move into this new binding pocket of the receptors addressed in Fig. 3. ## New putative ligand binding sites in other GPCRs To further investigate whether our newly discovered binding sites of the M3 and M4 receptors are also present in other GPCRs, we inspected 39 unique crystal structures of family A GPCRs (Fig. S9 † and 4). Here, we used ConCavity 53 for ligand site prediction and found that all investigated GPCRs also potentially have an additional ligand binding site between positions 6.48 and 2.50. To further characterize this binding site for each receptor, we calculated the volume of this region using the POVME tool. 54 The calculation indicates that most receptors have a reasonable volume of 30-85 3 for this newly discovered binding site. Specifcally, 5-HT 2B , M1, AT1R, APJ, H1R, kOR, OX1, P2Y 12 , PAR2 and LPA6 receptors all have a relatively small binding site, with volumes in the range of 30-40 3 . In contrast, the volumes of the putative additional binding pockets in 5-HT 1B , 5-HT 2C , AT2R, M3, M4, C5aR, CCR9, CXCR4, D2R, D3R, ETB, LPA1, dOR, mOR, RHO, OX2, P2Y 1 and S1P 1 are in the range of 40-60 3 . Finally, receptors A 1A R, M2, b 1 AR, b 2 AR, CCR2, CCR5, D4R, NOP, LTB4 and CX3CL1 have distinct larger putative ligand binding pockets, in the range of 60-85 3 which is the size of two water molecules or a small ligand. However, the volume of this pocket in each receptor might be substantially enlarged upon agonist binding as indicated by our MD simulation in this work. Interestingly, the fnal position of the sidechain of W236 6.48 observed in LTB4 is identical to that of the corresponding W400 6.48 in the agonist bound activated M2 receptor (Fig. S5 †): both sidechains underwent rotamer changes. The superimposed structures of all family A GPCRs in the data bank indicate that the c1 angles of W 6.48 can undergo changes as large as 120 (Fig. 3C). Although these new potential orthosteric sites are small in size, their volumes can be increased accordingly after the rotation of the highly conserved W 6.48 upon proper ligand binding. This is further confrmed by our previously performed long time scale MD simulations on FPR1, 61 S1P 1 R 62 and A 2A R, 55 Finally, we inspected the conservations of residues in this newly discovered cavity, based on all resolved crystal structures of family A GPCRs (Fig. 5). We found that most residues of this Fig. 4 The volumes of the newly discovered putative binding sites in the absence of ligands in 39 unique structures of family A GPCRs. Green region: the volume of the newly discovered binding site located between the highly conserved W 6.48 ## Conclusions In summary, discovering a new ligand binding site for a particular protein is an essential frst step for structure-based drug discovery. Computer modelling is a very useful tool for modern structural biology. By using MD simulations coupled with mutagenesis experiments, we identifed a potential new binding pocket for the endogenous agonist ACh in M3 and M4 muscarinic acetylcholine receptors (Fig. 6), whereas the antagonist molecule TTP is trapped in the traditional orthosteric binding site. The small and flexible ACh diffuses deeper inside the receptor towards a new binding site next to the highly conserved residue D 2.50 . Molecular switches take place at the highly conserved W 6.48 induced by ligand binding. As the volumes of both orthosteric and new sites expand upon the binding of ACh in the case of the M3 and M4 receptors, we predict similar volume changes in the other putative binding sites upon ligand binding. In our present work, we systematically inspected the binding sites of all published family A GPCR crystal structures. We discovered that most residues, forming this newly discovered pocket, are highly conserved. We also found that most GPCRs potentially comprise these newly discovered binding sites, which have been confrmed by both a recently resolved crystal structure of LTB4 and computer modeling. Our newly uncovered ligand binding sites open new opportunities for GPCR drug discovery and design. to SY and G07-13 & GA65-23 to SF), the National Center of Science, Poland (grant 2013/08/M/ST6/00788) to SF), the European Community (project SynSignal, grant no. FP7-KBBE-2013-613879) and the EPFL to HV is acknowledged. Additional funding came from the Arnold and Mabel Beckman Foundation to KP. HV and SF participated in the European COST Action CM1207 (GLISTEN). KP is the John H. Hord Professor of Pharmacology. Dedicated to Prof. Horst Vogel on the occasion of his 70th birthday.

chemsum

{"title": "Exploring a new ligand binding site of G protein-coupled receptors", "journal": "Royal Society of Chemistry (RSC)"}

calcium_looping_co₂_capture_system_for_back-up_power_plants

## Abstract: This paper analyses a CO 2 capture system based on calcium looping, designed for power plants that operate with very low capacity factors and large load fluctuations, including shut-down and start-up periods. This can be achieved by decoupling the operation of the carbonator and calciner reactors and connecting them to piles filled with CaO or CaCO 3 . When the power plant enters into operation, calcined solids are fed into the carbonator to provide the necessary flow of CaO for capturing CO 2 and storing the carbonated solids. An oxy-CFB calciner designed to have a modest thermal capacity and operate continuously refills the CaO pile. Mass and energy balances of the entire system, combined with state-of-the-art performance criteria for reactor design, have been solved to identify suitable operating windows. An analysis of the effect of the CaO reactivity of the material stored in the piles indicates that temperatures of around 500-600 1C in the carbonator are compatible with the storage of solids at low temperature (o250 1C). This, together with the low inherent cost of the material, allows large piles of stored material. Electricity costs between 0.13-0.15 $ per kWh e are possible for the system proposed in contrast to standard CaL systems where the cost would increase to above 0.19 $ per kWh e when forced to operate at low capacity. The proposed concept could be integrated into existing power plants operating as back-ups in renewable energy systems in preference to other CO 2 capture technologies that are heavily penalized when forced to operate under low capacity factors. Broader contextThe progressive adoption of low-carbon technologies is essential for tackling climate change and achieving the 1.5 1C desirable target as agreed by the COP21 in 2015. In this context, renewable energies and CO 2 capture and storage technologies will play an essential role in the power generation sector. Most of the efforts on developing CO 2 capture processes assume a base load operation. However, renewable energies are inherently intermittent and CO 2 capture systems must be able to treat flue gases from power plants operated as back-up to cover periods of time with no renewable energy production. Thus, the development of flexible CO 2 capture systems is increasingly recognized as a key point for the positioning of these technologies. While standard CO 2 capture systems developed to date are very limited for their integration in back-up power plants, this work presents a highly flexible CO 2 capture system based on calcium looping. This includes low cost solid storage piles within its boundaries allowing the capture of all the CO 2 from the existing coal power plants operated under very low capacity factors while maintaining reasonable costs. ## Introduction The power sector in many countries is rapidly evolving as lowcarbon technologies are increasingly deployed to reduce greenhouse emissions. 1,2 However, the large share of solar and wind power requires the implementation of energy storage systems 3 and/or back-up fossil power generation in order to fill the time periods when no renewable energy is available. The current need for flexible fossil power has already produced a major shift in the operating configurations of many fossil-fired power plants: 4 both existing and new fossil fuel power plants are forced to operate with significant load variations, ramping-up-and-down or even having shut-down periods to satisfy the electricity demand. As a result, very low capacity factors are to be expected (leading to significant extra costs) as well as many other technical penalties (e.g. increased wear and tear due to cycling operation). On the other hand, the use of fossil fuels as a back-up power generation system, even for limited periods of time, is incompatible with the possible long-term scenario of a deeply decarbonized energy supply. Consequently, flexible power plants equipped with CO 2 capture and storage (CCS) technology are needed to substitute the current fossil infrastructure of back-up power generation. The previously mentioned challenges are exacerbated due to Spanish National Research Council (CSIC-INCAR), C/ Francisco Pintado Fe 26, 33011, Oviedo, Spain. E-mail: yolanda.ac@incar.csic.es the great complexity of CCS systems and the capital-intensive character of sub-systems originally designed only for base load operation. This is now recognized as a major weakness of CCS technologies and explains the growing awareness of the need for more flexible CCS processes. 8, In the case of the most developed CO 2 capture technologies (i.e. post-combustion, oxy-combustion and pre-combustion), several process alternatives based on different energy and material storage options have been proposed to improve their flexibility. For example, in post-combustion amine-based CO 2 capture systems it is possible to reduce the power consumption associated with solvent regeneration and CO 2 compression units during peak demand periods. 12,16 This can be done by storing a fraction of the rich solvent leaving the absorber and operating the regenerator at lower loads during these periods. As a result, the regeneration of the stored sorbent and the compression of the CO 2 captured can be postponed to low power demand periods. 17 However, this solution implies economic penalties due to the storage of large masses of relatively costly amine, and the need for an oversized regenerator and CO 2 compression unit if the power plant is required to operate continuously at base load. 11 For oxy-combustion systems, the use of O 2 cryogenic tanks has been proposed in order to operate the air separation unit (ASU) at base load and to store the O 2 during low-energydemand periods (i.e. when the boiler operates at partial load) or by reducing the ASU capacity while the plant load is increased to respond to varying electricity requirements. 11,13,18,19 One of the main drawbacks of these approaches is the noticeable increase in O 2 production costs as a consequence of the need to liquefy and re-evaporate the stored O 2 . The storage of the H 2 produced in the pre-combustion CO 2 capture systems has also been proposed 8,11 using suitable geological structures due to their relatively low cost and large storage capacity. 20,21 The use of syngas in other industries (leading to poly-generation systems) has also been presented as an alternative to the storage of H 2 . Both the alternatives increase flexibility through the decoupling of the power generation and hydrogen production blocks. Other processes have been proposed to adapt the variable power output irrespective of the power input by incorporating a thermal energy storage system within the power plant and/or the CO 2 capture system boundaries. This concept has already been considered in previous works 25,26 for example by using molten salts as a storage medium to meet intermediate loads in coal-fired power generation. In a similar way, thermal energy storage using solids at high temperature has recently been proposed in order to increase the flexibility of power plants based on oxy-fired circulating fluidized bed (CFB) combustors with CO 2 capture taking advantage of the availability of these reactors for the circulation of large flows of high temperature solids between silos. 27,28 Another CCS technology that can benefit from the energy storage potential of reacting solids is post-combustion CO 2 capture by calcium looping (CaL). A conventional CaL configuration uses two interconnected circulating fluidized bed reactors, a carbonator and a calciner. The flue gas from a power plant is fed to the carbonator in order to capture the CO 2 using a stream of CaO particles at temperatures close to 650 1C. The carbonated stream of solids is then separated from the lean-CO 2 flue gas leaving the carbonator and sent to the calciner. In this reactor, the CaCO 3 is decomposed at temperatures of around 900-920 1C by the combustion of coal under oxy-firing conditions and the regenerated CaO is sent back to the carbonator. Thus, the CO 2 captured in the carbonator and the CO 2 released during combustion in the calciner are obtained in a concentrated stream. Post-combustion CO 2 capture by CaL has developed rapidly in the last five years up to the scale of the MW th with several pilot plants due to the similarity between its reactors and those of existing circulating fluidized bed power plants and the lower energy penalty compared to that of other CCS systems since extra power can be obtained from the additional heat input required to drive the sorbent regeneration reaction. 36,37 The use of CaO/CaCO 3 in CaL systems makes it possible to integrate a thermochemical energy storage system by taking advantage of the reversible reaction of CaO with CO 2 at high temperature. There is already background knowledge about the use of CaO/CaCO 3 to store nuclear 38,39 or solar energy or to increase the flexibility of the integrated gasification combined cycle plants. 43 Also, basic schemes of energy storage systems based on CaO/CaCO 3 post-combustion CaL have been put forward. 44,45 In relation to these concepts of energy storage, a recent European project, FlexiCaL (www.flexical.eu), is being carried out to investigate in more detail the viability of flexible CaL systems. The aim of the present work is to analyze and evaluate the operation variables of a CaL system that incorporates a largescale energy storage system and exploits the thermo-chemical energy of CaO/CaCO 3 . Specific reference process schemes for CO 2 capture in power plants operating in back-up mode are assessed. The main elements of the resulting CaL CO 2 capture system (reactors, solids piles, steam cycle, elements linked to the capture system and auxiliary components for O 2 production and CO 2 compression and purification) can then be dimensioned and a preliminary estimation of the electricity and CO 2 avoidance costs can be calculated to identify the scenarios where the proposed system would be most competitive. ## Process description The flexible post-combustion CO 2 capture concept by calcium looping (FlexiCaL), proposed in this work is presented in Fig. 1. It is based on a conventional CaL system 46 that connects a CFB carbonator to an existing power plant and includes an oxy-CFB calciner fired with coal using pure oxygen from an air separation unit (ASU). In addition, the proposed system in Fig. 1 includes features designed to allow the integration of the energy storage system: in particular, two solid storage piles to store the solids coming from the carbonator and the calciner (named for the sake of simplicity ''CaCO 3 '' and ''CaO'' respectively in Fig. 1, although the piles include other components as will be discussed later). These allow the circulation of solids between the carbonator and the calciner to be decoupled and the oxy-fired calciner to operate continuously irrespective of the CO 2 load to the carbonator. At the exit of both the CFBs, the sensible heat from the flue gas depleted in CO 2 and from the CO 2 -rich gas is recovered using convection pass heat exchangers, labelled HX1 and HX2. After the heat has been recovered, the CO 2 gas stream leaving the calciner is delivered to the CO 2 compression and purification unit (CPU) before transport and storage. The system also includes two series of external fluidized bed heat exchangers (FBHX1 and FBHX2) to cool down the solids coming from the carbonator and the calciner before they are stored at low temperature (T CaCO 3 and T CaO , respectively). These FBHXs could be based on a series of fluidized beds in countercurrent flow to the water-steam flows, as suggested by Schwaiger et al. 47 In the system depicted in Fig. 1, when the power plant enters into operation and the flue gas is fed into the carbonator, calcined solids are fed from the CaO pile into the carbonator to provide the necessary amount of sorbent during these periods. The carbonated solids leaving this reactor are then cooled down and stored in the CaCO 3 pile. This allows the capture of the CO 2 and the production of additional power from the HX1 and FBHX1 when the power plant enters into operation during high power demand periods. To fill the CaO pile with calcined solids, the oxy-CFB calciner operates at base load independently of the power plant. Thus, the solids from the CaCO 3 pile are fed continuously to the calciner together with a make-up flow of limestone (see F 0 in Fig. 1) needed to compensate for the decay of the CO 2 carrying capacity of the particles of CaO during cycling and to purge the inerts (ashes and CaSO 4 ) from the inventory of solids. The calcined solids leaving this reactor are stored in the CaO pile after being cooled down. As a result, power is continuously provided by both HX2 and FBHX2. The operation of the oxy-CFB calciner (including the ASU and CPU units) in base-load mode, i.e. disconnected from the main power plant and the carbonator, has inherent advantages in that it is simpler to control and more stable. In addition, the calciner footprint and the associated combustion equipment, which are the main cost components in CaL systems, can be considerably reduced in scale by taking advantage of large and low-cost storage piles of Ca-material. With respect to the storage conditions of the solid materials, this work mainly focused on operating at low temperatures (o300 1C) and under atmospheric pressure to facilitate verylong duration energy storage. Conventional solid handling and storage equipment, such as that used in the cement industry, is employed to store and handle large masses of this type of material. The main drawback of operating with solid piles at low temperatures is that the storage system will have lower energy storage densities as it can only exploit chemical energy (using the enthalpy of the CaO carbonation reaction, i.e. 168 kJ mol 1 CO 2 ). A storage temperature for the carbonated solids (T CaCO 3 ) of 150 1C has been fixed and a CaO pile temperature (T CaO ) objective of 200-250 1C has been targeted. Furthermore, due to the low temperature of the calcined solids entering the carbonator, the latter is assumed to be an adiabatic reactor (in contrast with the boiler-type carbonators usually assumed for standard CaL systems). An analysis of the main variables affecting the design of the scheme in Fig. 1 Energy & Environmental Science Paper energy balances, using in-house CaL process models previously validated against published works that simulate CaL processes using commercial software. This analysis is focused only on steady-state modes. The main variables affecting the process shown in Fig. 1 are the capacity factor (CF), directly related to the fraction of time that the power plant and the CFB carbonator are in operation, the activity of the CaO material (X ave ) and the solid storage temperature in the piles. Several reasonable assumptions following the typical values reported in the literature for similar systems have been made as described later. Although these assumptions could introduce uncertainties in the results, a more detailed sensitivity study of other parameters has been considered outside the scope of this first conceptual design. ## has been carried out by solving mass and An average-sized coal-fired power plant with a thermal input (P coal,power plant ) of 500 MW th has been chosen as the basis for calculations. This produces a total flue gas of 6.8 kmol s 1 with 15.9% v CO 2 when operating at full power. For simplification purposes, it has been considered that the flue gas enters the carbonator free of ashes and sulfur after having been cooled down to a temperature of 150 1C. The composition of the coal fed into the oxy-CFB calciner is 75.5% C, 3.0% H, 0.5% S, 8.0% O, 7.0% H 2 O, 1.0% N and 5.0% ash with a lower heating value (LHV) of 25.0 MJ kg 1 . This is burnt in the calciner at 900 1C using pure O 2 preheated up to 350 1C with an excess of 6%. Under these conditions, the total calcination of the CaCO 3 and a SO 2 capture efficiency of 95% in the oxy-CFB calciner can be assumed. It has been considered that 50% of the ashes leave the system through the calciner cyclone with the flue gas while the remaining ashes accumulate with the solids and need to be removed in the solid purge. With respect to the carbonator, the mass and energy balances have been solved assuming a 90% CO 2 capture efficiency. The flow of solids needed from the CaO pile is linked to the maximum average CO 2 carrying capacity of the particles (X ave ), which is directly related to the make-up flow of limestone introduced to the calciner. To estimate the X ave of the inventory of solids the expression proposed by Rodrı ´guez et al. 48 has been used. where F 0 represents the molar make-up flow of fresh limestone, F CO 2 is the molar flow of CO 2 coming with the flue gas from the power plant, F R is the molar flow of sorbent entering the carbonator reactor (from the CaO storage pile in the scheme of Fig. 1) and f carb characterizes the extent of carbonation of the particles expressed as the ratio X carb /X ave . For simplification purposes, a constant f carb value of 0.8 has been considered for all the cases presented in this work. The last term of eqn (1) refers to the impact of the SO 2 on the activity of the sorbent assuming that sulfur reacts only with the active CaO, X S being the sulfate conversion of the sorbent. Finally, a 1 , f 1 , a 2 , f 2 and b represent the sorbent deactivation constants according to eqn (2): 49 Values of a 1 = 0.1045, f 1 = 0.9822, a 2 = 0.7786, f 2 = 0.7905 and b = 0.07709 were used in this work as representative fitting parameters of typical high purity lime materials. As mentioned above, it is assumed that CO 2 is fed into the carbonator when the power plant enters into operation, while the calciner operates continuously in order to calcine the stored carbonated solids and the required make-up flow of limestone. On the basis of this assumption, F 0 /F CO 2 used in eqn (1) has been calculated as the molar ratio of the limestone fed into the calciner and the CO 2 fed into the carbonator during the lifetime of the power plant. As the carbonator is adiabatic, the reactor operation temperature and T CaO are directly linked by the activity of the solids. Thus, in order to find feasible operation windows, the effect of sorbent activity (i.e. different F 0 /F CO 2 ratios) on the CaO storage temperature was analyzed. A minimum operation temperature in the carbonator of 500 1C was chosen on the basis of previous experimental studies that indicate that temperature has only a modest effect on the carbonation reaction rates in the range of 500-650 1C due to the low activation energy. From the heat recovery point of view, lower than usual (650 1C) temperatures in the carbonator (i.e. 500-600 1C) are possible at the expense of modest efficiencies and larger heat transfer areas in HX1 and FBHX1. Fig. 2 shows the effect of the F 0 /F CO 2 ratio on the temperature of the storage pile of CaO (T CaO ). Mass and energy balances have been solved for four different carbonator temperatures (500/550/ 600/650 1C). As can be appreciated in this figure, a carbonator operation temperature of 650 1C is only possible for CaO storage temperatures well above 400 1C. This high T CaO is compatible for process schemes with lower storage capacity requirements if refractory silos are used (i.e. for the capture of CO 2 from power plants operating continuously but with load changes 44,45 ), but it ## Paper Energy & Environmental Science is considered less feasible for the storage and handling of very large masses of solids. On the other hand, it is possible to store the CaO at temperatures around 250 1C and operate the carbonator at 600 1C if large ratios of F 0 /F CO 2 (40.9) are allowed. This option is possible despite the large F 0 /F CO 2 ratios shown in Fig. 2 since the total consumption of fresh limestone and purge produced is reasonable and in the range of that required in cement plants, and furthermore when taking into account the moderate F CO 2 fed into the carbonator if the power plant operates with a low capacity factor. However, such large F 0 /F CO 2 scenarios need to be analyzed by making better use of the potential synergy between the FlexiCaL system and a cement plant (or any other large scale CaO user) which is beyond the scope of this study. If synergy with a cement plant is not feasible, lower F 0 /F CO 2 ratios are needed (0.25-0.35) to reduce the consumption of limestone in the CO 2 capture system. In this case, the carbonator must operate at temperatures below 550 1C to ensure the CaO storage temperatures stay within the targeted range (see Fig. 2). Another important parameter affecting the process depicted in Fig. 1 is the capacity factor of the existing power plant. Fig. 3 shows the fraction of thermal input to the calciner with respect to the total thermal input for different capacity factors as a function of the F 0 /F CO 2 ratio. In standard CaL systems, approximately 45-50% of the thermal input into the entire system (power plant and CaL) is introduced in the oxy-fired CFB calciner. 37 As can be seen in Fig. 3, the thermal input to the calciner can be drastically reduced by integrating the energy storage system. Thus, the fraction of the thermal input to the calciner could be lower than 0.20-0.25 if the existing power plant is operated with a capacity factor of 0.2. Other improved configurations for further minimizing the energy demand by preheating the carbonated solids entering the calciner could be implemented to further reduce the thermal input to the calciner. 44 In order to illustrate the operation of the conceptual design of the system proposed in this work, a capacity factor of 0.2 for the existing power plant has been chosen. This low capacity factor would be extremely challenging for any CO 2 capture system, as the capital cost component of electricity is inversely proportional to the capacity factor. However, it can be considered a reasonable mid-term assumption for fossil plants without capture, considering the trend observed in the operation hours of existing coal power plants connected to electrical markets with a large share of renewable energy (e.g. in the Spanish electricity market, where there has been a significant increase in the share of renewable energy during the last decade from 20% of the total electricity produced in 2007 to 40% in 2016 and where transnational network connections are limited, the fraction of operation time of fossil fuel power plants has dropped from 0.85 in 2007 to 0.53 in 2016, see www.ree.es). A F 0 /F CO 2 of 0.95 has been adopted on the assumption that the purge can be used in a cement plant. This results in a maximum CO 2 carrying capacity of the solids of 0.52 (configuration FlexiCaL/X ave = 0.52). In accordance with the discussion above, a carbonator temperature of 600 1C could be achieved by allowing a CaO solid storage temperature of 250 1C. The main mass flow streams, temperatures and power available from the different heat exchangers are summarized in Table 1. When the power plant is in operation, the flow of calcined solids is fed into the carbonator at 133.7 kg s 1 while the flow of carbonated solids sent to the CaCO 3 pile is 176.6 kg s 1 . Assuming a bulk density for the CaO solids of 1000 kg m 3 , a total of 9243 m 3 of calcined solids are needed in the CaO pile per day of power plant operation. In the case of the operation conditions in the calciner, a thermal input of 132 MW th is supplied continuously to calcine a 20.6 kg s 1 flow of fresh limestone and a 35.3 kg s 1 flow of carbonated solids coming from the CaCO 3 pile. In order to ensure that the inventory of solids in the storage system is kept constant, a 26.8 kg s 1 flow of calcined solids is sent to the CaO pile. As a result, a 11.7 kg s 1 purge of rich CaO solids is produced continuously. This translates into an annual production of around 0.37 Mton which represents approximately half of the CaO requirements of a typical cement plant with a clinker capacity of 1 Mton clinker per year assuming 0.63 kg CaO kg 1 cement . Table 1 also includes another example of a conceptual design (FlexiCaL/X ave = 0.26) where synergy with a cement plant is not possible. In this case, a F 0 /F CO 2 of 0.25 has been fixed to reduce limestone consumption to 5.4 kg s 1 . The average CO 2 carrying capacity of the solids under these conditions is 0.26. As a result, the carbonator can be operated at a lower temperature (500 1C) to allow the storage of the CaO solids at 230 1C. Due to the lower activity of the solids, the flow of solids required to be fed into the carbonator increases up to 277.1 kg s 1 while the storage capacity needed per day of operation increases up to 19155 m 3 . In this case, the calciner's capacity decreases slightly and a thermal input of 112 MW th is needed. A detailed integration of the heat sources of the system depicted in Fig. 1 into a steam cycle is considered to be outside the scope of this work. However, since HX1 and FBHX1 only enter into operation with the power plant, the use of two different steam cycles will be required (e.g. one associated with the carbonator and one linked to the calciner). The steam cycle associated with the calciner must be able to satisfy the power requirements of the calciner (including the ASU, CPU and auxiliaries). In order to estimate the power needed, average specific consumptions of 200 kWh e per t O 2 and 120 kWh e per t CO 2 for the ASU and CPU respectively have been adopted on the basis of the values reported in the literature. 55 It is assumed that 5% of the gross electric power output is consumed in the auxiliaries. 56 Assuming a reasonable efficiency value of 45% for the steam cycle associated with the oxy-calciner, the power produced would be 27.4 and 29.4 MW e for the FlexiCaL configurations with values of X ave = 0.52 and X ave = 0.26 respectively. Thus, after accounting for the ASU, CPU and auxiliaries consumptions, the power available for export from the system proposed could be drastically reduced to 2 and 9 MW e during the power plant shut down periods (i.e. when no electricity is required). On the other hand, additional power of up to 194 and 202 MW th is available from HX1 and FBHX1 (P carbonator ) for X ave = 0.52 and X ave = 0.26 respectively. By assuming a net efficiency of 0.39 for the operation conditions of the heat sources in the CFB carbonator (500-600 1C) as in similar steam cycles integration schemes, 57,58 74.1 and 84.2 MW e of extra power can be produced when the power plant enters into operation. ## Costs analysis A basic economic study has been carried out to estimate the impact of retrofitting the CO 2 capture system of Fig. 1 into an existing power plant. The economic parameters used in this analysis are the cost of the electricity (COE) and the CO 2 avoidance costs (AC). The COE can be calculated as follows: where TCR is the total capital required, FCF is the fixed charge factor, FOM is the fixed operating cost, VOM is the variable operating cost, FC is the fuel cost and Z plant is the net plant efficiency including that of the CO 2 capture system. The CO 2 avoidance costs (AC) represent the cost of reducing one ton of CO 2 while providing the same amount of power as a reference plant without CO 2 capture and can be calculated as follows: where CO 2 /kWh e is the CO 2 emission factor of each plant. Two different reference power plants were considered, a new plant with a high net efficiency of 0.45 and an existing plant with its capital already amortized and with a much lower net efficiency a Refers to the volume of solids stored during one day with the power plant operating at full load and assuming a solid bulk density of 1000 kg m 3 . b Available heat after the preheating of the stream (8). ## Paper Energy & Environmental Science of 0.35. Based on reference cost data available for coal power plants, 59 the representative data shown in Table 2 have been chosen to carry out the cost analysis. The calculated costs of the electricity for the new and the amortized coal power plant operating at base load (CF = 0.9) are 0.050 and 0.031 $ per kWh e respectively which is in agreement with the data reported in the literature. 59 However, if the new power plant is operating as a back-up system with a low capacity factor (CF = 0.2), the COE increases up to the uneconomic value of 0.152 $ per kWh e . Using large coal power plants as back-ups seems a more reasonable strategy for amortized systems, as the cost of the electricity is mainly dependent on the operation and fuel costs. Thus, as shown in Table 2, the COE of an amortized power plant operating with a capacity factor of 0.2 could be as low as 0.048 $ per kWh e . Accordingly, the analysis that follows will be focused only on the integration of the CO 2 capture system of Fig. 1 into an existing power plant whose initial capital cost expenditure has been fully amortized. For comparison purposes, a conventional CaL system has been also included in this analysis assuming that it retrofits an amortized coal power plant (Standard CaL in Table 2). In order to calculate the capital cost of the equipment, the specific cost of the equipment is expressed in terms of unit of thermal input as explained in similar works for other CaL systems. 36 In the case of the FlexiCaL system in Fig. 1, we have attempted to estimate the specific costs by exploiting similarities with the elements already developed at the commercial level. Four major components can be identified for the cost analysis: the power plant, the CFB carbonator, the oxy-CFB calciner, and the storage system. In order to facilitate discussion, the specific costs of the equipment, TCR Total , have been defined as per unit of thermal power. Thus, the total cost of the whole system (TCR) can be estimated as TCR = TCR total P total = TCR total (P coal,power plant + P carbonator + P coal,calciner ) Regarding the cost of the existing power plant, we considered that no additional modifications are needed to satisfy the requirements of the CO 2 capture step. Therefore, the specific cost of the power plant in Fig. 1 can be assumed to be zero. The CFB carbonator is linked to the thermal input to the power plant. This component can be considered as an adiabatic combustor whose TCR refractory is assumed to be 125 $ per kW th based on the cost of pre-calciners in cement plants. 60 Another element that is related to the thermal input of the power plant is the fraction of CPU that treats the flue gas captured in the carbonator. A specific cost of 80 $ per kW th has been assumed for the CPU unit 61 considering that the CO 2 captured as CaCO 3 from the power plant is then released from the calciner operating at base load. The process of Fig. 1 also includes the equipment needed to extract the thermal power associated with HX1 and FBHX1 (P carbonator ) and to generate power. A TCR P carbonator of 450 $ per kW th has been adopted assuming that these elements represent around 50% of the TCR of a conventional power plant. 59 In the case of the energy storage system, only the cost associated with the inventory of solids in the solids piles has been included in the specific capital cost of the solid storage system, TCR storage , as the handling and transport of solids between the solid piles has already been included in the cost of the carbonator and calciner circulating fluidized reactor. TCR storage has been calculated assuming 10 $ per ton of limestone and taking into account the fact that the CO 2 carrying capacity of the solids and the enthalpy of the carbonation reaction (467 kWh th per ton of active sorbent) will result in 0.05-0.10 $ per kWh th depending on the value of X ave in eqn (1). Finally, the oxy-CFB calciner in Fig. 1 resembles an oxy-CFB power plant. According to the data available in the literature, 61 a total capital requirement for an oxy-CFB power plant of 3700 $ per kW e is assumed, which results in 1296 $ per kW th by assuming a thermal efficiency of 0.35. Approximately 60% of this cost corresponds to the combustion equipment (i.e. coal feeding system, ash removal systems, flue gas cleaning, ASU and CPU units) while the rest can be associated with the power block (including the heat exchangers HX2 and FBHX2 and the corresponding entire steam cycle). The thermal power transferred to the power block in the oxy-CFB calciner (P calciner ) in Fig. 1 represents only around 50% of the thermal input fed into the calciner (P coal,calciner ) when compared with conventional oxy-fired CFB power plants. Therefore, the cost of the TCR calciner is assumed to be 1037 $ per kW th . Once the cost associated with the different components has been defined, the total capital requirement of the system of Fig. 1 can be calculated as follows: TCR total = (TCR power plant + TCR CPU ÁCF + TCR refractory )Áf power plant + (TCR P carbonator + 2ÁTCR storage Át max )Áf carbonator + TCR calciner Áf calciner (6) where f power plant , f carbonator and f calciner are the power fractions of the power plant, carbonator and calciner with respect to the total power input. In order to calculate the TCR Total corresponding to the appropriate units of eqn (3) ($ per kW e ), the net efficiency of a For a make-up flow consumption of (F 0 /F CO 2 = 0.1) and a thermal input into the calciner of 410 MW th 36 assuming that it retrofits an amortized coal power plant. b Including a make-up limestone cost of 10 $ per ton. Energy & Environmental Science Paper the system depicted in Fig. 1 (Z plant ) can be estimated by means of eqn (7): The thermal power associated with the calcination of the purge (P calc , F 0 ) is discounted from the estimation of the efficiency to account for the credits associated with the utilization of the purge. According to this equation, efficiencies of 0.26 and 0.29 are calculated for the FlexiCaL system with X ave values of 0.52 and 0.26 respectively. The main cost parameters for the cases chosen as examples are summarized in Table 3. In order to focus the discussion on the specific capital requirements, the same fuel and fixed costs as those of the conventional CaL system have been assumed. The cost associated with the consumption of limestone has been included under variable costs (VOM). Based on the assumptions summarized in Table 3 the COE calculated by means of eqn (3) for the FlexiCaL system are 0.150 and 0.126 $ per kWh e for X ave values of 0.52 and 0.26 respectively. The higher COE obtained for FlexiCaL/X ave = 0.52 is mainly due to the higher make-up flow of limestone and the slightly lower efficiency. The potential advantages of the system proposed in this work can be appreciated when compared with the COE calculated for a conventional CaL system retrofitted into a power plant operating with a CF value of 0.2. In this case, the system is heavily penalized by the higher capital requirements, with the COE increasing up to 0.192 $ per kWh e (see Table 2). If the system in Fig. 1 is designed for capturing CO 2 from a power plant operating with a high capacity factor, the contribution of the energy storage system to the whole process is reduced. As a result, the contribution of the costly oxy-fuel calciner is higher and the total capital requirements increase. In that case, the FlexiCaL system and the standard CaL would show similar COE values of around 0.064-0.082 $ per kWh e at the high CF values of the power plant. However, as the capacity factor decreases (see Fig. 4a), the COE of the FlexiCaL integrated with an energy storage system becomes much more favorable, with a reduction of 35% in the case of FlexiCaL/X ave = 0.26 when compared to a standard CaL system at CF = 0.2. To calculate the CO 2 avoidance costs, an existing amortized power plant operating with the same capacity factor of 0.2 has been chosen as a reference. It must be noted here that, as it has been assumed that both the reference CaL and FlexiCaL systems retrofit an amortized power plant with a lower COE than a new power plant (see Table 2), the AC calculated by means of eqn ( 4) are increased when compared with systems coupled to new power plants (e.g. 56 $ per t CO 2 vs. 26 $ per t CO 2 for a standard CaL system operated at base load). As can be seen in Fig. 4b in all the cases the AC are higher than those corresponding to a standard CaL system operating at base load (for a typical capacity factor of 0.9). However, if the standard CaL operates at CF = 0.2, the cost of CO 2 avoided increases almost 5-fold up to a value of 251 $ per t CO 2 . This very high cost can make this kind of system economically unviable for such low capacity factors. In contrast, the increase in the CO 2 avoidance costs for the FlexiCaL cases operating with capacity factors of 0.2 is relatively low for the examples presented in this work. For example, an increment of only 26 $ per t CO 2 is obtained for the FlexiCaL/X ave = 0.26 configuration compared to the standard CaL system operating at base load. It should be noted however that it is an immense challenge to operate any other CO 2 capture system at such low capacity factors as those required for power plants used for back-up purposes. The need to store very large quantities of the functional material (such as amines in post-combustion or cryogenic O 2 in oxy-fuel systems) would be at the expense of substantial economic penalties due to the inherently higher costs of these materials. In contrast, the FlexiCaL system proposed in this work allows the storage of very large masses of sorbent at a very low cost, due to the low cost of the CaO precursor (limestone), the lack of environmental risks, and the low thermal conductivity of stagnant solids. Another important advantage of this concept is the possibility of retrofitting to existing coal power plants as there is no need for any modification to the existing power plant as with other technologies (e.g. the need to extract steam in the case of post-combustion amine-based systems, to desulfurize the flue gas, to modify the boiler in the case of oxy-combustion etc.). Most important of all, these plants do not have much chance of being incorporated into future energy systems where there is an increasing share of renewable power, other than as back-up systems for brief periods of time, unless a competitive method for CO 2 capture during the brief periods of operation can be found.

chemsum

{"title": "Calcium looping CO₂ capture system for back-up power plants", "journal": "Royal Society of Chemistry (RSC)"}