Datasets:

ghezalahmad

/

training

like 0

See discussions, st ats, and author pr ofiles f or this public ation at : .researchgate.ne t/public ation/335404336 MECHANICAL AND MICRO-STRUCTURE CHARACTERIZATION OF STEEL FIBER- REINFORCED GEOPOLYMER CONCRETE Conf erence Paper · May 2019 CITATIONS 2READS2 author s: Some o f the author s of this public ation ar e also w orking on these r elat ed pr ojects: Geopolymer s for W ater Purific ation View pr oject Progressiv e collapse analysis of RC struct ures View pr oject S. A. Elkholy Fayoum Univ ersity 68 PUBLICA TIONS    226 CITATIONS     SEE PROFILE Hilal El-Hassan Unit ed Ar ab Emir ates Univ ersity 97 PUBLICA TIONS    1,067 CITATIONS     SEE PROFILE All c ontent f ollo wing this p age was uplo aded b y S. A. Elkholy on 23 Dec ember 2021. The user has r equest ed enhanc ement of the do wnlo aded file. Performance Evaluation and Microstructure Characterization of Steel Fiber –Reinforced Alkali-Activated Slag Concrete Incorporating Fly Ash Hilal El-Hassan, Ph.D., M.ASCE1; and Said Elkholy, Ph.D.2 Abstract: This paper investigates the performance of steel fiber –reinforced alkali-activated slag concrete incorporating different fly ash replacement percentages. Three different molarities of sodium hydroxide (SH) were combined with sodium silicate to activate the bindingphase. Double hooked-end steel fibers were incorporated into the alkali-activated mix in varying volumetric proportions up to 3% to enhance its ductility. Blended binder, alkali-activator solution, dune sand, and coarse aggregate contents were proportioned and samples were cured at ambient conditions. Results showed that higher slag content, molarity of SH, and fiber addition led to less-workable concretes but withimproved mechanical properties, especially at early ages. Fly ash replacement of 25% could enhance mechanical performance after 28 days. Analytical models correlating mechanical properties were developed for alkali-activated slag concretes with fly ash. Scanning electron microscopy, differential scanning calorimetry, and Fourier transform infrared spectroscopy highlighted the coexistence of calcium alumi-nosilicate hydrate and sodium aluminosilicate hydrate gels. DOI: 10.1061/(ASCE)MT.1943-5533.0002872 .© 2019 American Society of Civil Engineers. Author keywords: Fiber-reinforced concrete; Mechanical properties; Rheological properties; Microstructure analysis; Analytical models. Introduction The concrete industry has been facing global pressure to reduce its greenhouse gas emissions and consumption of natural resources. To alleviate these environmental concerns, scientists and environ- mentalists recommend the replacement of cement by supplemen-tary cementitious materials (SCMs). If integrated into concrete mixes, they have the dual benefit of reducing cement usage while also usefully disposing of solid industrial byproducts. Fly ash (FA) and ground granulated blast furnace slag (referred to hereafter as slag) are industrial byproducts of the combustion of coal and pro- duction of steel, respectively. With their pozzolanic properties and global abundance, they are highlighted as primary replacements ofordinary portland cement (OPC) in concrete. Fly ash –based geopolymer concretes exhibit promising per- formance as an alternative to conventional concrete. However, their activation depends on curing at elevated temperatures. Such energy-intensive curing is considered a hindrance to its adoption by the concrete industry. In comparison, alkali-activated slag mortars and concretes cured in ambient conditions have been reported to exhibit limited workability but high compressive strengths up to 150 MPa ( Aydın and Baradan 2014 ;El-Hassan and Ismail 2018 ; El-Hassan et al. 2018 ). However, they experience deterioration that is attributed to the development of microcracks over time(Collins and Sanjayan 2001 ). Ballekere Kumarappa et al. ( 2018 ) explained that such shrinkage-induced microcracks are directly re-lated to the properties of the alkali-activator solution. The differentmicrostructures of fly ash geopolymer and alkali-activated slagconcretes partially explain these results. The main reaction product of the former is a three-dimensional sodium aluminosilicate hydrate (N─A─S─H) gel, resulting from the dissolution of SiO 2(Nath et al. 2016 ;Palomo et al. 1999 ;Škvára et al. 2009 ), while those of the latter are phases of hydrotalcite and calcium silicate hydrateswith traces of aluminum (C ─A─S─H) gel ( Ismail et al. 2014 ;Sun et al. 2018 ). Researchers have investigated the incorporation of slag in fly ash–based geopolymers to eliminate the need for heat curing. Research findings have highlighted an increase in mechanicalperformance but a decrease in slump/flow and setting times(Nath and Sarker 2014 ;Saha and Rajasekaran 2017 ;van Deventer et al. 2015 ;Yang et al. 2012 ). These blended geopolymers offer good resistance to water permeation, abrasion, acid and sulfate at-tack, elevated temperatures, corrosion, and fire ( Bernal et al. 2012 ; Karahan and Yakupo ğlu 2011 ;Shi 2003 ;van Deventer et al. 2015 ; Zhang et al. 2018 ). However, the chemical composition of slag seems to affect the hydration kinetics and strength development,especially the quantities of magnesia and alumina ( Winnefeld et al. 2015 ). While extensive work has been conducted on fly ash –based geopolymers incorporating slag (slag-to-fly ash ratio between 0and 1), limited work has been performed on alkali-activated slagconcretes with fly ash (slag-to-fly ash ratio >1)(Rodrigue et al. 2018 ;Wardhono et al. 2015 ). The effect of fly ash replacement on the microstructure of alkali-activated slag concrete has alsonot yet been investigated. Despite their superior performance, geopolymer and alkali- activated concretes have shown little resistance to cracking dueto their brittle nature. For application in the construction industry,their tensile and flexural properties should be improved. Steel fibershave been proposed for their ability to reduce crack propagation and cause more ductile behavior. Recent studies investigated the1Assistant Professor, Dept. of Civil and Environmental Engineering, United Arab Emirates Univ., P.O. Box 15551, Al Ain, United Arab Emirates (corresponding author). ORCID: Email: helhassan@uaeu.ac.ae 2Assistant Professor, Dept. of Civil and Environmental Engineering, United Arab Emirates Univ., P.O. Box 15551, Al Ain, United Arab Emirates; Associate Professor, Fayoum Univ., Fayoum 63514, Egypt. Note. This manuscript was submitted on June 11, 2018; approved on April 17, 2019; published online on July 26, 2019. Discussion period open until December 26, 2019; separate discussions must be submitted for individual papers. This paper is part of the Journal of Materials in Civil Engineering , © ASCE, ISSN 0899-1561. © ASCE 04019223-1 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. mechanical properties of steel fiber –reinforced geopolymer and alkali-activated composites. An increase in the modulus of elastic- ity and the compressive, tensile, and flexural strength was reportedupon using steel fibers in fly ash –based geopolymer with dosages up to 1% by volume ( Ganesan et al. 2015 ). Geopolymers incorpo- rating slag and palm oil fuel ash (POFA) or silica fume reported enhanced mechanical properties, including tensile and flexural strength, first crack load, and impact resistance upon the additionof 0.5% and 2% steel fibers ( Al-Majidi et al. 2017 ;Aydın and Baradan 2013 ;Islam et al. 2017 ). Bernal et al. ( 2010 ), Beglarigale et al. ( 2016 ), and Guo and Pan ( 2018 ) noted a significant improve- ment in the performance of alkali-activated slag mortar and con- crete with the use of 0.5% steel fibers by volume. Yet, no work has been conducted on the addition of steel fibers (up to 3%) tofly ash –slag blended alkali-activated concrete. Summarizing the review, while many authors have investigated the development and performance of steel fiber –reinforced geopol- ymer and alkali-activated concrete made with single binders (slagor fly ash), very few studies have examined the mechanical proper- ties and microstructure characteristics of blended alkali-activated concretes. In this research, steel fiber –reinforced alkali-activated slag concrete blended with fly ash and cured under ambient con- ditions were produced. Different blends of fly ash with slag wereused (fly ash ∶slag of 50∶50,25∶75, and 0∶100) and served to elimi- nate heat curing and age-related microcrack development while promoting the adoption of alkali-activated concrete by the con-struction industry for in situ applications. Unlike past research with the highest fiber volume fraction used being 2%, in this work, fibers were added up to dosages of 3% to transform the ceramic-like brittle behavior of alkali-activated concretes into a more duc- tile one. Three different molar concentrations of sodium hydroxide (SH) were used to assess its combined effect with fiber incorpo- ration. While past work utilized manufactured crushed fine sand,this study employed desert dune sand as a sustainable fine aggre- gate. The performance of alkali-activated concretes made withdifferent blends of slag and fly ash were evaluated in terms offresh properties, such as workability, and hardened properties, in-cluding modulus of elasticity and compressive, tensile, and flexu-ral strengths. An analytical study relating mechanical properties to codified and literature equations was also performed. Such a correlation has not been established for steel fiber –reinforced alkali-activated slag concrete (with and without fly ash), and itis important to provide a prediction of mechanical performancewithout the need for conducting extensive experimental testing.The microstructure of plain alkali-activated concretes has alsobeen analyzed by differential scanning calorimetry (DSC), Fouriertransform infrared spectroscopy (FTIR), and scanning electronmicroscopy (SEM). Experimental Program Materials Class F ( ASTM 2015a ) fly ash and slag were used in this study as the blended alkali-activated binding material. The specific gravityand Blaine fineness of fly ash were determined to be 2.32 and3,680 cm 2=g, respectively, while those of slag were 2.70 and 4,250 cm2=g, respectively. Figs. 1and2present the particle size distribution, SEM micrograph, and X-ray diffraction (XRD) spec-trum of fly ash and slag, respectively. Crushed stone was used ascoarse aggregate with a nominal maximum size of 10 mm, densityof1,660 kg=m 3, and water absorption of 0.7%. Prior to incorpo- ration into alkali-activated concrete mixes, the aggregates were pre-pared to surface-saturated dry (SSD) condition. Desert dune sand served as fine aggregate. Its specific gravity and unit weight were 2.57 and 1,670 kg=m 3, respectively. Fig. 3shows the particle size (b) (a) (c) 020406080100 0.1 1 10 100Percentage passing (%) Particle size (µm)Fly ash 10 20 30 40 50 60Intensity (counts) Position (2 θ)Q: Quartz M: Mullite H: Hematite QQ MM HQ QQ Q Fig. 1. (a) Particle size distribution; (b) SEM micrograph; and (c) XRD spectrum of fly ash. © ASCE 04019223-2 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. distribution, scanning electron micrograph, and X-ray diffraction spectrum. The chemical compositions of as-received fly ash, slag, and dune sand are shown in Table 1. An alkali-activator solution was prepared as a mixture of sodium silicate (SS) and SH.The mass chemical composition of the Grade N SS solution was 26.3% SiO 2, 10.3% Na 2O, and 63.4% H 2O. The SH solution was formulated to molarities of 8, 10, and 14 by dissolving97%–98% pure NaOH flakes in tap water. Past work by(b) (a) (c) 020406080100Percentage passing (%) Particle size (µm)Slag 1 10 100 10 20 30 40 50 60Intensity (counts) Position (2 θ)Q: Quartz M: Mullite G: Gehlenite GQ QGM Fig. 2. (a) Particle size distribution; (b) SEM micrograph; and (c) XRD spectrum of slag. (b) (a) (c)020406080100Percentage passing (%) Particle size (µm)Dune sand 10 100 1000 10 20 30 40 50 60Intensity (counts) Position (2 θ)Q Q C Q QCFA FQ: Quartz C: Calcite F: Iron oxide A: Aluminum oxide Fig. 3. (a) Particle size distribution; (b) SEM micrograph; and (c) XRD spectrum of dune sand. © ASCE 04019223-3 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. Kanesan et al. ( 2017 ), Patankar et al. ( 2014 ), and Sani et al. (2016 ) showed that a molarity below 8 and above 14 reduced the reaction efficiency, while a molarity of 10 acted as a bettermedian than 12. Double hooked-end steel fibers with a specificgravity of 7.9, an aspect ratio of 65, and length of 35 mm wereused. To ensure sufficient workability for mixes with high steelfiber content, a polycarboxylic ether polymer –based superplasti- cizer (SP) was used. It has been reported that such SPs can im-prove the workability of alkali-activated composites withoutcompromising the mechanical properties ( Montes et al. 2012 ; Palacios and Puertas 2005 ). Alkali-Activated Concrete Mix Design A total of seventeen (17) alkali-activated concrete mixes were pre- pared with varying fly ash replacement proportions (0%, 25%, and50%), SH molar concentrations (8, 10, and 14 M), and fiber volumefractions (0%, 1%, 2%, and 3%), as shown in Table 2. Mixtures were designated as GX-YM-ZSF, where X is the slag content (per-centage of binder by mass), Y is the molar concentration of SH, andZ is the steel fiber volume fraction (%). For instance, Mix 9, des-ignated as G75-8M-1.0SF, represents a concrete mix made with aslag:fly ash ratio of 75:25, 8 M sodium hydroxide solution, andsteel fiber volume fraction of 1%. Binder, aggregate and activatorsolution contents were optimized based on previous work con-ducted by the authors ( El-Hassan and Ismail 2018 ;Ismail and El-Hassan 2018 ). For all mixes, the total binder, dune sand, andcoarse aggregate contents were 500, 550, and 1,100 kg=m 3, respec- tively. An alkali-activator solution-to-binder ratio of 0.40 was se-lected with a SS-to-SH ratio of 2.5. The SP content was 2% bybinder mass, except for samples made with equal proportions offly ash and slag, where no SP was added. Steel fibers were incor-porated into the mix in the range of 0 –2% for samples with 14 M SH and up to 3% for 8 M SH. Preliminary test trials showed thatmixes with higher SH molar concentration had reduced workabilityand compactability due to an increased viscosity and thus could notincorporate more than 2% steel fibers (see the section “Workabil- ity”). Other than Mixes 1 –6, all concretes were made with a slag ∶fly ash ratio of 75∶25due to the superior performance after 28 days (see the section “Results and Discussion ”). Also, it should be noted that Mixes 13 and 14 (G75-10M-0.0SF and G75-10M-2.0SF)served as median samples with 10 M SH to assess the effect of SHmolar concentration on the performance of steel fiber –reinforced alkali-activated slag/fly ash blended concrete. Sample Preparation Concrete samples were cast under ambient conditions (temperature of24°C/C62°C and relative humidity of 50% /C65%). The alkali- activator solution was prepared 24 h prior to casting to allow fordissipation of heat associated with the exothermic chemical reac-tions of SH flakes with water and SH solution with SS solution. Thedry components, comprising slag, fly ash, dune sand, and coarseaggregates, were mixed in a pan mixer for 3 min. The preparedsolution was gradually incorporated into the dry components andmixed for another 3 min to ensure homogeneity and uniformity.Superplasticizer was added a few seconds after the activatorsolution. Freshly mixed alkali-activated concrete samples wereprepared as 100×200mm (diameter ×height) cylinders for com- pressive strength testing, 150×300mm (diameter ×height) cylin- ders for tensile splitting testing, and 100×100×500mm prisms for flexural strength testing. One additional cylindrical sample(100mm diameter ×200mm height) was prepared for micro- structure testing. Specimens were cast into two to three layers, compact-vibrated for 10 s on a vibrating table, left to rest for 24 hat ambient conditions, and then demolded. So-produced concreteswere kept under ambient conditions until testing age to simulateon-site construction scenarios.Table 1. Chemical composition of as-received materials OxideMaterial (%) Slag Fly ash Dune sand CaO 42.0 3.3 14.1 SiO 2 34.7 48.0 64.9 Al2O3 14.4 23.1 3.0 MgO 6.9 1.5 1.3Fe 2O3 0.8 12.5 0.7 Na2O 0.0 0.0 0.4 K2O 0.0 0.0 1.1 Loss on ignition 1.1 1.1 0.0 Others 0.2 10.5 15.5 Table 2. Mix proportion of alkali-activated concrete (kg =m3) Mix No. Mix designationBinder Aggregate Activator SPSteel fiber volume (%) Fly ash Slag Dune sand Coarse SS SH (molarity) 1 G50-14M-0.0SF 250 250 550 1,100 143 57 (14M) 0 0.0 2 G50-14M-2.0SF 250 250 550 1,100 143 57 (14M) 0 2.0 3 G75-14M-0.0SF 125 375 550 1,100 143 57 (14M) 10 0.0 4 G75-14M-2.0SF 125 375 550 1,100 143 57 (14M) 10 2.0 5 G100-14M-0.0SF 0 500 550 1,100 143 57 (14M) 10 0.0 6 G100-14M-2.0SF 0 500 550 1,100 143 57 (14M) 10 2.07 G75-8M-0.0SF 125 375 550 1,100 143 57 (8M) 10 0.0 8 G75-8M-0.5SF 125 375 550 1,100 143 57 (8M) 10 0.5 9 G75-8M-1.0SF 125 375 550 1,100 143 57 (8M) 10 1.0 10 G75-8M-1.5SF 125 375 550 1,100 143 57 (8M) 10 1.5 11 G75-8M-2.0SF 125 375 550 1,100 143 57 (8M) 10 2.012 G75-8M-3.0SF 125 375 550 1,100 143 57 (8M) 10 3.0 13 G75-10M-0.0SF 125 375 550 1,100 143 57 (10M) 10 0.0 14 G75-10M-2.0SF 125 375 550 1,100 143 57 (10M) 10 2.0 15 G75-14M-0.5SF 125 375 550 1,100 143 57 (14M) 10 0.5 16 G75-14M-1.0SF 125 375 550 1,100 143 57 (14M) 10 1.017 G75-14M-1.5SF 125 375 550 1,100 143 57 (14M) 10 1.5 © ASCE 04019223-4 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. Performance Evaluation The workability of fresh alkali-activated slag/fly ash blended concrete was characterized by the slump of ASTM C143 ( ASTM 2015c ). The mechanical performance of plain and steel fiber – reinforced concrete samples was evaluated following standardizedprocedures. ASTM C39 ( ASTM 2015b ) was employed to measure the compressive strength at 1, 7, and 28 days. The static chord modu- lus of elasticity was determined as per the procedure of ASTM C469 (ASTM 2014 ). In this test, a 500-kN compression load cell was used to record the applied compressive load, while four 60-mm-longstrain gauges were attached at midheight of the cylinder on diamet- rically opposite points. The tensile strength of 28-day specimens was measured indirectly following the method of tensile splitting strength of ASTM C496 ( ASTM 2011 ). The flexural strength (also referred to as the modulus of rupture) was calculated by performing four-pointloading tests on 28-day concrete prism samples as per ASTM C78 (ASTM 2016 ). For each experimental performance test, three repli- cates were prepared to obtain an average. Analytical Techniques The mechanical performance of alkali-activated slag concrete made with up to 50% fly ash replacement can only be properly under- stood through microstructure analysis. With limited information available in the literature, the microstructure of untested alkali-activated slag concrete with different fly ash replacement percent- ages was analyzed at the ages of 7 and 28 days. Three specimens were tested for each experimental test to ensure credibility andconsistency of results. Differential scanning calorimetry was per- formed using a DSC-Q200 calorimeter (TA Instruments, New Castle, Delaware) at a heating rate of 10°C=min in a constant flow of nitrogen gas ( 50mL=min). Concrete samples were finely ground and sieved through an 80- μm sieve to remove coarse particles. The obtained powder specimen was weighed (10 mg) and packed into aluminum crucibles for DSC testing. Fourier transform infrared spectroscopy was conducted using a Varian 3100 FT-IR spectrometer (Walnut Creek, California) in transmittance mode at a resolution of 1cm −1from 400 to 4,000 cm−1. Concrete samples, similar to those used in DSC test- ing, were mixed with potassium bromide at a ratio of 3:1, by mass, and pelletized for FTIR testing. Scanning electron microscopy was performed using a JEOL- JSM 6390A microscope (Tokyo) with energy dispersive X-ray(EDX). Concrete chunks (25 mm in diameter) were obtained from the core of untested original alkali-activated concrete specimens, sputter-coated with a thin gold layer (99% purity) to ensure con- ductivity, and examined in high-vacuum SEM mode.Results and Discussion Workability The slump values, characterizing the workability, of plain and steel fiber–reinforced alkali-activated concrete are shown in Fig. 4.A s illustrated in Fig. 4(a), the replacement of slag by 25% and 50% fly ash increased the slump by 17 and 45 mm, representing increases of74% and 196%, respectively. This improvement in workability isattributed to the lubricating effect of spherical fly ash particles,which produced less interparticle frictional forces ( Sun et al. 2003 ). The addition of 2% steel fibers by volume resulted in zero slump for 75% and 100% slag concrete samples. The slump was 10 mm for concretes with a fly ash content of 50%. Fig.4(b)shows that plain concrete samples made with a slag ∶fly ash ratio of 75∶25and 8, 10, and 14 M SH solution observed slumps of 60, 52, and 40 mm, respectively. As the molarity increased, a13% and 33% reduction in slump was noted. This is owed to the higher viscosity of the produced activator solution. The incor- poration of steel fibers caused a significant decrease in the work-ability. On average, every 0.5% steel volume fraction additionreduced the slump by 11 mm as a result of more mortar adheringto the fibers, which caused an increase in viscosity and decrease inslump ( Chen and Liu 2005 ). A zero slump was noted for 14-M-SH samples made with 2% steel fibers by volume. Evidently, with higher SH molar concentration, a less workable and compactableconcrete was produced, to the extent that only up to 2% steel fiberscould be incorporated. Further additions of steel fibers (3%) wasthus not possible in 14-M-SH concretes. Compressive Strength The compressive strength of alkali-activated concrete samples was tested at the ages of 1, 7, and 28 days. Fig. 5(a)presents the com- pressive strength development of 14-M-SH concretes incorporating0%, 25%, and 50% fly ash and 0% and 2% steel fiber reinforce- ment. At an early age of 1 and 7 days, the compressive strength decreased as the fly ash content increased from 0% to 50%. Thisis mainly due to a slower fly ash polymerization reaction at ambienttemperatures ( Ismail et al. 2014 ). In comparison, samples with high slag content have a larger reactant surface area (i.e., more ac-tivation sites), leading to an accelerated reaction of calcium and production of aluminosilicate hydrate and calcium silicate hydrate gels ( Al-Majidi et al. 2016 ;Chi 2012 ;Davidovits 2008 ;Palomo et al. 1999 ). However, at a later age of 28 days, the compressive strength slightly increased as the fly ash content increased from 0%to 25% but decreased when the content was further increased to50%. Apparently, fly ash contributes to later strength through(b) (a)01020304050607080 02 0 4 0 6 0Slump (mm) Fly Ash Content (%)SF=0% SF=2%01234Slump (mm) Steel Fiber Volume (%)8M 10M 14M Fig. 4. Workability of alkali-activated slag concretes with different: (a) fly ash content; and (b) volumetric fiber fractions. © ASCE 04019223-5 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. the development of N ─A─S─H gel, while slag is the main con- tributor to early-age strength with the formation of C ─A─S─H gel. Similar findings have been reported by Ismail et al. ( 2014 ). It is also worth noting that the highest increase in strength between7 and 28 days was 23.8% and 19.5% for samples with 50% and75% slag, respectively. Nevertheless, 50% slag replacement byfly ash could hinder the development of strength, remaining 12% lower than its 25% fly ash concrete counterpart. Furthermore, the addition of 2% steel fibers increased the compressive strengthby 20% –52% (average of 33.3%), with highest increase reported for alkali-activated concrete mixes with 25% fly ash. The compressive strength development of plain and steel fiber – reinforced alkali-activated slag concrete with 25% fly ash and 8 M SH is shown in Fig. 5(b). On average, 78% and 89% of the 28-day compressive strength was obtained at 1 and 7 days of age, respec-tively. As for 14-M-SH concretes of Fig. 5(c), an average of 81% and 97% of the respective 28-day compressive strength wasreached within 1 and 7 days. This is a clear indication of the rapid strength development of alkali-activated slag/fly ash blended con- crete. Temuujin et al. ( 2009 ) reported that such a phenomenon is mainly due to the high SH concentration, which improves theefficiency of activation reaction and increases the early-age com- pressive strength. Fig. 5(d) shows the strength development of 0.0% and 2.0% steel fiber –reinforced concretes made with a slag ∶fly ash ratio of 75∶25and different SH molar concentrations (8, 10, and 14 M). A superior performance can be noted with higher concentrationof SH solution. It is believed that the increase in SH molarity provides more sodium ions that could balance the charges, thusforming aluminosilicate networks in the alkali-activated mix (Sathonsaowaphak et al. 2009 ). On the other hand, a low molarity of SH could limit the activation reaction due to less leaching ofsilica and alumina from the source material ( Alonso and Palomo 2001 ). The addition of steel fibers to the alkali-activated slag/fly ash blended concrete created a significant variation in the developmentof compressive strength. The strength index, shown in Eq. ( 1), was used to evaluate the increase in strength owing to the incorporationof different volumes of steel fibers. Thus, the strength index plottedin Fig. 6is the relative increase in strength of steel fiber –reinforced concretes compared to their unreinforced counterparts Strength index ð%Þ¼½ ð fcf−f0cpÞ=f0cp/C138ð 1Þ where f0 cf= average compressive strength of fiber-reinforced alkali-activated concrete; and f0cp= average compressive strength of plain alkali-activated concrete. Fig.6(a)presents the compressive strength increase due to the addition of 2% steel fibers to alkali-activated slag concretes madewith different fly ash replacement percentages and a 14-M-SHsolution. For 1- and 7-day samples, the strength increased inthe range of 20% –39% and 36% –52%, respectively, with highest value associated with 25% fly ash concretes (G75-14M-2.0SF).Among mixes with 0% and 25% fly ash replacement (G100-14M-0.0SF and G75-14M-0.0SF), the latter was weaker at 1 and7 days [Fig. 5(a)]. Thus, the addition of steel fiber contributed more to its strength, owing to a less dense matrix at an early age. Similar conclusions were noted for samples after 28 days, with an increase(b) (a) (c) (d) 01020304050607080 G50-14M- 0.0SF G75-14M- 0.0SF G100-14M- 0.0SF G50-14M- 2.0SF G75-14M- 2.0SF G100-14M- 2.0SFCompressive Strength (MPa) Mix Designation1-day 7-day 28-day SF = 0.0%SF=2.0%G75-8M- 0.0SF G75-8M- 0.5SF G75-8M- 1.0SF G75-8M- 1.5SF G75-8M- 2.0SF G75-8M- 3.0SFCompressive Strength (MPa) Mix Designation1-day 7-day 28-dayG75-14M- 0.0SF G75-14M- 0.5SF G75-14M- 1.0SF G75-14M- 1.5SF G75-14M- 2.0SFCompressive Strength (MPa) Mix Designation1-day 7-day 28-dayG75-8M- 0.0SF G75-10M- 0.0SF G75-14M- 0.0SF G75-8M- 2.0SF G75-10M- 2.0SF G75-14M- 2.0SFCompressive Strength (MPa) Mix Designation1-day 7-day 28-day Fig. 5. Strength development of alkali-activated concrete made with (a) 14-M-SH solution and different fly ash content and SF ratios; (b) 8-M-SH solution and different SF ratios; (c) 14-M-SH solution and different SF ratios; and (d) different molarities of SH solution. © ASCE 04019223-6 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. in strength between 23% and 30%. This shows that the effect of steel fibers was less significant at a later age of 28 days due to a densified alkali-activated structure, reduced porosity, and en- hanced binding ability of the aluminosilicate hydrate and calcium silicate hydrate gels, as discussed later in the microstructure analy- sis sections. As shown in Fig. 6(b), the addition of steel fibers led to an increase in the load carrying capacity of 8M-SH concretes. An increase between 7% and 22% was observed after 1 day when 0.5% –3% steel fibers were added (G75-8M-0.5SF, G75-8M- 1.0SF, G75-8M-1.5SF, G75-8M-2.0SF, and G75-8M-3.0SF). At 7 and 28 days, up to 25% improvement in compressive strength was recorded with 3% steel fibers (G75-8M-3.0SF). At these ages, the effect was much less apparent with low steel fiber volume fractions of 0.5% and 1% (G75-8M-0.5SF and G75-8M-1.0SF). This shows that the incorporation of steel fibers could significantly enhance the compressive strength of 25∶75fly ash ∶slag blended alkali-activated concretes if more than 1.5% by volume was added to the mix. This is mainly due to the ability of steel fibers to restrain the expansion of cracks, reduce the stress concentration within cracks, and delay the growth of cracks by redirecting them ( Afroughsabet and Ozbakkaloglu 2015 ). Analogous findings have been reported by Aydın and Baradan ( 2013 ). Similar observations can be noted for 14-M-SH alkali-activated concrete in Fig. 6(c). At 1 and 7 days of age, the addition of steel fibers enhanced the compressive strength by 7% –39% and 1% – 46%, respectively. It was less effective at 28 days with an increase of 13% –30% owing to a superior alkali-activated matrix. Fig. 6(d) examines the strength increase in 75∶25slag∶fly ash alkali-activated concrete made with 2.0% steel fiber volume fraction and different molarities of SH solution (G75-8M-2.0SF,G75-10M-2.0SF, and G75-14M-2.0SF). It is clear that incorporat- ing steel fibers could further improve the compressive strength of samples with higher SH molar concentration. Nonetheless, its ef-fect was more significant at an early age of 1 day. The respectiveincrease of G75-8M-2.0SF, G75-10M-2.0SF, and G75-14M-2.0SF was 12.6%, 21.8%, and 32.8% after 1 day and 7.0%, 17.9%, and 20.4% after 7 days. This is possibly due to the increase in activationreaction efficiency at a later age of 28 days to the extent that theaddition of steel fibers caused limited improvement to compressive strength. While 1- and 7-day results are novel, the 28-day findings are not unexpected. Indirect Tensile Strength The 28-day tensile strength of steel fiber –reinforced alkali- activated slag concrete incorporating fly ash was measured indirectly as the tensile splitting strength ( f sp). The effect of fly ash replacement and steel fiber addition on fspis shown in Fig.7(a). Values ranged between 3.53 and 3.81 MPa for plain sam- ples and between 4.16 and 4.64 MPa for steel fiber –reinforced counterparts (SF ¼2.0%). The substitution of slag with 25% fly ash resulted in a slight increase in the 28-day tensile strength.Apparently, the mechanical performance was enhanced with an intermix of sodium aluminosilicate hydrate (N ─A─S─H) and cal- cium aluminosilicate hydrate (C ─A─S─H) gels from the alkali ac- tivation of fly ash and slag. This intermix has been reported todensify the concrete matrix and increase the strength ( Brough and Atkinson 2002 ). The addition of steel fibers led to a significant increase in tensile strength, with the highest increase of 22% rep-resenting that of mix with 75% slag and 25% fly ash. It is clear thatwhile steel fibers could enhance the mechanical performance of (b) (a) (d) (c)01020304050 0 1 02 03 04 05 06 0Strength Index (%) Fly Ash Content (%)1-day 7-day 28-day01020304050 0.0 1.0 2.0 3.0 4.0Strength Index (%) Steel Fiber Volume (%)1-day 7-day 28-day0.0 0.5 1.0 1.5 2.0 2.5Strength Index (%) Steel Fiber Volume (%)1-day 7-day 28-day6 8 10 12 14 16Strength Index (%) Molarity of SH solution (M)1-day 7-day 28-day Fig. 6. Strength index of alkali-activated concretes over time: (a) 2.0% SF with different slag-to-fly ash ratio and 14-M-SH solution; (b) 8-M-SH; (c) 14-M-SH; and (d) 2.0% SF with different molarities of SH. © ASCE 04019223-7 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. alkali-activated slag concrete, they were most effective in mixes with 25% fly ash replacement. It is also worth noting that speci- mens with 50% fly ash were somewhat weaker than the others owing to a slower activation chemical reaction of fly ash at ambienttemperatures ( Palomo et al. 1999 ). Fig.7(b) presents the tensile splitting strength as a function of steel fiber volume fraction in 8-, 10-, and 14-M-SH concretes (75%slag and 25% fly ash). It can be seen that the 14-M-SH concretesamples were consistently superior to counterparts made with alower molarity of SH. This is mainly due to the enhanced perfor-mance of the activation matrix made with a higher SH molar con-centration, which is also reflected in the compressive strengthresults. Specimens with 8 M SH increased by 14%, 17%, 24%,29%, and 31% when 0.5%, 1%, 1.5%, 2%, and 3% steel fibers wereadded, respectively. Clearly, the contribution of steel fibers to ten-sile strength was more significant at higher fiber dosages. On theother hand, 14-M-SH counterparts increased by 13%, 16%, 17%,and 22% for similar steel fibers volume fractions. The effect of steelfibers on splitting tensile strength seemed to be similar regardless ofthe molarity of SH solution. It is also worth noting that the tensilesplitting strength was approximately 5% –8% of the compressive strength. A previous study reported similar results due to the incor-poration of steel fibers in fly ash –based geopolymer concrete (Islam et al. 2017 ). The tensile splitting strength of concrete is typically estimated using codified and previously published equations, including thosein AS3600 ( Standards Australia 2009 ), ACI Commitee 318 ( ACI 2014 ), CEB-FIP ( 1990 ), Choi and Yuan ( 2005 ), Xu and Shi ( 2009 ), Perumal ( 2015 ), Ahmed and Shah ( 1985 ), and ACI Committee 363 ( ACI 1992 ). Such equations, shown in Table 3, were applied to determine the tensile splitting strengths of plain and steelfiber–reinforced alkali-activated concretes incorporating slag and fly ash using the average 28-day compressive strength. The asso-ciated plots are shown in Fig. 8(a). The equations provided by ACI Commitee 318 ( ACI 2014 ) and Ahmed and Shah ( 1985 ) are suitable for 10-M-SH and 14-M-SH steel fiber –reinforced con- crete mixes but slightly overestimate the tensile strength for plaincounterparts owing to its ceramic-like brittleness. Apparently, theaddition of steel fibers modified the tensile behavior of these mixesand rendered them more similar to that of conventional concrete.Furthermore, the AS3600 ( Standards Australia 2009 ) equation could be employed to predict the performance of plain 8-M-SHconcretes only. Nevertheless, it is should be noted that the ACICommitee 318 ( ACI 2014 ) equation could be used for all alkali- activated concrete mixes produced in this work if such a codifiedequation were to be multiplied by a factor α. To determine the value ofα, the regression equations were obtained based on plotted trend lines, as shown in Fig. 8(b). Each regression equation can accu- rately predict the tensile splitting strength of alkali-activated slagconcretes produced in this work with R 2>0.90. Thus, a modified version of the ACI Commitee 318 ( ACI 2014 ) equation is proposed in the form of Eq. ( 2), where αis 1 for 10-M-SH and 14-M-SH samples and 0.75 for 8-M-SH specimens: fsp¼0.56αffiffiffiffiffi f0cp ð2Þ The tensile splitting behavior of 8-, 10-, and 14-M-SH alkali- activated concrete was significantly improved due to the incorpo- ration of steel fibers. The postcracking behavior was also affected as shown in Fig. 9. The tensile strength development in plain con- cretes could be associated to the formation of alkali-activationproducts in the matrix with limited postcracking performance. (b) (a)0123456 0 1 02 03 04 05 06 0Tensile Splitting Strength (MPa) Fly Ash Content (%)SF=0% SF=2%0.0 0.5 1.0 1.5 2.0 2.5 3.0Splitting Tensile Strength (MPa) Steel Fiber Volume (%)8M 10M 14M Fig. 7. Splitting tensile strength of alkali-activated slag concrete with different: (a) fly ash content; and (b) molar concentration of SH solution. Table 3. Published equations relating mechanical properties Reference Tensile splitting strength ( fsp) Flexural strength ( fr) Modulus of elasticity ( Ec) ACI Commitee 318 ( ACI 2014 ) 0.56f00.5c 0.62f00.5c 0.043w1.5f00.5c AS3600 ( Standards Australia 2009 ) 0.36f00.5c 0.60f00.5c 0.024w1.5ðf00.5cþ0.12Þ CEB-FIP ( 1990 ) 0.3f00.67c 0.81f00.5c 9,979 .4f00.33c Choi and Yuan ( 2005 ) 0.6f00.5c —— Xu and Shi ( 2009 ) 0.21f00.83c 0.39f00.59c — Perumal ( 2015 ) 0.188f00.84c 0.259f00.843c — Ahmed and Shah ( 1985 ) 0.46f00.55c 0.44f00.5c 3.38w2.5f00.65c Dilli et al. ( 2015 ) —— 10,750 ðf0c=10Þ0.5 Mendis ( 2003 ) —— 0.043w1.5ð1:1−0.002f0cÞf00.5c ACI Committee 363 ( ACI 1992 ) 0.59f00.5c 0.94f00.5c 3,320 f00.5cþ6,900 Note: w= density of concrete (kg =m3); and f0c= compressive strength (MPa). © ASCE 04019223-8 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. Nevertheless, the tensile splitting strength and the postcracking performance were enhanced with higher steel fiber volume frac- tions. It is believed that the incorporation of ever more fibers increased the energy requirements for crack propagation and bridged microcracks. Hence, the fracture process seemed to bearrested or retarded, producing an alkali-activated concrete with higher splitting tensile load capacity and improved postcrackingperformance. Flexural Strength The flexural strength of alkali-activated slag concretes made with different fly ash replacement percentages and steel fiber volumefractions is shown in Fig. 10(a) . Values ranged from 3.99 to 4.31 MPa for plain concretes and from 4.72 to 5.27 MPa for steelfiber–reinforced equivalents. A similar trend to tensile splitting strength is noted, whereby the highest flexural strength was thatof alkali-activated slag concrete mix with 25% fly ash replacement.Also, the addition of steel fibers enhanced the flexural performanceby up to 22%. Fig.10(b) presents the flexural strength ( f r) of plain and steel fiber–reinforced alkali-activated concrete after 28 days. An increase in flexural strength can be noted upon the incorporation of steelfibers. For 8-M-SH samples, a volumetric addition of fiber up to1.5% could enhance the flexural capacity by 16%. However, theeffect of fiber addition was more pronounced at higher dosagesof 2% and 3%. The associated increase in strength was 23% and25%, respectively. Apparently, crack propagation through alkali-activated concrete was retarded or arrested due to the incorporationof 2% –3% steel fibers. Similar results were presented for steel fiber–reinforced fly ash-based geopolymer concretes ( Bernal et al. 2010 ). An increase of 19% was also reported for 10-M-SH con- cretes upon the addition of 2.0% steel fiber volume fraction. In con-trast, a near-linear relation between the volume fraction of steelfibers and flexural strength increase was noted for 14-M-SH sam-ples. On average, the flexural strength increased by 7% for every0.5% volume fraction of steel fiber added. A comparison betweenalkali-activated concretes with constant steel fiber content and dif-ferent molarity of SH (8M-2.0SF, 10M-2.0SF, and 14M-2.0SF)showed that a higher molarity resulted in further increases in flexu-ral strength. Others reported that the compression zone of concreteprisms placed under flexural loading played an insignificant role infailure ( Gencel et al. 2011 ). On the other hand, the tension zone was critical to the development of flexural strength. It is thus believed that the addition of steel fibers could allow alkali-activated concrete to withstand more crack generation in the tension zone and, as such,result in a higher flexural strength. Codified and past literature equations, shown in Table 3, from AS3600 ( Standards Australia 2009 ), ACI Commitee 318 ( ACI 2014 ), Ahmed and Shah ( 1985 ), Xu and Shi ( 2009 ), Perumal (2015 )(CEB-FIP 1990 ), and ACI Committee 363 ( ACI 1992 ) were employed to predict the flexural strength of steel fiber –reinforced alkali-activated concrete. The results displayed in Fig. 11(a) show that the equations of ACI Commitee 318 ( ACI 2014 ) and ACI Committee 363 ( ACI 1992 ) provide the closest prediction with rea- sonable accuracy. Plain alkali-activated concrete samples were gen-erally lower than predictions provided by most codified equationsdue to its ceramic-like brittle behavior. However, it is possible toapply a modification factor, α, to the ACI Commitee 318 ( ACI 2014 ) equation based on the molarity of SH. To obtain the value ofα, the plots of linear regression analyses equations were devel- oped, as shown in Fig. 11(b) . Accordingly, Eq. ( 3) is proposed as a modified version of the ACI Commitee 318 ( ACI 2014 ) equa- tion to accurately predict ( R 2>0.90) the flexural strength, with α being 0.75 for 8-M-SH and 1 for 10- and 14-M-SH alkali-activatedconcretes: f r¼0.62αffiffiffiffiffi f0cp ð3Þ(a) (b) 234567 6.5 7.0 7.5 8.0 8.5Splitting Tensile Strength (MPa) Compressive Strength, √f'c (MPa) 8M 10M 14M AS3600 (2009) ACI Committee 318 (2014) CEB-FIP (1990) Choi and Yuan (2005) Xu and Shi (2009) Perumal (2015) Ahmed and Shah (1985) ACI Committee 363 (1992) fsp= 0.42 √f'cfsp= 0.55 √f'cfsp= 0.56 √f'c6.5 7.0 7.5 8.0 8.5Splitting Tensile Strength (MPa) Compressive Strength, √f'c (MPa)8M 10M 14M Fig. 8. Relationship between splitting tensile and compressive strength: (a) in comparison to literature and codified equations; and(b) using linear regression analysis. 050100150200Applied Load (kN) Loading Time (s)0.00% vf 0.50% vf 1.00% vf 2.00% vf Fig. 9. Development of splitting tensile load over time of 14-M alkali- activated slag concretes with 25% fly ash. © ASCE 04019223-9 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. Modulus of Elasticity The modulus of elasticity ( Ec) is the ability of a material to sustain induced stress while increasing the strain within the elastic limit.Table 4shows the experimentally determined modulus of elas- ticity for produced alkali-activated slag concretes after 28 days. The replacement of slag by 25% fly ash resulted in higher E c.H o w - ever, further replacement up to 50% reduced the modulus. Theseresults are well aligned with those of other mechanical properties.The experimental modulus of steel fiber –reinforced alkali-activated concrete made with 8-M-SH solution could reach 28.5 GPa com- pared to 25.5 GPa for nonreinforced counterparts, representing an 11.8% increase due to the addition of 3% steel fiber by volume. However, the modulus was almost unchanged when less than 1.5% steel fiber was added. An increase in the modulus of 9.2% was also noted for 10-M-SH alkali-activated concrete upon the incorporation of 2% steel fibers by volume. This shows that steel fibers can enhance the elastic behavior of alkali-activated concrete.For 14-M-SH, 100% slag concrete samples, the modulus increased by 14.4% when 2% fiber volume fraction was incorporated into the mix. The effect of adding steel fibers seemed to be more significant in alkali-activated slag concrete with 25% fly ash and 14-M-SH solution. Fig.12(a) shows a comparison between the experimental results and prediction equations provided by AS3600 ( Standards Australia 2009 ), ACI Commitee 318 ( ACI 2014 ), Dilli et al. ( 2015 ), Ahmed and Shah ( 1985 ), ACI Committee 363 ( ACI 1992 ), Mendis ( 2003 ), and CEB-FIP ( 1990 ). Most equations seem to overestimate the modulus of elasticity of 8-M-SH concretes, except for that given(b) (a)0123456 0 1 02 03 04 05 06 0Flexural Strength (MPa) Fly Ash Content (%)SF=0% SF=2%0.0 1.0 2.0 3.0 4.0Flexural Strength (MPa) Steel Fiber Volume (%)8M 10M 14M Fig. 10. Flexural strength of alkali-activated slag concrete with different: (a) fly ash contents; and (b) SH molar concentrations. (a) (b) 012345678910 6.5 7.0 7.5 8.0 8.5Flexural Strength (MPa) Compressive Strength, √f'c (MPa) 8M 10M 14M AS3600 (2009) ACI Committee 318 (2014) Ahmed and Shah (1985) Xu and Shi (2009) Perumal (2015) CEB-FIP (1990) ACI Committee 363 (1992) fr= 0.46 √f'c fr= 0.61 √f'c fr= 0.62 √f'c 6.5 7.0 7.5 8.0 8.5Flexural Strength (MPa) Compressive Strength, √f'c (MPa)8M 10M 14M Fig. 11. Relationship between flexural and compressive strength: (a) in comparison to literature and codified equations; and (b) using linearregression analysis.Table 4. Modulus of elasticity of alkali-activated concrete samples Mix No. Mix designationModulus of elasticity Experimental PredictedPercentage increaseaPercentage differenceb 1 G50-14M-0.0SF 33.5 34.5 — −3.0 2 G50-14M-2.0SF 37.8 38.2 12.8 −1.1 3 G75-14M-0.0SF 36.8 36.8 — 0.0 4 G75-14M-2.0SF 42.1 42.0 14.4 0.2 5 G100-14M-0.0SF 35.8 36.2 — −1.0 6 G100-14M-2.0SF 40.5 40.8 13.1 −0.7 7 G75-8M-0.0SF 25.5 25.5 — 0.0 8 G75-8M-0.5SF 25.6 25.6 0.4 0.0 9 G75-8M-1.0SF 25.7 25.7 0.8 0.0 10 G75-8M-1.5SF 26.5 26.4 3.9 0.411 G75-8M-2.0SF 27.6 27.7 8.2 −0.4 12 G75-8M-3.0SF 28.5 28.5 11.8 0.0 13 G75-10M-0.0SF 35.7 35.7 — 0.0 14 G75-10M-2.0SF 39.0 38.8 9.2 0.5 15 G75-14M-0.5SF 39.0 39.1 6.0 −0.3 16 G75-14M-1.0SF 40.0 40.2 8.7 −0.5 17 G75-14M-1.5SF 40.5 40.3 10.1 0.5 aPercentage increase of steel fiber –reinforced concrete with respect to plain counterparts. bPercentage difference between experimental and predicted modulus of elasticity values. © ASCE 04019223-10 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. by Dilli et al. ( 2015 ) and AS3600 ( Standards Australia 2009 ). This is in agreement with work reported by Pan et al. ( 2011 ), whereby the modulus of plain fly ash –based geopolymer concrete was 27% lower than conventional cement-based counterparts. To accurately predict the modulus of elasticity from the compressive strength,linear regression analysis was employed. The regression equationsand trend lines are shown in Fig. 12(b) . Based on these equations, it is possible to apply a modification factor, α, of 0.75 to the ACI Commitee 318 ( ACI 2014 ) equation to be suitable for 8-M-SHmixes. The case was different with 10- and 14-M-SH samples. The equation proposed by ACI Commitee 318 ( ACI 2014 ) could accurately predict E c(α¼1). It is thus proposed that Eq. ( 4)b e employed to estimate the modulus of elasticity, with αbeing 0.75 for 8-M-SH and 1 for 10- and 14-M-SH alkali-activated concretes: Ec¼5:056αffiffiffiffiffi f0cp ð4Þ Because concrete is a heterogeneous material, its elastic modu- lus depends on its different components. Although the compositionis dissimilar to conventional concrete, the ACI Commitee 318 ( ACI 2014 ) equation may be used with some modification for 8-M-SH alkali-activated slag concrete. Table 4presents the percentage dif- ference between moduli obtained from the prediction equation andexperimental testing. Besides the correlation coefficient exceeding0.97, the difference ranged between −3%andþ1%. This indicates an excellent performance of the regression model, with the pre-dicted modulus being similar to the experimental values. DSC Analysis The microstructure of alkali-activated slag concretes with 0% –50% fly ash replacement was analyzed at the ages of 7 and 28 days tostudy the development of reaction products. Such a comparativeanalysis has not yet been performed and is critical to correlating the microstructure to the mechanical performance. The DSC ther- mograms of 7-day alkali-activated slag concretes incorporating upto 50% fly ash are shown in Fig. 13(a) . A broad endotherm between 110°C and 130°C is detected in samples made with equal propor-tions of slag and fly ash (G50). It is associated with the decompo-sition of alkali-activation reaction products as combined water insodium aluminosilicate hydrate (N ─A─S─H) gel ( Colella 1999 ) or as interstitial water ( Perera et al. 2006 ). At lower fly ash replace- ments (0% and 25%), the peak shifted toward higher temperatures,became narrower, and increased in intensity. This providesevidence of the formation of calcium aluminosilicate hydrate(C─A─S─H) gel ( Al-Majidi et al. 2016 ;Nath and Kumar 2013 ) and explains the higher 7-day strength of alkali-activated concretesmade with 100% slag. Fig. 13(b) presents the DSC thermograms of concretes after 28 days. The endotherm in the range of 120°C –140°C in concrete made with 25% fly ash was more intense than samples with 0% and50% fly ash. Compared to 7-day equivalents, these peaks had ahigher intensity and narrower shape, indicating the progressionof the activation reaction and the formation of C ─A─S─H gel (Al-Majidi et al. 2016 ;Nath and Kumar 2013 ). Nevertheless, N─A─S─H gel could also be detected within the same endotherm (Colella 1999 ). Clearly, an intermix of reaction products between N─A─S─H and C ─A─S─H existed at a later age of 28 days,(a) (b) 01020304050 0 1 02 03 04 05 0Predicted Ec (GPa) Experimental Ec (GPa)ACI Committee 318 (2014) ACI Committee 363 (1992) Dilli, Atahan and Sengül (2015) Ahmed and Shah (1985) Mendis (2003) AS3600 (2009) CEB-FIP (1990) Ec= 3.8027 √f'cEc= 5.1063 √f'c Ec= 5.0913 √f'c6.0 6.5 7.0 7.5 8.0 8.5Modulus of Elasticity (GPa) Compressive Strength0.5(MPa0.5)8M 10M 14M Fig. 12. (a) Predicted versus experimental modulus of elasticity; and (b) relationship between modulus of elasticity and compressivestrength using linear regression analysis. (b) (a)50 100 150 200 250 300 350 400 450 500Heat Flow (mW) Temperature (°C)G50 G75 G100 50 100 150 200 250 300 350 400 450 500Heat Flow (mW) Temperature (°C)G50 G75 G100 Fig. 13. DSC thermograms of (a) 7-day; and (b) 28-day alkali-activated concretes. © ASCE 04019223-11 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. providing superior mechanical performance. Also, a broad endo- thermic shoulder was identified in 100% slag specimens at around400°C corresponding to the decomposition of a hydrotalcite-typephase ( Ben Haha et al. 2011a ,b). It has been reported that hydro- talcite may have a higher pore-filling ability and, thus, could pro-duce a denser and stronger concrete structure ( Ben Haha et al. 2011a ,b). Yet, the intensity of this band decreased with more than 25% fly ash replacement, confirming the reduced compressive strength. FTIR Analysis Fig. 14(a) presents the FTIR spectra of alkali-activated slag con- crete with different fly ash replacement percentages. Samples with 100% slag show a broad transmittance band in the range of600–700cm −1, signifying the stretching vibrations of Al ─O bonds in the AlO 4groups, which could possibly indicate the formation of C─A─S─H gel ( Jiao et al. 2018 ;Puertas and Torres-Carrasco 2014 ). The higher intensity of this band in 100% slag concrete compared to counterparts blended with fly ash explains the higher7-day strength of the former concrete. This is in good agreementwith DSC analysis results. Bending vibrations of O ─Ha t 1,600 –1,700 cm −1and stretching vibrations of H ─OH groups at 3,200 –3,600 cm−1(Ismail et al. 2014 ) decreased with higher slag replacement by fly ash, indicating a lower degree of reaction. As aresult, their mechanical performance was inferior to 100% slagconcretes. Fig.14(b) presents the FTIR spectra of concretes after 28 days. Compared to 7-day samples, the intensity of bands increasedthroughout the spectra. This signifies that larger quantities of reac-tion products had developed during subsequent curing. A narrow-shaped asymmetric vibration of Si ─O─Si is detected between 895 and 1,050 cm −1in the 100% slag concretes, indicating the formation of an ordered calcium-based structure ( Nath 2018 ). In comparison, concrete mixes with 25% fly ash noted a similar vi-bration of Si ─O─Si but at higher wavelengths of 900–1,100 cm −1, centered at 946cm−1. Similar findings have been reported else- where ( Al Bakri et al. 2012 ;Ismail et al. 2014 ;Rosas et al. 2014 ; Škvára et al. 2009 ). This shift in wavelength is due to the simulta- neous activation of fly ash and slag, creating a more cross-linkedgel structure, comprising an intermix of reaction products ofN─A─S─H and C ─A─S─H. As a result, the activation reaction accelerated and the compressive strength increased. For concreteswith 50% fly ash, the Si ─O─Si vibration became broader and fur- ther shifted toward higher wavelengths. Ismail et al. ( 2014 ) ex- plained this higher wavelength as being due to the presence of unreacted fly ash components. Consequently, its mechanical per- formance was inferior to its counterparts with low fly ash content(0% and 25%). In addition, the broad band at 1,400 cm −1repre- sents O ─C─Oi nC O2− 3(Ismail et al. 2014 ;Samantasinghar and Singh 2018 ), which indicates the formation of carbonation products from the reaction of alkali-activated slag with CO 2in the air over the duration of curing ( Zhang et al. 2017 ). It should also be noted that water peaks ( 1,600 –1,700 cm−1and3,200 –3,600 cm−1) were highest in the sample with 75% slag and 25% fly ash. Accordingly, its mechanical properties were superior to 50% and 100% slag sam- ples. This is in agreement with the findings of DSC analysis. SEM Analysis The morphology of alkali-activated slag concretes with different fly ash replacements was characterized using SEM with EDX. Fig. 15(a) shows the micrograph of 7-day concrete samples with 100% slag. A dense matrix can be noted by randomly distributedfibrous spherical shapes. Microcracks with an average width of5μm spread on the sphere ’s surface. Work by Beglarigale and Yazici ( 2014 ), El-Hassan et al. ( 2018 ), and Yaz ıcı(2012 ) confirmed that such spherical shapes are associated with calcium aluminosil-icate hydrate (C ─A─S─H) gel. The effect of replacing 25% slag by fly ash on the microstructure is shown in Fig. 15(b) . While the spherical shapes are still evident in the matrix, they are smaller.EDX analysis shows higher Ca/Si and Si/Al ratios with largerspheres, which indicate more calcium hydrates and silicate phases,respectively. With more phases being produced, the matrix densi-fies, creating a higher-strength concrete ( Brough and Atkinson 2002 ). Further replacement of slag by fly ash up to 50% created a much different microstructure, as seen in Fig. 15(c) . Microgaps are detected in a matrix intermixed with angular slag and sphericalfly ash particles, creating a more porous microstructure and weakerconcrete. EDX analysis of this concrete (50% fly ash replacement)in Fig. 15(d) highlights the formation of a compound with calcium, silicon, sodium, aluminum, and oxygen, signifying an intermixof C─A─S─H and N ─A─S─H gels. Yet, it should be noted that the latter seems to dominate the former given the low calciumcontent. Accordingly, while C ─A─S─H is known to be a strength- contributing product, the weaker aluminosilicate compound gov-erned the mechanical performance of alkali-activated concretemixes with 50% fly ash. Fig.16shows the micrographs of alkali-activated slag concrete samples after 28 days. In general, the microstructure did notsignificantly change between 7 and 28 days. Fig. 16(a) presents a fibrous spherical shape similar to that noted at 7 days. In com-parison, concretes with 25% fly ash showed a smaller sphericalshape covered in white dots, which may possibly be developed N─A─S─H gel particles, as shown in Fig. 16(b) . The 28-day concrete sample with 50% fly ash replacement is presented in(b) (a)500 1000 1500 2000 2500 3000 3500 4000 Transmittance (%) Wavelength (cm-1)G50 G75 G100 500 1000 1500 2000 2500 3000 3500 4000 Transmittance (%) Wavelength (cm-1)G50 G75 G100 Fig. 14. FTIR spectra of (a) 7-day; and (b) 28-day alkali-activated concretes. © ASCE 04019223-12 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. Fig. 15. SEM micrograph of 7-day alkali-activated concretes with (a) 100% slag; (b) 75% slag and 25% fly ash; (c) 50% slag and 50% fly ash; and (d) EDX spectrum of 50∶50slag∶fly ash sample. Fig. 16. SEM micrograph of 28-day alkali-activated concretes with (a) 100% slag; (b) 75% slag and 25% fly ash; (c) 50% slag and 50% fly ash; and (d) EDX spectrum of 50∶50slag∶fly ash sample. © ASCE 04019223-13 J. Mater. Civ. Eng. J. Mater. Civ. Eng., 2019, 31(10): 04019223 Downloaded from ascelibrary.org by United Arab Emirates University on 10/29/19. Copyright ASCE. For personal use only; all rights reserved. Fig. 16(c) . Most slag and fly ash particles are integrated into the matrix, producing an intermix of activation and hydration products,as noted in the EDX analysis in Fig. 16(d) . A comparison between the EDX spectra of 7- and 28-day alkali-activated slag concrete with 50% fly ash replacement was carried out. The Ca/Si andSi/Al ratios increased from 0.29 and 2.50 to 0.56 and 3.40, re-spectively, during subsequent curing. This indicates further de-velopment of calcium hydrates and silicate phases, respectively,providing evidence of the 23.8% increase in compressive strength from 7 to 28 days. Nevertheless, some microcracks still exist in the 28-day microstructure of 50∶50slag∶fly ash concrete due to solu- tion evaporation and slow polymerization reaction of fly ash inambient conditions. This explains its inferior compressive strengthat 28 days. Conclusions The effects of slag replacement by fly ash, sodium hydroxidemolarity, and steel fiber incorporation on the rheology, mechanicalperformance, and microstructure characteristics of alkali-activatedslag concrete containing dune sand were investigated. What followsare the key findings of the study: 1. The replacement of slag by fly ash led to an increase in work- ability due to its more spherical shape, causing less interparticlefriction. The molarity of SH solution was inversely proportionalto workability. An average 23% decrease in slump was notedupon increasing the molarity of SH from 8 to 14 M, owingto a more viscous alkaline activator solution. Further, for every0.5% of steel fiber added to the concrete, an average 11-mm reduction in slump was noted. Concrete samples made with 14 M SH could not incorporate more than 2% steel fiber volumefraction. 2. At 1 and 7 days, the addition of fly ash resulted in a decrease in mechanical properties, i.e., modulus of elasticity, compressive,tensile, and flexural strength, owing to the slow activation of flyash in ambient conditions. After 28 days, alkali-activated slagconcrete samples with 25% fly ash had superior strength, withan intermix of activation and hydration products. Higher molarconcentration of sodium hydroxide resulted in better mechanicalperformance. This was mainly due to the availability of moresodium ions that could balance the charges, thereby formingaluminosilicate networks in the alkali-activated matrix. 3. The compressive strength rapidly developed, with, on average, 78% and 89% of the 28-day compressive strength being ob-tained at the ages of 1 and 7 days. The addition of steel fiberscould increase the compressive strength by up to 39% and 52%for 1- and 7-day samples. Its effect, however, was less signifi-cant (23% –30% increase) for 28-day samples. With a densified alkali-activated structure and enhanced binding ability of thealuminosilicate hydrate and calcium silicate hydrate gels, thealkali-activated matrix gained sufficient strength to the extentthat steel fibers caused limited improvement. 4. The tensile splitting and flexural strength could increase by up to 31% and 25%, respectively, upon the addition of steel fibers,with more contribution at higher dosages. Its effect was morepronounced in concretes made with 25% fly ash replacement,especially at an early age and 8-M-SH concrete samples witha weaker alkali-activated matrix. Clearly, adding steel fibersincreased the energy requirements for crack propagation andbridged microcracks in the alkali-activated concretes, resultingin a delayed fracture process and improved postcracking perfor-mance. The tensile and flexural behavior of these steel fiber – reinforced mixes was similar to that of conventional concrete.5. Experimental results of the modulus of elasticity of alkali- activated concretes showed a gradual increase as more steel fibers were incorporated. The increase could reach up to 12%,10%, and 14% for 8-M-, 10-M-, and 14-M-SH samples, respec-tively. The effect of adding steel fibers seemed to be more sig-nificant in alkali-activated slag concrete with 25% fly ash and14-M-SH solution. 6. Published and codified equations relating the modulus of elasticity and the tensile splitting and flexural strength to compressive strength could be suitable for predicting, withreasonable accuracy, the mechanical properties of steel fiber – reinforced alkali-activated slag concretes. In fact, ACI Commit-tee 318 ( ACI 2014 ) equations for f r,fsp, and Eccould be employed as-is for 10- and 14-M-SH alkali-activated concretes,regardless of the steel fiber volume fraction and slag:fly ash ra-tio. However, a modification factor was introduced to ACI 318prediction equations for 8-M-SH samples. This factor, α, was found to be 0.75. 7. Microstructure analysis was performed on plain alkali-activated slag concretes with different fly ash replacements using DSC, FTIR, and SEM. At 7 days, calcium aluminosilicatehydrate (C ─A─S─H) gel was found to be the dominant, strength-contributing reacti on product. Sodium aluminosili- cate (N ─A─S─H) gel was reported, but in small quantities due to the slower activation reaction of fly ash particles in am-bient conditions. Water peaks in FTIR and endothermic bands inDSC were more intense with higher slag content, indicating ahigher degree of activation reaction. This provides evidence ofthe superior performance of 100% slag at 7 days. After 28 days,N─A─S─H further developed and was identified using SEM as white nanosized dots on the surface of the activated matrix,being most prevalent in samples with 25% fly ash replacement.As such, these samples ( 75∶25slag∶fly ash) had superior perfor- mance at 28 days of age. Based on the current study, the mechanical performance of alkali-activated slag concrete cured in ambient conditions improvedwith the replacement of slag by 25% fly ash and the addition ofsteel fibers. ACI Committee 318 equations could be employed,with a newly proposed modification factor, for plain and steelfiber–reinforced alkali-activated slag/fly ash blended concretes. A more in-depth study of the pore structure, fiber –matrix interfacial properties, and mechanism of fiber reinforcement in alkali-activated concretes should be conducted in the future. Acknowledgments The authors are grateful to the United Arab Emirates University (UAEU), Al Ain, UAE, for the financial support through Grants 31N282 and 31N305. The assistance of students and laboratorystaff at UAEU is gratefully acknowledged.

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 04 Issue: 08 | Aug -2017 www.irjet.net p -ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 18 EXPERIMENTAL STUDY ON MECHANICAL PROPERTIES OF FLY ASH AND GGBS BASED GEOPOLYMER CONCRETE Ajay Takekar1, Prof.G.R.Patil2 1 PG Student, Department of Civil Engineering, JSPM’s Rajarshi Shahu college of Engg., Pune, MH, India 2 Professor, Department of Civil Engineering, JSPM’s Rajarshi Shahu college of Engg., Pune, MH, India --------------------------------------------------------------------- ***--------------------------------------------------------------------- Abstract - The present investigation aims at studying the mechanical properties of fly ash and GGBS based geopolymer concrete. In this study, fly ash was replaced at different levels (0%, 25%, 50%, 75% and 100%) by GGBS. Sodium hydroxide (14 M) and Sodium silicate solution is used as alkaline activator in 1:1 ratio. M -25 grade of geopolymer concrete is tested for mechan ical properties Viz, compressive strength, split tensile strength and flexural strength at 3,7 and 28 day and compared with normal OPC M -25 grade concrete and also the effect of ambient curing and oven curing is studied. The result shows that replacement o f fly ash by GGBS eliminates heat curing of geopolymer concrete. Geopolymer concrete shows better result than normal conventional concrete. Rate of gain of strength of geopolymer concrete is high at early stage however geopolymer concrete is more advantage ous, economical and ecofriendly. This experimental investigation is for research purpose for strength properties of geopolymer concrete using fly ash and ground granulated furnace slag (GGBS). Key Words : Geopolymer Concrete, Fly ash, Ground granulated fur nace slag (GGBS), Alkaline solutions, oven curing, ambient curing, compressive strength, Split tensile strength, Flexural strength. 1. INTRODUCTION The major problem the world is facing today is environmental pollution. In the construction industry main ly the production of Portland cement causes emission of pollutants results in environmental pollution. It is widely known that the production of Portland cement consumes considerable energy and at the same time contributes a large volume of CO2 to the atmo sphere. However, Portland cement is still the main binder in concrete construction prompting a search for more environmentally friendly materials .One possible alternative is the use of alkali -activated binder using industrial by -products containing sil icate materials. Davidovits proposed that an alkaline liquid which can react with the silicon (Si) and aluminum (Al) in a source material of geological origin or industrial by product can be used to produce binders. Since chemical reaction in this process is of polymerization therefore, he termed it as “Geopolymer” [2,3]. Thus geopolymer constitutes of two main compounds namely source materials and alkaline liquids. The alkaline liquids are f rom soluble alkali metals which are mainly sodium or potassium based. Sodium hydroxide (NaOH) or Potassium Hydroxide (KOH) and Sodium silicate or Potassium silicate are most widely used alkaline liquid. The primary difference between concrete produced usin g Portland cement and geopolymer concrete is the binder. Geopolymer consists of silicon and aluminum atoms bonded via oxygen into a polymer network. Geopolymer are prepared by dissolution and poly condensation reactions between alumino silicate binder and an alkaline silicate solution such as a mixture of an alkali metal silicate and metal hydroxide is obtained. The most common industrial by -products used as binder materials are fly ash (FA) and ground granulated blast furnace slag (GGBS). GGBS has been widely used as a cement replacement material due to its latent hydraulic properties, while fly ash has been used as a pozzolanic material to enhance the physical, chemical and mechanical properties of cement and concrete. Increasing emphasis on the environmental impacts of construction materials such as Portland cement has provided immense thrust in recent years to the increased utilization of waste and by -product materials in concretes. Activation of alumina silicate materials such as fly ash, b last furnace slag, and metakaolin using alkaline solutions to produce binders free of Portland cement is a major advancement towards increasing the beneficial use of industrial waste products and reducing the adverse impacts of cement production. It has been reported that fly ash and ground granulated blast furnace slag (GGBS) are very effective as starting materials for cement - free binder concretes because of the soluble silica and alumina contents in these materials that undergo dissolution, p olymerization with the alkali, condensation on particle surfaces, and solidification that eventually provides strength and stability to these matrices. Investigation and discussions for mix design code for geopolymer concrete are in process to add up in IS standards. However in this study mix design for geopolymer concrete is used from theory of mix design proposed by Subash V.Patankar. International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 04 Issue: 08 | Aug -2017 www.irjet.net p -ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 19 2. EXPERIMENTAL WORK Entire experimental work carried out is briefly explained as follows. 2.1 Materials Fly ash used in this study is low calcium class F Fly ash from Dirk India private limited under the name of the product POZZOCRETE 60. Ground granulated blast furnace slag (GGBS) used is obtained from JSW cements. The chemical and physical properties of GGBS a nd Fly Ash used are shown in the Table 1 and Table 2 respectively. The most commonly used alkaline activators are a mixture of sodium hydroxide (NaOH) with sodium silicate (Na2SiO3). For preparation of alkaline liquids, sodium hydroxide with 98% purity in the form of flakes and sodium silicate were obtained from local manufacturer. Locally available 20 mm crushed aggregates have been used as coarse aggregates. Locally available river sand is used as fine aggregate in the mixes. Crushed granite stones of size 20mm are used as coarse aggregate. As per IS: 2386 (Part III) -1963, the bulk specific gravity in oven dry condition and water absorption of the coarse aggregate are 2.58 and 0.3% respectively. The fineness modulus of 20mm coarse aggregates are 6.68. A s per IS: 2386 (Part III) -1963, the bulk specific gravity in oven dry condition and water absorption of the sand are 2.62 and 1% respectively. The fineness modulus of sand is 2.47. TABLE 1 . Chemical properties of Fly Ash and GGBS TABLE 2. Physical properties of Fly Ash and GGBS Properties Class F fly Ash GGBS Specific Gravity 2.24 2.86 Fineness (m²/kg) 360 400 2.2 Specimen Details and its Schedule Specimen size and its number are determined Table 3. Schedule of Specimen As per the respective IS guidelines used for determining mechanical properties of Concrete. Thus total 99 cubes + 66 beams + 66 cylinders = 231 specimens are casted and tested as per their IS provisions. Cubes are tested as per IS 516 -1959. Chemical Composition Class F fly ash GGBS % silica (SiO 2) 65.6 30.61 % Alumina (Al 2O3) 28.0 16.24 % Iron oxide (Fe 2O3) 3.0 0.584 % Lime (CaO) 1.0 34.48 % Magnesia (MgO) 1.0 6.79 % Titanium (TiO 2) 0.5 - % Sulpher Trioxide (SO 3) 0.2 1.85 Loss on Ignition 0.29 2.1 Concrete Designation Concrete Grade Binding Material Curing Type Cube (Compressive Strength) Beam (flexural Strength) Cylinder (Split Tensile Strength) 3 Day Test 7 Day Test 28 Day Test 7 Day Test 28 Day Test 7 Day Test 28 Day Test Batch -O M25 OPC -43 Cement -100% Water curing 3 3 3 3 3 3 3 Batch -A G25 Fly Ash -100% GGBS - 0 % Hot Curing 3 3 3 3 3 3 3 Ambient Curing 3 3 3 3 3 3 3 Batch -B G25 Fly Ash - 75 % GGBS - 25 % Hot Curing 3 3 3 3 3 3 3 Ambient Curing 3 3 3 3 3 3 3 Batch -C G25 Fly Ash - 50 % GGBS - 50 % Hot Curing 3 3 3 3 3 3 3 Ambient Curing 3 3 3 3 3 3 3 Batch -D G25 Fly Ash - 25 % GGBS - 75 % Hot Curing 3 3 3 3 3 3 3 Ambient Curing 3 3 3 3 3 3 3 Batch -E G25 Fly Ash - 0 % GGBS - 100 % Hot Curing 3 3 3 3 3 3 3 Ambient Curing 3 3 3 3 3 3 3 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 04 Issue: 08 | Aug -2017 www.irjet.net p -ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 20 Beams are tested according to IS 516 -1959 provisions and cylinders are tested according to IS 5816 -1999 provisions. 2.3 Mix Design Mix design of geopolymer concrete of G -25 grade is made according to theory of mix design proposed by Subash V.Patankar . Alkaline liquid ratio is kept constant at 1. The sodium hydroxide solution of 14 M concentration is used and it is kept constant throughout investigation. Cubes of size 150 mm × 150 mm × 150 mm are used for compressive strength test. Beams of size 100 mm × 100 mm × 500 mm are used to determine flexural strength and cylinder of 150 mm (dia) × 300 mm are used to determine split tensile strength. Table 4. Mix proportions of Geopolymer Concrete (G -25) FA or GGBS Sand Coarse Agg. (20mm) NaOH Na 2SiO 3 Water Kg/m³ 587.3 Kg/m³ 1283.08 Kg/m³ 70.88 Kg/m³ 70.88 Kg/m³ 82.42 Kg/m³ 1 1.45 3.16 0.35 0.20 2.4 Curing In this experimental work effect of oven curing and ambient curing on mechanical properties of geopolymer concrete is also studied. After casting the test specimens, specimens are demoulded after 24 hours and kept for oven curing at 60°C for 24 hours and then after are kept at ambient surroundings while some specimens are directly kept for ambient curing. 2.4 Testing of Specimens Specimens are tested at 3, 7 and 28 day of testing for compressive strength, flexural strength and split tensile strength. 2.4.1 Compressive Strength testing of Cubes Cubes as casted of size 150 x 150 x 150 mm were tested using Compression testing machine (CTM) of capacity 250 ton, capable of giving load at the rate of 140 kg/sq.cm/min. Testing of the cubes was done at the age of 3rd, 7th and 28th day. Cubes were placed in the machine between wiped and cleaned loading sur faces and load is given approximately at the rate of 140 kg/sq.cm/min. and ultimate crushing load is noted to calculate crushing strength of concrete according to IS: 516 -1959. The measuring strength of specimen is calculated by dividing the maximum load applied to the specimen during the test by the cross section area. Fck= = =………N/mm² (Eq.1) Fig.1 Compressive testing 2.4.2 Split Tensile Testing of Cylinder Cylinders as casted of size 150 mm (dia) × 300 mm were tested using CTM machine of capacity 250 ton, capable of giving load at the rate of 140 kg/sq.cm/min. Testing of the cubes was done at the age of 7th and 28th day. The cylinders were placed in the machine between wiped and cleaned loading surfaces and load is given approximately at the rate of 140 kg/sq.cm/min. and ultimate crushing load is noted to calculate crushing strength of concrete according to IS: 516 -1959. The measuring strength of specimen is calculated by dividin g the two times maximum load applied to the specimen during the test by the cross section area Ft= =…….. N/mm² (Eq.2) Fig.2 Split tensile testing International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 04 Issue: 08 | Aug -2017 www.irjet.net p -ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 21 2.4.2 Flexural Testing of Beams Beams casted as of size 100 mm × 100 mm × 500 mm are tested for flexural strength under the UTM machine of 600 KN capacity as per the guidelines given in IS 516 -1959.The flexural strength of beam is calculated by following equation Fb= ……….N/mm² (Eq.3) 3. RESULTS AND DISCUSSION All the specimens are tested according to their respective IS provisions and results are shown as below, Table 6. Compressive Strength Results Batc h 3 Day Result (N/mm²) 7 Day Result (N/mm²) 28 Day Result (N/mm²) Ambient Curing Oven Curing Ambient Curing Oven Curing Ambient Curing Oven Curing Batc h- O 11.81 - 19.17 - 26.67 - Batc h- A 5.71 12 8.33 14.8 9 14.66 19.33 Batc h- B 20.94 26.60 25.48 38.3 3 32.64 41.04 Batc h- C 29.65 32.81 33.12 40.0 7 39.52 43.24 Batc h- D 37.55 40.56 41.57 46.1 1 47.91 50.47 Batc h- E 40.13 41.86 43.75 48.4 9 47.98 53.11 Fig.3 3 Day Compressive Strength Fig.4 7 Day Compressive Strength Fig.5 28 Day Compressive Strength The Above graph shows Compressive Strength results of concrete. From the graph it is clearly seen that Fly ash based geopolymer concrete doesn’t show good strength but however GGBS based geopolymer concrete shows the best resul t of compressive strength. Also, oven cured specimens shows more strength than the ambient cured specimens. From the above results batch B gives optimum mix design with satisfactory results however Batch E gives the highest strength for the mix design combination.From Batch B onwards we can replace Conventional Concrete by geopolymer concrete since it is giving better results than it. Fig.6 7 Day Split Tensile Strength in N/mm² International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 04 Issue: 08 | Aug -2017 www.irjet.net p -ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 22 Fig.7 28 Day Split Tensile Strength in N/mm² The above graph shows the results of Split tensile strength of concrete. From the graph it is clearly seen that split tensile strength increases as we increase replacement of fly ash by GGBS. The highest split strength is achieved in Batch E that is in GGB S based geopolymer concrete giving highest strength. From the graph it is seen that oven cured specimen gives more strength than ambient cured specimen. Also early stage strength gain is high in geopolymer concrete than the normal conventional concrete. Table 7. Flexural Strength Results Batch 7 Day Result (KN) 28 Day Result (KN) Ambient Cured Oven Cured Ambient Cured Oven Cured Batch - O 10.38 - 13.50 - Batch - A 8.85 9.12 9.25 9.90 Batch - B 10.11 10.83 12.20 12.94 Batch - C 12.84 13.08 14.92 15.20 Batch - D 13.08 14.80 15.46 16.10 Batch -E 14.40 15.08 16.46 17.10 Flexural strength of geoplymer concrete gives much more flexural strength than the conventional concrete from Batch B onwards. Also the same nature that oven cure specimen gives more strength than the ambient cured specimen. Highest flexural strength from above is given by GGBS based g eopolymer concrete that by Batch E. Fly ash based geopolymer concrete strength results are less than the normal conventional concrete. Flexural strength result is shown as load taken in KN. Fig.8 7 Day Flexural Strength Results in KN Fig.9 28 Day Flexural Strength Results in KN Failure of geopolymer concrete and conventional concrete was brittle failure. Flexural strength of geopolymer concrete is more than the conventional concrete except in Batch A. however the same nature of being more st rength in oven cured than the ambient cured is found. Maximum of flexural strength is achieved in Batch E that GGBS based and oven cured specimen. 4. CONCLUSIONS Based on above experimental work carried out, following conclusions are made 1) It is observed that an increasing trend has been observed in compressive strength of GPC mixes up to full replacement level of fly ash by GGBS. 2) The rate of gain in compressive strength, split tensile strength and flexural strength of geopolymer concrete is very fast at 7 days curing period and the rate gets reduces with age. 3) Almost 90 % of strength is achieved at 7th day. Therefore speedy construction is possible with geopolymer concrete. 4) Oven curing gives higher result than ambient curing International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 04 Issue: 08 | Aug -2017 www.irjet.net p -ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 23 5) Addition of GGBS eliminate the necessity of oven curing. 6) The geopolymer concrete using GGBS as a sole binder achieves more strength than that of normal control concrete when oven curing is done. A higher concentration of GGBS result in higher compressive strength, split tensile strength and flexural strength of geopolymer concrete. 7) High Strength Concrete is achieved by using geopolymer concrete. 8) Fly ash and GGBS based geopolymer concrete can be used for structural use. 9) Mix design with 75% fly ash and 25% GGBS as binding ma terial gives economic design with better strength. 10) However GGBS based geopolymer concrete give best results in all aspect. 11) Thus Geopolymer concrete can be recommended as an innovative construction material for the use of construction. ACKNOWLEDGEMENT I wish to express my deep sense of gratitude to my guide and PG co coordinator Prof.G.R.Patil, Department of Civil Engineering, for his valuable time, sustained guidance and useful suggestions, which helped me in completing the work, in time. Last, but not the least, I would like to express my heartfelt thanks to my beloved parents for their blessings, my friends/classmates for their help and wishes for the successful completion of this work. REFERENCES Malhotra V. M., 2002, Introduction: Sustainable development and concrete technology, ACI Concrete International, 24(7). The Proceedings of Geopolymer 2005 World Congress , 4th Int. Conference on geopolymers, edited by J. Davidovits, Geopolymer Institute, France 2006 Davidovits J, 1994, “Propertie s of Geopolymer cement,”, Proceedings first international conference on Alkaline cements and concretes, scientific research institute on binders and materials”, Kiev state technical university, Kiev, Ukraine, pp. 131 -149. D. Hardjito and B. V. Rangan, 2005,”Development & Properties of Low calcium Fly ash based Geopolymer concrete” Research report Curtin University, Australia. Adam, A.A, 2009, “Strength and Durability Properties of Alkali Activated Slag and Fly Ash -Based Geopolymer Concrete”, Thesis. RMIT University. Talling, B and Brandstetr, J.,1989,“ Present state and future of alkali -activated slag concretes, Proceedings of the Third International Conference, Trondheim, Norway ”ACI SP -114, Vol. 2, pp 1519 -1546 Subash V.Patankar, Springer India 2015, advances in structural Engineering DOI 10.1007/978 -81-322 -2187 - 6_123. BIOGRAPHIES Ajay M. Takekar (PG 2nd year student) Department of Civil (Structural) Engg. JSPM’ s Rajarshi Shahu college of Engineering, Tathawade,pune -411033 Prof.G.R.Patil (Guide & PG coordinator) Department of Civil (Structural) Engg. JSPM’s Rajarshi Shahu college of Engineering, Tathawade, Pune -411033

See discussions, st ats, and author pr ofiles f or this public ation at : .researchgate.ne t/public ation/316566650 Mechanical Properties of Geopolymer Concrete Composites Article    in  Mat erials T oday: Pr oceedings · Dec ember 2017 DOI: 10.1016/ j.matpr .2017.02.175 CITATIONS 37READS 3,922 2 author s: Some o f the author s of this public ation ar e also w orking on these r elat ed pr ojects: Strength and dur ability st udies of multi blended c oncr etes c ontaining fly ash and silic a fume View pr oject Thesis View pr oject Kolli R amujee VNR V ignana Jy othi Instit ute of Engineering & T echnolog y 21 PUBLICA TIONS    92 CITATIONS     SEE PROFILE Pothar aju Malasani GITAM Univ ersity 34 PUBLICA TIONS    151 CITATIONS     SEE PROFILE All c ontent f ollo wing this p age was uplo aded b y Kolli R amujee on 10 Mar ch 2018. The user has r equest ed enhanc ement of the do wnlo aded file. Available online at www.sciencedirect.com ScienceDirect Materials Today: Proceedings 4 (2017) 2937 –2945 www.materialstoday.com/proceedings 2214 -7853©2017 Elsevier Ltd. All rights reserved . Selection and peer -review under responsibility of Conference Committee Members of 5th International Conference of Materials Processing and Characterization (ICMPC 2016) . 5th International Conference of Materials Processing and Characterization (ICMPC 2016) Mechanical Properties of Geopolymer Concrete Composites Kolli.Ramujeea, M.PothaRajub a Professor, VNR Vignana Jyothi Institute of Engineering & Tahnology,Bachupally,Hyderabad -500090, India bProfessor of Civil Engineering,GITAM University,Rushikonda,Vishakapatnam -530045,India Abstract The increasing demand of environment friendly constructi on has been the driving force for developing sustainable and economical building materials. The critical aspects influencing the development are performance of the materials under different and special user conditions, economic aspects as well as environme ntal impact aspects. Cement is an energy consuming and high green house gas emitting product. Geopolymers are gaining increased interest as binders with low CO 2-emission in comparison to Portland cement. In the present investigation, the mechanical propert ies of flyash based geopolymer concrete(GPC) were studied. Experimentally measured values of the compressive strength and split tensile strength of GPC specimens made from low, medium and higher grades compared with reference to the control mixes(OPC). The regression model analysis was carried out to study the relationship between the Compressive strength and Split tensile strength and It was found that the mechanical behaviour of GPC is similar to that of ordinary Portland cement(OPC) concrete. ©201 7 Elsevier Ltd. All rights reserved . Selection and peer -review under responsibility of Conference Committee Members of 5th International Conference of Materials Processing and Characterization (ICMPC 2016) . Keywords: Geopolymers; OPC concrete;class -F flyash;Alkaline actvators; Strength,Correlation 1. Introduction Increasing emphasis on the environmental impacts of construction materials such as Portland cement has provided immense thrust in recent years to the increased utilization of waste and by -product materials in concretes. Activation of aluminosilicate materi als such as fly ash, blast furnace slag, and metakaolin using alkaline solutions to produce 2938 Kolli.Ramujee and M.PothaRaju / Materials Today: Proceedings 4 (2017) 2937– 2945 binders free of Portland cement is a major advancement towards increasing the beneficial use of industrial waste products and reducing the adverse impacts of cement production. These alkali -activated binder systems appear in the vast available body of literature in a variety of names: alkali- activated cement (predominantly used for binders containing large amounts of calcium such as slags), geopolymer [2 –6], inor ganic polymer , hydroceramic, or low temperature aluminosilicate glass . They are sometimes called polysialates also because of the polymeric silicon –oxygen –aluminum framework . In general, concretes made using alkali activation of the starting aluminosilicate material do not contain Portland cement at all, and hence they are referred to generically as cement - free binder concretes in this study. The concretes containing alkali -aluminosilicate gel as the binder has been shown to have high compressi ve strengths, and resistance to fire, and chemical attack [10 –13]. The starting material and activating agent type and concentration are the most important parameters that influence the properties of the alkali - activated end product [14,15] . The production of cement -free binders requires a starting material containing aluminum and silicon species that are soluble in highly alkaline solutions. Previous studies have shown that the amount of vitreous silica and alumina present in the starting material plays a significant role in activation reactions and the properties of the reaction product [10,16] . The presence of calcium oxide in the source material also influences the properties, especially because of the formation of more than one reaction product . Sodium hydroxide and sodium silicate are the more commonly used alkaline activating agents [18–20] even though few studies have also been carried out with potassium hydroxide or sodium carbonate as the activating agent .In the present investigation, Geo polymer concrete is manufactured using source materials that are rich in silica and alumina. While the cement -based concrete utilizes the formation of calcium -silica hydrates (CSHs) for matrix formation and strength, geopolymers involve the chemical reacti on of alumino -silicate oxides with alkali polysilicates yielding polymeric Si– O–Al bonds. 2.0 Source materials and mixture proportions 2.1. Materials used In this experimental work, fly ash is used as the source material to make geopolymer paste as the b inder, to produce concrete. The manufacture of geopolymer concrete is carried out using the usual concrete technology methods. The role and influence of aggregates are considered to be same as in the Portland cement concrete.The mass of combined aggregate s may be taken to be between 75% and 80% of the mass of the Geopolymer concrete. Sodium -based activators were chosen because they were cheaper than Potassium -based activators. The sodium hydroxide was used, in flake or pellet form. It is recommended that the alkaline liquid is prepared by mixing both the solutions together at least 24 hours prior to use. The mass of NaOH solids in a solution varied depending onthe concentration of the solution expressed in terms of Molarity, M. The concentration of sodium hydroxide solution (NaOH) liquid measured in terms of Molarity(M), in the range of 8 to 16 M. The mass of water is the major component in both the alkaline solutions. In order to improve the workability, high performance Polycarboxylic based super plasticizer purchased from BASF under trade name GLENIUM B233 has been added to the mixture. The primary difference between Geopolymer concrete and Portland cement concrete is the binder. The silicon and aluminium oxides in the low – calcium fly ash reacts with th e alkaline liquid to form the geopolymer paste that binds the loose coarse and fine aggregates and other unreacted materials to form the geopolymer concrete. As in the case of Portland cement concrete the coarse and fine aggregates occupy about 75% to 80% of the mass of Geopolymer concrete. This component of Geopolymer concrete mixtures can be designed using the tools currently available for Portland cement concrete. The compressive strength and workability of geopolymer concrete are influenced by the proportions and properties of the constituent materials that make the geopolymer paste. 2.2 Mixture Proportions The mix proportions for Geopolymer concretes of three grades namely G20,G40 & G60 were arrived based on the Kolli.Ramujee and M.PothaRaju / Materials Today: Proceedings 4 (2017) 2937 –2945 2939 trial mixes carried out by the author i n the laboratory and for control mix the corresponding grades were taken in equivalent mix proportions of geopolymer concrete. the details of properties of materials ( table 1to 5) and the the mix proportions for GPC of three grades((shown in tab le .6 ) and the control mix proportions were given in table7. Table 1. Chemical composition of flyash(%mass) Oxides Percentage SiO2 Al203 Fe2O3 CaO MgO K2O+Na 2O SO 3 Loss on ignition 60.54 5.87 0.38 0.23 Table 2. Properties of Sodium Hydroxide (NaOH) Property Fine Aggregate Coarse Aggregate Specific gravity Water absorption Fineness modulus Bulk density Kg/m3 Source 2.78 0.50% 7.21 Crushed granite stone 2.63 1% 2.40 River sand Table 3 . Properties of Sodium silicate solution Property Value Specific gravity Molar mass (gr/mol) Na2O ( by mass) SiO 2 ( by mass) Weight of solids (by mass) Water ( by mass) Weight ratio (SiO 2 to Na 2O) Molar ratio 1.6 14.70% 29.40% 44.1% 55.90% 2.06 2940 Kolli.Ramujee and M.PothaRaju / Materials Today: Proceedings 4 (2017) 2937– 2945 Table 4. Properties of Sodium hydroxide(NaOH) Property value Molar mass Appearance Density Melting point Boiling point Amount of heat liberatedhen dissolved in water 40 gr/mol White solid 2.1gr/cc 3180C 13900C 266 cal/gr Table 5 . Properties of Cement Property Value Standard consistency Initial setting time Final setting time Specific gravity Soundness Compressive strength (MPa) 7 days 28 days 32% 115 min 240 min 1mm 40.1 Table 6. Mix proportions for various grades of Geopolymer concrete Grade of GPC G20 G40 G60 Fly ash (kg/m3) Fine Aggregate (kg/m3) Coarse Aggregate (kg/m3) NaOH(kg/m3) (M) Concentration Na2SiO 3(kg/m3) Extra water (kg/m3) Super plasticizer (GLENIUM)(kg/m3) Ratio of Mixture proportions Liquid/binder ratio Water/geopolymer Solids ratio Workability (Slump) 327 1248 54.33( 8M) 108.67 ----- 1:2.05:3.81 0.31 100 mm 394.3 1201.2 45.06(16M) -- 1:1.64:3.04 0.21 50mm 408.9 1294.0 40.89(16M) -- 1:1.35:3.16 0.158 50mm Kolli.Ramujee and M.PothaRaju / Materials Today: Proceedings 4 (2017) 2937 –2945 2941 Table 7. Mix proportions of OPC Conventional concrete expressed as Equivalent Proportions of GPC Grade of concrete M20 M40 M60 Cement (kg/m3) Fine Aggregate(kg/m3) Coarse Aggregate(kg/m3) Water(kg/m3) SP(kg/m3) Mix Proportion 327 1248 2 1:2.05:3.81 394.3 1201.2 4 1:1.64:3.04 408.9 1294.0 143.11 1:1.35:3.16 W/C ratio 0.50 0.40 0.35 2.2 Specimen preparation The sodium hydroxide flakes were dissolved in distilled water to make a solution with a desired concentration at least one day prior to use. The fly ash and the aggregates were first mixed together in a 80 litre pan mixer shown in fig.1 for about three minutes. The sodium hydroxide and the sodium silicate solutions were mixed together and then added to the dry materials and mixed for about four minutes shon in fig.2 , then the super plasticizer of required dosage was added. Even though af ter addition of super plasticizer, the required slump is not achieved ,then the extra water was added to the mix. After mixing , the slump of the fresh geopolymer concrete was determined in accordance with slump test . After determination of slump , the fresh concrete was cast into the mould. The specimens were compacted with three layer placing and tamping using a rod. This was followed by an additional vibration of 10 seconds using a vibrating table. Fig.1 Pan mixer used for mixing GPC Fig.2 addition of Alkaline liquids to the Mix 2.3. Curing regime The specimens were wrapped with thin vinyl sheet to avoid loss of water due to evaporation, All the specimens were then transferred to an oven set at a temperature of 60 °C and stored for 24 hours. After curing ,the specimens were allowed to cool in air ,demoulded and kept in open until the day of testing. Both types of heat curing and 2942 Kolli.Ramujee and M.PothaRaju / Materials Today: Proceedings 4 (2017) 2937– 2945 ambient curing were adopted. In case of heat curing shown in fig.3 , the specimens were cured in an oven and in ambient curing the specimens were left to air for required period. For each grade of Geopolymer concrete mix about 7 mix proportions were considered and the mix which exhibited the required compressive strength consistently at 28 days un der both curing conditions was adopted for further investigations. Fig.3 Hot Air Oven for curing GPC Fig.4 Computerized 100 T UTM for testing 2.4.Strength tests The compressive strength and split tensile strength tests on GPC concretes for G20,G40 and G60 grades and their counterparts i.e control mixes have been carried out shown in fig.4 .A summary of the mechanical properties measured for each of the sample is presented in the Table No.8. 3.0 Results Table: 8 Mechanical proper ties of Geopolymer concrete & OPC concrete S. N o Mix designaton Compressive Strength (MPa) Split Tensile strength (MPa) Heat cured specimens @ 28 days 1 MH2G 20 30.4 1.97 2 31.2 2.34 3 32.4 2.68 4 MH2G 40 50.0 3.02 5 49.69 2.98 6 52.28 3.45 7 MH2G 60 71.2 4.20 8 70.6 4.20 9 71.4 4.60 Control mix water cured @ 28 days 10 M1M20 27.5 2.21 11 M1M40 48.8 3.54 12 M1M60 68.6 4.52 Kolli.Ramujee and M.PothaRaju / Materials Today: Proceedings 4 (2017) 2937 –2945 2943 4.0 D iscussions The cube compressive strength ranged from 30.4Mpa to 71.4Mpa for GPC specimens under heat cured conditions are shown in Table 8.,where as for control mix the results varied from 27.5Mpa to 68.6Mpa is also shown in table 8. The split tensile strength values for GPC for heat cured were shown in the table 8.I t has been widely reported that splitting tensile strength can be estimated from compressive strength of concrete through various empirical relations proposed by different concrete institutes and researchers . These empirical relations can be summarized by the following general equ ation: f ts = A( fck)B where f ts is splitting tensile strength, MPa; f ck iscompressive strength, MPa; A and B are regression coefficients. A Regression model analysis was carried out using MATLAB software for analyzing the results between the compressive strength and split tensile strength . Correlation between the split tensile strength (f ts) and compressive strength of GPC is also analysed by regression analysis was shown in fig.3, by regression analysis the following empirical relation can be expressed as f st = 0.08 f cs0.92 with R2of this proposed relation as 0.80 suggests variabi lity among the relation may be due to various factor which were affecting the properties of GPC with heat curing and ambient curing condtions. However , the relationship was supported by the following empirical formula suggested by ACI318- 99 , Fig. 5, f st = 0.095 f cs0.89 Fig.5 Relationship between compressive strength and split tensile strength 2944 Kolli.Ramujee and M.PothaRaju / Materials Today: Proceedings 4 (2017) 2937– 2945 5.0 C onclusions Based on the applicability evaluation and empirical relations to GPC and the regression analysis on experimental data of mechanical properties of Geopolymer Concrete , the following conclusions can be drawn. 1) The GPC attains its target strength much faster under heat cured condition compared to ambient cured condition.. 2) The GPC under heat cured conditio n exhibited almost same split tensile strength as that of corresponding strengths of OPC concrete. 3) The Relationship between Split tensile strength and compressive strength of GPC can be expressed using regression model analysis that resembles that given b y ACI -318-99 for Ordinary concrete. 4) Mechanical behavior of Geopolymer concrete is similar to that of OPC concrete. 5) Geopolymer concrete exhibits some of the characteristics required for an Engineered material based on its mechanical properties and behavio r.

Construction and Building Materials 301 (2021) 124380 Available online 31 July 2021 0950-0618/© 2021 Elsevier Ltd. All rights reserved.Influence of molarity and alkali mixture ratio on ambient temperature 303007, India bDepartment of Civil Engineering, Malaviya National Institute of Technology, Jaipur 302017, India ARTICLE INFO Keywords: Construction and demolition waste Molarity Alkali mixture Geopolymer mortar ANOVA ABSTRACT Generation of an enormous amount of construction and demolition (C&D) waste due to rapid growth in the infrastructure segment globally is a matter of dire concern for the present decade due to poorly managed disposal to re-utilization. The present study focuses on the formulation of an effective utilization regime of C&D waste (cement concrete) for its re-consumption as geopolymer mortar. Formulation of different molarities of sodium hydroxide (6 M to 16 M) and alkali mixture ratio of sodium silicate to sodium hydroxide solution (1.5, 2.5 and 3.5) have been done and cured for 90D at controlled ambient temperature. The samples were tested for fresh and hardened properties. The effect of silicate modulus and Na2O% with different alkali proportions on the hardened properties were studied. The results reveals that the optimum compressive strength of 20.7 MPa is achieved at a silicate modulus of 1.54 and Na2O% of 12.5%. Further microstructural analysis (mineralogical, elemental, thermal and bonding types) has been done to better comprehend the interactive performance of C&D waste with the change in alkali quantities. The physicochemical and microanalysis reveals that molarity, dissolution of aluminosilicates, and mineralogical compositions are the main factors governing the outcomes while keeping the curing regime at ambient temperature. 1.Introduction The construction industry is closely related to any country ’s socio- economic growth. It showcases the prosperity and livelihood of the countrymen. When we closely observe the construction industry, the foremost thing is the construction materials, and the most used con- struction material is cement concrete. Cement concrete is the second- highest utilized material after water throughout the globe . Also, it is a fact that Ordinary Portland Cement (OPC) production is an energy- intensive process and contributes approximately 5–8% of the total anthropogenic CO2 (major greenhouse gas) emission globally per annum . The total CO2 emission from cement production can further be bifurcated into CO2 emitted from cement clinker production (50–60%) and the burning of fossil fuel (40%) . Other than the OPC production to manufacture concrete, a substantial amount of natural aggregates (NA) are also consumed, which leads to the scarcity of natural aggregate and increases its overall cost . Other than the shortage of natural aggregate and emission of CO2 due to OPC manufacturing, continuous repair and rehabilitation of old structures generate a considerable amount of construction and demolition (C&D) waste . Due to the vast heterogeneity of the C&D waste, it is typically tough to recycle, and unmanaged dumping of the C&D waste occupies a massive area for landfills and creates an envi- ronmental concern due to the different levels of contamination of waste [5,6] . Therefore, reuse of the C&D waste is encouraged by various government agencies throughout the globe, and a considerable amount of research works have already been conducted utilizing C&D as a partial to complete replacement of NA in cement concrete production [7-9] . However, due to the adhesion of old mortar with the recycled aggregate (RA), the quality of the RA is much inferior to the NA and subsequently resulted in at least 10–40% strength reduction with about 50% increased water absorption and drying shrinkage [10-13] . Addi- tionally, utilization of RA, as a replacement of NA, reduces the bulk density, specific gravity, and compressive strength, limiting its con- sumption to the non-structural application or non-important structural part like base course filling work. The non-structural applications also create environmental hazards due to leaching out of the contaminants and altering the surrounding earth ’s pH level . Geopolymers, a series of inorganic polymers generated from alumi - nosilicate sources in a concentrated alkali environment, make their way in the effective utilization of C&D waste. The most often used waste *Corresponding author. E-mail address: san.civil@gmail.com (S. Shrivastava). Contents lists available at ScienceDirect Construction and Building Materials u{�~zkw! s{yo| kro>! ÐÐÐ1ow �o�to~1m{y2w {mk�o2m{zl� twnyk�! Received 8 January 2021; Received in revised form 20 July 2021; Accepted 26 July 2021 Construction and Building Materials 301 (2021) 124380 2materials in the production of geopolymers are fly ash and slag. Recently, building and demolition wastes such as cement concrete, clay brick, and glass have also been investigated. [4,15-19] . In a highly concentrated alkali environment (either sodium (Na-) or potassium (K-) based), the silica (Si) and alumina (Al) from the aluminosilicate source precipitate and form the Si-O-Al-O monomers . The entire geo- polymerization process begins with the dissolution of Si and Al, followed by the production of oligo-sialate monomers, and finally, the develop - ment of geopolymers by reticulation, networking, and solidification. . Eq 1 and 2 represent the chemical reactions behind geo- polymerization . Researchers have proved that geopolymers have shown superior quality in both mechanical and durability aspects over cement mixtures , leading to sustainable development and a greener environment by utilizing the industry-generated wastes. Further, the wastes having a considerable amount of calcium enhance the geo- polymer ’s mechanical properties due to the coexistence of C-S-H gel and the sialate bond (Si-O-Al-O) [23,24] . Focusing on utilizing different types of C&D wastes, researchers have used waste cement concrete as a precursor for producing geopolymers. Still, it was either a binary or ternary blend mixture. Zaharaki et al., in their study, examined both sole recycled cement concrete based geopolymers and four other C&D waste blended geopolymers and they found that upon heat curing at 80 •C for 24 hrs, the control mix (10 M NaOH solution) was able to achieve a compressive strength of 8 MPa only . While substituting the recy- cled cement concrete content with other precursors (slag, tiles, and red clay bricks), only 10% of recycled cement concrete content gives an optimum compressive strength of 75 MPa at the same curing condition. Allahverdi and Najafi examined a binary blend of waste brick with waste concrete, and they encountered a maximum compressive strength of 16.5 MPa at 28D with 8% Na2O content when the optimum amount of waste concrete was 60% of the total precursor . Another researcher H. M. Khater utilized crushed walls which contains bricks with adhered cement mortar (55% wt.), waste concrete (40% wt.), and hydrated lime (5% wt.) to assess the effect of calcium content on geopolymerization as a precursor and reported 5 MPa 28D compressive strength at ambient temperature curing . Minimal investigations have been reported utilizing waste cement concrete as a sole precursor for geopolymer production. Robayo-Salazar et al. reported a compressive strength of 7.5 MPa with only NaOH (6% Na2O) and 25 MPa with NaOH and Na2SiO3 mixture having silicate modulus of 11.11 at 25 •C temperature . Whereas utilizing 100% concrete waste as a precursor for geopolymer, Zaharaki et al. reported a compressive strength of 7.8 MPa with 10 M NaOH solution at ambient temperature curing . In contrast, V˘asquez et al. reported a compressive strength of 25 MPa at ambient temperature. Interestingly, the compressive strength reduces from 37 MPa at 7D to 25 MPa at 28D when heat cured at a temperature of 70 •C for 24 h . In addition, using grounded waste concrete as the sole precursor, Ahmari et al. re- ported 7D compressive strength of 7 MPa and 11 MPa using 5 M and 10 M with an alkali mixture ratio of 1 . Due to the lack of reported extensive studies on the utilization of finely grounded waste cement concrete as a sole precursor for geo- polymer production, the present study aims to analyse the effect of the molarity of NaOH solution and alkali mixture ratio (NaOH/Na 2SiO3) on the fresh properties and mechanical strength of geopolymer mortar at ambient temperature curing. Different molarities of NaOH solution (6–16 M) and AM ratio (1.5–3.5) were formulated. Further, the List of abbreviations and notations: AM Alkali mixture (Na 2SiO3 / NaOH solution) ANOVA Analysis of Variance C&D Construction and Demolition COD Coefficient of Determination, denoted as R2 C-S-H Calcium Silicate Hydrate D Age of casting in days (Ex.- 28D means 28 days of casting) dF Degrees of Freedom DTG Derivative Thermogravimetry EDX Energy Dispersive X-ray Edyn Dynamic Modulus of Elasticity FTIR Fourier Transform Infrared Spectroscopy GW Glass Waste IR Infrared JCPDS Joint Committee on Powder Diffraction Standards M Molarity MS Mean of Squares MSW Municipal Solid Waste NA Natural aggregate N-A-S-H Sodium Alumino Silicate Hydrate OPC Ordinary Portland Cement RA Recycled aggregate Rh Relative humidity RMSE Root Mean Square Error RS River sand Sample symbol (Ex. - 6M1.5AM) Here ‘6M’ is the molarity of NaOH solution and ‘1.5AM ’ is the AM ratio value SD Standard Deviation SEM Scanning Electron Microscope SM Silicate Modulus, the ratio of SiO2/Na 2O SS Sum of Squares TGA Thermogravimetric Analysis UPV Ultrasonic Pulse Velocity WCC Waste Cement Concrete XRD X-ray Diffraction XRF X-ray Fluorescence Spectroscopy Å Angstrom v Velocity ρ Density n…Si2O5∙Al2O2†‡2nSiO 2‡4nH 2O⎬⎬⎬⎬↗NaOHorKOHn…OH†3SiO…† Al …OH†2OSi…OH†3 (1) n…OH†3SiO…† Al …OH†2OSi…OH†3⎬⎬⎬⎬↗NaOHorKOH…NaCK†…‡†⎫ ⎬⎭† Si † OO† Al † OO† Si † OO⎩ ⎪⎨‡4nH 2O (2) S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 3mineralogical, chemical, and microstructural analysis for the identified mixtures from the mechanical test results were analysed by conducting Scanning Electron Microscopy (SEM), Energy Dispersive X-ray analysis (EDX), X-ray Diffraction (XRD), Fourier Transform Infrared Spectros - copy (FTIR), and Thermogravimetric Analysis (TGA) with Derivative Thermogravimetry (DTG). 2.Materials and Methodology 2.1. Materials used Waste cement concrete (WCC) originated from demolished plain cement concrete slabs as construction and demolition (C&D) waste was used in this study as a precursor. Table 1 shows the X-ray fluorescence spectroscopy (XRF) of WCC and river sand. The particle size distribution of WCC powder is presented in Fig. 1a. River sand was used as fine aggregate, and its gradation curve is shown in Fig. 1b. Gradation of the river sand was tested for the limit specified for the use in masonry mortar as per IS:2116 –1980 , and it fits within the limit (Fig. 1b). The scanning electron microscopy (SEM) and energy dispersive X-ray analysis (EDX) of WCC is shown in Fig. 2a and 2b. Table 2 presents the physical properties of the materials. XRD analysis (Fig. 3) shows quartz, albite, portlandite and calcite in the WCC. Further, the presence of quartz and albite is due to the ag- gregates ; portlandite and calcite are due to cement in the WCC . The XRF analysis shows that WCC contains 63.70 wt%, 7.13 wt% and 19.70 wt% of SiO2, Al2O3 and CaO, respectively. Whereas EDX also confirms that the vital elements in the WCC are silica (Si) and calcium (Ca) with a low amount of aluminium (Al). Microstructural analysis through SEM presents the irregular particle dimension of WCC. This varying particle size impacts the workability of the mixtures as discussed in the results section (Section 3.1). FTIR analysis (Fig. 4a) of the WCC was conducted within an infrared region of 500–4000 cm1. A medium sharp peak observed around 3725 –3850 cm1 is due to the presence of Table 1 XRF analysis of the base materials. Material Oxides (wt.%) SiO2 Al2O3 CaO TiO 2 MnO Fe2O3 MgO Na2O K2O P2O5 SO3 Cl WCC 63.70 7.13 19.70 0.07 0.05 3.49 1.03 0.59 1.02 0.15 1.23 0.13 River Sand 96.48 0 1.13 0 0.03 2.01 0.34 0 0 0 0 0 Fig. 1.Particle size distribution curve of (a) WCC and (b) Fine aggregate. Fig. 2.(a) SEM image and (b) EDX spectra of WCC. Table 2 Physical properties of the materials used. Material Bulk density (kg/m3) Water absorption (%) Fineness modulus River Sand 1599 1.80 2.74 WCC 1450 S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 4the O–H stretching bond . The band around 1650 –1700 is due to the bending O–H bonds . The highest intensity peak was observed at 1040 cm1 due to a high concentration of Si-O-Si bonds in the WCC . The presence of quartz found in the XRD analysis (Fig. 3) is confirmed by the sharp peak at 782 cm1 and 692 cm1 [34,35] . The peaks in the lower wavenumber 500–650 cm1 are assigned to the covalent bonding of Si-O-Si . The TGA analysis (Fig. 4b) shows a continuous weight loss till 650 •C of 5.5% when heated at a rate of 10 •C per minute till 900 •C in air medium. Also, the TGA graph shows three significant mass losses. The first phase of mass loss ranges from room temperature to 350 •C due to the evaporation/dehydration of different hydrates (C-S-H, ettringite, carbo-aluminates). Within this range, the initial mass loss from room temperature to 110 •C is due to the evaporation of surface to some extent of bound water; 110–170 •C temperature range decomposes the ettringite and part of carbo-aluminate hydrates [37,38] . Finally, the range between 180 and 350 •C governs the bound water mass loss by dehydration of C-S-H and total carbo-aluminate hydrates [39,40] . The second phase occurs in the temperature range of 400–550 •C, where the dihydroxylation of another hydration product, i.e., portlandite occurs. Its presence is also confirmed by the XRD analysis . The third phase of decomposition around temperature ranging from 620 to 680 •C is due to CaCO 3 (calcite) . 2.1.1. Selection procedure and preparation of precursor Though ample research has been reported utilizing WCC as fines , fine aggregates [42-44] , and coarse aggregates [45-47] for pro- ducing geopolymers but most of them used fly ash and GGBS as a primary precursor and substituted some portion of it with WCC. Very few articles utilize WCC of unknown quality as a primary precursor for geopolymer mortar preparation . Few articles also reported known laboratory-made cement concrete as C&D waste as precursors for geo- polymers . Hence, in this study, WCC, a plain cement concrete slab, was used and collected from a landfill site located in the Jaipur city of the state of Rajasthan in India. After collecting the waste, it was grounded in a D-crusher and then in a pulveriser to obtain a finely grounded particle sizeD90 µm (Fig. 1a). The C&D waste was not chemically treated before grinding. Most of the reported studies on utilizing WCC as the precursor employ heat curing regime. Hence, in the present study, controlled ambient temperature curing was conducted along with WCC as a sole precursor for geopolymerization. 2.2. Research methodology 2.2.1. Experimental process 2.2.1.1. Physical testing. In this study, different mix combinations were formulated by altering the alkali proportions only. A combination of sodium hydroxide (NaOH) and sodium silicate (Na2SiO3) was used for alkalination. Laboratory-grade (99% purity) NaOH and commercial- grade Na2SiO3 (Na2O ˆ8.40%, SiO2 ˆ27.22% and H2O ˆ64.38% as provided by manufacturer) was used. The mix ratio of Na2SiO3 to NaOH solutions was varied from 1.5 to 3.5 with an increment of 1. The molarity of NaOH was also varied from 6 M to 16 M with an increment of 2 M. A detailed mix combination is shown in Table 3. The NaOH solution was prepared 24 h before casting as the rest period will help the solution to cool down naturally. After 24 h, two different alkali solutions were mixed and allowed to rest for another day before utilizing it for exper - imental work. Other researchers have also followed a similar approach [49-51] . In the mixing process, initially, the precursor and fine aggregates were mixed in a Hobart mixer for 2 min till a homogeneous mixture was obtained by visual inspection. Then the alkali solution was poured uniformly into the dry mixture, and the mixing was continued for another 2–3 min to get a final consistent mixture. After mixing, the fresh mortar was cast into different sized moulds for compression (cube mould 50 ×50 ×50 mm), flexural (prism mould – 40 ×40 ×160 mm) and drying shrinkage (beam size – 25 ×25 ×250 mm) test. All samples were mixed, cast, and vibrated (for one minute on a table vibrator to expel the entrapped air bubbles within the mixture) at room temperature ranging from 24 •C to 26 •C and kept in the same condition for 24 h. De-moulding was performed after 24 h (except shrinkage test specimens), and the samples were placed at 30 •C with Rh of 85 ±5% in a temperature-controlled oven. The shrinkage test specimens were de- moulded after 48 h of casting. The flow of the fresh mixtures was Fig. 3.XRD analysis of WCC. Fig. 4.(a) FTIR and (b) TGA with DTG plot for WCC. S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 5measured using a flow table test as per ASTM C1437 , and the flow table used is as per the specifications mentioned by ASTM C230/C230M . Setting time test was also conducted at room temperature of 24–26 •C using Vicat needle apparatus according to Indian Standard code IS:4031 (Part-5)-1988 . The destructive and non-destructive test results reported in this study were the average of three tested specimens with the standard deviation as an error. Compressive strength tests were carried out on all specimens at a specified time using a servo-controlled universal testing machine with a load rate of 0.9 kN/s to 18 kN/s according to ASTM C109/109 M . Flexural tests were carried out at a load rate of 0.044 kN/s following ASTM C348 on the flexural prism mould using a three- point loading setup. The flexural strength was calculated using Eq (3), where Sf is the flexural strength in MPa, and P is the maximum load at failure in N. Sfˆ0B0028 P (3) By adopting the procedure defined in ASTM C1148 and ASTM C490 , the drying shrinkage test was conducted using an automated length comparator with precision to quantify the length shift up to 0.001 mm. Ultrasonic pulse velocity (UPV) tests were performed on the 28D, 56D and 90D cured beam prism specimen of size 16 ×4 ×4 cm. Transducers of frequency 55 kHz were pressed on the two ends of the specimens, and the time required by the ultrasonic wave to travel from one end to another were measured in microseconds. The Dynamic Modulus of Elasticity (Edyn) of the mortar specimens were calculated by the ultrasonic pulse velocity method according to BS EN 12504-4:2004 by using the obtained UPV values. Eq (4) is used to calculate the Edyn values from the obtained UPV values. Edynˆ…v2×ρ† (4) The velocity (km/s) of the longitudinal ultrasonic wave was obtained from the UPV test, and the density (kg/m3) was calculated by measuring the weight and the dimensions of each beam mortar specimen. 2.2.1.2. Microstructural analysis. To better comprehend the physical test, microstructural analysis (SEM, EDX, XRD, FTIR, TGA) was per- formed by obtaining samples from various portions of the 28D cured specimens after conducting the compressive strength test. All collected samples were quenched in acetone for 7D and transformed to different levels of fineness as per the requirements of each microstructural analysis. JSM-7610FPlus FESEM coupled with EDX detector was used to examine the microstructure and elemental composition changes for the change in molarity and alkali solution mixture ratio. Gold sputtering was applied on the specimens for one minute to coat the samples with a thin layer of 100 Å before testing. Rigaku Miniflex 600 was used to conduct the XRD analysis. This analysis determines the mineralogical variations. Scanning of the sam- ples was performed for Bragg ’s angle (2-theta) from 5•to 80•at a speed of 1•per minute. Cu-K α radiation and graphite monochromator was used, and a diffractometer records the diffracted beams at various an- gles. Philips X’Pert Highscore software was used to plot the data in in- tensities vs 2θ (Bragg ’s angle), and the peaks were identified by comparing the peaks with the data files available in the JCPDS database. Fourier-transform infrared spectroscopy (FTIR) works on the prin- ciple that each element either absorbs or transmits some wavelengths in the wide infrared zone, which will precisely help to identify the avail- able bonding or functional groups within the elements. Mostly the wavenumbers for the mid-infrared (IR) zone of 400–4000 cm1 was utilized to determine the vibrational or rotational / bending vibrational structures. The Bruker ’s Alpha II model was used to measure the same within a range of 500–4000 cm1 wavenumber. The background IR spectra were initially recorded, followed by the samples ’ emission s- pectra. The ratio of the sample to the background spectrum gives the absorbance spectrum of the sample. The absorbance spectra are then converted into transmittance. The resulting transmittance spectral Table 3 Description of mixture details. Symbol Waste material Quantity (g) Fine aggregate Quantity (g) NaOH solution (wt.% of WCC) Na2SiO3 solution (wt.% of WCC) Water* (wt.% of WCC) SM Na2O (%) 6 M1.5AM WCC 760 RS 2280 32 49 57 1.45 11.2 8 M1.5AM WCC 760 RS 2280 32 49 56 1.28 12.7 M1.5AM WCC 760 RS 2280 32 49 54 1.16 14 M1.5AM WCC 760 RS 2280 32 49 53 1.07 15.2 M1.5AM WCC 760 RS 2280 32 49 52 0.99 16.3 M1.5AM WCC 760 RS 2280 32 49 51 0.94 17.3 6 M2.5AM WCC 760 RS 2280 23 58 56 1.86 10.4 8 M2.5AM WCC 760 RS 2280 23 58 55 1.68 11.5 M2.5AM WCC 760 RS 2280 23 58 54 1.54 12.5 M2.5AM WCC 760 RS 2280 23 58 53 1.45 13.3 M2.5AM WCC 760 RS 2280 23 58 52 1.37 14.1 M2.5AM WCC 760 RS 2280 23 58 51 1.30 14.8 6 M3.5AM WCC 760 RS 2280 18 63 55 2.10 10 8 M3.5AM WCC 760 RS 2280 18 63 54 1.91 11 M3.5AM WCC 760 RS 2280 18 63 53 1.81 11.6 M3.5AM WCC 760 RS 2280 18 63 53 1.71 12.3 M3.5AM WCC 760 RS 2280 18 63 52 1.63 12.9 M3.5AM WCC 760 RS 2280 18 63 51 1.57 13.4 *Water – The total amount of water (from alkali solutions) present in each batch of mixture. S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 6graph shows various modes of vibration/rotation/bending in the structural bonds and functional groups. The thermogravimetric analysis (TGA) was used to observe weight loss or mass change when a sample is exposed to a uniform increase in temperature. It is represented as a percent change in mass vs tempera - ture. It also provides information on oxidation, decomposition, ab- sorption, adsorption, and change in phase. Shimadzu DTH-60H was used to analyse the samples for TGA at a rate of 10 •C per minute to 900 •C. The apparatus was initially calibrated at room temperature with an empty crucible for temperature and weight. Then approximately 12–15 mg of the sample was placed in the crucible, and the test was carried out. 2.2.2. Statistical analysis A full factorial design was conducted at a 95% confidence level for ANOVA and F-test using Minitab ® 18 to evaluate the statistical signif - icance of the identified factors influencing the response . Identified factors with their different levels have been listed in Table 5. 3.Results and discussion 3.1. Effect on flowability Fig. 5 presents the variation in flowability. It was observed that the flow or cohesiveness of the fresh mixtures depends on the molarity and AM ratio. Increasing the molarity of the NaOH solution and AM ratio reduces the flowability as well as increases the cohesiveness of the mixture. A similar kind of behaviour was also found by other researchers . With an increase in the Na2SiO3 content, the total amount of water in the mixture remains almost the same (Table 3), but the incremental presence of a highly viscous Na2SiO3 solution increases the AM ratio, and results in a reduction of the flowability . Although maintaining a total solid (precursor and fine aggregate) to alkali solution ratio at 0.2, the overall flowability remains above 105%, which is aligned with the requirements specified in ASTM C1437 . 6 M1.5AM and 16 M2.5AM mixture reported the maximum and minimum flow of 140% and 102%, respectively. The possible reason for the 102% flow for the 16 M2.5AM mixture can be attributed to the mixing of a greater quantity of finer particles in the fine aggregate. Ghosh and Ghosh reported that increasing the water content increases the flow of the mixture by reducing the viscosity and results in delayed geopolymerization and lower strength . Table 3 and Fig. 5 show that Na2O content also plays an essential role in the flow percentage. For each group of AM ratios with an increase in the Na2O content, the flow percentage decreases due to an increase in the viscosity. Further, higher Na2O and lower water content increased the rate of geopolymerization and initi- ated the rapid deterioration of the plasticity of the mixture . 3.2. Effect on the setting time Fig. 6 shows the variations in the initial and final setting time of the mixtures with the variation in the molarity of NaOH and AM ratio. Higher initial and final setting time is seen for low molarity and AM ratio mixes. A significantly small quantity of Na2O slows down the geo- polymerization process at ambient temperature and takes a longer time to set . In contrast, increased molarity as well as Na2O content in- tensifies the geopolymerization and lowers both setting times. This Table 4 Ultrasonic Pulse Velocity (UPV) and Dynamic Modulus of Elasticity (Edyn) values for different mixtures at different curing age. Symbol Average Ultrasonic Pulse Velocity (km/s) Average Dynamic Modulus of Elasticity, Edyn (GPa) 28D 56D 90D 28D 56D 90D 6 M1.5AM 2.92 3.01 3.01 20.72 21.28 21.11 8 M1.5AM 3.23 3.31 3.27 20.78 22.02 21.86 10 M1.5AM 3.29 3.39 3.46 22.35 23.82 24.85 12 M1.5AM 3.58 3.72 3.72 26.25 27.59 27.94 14 M1.5AM 3.49 3.55 3.61 25.62 25.55 26.67 16 M1.5AM 3.61 3.57 3.72 27.08 26.85 28.37 6 M2.5AM 3.29 3.28 3.26 21.29 21.62 21.35 8 M2.5AM 3.61 3.75 3.75 26.35 28.97 28.86 10 M2.5AM 4.07 4.15 4.21 34.02 34.91 36.10 12 M2.5AM 3.71 3.67 3.76 29.11 29.25 30.42 14 M2.5AM 3.47 3.55 3.56 25.16 25.72 25.54 16 M2.5AM 3.61 3.71 3.81 26.85 28.11 29.27 6 M3.5AM 3.03 3.12 3.09 18.21 19.51 19.50 8 M3.5AM 3.48 3.53 3.46 24.74 24.47 24.68 10 M3.5AM 3.45 3.47 3.61 24.36 24.76 26.25 12 M3.5AM 3.75 3.85 3.89 29.01 30.02 31.31 14 M3.5AM 3.49 3.66 3.75 24.83 27.56 28.75 16 M3.5AM 3.57 3.62 3.75 25.64 27.23 28.41 Table 5 Factors and levels for ANOVA and F-test. Factor Level 1 Level 2 Level 3 Level 4 Level 5 Level 6 M 6 8 10 12 14 16 AM 1.5 2.5 3.5 D 3 7 28 56 90 Fig. 5.Flowability of different mixtures. Fig. 6.Variation in setting time of different mixtures. S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 7behaviour of geopolymer is in good agreement with the findings of other researchers where they had reported reduced setting times when the Na2O content is increased in the mixture [64,65] . Though from the point of ease of application of the mixture, a higher initial setting time is desired but increasing the setting time with water reduces the compressive strength. Hence, it can be concluded that an increased setting time can be achieved with lower molarity of NaOH solution or by lowering the Na2O content in its solution with Na2SiO3 when the two alkalis were used for geopolymerization. However, it also affects other properties viz. compressive and flexural strengths. 3.3. Compressive strength The change in the compressive strength for different molarity (from 6 M to 16 M) with the variation in the AM (from 1.5 to 3.5) ratios is presented in Fig. 7. Observations from Fig. 7 show that both molarity of the NaOH solution and curing age are the influential factors for the generation of compressive strength when the curing temperature, pre- cursor, and fine aggregate are kept fixed. It is a proven fact that higher alkalinity (higher amount of Na‡cations) increases the dissolution of Si and Al from the aluminosilicate source [61,66] . However, excess hy- droxide (OH–) ions also lead to rapid precipitation of the aluminosili - cates at an early age and hinder the availability of Si and Al in the mixtures for geopolymerization at different curing ages . Our study shows a maximum compressive strength of 20.55, 20.65 and 20.70 MPa at 28D, 56D and 90D for the 10 M2.5AM mixture. Whereas for the same molarity with AM ˆ1.5, it was 12.6, 17.5 and 19 MPa; for AM ˆ3.5, it was 14.5, 14.65 and 15 MPa at 28D, 56D and 90D casting age, respectively. Interestingly, Fig. 7 shows no specific pattern with the variation in molarity and AM ratio. However, increasing the Na2SiO3 content and the molarity of NaOH solution leads to lower compressive strength. A considerable variation in the compressive strength between 7D and 28D was observed for mixtures with both AM 1.5 and 3.5, which shows the delayed geopolymerization effect. Whereas the mixtures having AM ˆ2.5, the variation was more diminutive, and initial strength gain was much higher than the other two AM ratio mixtures. The effect of SiO2/Na 2O (SM) was calculated and analyzed. Fig. 8 shows the variation in compressive strength with the SiO2/Na 2O ratio value in ascending order. The optimum SiO2/Na 2O value is 1.54 (10 M2.5AM), where the maximum compressive strength was observed. Moreover, the second-highest compressive strength was obtained around SiO2/Na 2O ratio of 0.94. Fig. 8 indicates a threshold value of SiO2/Na 2O ratio is required to get the optimum result. Similar obser - vations regarding the threshold limit are also reported by other re- searchers [68,69] . Cho et al. found that increasing the SiO2/Na 2O from 0.8 to 2.0 decreases the compressive strength due to an increase in the percentage of porosity, and the optimum SiO2/Na 2O value is 1.4 when fly ash of class F grade was utilized as a precursor . Our study also found the optimum SiO2/Na 2O content to be around 1.54, while increasing the Na2SiO3 content simultaneously increases the SiO2/Na 2O ratio value, decreasing the compressive strength by reducing the reac- tivity of the geopolymeric gel . Further, depending on the molarity and AM ratio, Na2O% is calcu - lated for each mixture, and the corresponding compressive strength with age is presented in Fig. 9 in ascending order of Na2O%. While it is well known that increasing the Na2O% enhances the geopolymerization by increasing the solubility of amorphous silica and alumina from the precursor but also the high rate of dissolution of silica and alumina from the precursor could not end up to the geopolymeric chain and causes a blocking effect to the synthesis of geopolymers and further strength gain [69,70] . It can be observed from Fig. 9 that 12.5% Na2O content leads to maximum compressive strength, beyond which there is a drastic fall in the compressive strength, and again it increases at 17.5%. Further variation in compressive strength with increment in the percentage of Na2O content could not be studied as the scope of this research is limited Fig. 7.Change in compressive strength for different mixtures with curing age. Fig. 8.Change in compressive strength with the variation in SiO2/ Na2O content. Fig. 9.Change in compressive strength with the variation in Na2O content. S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 8to this level. So it can be concluded that the optimum Na2O content is 12.5% for the highest compressive strength. 3.4. Flexural strength Fig. 10 shows the change in the flexural strength of each mixture. The maximum and minimum flexural strength was also achieved by 10 M2.5AM and 6 M1.5AM mixture similar to the observations for the compressive strength. It is observed from Fig. 10 that increasing the molarity can achieve higher flexural strength for all the three different Na2SiO3 content, but AM2.5 mixtures almost attain its 60% of 28D flexural strength within the first 7D of curing. In contrast with other silicate content mixtures, the initial 7D flexural strength was merely 25–35% of its 28D flexural strength. It is also observed that AM2.5 is the optimum Na2SiO3 content. Any variation from it results in delayed geopolymerization, which is also observed from the flexural strength difference between 7D and 28D, where there is no considerable surge in the flexural strength. A best-fitted correlation plot between compressive and flexural strength depicts a COD (R2) value of 0.96, proving an excellent corre - lation between the physical strength parameters computed through destructive testing. Also, the correlation can be helpful to ascertain the flexural or compressive strength from one test result without any destructive testing for the other one. 3.5. Drying shrinkage analysis Shrinkage in specimens occurs due to drying when there is a differ - ence between the atmospheric humidity and surface moisture of the specimen. Less atmospheric humidity than the surface of the specimen absorbs the water from the specimen, which further leads to a reduction of the water content within the specimen and reduces the water voids and changes the shape and size of the specimen. Every measurement for a specific specimen was conducted for a particular side and direction which was marked on the specimen at the very first length comparator test to minimize the error due to the warping effect with curing age . The drying shrinkage variation of each sample is presented in Fig. 11. It is seen that increasing the AM ratio from 1.5 to 3.5 increases the maximum shrinkage percentage with age with the variation in the molarity of NaOH solution. The minimum shrinkage value was obtained by 10 M2.5 AM and 12 M2.5AM, whereas the maximum was by 16 M3.5AM. Drying shrinkage results were obtained till 180D of casting to get the ultimate drying shrinkage value. During experimentation, it was observed that all the geopolymers took an average of 90D to stabilize the drying shrinkage. Beyond 90D, the drying shrinkage value becomes stable. As the quantity of precursor and fine aggregate were fixed so the Fig. 10.Flexural strength plot of different mixtures. Fig. 11.Change in the drying shrinkage values for different mixtures with curing age. S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 9variation in drying shrinkage was mainly governed by the alkali content and secondarily by the particle size of the fine aggregates. High Na2SiO3 content increases the cohesiveness of the mixture, which could be observed from the flowability of the mixtures (Fig. 5). From Fig. 11, the maximum drying shrinkage was observed for AM1.5 mixtures as 0.002% after 180D, whereas for AM3.5 mixtures, the minimum drying shrinkage value as 0.0022% after 180D irrespective of the molarity of NaOH. The high cohesiveness of the fresh mixture and adhesive nature of Na2SiO3 tend to create honeycombs, which with age intensifies the drying shrinkage. Further, it was observed that 10 M and 12 M mixtures resulted in the most negligible drying shrinkage value and could be taken as the optimum molarity to get a minor drying shrinkage effect. 3.6. Ultrasonic pulse velocity and Dynamic modulus of Elasticity Table 4 presents the mean of UPV (km/s) and Edyn (GPa) of three test specimens for each symbol. Where it was observed that the UPV in- creases after 28D till 56D but beyond 56D until 90D it remains almost the same. A similar phenomenon was observed for Edyn also as it had been computed using the method mentioned in BS EN 12504 –4:2004 using the UPV values. Fig. 12 shows that the UPV value increases with the curing age, though the increment is minimal for almost half of the mixtures. This phenomenon justifies that with curing age at room temperature curing conditions, the geopolymerization rate was slow but steady when WCC was used as a sole precursor. Maximum UPV value is observed for 10 M 2.5AM mixture, and the same mixture shows the highest compressive strength. Other than 6 M mixtures, almost all different mixtures reported UPV within a range of 3.5 to 3.75 km/s which depicts a good quality zone . The R2 value was found to be 0.77 when the UPV was compared with the compressive strength for the test results obtained at 28D, 56D and 90D age. Dynamic modulus of elasticity (Edyn) is also an essential factor in assessing the mortar quality and have been calculated as mentioned under section 2.2.1.1. The lowest and highest Edyn value is shown by 6 M3.5AM and 10 M2.5AM respectively. Variation in the Edyn value represents the stiffness of the mortar, where lower values indicate lower stiffness due to the presence of porosity in it and vice versa. Fig. 13 shows that Edyn value increases with the curing age, indicating increased compactness of the mixture with age. This was possible due to the delayed geopolymerization of the WCC due to the absence of heat curing. Further, analysing Fig. 13, it is observed that for AM ˆ1.5, increasing the molarity, the Edyn value increases, which indicates the increased compactness of the final product. Whereas for AM 2.5 and 3.5, the outcome was mixed with the variation in the molarity of NaOH so- lution. So, increasing the AM ratio with the molarity does not affect the Edyn value significantly, and it remains within a range of 25 to 30 GPa. However, AM2.5 mixtures show better Edyn values when compared to AM3.5 mixtures, which signifies the effect of Na2SiO3. Increasing the Na2SiO3 content leads to a highly porous structure and simultaneously decreases the Edyn value. An optimum value for Edyn was obtained for AM 2.5 mixtures among the three different Na2SiO3 content. An R2 value of 0.76 was found when the compressive strength and Edyn values were correlated. 3.7. Statistical analysis of the results For the statistical assessment of the different factors affecting the geopolymer mortar compressive and flexural strength, ANOVA and F- test were conducted. A confidence level of 95% was assigned. A higher F-value demonstrates the variation of the parameters, which signifi - cantly affects the samples ’ performance characteristics . A full factorial design with three factors and different levels of each factor was applied to analyse the influence of the factors. In Table 5, the factors and their levels are listed where M is the molarity of the NaOH solution; AM is the mixing ratio of Na2SiO3 solution to NaOH solution; and D is the age of casting in days. Tables 6 and 7 represents the ANOVA results of the geopolymers mortars concerning their compressive and flexural strength, respectively. From Table 6, it is observed that both M and D have a significant effect on the compressive strength with a p-val - ueD0.05, whereas AM is not significant. It further supports the destructive test results, where no pattern was found between M and AM Fig. 12.Change in the UPV values for different mixtures at 28D, 56D and 90D curing age. Fig. 13.Change in the Edyn values for different mixtures at 28D, 56D and 90D curing age. Table 6 Results of ANOVA for compressive strength (MPa). Factor SS dF MS F-value p-value Significance M 89.70 1 89.70 17.67 0.00 D0.05 Significant AM 0.73 1 0.73 0.15 0.71 D 2360.30 1 2360.30 464.89 0.00 D0.05 Significant M ×M 23.32 1 23.32 4.59 0.03 D0.05 Significant AM ×AM 102.86 1 102.86 20.26 0.00 D0.05 Significant D ×D 598.16 1 598.16 117.81 0.00 D0.05 Significant M ×AM 13.31 1 13.31 2.62 0.11 M ×D 8.23 1 8.23 1.62 0.21 AM ×D 8.08 1 8.08 1.59 0.21 Error 406.17 80 5.08 Total 3610.86 89 S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 10as shown in Fig. 7. For the predicted compressive strength (Fig. 14) after the ANOVA, the RMSE was found to be 2.25 MPa, and R2 was 0.87, which shows a good level of correlation. Similarly, from Table 7, M and D have a significant effect on the response of the experiment regarding flexural strength. Also, from Fig. 14, the predicted flexural strength has an R2 value of 0.90 with RMSE as 0.49 MPa. Fig. 15 shows the main and interaction effect of SM and Na2O regarding the compressive strength of the geopolymer mortar. Though the factors alone are not significant, the effect of their interaction is significant, which is observed from the Pareto chart. Also, a zone was found after the ANOVA test, which results in the achievement of the maximum compressive strength. 4.Microstructural analysis Microstructure analysis was carried out to complement the assess - ment and correlate the observations made during mechanical testing. For all the microstructural analyses, 28D cured specimens were used. The samples were collected from different parts of the crushed samples after the compression testing. Three samples, i.e., 10 M2.5AM, 12 M1.5AM and 16 M1.5AM, were analysed for microstructure analysis. 4.1. Microstructure examination by SEM and elemental analysis by EDX For a proper understanding of the microstructure, three different mix proportions (10 M2.5AM, 16 M1.5AM and 12 M1.5AM) were identified depending on the highest compressive strength and a series of SEM images were taken and EDX were conducted. The results are presented in Fig. 16 and Table 8. Fig. 16 presents two series of SEM images where the first series (a, b, and c) shows the crushed/semi-powdered specimen images, and the second series (d, e, and f) shows the saw-cut surface images of the identical specimen. Fig. 16d presents the most compact microstructure for the batch designated as 10 M2.5AM. In contrast, Fig. 16e shows multiple levels of voids with fracture lines between the sand particles and geopolymer paste which reveal an improper bonding between them and consequently result into lower mechanical strength. Fig. 16f shows a relatively compact structure. Fig. 16e shows a clear interfacial transition zone between the sand particle and geopolymer paste. From Fig. 16b and c, the samples prepared by using a 1.5AM ratio presented voids of different sizes and show a weaker bonding between the particles. Whereas the 2.5AM mixture images (Fig. 16a and d) show the most compact matrix. These voids can be directly correlated to their compressive strength, where it was found that the most compact matrix represents the highest compressive strength at different curing ages. Upon close inspection, multiple hairline cracks are observed for each mixture due to the drying shrinkage effect. Though the 10 M2.5AM mixture achieved the highest physical strength through proper bonding, drying shrinkage cracks were visible within the geopolymer paste area. Along with the sialate bonds, the high concentration of NaOH can dissolve the gypsum present in the mixture and convert it into calcium hydroxide (Eq (5)) . This also strengthens the pozzolanic reactivity Table 7 Results of ANOVA for flexural strength (MPa). Factor SS dF MS F-value p-value Significance M 3.29 1 3.29 13.48 0.00 D0.05 Significant AM 0.08 1 0.08 0.35 0.55 D 99.82 1 99.82 621.95 0.00 D0.05 Significant M ×M 12.31 1 7.31 32.02 0.04 D0.05 Significant AM ×AM 11.75 1 11.75 48.02 0.00 D0.05 Significant D ×D 41.12 1 41.12 168.04 0.00 D0.05 Significant M ×AM 0.12 1 0.12 0.48 0.49 M ×D 0.14 1 0.14 0.58 0.45 AM ×D 0.01 1 0.01 0.02 0.91 Error 19.58 80 0.24 Total 188.22 89 Fig. 14.(a) Pareto chart of the standardized effects of compressive strength depending on the factors (molarity, AM and curing age); (b) Best fitted plot for compressive strength; (c) Pareto chart of the standardized effects of flexural strength depending on the factors (molarity, AM and curing age); (d) Best fitted plot for flexural strength. S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 11between the formed calcium hydroxide and aluminosilicates and leads to calcium silicate hydrate gel . CaSO 4∙2H 2O…s†‡NaOH…aq†→4Ca…OH†2…s†‡Na2SO 4…aq† (5) From the examination of the parent WCC through EDX analysis, it is found that silica (Si), alumina (Al) and calcium (Ca) are the predominant minerals present. From Table 8 it is observed that after geo- polymerization there is a huge change in the atomic percentage of the minerals. This can be attributed to the ionic reactions that happen within the minerals due to the formation of covalent bonding of the polymers and hence, variations in the cohesiveness and strength prop- erties are observed . The increase in the quantity of sodium (Na) is due to the alkalis present in the mixtures for geopolymerization. From Table 8 it is observed that the Si/Al ratio value maintains 16 M1.5AM (1.80) D10 M2.5AM (3.55) D12 M1.5AM (5.54) DWCC (6.48). The optimum strength was obtained by Si/Al ˆ3.55, which is almost equivalent to the poly-sialate di-siloxo formation of the geopolymers, whose theoretical value stands at Si/Al ˆ3. The second highest compressive strength is obtained by Si/Al ˆ1.80, which stands close to poly-sialate siloxo, i.e., Si/Al ˆ2. The least compressive strength is observed for Si/Al ˆ5.54, which comes under poly-sialate multi-siloxo, i.e., Si/Al F3 among the three-mix proportion analysed. Ca in the geopolymer mixture also reinforces the geopolymer network by acting as a charge balancing cation. From Table 8, the optimum Ca/Si ratio is found to be 0.24 and it is aligned with the research findings reported by other researchers [76,77] . 4.2. X-ray Diffraction (XRD) analysis Fig. 17 shows the XRD pattern of the three selected samples at 28D curing age along with a comparison with the WCC. The peaks observed at the Bragg ’s angle (2θ) values of 21•, 26.5•, 36.5•, 43.5•, 60.1•and 68.3•are for the different crystal planes of quartz which also suggest that the predominant mineral in WCC is quartz [78,79] . These high- intensity peaks for all the three mixtures at the Bragg ’s angle mentioned above are due to the quartz, which remains in an inert phase and does not participate in the geopolymerization and further addition of fine aggregate to WCC for mortar preparation intensifies the peaks of quartz [79,80] . Nevertheless, it was also reported by Lecomte et al. that though quartz remains in an inert phase and high molarity could reduce its intensity due to dilution effect, which can be observed for 16 M1.5AM mixture at 21•Bragg ’s angle . Furthermore, small intensity peaks are observed in between 20•to 30•which indicate the formation of geopolymers (except the quartz peaks at 21•and 26.5•) . Also, it has been reported by researchers that halo peaks around 28•to 30•repre - sent the predictable amorphous structures of geopolymers . Fig. 17 also depicts that, after alkali activation of WCC, the peaks shifted towards the right (i.e., towards higher Bragg ’s angle value). This phenomenon indicates the formation of the amorphous sodium alumi - nosilicates which is the key binding agent in geopolymers. High- intensity peaks for the 10 M2.5AM combination at higher Bragg ’s angle values indicate optimum geopolymer production, which is also demonstrated during mechanical testing by achieving maximum strength among the various mixtures. Also, peaks observed at Bragg ’s angle 32•, 42.5•, 50.5•and 52•correspond to zeolites which get formed due to geopolymerization upon alkali activation, and it is also observed that with the increase in the molarity of the mixture, the intensity of the peaks reduces . Similar behaviour upon alkali activation was observed by Heah et al. . Fig. 15.(a) Main effects plot for compressive strength; (b) Pareto chart of the standardized effects of compressive strength depending on the factors (SM and Na2O %); (c) Interaction effects plot for compressive strength; (d) Contour plot of compressive strength vs SM, Na2O%. S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 124.3. Fourier transform infrared spectroscopy (FTIR) analysis Fig. 18 presents the FTIR analysis of WCC and geopolymers. The band assignments to different types of bonding are tabulated in Table 9. The change or shift of FTIR spectra before and after geopolymerization of WCC are observed primarily below 2000 cm1, around 2350 cm1 and within 3500 –4000 cm1 wavenumbers. Spectra below 2000 cm1 are sharper for WCC than the three geopolymers (10 M2.5AM, 12 M1.5AM and 16 M1.5AM), confirming more amorphous nature of the geopolymers in comparison to WCC. The bands at higher wavenumbers, i.e., 3500 –4000 cm1, are assigned to the stretching vibrations of the O–H bond, which has sharp and broad peaks for the geopolymers due to the presence of water molecules within the geopolymers. A flat peak occurred around 2350 cm1 due to the formation of C-O vibrational bonds in CO2 constricted in the amorphous phase of the geopolymer . This kind of behaviour is attributed to the use of WCC as a pre- cursor. The subsequent sharp peak was observed between 1500 and 1700 cm1 related to the stretching vibration of the O-C-O bonds. It is observed that with this range, the maximum and minimum intensity peak is obtained by 16 M1.5AM and WCC, respectively. A higher peak is attributed to the capacity of the material to absorb environmental CO2 due to the presence of Ca within the precursor and formed CaCO 3 [66,85] . And peaks around 1463 cm1 close to the 1500 –1700 cm1 range is assigned to the C–O stretching vibrational bond, which is assigned to the formation of CaCO 3. Further, it is also observed that the lesser the peak intensity, the higher the compressive strength. This phenomenon can also be restated as the lower the CO2 absorption ca- pacity higher will be the compressive strength for geopolymers. Samples Fig. 16.SEM images of 28D sample of 10 M2.5AM (a and d), 12 M1.5AM (b and e) and 16 M1.5AM (c and f). Table 8 Change in the elemental atomic percentage before and after geopolymerization. Symbol Minerals (Atomic, wt. %) Si Al Ca Na K Fe Si/Al Na/Al Ca/Si WCC 14.90 2.30 5.76 0 0.31 0.31 6.48 0 0.38 16 M1.5AM 0.56 0.31 0.28 2.47 0.07 0.23 1.80 7.96 0.50 12 M1.5AM 1.33 0.24 0.28 2.93 0.07 0.18 5.54 12.21 0.21 10 M2.5AM 1.21 0.34 0.29 2.67 0.11 0.22 3.55 7.85 0.24 S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 13showing a sharp peak centred around 1000 cm1 is assigned to the vibrational stretching mode of Si-O-Si of SiO4 tetrahedral structure [24,25,86] . The key observation for the wavenumbers D1000 cm1 before and after geopolymerization of WCC is the peak formation pattern of Si-O vibrational bonds. A peak broadening for the geo- polymers around 1000 cm1 and shifting of peaks of the geopolymers towards the lower wavenumbers, i.e., around 580–600 cm1, indicates the random formation of Si-Al bonds . This shifting of Si-O related bands to lower wavenumber is due to the depolymerization of silicates and the exchange of Si with the Al atoms. A similar kind of behaviour was also reported by other researchers when using WCC as a precursor for geopolymers [4,87-89] . 4.4. Thermogravimetric analysis (TGA) Fig. 19 shows the TGA graph of the specimens before and after geopolymerization of the parent material (WCC) at different molarity (10 M, 12 M and 16 M) and AM ratios (1.5 and 2.5). TGA and DTG graphs (Fig. 20) show three major areas of mass loss for the WCC which correspond to three temperature ranges: i) room temperature 200•C, ii) 300-400•C, and iii) 600–700 •C They correspond to loss of bound water, dehydration of calcium hydroxide, and decomposition of CaCO 3 [92-94] . The mass loss encountered by WCC for the above-mentioned temperature zones are 1.82%, 3.1% and 5.45%, whereas for the geo- polymers within the same temperature zone the mass loss is 3.22%, 3.36% and 4.18% for 10 M2.5AM; 3.29%, 3.38% and 4.13% for 12 M1.5AM; 3.54%, 3.61% and 4.59% for 16 M1.5AM. The weight loss was higher for the geopolymers to 600 •C than WCC due to the involvement Fig. 17.XRD pattern of 28D sample of WCC, 10 M2.5AM, 12 M1.5AM and 16 M1.5AM. Fig. 18.FTIR spectra of 28D sample of WCC, 10 M2.5AM, 12 M1.5AM and 16 M1.5AM. Table 9 Assignments of different FTIR transmittance peaks. Sl. No. Band range (cm1) Type of bonding Reference 1 525 O-Si-O (symmetric bending) Carrasco et al. 2 500–650 Si-O-Alvi and AlO 6 (Polyhedral form) Z. Zuhua et al. 3 580 Si-OH and Siloxane backbone (vibrational) H. Yoshino et al. 4 780, 1040 Si-O-Si (stretching vibration) Allahvardi and Kani , F. Rubio et al. 5 1463 C–O stretching vibrational bond may be attributed to CaCO 3 I. García Lodeiro et al. 6 1500 – 1700 O-C-O stretching vibrational bond S.K. Das and S. Shrivastava 7 2350 C-O vibrational bond Z. Li and L. Zhang 8 3600 –3850 O–H stretching vibrational bond S.K. Das and S. Shrivastava Fig. 19.TGA curve of 28D sample of WCC, 10 M2.5AM, 12 M1.5AM and 16 M1.5AM. S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 14of multiple reaction phases within the geopolymer. These reaction phases can be summarized as (a) up to 250–300 •C the weight loss is predominantly for the evaporation of molecularly bound water and zeolitic water present within the geopolymers and the de-hydroxylation phase of the N-A-S-H gel available within the geopolymers [95,96] ; (b) beyond 300 •C to 600 •C the de-carbonation governs the weight loss and de-hydroxylation of secondary carbonates and crystalline zeolitic pha- ses, respectively [97,98] . Further, above 600 •C, a sudden weight loss is observed for WCC due to the decomposition of CaCO 3 in higher quantity than the geopolymers which is attributed to the existence of small amount of CaCO 3. Similar observation is also reported by other re- searchers using slag as a precursor for geopolymer production [98-100] . However, in our study, 16 M1.5AM shows the maximum mass loss beyond 600 •C. The same phenomenon is also observed in the FTIR analysis (Fig. 18). 5.Conclusion A thorough experimental and microstructural analysis was con- ducted to understand the geopolymeric potential of waste cement con- crete used as a single precursor for geopolymer production. A series of destructive and non-destructive tests were conducted at different curing ages. The summary and critical conclusions of this study are mentioned below: ≡Increasing the amount of Na2SiO3 decreases the flowability and setting time of the mixtures by increasing their cohesiveness and rapid hardening. It impacts the workable time. However, it does not significantly affect the compressive or flexural strength. A lower quantity of Na2SiO3 would be helpful to have good workability and workable time. Also, a substantial increase in the amount of Na2SiO3 increases the drying shrinkage effect in the mortar. ≡12.5% Na2O was found to be the optimum limit for WCC based geopolymers. Further variation from it reduces the mechanical strengths. While for silicate modulus, SiO2/Na 2O ratios between 1 and 1.5 were found optimal. ≡An alkali mixture ratio of 2.5 is optimum among the three ratios as proved by experimental observations. It expedites the geo- polymerization by gaining good compressive strength within 7D of casting. In comparison, 1.5AM and 3.5AM delays the geo- polymerization. Optimum molarity is found to be within 10 M to 12 M. ≡ANOVA and F-test validates molarity and curing age as a significant factor influencing the mechanical strength. In contrast, AM is not a significant factor for achieving the same. ≡The presence of significant amount of CaO in WCC results in the formation of a hybrid geopolymer by producing both the C-S-H gel and the poly-sialate phase. At ambient temperature curing condi - tions, the former coexists with the geopolymers and enhances the particle connection. Overall, based on the results of the experiments, 100% WCC may be utilized as a single precursor for manufacturing geopolymers at ambient temperature curing, broadening its applicability from precast material to cast-in-situ applications. However, the presence of CaO and the availability of C-S-H gel phase may affect the durability of the manu - factured geopolymers, which must be extensively investigated. 6.Data statement All the data related to this article can be obtained from the corre - sponding author upon stating the reason for collecting the data through email. 7.Funding information This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Fig. 20.TGA with DTG curve of 28D sample of WCC, 10 M2.5AM, 12 M1.5AM and 16 M1.5AM. S.K. Das and S. Shrivastava Construction and Building Materials 301 (2021) 124380 15CRediT authorship contribution statement Sourav Kumar Das: Investigation, Methodology, Writing - original draft. Sandeep Shrivastava: Conceptualization, Methodology, Super - vision, Writing - review & editing. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements Authors would like to acknowledge the help provided by NCESS, Thiruvananthapuram for XRF analysis; Central Instrumentation Facility, MIT Manipal and Sophisticated Analytical Instrument Facility, Manipal University Jaipur for SEM and EDX analysis; Central Analytical Facility, Manipal University Jaipur for FTIR and TGA analysis; and Manipal University Jaipur for providing their Material Testing Laboratory resources.

Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 5 | May -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1306 EXPERIMENTAL STUDY OF FACTORS INFLUENCING COMPRESSIVE STRENGTH OF GEOPOLYMER CONCRETE Ajay Sharma 1, Juned Ahmad 2 1 PG Student, Structural Engineering, Integral University, Uttar Pradesh, India 2 Assistant professor, Department Of Civil Engineering, Int egral University, Uttar Pradesh, India ------------------------------------------------------------------- --***-------------------------------------- ------------------------------- Abstract - Manufacture of Portland cement produces large of volumes of carb on dioxide and other gases. Releasing these gases causes atmospheric pollution and subsequent environmental degradation. Finding a suitable alternative solution to mitigate the environmental degradation caused by using Portland cement is very important for environmental sustainability. On the other side, fly ash is the waste material of coal based thermal power plant, available abundantly but pose disposal problem. There are environmental benefits in reducing the use of Portland cement in concrete, and usin g a cementitious material, such as fly ash. Geopolymer concrete is new sustainable concrete which is manufactured by replacing cement 100% with processed fly ash which is chemically activated by alkaline solutions made from sodium silicate (Na2SiO3) and so dium hydroxide (NaOH). This thesis presents the effect of several factors like alkaline liquid to fly ash ratio, molarity of NaOH, curing hours and curing temperature on the compressive strength of fly ash based geopolymer concrete. Fly ash is taken as the basic material to develop the geopolymer concrete and it is activated by the alkaline solution of sodium silicate and sodium hydroxide. The test variables were molarities of sodium hydroxide (NaOH) 12M, 4M, 16M, and 18M, ratio of NaOH to sodium silicate (Na2SiO3) 2.5, alkaline liquid to fly ash ratio 0.35, 0.40, 0.45 and 0.50 were used in the present study. The experiment were also conducted on GPC cubes for curing temperature of 75° C, 90° C and 105° C with curing period of 12, 18 and 24 hours by adoptin g hot oven curing method. The test result indicated that the compressive strength increases with increase in molarity of NaOH but it decreases with increases in water content. It is also absorbed that compressive strength is remarkably affected by the curi ng hours and curing temperature. When curing temperature is increases, the compressive strength is also increases and it requires less curing period to gain the higher strength. 1.INTRODUCTION 1.1 General Production of cement is one of the major contr ibutors to the emission of green -house gasses like carbon dioxide. Day by day the World’s Portland cement production increases with the increasing demand of construction industry. Cement is the main ingredient for the production of concrete. But the produc tion of cement requires large amount of raw material. During the production of cement burning of lime stone take place which results in emission of carbon dioxide (CO2) gas into the atmosphere. There are two different sources of CO2 emission during cement production. Combustion of fossil fuels to operate the rotary kiln is the largest source and other one is the chemical process of burning limestone. In 1995 the production of cement was 1.5 billion tons which goes on increasing up to 2.2 billion tons in 201 0. One ton of production of cement causes one ton of emission of CO2 into the atmosphere. It is e stimated that the emission of carbon dioxide due to cement production to be nearly about 7% of the total production of carbon dioxide, which make required to g o for other greener alternative binder from Portland cement [1 ]. Fly ash is the waste residue that results from the combustion of coal in thermal power station is available at large scale all over the world. In India more than 100 million tons of fly as h is produced annually. Out of this, only 17 – 20% is utilized either in concrete or in stabilization of soil. Most of the fly ash is disposed off as a waste material that coves several hectors of valuable land. As fly ash is light in weight and easily fli es, this creates severe health problems like asthma, bronchitis, and so forth. There are environmental benefits in reducing the use of Portland cement in concrete, and using a by -product material, such as fly ash as a substitute. With silicon and aluminium as the main constituents, fly ash has great potential as a cement replacing material in concrete. For every ton of fly ash used in place of Portland cement saves about a ton of carbon dioxide emission to the atmosphere [2 ]. Davidovits proposed a new term geopolymer in 1978 to represent the mineral polymers resulting from geochemistry. Geopolymers are members of the family of inorganic polymers in which the mineral molecules ar e linked with covalent bond. Geopolymers are produced by source materials or by -product of geological origin that is rich in silica and alumina like fly -ash when react with alkaline solution at elevated temperature. The chemical reaction that takes place in this case is polymerization, so this binder is called geopolymer. Geopolymer concr ete is a new type of concrete in which cement is fully replaced by the pozzolanic materials that is rich in Silicon (Si) and Aluminium (Al) like fly ash. It is activated by highly alkaline liquids to produce the binder which binds the aggregates in concrete when subjected to elevated temperature. The Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 5 | May -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1307 concrete made with such industrial waste is eco -friendly and so it is called as “Green concrete”. The chemical composition of the geopolyme r material is similar to zeolitic materials, but t he microstructu re is irregular. The polymerization process involves a fast chemical reaction under alkaline condition on Si -Al minerals, those results in a three -dimensional polymeric chai n and ring structure existing of Si -O-Al-O bonds. Poly (sialates) as the resulted m aterial has the empirical formula as below: Mn [ -(SiO2)z -AlO2]n. wH2O Where: M = the alkaline element or cation such as potassium, sodium or calcium n = the degree of polycondensation or polymerisation z = 1,2,3 or higher The main concept behind this geopolymer material is the polymerization of the Si -O-Al-O bond which develops when Al-Si source materials like Fly ash is mixed with alkaline liquids . The geopolymer can be one of these basic form:  Poly (sialate), [-Si-O-Al-O-]  Poly (sialate -siloxo), [-Si-O-Al-O-Si-O-]  Poly (sialate -disiloxo), [-Si-O-Al-O-Si-O-Si-O-] The schematic formation of geopolymer mat erial can be shown by equations (1) and (2). n( Si2O5, Al2O2 ) + 2nSiO2 + 4nH2O +NaOH/KOH (Si-Al materials) | | Na+, K+ + n(OH)3 -Si-O-Al--O-Si-(OH | (OH)2 (Geopolymer precursor) ---------------------- n(OH)3 -Si-O-Al--O-Si-(OH)3 + NaOH/KOH | | (OH)2 | ( Na+, K+) -(-Si-O-Al--O-Si-O-) + 4nH2O | | | O O O ( Geopolymer Backbone) ----------------------- The exact mechanism of setting and hardening of the geopolymer material is not clear. However, most proposed mechanism consists the chemical reaction may comprise the following steps: 1. Dissolution of Si and Al atoms from the source material through the action of hydroxide ions. 2. Transportation or orientation or condensation of precurso r ions into monomers. 3. Setting or polymerisation of monome rs into polymeric structures [3 ]. Water is released during the chemical reaction that occurs in the formation of ge opolymers. This water, released from the geopolymer concrete during the rest per iod, oven curing and further drying periods, leaves behind dis continuous nano -pores in the concrete , which provide benefits t o the performance of geopolymer concrete . The water in a geopolymer mixture, plays no role in the chemical react ion that takes plac e; it provides the workability to the mixture during handling [ 4] 1.2 Constituents of Geopolymer Concrete There are two main constituents of Geopolymers, namely the source mate rials and the alkaline liquids. By -product materials such as fly ash, silica fume, slag, r ice husk ash, GGBS, red mud, etc can be used as source materials. The most common alkaline liquid used in g eopolymer concrete is a combination of sodium hydroxide or potassium hydroxide and sodium silicate or potassium silicate. It is importa nt that the alkali ne liquid is prepared by mixing both the solutions together at l east 24 hours prior to use. The mass of N aOH solids in a solution depend on the molarity of the solutio n. 1.3 Advantages of Geo -Polymer Concrete  High early compressive strengt h gain  Good abrasion resistance  Rapid controllable setting and hardening  Fire r esistance up to 1000 degree centigrade and no emission of toxic fumes when heated  High level of resistance to a range of different acids and salt solutions  Low thermal cond uctivity and low shrinkage  Impermeable like normal OPC concrete  Bleed free  High surface definition that replicates mould patterns . 1.4 Objective of the Present S tudy To study the effect of factors like Molarity, alkaline liquid to fly ash ratio, curing hour and curing temperature that affects the compressive strength of fly ash based geopolymer concrete. Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 5 | May -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1308 2 EXPERIMENTAL PROGRAMME 2.1 Materials Used A. Fly A sh In this study, fly ash (ASTM Class F) obtained from Rosa Thermal Power Station Shahajahanpur was used. The chemical and physical properties of fly ash shown below in table 2.1 and 2.2 respectively. Table -1: Chemical Properties of Fly Ash Characteristics Observed Value (%) Range Specified for Class - F Fly ash (as per ASTM C -618) Silicon dioxide (SiO2) 56.31 46-60 Alumina (Al2O3) 31.82 21-28 Iron oxide (Fe2O3) 4.77 5-9 Calcium oxide (CaO) 2.33 0.5-8 Magnesium Oxide (MgO) 1.09 0.2-4 Sulphur trioxide (SO3) 0.16 0-0.4 Sodium oxide (Na2O) 0.042 0-0.3 Potassium oxide (K2O) 0.013 0-0.2 Titanium (TiO2) 2.01 1-2.1 Table -2: Physical Properties of Fly A sh Characteristics Observed Value Specific Gravity 2.18 Fineness(m2/kg) 340 Lime reactivity(N/mm2) 4.8 Loss on Ignition(% by mass) 0.70 Soundness by auto -cleave method 0.12 B. Alkaline S olution The combination of sodium silicate to sodium hydroxide was used as alkaline solution in the present study and the ratio of both was maintained to 2.5 throughout the study. The solution was prepared one day prior to be used. I. Sodium H ydroxide Generally N aOH is available in market in pellets or flakes form with 96% to 98% purity where the cost of the product depends on the purity of the material. S odium hydroxide in pellet form was used in this work of 97% purity. The solution of NaOH was formed by dissolv ing it in deionised water for the molarity of 12M, 14M, 16M & 18M. The NaOH solution was prepared 24 hours before casting of specimens. II. Sodium S ilicate Sodium Silicate ( Na2SiO3 ) is also known as waterglass which is available in the market in gel form and also in the solid form. Sodium silicate in gel form was used in this study having 31% of SiO2, 14% of Na2O & 55% of H2O. The ratio of silicon dioxide ( SiO2) and sodium oxide (Na2O) in sodium silicate gel is 2.21. Fig.-1: Sodium Hydroxide Fig.-2: Sodium Silicate Liquid Pellets 2.2 Mix Proportioning Details The geopolymer concrete was made for sixteen different mix proportion of fly ash, alkaline solution, fine aggregate and coarse aggregate with variation in alkaline liquid to f ly ash ratio as 0.35, 0.40, 0.45 & 0.50 for mo larities of 12M, 14M, 16M, &18M. Table -3: Mix details of fly ash based geopolymer Concerte S.No. Alkaline sol./Fly ash ratio Fly ash (Kg/m3) Fine aggregate (Kg/m3) Mix-1 Mix-2 Mix-3 Mix-4 0.35 566 Mix-5 Mix-6 Mix-7 Mix-8 0.40 556 Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 5 | May -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1309 Mix-9 Mix-10 Mix-11 Mix-12 0.45 546 Mix-13 Mix-14 Mix-15 Mix-16 0.50 536 Coarse aggregate (Kg/m3) Sodium hydroxide (Kg/m3) Sodium silicate (Kg/m3) Molarity (M) 58 145 12 16 1032 165.71 12 16 1013 186.43 12 16 994 207.14 12 16 2.3 Mixing Casting and Curing of Geopolymer Concrete In the laboratory, the fly ash and the aggregates was first mixed together in dry state 2 -4 minutes to get hom ogeneou s mix. The alkaline solution was mixed with the extra water and t his liquid components were added to the mixed aggregate and the mixing continued usually for another 12 - 15 minutes so that binding paste covered all the aggregates and mixture become homogeneous and uniform in colour. The fresh concrete could be handled up to 120 minutes without any sign of setting and without any degradation in the compressive strength. After the mixing is done, the fresh geopolymer concrete was filled in the moulds in three layers with required compaction same as the usual methods used in the case of Portland cement concrete and the specimens are kept on a vibrating table so that to minimize amount of voids present in the fresh concrete . The workability of the fresh concrete was measured by means of the conventional slump test. For the polymerisation process, the high temperature curing is required in geopolymer concrete . The required temperature may be provided either by oven curing or by steam curing. In t he presen t study, oven curing was used. The GPC cubes were placed in an oven for the period of 12, 18 and 24 hours. After the curing period, the test specimens were l eft in the moulds for at least 5 -6 hours in order to avoid a major change in the environmental cond itions. After de-moulding, the concrete specimens were allowed to become air -dry in the laboratory until the day of the compressive strength testing. Fig.-4: Addition of Alkaline Solution Fig.-5: Fresh Geopolymer Concrete Fig.-6: Oven Curing of GPC Cubes Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 5 | May -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1310 2.4 Testing of Geopolymer Concrete C ubes The geopolymer hardened concrete cubes were tested for compressive strength. The compressive strength test was performed according to IS -5160: 1959. Cube specimens of size 150mm × 150mm× 150mm were pre pared for each mix. After one day of rest period, they were cured in oven for 12, 18 & 24 hours and were demoded and stored until the day of testing. For specimens with uneven surfaces, capping was used to minimize the effect of stress concentration. The compressive strength reported is the average of three results obtained from three identical cubes. Fig.-7: Compressive Strength Testing of GPC Cubes 3 . EXPERIMENTAL RESULTS AND DISCUSSION 3.1 Effect of Molarity and AL (Alkaline Liquid)/Fly -Ash Ratio on the Compressive Strength of GPC Table -4: Compressive Strength of Alkaline solution to fly ash ratio=0.35 Mix No. Molarity (M) Compressive Strength (MPa) @ 7 days Average comp. Strength (MPa) Mix -1 Mix-2 Mix-3 Mix-4 12 26.60 26.30 24.90 14 29.70 31.2 32.90 16 35.60 34.8 32.70 18 31.70 33.9 35.90 34.10 Chart -1: Effect of molarity on compressive strength of geopolymer concrete for AL/Fly -ash ra tio 0.35 Table -5: Compressive strength for Alkaline solution to fly ash ratio=0.40 Mix No. Molarity (M) Compressive strength (MPa) @ 7 days Average comp. Strength (MPa) Mix-5 Mix-6 Mix-7 Mix-8 12 34.8 32.1 33.2 14 36.4 36.7 34.6 16 41.8 40.3 38.4 18 37.9 40.8 42.9 Chart -2: Effect of molarity on compressive strength of geopolymer concrete for AL/Fly -Ash ratio 0.40 Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 5 | May -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1311 Table -6: Compressive stre ngth for Alkaline liquid to fly ash ratio=0.45 Table -7: Compressive strength for Alkaline liquid to fly - ash ratio=0.50 Mix No. Molarity (M) Compressive Strength (MPa) @ 7 days Average comp. Strength (MPa) Mix-9 Mix-10 Mix-11 Mix-12 12 36.2 36.8 39.2 14 42.5 40.4 38.2 16 44.4 43.5 43.2 18 43.4 42.9 Chart -3: Effect of molarity on compressive strength of geopolymer concrete for AL/Fly -ash ratio 0.45 Cha rt-4: Effect of molarity on compressive strength of geopolymer concrete for AL/Fly -ash ratio 0.50 The results indicated that the compressive strength increases with increase in molarity of NaOH solution upto 16M but beyond the molarity 16M, slightly var iation in compressive strength is observed for the molarity 18M. Results also show that the compressive strength is increases with increment in alkaline liquid to fly ash ratio but for ratio value beyond 0.45, slightly decrease in compressive strength is o bserved. It means w hen this ratio is increases, the water content in the solution is increases which affect the compressive strength. 3.2 Effect of Curing Hour and Curing Temperature on Compressive Strength of GPC Table -8: Compressive strength of GPC spe cimens at curing temperature of 75°C Mix No. Curing Tempe - rature (°C) Curing Period (hours ) Compressiv e Strength (MPa) Mean Compressiv e Strength (MPa Mix -11 12 26.9 27.8 27.5 18 34.6 34 35.4 24 37.2 37.4 35.1 Mix No. Molarity (M) Compressive Strength (MPa) @ 7 days Average comp. Strength (MPa) Mix-13 Mix-14 Mix-15 Mix-16 12 33.4 33.4 31.9 14 37.9 37.0 35.2 16 39.2 39.7 41.1 18 37.6 38.9 40.1 Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 5 | May -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1312 Chart -5: Effect of curing hours on compressive strength of GPC for temperature 75°C Table -9: Compressive strength of GPC specimens at curing temperature of 90°C Mix No. Curing Tempe - rature (°C) Curing Period (hours) Compressive Strength (MPa) Mean Compressive Strength (MPa) Mix-90 12 38.1 38.4 36.7 18 42.2 42 43.7 24 44.4 43.5 42.9 Chart -6: Effect of curing hours on compressive strength of GPC for temperature 90°C Table -10: Compressive strength of GPC specimens at curing temperature of 105°C Mix No. Curing Tempe - Raptur e (°C) Curing Period (hours ) Compressiv e Strength (MPa) Mean Compressiv e Strength (MPa) Mix -11 12 43.9 42.9 44.7 18 40.7 41.2 40.9 24 37.9 38.7 38.1 Chart -7: Effect of curing hours on compressive strengt h of GPC for temperature 105°C The compressive strength of GPC concrete is observed to increases at curing temperatures 75°C and 90°C for 12, 18 and 24 hr of curing period, but when curing is done at 105°C, compressive strength increasing up to 12 hr of curing, after that it start decreasing. Hence it is observed that when curing temperature is increa sed, it requires less curing hours to gain same compressive strength. 3. CONCLUSIONS Based on the experimental work reported in this study, the following conclusions are drawn: 1. Compressive strength of geopolymer concrete is increases with increase in al kaline solution to fly ash ratio upto 0.45, but at 0.50 the strength slightly decreases because of increase in water content present in alkaline solution. Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 5 | May -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 1313 2. Compressive strength of GPC is also increases with increase in concentration of NaOH up to 16 M, b ut for 18 M, there is no remarkable changes in compressive strength but the cost of production is increases. 3. At 75 °C temperature , the compressive strength increases gradually with the increases in curing period up to 24 hr. 4. At 90°C temperature, the sample s gained 95 % strength at 18 hours of curing, beyond this curing period, minor increment in strength is recorded. 5. Compressive strength of GPC samples at 105°C temperature certain increases up to 12 hours of curing period but it decreases when it cured for 24 hr at 105°C. It has been observed from the above discussion that there are various parameters that affect s the compressive strength of the geopolymer concrete. Therefore, parametric study of variou s factors influncing the compressive strength of the ge opolymer concrete is strongly recommended first before conducting any further investigations related to mechanical properties and durability of the geopolymer concrete in order to get the desirable benefits from the further investigations. ACKNOWLEDGEMENT Author thanks to Asst. Prof. Juned Ahmed and friends for encouraging me to work on this topic and their constant assistant. REFERENCES Shivaji S. Bidwe and Ajay A. Hamane “Effect of different molarities of Sodium Hydroxide solution on the Strength of G eopolymer concrete” American Journal of Engineering Research (AJER) e -ISSN : 2320 -0847 p - ISSN : 2320 -0936 Volume -4, 2015, Issue -3, pp -139-145. Satpute Manesh B et al., “Effect of Duration and Temperature of Curing on Compressive Strength of Geopolymer Conc rete” International Journal of Engineering and Innovative Technology (IJEIT) Volume 1, Issue 5. May 2012. M. M. A. Abdul lah et al “Mechanism and Chemical Reaction of Fly Ash Geopolymer Cement - A Review” Int. J. Pure Appl. Sci. Technol., 6(1) , 2011 , pp. 35 -44. Kolli Ramujee and M. Potharaju “Development of Low Calcium Flyash Based Geopolymer Concrete ” IACSIT International Journal of Engineering and Technology, Vol. 6, No. 1, February 2014, A.Maria Rajesh, M.Adams Joe, Roy Mammen “Study of the Strength Geopol ymer Concrete with Alkaline Solution of Varying Molarity” IOSR Journal of Engineering (IOSRJEN) ISSN (e): 2250 -3021, ISSN (p): 2278 -8719 Vol. 04, Issue, June. 2014, ||V1|| PP 19 -24. Prakash R. Voraa , Urmil V. Daveb “Parametric Studies on Compressive Stre ngth of Geopolymer Concrete” Chemical, Civil and Mechanical Engineering Tracks of 3rd Nirma University International conferenc e on Engineering, 2012. Subhash V. Patankar et al “Effect of water -to- geopolymer binder ratio on the production of fly ash based g eopolymer concrete” Research Scholar, Department of Applied Mechanics, Government Engineering College, Aurangabad, (M.S.) India , 2013. Yasir Sofi, Iftekar Gull “Study Of Properties Of Fly Ash Based Geo Polymer Concrete” A Peer Reviewed International Journa l, Vol.3., Issue.1 , 2015. Raijiwala D.B. Patil H. S. “Geopolymer Concrete: A Concrete Of Next Decade” JERS/Vol.II/ Issue I/January - March 2011/19 -25. Dali Bondar “Geo -polymer Concrete as a New Type of Sustainable Construction Materials” third international conference on sustainable construction materials and technologies. Shriram Marathe et al “A Review on Strength and Durability Studies on Geopolymer Concrete” International Journal of Innovative Research in Science, Engineering and Tec hnology,Vol. 5, Speci al Issue 9, May

1333 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 An Experimental Investigation on Volcanic Ash – Based Geopolymer Concrete Mr. Sultan Saleh Al -Tais1, Mrs. D. Annapurna2 1. PG scholar, Civil Engineering Department, University College of Engineering, Osmania University, Hyderabad 2. Assistant Professor, Civil Engineering Department, University College of Engineering, Osmania University , Hyderabad ABSTRACT The main purpose of this study is to investigate, through experimental testing the utility and efficiency of using volcanic ash with other source materials as a binder for civil engineering applications. Volcanic ash powder from YEMEN was used as the princ iple source of aluminusilicate. Investigation of the effect of replacing volcanic ash by different source materials namely rice husk ash, fly ash, silica fume, and GGBS with a percentage of 0%, 10%, 20%, 30%, 40%, and 50% was done. The Alkali activators u sed in this study were sodium hydroxide (NaOH) mixed with sodium silicate (Na 2SiO 3) in the ratio of 2.5 and the concentration of sodium hydroxide was 8 M. The geopolymer concrete specimens were casted and tested for 28 days at ambient temperature. From com pressive strength results, replacement of volcanic ash by 50% GGBS gave greater strength (52MPa). Further investigation was done for the mixture of 50% volcanic ash and 50% of GGBS for mechanical properties, sorptivity, water absorption and microstructure by using SEM. Hence, the study concludes that volcanic ash – GGBS geopolymer concrete can be used for the development of a sustainable construction material to replace OPC for the production of economically and echo -friendly geoplymer concrete. KEY WORDS: Geopolymer Concrete, Volcanic Ash, GGBS, RHA, OPC, SEM, Aluminusilicat 1. INTRODUCTION Concrete is the most popular material for construction on earth. It is widely known that the production of Portland cement consumes considerable energy and at the same time contributes a large volume of CO 2 to the atmosphere. However, Portland cement is still the main binder in concrete construction prompting a search for more environmentally friendly materials. The contribution of Ordinary Portland Cement (OPC) producti on worldwide to greenhouse gas emissions is estimated to be approximately 1.35 billion tons annually or approximately 7% of the total greenhouse gas emission to the earth’s atmosphere [ 1]. Geopolymer is such an alternative construction material which can act as a binder replacing cement. Geopolymers are members of the family of inorganic alumino -silicate polymer synthesized from alkaline activation of various aluminosilicate materials or other by -product materials like fly ash, metakaoline, blast furnace s lag etc. [ 2]. It mainly makes use of waste or by -product substances like fly ash, metakaolin, silica fume, and rice husk ash which are cheap and will reduce environmental pollution to a large extent. The amorphous nature of volcanic ash due to its high no n-crystalline silica content accounts for the dissolution of this material in alkaline solution. This led to the idea of using volcanic ash for the development of strengthened structural materials by dissolution, polymerisation and polycon - densation. It is reasonable to assume that volcanic ash, which is an amorphous alumino -silicate, will form geopolymers similar to metakaolin or fly ash. The present work investigated volcanic ash as an 1334 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 alternative raw material for geopolymer structural products. Among the raw materials investigated up to now for geopolymerisation, metakaolin has been identified as the alumino -silicate with the highest release per weight of both silicate and aluminate even after only a few hours in alkaline solution . 2. OBJECTIVES  To assess the optimum mix proportion of geopolymer concrete with resource material replaced in various percentages by volcanic ash.  To assess the strength development, namely compressive strength, flexural strength, modulus of elasticity of the optimized mixt ures.  To investigate some durability properties such as water absorption, and sorptivity of the optimized mixtures.  To study the microstructure aspects of geopolymer concrete for different source materials using Scanning Electron Microscopy (SEM) 3. EXPER IMENTAL PROGRAM An experimental program has been planned. Samples of each and every constituent material are tested in the laborator y for their physical properties . The casting and testing of volcanic ash -based geopolymer concrete specimen were done according to the specification followed for ordinary Portland cement concrete. 3.1 Materials The materials used in present investigational study are volcanic ash ,rice husk ash, fly ash, silica fume, and GGBS as source material, aggregate (coarse and fine), alkaline liquid (sodium hydroxide and sodium silicate), superplasticizer, and water. The properties of materials used are presented in details in tables below. Volcanic Ash Volcanic ash used in this work comes from the volcanic deposit of Hamdan, SA NAA – Republic of Yemen located N 15° 21' 7.3044''andaltitude 1250 meters (Near the capital city of Yemen).This volcanic ash is currently used by a local cement industry (Ammran Cement Industries) aspozzolanic materials in blended cement. Table -1Physical and chemical properties of volcanic ash Ground Granulated Blast Furnace Slag (GGBS) GGBS is the by -product of the manufacture of the iron and steel manufacture. For the present investigation GGBS was obtained from JSW steel plant, Bellary, Karnataka. The physical and chemical properties of GGBS are supplied by manufacturer and found satisfying the values as per IS : 12089 -1987 as given in table -2. Properties Specific gravity Bulk density, g/cc Color Experimental values 2.77 1.67 black Oxide composition (% by mass) CaO SiO 2 Al2O3 Fe2O3 SO 3 MgO Na2O K2O P2O5 TiO2 LOI 10.558 45.009 15.856 11.007 0.034 9.486 3.653 1.131 0.426 2.40 1.0 1335 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 Table - 2 Physical and chemical properties of GGBS Rice husk ash (RHA) Rice husk ash is one of the agricultural wastes and pozzolanic material .for the present study rice husk ash was obtained from Galaxy Agro Industries – Nizamabad the specific gravity and bulk density were 1.91and 0.3 g/cc respectively. The physical and chemical properties of Rice husk ash are given in table (3). Table -3 Physical and chemical properties of Rice Husk Ash Properties Specific gravity Bulk density, g/cc Color Experimental values 1.97 0.3 black Oxide composition (% b y mass) CaO SiO 2 Al2O3 Fe2O3 MgO 0.55 95.58 0.56 0.43 0.4 Silica Fume The commercially available silica fume obtained from local market (shown in figure 3.4) was used in this study. The physical properties and chemical composition of silica fume is given in the following table - 4. Table -4 Physical and chemical properties of Silica fume Properties Specific gravity Bulk density, g/cc Color Experimental values 2.163 1.08 dark Oxide composition (% by mass)* CaO SiO 2 Al2O3 Fe2O3 MgO LOI 2.25 90.58 2.77 1.43 1.14 1.14 *Provided by Manufacturer Fly Ash For the present study fly ash used conforming to ASTM class F and obtained from KTPC Kaktpally thermal project . Fly ash passing 90 micron sieve was used.The physical properties of the fly ash were tested in the laboratories of Osmania University whereas chemical properties were tested IICT, Tarnaka, Hyderabad. The properties tested were found to satisfy the code p rovisions of IS 3612 -2003 as given in table -5. Specific gravity Bulk density, g/cc Fineness,m2/kg Soundeness Le-chatelier expansion (mm) 2.88 1230 375 1.6 Oxide composition (% by mass) CaO SiO 2 Al2O3 Fe2O3 SO 3 MgO Na2O P2O5 LOI 40.3 35 10 1.3 2.39 8 0.15 0.6 1.8 1336 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 Table -5 Physical and chemical properties of fly ash Properties Specific gravity Bulk density, g/cc Fineness, m2/kg Colour Experimental values 2.08 1.16 290 Cream white Oxide composition (% by mass) SiO 2 Fe2O3 Al2O3 CaO MgO) TiO 2 K2O SO 3 K2O LOI 0.52 4 33.9 1.2 0.81 0.27 0.83 0.4 0.83 1.45 Aggregates The aggregates are the main components of the concrete .Local aggregates, comprising coarse aggregates of size 20mmand fine aggregates, in saturated surface dry condition (SSD) condition were used Alkaline Activators A combination of sodium hydroxide and sodium silicate was used in this study. Both of these chemicals are commercially available in the local market. The alkaline liquid should be made prior to one day before mixing due to the huge amount of heat generated while mixing NaOH and Na 2SiO 3. Super plasticizer Superplasticizer is used in the production of geopolymer concrete to improve workability of concrete mixtures. Conplast SP 430 was used. 3.2Mixture Proportions The mix design described in previous section was followed. Five geopolymer mixtures were prepared by varying the ratio of (FA, SF, GGBS, and RHA ) quantity . The quantity of alkaline activator and the aggregate content were kept constant for all mixtures. Superplasticizer and the water were added according to the mix design data outlined in section. The proportioning of ingredients was conducted based onthe w eight method. The mixture proportions of geopolymer concrete are given in Table -7. Table -6 Mix Proportion of Geopolymer concrete 3.3 Preparation of Alkaline Liquid The alkaline activator was a combination of sodium silicate and sodium hydroxide solutions. Sodium hydroxide solution of 8M concentration was prepared by mixing 97 -98% pure pellets with tap water. The sodium silicate was added to enhance the formation of g eopolymer precursors or the polymerization process .The mass of NaOH solids was measured as 8×40 = 320 grams per litre of NaOH solution of 8M concentration. Sodium silicate solution with SiO 2to Na2O ratio by mass of 2. 5 Replacement of(FA,RHA,SF,GGBS) 0% [Kg/m3] 10% [Kg/m3] 20% [Kg/m3] 30% [Kg/m3] 40% [Kg/m3] 50% [Kg/m3] Volcanic ash 384 384 307.2 268.8 230.4 192 (FA,RHA,SF,GGBS) 0 38.4 76.8 115.2 153.6 192 Coarse aggregate 1185.6 1185.6 1185.6 1185.6 1185.6 1185.6 Fine aggregate 683.4 683.4 683.4 683.4 683.4 683.4 Sodium hydroxide 54.85 54.85 54.85 54.85 54.85 54.85 Sodium silicate 103 103 103 103 103 103 1337 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 (SiO 2=30.0%, Na 2O= 11.5% and water=5 8.5%) was used in this study.The sodium silicate solution and the sodium hydroxide solution were mixed together at least one day prior to use to prepare the alkaline liquid as shown in figure -1.On the day of casting of the specimens ,the alkaline liquid was mixed together with the superplasticizer and the extra water (if any) to prepare the liquid component of the mixture. Figure -1 Preparation of alkaline liquid 4. RESULTS AND DISCUSSION 4.1 Compressive Strength Strength developments of the concrete mixtures with different replacement percentages are plotted in Figures -2. Figure -4 show s the efflorescence phenomenon of VA -FA geopolymer concrete. Figure -5 shows bulging of VA -SF geo polymer concrete after casting. Figure -2 Compressive strength of geopolymer concrete 0% 10% 20% 30% 40% 50% 60%Compressive strength MPa Percentage of replacement RHA FA GGBS 1338 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 Figure -3 Unit weights densities of different sources materials Figure - 4 Efflorescence of VA -FA GPC Figure -5 Bulging of VA -SFGPC  Discussion on compressive strength  Figures -2 shows a graph plotted between different percentages replacements (x - axis) versus compressive strength of different source materials (y -axis). F igure -3 show bar chart for different percentages replacement (x -axis) versus unit weight densities of different source materials (y - axis).  It is observed that 10% replacement of volcanic ash by RHA decreases 52.63% and 15% of compressive strength and unit weight density respectively whereas 50% replacement decreases 81.57% and 21 % of compressive strength and unit weight respectively  The variations of compressive strength and unit weight densities with different percentages of fly ash replacement were given in table 4.1. From figure 4.1 it was observed that the maximum 1 2 3 4 5 Unit Weight Denisity % of replacement RHA FA GGBS 1339 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 compressive strength was obtained for mix GVF3 with 70% volcanic ash and 30 % fly ash. It is observed that 10%, 20%, and 30% replacement of volcanic ash by fly ash slightly increase the compressive strength by 5.26%, 15.78%, and 26.31% respectively. Furthe r replacement of 40 % of volcanic ash by fly ash, the strength dropped marginally to 16.3 Mpa.  Figure -4 shows formation of white salts deposits on the surfaces of specimens which is known as efflorescence phenomenon of VA -FA geopolymer concrete. Because ge opolymer concrete contains high soluble alkali content, efflorescence can be significant matter when the geopolymer concrete exposed to humid air or in contact with water. It is reported that efflorescence is mainly sodium carbonate nephthahydrate (Na 2.CO 3.7H 2O). Reasons of efflorescence might be due to much amount of soluble salts in the concrete, bad mixing or too much free water. Efflorescence has negative effects on the compressive strength of geopolymer due to the mechanical influence on binder.  Figure -5 shows bulging of specimens that have volcanic ash and silica fume content. This phenomenon may be related to the silica fume and volcanic ash nature minerals and because of swelling properties of their minerals during absorbing moisture. When the sample s are subjected to wetting condition, they start swelling. Swelling is due to the adsorption of water by the natural minerals in VA -SF gel. The water molecules act against the cohesive force and tend to force the gel particles further apart as a result of which swelling takes place. In addition, the ingress of water decreases the surface tension of the gel.  Figure -2 show s that compressive strength of volcanic ash -GGBS geopolymer concrete get increased as the percentage of the GGBS increased in rate of 10 %,20%30%40%50% of volcanic ash volume.  It is observed that 10% replacement of volcanic ash by GGBS increases 10.52% and 0.17% of compressive strength and unit weight density respectively whereas 50% replacement increases 168% and 6 % of compressive streng th and unit weight respectively.  Because of good compressive strength and reasonable setting time obtained, volcanic ash and GGBS geopolymer concrete was selected for further investigations.  50% replacement of volcanic ash by GGBs was selected to be investigated for mechanical properties, sorptivity test, water absorption test and microstructure of concrete by using SEM. 4.2 Flexural Strength ( As per IS: 9399 -1979) The flexural strength of volcanic ash was calculated by testing 3 specimens of two point loaded prism with standard size of 100 mm x100 mmx500 at the age of 28 days from day of casting. The results of flexural strength are persented in Table -7. Table -7 Flexural strength of VA -GGBS Geopolymer concrete with 50% GGBS S.No Sample Flexural strength (MPa) 1. GVA501 5.5 2. GVA502 5 3. GVA503 5 Average 5.17 MPa From the table 4.2 it shows that the average flexural strength of VA -GGBS geopolymer concrete with 50 % of GGBs at 28 days is 5.17 MPa and the limiting value of flexural strength is 0.7 √𝑓𝑐 1340 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 which is 4.49. Hence it is observed that geopolymer concrete achieved 10% more than the limiting value from the equation mentioned above. 4.3 Modulus of elasticity (As per IS: 516 -1959) The stress - strain values of the specimens are shown in table -8 below . Table -8 Modulus of elasticity VA-GGBS GPC with 50%GGBS The average Young’s Modulus of elasticity of 28 days for the GVA -GGBS geopolymer concrete is 2.17 MPa. A normal weight concrete usually has a Young’s modulus in range of 2.1 x 104 – 4.2 x 104 MPa. GVA -GGBS geopolymer performed a lower Modulus of elasticity than ordinary Portland cement concrete although its compressive strength is high. It can be concluded that modulus of elasticity relies on the microstructure of geopolmer concrete not on aggregate or volcanic ash type. 4.4 Water absorption test (As per A S 1012.21) Table -9 shows that the average water absorption (WA) = 5.75 %.The water absorption of 50% replacement of volcanic ash by GGBS was lower than 6% which can be classified as excellent as per AS 1012.21. Water absorption is dependent on the mixture composition and cur ing. Table -9 Water absorption of VA -GGBS GPC with 50% GGBS Sample Mass of air -dried sample (kg) Mass of surface - dried sample (kg) Mass Oven -dried sample (kg) Water absorption GVA1 2.455 2.492 2.325 6.7 GVA2 2.482 2.465 2.335 5.27 GV3 2.452 2.490 2.330 6.4 GV4 2.475 2.496 2.355 5.64 GV5 2.422 2.460 2.332 5.2 GV6 2.400 2.485 2.356 5.29 Average 5.75% 4.4 Sorpitivity test (As per ASTM C1585) . In this research, only the initial sorptivityduring the first six hours was measured. Table -10 Results of Sorptivitywith 50% GGBS Sample Code Sorptivity (S) mm/min0.5 GVI 0.0744 GV2 0.1057 GV3 0.0962 GV4 0.076 GV 0.0654 `GV6 0.0822 Average 0.0733 S.No Specimen Modulus of elasticity (N/mm2) 1. Sample A 2 x 104 2. Sample B 2.1 x 104 3. Sample C 2.4 x 104 Average 2.17 x 104Mpa 1341 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 Table -10 shows the results of sorptivity for six specimens. The initial slopes from the sorptivity results were measured by calculating the average rate of uptake of water during the first 6 hours of the measurement. The sorptivity of the s amples of VA -GGBS geopolymer concrete containing 50% GGBS varied in the range of 0.0761 -0.0845 mm/min0.5. Based on the experimental values, the geopolymer concrete in this research was classified as having a good quality. It was recommended to limit sorpt ivity index less than 0.2000 mm/ min0.5as per ASTM C1585 to maintain its water tightness. The same finding was also confirmed by previous researchers. 4.5 SEM (scanning electron microscopy) SEM analysis was performed to study the pore morphology and to view the reacted and unreacted regions of the specimen. Figures 6 to 8 present the SEM micrograph for geoplymer concrete specimens GV50, SGV50, and WGV50, along with their EDAX traces which cur ed at ambient temperature for 28 days. Figure -6 (a) SEM image of GVA50 ,(b) EDX spectrum of GVA50 Figure -7 (a) SEM image of S GVA50 specimen , (b) EDX spectrum of SGVA50 1342 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 Figure -8 (a) SEM image of WGVA50 specimen , (b) EDX spectrum of WGVA50 In all the micrographs of specimens whether volcanic ash with GGBS , it depicts a microstructure having some unreacted or partially reacted particles embedded in the Geopolymer gel.  Unlike fly ash whose particles have spheres shapes, volcanic ash particles do not have common shapes which cannot lead to specific morpholog y. Figure -6a shows the micrograph of specimen with 50 % VA and 50% GGBS. The micrograph indicates that there is a strong bond and particles appeared as flake -like structures. A grain structure is still appeared, but the intervening materials is of more hom ogenous nature with small pores and this the reason of getting high compressive strength. EDX spectra of GVA50 shows major elements such as oxygen (O), aluminum (Al) , silicone (Si) , calcium (Ca), and Sodium (Na). The weight percentages of the elements ar e shown in figure -6b.  Figures -7 and 8 show the micrograph of specimens after sorptivity and water absorption tests respectively. They show that the sample consist of both agglomerated and angular properties with cracks detected along the interface. The m ajor components found on both samples from EDX spectra are oxygen (O), aluminum (Al), silicone (Si), calcium (Ca), and Sodium (Na) other elements such as Mg, K, and Pb were also found in much lower quantity. 5. CONCLUSION The following conclusions can be d rawn  With replacement of 10% to 50% of volcanic ash by rice husk ash, the compressive strength gradually decreases from 19 MPa to 3.5MPa. Hence, rice husk ash cannot be used as an alternative material replaced with volcanic ash due to low strength obtained, slow setting and reduced unit weight.  For volcanic ash and fly ash geopolymer concrete, strength increased gradually until 30% replacement (24 MPa), while further replacement of fly ash beyond 30% showed a significant reduction in the strength (1 3 MPa) beside formation of efflorescence on the surfaces of specimens .  It is observed that 10% replacement of volcanic ash by GGBS increases 10.52% and 0.17% of compressive strength and unit weight density respectively whereas 50% replacement increases 168% and 6 % of compressive strength and unit weight respectively. 1343 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018  Bulging phenomenon of volcanic ash and silica fume specimens may be related to the silica fume and volcanic ash nature minerals and because of swelling properties of their minerals during absorbing moisture.  The density of the GGBS -VA was in the range of 2440 -2550. The values are still in the range of normal concrete, which are 2200 -2600 kg/m3.  Mixture with 50% of volcanic ash and 50% of GGBS considered as the optimum mixture for high compres sive strength (52 MPa) in ambient curing condition with a setting time comparable to that of OPC.  The flexural strength of VA -GGBS geopolymer concrete at 28 days was 5.17 MPa and the limiting value of flexural strength is 0.7 √𝑓𝑐/ which is 4.49. Hence it is observed that achieved geopolymer concrete flexural strength is more by 10%. This might be due to good bonding between aggregate and geopolymer gel.  The Young’s modulus of elasticity of 50%VA -50%GGBS geopolymer concrete at 28 days was 2.1x104 MPa and th e young’s modulus range for a normal weight concrete is 2.1 -4.2 x104 MPa.  Sorpitivity of 50% replacement of volcanic ash by GGBS was 0.0733 mm/min0.5 which is very low comparing to OPC.  The water absorption of 50% replacement of volcanic ash by GGBS was lower than 6% which can be classified as excellent. Water absorption is dependent on the mixture composition and curing.  The microstructure of the optimum strength of geopolymer appeared to be homogenous and contained the minim um proportions of unreacted volcanic ash.  The good densification behavior, good mechanical properties and lower sorptivity and water absorption of volcanic ash based geopolymer concrete seem to be suitable for building application.  The geopolymer concrete produced with different combinations of volcanic ash and GGBS are able to produce structural concrete by ambient curing only. REFERENCES 1) Malhotra, V. M. (1999) . Making concrete" greener" with fly ash. Concrete international , 21(5), 61 -66.‏ 2) Deb, P., Nath, P., &Sarker, P. (2013). Properties of fly ash and slag blended 3) Glukhovsky, V. D. (1965). Soil silicates. Their properties, technology and manufacturing and fields of application, Doct. Tech. Sc. Degree Thesis. Civil Engineering Institute, Kiev .‏ 4) Olivia, M., &Nikraz, H. (2012). Properties of fly ash geopolymer concrete designed by Taguchi method. Materials & Design (1980 -2015), 36, 191 -198. 5) Schneider, M., Romer, M., Tschudin, M., &Bolio, H. (2011). Sustainable cement production —present and future. Cement and Concrete Research , 41(7), 642 -650.‏ 6) Leonelli, C., Kamseu, E., Boccaccini, D. N., Melo, U. C., Rizzuti, A., Billong, N., &Miselli, P. (2007) . Volcanic ash as alternative raw materials for traditional vitrified ceramic products. Advances in applie d ceramics, 106(3), 135-141.‏ 7) Djobo, J. N. Y., Elimbi, A., Tchakouté, H. K., & Kumar, S. (2017). Volcanic ash -based geopolymer cements/concretes: the current state of the art and perspectives. Environmental Science and Pollution Research, 24(5), 4433 -4446. 8) Robayo, R. A., de Gutiérrez, R. M., & Gordillo, M. (2016). Natural pozzolan -and granulated blast furnace slag -based binary geopolymers. Materiales de Construcción , 66(321), 077. ‏ 9) Moon, J., Bae, S., Celik, K., Yoon, S., Kim, K. H., Kim, K. S., & Monteiro, P . J. (2014) . Characterization of natural pozzolan -based geopolymeric binders. Cement and Concrete Composites , 53, 97-104.‏ 1344 Mr. Sultan Saleh Al -Tais, Mrs. D. Annapurna International Journal of Engineering Technology Science and Research IJETSR www.ijetsr.com ISSN 2394 – 3386 Volume 5, Issue 1 January 2018 10) Risdanareni, P., Karjanto, A., &Khakim, F. E. B. R. I. A. N. O. (2016, January). Physical Properties of Volcanic Ash Based Geopolymer Concrete. In Materials Science Forum (Vol. 841). ‏ 11) Kani, E. N., Allahverdi, A., &Provis, J. L. (2012). Efflorescence control in geopolymer binders based on natural pozzolan. Cement and Concrete Composites , 34(1), 25 -33.‏ 12) Lemougna, P. N., Melo, U. C., Delpl ancke, M. P., &Rahier, H. (2014). Influence of the chemical and mineralogical composition on the reactivity of volcanic ashes during alkali activation. Ceramics international , 40(1), 811 -820.‏ 13) IS: 383 -1970 Specification for Coarse and Fine Aggregates from natural sources for concrete (Second revision). 14) IS 456 -2000 , Specifications for plain and reinforced concrete. 15) IS: 516 -1959 , method of test for strength of concrete Bureau of I ndian standards, New Delhi, India 16) IS: 9399 – 1979 , “Specification for apparatus for flexural testing of concrete”.BIS New Delhi.

Fracture properties of slag/fly ash-based geopolymer concrete cured in/C15The fracture behaviors of SFGC are influenced by the material parameters. /C15The Baz ˇant and Becq-Giraduon model predicts well the fracture energy of SFGC. /C15The characteristic length of SFGC is higher than the prediction for PCC. /C15The SFGC beam with Ms= 2.0 exhibits more ductility behavior. article info Article history: Received 3 June 2018Received in revised form 20 September2018 Accepted 20 September 2018 Keywords: Slag/FA-based geopolymer concrete (SFGC)Ambient temperature curingFracture propertyMechanical propertyThree-point bending (TPB) testabstract Slag/fly ash (FA)-based geopolymer cured in ambient temperature as a green alternative to Portland cement is attracting increasing attentions. The fracture properties of slag/FA-based geopolymer concrete (SFGC) was studied by conducting three-point bending (TPB) tests on precut beams. The effects of material parameters including the alkali concentration, the modulus of alkali activator, the slag/FA massratio and the liquid/binder ratio on the fracture properties of SFGC were assessed. The results exhibit that the fracture behaviors of SFGC are influenced significantly by the material parameters. The fracture energy and the ultimate load of TPB tests of SFGC beams increase with the increase of the alkali concen-tration, the modulus of alkali activator as well as the slag/FA ratio while decrease with the increase of liquid/binder ratio. The Baz ˇant and Becq-Giraduon model predicts well whereas the CEB-FIP model underestimates the fracture energy of SFGC beams. Besides, the characteristic length of SFGC decreaseswith the increase of compressive strength regardless of the mix proportion, and is higher than the pre-diction for Portland cement concrete (PCC) given the similar compressive strength, suggesting that SFGC might be more ductile. In addition, the relationships between compressive strength, splitting tensile strength, elastic modulus and material parameters of SFGC specimens are also discussed. /C2112018 Elsevier Ltd. All rights reserved. 1. Introduction It has been known that pozzolanic materials containing silica and/or alumina can be chemically activated by using alkali solution to form an environmentally friendly cementitious material desig- nated as geopolymer cement. This binder with the advantages of Portland cement (PC) whereas lower cost and CO 2emissions has been regarded as a potential alternative to PC [46,14] . In addition, it could recycle industrial by-products, i.e., ground granulated blast furnace slag (GGBFS) and fly ash (FA) or use natural minerals such as metakaolin [50,40] as raw materials. Typical examples of this type of binder include Class F (low-calcium content) FA-basedgeopolymer (FFG) and slag-based geopolymer (SG). Although the synthesis conditions for FFG and SG are quite similar, their hydration products are different depending on whether calcium is necessary. Analogously, geopolymer cement is referred to as a quasi-brittle material as PC [29,43,31,15] . Reinforcing the geopoly- mer matrix with fibers could efficiently increase the ductility or even make it exhibit strain-hardening behavior [34–36,13] , which is similar to that of the traditional cementitious composites [16,17,52] . Fracture characteristics of concrete are believe to be influenced by the material properties such as strength, mixture constituents,and types of aggregate used. As a relatively new material, extensive research on various properties of geopolymer cement is required to ensure its suitability for structural applications, among which fracture property is of critical importance. Previous researches 0950-0618/ /C2112018 Elsevier Ltd. All rights reserved.⇑Corresponding author. E-mail address: cshi@hnu.edu.cn (C.-J. Shi).Construction and Building Materials 190 (2018) 787–795 Contents lists available at ScienceDirect Construction and Building Materials journal homepage: www.elsevier.com/lo cate/conbuildmat [37,7,20,25] have stated that due to the high activation energy of Class F FA, high temperature curing is necessarily needed in producing Class F FA-based geopolymer concrete (FFGC) with desirable strength. The fracture properties of FFGC cured in elevate temperature have been reported. Pan et al. compared the fracture behaviors of FFGC cured in elevated temperature with Portland cement concrete (PCC). They found that given a similar compressive strength, the fracture energy of PCC was around 24% larger and its characteristic length was almost three times higher than those of FGC, suggesting that PCC exhibited more ductility than FFGC. Sarker et al. also reported the fracture properties of high temperature cured FFGC. They found that the descending branches of the load-displacement curves of FFGC specimens were steeper and the fracture planes were smoother than those of the PCC specimen with similar compressive strength, indicating thatPCC exhibited more ductility than its FFGC counterpart which was consistent with the results from Pan et al. . Xie et al. used digital image correction (DIC) method to monitor the crack propagation in FFGC and stated that the normalized crack extension length could be used to describe the unstable crack propagation in FFGC. Reports on the fracture behavior of ambient-cured geopolymer concrete are scarce in literature. Nath and Sarker examined the fracture properties of slag/Class F FA-based geopolymer con- crete (SFGC) cured in ambient temperature and found that SFGC achieved higher fracture energy compared with its PCC counter- part. Besides, the deflection before ultimate failure of SFGC speci- men was slightly larger which might suggest that SFGC cured in ambient temperature was more ductility than its PCC counterpart. Recently, the fracture property of slag-based geopolymer concrete (SGC) cured in ambient temperature has been studied by the authors . They found that the fracture energy of SGC was larger than its PCC counterpart because of its more homogeneous and denser interfacial transition zones. The smaller characteristic length values of SGC than those of PCC indicated that SGC was more brittle than its PCC counterpart. Above all, it can be seen that the fracture properties of FFGC cured in elevated temperature has been reported. Since that curing method has a significant influence on the mechanical properties of the hardened geopolymer concrete, it is necessary to study the fracture behavior of geopolymer concrete cured in ambient condition. Although Nath and Sarker stated that the fracture properties of SFGC cured in ambient temperature were affected by the mix proportions, only a small replacement percentage of FA with slag (i.e., up to 15%) and two liquid/binder ratios were used. Hence, this present study aims to further investigate the influence of material parameters on the fracture properties of SFGC cured in ambient temperature systematically, including a largerslag replacement percentage (50–100%), wider ranges of alkali con- centration (3–5%), modulus of alkali activator (1.0–2.0) and liquid/ binder ratio (0.40–0.50) compared with former research . Through three-point bending (TPB) tests, the fracture properties of SFGC including the load vs. displacement curve, the peak load of TPB test, the fracture energy and the characteristic length were obtained. In addition, the basic mechanical properties of hardened SFGC such as the compressive strength, the splitting tensile strength and the elastic modulus were also studied.2. Experiment program 2.1. Raw materials The geopolymer binder was prepared using the combination of the GGBFS pro- duced by a cement company, and the Class F FA (ASTM C 618) collected from apower plant in Hunan, China. Their chemical compositions are presented in Table 1 . The contents of CaO by mass were 38.8% and 1.32% in GGBFS and FA, respectively.Fig. 1 exhibits the particle size distributions of the GGBFS and the FA. The particle size of the former was mainly in the range from 2 lmt o5 0 lm while the latter had a wider range from 1 lm to 300 lm. The X-ray diffraction patterns of raw GGBFS and FA are depicted in Fig. 2 . It shows that the slag contained large amounts of glass, as indicated by the broad diffuse bands between 24 /C176and 37 /C176(2h). In addi- tion, the amorphous character of FA was illustrated by the broad diffuse bandsbetween 15 /C176and 30 /C176(2h), and signals corresponding to the crystalline phases, such as quartz and mullite can also be found. The glass contents of FA and GGBFS deter-mined by Rietveld quantitative XRD method were 74.3% and 98.6%, respectively. Table 1 Chemical composition of raw materials (wt%). CaO Al 2O3 SiO 2 SO3 P2O5 MgO Na 2OK 2O TiO 2 GGBFS 38.8 14.8 33.8 2.49 0.05 7.09 0.25 0.44 1.23FA 1.32 34.1 52.6 0.33 0.17 0.50 0.34 1.37 1.720.1 1 10 100 1000020406080100 FA GGBFSCumulative volume (%) Size of particle ( m) Fig. 1. Particle size distribution of GGBFS and FA. 10 20 30 40 50 60 70MMM FA GGBFSQ Fig. 2. X-ray diffraction patterns of GGBFS and FA. Q = quartz; M = mullite.788 Y. Ding et al. / Construction and Building Materials 190 (2018) 787–795 The alkali activator solution was a mixture of sodium hydroxide and sodium sil- icate solution. Sodium silicate solution used was a commercially available productwith chemical composition of 8.9% Na 2O, 32.1% SiO 2and 59% H 2O by mass, respec- tively. The addition of 99% purity sodium hydroxide (NaOH) flakes helped to adjustthe modulus of the alkali activator to the targeted values. The fine aggregate was medium sand with fineness modulus of 2.47. The speci- fic density and the water absorption of the fine aggregate were 2340 kg/m 3and 2.75%, respectively. Besides, gravel from local river with a maximum size of10 mm was selected as coarse aggregate. The bulk specific density and the waterabsorption of coarse aggregate were 2530 kg/m 3and 1.83%, respectively. 2.2. Mix proportions The mix proportions of SFGC are summarized in Table 2 . The binder amounts and the sand ratios of SFGC were constant at 400 kg/m3and 0.40 for all mixtures. The alkali concentration (the percentage of Na 2O by mass of binder, n, from 3% to 5%), the modulus of the alkali activator (the mole ratio of SiO 2to Na 2O,Ms, from 1.0 to 2.0), the replacement percentage of GGBFS with FA (from 0% to 50%) andthe liquid/binder ratio (from 0.40 to 0.50) were considered as material parameters. 2.3. Specimen preparation The alkali activator solution was generated by blending sodium hydroxide, sodium silicate solution and water together 24 h before concrete mixing to assurethe solution to cool down to room temperature. The weighed FA, GGBFS, fine andcoarse aggregates were firstly mixed in a mixer for 2 min. Then, the alkali activatorsolution was slowly pouring into the mixer and mixed with the dry fractions foranother 3 min until a uniform mixture was formed. The final mixture was cast inthe prepared molds and solidated on a vibrating table. All the specimens were cov-ered with plastic sheets for 24 h and then demolded. After that, the specimens were stored in a curing room with constant temperature of 21 ± 1 /C176C and a related humidity of 90 ± 5% until testing ages. 2.4. Testing procedure 2.4.1. Compressive strength, splitting tensile strength and elastic modulus tests A 2000 kN capacity universal testing machine was adopted to conduct the com- pressive strength, the splitting tensile strength and the elastic modulus tests of hardened SFGC specimens. The dimensions of the cylinder specimens were 100/C2200 mm. The loading rates adopted in the compressive strength tests and the splitting tensile tests were 0.30 MPa/s and 0.02 MPa/s , respectively. In addition, the elastic modulus tests were conducted referring to the ASTM C496 . 2.4.2. Three-point bending (TPB) test Three-point bending (TPB) test was conducted following the RILEM TC50-FMC recommendation. The beam dimensions were 100 /C2100/C2515 mm, and the span/depth ratio was 4.0. All the specimens were precut in the middle of the beams with a notch of 40 mm deep and 3 mm wide. The beam was simply supported with the notched face down. The geometry of the beam is shown in Fig. 3 . Four identical specimens were prepared for each mixture listed in Table 2 . The crack mouth open- ing displacement (CMOD) was measured using clip gauge. Two high-precise dis-placement transducers (HPDTs) were installed to record the mid-spandisplacement ( d) of the beam, meanwhile the influence of support settlement on the mid-span deflection was removed by measuring the deflections of the two sup-ports. A 100 kN capacity hydraulic jack operated at 0.02 mm/min was used for the sake of obtaining the complete load-displacement (including P-dand P-CMOD) curves,3. Test results and discussion 3.1. Compressive strength The effects of material parameters on compressive strength developments of SFGC are shown in Fig. 4 . The 28-day compressive strengths of SFGC are also summarized in Table 3 . Obviously, the compressive strength of SFGC increased with the increase of alkali concentration, the modulus of alkali activator and the slag/FA ratio while decreased with the increase of the liquid/binder ratio, which are consistent with previous researches [29,31] . It can be seen from Fig. 4 a that the 28-day compressive strength of SFGC increased from 31.7 MPa to 58.2 MPa with a 83.6% increase, when the alkali concentration increased from 3% to 5%. Regardless of curing age, higher alkali concentration always led to higher compressive strength. This might be due to that higher alkali concentration would accelerate and increase the dissolution of silicon and alu- minum ions in the source materials and the formation of the geopolymer paste, and thus benefit the early age compressive strength and strength development. In addition, the compressive strength of SFGC increased from 44.4 MPa to 65.0 MPa (i.e., a 46.4% increase) as the modulus of alkali activator improved from 1.0 to 2.0 at 28d (see in Fig. 4 b). Increasing the modulus of alkali activator could promote the hydrolysis of the siliceous and alu- minum species of the raw materials and provide additional sil- icate anions to react with Ca2+dissolved from the source materials to form C-S-H. All these could contribute to the enhancement of compressive strength. Moreover, as shown in Fig. 4 c, the improve- ment in compressive strength with the increase of the amount of GGBFS could be attributed to the high activity and calcium content in GGBFS [29,43,31] . As expected, the influence of liquid/binder ratio on compressive strength of SFGC was similar with that of the water/cement ratio on PCC (see in Fig. 4 d). Table 2 Mix proportions of SFGC. N(%) Ms Slag kg/m3FA kg/m3Fine aggregate kg/m3Coarse aggregatekg/m 3Water kg/m3Alkali activator l/b Sodium silicate solution (kg/m3)Sodium hydroxide (kg/m3) SFGC-1 3 1.5 200 200 716 1074 145 56 9 0.45 SFGC-2 4 1.5 200 200 712 1068 133 74 12 0.45SFGC-3 5 1.5 200 200 708 1062 122 93 15 0.45SFGC-4 5 1.0 200 200 712 1068 139 62 19 0.45SFGC-5 5 2.0 200 200 704 1056 104 124 12 0.45 SFGC-6 5 1.5 200 200 716 1074 102 93 15 0.40 SFGC-7 5 1.5 200 200 700 1050 142 93 15 0.50SFGC-8 4 1.5 300 100 712 1068 133 74 12 0.45SFGC-9 4 1.5 400 0 712 1068 133 74 12 0.45Fig. 3. Configuration of TPB test beams.Y. Ding et al. / Construction and Building Materials 190 (2018) 787–795 789 Additionally, as shown in Fig. 4 a, the average compressive strengths of SFGC with alkali concentration of 5% were 34.7 MPa, 46.0 MPa, 58.2 MPa and 68.4 MPa, respectively, corresponding to the curing ages of 3, 7, 28 and 90 days. A further 17.5% increase of compressive strength was observed with the curing age increased from 28 days to 90 days. Besides, the 3-day compressive strength of SFGC could achieve almost 35 MPa in many cases under room temperature curing. The significant strength development and the high early age strength are attributed to the formation of C-S-H gels and the condensation of microstructure with the incor- poration of GGBFS [28,23] . 3.2. Splitting tensile strength As summarized in Table 3 , the splitting tensile strength of SFGC was only a portion of the corresponding compressive strength.Similarly, the splitting tensile strength of SFGC increased with the alkali concentration, the modulus of alkali activator and the slag/FA ratio while decreased with the increase of liquid/binder ratio . The splitting tensile strength of SFGC increased from 2.88 MPa to 3.81 MPa, with a 32.3% increase, when the alkali con- centration increased from 3% to 5%. In general, codes of practice provide empirical formulae to esti- mate the splitting tensile strength of PCC by its compressive strength. The splitting tensile strength of PCC can be predicted by ACI Building Code 318 using Eq. (1). fst¼0:56ffiffiffiffi fcq ð1Þ where fstis the splitting tensile strength (MPa) and fcis the charac- teristic compressive strength (MPa).01020304050607080 n=3% n=4% n=5%Compressive strength (MPa) Age (days)37 28 90Ms=1.5 l/b=0.45 slag/FA=50/50Ms=1.0 Ms=1.5 Ms=2.0Compressive strength (MPa) Age (days)37 28 90n=5% l/b=0.45 slag/FA=50/50 (a) Alkali concentration (b) Modulus slag/FA=50/50 slag/FA=75/25 slag/FA=100/0Compressive strength (MPa) Age (days)37 28 90n=4% Ms=1.5 l/b=0.45l/b=0.50 l/b=0.45 l/b=0.40Compressive strength (MPa) Age (days)37 28 90n=5% Ms=1.5 slag/FA=50/50 (c) Slag/FA ratio (d) Liquid/binder ratio Fig. 4. Influences of material parameters on compressive strength development of SFGC. Table 3 Compressive strengths, splitting tensile strengths and elastic moduli of SFGC at 28 days. n(%) Ms Slag kg/m3FA kg/m3l/b f cMPa fstMPa EGPa SFGC-1 3 1.5 200 200 0.45 31.7 ± 2.5 2.88 ± 0.4 22.7 ± 1.1 SFGC-2 4 1.5 200 200 0.45 49.3 ± 0.8 3.32 ± 0.5 24.4 ± 0.8SFGC-3 5 1.5 200 200 0.45 58.2 ± 1.8 3.81 ± 0.2 24.6 ± 0.7 SFGC-4 5 1.0 200 200 0.45 44.4 ± 1.0 3.58 ± 0.3 27.4 ± 0.6 SFGC-5 5 2.0 200 200 0.45 65.0 ± 1.3 4.01 ± 0.6 24.2 ± 0.6SFGC-6 5 1.5 200 200 0.40 65.7 ± 1.0 4.32 ± 0.6 25.7 ± 0.5SFGC-7 5 1.5 200 200 0.50 42.3 ± 0.3 3.37 ± 0.3 22.7 ± 1.1SFGC-8 4 1.5 300 100 0.45 52.1 ± 1.8 3.56 ± 0.4 23.8 ± 0.5SFGC-9 4 1.5 400 0 0.45 63.1 ± 2.8 3.97 ± 0.2 22.3 ± 0.4790 Y. Ding et al. / Construction and Building Materials 190 (2018) 787–795 Besides, Eurocode 2 suggests a relationship between the characteristic compressive strength fckand the mean tensile strength ftmof PCC as follows: ftm¼0:3fckðÞ2=3;fck/C2050 MPa ftm¼2:12/C2ln 1þfcm=10ðÞðÞ ;fck>50 MPa( ð2Þ fcm¼fckþ8 ð3Þ ftm¼0:9fst ð4Þ where fcmis the mean compressive strength (MPa). Recently, Ding et al. proposed an empirical formula to esti- mate the splitting tensile strength of geopolymer concrete by itscompressive strength according to the published data by regres- sion analysis. The equation is in the similar form to that of the ACI Building Code 318 and written as follows: fst¼0:527ffiffiffiffi fcq ð5Þ The splitting tensile strengths of SFGC specimens with different material parameters obtained directly from tests, and those pre- dicted by ACI 318 Model (Eq. (1)), EC2 Model (Eqs. (2)–(4) ) and Ding et al. (Eq. (5)) are compared in Fig. 5 . It can be seen that the splitting tensile strengths of SFGC were underestimated by EC2 Model and overestimated by ACI 318 Model. Nevertheless, the equation proposed by Ding et al. predicted reasonably well of the splitting tensile strength of SFGC. Given the compressive strength of 65.0 MPa, the splitting tensile strength of SFGC was 4.0 MPa. The predictions provided by EC2 Model and ACI 318 Model were 12.5% lower (i.e., 3.5 MPa) and 12.5% higher MPa (i.e., 4.5), respectively, than that of the testing value. However, the splitting tensile strength estimated by Ding et al. (i.e., 4.2 MPa) was only 5.0% higher compared with the experimental result. 3.3. Elastic modulus The results in Table 3 indicate that the material parameters have ignorable influence on the elastic modulus of SFGC. In other words, the variation of elastic modulus of SFGC with compressive strength is not significant as shown in Fig. 6 . Comparing the modulus of elasticity of SFGC-1, SFGC-2 and SFGC-3, it is clear that the modulus of elasticity increased marginally from 22.7 GPa to 24.6 GPa with a 8.4% increase when the alkali concentration increased from 3% to 5%, corresponding to the compressivestrength increased from 31.7 MPa to 58.2 MPa (i.e., a 83.6% increase). This might be due to that higher compressive strength geopolymer is always accompanied with more serious drying shrinkage and shrinkage cracks [6,51] which would reduce the elastic modulus. The experimental results of the elastic modulus of SFGC and the predictions by the empirical formulae for PCC suggested by CEB- FIP Model (Eq. (6)) are compared in Fig. 6 . As shown in Fig. 6 , the elastic modulus of SFGC was about 20–30% lower than the predictions for PCC using CEB-FIP Model. When the compressive strength of SFGC was 65.7 MPa, the corresponding modulus of elasticity was 25.7 GPa which was 24.6% lower than the CEB-FIP Model prediction (i.e., 34.1 GPa). Published researches [41,39] have reported that the main hydration products of SFGC are the co-existence of C-A-S-H gel and N-A-S-H gel. Although the intrinsic Young’s modulus of C-A-S-H gel formed in slag-based geopolymer (i.e., 12–43 GPa) is in the similar range with that of the C-S-H gel formed in cement (i.e., 16–44 GPa), the intrinsic Young’s modulus of N-A-S-H gel formed in low-calcium FA-based geopolymer is around 17–18 GPa which is significantly smaller than that of the C-S-H gel formed in cement. In addition, previous researches have stated that the slag/FA-based geopolymer paste exhibited higher shrinkage than cement [11,30] which could generate more serious initial micro-cracks in the former matrix and result in lower elastic modulus. Therefore, it is reasonable that SFGC specimens exhibited lower elastic modulus than their PCC counterparts with similar compressive strengths. E¼0:85/C22:15/C2104fcm=10ðÞ1=3ð6Þ 3.4. Load-midspan displacement curve The complete P-dcurve can reflect the fracture behavior of con- crete. Fig. 7 shows the average P-dcurves of SFGC beams. It can be observed from the curves that the load increased linearly until the crack initiated, then followed by a non-linear increase to the peak load. After the peak load was reached, the crack continuously prop- agated which resulted in the post-peak downward. The post-peak branch of the curve reflected the ductility of the material. The mid- span displacements at failure for most of the beams were around 0.5 mm except those with higher modulus of alkali activator with Ms= 2.0, for which the average midspan displacement was about 0.7 mm (see in Fig. 7 b), indicating a less brittle behavior. With the increase of modulus of alkali activator, the ratio of Ca/Si decreased and the calcium-silicate layers became less defective15 20 25 30 35 40 45 50 55 60 65 701234567 Splitting tensile strength EC2 ACI 318 Ding et al. (2017)Splitting tensile strength (MPa) Compressive strength (MPa) Fig. 5. Relationship between splitting tensile strength and compressive strength of SFGC.20 30 40 50 60 7015202530354045 Experimental results CEB-FIPElastic modulus (GPa) Compressive strength (MPa) Fig. 6. Relationship between elastic modulus and compressive strength of SFGC.Y. Ding et al. / Construction and Building Materials 190 (2018) 787–795 791 which would lead to higher strength . Nevertheless, the improvement in modulus of alkali activator meant more silica was provided and more silica-rich gel was formed during hydra- tion. A silica-rich gel had a high water content generating greatshrinkage when water expelled [27,47] . As a result, the greater shrinkage would lead to more initial micro-cracks and make the matrix less brittle. 3.5. Fracture energy and peak load The fracture energy is a critical parameter describing the energy consumption capacity during crack propagation and could be cal- culated by Eq. (7) according to RILEM TC50-FMC recommendation. GF¼W0þmgd0 ðÞ =Alig ð7Þ where mis the beam mass between two supports (kg); W0is the external work (N ∙m); d0is the final mid-span deformation (m); g is the gravitational acceleration, 9.81 m/s2and Aligis the ligament area (m2). The peak loads Puof the TPB tests and the fracture energy G F calculated using Eq. (7)of SFGC beams with different material parameters are shown in Fig. 8 . As expected, the peak loads Pu and the fracture energy G Fof SFGC specimens increased with the alkali concentration, the modulus and the slag/FA mass ratio while reduced with the increase of liquid/binder ratio, that is to say, increased with compressive strength. As shown in Fig. 8 a, theimprovement in the average peak load Puof SFGC beams with the increase of alkali concentration from 3% to 5% was 31.5%, and the fracture energy increased from 118.3 N/m to 137.6 N/m with a 16.3% increase. Although the peak loads Puof SFGC beams increased insignificantly from 3.20 kN to 3.34 kN with only a 4.4% increase, the average fracture energy G Fstill increased from 128.8 N/m to 157.7 N/m (i.e., a 22.4% increase) when the modulus of alkali activator increased from 1.0 to 2.0 ( Fig. 8 b). It is known that with the increase of modulus of alkali activator, a greater shrinkage was found which would lead to more initial micro- cracks [27,47] . Hence, the increased initial defects would generate adverse influence on strength improvement, whereas result in a more ductility matrix which could absorb more energy during crack propagation. Similarly, the improvements of Puand G Fwere 19.2% and 14.1%, respectively (i.e., from 2.99 kN to3.57 kN and from 118.7 N/m to 135.4 N/m) with the increase of slag/FA mass ratio from 50/50 to 100/0 ( Fig. 8 c). In addition, the increases of Puand G Fwith the decrease of liquid/binder ratio from 0.50 to 0.40 were 30.0% and 26.4%, respectively ( Fig. 8 d). The fracture energy of PCC can be predicted using its maximum aggregate size and compressive strength by CEB-FIP Model as Eq. (8). Besides, Baz ˇant and Becq-Giraduon proposed a formula as Eq.(9)by statistical analysis accounting the influences of the water/cement ratio and the aggregate shape into the fracture energy. GF¼0:0469 /C2D2 max/C00:5Dmaxþ26/C16/C17 /C2ðfc=10Þ0:7ð8Þ0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.801234Ms=1.5 l/b=0.45 slag/FA=50/50 n=3% n=4% n=5%Load (kN) Displacement (mm)0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.801234 n=5% w/b=0.45 slag/FA=50/50 Ms=1.0 Ms=1.5 Ms=2.0Load (kN) Displacement (mm) (a) Alkali concentration (b) Modulus 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.801234n=4% Ms=1.5 w/b=0.45 slag/FA=50/50 slag/FA=75/25 slag/FA=100/0Load (kN) Displacement (mm)0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.801234n=5% Ms=1.5 slag/FA=50/50 l/b=0.50 l/b=0.45 l/b=0.40Load (kN) CMOD (mm) (c) Slag/FA ratio (d) Liquid/binder ratio Fig. 7. P-dcurves of SFGC beams with different material parameters.792 Y. Ding et al. / Construction and Building Materials 190 (2018) 787–795 GF¼2:5a0fc 0:051/C18/C190:46 1þDmax 11:27/C18/C190:22w c/C16/C17 /C00:30 ð9Þ where Dmaxis the maximum size of aggregate (mm), a0is the aggre- gate shape factor ( a0¼1 for rounded aggregates, a0¼1:44 for angular aggregates), and w=cis the water/cement ratio of concrete. The fracture energy G Fof SFGC with different compressive strengths are compared with those predicted by CEB-FIP Model and Baz ˇant and Becq-Giraduon Model in Fig. 9 . The liquid (includ- ing the water in sodium silicate solution and extra added water) tobinder (the sum of GGBFS and FA) ratio in the SFGC replaced the water/cement ratio in PCC. It is clear that CEB-FIP Model signifi- cantly underestimated the fracture energy of SFGC. However, the predictions obtained by using Baz ˇant and Becq-Giraduon Model were close to the fracture energy of SFGC beams. For instance, the fracture energy G Fof SFGC-2 with compressive strength of 49.3 MPa and liquid/binder ratio of 0.45 was 118.7 N/m which was just 4.9% lower than that predicted by the Baz ˇant and Becq- Giraduon Model (i.e., 124.8 N/m). 3.6. Characteristic length Hillerborg et al. proposed characteristic length to reflect the brittleness of concrete as Eq. (10): lch¼GFE f2 stð10Þ The value of characteristic length can reflect the toughness of concrete. A brittle concrete always has a small characteristic length. Referring to the CEB-FIP Model, it is known that the fracture energy, the modulus of elasticity and the tensile strength of PCC could all be predicted by its compressive strength as Eqs. (11)– (13) (ad¼4:5 when the maximum size of aggregate is 10 mm). Hence, an expression using the compressive strength to estimate the characteristic length can be derived as Eq. (14). Hilsdorf and Brameshuber compared the experimental characteristic3% 4% 5%1.01.52.02.53.03.54.0 Ms=1.5 l/b=0.45 slag/FA=50/50 Peak load Alkali concentrationPu(kN) 90100110120130140150160170 Fracture energy GF (N/m) 11 . 521.01.52.02.53.03.54.0 n=5% l/b=0.45 slag/FA=50/50 Peak load ModulusPu(kN) Fracture energy GF (N/m) (a) Alkali concentration (b) Modulus 50/50 75/25 100/01.01.52.02.53.03.54.0 n=4% Ms=1.5 l/b=0.45 Peak load Slag/FA ratioPu(kN) Fracture energy GF (N/m) 0.5 0.45 0.41.01.52.02.53.03.54.0 n=5% Ms=1.5 slag/FA=50/50 Peak load Liquid/binder ratioPu(kN) Fracture energy GF (N/m) (c) Slag/FA ratio (d) Liquid/binder ratio Fig. 8. Parametric effects on fracture energy and peak load of SFGC. 30 35 40 45 50 55 60 65 70 75 800306090120150180 l/b=0.45 l/b=0.50 l/b=0.40 Eq.8 Eq.9 w/c=0.40 Eq.9 w/c=0.45 Eq.9 w/c=0.50Fracture energy (N/m) Compressive strength (MPa) Fig. 9. Comparison between fracture energy of SFGC beams and model predictions.Y. Ding et al. / Construction and Building Materials 190 (2018) 787–795 793 lengths of PCC with those estimated using the above method, the correlation coefficient was 0.72, demonstrating the creditability of this method in estimating the characteristic length of PCC. Sim- ilarly, Baz ˇant and Becq-Giraduon Model (Eq. (9)) can be combined with CEB-FIP Model (Eqs. (12) and (13) ) to predict the characteris- tic length of PCC as Eq. (15). GF¼adf0:7 c ð11Þ E¼104f1=3 c ð12Þ ft¼0:3f2=3 c ð13Þ lch¼1000 9ad/C1f/C00:3 c ð14Þ lch¼1596 f/C00:54 c ð15Þ The characteristic lengths lchof SFGC in this study are drawn against their compressive strengths in Fig. 10 . Regardless of the material parameters, the characteristic length of SFGC decreased as the compressive strength increased. Therefore, SFGC exhibited more brittle behavior as the compressive strength increased. A for- mula is proposed to predict the characteristic length of SFGC with R2= 0.895 as follows: lch¼2114 f/C00:54 c ð16Þ It can be seen from Fig. 10 that the characteristic lengths of SFGC were significantly higher than the evaluations using the CEB-FIP Model alone (Eq. (14)). According to earlier discussion, it is known that Baz ˇant and Becq-Giraduon Model predicted more precisely of the fracture energy of SFGC. Hence, comparing the characteristic length values of SFGC with the predictions using the combination of Baz ˇant and Becq-Giraduon Model and CEB- FIP Model (Eq. (15)) was more suitable. However, as shown in Fig. 10 , although the predictions of characteristic lengths using the combination of Baz ˇant and Becq-Giraduon Model and CEB- FIP Model were relatively closer to those of the experimental val- ues of SFGC, they were still smaller. Therefore, SFGC might feature higher ductility than PCC given the similar compressive strength which might be caused by the more shrinkage micro-cracks generated in SFGC matrix. 4. Conclusions This paper conducted a systematic experimental study on the fracture behaviors of SFGC cured in ambient temperature withdifferent alkali concentrations, moduli, slag/FA mass ratios and liq- uid/binder ratios through TPB tests. According to the experimental results and parametric analyses, the following conclusions could be obtained. (1) The compressive and the splitting tensile strengths of SFGC specimens as well as the peak load of the TPB tests and the fracture energy of SFGC beams increase with the increase of alkali concentration, the modulus of alkali activa- tor and the slag/FA ratio while decrease with the increase of liquid/binder ratio. (2) The splitting tensile strengths of SFGC fall within the range provided by EC2 Model and ACI 318 Model. The formula pro- posed for geopolymer concrete previously based on pub- lished data fits reasonably well of the splitting tensilestrength of SFGC. (3) CEB-FIP Model overestimates the elastic modulus of SFGC for about 20–30%. The lower modulus of elasticity of SFGC than its PCC counterpart might be caused by the lower intrinsic Young’s modulus of N-A-S-H gel and more serious initial micro-cracks formed in the SFGC. (4) The peak loads of SFGC beams improve marginally while the fracture energy increases obviously when the modulus of alkali activator increases from 1.0 to 2.0, and beam with M s= 2.0 exhibits more ductility behavior. (5) CEB-FIP Model underestimates while Baz ˇant and Becq- Giraduon Model obtains close predictions of the fracture energy of SFGC. (6) The characteristic lengths of SFGC cured in ambient temper- ature are larger than the predictions using the combination of Baz ˇant and Becq-Giraduon Model and CEB-FIP Model, indicating that SFGC might feature more ductility than PCC given the similar compressive strength. Conflict of interest None. Acknowledgements The authors are grateful for the financial support received from the National Science Foundation of China (NSFC) Project Nos. 51638008 and Construction Industry Council Fund (Project code: K-ZJK2).

Investigation of engineering properties of normal and high strength fly ash based geopolymer and alkali-activated slag concrete compared to ordinary Portland cement concrete Nabeel A. Farhan, M. Neaz Sheikh, Muhammad N.S. Hadi⇑ School of Civil, Mining and Environmental Engineering, University of Wollongong, Australia highlights /C15Engineering properties of FAGP and AAS concrete have been investigated. /C15The FAGP concrete has been produced by blending an alkaline activator with FA. /C15The AAS concrete has been produced by blending an alkaline activator with GGBFS. /C15The FAGP concrete has been producedusing heat curing at 80 /C176C for 24 h. /C15The AAS has been produced at ambient curing condition.graphical abstract article info Article history: Received 1 August 2018Received in revised form 7 November 2018Accepted 11 November 2018Available online 20 November 2018 Keywords: Fly ash-based geopolymer concreteAlkali-activated slag concreteEngineering propertiesHigh strengthNormal strengthabstract Fly ash-based geopolymer (FAGP) and alkali-activated slag (AAS) concrete are produced by mixing alka- line solutions with aluminosilicate materials. As the FAGP and AAS concrete are free of Portland cement,they have a low carbon footprint and consume low energy during the production process. This paper compares the engineering properties of normal strength and high strength FAGP and AAS concrete with OPC concrete. The engineering properties considered in this study included workability, dry density,ultrasonic pulse velocity (UPV), compressive strength, indirect tensile strength, flexural strength, directtensile strength, and stress-strain behaviour in compression and direct tension. Microstructural observa- tions using scanning electronic microscopy (SEM) are also presented. It was found that the dry density and UPV of FAGP and AAS concrete were lower than those of OPC concrete of similar compressivestrength. The tensile strength of FAGP and AAS concrete was comparable to the tensile strength of OPC concrete when the compressive strength of the concrete was about 35 MPa (normal strength concrete). However, the tensile strength of FAGP and AAS concrete was higher than the tensile strength of OPC con-crete when the compressive strength of concrete was about 65 MPa (high strength concrete). The mod- ulus of elasticity of FAGP and AAS concrete in compression and direct tension was lower than the modulus of elasticity of OPC concrete of similar compressive strength. The SEM results indicated thatthe microstructures of FAGP and AAS concrete were more compact and homogeneous than themicrostructures of OPC concrete at 7 days, but less compact and homogeneous than the microstructures of OPC concrete at 28 days for the concrete of similar compressive strength. /C2112018 Elsevier Ltd. All rights reserved. 0950-0618/ /C2112018 Elsevier Ltd. All rights reserved.⇑Corresponding author. E-mail address: mhadi@uow.edu.au (M.N.S. Hadi).Construction and Building Materials 196 (2019) 26–42 Contents lists available at ScienceDirect Construction and Building Materials journal homepage: www. elsevier.com/loc ate/conbuildmat 1. Introduction Cement is the main material used in the production of concrete. The production process of cement is associated with the consump- tion of high energy and natural resources. The production of cement is associated with the emission of greenhouse gases including methane, nitrous oxide and carbon dioxide into the atmosphere. Indeed, it is estimated that the production of one tonne of cement requires about 1.5 tonnes of raw materials and releases nearly one tonne of carbon dioxide into the atmosphere [1–4] . Thus, the use of aluminosilicate materials as an alternative to the cement has become necessary, especially to reduce the car- bon dioxide emissions into the atmosphere. Many research studies were carried out to develop new and greener materials as alterna- tives to cement such as geopolymer and alkali activated binder. Fly ash (FA) and Ground Granulated Blast Furnace Slag (GGBS) are the most common aluminosilicate materials used in the production of fly ash based geopolymer (FAGP) and alkali-activated slag (AAS) concrete. The FAGP and AAS concrete are green concrete without Portland cement. The FAGP and AAS concrete can be produced by blending an alkaline solution with aluminosilicate materials such as FA and GGBS. The FAGP and AAS concrete are proven to have comparable mechanical properties to the OPC concrete but with reduced greenhouse gas emissions. The use of FAGP or AAS con- crete can reduce CO 2emissions into atmosphere associated with the production of concrete by 60–80% [5–7] . Fernandez-Jimenez et al. studied the engineering properties of heat cured FAGP concrete and compared with the engineering properties of OPC concrete. The test results showed that the indi- rect tensile and flexural strengths of FAGP concrete were higher than those of OPC concrete. However, the modulus of elasticity of FAGP concrete was lower than the modulus of elasticity of OPC concrete. Hardjito and Rangan showed that FAGP concrete achieved similar compressive strength, higher indirect tensile and flexural strengths and lower modulus of elasticity than OPC concrete. Neupane et al. studied the engineering properties of heat cured FAGP concrete and compared with the engineering properties of OPC concrete. It was found that the indirect tensile and flexural strengths of FAGP concrete were higher than those of OPC concrete, whereas the modulus of elasticity of FAGP con- crete was similar to the modulus of elasticity of OPC concrete. Diaz-Loya et al. investigated the engineering properties of heat cured FAGP concrete. The engineering properties of heat cured FAGP concrete were found to be similar to those of OPC concrete.The test results also showed that the equations in the existing design standards for OPC concrete could be used for FAGP concrete to determine the indirect tensile strength, flexural strength, and the modulus of elasticity. Several studies investigated the engineering properties of AAS concrete and compared with the engineering properties of OPC concrete. Bernal et al. studied the engineering properties of AAS concrete produced in the laboratory at an ambient condition and compared with the engineering properties of OPC concrete. The compressive strength of AAS concrete was found to be compa- rable to the compressive strength of OPC concrete, but the indirect tensile and flexural strengths were slightly higher than those of OPC concrete. Lee et al. studied the engineering properties of AAS concrete produced in the laboratory at an ambient condi- tion and showed that the indirect tensile strength and modulus of elasticity of AAS concrete were slightly lower than those of OPC concrete. Chi investigated the mechanical and durability performance of AAS concrete and compared with the mechanical and durability performance of OPC concrete. The test results showed that AAS concrete could be produced with superior engi- neering properties (compressive strength, splitting tensilestrength, drying shrinkage, sulphate attack resistance, and high- temperature resistance) and the durability to those of OPC concrete. Most of the previous studies focused either on the engineering properties of FAGP concrete or the engineering properties of AAS concrete and compared with the engineering properties of OPC concrete. The engineering properties of FAGP and AAS concrete compared to the OPC concrete have not been adequately investi- gated in the available literature. Very limited information is cur- rently available for the engineering properties of FAGP and AAS concrete compared to the OPC concrete. An extensive review of lit- erature revealed, none of the research studies investigated the engineering properties of normal strength and high strength FAGP and AAS concrete compared with the engineering properties of OPC concrete. A complete understanding of the engineering prop-erties of FAGP and AAS concrete is important for the design and field implementation of eco-friendly concrete structures. This paper compares the engineering properties of normal strength and high strength FAGP and AAS concrete with the engineering properties of normal strength and high strength OPC concrete. Microstructural investigations using scanning electronic micro- scopy (SEM) are also carried out. The equations in the existing standards for OPC concrete were used to calculate indirect tensile strength, flexural strength and modulus of elasticity of FAGP and AAS concrete and compared with the experimental results. It is noted that the development of mathematical models for the engi- neering properties of FAGP and AAS concrete is considered beyond the scope of this paper. 2. Experimental investigation 2.1. Materials used The materials used in this study were FA, GGBS and General- purpose cement. The FA supplied by Gladstone Power Station, Aus- tralia was used as the source material for FAGP concrete. The GGBS supplied by the Australian Slag Association was used as the source material for AAS concrete. General purpose cement was used as the binder for OPC concrete. The chemical composition of FA and GGBS was determined by X-ray Fluorescent (XRF) and is shown in Table 1 . Chemical analyses of FA and GGBS were carried out in the School of Earth and Environmental Sciences at the University of Wollongong, Australia. Table 1 shows that FA contains less than 5% calcium oxide (CaO). The sum of Al 2O, SiO 2and Fe 2O3contents was higher than 70% of the FA components. The CaO content was less than 8% of the FA components. Hence, the FA used in this study can be classified as Type ‘F’ according to ASTM C618-08 . The Table 1 The chemical composition FA and GGBS. Composition (mass) Mass content (%) FA GGBS SiO 2 62.2 32.4 Al2O3 27.5 14.96 Fe2O3 3.92 0.83 CaO 2.27 40.70 MgO 1.05 5.99K 2O 1.24 0.29 Na2O 0.52 0.42 TiO 2 0.16 0.84 P2O5 0.30 0.38 Mn 2O3 0.09 0.40 SO3 0.08 2.74 Loss on ignition 0.89 NAN.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 27 chemical compositions of the OPC provided by cement Australia are shown in Table 2 . Crushed coarse aggregate with 10 mm maximum aggregate size in the saturated surface dry condition and locally available river sand (fine aggregate) were used to prepare all the test specimens. The alkaline activator was a mixture of sodium hydroxide (NaOH) and sodium silicate (Na 2SiO 3) solution. Sodium hydroxide (NaOH) pellets were dissolved in potable water to prepare the sodium hydroxide (NaOH) solution with different concentrations. Sodium silicate solution (Na 2SiO 3) with a specific gravity of 1.53 and an activator modulus (Ms) of 2.0 (Ms = SiO 2/Na 2O; SiO 2= 29.4% and Na2O = 14.7%) was supplied by PQ Australia. To obtain fresh con- crete with high workability, commercially available high range water reducer (Glenium 8700) supplied by BASF, Australia was used in this study. 2.2. Preparation of concrete mixes Three types of concrete were used in this study: FAGP, AAS and OPC concrete. The design compressive strengths of the concrete at 28 days were 35 MPa (normal strength concrete, NSC) and 65 MPa (high strength concrete, HSC). The total amount of aggregate in the FAGP and AAS concrete was between 60 and 80% of the mass of the concrete. The amount of aggregate varied depending on the amount of binder (FA and GGBS) and alkaline activator. The con- centration of NaOH used to prepare the normal strength and high strength FAGP concrete was 12 mol/litre (M) and 14 mol/litre (M), respectively. The ratio of sodium silicate (Na 2SiO 3) to sodium hydroxide (NaOH) was fixed at 2. The concentration of NaOH used to prepare the normal strength and high strength AAS concrete was 12 M and 14 M, respectively. The ratio of sodium silicate (Na 2SiO 3) to sodium hydroxide (NaOH) was fixed at 2.5. Extra water and highrange water reducer were added into the concrete mixes to obtain consistent workability during the casting of concrete. For the normal strength OPC concrete, the mix proportions by weight of cement, fine aggregate, and coarse aggregate were 1:2.2:3.3 with a maximum aggregate size of 10 mm and water to cement ratio of 0.52. For the high strength OPC concrete, the mixproportions by weight of cement, fine aggregate, and coarse aggre- gate were 1:1.3:2.3 with a maximum aggregate size of 10 mm and water to cement ratio of 0.30. Table 3 shows the mix proportions of FAGP, AAS and OPC concrete mixes. The concrete was mixed in an electrical pan mixer with a capac- ity of 0.1 m 3in the High Bay Laboratory at the University of Wol- longong, Australia. To produce FAGP and AAS concrete, the dry materials including FA or GGBS, fine aggregates and coarse aggre- gates were mixed for about four minutes. Afterwards, alkaline acti- vator, water and the high range water reducer were added to the dry mix, which was then mixed for another five minutes for a uni- form consistency of concrete. These fresh mixes were then poured into Polyvinyl chloride (PVC) moulds to prepare specimens to test the dry density, ultrasonic pulse velocity (UPV), compressive strength, indirect tensile strength and stress-strain behaviourunder compression. Also, the fresh concrete was poured into ply- wood moulds to prepare the specimen for the flexural and direct tensile strength tests. These mixes were then vibrated on a vibra- tion table for 1 min to remove air bubbles and to ensure that the concrete was adequately compacted. In total, 24 cylinder speci- mens with 100 mm diameter and 200 mm height were cast to test the dry density, ultrasonic pulse velocity (UPV) and compressive strength of FAGP and AAS concrete. In addition, 48 cylinder speci- mens with 150 mm diameter and 300 mm height were cast to test the indirect tensile strength and stress-strain behaviour. Moreover, 48 prism specimens with a cross-section of 100 mm /C2100 mm and a length of 500 mm were cast for the flexural and direct tensile strength tests. After casting, the FAGP and AAS concrete specimens were kept in the moulds and left in the laboratory at the ambient condition (temperature of 23 ± 3 /C176C) for 24 h. The FAGP concrete specimens were heat cured at 80 /C176C for 24 h. Then the specimens were removed from the moulds and left in the laboratory until the time of testing. The AAS concrete specimens were removed from the moulds after 24 h of casting and were left in the labora- tory at the ambient condition until the time of testing. The dry material (cement, fine and coarse aggregates) for OPC concrete were mixed for about four minutes and water and high range water reducer were slowly added. The mixing continued for another five minutes for a uniform consistency of concrete. The fresh mix was then poured into the steel moulds and vibrated for 1 min on a vibration table to remove any air bubbles and ensure that the concrete was adequately compacted. Twelve cylinder specimens of 100 mm diameter and 200 mm height were cast with OPC concrete to test dry density, ultrasonic pulse velocity (UPV) and compressive strength. In addition, 24-cylinder specimens of 150 mm diameter and 300 mm height were cast to test the indirect tensile strength and stress-strain behaviour under compression. Twenty-four prism specimens with a cross-section of100 mm /C2100 mm and a length of 500 mm were cast for the flex-Table 2 Chemical composition of cement. Composition (mass) Mass content (%) Portland cement Clinker <97 Gypsum (CaSO 42H2O) 2–5 Limestone (CaCO 3) 0–7.5 Calcium oxide (CaO) 0–3Hexavalent Chromium Cr (VI) <20 ppmCrystalline silica (Quartz) <1 Table 3 Mix proportion of FAGP, AAS and OPC concrete. Concrete mix Normal strength concrete (NSC) High strength concrete (HSC) FAGP AAS OPC FAGP AAS OPC Cement (kg/m3) – – 350 – – 461 GGBS (kg/m3) – 400 – – 450 – FA (kg/m3) 410 – – 480 – 29 Alkaline activator/Binder 0.45 0.45 – 0.35 0.35 – Fine aggregate (kg/m3) 627 636 760 606 625 650 Coarse aggregate (kg/m3) 1164 1169 1138 1140 1154 1150 Na2SiO 3/NaOH 2 2.5 – 2 2.5 – Na2SiO 3(kg/m3) 123 128 – 112 106 – NaOH (kg/m3) 61.5 52 – 56 53 – NaOH (moles/liter) 12 12 – 14 14 –Water (kg/m 3) 45 48 182 35 40 148 Superplasticizer (kg/m3) 22.5 20 8 17.5 12.5 6.528 N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 ural and direct tensile strength tests. After casting, the OPC con- crete specimens were kept in the moulds and left in the laboratory at the ambient condition (temperatures of 23 ± 3 /C176C) for 24 h. Afterwards, the specimens were removed from the moulds and cured in water until the time of testing. The preparation of FAGP and AAS concrete specimens are shown in Fig. 1 . 3. Test methods 3.1. Microstructural analysis The microstructure of primary materials (i.e. FA, GGBS and OPC) and the microstructure of FAGP, AAS and OPC concrete specimens were assessed using a Scanning Electron Microscope (SEM). The SEM analysis were carried out using JEOL-JSM 6490LV at the Elec- tron Micro Centre (EMC), University of Wollongong, Australia. The samples for SEM investigation of FAGP, AAS and OPC concrete specimens were taken from the broken particles of the specimens which were tested under compressive strength. The samples were cut for 20 mm in diameter and 10 mm high. The samples were left in the laboratory at the ambient condition for 7 days before testing to ensure that the samples were adequately dried and then coated with gold for SEM imaging. 3.2. Tests for fresh concrete Slump tests were carried out according to AS 1012.3.1-1998 to determine the consistency of the mixes. The workabilityof fresh concrete was determined by the slump test using a steel cone with a top diameter of 100 mm and a bottom diameter of 200 mm and a height of 300 mm. 3.3. Tests for hardened concrete To evaluate the engineering properties of hardened FAGP and AAS concrete and compare with the engineering properties of OPC concrete, dry density, ultrasonic pulse velocity, compressive strength, indirect tensile strength, flexural strength, direct tensile strength and stress-strain behaviour tests were carried out. The density of the hardened concrete was measured according to AS 1012.12.2-1998 . The density test was carried out on three specimens of 100 mm in diameter and 200 mm in height for each mix and the average density was recorded. Ultrasonic Pulse Veloc- ity (UPV) tests were carried out in accordance with ASTM C597- 2009 . The UPV test was carried out on three specimens of 100 mm in diameter and 200 mm in height for each mix and the average UPV was recorded. Three specimens were tested and the average result has been reported to evaluate the compressive strength and quality of the concrete based on the speed of a stress wave passing through a solid medium. The speed of the stress wave is related to the density of the concrete. The UPV test was carried out with a Portable Ultrasonic Non-destructive Digital Indi- cating Test set up. The compressive strength tests were carried out with the Avery compression testing machine of 1800 kN capacity according to AS 1012.9-1999 . Before testing, the specimens were capped with Fresh concrete Compressive strength specimens Failure mode (b) FAGP (a) AAS Fig. 1. Preparation and failure for: (a) FAGP concrete and (b) AAS concrete.N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 29 high strength plaster to ensure a uniform loading surface. Three specimens from each mix were tested and the average compressive strength was recorded. Indirect tensile strength tests were carried out to determine the tensile strength of concrete according to AS 1012.10-2000 . The specimens were tested with the Avery compression testing machine at a loading rate of 106 kN/min until the specimen failed. Three specimens from each mix were tested and the average indirect tensile strength was recorded as the ten- sile strength of concrete. The four-point bending tests were carried out according to AS 1012.11-2000 using an Avery 50 tonne testing machine at a loading rate of 2 kN/s. The specimens were tested until failure. The average measurement of three specimens was recorded as the flexural strength of concrete. The direct tensile strength of the specimens was determined according to the test setup proposed by Alhussainy et al. . The direct tensile test was carried out with a 500 kN Universal Instron testing machine at 0.1 mm/min. To ensure that the speci- mens fractured in the middle, the cross-sectional area in the mid- dle was reduced by 20% using two wooden triangular prisms. Three specimens were tested for each mix and the average direct tensile strengths have been reported. The stress-strain behaviour of specimens (150 mm diameter by 300 mm high) under compression was determined according to AS 1012.17-2014 with a 5000 kN Denison compression testing machine at a loading rate of 0.3 mm/min. Three linear variable dif- ferential transducers (LVDT) were used to record the axial defor- mation of the specimens. The specimens were capped before testing with high strength plaster to ensure uniform loading surfaces. 4. Results and discussion 4.1. Microstructural development The microscopic characteristics of primary materials (i.e., FA, GGBS and OPC) used in the production of FAGP, AAS and OPC con- crete are shown in Fig. 2 .Fig. 2 (a) shows that the FA consists mainly of glassy, spherical particles. The surfaces of the particles appear to be dense and smooth. The OPC and GGBS particles con- sist mainly of clear edges and angular shapes ( Fig. 2 b and c). The microstructural development of normal strength and high strength FAGP, AAS and OPC concrete are shown in Figs. 3–5 . The microstructure of normal strength and high strength FAGP con- crete showed an abundance of unreacted spherical shaped parti- cles of fly ash and a loose amorphous structure with visible micro-cavities in the FAGP concrete specimens at 7 days ( Fig. 3 ). These visible micro-cavities at 7 days are due to the evaporationof water from FAGP concrete specimens during the heat curing stage. The microstructure of FAGP concrete at 28 days showed less unreacted particles of fly ash. The structures of the geopolymer mixes look denser and more compact due to some additional geopolymerisation and the formation of aluminosilicate gel in the FAGP concrete specimens. The aluminosilicate gel diffused through the micro-cavities to fill the interior voids in the FAGP concrete specimens and increase adhesion with particles of geopolymer matrices, which resulted in a highly compacted and homogeneous structure . The microstructural development of normal strength and high strength AAS concrete displayed heterogeneous gel matrices at 7 days ( Fig. 4 ).Fig. 4 shows that most of the GGBS particles were partially dissolved by the alkaline activator to form C-S-H gel. Small microcracks were formed on the surface of the AAS microstructure due to a rapid reaction between the alkaline activa- tor and GGBS particles in the initial period [12,25] . After 28 days, the microstructural development of AAS concrete showed moreC-S-H gel due to the dissolution of the remaining unreacted GGBS particles. It is noted that, as the reaction continued, the small microcracks on the surface of the AAS microstructure were filled with C-S-H gel. This helped to bridge the microcracks on the sur- face of AAS microstructure. Hence, the density and uniformity of (a) FA (b) GGBS (c) OPC Fig. 2. SEM images for (a) FA, (b) GGBS and (c) OPC binder.30 N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 AAS microstructure increased and a more compacted and homoge- neous structure was formed between 7 and 28 days. The findings demonstrated in this study are consistent with those reported in few previous studies [25,26] . The microstructure of normal strength and high strength OPC concrete was less compact and homogeneous than FAGP and AAS concrete at 7 days ( Fig. 5 ). However, the microstructural develop- ment of OPC concrete at 28 days achieved denser microstructures and was more homogeneous than FAGP and AAS concrete at 28 days. Less unreacted OPC particles and no cracks were observed in the OPC matrices at 28 days. 4.2. Workability The workability of fresh FAGP, AAS and OPC concrete was mea- sured using slump test. The workability of fresh FAGP, AAS and OPC concrete was determined immediately after mixing the ingredients of the concrete. For the normal strength concrete (NSC), the fresh FAGP, AAS and OPC concrete were handled, placed, compacted and finished easily. It was observed that FAGP concrete exhibited the highest workability compared to AAS and OPC concrete. During the slump tests, it was observed that the FAGP concrete collapsed during the slump test as soon as the slump cone was lifted. Thiswas attributed to the spherical shaped particles of fly ash, which increased the followability of the mixes ( Fig. 2 a). In addition, the sodium silicate solution and the added water contributed further to the high flowability [27,28] . For the high strength concrete (HSC), the workability of FAGP, AAS and OPC concrete decreased with the decrease in the liquid/ binder and increase in the binder content. The decrease in the workability was more significant for AAS and OPC concrete. This can be attributed to the angular shape of the GGBS and OPC parti- cles, which increased the internal shear friction of the mixture . It was also observed that, with the increase in the NaOH concentra- tion, the viscosity of the alkaline activator solution was increased,which made the mix very sticky. As a result, the workability of the FAGP and AAS concrete decreased. 4.3. Dry density The dry density of FAGP, AAS and OPC concrete at 7 and 28 days are presented in Table 4 . For the NSC, the average dry density of FAGP, AAS and OPC concrete at 7 days was 2373 kg/m 3, 2389 kg/ m3and 2368 kg/m3, respectively. The dry density of FAGP, AAS and OPC concrete increased as the age of the concrete increased. The average density of FAGP concrete increased from 2373 kg/m3 (a) (b) 7 day s yad82 s Micro cavitiesN-A-S-H gel Unreacted FA Micro cavities Micro-cracks Unreacted FA 7 days N-A-S-H gel 28 days Fig. 3. SEM images of FAGP concrete: (a) Normal strength concrete and (b) High strength concrete.N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 31 at 7 days to 2378 kg/m3at 28 days with an overall increase of 0.21%. The average density of AAS concrete increased from 2389 kg/m3at 7 days to 2403 kg/m3at 28 days with an overall increase of 0.58%. The average density of OPC concrete increased from 2368 kg/m3at 7 days to 2415 kg/m3at 28 days with an over- all increase of 1.98%. The OPC concrete achieved the highest dry density compared to the dry density of FAGP and AAS concrete at 28 days. For the HSC, the average dry density of FAGP, AAS and OPC con- crete at 7 days were 2381 kg/m3, 2420 kg/m3and 2401 kg/m3, respectively. The dry density of FAGP, AAS and OPC concrete increased as the concrete age increased. The average density of FAGP concrete increased from 2381 kg/m3at 7 days to 2384 kg/ m3at 28 days, while the average density of AAS concrete increased from 2420 kg/m3at 7 days to 2432 kg/m3at 28 days. This increase in density was about 0.13% and 0.50% for FAGP and AAS concrete, respectively. The average density of OPC concrete increased from 2401 kg/m3at 7 days to 2443 kg/m3at 28 days with an overall increase of 1.75%. These results indicated that there were slight increases in the density of normal strength and high strength FAGP, AAS and OPC concrete over time. Whereas, the average density of FAGP and AAS concrete was less than the average density of OPC concrete with similar compressive strengths. These findings wereconfirmed by SEM analyses. The SEM images showed that FAGP and AAS concrete were less dense, less compacted, and had less homogeneous microstructures than OPC at 28 days ( Figs. 3–5 ). 4.4. Ultrasonic pulse velocity The ultrasonic pulse velocity (UPV) test is used to evaluate the strength and quality of concrete. The pulse velocity depends mostly on the density and properties of concrete. The pulse veloc- ity of FAGP, AAS and OPC concrete at 7 and 28 days are shown in Table 4 .Table 4 indicates that the pulse velocity of FAGP, AAS and OPC concrete increased as the concrete age increased. For the NSC, the average pulse velocity of FAGP concrete increased from 3.14 km/s at 7 days to 3.20 km/s at 28 days, while for AAS concrete the average pulse velocity increased from 3.18 km/s at 7 days to 3.31 km/s at 28 days. The increase in the pulse velocity of FAGP and AAS concrete was about 1.91% and 4.1%, respectively. The average pulse velocity of OPC concrete increased from 3.30 km/s at 7 days to 3.52 km/s at 28 days with an overall increase of 6.67%. The ultrasonic pulse velocity test results indicated that the quality of the concrete improved over time. The quality of the concrete can be evaluated according to the International Atomic Energy Agency , as shown in Table 5 . Based on the IAEA, (a) (b) Micro-cracks Dense microstructureC-S-H gel 28 days 7 day s yad82 s C-S-H gel Dense microstructure 7 days Micro-cracks Fig. 4. SEM images of AAS concrete: (a) Normal strength concrete and (b) High strength concrete.32 N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 OPC concrete can be classified as ‘‘medium” quality at 7 days, because the pulse velocity was 3.30 km/s. As the pulse velocity increased to 3.52 km/s at 28 days, the concrete can be classified as ‘‘good” quality. The average pulse velocity of FAGP and AAS con- crete is less than the average pulse velocity of OPC concrete, which was between 3 and 3.5 km/s at 7 and 28 days. Hence, the FAGP and AAS concrete are classified as ‘‘medium” quality concrete . For the HSC, the average pulse velocity of FAGP concrete increased from 3.82 km/s at 7 days to 3.93 km/s at 28 days with (a) (b) Pores Dense microstructure Dense microstructure7 days 28 days 7 days 28 days Fig. 5. SEM images of OPC concrete: (a) Normal strength concrete and (b) High strength concrete. Table 4 Engineering properties of FAGP, AAS and OPC concrete at 7 and 28 days. ConcreteMixDesign compressive strength (MPa) at 28 daysDry density (kg/m 3)Ultrasonic pulse velocity (km/s)Compressive strength (MPa)Indirect tensile strength (MPa)Flexural strength (MPa)Direct tensile strength (MPa) 7 days 28 days 7 days 28 days 7 days 28 days 7 days 28 days 7 days 28 days 7 days 28 days FAGP-35 35 2373 2378 3.14 3.20 33.90 35.91 3.37 3.58 3.57 3.81 2.33 2.43 AAS-35 2389 2403 3.18 3.31 29.03 36.44 2.93 3.55 3.21 3.79 2.02 2.42OPC-35 2368 2415 3.30 3.52 26.51 35.82 2.66 3.51 3.06 3.78 1.91 2.41 FAGP-65 65 2381 2384 3.82 3.93 61.71 65.28 5.32 5.73 6.07 6.42 3.36 3.52 AAS-65 2420 2432 3.78 3.98 53.68 66.12 4.49 5.23 5.40 6.31 2.93 3.52 OPC-65 2401 2443 3.87 4.15 50.73 66.69 3.78 4.94 4.57 5.81 2.79 3.51 Table 5 Classification of the quality of concrete based on ultrasonic pulse velocity. Longitudinal pulse velocity (km/s) Quality of concrete >4.5 Excellent 3.5–4.5 good3.0–3.5 medium2.0–3.0 Poor <2.0 Very poorN.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 33 an increase of 2.88%. The average pulse velocity of AAS concrete increased from 3.78 km/s at 7 days to 3.98 km/s at 28 days with an increase of 5.29%. The average pulse velocity of OPC concrete increased from 3.87 km/s at 7 days to 4.15 km/s at 28 days with an increase of 7.23%. The pulse velocity of FAGP concrete was lower than the pulse velocity of OPC concrete at 7 and 28 days. Similarly, the pulse velocity of AAS concrete was lower than the pulse veloc- ity of OPC concrete at 7 and 28 days. Since the pulse velocity of FAGP, AAS and OPC concrete at 7 and 28 days ranged between 3.5 and 4.5 km/s, they can be classified as ‘‘good” quality concrete . 4.5. Compressive strength The average compressive strength of FAGP, AAS and OPC con- crete at 7 and 28 days are shown in Table 4 . The compressive strength of AAS and FAGP concrete is comparable to the OPC con- crete at 28 days ( Table 4 ). For the NSC with the design compressive strength of 35 MPa, the average compressive strength of FAGP, AAS and OPC concrete at 7 days was 33.90 MPa, 29.03 MPa and 26.51 MPa, respectively. The FAGP concrete achieved the highest initial compressive strength at 7 days, which was 94.44% of the compressive strength at 28 days. However, AAS and OPC concrete obtained a lower initial compressive strength than FAGP concrete at 7 days, which were 79.66% and 74.01% of the compressive strength at 28 days. The compressive strength of FAGP, AAS and OPC concrete increased with time ( Table 4 ), the average compres- sive strength of FAGP, AAS and OPC concrete at 28 days was 35.91, 36.44 MPa and 35.82 MPa, respectively. For the HSC with the design compressive strength of 65 MPa, the average compressive strength of FAGP, AAS and OPC concrete at 7 days was 61.71 MPa, 53.68 MPa and 50.73 MPa, respectively. The FAGP concrete achieved the highest initial compressive strength at 7 days, which was 94.53% of the compressive strength at 28 days. The compressive strength of AAS and OPC concrete at 7 days were 81.20% and 76.06%, respectively, of the compressive strength at 28 days. The compressive strengths of FAGP, AAS and OPC concrete increased with time. The average compressive strengths of FAGP, AAS and OPC concrete at 28 days were 65.28, 66.12 MPa, and 66.69 MPa, respectively. For the NSC and HSC, FAGP concrete developed most of its compressive strength at 7 days although there was a slight increase in the compressive strength at 28 days ( Table 4 ) due to heat curing, which accelerated the geopolymerisation (dissolution mechanism) reaction and increased the compressive strength. The findings of this study agree with Adam , in which it was shown that FAGP concrete developed most of its compressive strength at 7 days and there was a marginal increase in the compressive strength at 28 days. 4.6. Indirect tensile strength The indirect tensile strength of FAGP, AAS and OPC concrete was determined at 7 and 28 days, and the results are reported in Table 4 . For the NSC, the average indirect tensile strength of FAGP, AAS and OPC concrete at 7 days was 3.37 MPa, 2.93 MPa and 2.66 MPa, respectively. The FAGP concrete achieved the highest indirect tensile strength at 7 days. The indirect tensile strength of FAGP, AAS and OPC concrete increased as the concrete age increased. The average indirect tensile strength of FAGP, AAS and OPC concrete at 28 days was 3.58 MPa, 3.55 MPa and 3.51 MPa, respectively. The indirect tensile strength of FAGP, AAS and OPC concrete increased by 6.23%, 21.16% and 31.95% at 28 days, respec- tively, compared to the indirect tensile strengths at 7 days. When compared with the OPC concrete, the FAGP and AAS concrete achieved very similar indirect tensile strength at 28 days ( Table 4 ).For the HSC, the average indirect tensile strength of FAGP, AAS and OPC concrete at 7 days was 5.32 MPa, 4.49 MPa and 3.78 MPa, respectively. The FAGP concrete achieved the highest indirect ten- sile strength at 7 days. The indirect tensile strength of FAGP, AAS and OPC concrete increased with age. The average indirect tensile strength of FAGP, AAS and OPC concrete at 28 days was 5.73 MPa, 5.23 MPa and 4.94 MPa, respectively. The indirect tensile strength of FAGP, AAS and OPC concrete increased by 7.71%, 16.48% and 30.68% at 28 days, respectively. From the test results, it can be observed that the FAGP and AAS concrete achieved about 15.99% and 5.87%, respectively, higher indirect tensile strength at 28 days than OPC concrete of similar compressive strength. These results are consistent with previous studies carried out on FAGP and AAS concrete. Ryu et al. examined the indirect tensile strength of fly ash based geopolymer concrete and found that the indirecttensile strength of geopolymer concrete was higher than the indi- rect tensile strength of OPC concrete. Bernal et al. reported that AAS concrete achieved a higher indirect tensile strength than OPC concrete at 28 days. 4.7. Flexural strength The flexural strength is generally higher than the indirect ten- sile strength as specified in the ACI 318-14 and AS 3600- 2009 . The average flexural strengths of FAGP, AAS and OPC concrete at 7 and 28 days are shown in Table 4 . For the NSC, the average flexural strength of FAGP, AAS and OPC concrete at 7 days was 3.57 MPa, 3.21 MPa and 3.06 MPa, respectively. The FAGP con- crete achieved the highest flexural strength at 7 days. The flexural strength of FAGP, AAS and OPC concrete increased with age. The average flexural strength of FAGP, AAS and OPC concrete at 28 days was found to be 3.81 MPa, 3.79 MPa and 3.78 MPa, respectively. The flexural strength of FAGP, AAS and OPC concrete increased by 6.72%, 18.07% and 23.53%, respectively, at 28 days compared to the flexural strengths at 7 days. From the test results, it can be seen that a significant development in the flexural strength of FAGP concrete at 7 days (3.57 MPa), which was 93.70% of its flexu- ral strength at 28 days. The flexural strength of FAGP and AAS con- crete was very similar to the OPC concrete at 28 days, as shown in Table 4 . For the HSC, the average flexural strength of FAGP, AAS and OPC concrete at 7 days was 6.07 MPa, 5.40 MPa and 4.57 MPa, respec- tively. The FAGP concrete achieved the highest flexural strength at 7 days. The flexural strengths of FAGP, AAS and OPC concrete increased with the increase in the age of concrete. The average flexural strength of FAGP, AAS and OPC concrete at 28 days was 6.42 MPa, 6.31 MPa and 5.81 MPa, respectively. The flexural strength of FAGP, AAS and OPC concrete increased by 5.76%,16.85% and 27.13%, respectively, at 28 days compared to the flexu- ral strengths at 7 days. The FAGP concrete achieved the highest flexural strength at 7 days (6.07 MPa), which was 94.54% of its flex- ural strength at 28 days. The flexural strength of FAGP and AAS concrete was 10.5% and 8.6%, respectively, higher than the flexural strengths of OPC concrete at 28 days ( Table 4 ). These findings agree with previous studies which reported that FAGP concrete achieved higher flexural strength than OPC concrete for heat cured and ambient cured geopolymer concrete of similar compressive strengths [8,11,35,36] . Sarker et al. also reported that AAS con- crete had higher flexural strengths than OPC concrete of similar compressive strengths. 4.8. Stress-strain behaviour under uniaxial tension The stress-strain behaviour under uniaxial tension of normal strength and high strength FAGP, AAS and OPC concrete are shown inFigs. 6 and 7. It can be observed that the ascending branches of34 N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 the stress-strain curves of FAGP, AAS and OPC concrete exhibited similar behaviours up to the peak stress. After reaching peak stress, the FAGP, AAS and OPC concrete showed a brittle failure as soon as they reached the peak stress. The reduction of the cross-sectional area in the middle increased the stresses in the middle of the spec- imens and induced uniform failure in the middle of the specimens. 4.8.1. Direct tensile strength The direct tensile strength of normal strength FAGP, AAS and OPC concrete are presented in Table 4 . The average direct tensile strengths of FAGP, AAS and OPC concrete at 7 days was 2.33 MPa, 2.02 MPa and 1.91 MPa, respectively. The average direct tensile strength of FAGP, AAS and OPC concrete at 28 days was2.43 MPa, 2.42 MPa and 2.41 MPa, respectively. The direct tensile strength of FAGP, AAS and OPC concrete increased by 4.29%, 19.80% and 26.18% at 28 days, respectively, compared to the direct tensile strength at 7 days. The high strength FAGP, AAS and OPC concrete specimens achieved average direct tensile strengths at 7 days of 3.36 MPa, 2.93 MPa and 2.79 MPa, respectively ( Table 4 ). The direct tensile strength of FAGP, AAS and OPC concrete increased with the increase in the concrete age. The average direct tensile strength of FAGP, AAS and OPC concrete at 28 days was 3.52 MPa, 3.52 MPa and 3.51 MPa, respectively ( Table 4 ). The direct tensile strength of FAGP, AAS and OPC concrete increased by 4.76%, 20.14% and 25.81%, respectively, at 28 days compared to the direct tensile strength at 7 days. It was observed that the average direct tensile strength of FAGP, AAS and OPC concrete was less than the average indirect tensile and flexural strength of FAGP, AAS and OPC concrete, respectively. The lower direct tensile strength compared to the indirect tensileand flexural strengths was similar to the observation reported in Swaddiwudhipong et al. for normal strength OPC concrete. The average direct tensile strength of normal strength FAGP, AAS and OPC concrete was found to be 32%, 30% and 31% less thanthe average indirect tensile strength of FAGP, AAS and OPC con- crete at 28 days, respectively. Also, the average direct tensile strength of FAGP, AAS and OPC concrete was found to be 37%, 33% and 36% less than the average flexural strength of FAGP, AAS and OPC concrete at 28 days, respectively. For the HSC, the average direct tensile strength of FAGP, AAS and OPC concrete was found to be 38%, 32% and 29% less than the average indirect tensile strength of FAGP, AAS and OPC concrete at 28 days, respectively. Also, the average direct tensile strength of FAGP, AAS and OPC concrete was found to be 45%, 44% and 40% less than the average flexural strength of FAGP, AAS and OPC concrete at 28 days, respectively. 4.8.2. Peak stress and corresponding strain The peak stress and strain at peak stress of normal strength FAGP, AAS and OPC concrete are presented in Table 6 . It can be observed that the FAGP, AAS and OPC concrete specimens achieved peak stresses at 7 days of 2.33 MPa, 2.02 MPa and 1.91 MPa, respectively. The FAGP concrete achieved higher peak stress than OPC and AAS at 7 days. However, the peak stress of FAGP, AAS and OPC concrete specimens was similar at 28 days. The specimens of FAGP, AAS and OPC concrete achieved peak stresses at 28 days of 2.43 MPa, 2.42 MPa and 2.41 MPa, respectively. The peak stresses of FAGP, AAS and OPC concrete increased by 4.29%, 19.80% and 26.18% at 28 days, respectively. Also, the strain corresponding peak stress of FAGP, AAS and OPC concrete increased by 7.14%, 16.67% and 8.34%, respectively, at 28 days compared to the strain at peak stresses at 7 days.(a) (b) 01234 0 0.005 0.01 0.015 0.02Stress (MPa) Strain (%)FAGP AAS OPC0 0.005 0.01 0.015 0.02Stress (MPa) Strain (%)FAGP AAS OPC Fig. 6. Typical stress-strain behaviour under uniaxial tension for specimens of design compressive strength of 35 MPa: (a) at 7 days and (b) at 28 days.(a) (b) 01234 0 0.005 0.01 0.015 0.02 0.025Stress (MPa) Strain (%)FAGP AAS OPC0 0.005 0.01 0.015 0.02 0.025Stress (MPa) Strain (%)FAGP AAS OPC Fig. 7. Typical stress-strain behaviour under uniaxial tension for specimens of design compressive strength of 65 MPa: (a) at 7 days and (b) at 28 days.N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 35 For the HSC, the peak stress of FAGP, AAS and OPC concrete at 7 days was 3.36 MPa, 2.93 MPa and 2.79 MPa, respectively. The peak stress of FAGP, AAS and OPC concrete increased with time. The FAGP, AAS and OPC concrete achieved peak stresses of 3.52 MPa, 3.52 MPa and 3.51 MPa at 28 days ( Table 6 ). The peak stresses of FAGP, AAS and OPC concrete increased by 4.76%, 20.14% and 25.81%, respectively, at 28 days compared to the peak stresses at 7 days. Also, the strain corresponding peak stress of FAGP, AAS and OPC concrete increased by 17.64%, 12.5% and 13.34%, respectively, at 28 days compared to the strain at peak stresses at 7 days ( Table 6 ). 4.8.3. Modulus of elasticity The modulus of elasticity of FAGP, AAS and OPC concrete was calculated using the slope of ascending branches of tensile stress-strain curves. The modulus of elasticity of normal strengthand high strength FAGP, AAS and OPC concrete are presented in Table 6 . For NSC, the modulus of elasticity at 7 days was 16.59 GPa, 16.20 GPa and 16.23 GPa for the FAGP, AAS and OPC concrete specimens, respectively ( Table 6 ). The modulus of elastic- ity at 28 days was 16.63 GPa, 16.59 GPa and 17.98 GPa for the FAGP, AAS and OPC concrete specimens, respectively. The modulus of elasticity of FAGP, AAS and OPC concrete increased by 0.24%, 2.41% and 10.78% at 28 days, respectively, compared to the modu- lus of elasticity at 7 days. The OPC concrete achieved 8.12% and 8.38% higher modulus of elasticity than FAGP and AAS concrete at 28 days, respectively. The modulus of elasticity of high strength FAGP, AAS and OPC concrete was 19.22 GPa, 18.38 GPa and 18.66 GPa at 7 days, respectively ( Table 6 ). The modulus of elastic- ity of FAGP, AAS and OPC concrete at 28 days was found to be 19.46 GPa, 19.36 GPa and 20.95 GPa, respectively. The modulus of elasticity of FAGP, AAS and OPC concrete increased by 1.25%, 5.33% and 12.27% at 28 days, respectively; compared to the modu- lus of elasticity at 7 days. The OPC specimens achieved 7.65% and 8.21% higher modulus of elasticity than FAGP and AAS concrete at 28 days, respectively. 4.9. Stress-strain behaviour in compression For the NSC, the experimental stress-strain behaviour in com- pression of the specimens of FAGP, AAS and OPC concrete at 7 and 28 days are shown in Fig. 8 . It was observed that the ascending branch of the stress-strain curves of FAGP, AAS and OPC concrete was almost linear until the peak stress ( Fig. 8 ). After reaching peak stress, the FAGP and AAS concrete showed a more rapid decline in the descending branch of the stress-strain curves and failed in a brittle manner immediately after the peak stress. However, OPC concrete showed a softening decline in the descending branch of the stress-strain curves. The increase in the brittleness of FAGP and AAS concrete was also reported by Atis /C223et al. and can be attributed to the high micro-cracking in FAGP and AAS concrete . For the HSC, the experimental stress-strain behaviour of spec- imens of FAGP, AAS and OPC concrete at 7 and 28 days are showninFig. 9 . As the compressive strength increased, the slope of the ascending and descending branches of the stress-strain curves became steeper ( Fig. 9 ). In addition, the failure was more sudden and explosive rather than continual softening. 4.9.1. Peak stress and corresponding strain The peak stress and strain at peak stress obtained from the stress-strain curve are shown in Table 7 . For the NSC, the peak stress of FAGP, AAS and OPC concrete at 7 days was 32.40 MPa, 26.88 MPa and 24.81 MPa, respectively ( Table 7 ). The FAGP con- crete achieved higher peak stress than AAS and OPC concrete at 7 days. The peak stress for FAGP concrete increased slightly with time, whereas the peak stress of AAS and OPC concrete increased significantly with time. The peak stress of FAGP, AAS and OPC con- crete at 28 days was 33.39 MPa, 34.08 MPa and 33.06 MPa, respec- tively. The peak stress of FAGP, AAS and OPC concrete increased by 3.05%, 26.78% and 33.25%, respectively, at 28 days compared to the peak stresses at 7 days. While, the strain corresponding peak stress of FAGP, AAS and OPC concrete increased by 1.83%, 5.42% andTable 6 Experimental results of the peak stress, strain at peak stress, and modulus of elasticity of the tested specimens under uniaxial tension. Concrete Mix Average peak stress (MPa) Average strain at peak stress * 10/C03Average modulus of elasticity (GPa) 7 days 28 days 7 days 28 days 7 days 28 days FAGP-35 2.33 2.43 0.14 0.15 16.59 16.63 AAS-35 2.02 2.42 0.12 0.14 16.20 16.59OPC-35 1.91 2.41 0.12 0.13 16.23 17.98FAGP-65 3.36 3.52 0.17 0.20 19.22 19.46AAS-65 2.93 3.52 0.16 0.18 18.38 19.36OPC-65 2.79 3.51 0.15 0.17 18.66 20.95 (a) (b) 010203040 0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008Stress (MPa) Strain (mm/mm)FAGP AAS OPC0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008Stress (MPa) Strain (mm/mm)FAGP AAS OPC Fig. 8. Typical stress-strain behaviour under compression for specimens of design compressive strength of 35 MPa: (a) at 7 days and (b) at 28 days.36 N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 2.46%, respectively, at 28 days compared to the strain at peak stress at 7 days ( Table 7 ). For the HSC, the peak stress of FAGP, AAS and OPC concrete at 7 days was 59.36 MPa, 52.18 MPa and 48.56 MPa, respectively.The peak stress of FAGP concrete was higher than AAS and OPC concrete at 7 days. The peak stress of FAGP concrete slightly increased with time, whereas the peak stress of AAS and OPC con- crete increased significantly with time. The peak stress of FAGP, AAS and OPC concrete at 28 days was 63.07 MPa, 64.26 MPa and 63.34 MPa respectively. The peak stress of FAGP, AAS and OPC con- crete increased by 6.25%, 23.15% and 30.44%, respectively, at 28 days compared to the peak stresses at 7 days. The strain corre- sponding to the peak stress of FAGP, AAS and OPC concrete increased by 3.65%, 2.55% and 12.96%, respectively, at 28 days compared to the strain at peak stresses at 7 days ( Table 7 ). 4.9.2. Modulus of elasticity The modulus of elasticity was calculated according to ACI 318- 11 as the slope of the tangent of a stress-strain curve drawn from the origin to the stress equals 45% of the peak stress. Theslope of the tangent represents the modulus of elasticity of FAGP, AAS and OPC concrete. The modulus of elasticity of normal strength FAGP, AAS and OPC concrete are presented in Table 7 . The modulus of elasticity of FAGP, AAS and OPC concrete at 7 days was 17.34 GPa, 16.82 GPa and 18.78 GPa, respectively. The modu- lus of elasticity increased as the concrete age increased. The mod- ulus of elasticity of FAGP, AAS and OPC concrete at 28 days was 18.05 GPa, 17.95 GPa and 20.20 GPa, respectively. The modulus of elasticity of FAGP, AAS and OPC concrete increased by 4.09%, 6.72% and 7.56%, respectively, at 28 days compared to the modulus of elasticity at 7 days. The modulus of elasticity of high strength FAGP, AAS and OPC concrete was 21.35 GPa, 20.21 GPa and 22.10 GPa, respectively, at 7 days ( Table 7 ). The modulus of elasticity increased as the con- crete age increased. The modulus of elasticity of FAGP, AAS andOPC concrete at 28 days was found to be 24.47 GPa, 23.30 GPa and 27.63 GPa, respectively. The modulus of elasticity of FAGP, AAS and OPC concrete increased by 14.61%, 15.29% and 25.02%, respectively, at 28 days compared to the modulus of elasticity at 7 days. As such, the FAGP and AAS concrete had a lower modulus of elasticity than OPC concrete with similar compressive strength. The experimental results indicated that FAGP concrete had about 12–13% less modulus of elasticity than OPC concrete at 28 days. The AAS concrete had about 13–19% less modulus of elasticity than OPC concrete at 28 days. A similar observation was reported by Oli- via and Nikraz for heat cured fly ash based geopolymer con- crete which exhibited a modulus of elasticity of 14.9–28.8% less than OPC concrete with similar compressive strengths. Hardjito et al. reported that the modulus of elasticity of heat cured fly ash based geopolymer was about 10% lower than OPC concrete with similar compressive strengths. Yang et al. and Douglas et al. also reported that alkali-activated concrete generally had a lower modulus of elasticity than OPC concrete with similar compressive strengths. 5. Comparison between calculated and experimental results The design standards specified equations to calculate indirect tensile strength, flexural strength and modulus of elasticity from compressive strength of OPC concrete. The equations specified in the ACI 318-14 and AS 3600-2009 for OPC concrete and the equations proposed in the previous studies [11,42,44–46] for geopolymer concrete were used to calculate indirect tensile strength, flexural strength and modulus of elasticity of FAGP and AAS concrete and compared with the experimental results. 5.1. Indirect tensile strengths The ACI 318-14 specified Eq. (1)as the approximate rela- tionship between the indirect tensile strength and the compressive strength. fct:sp¼0:56ffiffiffiffiffiffi fC0q ðMPaÞð 1Þ(a) (b) 020406080 0 0.001 0.002 0.003 0.004 0.005Stress (MPa) Strain (mm/mm)FAGP AAS OPC0 0.001 0.002 0.003 0.004 0.005Stress (MPa) Strain (mm/mm)FAGP AAS OPC Fig. 9. Typical stress-strain behaviour under compression for specimens of design compressive strength of 65 MPa: (a) at 7 days and (b) at 28 days. Table 7 Experimental results of peak stress, strain at peak stress, and the modulus of elasticity of specimens tested under compression. Concrete Mix Average peak stress Average strain at peak stress Average modulus of elasticity (GPa) 7 days 28 days 7 days 28 days 7 days 28 days FAGP-35 32.40 33.39 0.00219 0.00223 17.34 18.05 AAS-35 26.88 34.08 0.00203 0.00214 16.82 17.95OPC-35 24.81 33.06 0.00203 0.00208 18.78 20.20 FAGP-65 59.36 63.07 0.00301 0.00312 21.35 24.47 AAS-65 52.18 64.26 0.00275 0.00282 20.21 23.30OPC-65 48.56 63.34 0.00216 0.00244 22.10 27.63N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 37 where fct:spis indirect tensile strength (MPa) and fC0is the specified compressive strength (MPa) at 28 days. The AS 3600-2009 specified Eq. (2)as the relationship between the indirect tensile strength and compressive strength. fct:sp¼0:36ffiffiffiffi f0 cq ðMPaÞð 2Þ Sofi et al. proposed Eq. (3)for the relationship between indirect tensile strength and compressive strength of fly ash based geopolymer concrete. fct:sp¼0:48ffiffiffiffiffiffi fC0q ðMPaÞð 3Þ Gunasekera et al. proposed Eq. (4)for the relationship between indirect tensile strength and compressive strength of concrete.fct:sp¼0:45ffiffiffiffiffiffi fC0q ðMPaÞð 4Þ The relationship between indirect tensile strength and com- pressive strength of the experimental and calculated values are shown in Fig. 10 . It can be seen that the experimental indirect ten- sile strength of normal strength FAGP and AAS concrete are close to the calculated indirect tensile strength using ACI 318-14 and mostly higher than those calculated using AS 3600-2009 , Sofi et al. and Gunasekera et al. . However, the experimental indirect tensile strength for high strength FAGP and AAS concrete were higher than the indirect tensile strength calculated using ACI 318-14 , AS 3600-2009 , Sofi et al. and Gunasekera et al. (Fig. 10 ). The results obtained using ACI 318-14 for OPC concrete provided a conservative estimate of normal strength (a) (b) 02468Indirect TensileStrength (MPa)Expermintal results (7 d) Expermintal results (28 d) ACI 318-14 AS 3600 Sofi et al. Gunasekera et al. 56789Indirect TensileStrength (MPa)Expermintal results (7 d) Expermintal results (28 d) ACI 318-14 AS 3600 Sofi et al. Gunasekera et al. Fig. 10. Indirect tensile strength versus compressive strength: (a) FAGP concrete and (b) AAS concrete.38 N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 FAGP and AAS concrete. However, the ACI 318-14 for OPC con- crete did not provide a conservative estimate of high strength FAGP and AAS concrete. 5.2. Flexural strengths The equations in the ACI 318-14 and AS 3600-2009 for OPC concrete and proposed in previous studies [11,44,46] for geopolymer concrete were used to calculate the flexural strength of FAGP and AAS concrete and compared with the experimental results. The ACI 318-14 recommended Eq. (5)for the relationship between the flexural strength and compressive strength of concrete. fct:f¼0:62ffiffiffiffiffiffi fC0q ðMPaÞð 5Þwhere fct:fis the flexural strength (MPa) and fC0is the specified compressive strength (MPa) at 28 days. The AS 3600-2009 recommended Eq. (6)for the relation- ship between the flexural strength and compressive strength of concrete. fct:f¼0:6ffiffiffiffiffiffi fC0q ðMPaÞð 6Þ Diaz-Loya et al. suggested Eq. (7)for the relationship between the flexural and compressive strength of geopolymer concrete. fct:f¼0:69ffiffiffiffiffiffi fC0q ðMPaÞð 7Þ Nath and Sarker proposed Eq. (8)for the relationship between the flexural strength and compressive strength of concrete. (a) (b) 02468Flexural Strength (MPa)Expermintal results (7 d) Expermintal results (28 d) ACI 318-14 AS 3600 Diaz-Loya et al. Nath and Sarker 56789Flexural Strength (MPa)Expermintal results (7 d) Expermintal results (28 d) ACI 318-14 AS 3600 Diaz-Loya et al. Nath and Sarker Fig. 11. Flexural strength versus compressive strength: (a) FAGP concrete and (b) AAS concrete.N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 39 fct:f¼0:93ffiffiffiffiffiffi fC0q ðMPaÞð 8Þ The relationship between flexural strength and compressive strength of the experimental and calculated values are drawn in Fig. 11 .Fig. 11 indicates that the experimental flexural strength of normal strength FAGP and AAS concrete are comparable to those calculated using ACI 318-14 and AS 3600-2009 . However, the experimental flexural strength of normal strength FAGP and AAS concrete are lower than those calculated using Diaz-Loya et al. and Nath and Sarker for geopolymer concrete. The experimental flexural strength of high strength FAGP and AAS con- crete are higher than those calculated using ACI 318-14 ,A S 3600-2009 and Diaz-Loya et al. and lower than those cal- culated using Nath and Sarker . This means that ACI 318-14 and AS 3600-2009 for OPC provided a conservative esti-mate of normal strength FAGP and AAS concrete in terms of flexu- ral strength. However, the ACI 318-14 and AS 3600-2009 for OPC concrete did not provide a conservative estimate of high strength FAGP and AAS concrete. 5.3. Modulus of elasticity The equations specified in the ACI 318-14 and AS 3600- 2009 for OPC concrete were used to calculate modulus of elas- ticity of FAGP and AAS concrete and compared with the experi- mental results. Also, the equations proposed in Hardjito et al. and Diaz-Loya et al. for geopolymer concrete were used to calculate the modulus of elasticity of FAGP and AAS concrete and compared with the experimental results. (a) (b) 01020304050Modulus of Elaciticity (GPa)Expermintal results (7 d) Expermintal results (28 d) ACI 318-14 AS 3600 Hardjito et al. Diaz-Loya et al. 56789Modulus of Elaciticity (GPa)Expermintal results (7 d) Expermintal results (28 d) ACI 318-14 AS 3600 Hardjito et al. Diaz-Loya et al. Fig. 12. Modulus of elasticity versus compressive strength: (a) FAGP concrete and (b) AAS concrete.40 N.A. Farhan et al. / Construction and Building Materials 196 (2019) 26–42 The ACI 318-14 specified Eq. (9)for the modulus of elastic- ity of OPC concrete. EC¼ðq1:5Þ/C2 0:043ffiffiffiffiffiffi fC0q /C16/C17 ð9Þ where ECis the modulus of elasticity, qis the density of concrete (kg/m3) and fC0is compressive strength at 28 days. The AS 3600-2009 specified Eq. (10) for the modulus of elasticity of OPC concrete. EC¼ðq1:5Þ/C2 0:024ffiffiffiffiffiffi fC0q þ0:12/C16/C17 when fC0>40MPa ð10Þ According to AS 3600-2009 , the modulus of elasticity can be calculated using a similar equation proposed in the ACI 318- 14 for OPC concrete of compressive strength less than 40 MPa. Hardjito et al. proposed Eq. (11) for the modulus of elastic- ity of geopolymer concrete. EC¼2707ffiffiffiffiffiffi fC0q þ5300 ð11Þ Diaz-Loya et al. proposed Eq. (12) for the modulus of elas- ticity of geopolymer concrete EC¼0:037/C2q1:5/C2ffiffiffiffiffiffi fC0q ð12Þ The calculated and experimental results of the modulus of elas- ticity of FAGP and AAS concrete are shown in Fig. 12 . The results obtained from the ACI 318-14 , AS 3600-2009 and Diaz- Loya et al. overestimated the experimental results of normal strength and high strength FAGP and AAS concrete ( Fig. 12 ). Simi- lar observations were reported in the previous studies conducted on the comparison between calculated and experimental modulus of elasticity. Yost et al. reported that the modulus of elasticity of FAGP concrete was 11–16% less than the calculated modulus of elasticity using ACI 318-14 . Yang et al. found that modu- lus of elasticity of AAS concrete was 12–15% lower than the values calculated using ACI 318-14 . The calculated modulus of elas- ticity using ACI 318-14 and AS 3600-2009 for OPC con- crete did not provide a conservative estimate of normal and high strength FAGP and AAS concrete in terms of modulus of elasticity. However, the results obtained using Hardjito et al. was very close to those obtained from experimental results. Therefore, themodulus of elasticity for normal strength and high strength FAGP and AAS concrete can be reasonably estimated using the equation proposed by Hardjito et al. . 6. Conclusions This paper compares the engineering properties of normal strength and high strength FAGP and AAS concrete with OPC con- crete. The following conclusions are drawn from the test results. 1. The average dry density and ultrasonic pulse velocity of FAGP and AAS concrete were lower than those of OPC concrete. This finding was confirmed by SEM analyses. The SEM images showed that at 28 days, FAGP and AAS concrete were less dense and less compacted with less homogeneous microstructures compared to OPC concrete. 2. The normal strength FAGP, AAS and OPC concrete have compa- rable indirect tensile, flexural and direct tensile strengths. How- ever, the indirect tensile, flexural strength and direct tensile strength of high strength (compressive strength of about 65 MPa) FAGP and AAS concrete were higher than those of high strength OPC concrete. 3. The equations recommended in ACI 318-14 for OPC con- crete can be used for the conservative prediction of the indirect tensile strength of normal strength (compressive strength ofabout 35 MPa) FAGP and AAS concrete. However, the current ACI 318–14 for OPC concrete does not provide a conserva- tive estimate of the indirect tensile strength of high strength (compressive strength of about 65 MPa) FAGP and AAS con- crete. The equations recommended in ACI 318-14 and AS 3600-2009 can be used for conservative prediction of the flexural strength of normal strength concrete (compressive strength of about 35 MPa) FAGP and AAS concrete. However, the equations recommended in ACI 318-14 and AS 3600- 2009 does not provide a conservative estimate of the flex- ural strength of high strength (compressive strength of about 65 MPa) FAGP and AAS concrete. 4. The modulus of elasticity of normal strength and high strength FAGP and AAS concrete under uniaxial tension was about 7–8% and 8–9% less than the modulus of elasticity of OPC with thesimilar compressive strengths at 28 days. The modulus of elas- ticity of normal strength and high strength FAGP and AAS con- crete under compression was about 12–13% and 13–19% less than the modulus of elasticity of OPC with a similar compres- sive strength at 28 days. 5. The modulus of elasticity of normal strength and high strength FAGP and AAS concrete calculated using ACI 318-14 ,A S 3600-2009 and Diaz-Loya et al. was higher than the experimental modulus of elasticity. However, the modulus of elasticity of normal strength and high strength FAGP and AAS concrete can be closely estimated reasonably using equation recommended in Hardjito et al. . Conflict of interest No conflict of interest. Acknowledgements The authors gratefully acknowledge the contribution of the technical staff at the High Bay Laboratories of the University of Wollongong in carrying out the experimental work. The authors also express thanks to the Australian Slag Association for providing aluminosilicate materials. The first author also thanks to the Iraqi government and UOW for the award of PhD scholarship.

Construction and Building Materials 301 (2021) 124330 Available online 2 August 2021 0950-0618/© 2021 Elsevier Ltd. All rights reserved.Mechanical and durability properties of alkali-activated fly ash concrete with increasing slag content Timothy A Aikena,*, Jacek Kwasnya, Wei Shaa, Kien T Tongb 55 Giai Phong, Hanoi, Vietnam ARTICLE INFO Keywords: Durability Geopolymer Concrete Fly Ash Slag Freeze-Thaw Chloride Migration Steel Corrosion ABSTRACT Alkali-activated concrete is a promising alternative to conventional Portland cement-based concrete. However, further understanding in relation to the durability of alkali-activated concrete is required. This paper explores the resistance of alkali-activated concrete mixes to aggressive media that concrete is expected to encounter in- service. Findings indicate that the chloride, freeze –thaw and acid resistance of alkali-activated concrete in- creases as the slag content and activator increased. However, performance against chloride and freeze –thaw was not as good as that of Portland cement concrete. Acid resistance seems to be a more promising characteristic of alkali-activated concrete compared with Portland cement concrete. 1.Introduction The uptake of cementless alkali-activated concrete has been rela- tively slow, despite a significant amount of research being conducted in this area. Many factors have been cited for this including vested in- terests, supply chain concerns, lack of incentive, lack of knowledge/ understanding within industry as well as technical concerns [1,2] . At the forefront of the technical concerns is uncertainty associated with dura- bility in aggressive environments . This study aims to assess the durability of alkali-activated concrete against some of the most common aggressive media that concrete is expected to encounter in-service. Specifically, performance when exposed to chloride, freeze –thaw and acid attack. A comparison with a traditional Portland cement concrete is also carried out. The binder for alkali-activated concrete is produced from the acti- vation of a solid aluminosilicate source powder under alkaline condi - tions with a solid or dissolved alkali metal [2,3] . This is then combined with appropriate quantities of aggregate and water to form hardened concrete. Two of the most commonly used aluminosilicate sources are fly ash and ground granulated blast furnace slag, later referred to as slag [4,5] . A wide range of alkali metal activators have been investigated, the most widely used being alkali hydroxides and silicates (usually Na or K) [6,7] . The potential environmental benefits of alkali-activated concrete are huge due to no Portland cement being present in the system. Port- land cement production is associated with many negative environmental impacts including substantial CO2 emissions which contribute around 8% of global emissions annually [8–10]. The potential reduction in CO2 emissions for alkali-activated concrete is a subject of much debate with values between 44 and 80% quoted [11–14]. It is worth noting that the actual environmental benefit will vary depending on source materials, transportation requirements and activator type [15,16] . Therefore, the continued drive towards a reduction in CO2 emissions has made alkali- activated concrete an interesting proposition in order to reduce the carbon footprint of the construction industry. The alkali-activation of fly ash and slag individually as single precursors each have specific chal- lenges. Alkali-activated fly ash usually requires curing at elevated tem- peratures to develop strength which is a potential barrier to widespread use in the concrete industry . On the other hand, alkali-activated slag can be cured at ambient temperature but there are concerns relating to workability and rapid setting time [18–20]. The blending of fly ash and slag together has the potential to eliminate these issues. Alkali-activated fly ash/slag blends do not require elevated curing temperatures and are less prone to fast setting [21,22] . Therefore, this study investigates 100% fly ash mixes (cured at high temperature) and mixes with increasing slag content (cured at ambient temperature). The corrosion of steel reinforcement is one of the main problems affecting the durability and safety of reinforced concrete structures . It is most commonly caused by the penetration of chloride ions into the concrete microstructure from seawater or de-icing salts [24,25] . The chloride content at the steel surface eventually reaches a threshold value causing *Corresponding author. E-mail addresses: taiken02@qub.ac.uk (T.A. Aiken), j.kwasny@qub.ac.uk (J. Kwasny), w.sha@qub.ac.uk (W. Sha), kientt@nuce.edu.vn (K.T. Tong). Contents lists available at ScienceDirect Construction and Building Materials u{�~zkw! s{yo| kro>! ÐÐÐ1ow �o�to~1m{y2w {mk�o2m{zl� twnyk�! Received 15 February 2021; Received in revised form 28 June 2021; Accepted 20 July 2021 Construction and Building Materials 301 (2021) 124330 2depassivation of the steel, which allows corrosion to begin . Previ - ous studies investigating the chloride resistance of alkali-activated concretes have at times provided conflicting results. Kupwade-Patil and Allouche reported that alkali-activated fly ash concretes dis- played lower chloride diffusion coefficients than corresponding Portland cement samples when using the method described in ASTM C1556 . Ganesan et al. used the same procedure and found that alkali- activated fly ash concrete had almost equal performance to that of Portland cement concrete. Babaee and Castel reported similar resistance for alkali-activated fly ash concrete and Portland cement concrete following exposure to chloride environments and assessment of their electrochemical performance. Ismail et al. and Mackechnie and Scott investigated the chloride migration coefficient of alkali- activated fly ash/slag blends and Portland cement mixes according to NT Build 492 . They reported that the chloride migration coefficient was larger for Portland cement materials. Olivia et al. investigated the chloride ion penetration in alkali-activated fly ash concrete using the method described in NT Build 443 . They found that the alkali- activated fly ash mixes had a higher chloride ion penetration than Portland cement concrete. The majority of studies are focused on chlo- ride penetration or migration and results are variable. Therefore, anal- ysis of durability performance of alkali-activated concrete mixes particularly related to protection of embedded steel requires further research . In many countries, harsh winters with cold temperatures provide challenges for concrete durability . Concrete is susceptible to freeze –thaw attack when temperatures rise above and fall below the freezing point of water. The two most common modes of failure due to freeze –thaw are surface scaling and internal structural damage. Both failure modes are detrimental to the long-term durability of concrete. Internal structural damage reduces the concrete mechanical properties due to cracking which can also lead to further durability issues [38,39] . Surface scaling leaves a rough concrete surface which diminishes the depth of cover to steel reinforcement and is particularly common in the presence of de-icing salts which are often used to remove snow and ice from roadways and footpaths [40–42]. Many theories exist to explain in more detail the precise mechanisms involved. These include the hy- draulic pressure theory , the microscopic ice lens theory [44,45] , the glue-spall mechanism , the osmotic pressure theory and the critical degree of saturation . Skvara et al. [49,50] suggested that alkali-activated fly ash materials possess excellent freeze –thaw resis- tance. They carried out an investigation where the mass of the samples did not change during freezing and thawing cycles in the presence of NaCl solution (164 g/dm3) according to Czech standard CSN 72 2452 with no visible defects or deformation noticed after 150 cycles. How- ever, after 150 cycles the compressive strength dropped to 70% of the samples which were not exposed to freeze –thaw cycles. Sun and Wu reported on the compressive strength loss of alkali-activated fly ash and Portland cement concrete after exposure to distilled water and testing according to ASTM C666. They found that the fly ash samples retained greater than 90% of their strength after 300 cycles whereas the Portland cement samples retained approximately 80% during the same period. On the other hand, Montes et al. investigated the freeze –thaw resistance of alkali-activated fly ash concrete and reported poor resis- tance as the samples were unable to complete the test (ASTM C666 ) without suffering complete degradation. No comparison with Portland cement concrete was carried out. Ionescu and Ispas prepared concrete containing fly ash, slag and sodium silicate as activator. They reported good freeze –thaw resistance after exposure to 50 freeze –thaw cycles . Puertas et al. compared the resistance of Portland cement, alkali-activated fly ash, alkali-activated slag and a combination of 50% fly ash and slag mortars. They prepared beams (40 ×40 ×160 mm) for each mix, which were subjected to freeze –thaw cycles with water, not de-icing salt. They studied the strength change after 50 freeze –thaw cycles and reported good resistance (i.e. no strength loss) for the Portland cement, alkali-activated slag and slag/fly ash combination. On the other hand, the alkali-activated fly ash samples suffered a 24% strength loss. There is not a large volume of work re- ported on the freeze –thaw resistance of alkali-activated materials. In particular, the literature available on surface scaling with the use of de- icing salts is scarce. Furthermore, comparisons with Portland cement materials are not often available in literature . Acid attack is a topic of increasing significance driven by increasing urban and industrial activities . There is a broad range of acid media which come into contact with concrete structures including organic acids from agri-food industries and mineral acids from wastewater treatment and acid precipitation [58–61]. Acid precipitation is gener - ated from the incomplete combustion of fuels which are released into the environment as gases (NO 2 and SO2) and react with water to form acidic precipitation (nitric and sulfuric acid) [62–66]. Therefore, nitric acid is the focus of the present study. Tahri et al. compared the resistance of alkali-activated fly ash samples with Portland cement samples when exposed to different concentrations of nitric acid. At acid concentrations of 10 and 20%, the fly ash samples performed better in terms of mass loss. On the other hand, when a concentration of 30% was used, the Portland cement samples displayed smaller mass losses than 100% fly ash samples. Shi studied the performance of pastes made with alkali-activated slag and Portland cement when exposed to nitric acid. After 580 days of immersion in nitric acid, the alkali-activated slag and Portland cement pastes were corroded to a depth of approximately 1 and 2.5 mm, respectively. The majority of studies conducted on acid resis- tance suggest improved resistance to nitric acid for alkali-activated materials when compared with Portland cement materials. However, in general there are a wide range of variables considered and many authors only consider a single deterioration indicator which can provide misleading findings. Rather, a multiscale approach is necessary is better understand the performance of different binders due to acid attack . Therefore, further work is needed in relation to the nitric acid resistance of fly ash/slag blends whilst considering a range of deterioration indicators. This study investigates the physical and durability properties of alkali-activated fly ash concretes with increasing slag content. The en- gineering properties examined include compressive strength, tensile strength, flexural strength and modulus of elasticity. Durability was investigated by assessing chloride migration, steel reinforcement corrosion, freeze –thaw resistance and acid resistance. A comparison with a traditional Portland cement mix is also carried out. This study compares the engineering properties and performance in a range of aggressive exposure conditions of the same mixes. This provides insight into the properties that need further work to allow the use of alkali- activated concrete to become more widespread. 2.Experimental programme 2.1. Materials The fly ash, slag and Portland cement were the same powder binding materials as used in previous studies. Their oxide compositions and loss on ignition (LOI) obtained by X-ray florescence are shown in Table 1. Further details regarding their mineralogy and particle size distribution are available in previous publications [70–72]. The alkali-activated binders were activated by solutions of sodium silicate and sodium hy- droxide. The sodium silicate solution contained 25.5% SiO2, 12.8% Na2O and 61.7% water. The sodium hydroxide solution was prepared at 30% w/w by the dissolution of solid commercial grade sodium hy- droxide with a 99% purity. The aggregates utilised in this study were 0–5 mm lough sand, 4–10 mm crushed basalt and 10–20 mm crushed basalt. The Lough sand was supplied by Creagh Concrete, Northern Ireland and was considered as fine aggregate. The crushed basalts were obtained from James Boyd & Sons, Northern Ireland and were consid - ered as coarse aggregates. The oven dry particle densities and water absorption of each aggregate at 1 and 24 h were obtained according to T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 3BS 812–2:1995 and are reported in Table 2. Prior to mixing, the aggregates were oven dried at 100 •C for at least 24 h to remove moisture. They were allowed to cool to room temperature before mixing began. Laboratory reagent grade nitric acid (≽65% HNO 3) was used to prepare the 0.52 mol/L acid solution by mixing with distilled water. For the freeze –thaw tests a 3% w/w NaCl solution was prepared by mixing with distilled water. 2.2. Mix proportions The concrete mix proportions used in this study are presented in Table 3. Five alkali-activated concretes (F100Ac, F100Bc, F80c, F60c, F30c) and one Portland cement concrete (PCc) were studied. F100Bc was considered suitable for low strength applications with a design 28- day compressive strength less than 20 MPa. F100Ac and F80c were considered suitable for medium strength applications with a design strength between 35 and 50 MPa. The mixes with increased slag content (F60c and F30c) and PCc were considered suitable for applications where strength greater than 55 MPa was required. The alkali-activated concrete mixes are based on mix design optimisation carried out by Rafeet et al. [22,74 –76] and Vinai et al. and the Portland cement mix proportions were obtained from work carried out by Kwasny et al. [78–80]. Alkali dosage (M‡) and alkali modulus (AM) were the pa- rameters used to determine the quantities of each activating solution used in each mix. M‡is the ratio of the Na2O and the dry powder component of the binder, expressed as a percentage, whilst AM is the mass ratio of Na2O to SiO2 in the activating solutions. The absorption water listed in Table 3 represents the amount of water needed to bring the aggregates to saturated surface dry conditions. The total added water includes the absorption water but does not include the water within the activating solutions. The water to solids ratio was calculated by dividing mass of the mixing water (i.e. mass of water in the activating solutions, mass of total added water reduced by mass of absorption water) and mass of the solid portion of the binder (i.e. mass of the powders used in the binder and mass of solids in the activating solu- tions). F100A has increased alkaline activator dosages (M‡of 11.5) compared with each of the other alkali-activated mixes which have the same dosages (M‡of 7.5). Therefore, F100B, F80, F60 and F30 can be compared to study the effect of fly ash/slag content as they all have the same activator dosages. PC is used as a reference for comparison with each of the alkali-activated mixes. Acid resistance was studied on equivalent mortar (physical tests) and paste (microstructural tests) samples. The composition of the paste was identical for each paste mortar and concrete mixes. The paste content of all the concrete mixes was fixed at 32.5%. The mortar mixes were prepared with 50% volume of paste and 50% volume of sand. Concrete mixes are labelled ‘c’, mortar mixes are labelled ‘m’ and paste mixes are labelled ‘p’. 2.3. Procedure for mix preparation, samples manufacturing and curing Concrete specimens were produced using a Croker RP50XD rotating pan mixer. Mortar and paste specimens were produced with a Hobart rotating paddle mixer. The mix procedure summarised in Fig. 1 and is described below: (1) The aggregates were placed into the mixing drum (concretes) or bowl (mortars) along with approximately half of the total water. Aggregates and water were mixed together for 1 min and then were allowed to absorb moisture for a further 30 min (concretes) or 15 min (mortars). (2) The appropriate powder binder component of each mix was then added and mixed together with the aggregate for 1 min. (3) The remaining water and activating solutions (alkali-activated mixes only) were then added and mixing continued for 5 min before casting. The mixes were cast into moulds in two layers and compacted using a vibrating table. They were wrapped in polyethylene film to avoid rapid moisture loss and placed in the appropriate curing conditions. The samples were demoulded after 24 h and immediately returned to the appropriate curing environment (described below). These curing re- gimes were established as part of previous research on mix optimisation by the Geopolymer Team at Queen ’s University Belfast. Two main curing regimes were employed depending on binder type: ≡The 100% fly ash mixes (F100A and F100B) were oven cured at 70 •C for 7 days then placed in constant room conditions (20 ±2 •C and 55 ±5% humidity) until testing. ≡The fly ash/ slag mixes (F80, F60 and F30) and Portland cement mixes were cured at 20 ±2 •C and a relative humidity of greater than 90% until testing. These conditions were achieved by placing the samples inside sealed plastic boxes on plastic supports with a height of 15 mm. The boxes were filled with water to a depth of 5 mm in order to avoid contact between the samples and water which may result in leaching . 2.4. Testing procedures A summary of tested properties sample details and testing procedures used are provided in Table 4. The volume of permeable voids (VPV) was assessed according to the standard procedure given in ASTM C642 . The standard procedure recommends an oven drying temperature of between 100 and 110 •C. In this study, a temperature of 40 •C was employed to avoid excessive drying which can cause changes to binding phases within alkali- activated concrete . The compressive strength of concrete mixes was obtained from 100 mm cubes after 1, 7 and 28 days of curing. A constant load rate of 200 kN/min was applied and the average was calculated from three samples. Concrete splitting tensile strength at 7- and 28-day was tested ac- cording to BS EN 12390 –6:2009 . Three cylinders, 100 mm in diameter and 200 mm in height, were tested at each testing age. Concrete flexural strength was tested at the age of 28 days using three concrete beams (100 ×100 ×500 mm) following the two-point loading method specified in BS EN 12390 –5:2009 . The tensile strength was determined with the pull-off test, according Table 1 Oxide compositions (wt %) and loss on ignition (LOI) of fly ash, slag and Port- land cement. Fly ash Slag Portland cement SiO2 46.78 29.38 20.21 Al2O3 22.52 11.23 4.79 Fe2O3 9.15 0.36 2.78 CaO 2.24 43.72 63.01 MgO 1.33 6.94 1.93 MnO 0.05 0.51 0.08 TiO 2 1.05 0.67 0.27 Na2O 0.89 1.05 0.19 K2O 4.09 0.93 0.59 SO3 0.90 1.76 2.60 LOI 3.57 2.40 3.16 Table 2 Oven dry particle density and absorption of aggregates. Aggregate Oven dry particle density (kg/m3) Absorption (%) 1 h 24 h 0–5 mm sand 2690 0.9 1.1 4–10 mm basalt 2790 1.4 2.2 10–20 mm basalt 2751 1.2 2.1 T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 4to BS 1881 –207:1992 . For each concrete mix, three 230 ×230 × 100 mm slabs were cast. After 14 days of curing, partial coring was taken (to a depth of approximately 40 mm) at two locations at the mould- finished surface of each slab. Then, slabs were returned to their desig - nated curing conditions. At 21-day slabs were moved to a conditioning room and were stored there at 20 ±1 •C and RHof 55 ±10% in order to remove the surface moisture. One day before testing, four 20 mm thick and 50 mm in diameter metal discs were fixed to the mould-finished surface (two to uncored surface and two to the partially cored surface) using an epoxy resin adhesive. At 28-day the pull-off apparatus (called Limpet ) was attached to the disc and a load was applied at a steady rate of approximately 2.4kN/min, until the failure occurred (via the disc coming off the concrete sample). The tensile strength was calculated by dividing the failure load by the adhered surface area of the disc. An average of six measurements is reported as the pull-off surface tensile strength (in MPa) for discs attached to both uncored and partially cored surface. Modulus of elasticity was determined according to BS 1881 –121:1983 using three cylinders (100 mm in diameter and 200 mm in height). Fourteen days prior to testing, the cylinders ’ trowel- finished ends were capped with a thin layer of a fine mortar, in order to create two parallel test planes. On the subsequent day the cylinders were returned to the designated curing conditions. At the age of 28 days three concrete cylinders were tested. The average static secant modulus of elasticity is reported. To determine the drying shrinkage of concrete, the length change of specimens was measured following ASTM C490-08 . For each concrete mix, three prisms were cast. Between the 3rd and the 5th day after casting, a steel ball was attached to the centre of each end of the prisms, using geopolymer or cement paste. The length of the prisms was measured initially at 7-day after casting and subsequently at convenient dates, i.e. 8, 10, 14, 21, 28, 35, 70, 140, 210, 280 and 364-day after casting (±2% accuracy). After the initial reading was taken, the prisms were placed in a conditioning room, maintained at RH 55 ±5% and 20 ±1 •C, on racks ensuring exposure of all surfaces to this environment. The drying shrinkage for each measurement, expressed in microstrain, was calculated for the nominal gauge length of 250 mm. The change of length was reported as drying shrinkage (average of three measure - ments). In addition, at each measurement age, the masses of the prisms were recorded to calculate a relative mass change due to the loss of evaporable water. The chloride migration coefficient was determined by carrying out the chloride migration test based on the NT Build 492 test method. This test was carried out on disk samples (50 mm thick and 100 mm in diameter) which were cored from concrete slabs and then cut to size. The test involved subjecting three samples from each mix to an electrical potential to facilitate the transfer of chloride ions through the concrete. The concrete disc was positioned between two cells containing testing solutions, on one side a 10% NaCl catholyte solution and on the other a 0.3 M NaOH anolyte solution. Cathode (stainless steel) and anode (mild steel) perforated plates were positioned inside the cells on either side of the sample. A sample was pressed into a silicone sleeve (to ensure no leakage during the test) and the two cells were tightened together with bolts. Immediately after the cells were filled with adequate solutions, an Table 3 Mix proportions for alkali-activated and Portland cement concretes. Mix composition F100Ac F100Bc F80c F60c F30c PCc Binder composition Fly ash (%) 100 100 80 60 30 – Slag (%) 0 0 20 40 70 – Portland cement (%) – – – – – 100 Paste content (%) 32.5 32.5 32.5 32.5 32.5 32.5 Water/solid ratio 0.37 0.37 0.38 0.40 0.42 – Water/cement ratio – – – – – 0.38 Alkali dosage (M‡) 11.5 7.5 7.5 7.5 7.5 – Alkali modulus (AM) 0.95 1.25 1.25 1.25 1.25 – Fly ash (kg/m3) 349 375 301 223 111 – Slag (kg/m3) – – 75 149 260 – Portland cement (kg/m3) – – – – – 468 Sodium silicate solution (kg/m3) 166 88 88 88 87 – Sodium hydroxide solution (kg/m3) 81 72 73 72 72 – Absorption water (kg/m3) 21 21 21 21 21 21 Total added water (kg/m3) 23 74 79 87 95 197 Sand (kg/m3) 728 728 728 728 728 739 Basalt 4–10 (kg/m3) 454 454 454 454 454 443 Basalt 10–20 (kg/m3) 666 666 666 666 666 666 1. Aggregates and approximately half of the total water were placed in the mixing drum (concrete) or bowl (mortar)2. The appropriate powder binder component of each mix was placed in the mixing drum (concrete) or bowl (mortar or paste)3. The remaining water and activating solutions (alkali- activated mixes only) were then added to the mixing drum (concrete) or bowl (mortar or paste) Mixed for 1 minuteMixer stopped for 30 minutes to allow aggregates to absorb moistureMixed for 1 minuteMixed for 5 minutes Paste mixing sequence Concrete and mortar mixing sequence Fig. 1.Mix procedure adopted during batching of concrete, mortar and paste mixes. T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 5initial electrical potential of 30 V was applied between the electrodes. Initial current was measured and, based on recommendations given in the standard, it was used to estimate electrical potential and duration of the test. When the test was completed (between 6 and 96 h depending on test conditions), the samples were removed from the apparatus and split. Silver nitrate solution (0.1 M AgNO 3) was then applied to the freshly split sample to obtain the chloride penetration depth. The presence of chloride is identified by a white/silver precipitate formed due to a chemical reaction between chloride and silver nitrate. The chloride penetration depths were more difficult to identify for the alkali- activated concretes than for the Portland cement concrete. A total of 7 readings were collected at 10 mm intervals from 20 to 80 mm along the 100 mm length of each split specimen. Additional information was also collected before and after the test procedure, including sample di- mensions, the anolyte and catholyte temperature. These were used alongside the average chloride penetration depth to calculate the chlo- ride migration coefficient according to Eq. (1). Dnssmˆ0B0239…273‡T†L …U2†t⌊ xd0B0238⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪⎪ …273‡T†Lxd U2̅ ⌋ (1) where: Dnssm ˆnon-steady-state migration coefficient (×1012 m2/s) T ˆaverage value of the initial and final temperatures in anolyte solution (•C) L ˆspecimen thickness (mm) U ˆapplied voltage (V) t ˆtest duration (hour) xd ˆaverage value of penetration depths (mm) The electrical resistivity was obtained for each concrete mix before the chloride migration test (28 days) according to the two probe method. The sample was placed between two conductive plates and moist sponges were used as contact media. A current was applied between the two plates and the potential was measured. The equipment used was a Tinsley prism instrument 6451 LCR Databridge. The corrosion of steel reinforcement was investigated by casting steel bars inside concrete slabs and ponding the slabs with NaCl solution. Six slabs were cast for each mix. The slabs were cast with a 15 mm rim around the outside of the ponding surface to contain the NaCl solution as shown in Fig. 2. Three 16 mm diameter carbon steel bars were cast in- side each slab at depths of 15, 25 and 40 mm from the ponding surface. Table 4 Summary of tested properties and sample details. Property type Tested property Sample details Testing procedure Physical VPV 100 mm long and 75 mm diameter cylinders taken from 100 mm concrete cubes (six samples for each mix) Test procedure according to ASTM C642 . Three samples were tested at the age of 28 and 180 days. Compressive strength 100 mm concrete cubes (nine samples for each mix) Three samples were tested at the age of 1, 7 and 28 days Tensile strength 200 mm long and 100 mm diameter concrete cylinders (three samples for each mix) Test procedure according to BS EN 12390 –6:2009 . Three samples were tested at the age of 7 and 28 days. Flexural strength 100 ×100 ×500 mm concrete beams (three samples for each mix) Test procedure according to BS EN 12390 –5:2009 . Three samples were tested at the age of 28 days. Tensile pull off strength 230 ×230 ×100 mm concrete slabs (three slabs for each mix) Test procedure according to BS 1881 –207:1992 . At the age of 28 days two cored and two uncored measurements were taken for each slab. Modulus of elasticity 200 mm long and 100 mm diameter cylinders (three samples for each mix) Test procedure according to BS 1881 –121:1983 . Three samples were tested at the age of 28 days. Drying shrinkage 250 ×75 ×75 mm concrete prisms (three samples for each mix) Test procedure according to ASTM C490-08 . Initial measurement taken at 7 days after casting and subsequently measurements up to the age of one year. Chloride migration Electrical resistivity 50 mm long and 100 mm diameter discs taken from concrete slabs (three samples for each mix) Three samples were tested at the age of 28 days Chloride migration coefficient 50 mm long and 100 mm diameter discs taken from concrete slabs (three samples for each mix) Test procedure according to NT Build 492 . Three samples were tested at the age of 28 days. Corrosion of steel reinforcement Mass loss of steel bars Steel bars embedded in concrete slabs at depths of 15, 25 and 40 mm (six slabs for each mix) Slabs were weekly ponded with a 3% NaCl solution for a duration of one year. Bars were extracted for mass loss measurements from a set of three slabs after 6 months of ponding and then at one year from the remaining three slabs. Chloride penetration depth Concrete slabs used for ponding test AgNO 3 solution sprayed on cross-section of concrete slabs following ponding test XRD Powdered samples obtained from below the ponding surface of concrete slabs Samples collected following 6 months of ponding Freeze-thaw resistance Visual appearance 100 mm long and 100 mm diameter cores taken from concrete slabs Three samples were assessed after every 7 freeze –thaw cycles up to 56 cycles Scaled mass 100 mm long and 100 mm diameter cores taken from concrete slabs (three samples for each mix) Three samples were assessed after every 7 freeze –thaw cycles up to 56 cycles Acid resistance Strength loss 50 mm mortar cubes Three samples were assessed after 8 weeks exposure to a 0.52 mol/L solution of nitric acid Mass loss 50 mm mortar cubes Four samples were monitored each week for 8 weeks during exposure to nitric acid XRD and SEM 50 mm paste cubes For XRD, samples were collected from the edges of paste samples exposed to acid. For SEM, samples perpendicular to the exposed face were obtained and polished prior to analysis. Pastes samples were exposed to a 0.52 mol/L solution of nitric acid for 3 weeks. Fig. 2.Photograph of concrete slab for ponding test. T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 6Three slabs were cast for each mix and the position of the bars inside the slabs were altered for each slab i.e. the bar cast at a depth of 15 mm was on the left, middle and right in one of each of the three slabs. The layout of the steel bars is shown in Fig. 3. Prior to placement inside the concrete and casting of concrete, the steel bars were cleaned with a steel wire brush and the mass of each bar was recorded. The slabs were stored at room conditions (20 ±2 •C and 55 ±5% humidity) and ponded with 3% NaCl solution every 7 days. After 24 h, the remaining ponded solution was removed with a sponge. Therefore, the samples experienced wetting (1 day) and drying (6 days) cycles. This procedure was continued for a duration of 1 year (52 ponding cycles). After 6 months (26 cycles) and 1 year (52 cycles) the steel bars were extracted from three slabs and inspected visually for signs of corrosion. Subsequently, the steel bars were cleaned with a steel wire brush and their mass was recorded. The results were reported as percentage mass losses of steel bars from the initial mass after 26 and 52 ponding cycles. The average depth of chloride penetration into each concrete slab was measured by applying AgNO 3 solution onto the split surface and measuring the depth of penetration from the ponding surface. A portion of each concrete slab from just below the ponding surface was taken for X-ray diffraction (XRD) analysis to determine if any chloride binding salts were formed. The freeze –thaw resistance of the concrete mixes was determined by performing a scaling test similar to DD CEN/TS 12390 –9:2006 . The test was carried out on 100 mm thick cores with a diameter of 100 mm obtained from concrete slabs and the test setup is shown in Fig. 4. The lateral surface of the cores were coated with epoxy resin to ensure scaling only took place on the test surface. A rubber seal was also placed on the top surface to avoid any moisture build up and deterioration. Three samples were tested for each mix and two different freeze –thaw mediums were used (distilled water and 3% NaCl solution). After 21 days of curing the samples were placed on 10 mm supports inside the polyethylene containers along with the appropriate freeze –thaw me- dium at 20 ±2 •C. The container was filled to a depth of 15 mm with the appropriate freeze –thaw medium. This allowed saturation of the sam- ples with the freeze –thaw medium before freeze –thaw cycles began. After another 7 days, the freeze –thaw medium was replaced and the 24 h freeze –thaw cycles were started according to the limits displayed in Fig. 5. After 7 cycles the scaled material was collected by filtration, oven dried and its mass recorded. The freeze –thaw medium was replaced and the samples were returned to the freeze –thaw chamber immediately. This process was repeated after every 7 cycles until 56 cycles or until the samples had lost more than 25% of their original mass by scaling. Acid resistance was tested based on the general guidelines provided in ASTM C267 and exactly the same procedure as previous studies [70–72]. To summarise, acid resistance was studied by immersing four 50 mm mortar cubes into a 0.52 mol/L solution of nitric acid for 8 weeks. The mass was recorded and the acid solution replenished each week. After 8 weeks of acid exposure, the compressive strength of three of the exposed samples was obtained for comparison with control samples stored in water. The fourth cube was used to assess the apparent depth of acid penetration. It was split and the freshly broken surface was sprayed with phenolphthalein solution to highlight the pH distribution. Fig. 3.Layout of steel reinforcement embedded inside three concrete slabs. Container Epoxy CoatingRubbe r seal Sample Freeze-thaw medium Test surfaceLidFig. 4.Freeze-thaw experimental setup. -25-20-15-10-50510152025 0 2 4 6 8 10 12 14 16 18 20 22 24Temperat ure (°C) Time (hours) Fig. 5.Upper and lower limit of temperature at the test surface during freeze –thaw cycles . T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 7XRD and SEM were used to assess the effect of nitric acid exposure on 50 mm paste cubes exposed to the sample cyclic exposure regime for 3 weeks. 3.Results 3.1. Physical properties Fig. 6 shows the VPV for each concrete mix after 28 and 180 days of curing. The error bars represent the standard deviation calculated from three specimens. The largest VPV values were observed for the 100% fly ash mixes which decreased as the slag content increased. This suggests a less porous microstructure as the slag content increased. This is due to the formation of space filling C-A-S-H and C-N-A-S-H gels compared with the more porous N-A-S-H gels formed in 100% fly ash mixes [22,92 –94]. The Portland cement concrete had a similar VPV as that of the blend of 30% fly ash and 70% slag. After 180 days the VPV value decreases for each mix. This reduction is more pronounced for the mixes with increased slag content and is likely due to continual development of their microstructure, whereas limited further reactions take place in the 100% fly ash mixes after curing was complete. Fig. 7 shows the compressive strength of each concrete mix after 1, 7 and 28 days of curing. The error bars represent the standard deviation calculated from three specimens. The 100% fly ash mixes attain more than 50% of their 28 day strength within the first 24 h. This is likely due to the high temperature curing employed. F100Ac had a 28 day strength of 48.5 MPa compared with 16.0 MPa for F100Bc which was due to the significantly higher alkaline activator content used in F100Ac. The 100% fly ash mixes displayed their maximum compressive strength after 7 days of curing. It could be related to their removal from high tem- perature curing after 7 days meaning no further reactions took place beyond 7 days. The slight reduction in compressive strength could be due internal stresses as a result of thermal shock or moisture re-entering the pore structure which had been removed during curing [95–97]. As the content of slag increased the compressive strength also improved to 64.0 MPa after 28 days for F30c. The compressive strength of PCc was 62.5 MPa after 28 days which was similar to that of F60c and F30c. Table 5 shows the tensile strength, flexural strength, pull off strength and modulus of elasticity of each concrete mix. The standard deviation calculated from three specimens is shown in brackets. The trends observed are similar to those observed for compressive strength. The strength values increased as the slag content increased. Interestingly, the 100% fly ash mix with the increased activator dosages performed very well in the tensile and pull off tests. This suggests the increased activator dosages provided an increased bond of aggregate and binder and could be due to the greater viscosity of the mixture. Fig. 8 shows the drying shrinkage or length change of each concrete mix over a one year period. The error bars represent the standard de- viation calculated from three specimens. The 100% fly ash mixes un- derwent expansion rather than shrinkage. The expansion was likely due to the high temperature curing causing moisture to be removed during the curing period. As moisture re-entered the pore structure, expansion F100AcF100Bc F80c F60c F30c PCcVPV (%)28 days 180 daysFig. 6.VPV for each concrete mix after 28 and 180 days of curing. F100AcF100Bc F80c F60c F30c PCcCompressi ve strengt h (MPa)1 day 7 days 28 daysFig. 7.Compressive strength of each concrete mix after 1, 7 and 28 days of curing. Table 5 Tensile strength, flexural strength, pull off strength and modulus of elasticity of each mix. Mix Tensile strength (MPa) Flexural strength (MPa) Pull off strength (MPa) Modulus of elasticity (GPa) days 28 days 28 days 28 days - surface 28 days - partially cored 28 days F100Ac 3.7 3.0 3.5 4.7 3.3 21.1 (0.6) (0.1) (0.3) (0.6) (0.3) (0.4) F100Bc 1.2 1.0 1.2 2.0 1.0 8.2 (0.0) (0.1) (0.1) (0.2) (0.2) (0.3) F80c 1.8 2.8 2.4 2.2 2.0 19.7 (0.3) (0.2) (0.1) (0.2) (0.3) (0.3) F60c 2.8 3.8 3.37 2.9 2.2 27.1 (0.7) (1.0) (0.1) (0.3) (0.2) (1.0) F30c 3.5 3.3 3.4 2.8 2.5 28.9 (0.4) (0.6) (0.2) (0.4) (0.3) (0.7) PCc 3.2 3.7 4.5 3.9 2.5 30.3 (0.1) (0.4) (0.5) (0.5) (0.4) (1.1) -2100-1750-1400-1050-700-35003507002870112154196238280322364Length change [micr ostrain] Testing age [days] F100Ac F100Bc F80c F60c F30c PCc shrinkageexpansion Fig. 8.Drying shrinkage or length change of each concrete mix over one year period. T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 8took place. It generally stabilised by 50 days for both mixes with less pronounced expansion observed for the mix with increased activator dosages. This could be due to a combination of its increased mechanical properties and decreased VPV. The shrinkage of the blended fly ash/slag concrete mixes decreased as the content of slag increased. This may be related to the increased mechanical properties. The drying shrinkage for PCc was similar to that of the alkali-activated mix with 30% slag and 70% fly ash (F30c). 3.2. Durability 3.2.1. Resistance to chloride migration Electrical resistivity is an indirect measurement of the transport properties of concrete and as corrosion is an electrochemical phenom - enon it has a bearing on the corrosion rate of steel [98,99] . Table 6 shows the electrical resistivity and chloride migration coefficient for each mix. The standard deviation obtained from three specimens is shown in brackets. The resistance of each concrete mix to chloride migration was investigated using the accelerated NT Build 492 procedure . The cross-section of one sample from each mix at the end of the chloride migration test is shown in Fig. 9. In the case of the 100% fly ash mixes, the depth of the entire cross-section (50 mm) appeared silver in colour suggesting the entire section was penetrated by NaCl. Therefore, it was not possible to calculate the chloride migration coefficient for these mixes as the exact penetration depth was unknown. However, we can conclude that the resistance of the 100% fly ash mixes to chloride penetration is much lower than each of the other mixes studied. If a penetration depth of 50 mm is assumed for F100Ac and F100Bc, the chloride migration coefficient would be 78.3 and 83.4 ×1012 m2/s, respectively. These values are significantly larger than those observed for each of the other mixes. The NT Build 492 test procedure is designed with the aim of ensuring that the front of chloride penetration is between 0 and 50 mm at the end of the test so that a penetration depth can be measured. This may not have worked for the fly ash concretes as the test was originally designed for the testing of Portland cement-based mate- rials. Furthermore, the VPV of the fly ash concretes was larger than any of the other mixes (Fig. 6). Zhu et al. also reported that alkali- activated mixes with high fly ash content were vulnerable to chloride penetration. As the content of slag increased the chloride migration coefficient decreased to 46.3, 7.1 and 3.9 ×1012 m2/s for F80c, F60c and F30c, respectively (Table 6). The chloride migration coefficient for the PC concrete was 11.1 ×1012 m2/s suggesting it has much better resistance to chloride penetration than F100Ac, F100Bc and F80c but slightly lower resistance to chloride penetration than F60c and F30c. The elec- trical resistivity results indicate a similar trend with a higher resistivity obtained for the mixes with increased slag content. These findings show good correlation with the VPV for each mix. The VPV decreased with increased slag content. Similarly, the resistivity increased, and chloride migration coefficient decreased as the slag content increased. This in- dicates that mixes with increased slag content have a greater resistance to chloride penetration due to their reduced volume of voids and increased resistivity. This is likely due to the more dense calcium aluminium silicate hydrate gel present in mixes with increased slag content compared with the more porous sodium aluminium silicate hydrate gel present in fly ash dominant mixes [22,70,93] . 3.2.2. Corrosion of steel reinforcement Fig. 10 shows photographs of the slab ponding surface of each con- crete mix after 6 months and 1 year of ponding with NaCl solution. The 100% fly ash mixes appear to have suffered some damage to their sur- face via scaling of paste and exposure of aggregates. A white coloured deposit on the surface is also visible and likely to be recrystallization of NaCl. The exposed surface of the blends of fly ash and slag concrete appear unaffected by ponding of NaCl solution after 6 months of ponding. After 1 year of ponding cracks have emerged on the ponding surface of F60c and F30c at the location of the steel reinforcement with 15 mm cover. The Portland cement concrete appears relatively unaf- fected by the ponding cycles. Following 6 months of NaCl ponding, the slabs were split and the depth of chloride penetration was measured by applying AgNO 3 solution to the cross-section of each mix. Fig. 11 shows a photograph of the cross- section of each mix following the application of AgNO 3. The chloride ingress starts form the top of each sample and the front of chloride penetration is marked by a dashed red line. The average measured depth of chloride penetration is also shown. The chloride solution penetrated through the entire cross-section of F100Ac and F100Bc. This is similar to the chloride migration test where these samples were also fully pene- trated. This is likely due to the large VPV (Fig. 6) and a well-connected pore network allowing rapid ingress of chloride ions. As the content of slag rose, the average chloride penetration depth decreased to 65, 20 and 9 mm for F80c, F60c and F30c, respectively. The average chloride penetration depth for PCc was 5 mm, similar to that of F30c. This is Table 6 Electrical resistivity and chloride migration coefficient of each concrete mix after 28 days. Mix Resistivity – 28 days (Ωm) Chloride migration coefficient (Dnssm) (×1012 m2/s) F100Ac 18.0 (0.7) F78.3 (1.3) F100Bc 13.3 (1.0) F83.4 (2.5) F80c 17.0 (0.1) 46.3 (4.2) F60c 69.0 (5.9) 7.1 (0.1) F30c 171.9 (6.1) 3.9 (0.3) PCc 85.7 (2.1) 11.1 (0.8) Increasing slag contentF100A c F80c F30c PCcF100B c F60c fly ash/slag blends 100% fly ash PCFig. 9.Chloride penetration depth in each concrete mix (arrow showing di- rection of chloride migration). T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 9similar to the VPV results which also decreased as the slag content increased and PCc also had a similar value as that of F30c (Fig. 5). This suggests that the chloride penetration depth is related to the VPV and pore network. It is worth noting that the chloride penetration depth measured is the average from a number of points. Therefore, the maximum chloride penetration depth would be larger . The steel bars were removed from the slabs and inspected following 6 months and 1 year of ponding with NaCl solution. They are shown in Fig. 12. The steel bars from F30c and PCc had very little visible corro - sion. Conversely, the steel bars from the other mixes had areas of brown/ orange coloured corrosion. This is known as pitting corrosion and is caused by a breakdown of the protective oxide film on steel reinforce - ment allowing chloride ions to attack . Some of the bars also had grey coloured paste attached. The paste and corrosion product were removed carefully using a steel wire brush before measuring the mass of each bar. This cleaning process was also applied to the steel bars before placement inside the concrete to ensure the cleaning procedure caused no additional mass loss . Fig. 13 shows the mass loss of steel bars embedded in each concrete mix at depths of 15, 25 and 40 mm after 6 months and 1 year of NaCl ponding on the concrete surface. The mass losses shown are averaged from three bars embedded in three different slabs at each depth. It is worth noting that the mass losses observed are relatively small (less than 0.6%) after 6 months. Angst et al. suggested that ideally a signifi - cant amount of corrosion should have taken place in such tests if the mass difference is to be detected. After 1 year a significant increase in the mass loss of the steel bars was observed. For each mix the mass loss decreases as the depth of concrete cover increases except for PCc as no mass loss was observed at any depth. The other exception is F100Bc after 1 year which is perhaps because the slab was fully penetrated at all depths for a significant period of time allowing as much or more corrosion to occur at depths of 25 and 40 mm. As the slag content increased the mass loss of the steel bars generally decreased. Particularly for the steel bars with 25 and 40 mm of concrete cover. On the other hand, after 1 year the bars in the 100% fly ash mixes were all signifi - cantly corroded regardless of the depth of concrete cover. The corrosion of the steel bars values are in agreement with the chloride penetration depths shown in Fig. 11. The mixes which result in the largest steel bar mass losses (F100Ac, F100Bc and F80c) at depths of 25 and 40 mm allowed chloride to penetrate the furthest after 6 months (Fig. 11). On the other hand, the steel bars embedded in F60c and F30c at 25 and 40 mm had much smaller mass losses. PCc had superior performance than any of the alkali-activated concrete mixes. The steel bars suffered no mass losses even with cover of only 15 mm. According to several authors [103 –105], the use of fly ash and slag as supplementary materials significantly improves the chloride binding capacity of Portland cement based concretes. This is due to the increased aluminium content in both slag and fly ash, particularly fly ash. How- ever, in Portland cement materials there is ample supply of calcium which is also needed to form chloride binding salts such as Friedel ’s salt. There is less calcium available in the alkali-activated concrete mixes. Fig. 14 shows the XRD patterns of each concrete following the ponding test. The samples were collected from directly below the ponding Increasing slag content fly ash/slag blends 100% fly ash PC1 year 6 months PCcF100A c F80c F30cF100B c F60cFig. 10.Photographs of the ponding surface of each mix after 6 months and 1 year of ponding with NaCl solution. Fig. 11.Photograph of the cross-section of each mix following the application of AgNO 3. T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 10surface. The majority of the phases present were due to aggregate par- ticles. They are clinochlore, muscovite, albite and quartz. Ettringite and calcium hydroxide were also identified in PCc but the peaks are small because the patterns are dominated by the peaks from aggregate parti- cles. The only phases observed that can be related to chloride penetra - tion are Friedel ’s salt in PCc and halite (NaCl) in F100Bc. The binding of chlorides due to the formation of Friedel ’s salt likely slows down the rate of chloride penetration and explains why PCc performed much better in the steel corrosion ponding test, than the chloride migration test. The chloride migration test described in the NT Built 492 is accelerated over 24 h so the ability of the concrete to bind chlorides would be largely negated. On the other hand, during the ponding test the binding of chlorides by Friedel ’s salt in PCc would have meant less chloride was 6 months 1 yearIncreasing slag contentF100A c F80c F30c PCcF100B c F60c fly ash/slag blends 100% fly ash PC15 mm 25 mm 40 mm15 mm 25 mm 40 mm 15 mm 25 mm 40 mm 15 mm 25 mm 40 mm 15 mm 25 mm 40 mmFig. 12.Steel bars extracted from concrete slabs (before cleaning) following 6 months and 1 year of ponding with NaCl solution. 0.00.20.40.60.81.01.21.41.61.8 F100Ac F100Bc F80c F60c F30c PCcMass loss (%)(b)0.00.20.40.60.81.01.21.41.61.8 F100Ac F100Bc F80c F60c F30c PCcMass loss (%)15 mm 25 mm 40 mm(a)Fig. 13.Mass loss of steel bars embedded in each concrete mix after a) 6 months and b) 1 year of NaCl ponding on the concrete surface. CL-clinochloreM-muscovite AB-albite Q-quartz E-ettringite CH-calcium hydroxide F-Friedel’s salt H-halite 5 10 15 20 25 30 35 40 45 50 55 60 65 2θ(degrees)F100A c ~ ~ ~ ~ ~ PCcF30cF60cF80cF100B cCLCLMAB ABQQ QQQQ Q Q EH CHF~ Fig. 14.XRD patterns of concrete mixes following NaCl ponding test. T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 11available to penetrate the concrete and initiate corrosion. In F100Bc, halite was observed which is a NaCl mineral. This suggests that sodium chloride recrystallized in the pores of the sample. This could be due to suitable sodium concentrations available in the pore solution. Halite was also observed by Ismail et al. in 100% fly ash mixes following chloride ponding. Monticelli et al. also reported that chloride binding in fly ash mixes took place via sodium chloride salts instead of Friedel ’s salt. There were no chloride binding minerals found in the fly ash/slag mixes (F80c, F60c and F30c) suggesting their ability to protect steel reinforcement is purely dependent on their porosity and pore connectivity. 3.2.3. Freeze-thaw resistance Fig. 15 shows the visual appearance of each concrete mix following exposure to 56 freeze –thaw cycles in the presence of water and 7 freeze –thaw cycles in the presence of 3% NaCl solution. The presence of NaCl resulted in a higher level of deterioration for each mix than water. All of the alkali-activated concrete mixes have experienced the loss of cement paste and the exposure of aggregates. Most significantly damaged are the mixes with 100% fly ash. The damage to the 100% fly ash mixes, particularly FA100Bc, was so severe that the test surface was very uneven following exposure cycles. As the content of slag increased the damage to the surface decreased significantly and the test surface remained relatively flat so that it was in contact with the freeze –thaw medium. However, the damage was still much greater for the alkali- activated concretes compared with the damage observed for PCc. Fig. 16 shows the cumulative scaled mass percentage of concrete mixes during exposure to 56 freeze –thaw cycles in the presence of water and NaCl. The error bars represent the standard deviation calculated from three specimens. In the presence of water F100Bc was significantly damaged after 21 cycles with a scaled mass of 28.2% observed. Each of the other mixes remained intact for the entire 56 cycles in the presence of water. F100Ac had a final scaled mass of 3.2% followed by F80c, F60c and F30c which had final scaled mass values of 0.6, 0.3 and 0.1% respectively. PCc displayed good resistance as no scaled mass loss was observed after 56 freeze –thaw cycles. The presence of NaCl resulted in larger scaled mass values for each mix compared with water. This has been reported for many years in the case of Portland cement concrete and also applies to the alkali-activated concretes investigated in this study. F100Bc suffered significant damage after 14 cycles, while F100Ac and F80c had scaled mass values of greater than 25% after 21 cycles so testing was stopped. F60c, F30c and PCc had scaled mass values of 7.3, 2.6 and 0.2%, respectively, after 56 cycles. Therefore, resistance to freeze –thaw was significantly increased as the slag content increased. This is likely related to the decrease in VPV (Fig. 6) and in- crease in tensile strength (Table 5) as the slag content increased. It is worth noting that air entrainment was not used in this study. The effectiveness of air entrainment to improve the freeze –thaw resistance of alkali-activated concrete should be considered in future investigations. 3.2.4. Acid resistance Fig. 17 shows the strength loss for each mortar mix following NaCl F100Bc – completely destroyedIncreasing slag contentF100A c F80c F30c PCcF100B c F60c fly ash/slag blends 100% fly ash PCF80c F60c F30c PCcWater F100A cFig. 15.Visual appearance of each concrete mix following exposure to 56 freeze –thaw cycles (water) and 7 freeze –thaw cycles (NaCl). 00.511.522.533.5 0 7 14 21 28 35 42 49 56Cumulative scaled mass (%) Freez e-thaw cycles(a)0 7 14 21 28 35 42 49 56Cumulative scaled mass (%) Freez e-thaw cyclesF100Ac F100Bc F80c F60c F30c PCc(b)Fig. 16.Cumulative scaled mass (%) of mixes during freeze –thaw cycles in the presence of a) water and b) 3% NaCl solution. T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 12exposure to a 0.52 mol/L solution of nitric acid for 8 weeks. The compressive strength of unexposed samples is also shown for compari - son and the percentage strength loss is also noted above each bar. The error bars represent the standard deviation calculated from three spec- imens. The mixes with 100% fly ash (F100Am and F100Bm) and 80% fly ash (F80m) suffered the largest percentage compressive strength losses of approximately 60%. As the slag content increased the percentage strength loss decreased to 37% for F30m. A similar percentage strength loss of 41% was observed for PCm. Fig. 18 shows a photograph of the cross-section of each mortar mix following exposure to a 0.52 mol/L solution of nitric acid for 8 weeks. Phenolphthalein solution has been applied to the cross-section of each sample to obtain an indication of the pH throughout the cross-section of each sample and possible depth of acid penetration. Unexposed samples are also shown for comparison. The mixes with the largest VPV (Fig. 5) appear to have suffered the largest pH change throughout their cross- section (F100Am, F100Bm and F80m). These mixes also suffered the largest strength losses which is likely caused by their larger VPV permitting a greater depth of acid penetration. Fig. 19 shows the mass change for each mortar mix during exposure to a 0.52 mol/L solution of nitric acid for 8 weeks. The error bars represent the standard deviation calculated from four specimens. Each mix was found to have undergone a decrease in mass due to nitric acid exposure. Throughout the 100% fly ash mixes displayed lower mass loss than each of their counterparts. Additionally, the mass losses observed were similar regardless of the alkali activator content used. Initially the mixes containing both fly ash and slag lost more mass than the PCm. However, the mass losses for the mixes containing both fly ash and slag appeared to decelerate whereas the mass loss for the PCm appeared to accelerate. As a result, PCm had the largest mass loss of 9.3% after 8 weeks exposure. The mixes containing fly ash and slag had mass losses between 6.3 and 7.4%. The mass loss decreased as the slag content increased. The 100% fly ash mixes had mass losses of only 2.6% after 8 weeks exposure. Fig. 20 shows photographs of each mortar mix after exposure to a 0.52 mol/L solution of nitric acid for 8 weeks. Unexposed samples are shown for comparison. Visual examination highlights that the surface of the 100% fly ash mixes is relatively unchanged following nitric acid exposure. On the other hand, the surface of the mixes con- taining slag and PCm appears damaged with the loss of paste and exposure of sand particles. Equivalent paste mixes were also investigated and exposed to 0.52 mol/L of nitric acid for 3 weeks. Fig. 21 shows the XRD patterns of F100Ap, F30p and PCp following exposure to nitric acid. Unexposed samples are also shown for com- parison. In F100Ap, quartz, mullite and hematite were identified due to unreacted fly ash particles . Following exposure to nitric acid, very little change was observed; quartz, mullite and hematite remained unaffected. The only change being that the broad hump between 15 and 35•2θ attributed to silicate and aluminosilicate gels shifted to be cen- tred at a lower angle. Centring of the peak in this region is typically observed for unreacted fly ash and may suggest that some of the reacted aluminosilicate phases have broken down leaving behind unreacted fly ash particles [108 –110]. This along with the suggested significant acid penetration would explain the larger strength losses observed for the 100% fly ash mixes. In F30p the main peaks identified in the unexposed sample are quartz, mullite and hematite. The intensities are much lower than for F100Ap due to only 30% fly ash being present in F30p. A broad peak centred at approximately 29•2θ was also observed and has been attributed to poorly crystalline calcium silicate hydrate type (C-S-H) gels [111 –113]. Following nitric acid exposure this peak was no longer present and a broad hump between 15 and 35•2θ was observed. This is likely due to the decalcification of C-S-H type gel leaving behind silicate and aluminosilicate gels as well as unreacted fly ash particles which are identified by this broad hump [108 –110]. In the unexposed PCp sample ettringite and calcium hydroxide were identified. These were not found after exposure to acid, with calcite being identified instead. This sug- gests the dissolution of ettringite and calcium hydroxide by nitric acid and explains the significant mass losses observed (Fig. 18) Fig. 22 shows SEM images of the outer edge (approximately 2.5 mm) of F100Ap, F30p and PCp following 3 weeks exposure to nitric acid. F100Ap shows little damage to the outer surface, this is in agreement with the mass loss results which showed very little change for the 100% fly ash mixes. The surface of F30p appears to have suffered damage to the edge and the loss of paste which agrees with the mass losses observed (Fig. 19). However, the region inside the outer edge appears unaffected by acid exposure. This may be due to the more dense microstructure observed and may explain the greater strength retention (Fig. 17) following acid exposure. PCp also seems to have suffered damage to the edge and cement paste appears close to falling off. PCp also had region which appears depleted and more porous than the undamaged area. 4.Discussion The paste content of each corresponding concrete, mortar and paste mix used in this study was identical which allowed comparison between performance in different environments. This provided insight into the performance of alkali-activated binders towards more widespread use as a replacement for traditional Portland cement-based binders. The 100% fly ash concretes (F100Ac and F100Bc) displayed poor resistance to chloride migration as the entire cross-section of the samples became penetrated with chloride. Similarly, the slabs with embedded steel were fully penetrated (Fig. 11) after 6 months of ponding cycles. This per- formance is likely related to the low electrical resistivity and high VPV. As the content of slag increased the electrical resistivity increased and the VPV decreased. This resulted in a decreased chloride migration co- efficient with increased slag content. The penetration depth of chloride into the ponding slabs also decreased, as did the corrosion of embedded steel bars. PCc had a slightly larger chloride migration coefficient than F60c and F30c but was more effective at protecting the embedded steel in the slabs which were ponded. This is likely due to the ability of PCc to bind chloride as Friedel ’s salt (Fig. 14) which reduced the chloride penetration depth. This highlights the importance of non-accelerated tests in the assessment of chloride binding capacity. The freeze –thaw resistance of the 100% fly ash concrete was poor with significant scaling and breakdown of the samples observed. The performance of the mix with the increased alkaline activator dosage was slightly better and was probably due to the increased tensile strength and ability to resist stresses created during freezing. As the slag content increased the performance improved significantly and could be due to the reduced VPV allowing less ingress of the freeze –thaw mediums. Previous work by Provis et al. has shown that mixes with less than 25% slag are dominated by sodium aluminium silicate hydrate gels which do not offer a similar level of pore network obstruction as space 59% 60%58%46%37% 41%F100AmF100Bm F80m F60m F30m PCmCompressi ve strengt h (MPa)Unexposed 0.52 mol/L nitric acidFig. 17.Compressive strength loss of each mortar mix following exposure to nitric acid for 8 weeks. T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 13filling calcium aluminium silicate hydrate gels found in slag dominant binders. The tensile strength also improved as the content of slag increased which may have increased resistance to tensile stresses created during freezing. The Portland cement concrete had much su- perior freeze –thaw resistance than 100% fly ash concrete and slightly better freeze –thaw resistance than F30c. The good performance for PCc may be due to its lower sorptivity and advanced pore features such as pore connectivity (which require further investigation) compared with alkali-activated concretes [115,116] . Alkali-activated 100% fly ash mortars displayed approximately a 60% reduction in compressive strength following 8 weeks exposure to nitric acid. As the content of slag increased the decline in strength decreased to 37% for F30m. This is closely linked to the apparent depth of acid penetration which also decreased as the slag proportion increased. Despite the lowered strength and degree of apparent pene- tration, the 100% fly ash mortars displayed very small mass losses and an undamaged outer surface. This is due to the low calcium content of these binders meaning they are more resistant to decalcification than fly ash/slag blends . The mass losses for the fly ash/slag blends was larger than the 100% fly ash mixes but decreased as the slag content increased. This decrease can be attributed to the more dense and stronger matrix formed allowing less ingress of acid. PCm suffered larger mass losses than any of the alkali-activated mixes. This is due to the breakdown of ettringite and calcium hydroxide, leaving behind a more porous matrix for further acid attack as observed by SEM (Fig. 22). A porous matrix was not observed at the front of acid attack in F30p due to the formation of silicate rich gel slowing the rate of further decalcifi - cation and further mass losses. In relation to the effect of activator dosages on alkali-activated fly ash concrete. The use of increased activator dosages resulted in enhanced mechanical properties for F100Ac compared with F100Bc. This was due to the increased availability of alkalis to react with fly ash and form a hardened matrix. In turn, this resulted in a decreased VPV and some improvement of durability properties. Albeit the improvement in durability performance may not be large enough to justify the increased cost and environmental impact associated with such an in- crease in the quantities of activators used. 5.Conclusions The main findings from this study can be summarised as follows: ≡As the slag proportion of alkali-activated fly ash/slag blended con- cretes increased the mechanical properties also increased and the volume of permeable voids decreased. ≡Neat fly ash concretes have poor resistance to chloride penetration due to their large volume of permeable voids. Mixes with increased slag content have similar or better resistance to chloride penetration F100A mNitric UnexposedF100B m F80m F60m F30m PCmFig. 18.Photograph of the cross-section of each mortar mix following exposure to nitric acid for 8 weeks and application of phenolphthalein solution. -11-10-9-8-7-6-5-4-3-2-10 0 1 2 3 4 5 6 7 8Mass change (%) Exposure time (weeks)F100Am F100Bm F80m F60m F30m PCm Fig. 19.Mass change of each mortar mix during exposure to nitric acid for 8 weeks. F100A mNitric UnexposedF100B m F80m F60m F30m PCm Fig. 20.Photograph of each mortar mix following exposure to nitric acid for 8 weeks. T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 14than that of Portland cement concrete. However, the ability of Portland cement concrete to bind chloride means it is more effective at protecting embedded steel from corrosion than alkali-activated mixes. ≡The alkali-activated concretes have lower freeze –thaw resistance than Portland cement concrete. As the slag proportion is increased, resistance to surface scaling is also increased. Further work is needed to investigate the effectiveness of air entrainment on the freeze –thaw resistance of alkali-activated concretes. ≡The alkali-activated fly ash/slag blends have increased acid resis- tance compared with Portland cement binders. This is due to their lower calcium content and the formation of silicate rich gel following the decalcification of calcium aluminium silicate hydrate gel which reduced the rate of further acid ingress and subsequent mass losses. Alkali-activated concrete mixes consisting of 100% fly ash have numerous barriers to their widespread application including the required curing conditions and need for significant activator dosages to achieve high strength properties. This study has shown that their durability is also a problem. They demonstrated inferior performance in relation to chloride migration, protection of steel reinforcement and freeze –thaw attack. This is mainly attributed to more open structure when compared with other mixes. However, due to their low calcium content, minimal damage and surface scaling was observed following acid attack. However, the pore structure allowed significant apparent acid penetration which caused a considerable reduction in compressive strength compared with other mixes. Alkali-activated fly ash/slag blends displayed increased performance in relation to chloride migration, protection of steel reinforcement and freeze –thaw attack. This is due to the more dense microstructure formed with less voids available for ion transport. The acid resistance of the fly/slag blends also increased as the slag proportion increased with lower strength and mass losses observed. In comparison with the control Portland cement mix, the alkali- activated concrete with increased slag content displayed comparable performance. The acid resistance was similar while the resistance to chloride migration was superior when using the NT Build 492 proced - ure. The ability to protect embedded steel was inferior due to the lack of chloride binding that was observed for Portland cement concrete. This requires further investigation, as does the freeze –thaw resistance for which the use of air entrainment techniques may be beneficial. CRediT authorship contribution statement Timothy A Aiken: Conceptualisation, Methodology, Investigation, Writing - Original Draft, Visualisation. Jacek Kwasny: Conceptualisa - tion, Methodology, Investigation, Writing – Review and Editing. Wei Sha: Conceptualisation, Methodology, Writing - Review and Editing. Kien T Tong: Investigation, Writing - Review and Editing. : . Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements The authors are grateful to the School of Natural and Built 5 101520253035404550556065 2θ(degrees)MM MMMMM MQ Q QQ H C CC CC CCCCF100A pUnexposed Unexposed UnexposedNitric Nitric NitricF30p PCp E E ECH CH CHCHCHCHCHM-mulliteQ-quartz H-hematite E-ettringite CH-calcium hydroxide C-calciteFig. 21.XRD patterns of F100Ap, F30p and PCp following exposure to ni- tric acid. F100A p PCpF30p1 mm Fig. 22.SEM image of the outer edge of F100Ap, F30p and PCp following exposure to nitric acid. T.A. Aiken et al. Construction and Building Materials 301 (2021) 124330 15Environment, Queen ’s University Belfast for provided facilities. The research studentship provided by the Department for the Economy (DfE), Northern Ireland is also gratefully acknowledged. The authors would also like to acknowledge an Invest Northern Ireland funded research project (REF. No.:RDO212970) under which some of the work related to the Portland cement concrete mix was carried out. The au- thors also appreciate the support received and useful discussions had with Prof. Marios Soutsos.

ORIGINAL PAPER Effect of GGBS Addition on Reactivity and Microstructure Properties of Ambient Cured Fly Ash Based Geopolymer Concrete S. Nagajothi1&S. Elavenil1 Received: 21 December 2019 /Accepted: 18 March 2020 #Springer Nature B.V. 2020 Abstract Geopolymer concrete is an eco-friendly alternate to conventional concrete that considerably lower green house gases emitting into the atmosphere. Fly ash based geopolymer concrete is reported to become hardened during heat curing process which comes as a major constraint for cast in in-situ applications. In this study, the aluminosilicate materials such as Ground Granulated BlastFurnace Slag (GGBS) with varying percentages such as 0%, 10%, 20%, and 30% replaces the fly ash (FA) in geopolymer concrete was used. Manufactured sand (M-sand) is used as full replacement material for natural sand as fine aggregate owing to its increasing demand. This work aims at investigating the effect of alumino silicate materials on strength properties, character-ization and micro structural analysis using Scanning Electron Microscopy (SEM), Energy-Dispersive X-ray spectroscopy (EDX), Fourier Transform Infrared spectroscopy (FTIR) and X-ray diffraction (XRD) in geopolymer concrete under ambient curing condition. The SEM and EDX results reveals that, the micro structural properties of fly ash, GGBS materials, CaO, Si/Al ratio,and gel formation have a significant effect on compressive strength and setting time of geopolymer concrete. The FTIR analysis reveals that the stretching vibration of fly ash shifts to low wave number values due to changes in geopolymerization. The X-ray diffraction (XRD) reports show that the C-S-H gel formed around 27 –30° 2theta value due to increase of GGBS in geopolymer concrete. Keywords Fly ash .GGBS .Geopolymer .Compressive strength .Microstructure 1 Introduction In construction industry, portl and cement has become the fore- most choice of binders in concrete. The huge production of portland cement annually exhausts 10 –11 EJ which is ~2 –3% of primary energy consumption globally [ 1]. Additionally, manufacturing of portland cement generates approximately one tonne of carbon dioxide for every tonne of cement produced [ 2] which contributes to 7% of CO 2emissions in global warming . Hence, most of the countries are consigning to reduce the green house gases to decrease their environmentally detrimental impact. In this context, exploring binder materials with low CO 2content or low energy materials, finding reuse possibilities for the byproduct materials from other industries gain attention. Already, a variety of byproduct materials such as fly-ash, slag and silicafume from coal, iron, and ferrosilicon production are used as an additional material for portland cement, usually on the order of 10–50% (and even in greater quantit ies sometimes). This attracts further development in obtain ing binders made completely or predominantly from waste materials [ 2]. The new technology of alkali activated binders (clinker-free) including geopolymers which are formed by the reaction be- tween aluminosilicate binder and alkali-activator solutions, help dissolution and polycondensatio no fr a wm a t e r i a l st op r o d u c ea hardened material [ 4,5]. Geopolymer concrete is mainly pro- duced by aluminosilicate materials like Fly ash, Metakaoline, Silica fume and GGBS reacts with alkaline activator solutionsuch as sodium or potassium based [ 6]. Geopolymer cement emits six times lesser CO 2than portland cement, which emits about 0.18 t of CO 2in geopolymer cement as against 1 t of CO 2 in cement [ 7]. During geopolymerizatio n process, the alumino- silicate materials are suspended into alkali activator solution to produce SiO 4and AlO 4tetrahedra (sialate network) linked*S. Nagajothi naga.jothis2014phd1138@vit.ac.in S. Elavenil elavenil.s@vit.ac.in 1School of Civil Engineering, Vellore Institute of Technology, Chennai Campus, Chennai 600127, India / Published online: 26 March 2020Silicon (2021) 13:507–516 alternately by all the oxygens [ 8]. The alkali metal cations pro- vides the charge balancing cations like Na, K, Ca to these SiO 4 and AlO 4tetrahedrons and provides polymeric precursors ( – SiO 4–AlO 4–SiO 4–SiO 4–,o r –SiO 4–AlO 4–,o r –SiO 4–AlO 4– SiO 4–) by sharing all oxygen atoms between two tetrahedral units and producing geopolymer [ 8,9]. The byproduct material of fly ash is mainly produced from thermal power plants during combustion of pulverized coalwhich contains SiO 2and Al 2O3along with the components of CaO, MgO, Fe 2O3etc. In geopolymer synthesis, fly ash has developed as a material of interest due to its availability, lowwater demand, high workability and alumino-silicate compo- sition [ 10,11].The fly ash geopolymer concrete gains strength slowly when it is in ambient temperature around 25 °C [ 12]. To achieve reasonable strength in fly ash geopolymer con- crete, the required curing temperature is 40 –75 °C [ 13]. The fly ash geopolymer concrete shows good mechanical proper-ties and enhanced durability [[ 10]; Fernandez and [ 14]]. However, the main limitation for using fly ash based geopolymers, are slow in setting and strength developmentdue to its slow reactivity [ 15]. To resolve the issue of low reactivity of fly ash and to improve the strength development, two different ways are suggested. One is addition of GGBSand the other one is mechanical processing of fly ash [ 16]. The effect of mechanical processing on its geopolymerization and reactivity has been reported [ 16–19]. While some researchers have used additives such as GGBS, flue gas desulfurization gypsum, and portland cement [ 12,20,21]. The byproduct material of GGBS is mainly produced from iron making plant. GGBS is a granular material having CaO, MgO, SiO 2and Al 2O3. Calcium Silicate Hydrate (C-S-H) gel is the main reaction product during the activation of GGBS [ 22], which results in achievement of suf ficient strength, also in ambi- ent curing condition [ 23,24]. The coupled materials of fly ash and GGBS are very efficient to give strength and stability sincethe alumina silicate materials undertake dissolution, polymeriza- tion with alkali, condensation, and solidification [ 25]. Now a days, M-sand is used for making concrete due to the lack ofnatural river sand and due to its quality controlled process. Also, full replacement of natural sand by M-sand didn ’ts h o w any adverse effect on the compressive strength of concrete [ 26, 27]. Nevertheless, only few studi es are available on using GGBSin fly ash based geopolymer concrete and the effect on strength properties and micro structural observations. The mechanical properties by using the byproduct mate- rials such as fly ash and GGBS as a replacement material for cement and M-sand as a replacement material for river sand ingeopolymer concrete have been validated by developing a model in Levenberg –Marquardt algorithm using artificial neu- ral network [ 28]. In this work, the mechanical property (com- pressive strength) of G30 grade geopolymer concrete have been determined using fly ash, GGBS, M-sand with 8 M con- centration of sodium hydroxide and observe the microstruc-ture of geopolymer concrete using Scanning Electron Microscopy (SEM) with Energy-Dispersive X-ray spectrosco- py (EDX), Fourier Transform Infrared spectroscopy (FTIR)and X-ray diffraction (XRD) analysis have been carried out under ambient curing condition. 2 Materials, Characterization and Techniques Used 2.1 Materials For making the geopolymer concrete, main source of alumino- silicate material of class F fly ash was used which was acquiredfrom North Chennai thermal power plant and GGBS, used as an additive [ 25] material was acquired from Astra chemicals, Chennai, Tamilnadu. The specific gravity of fly ash and GGBSare reported as 2.13 and 2.85, respectively. The fly ash and GGBS chemical composition obtained from X-Ray fluorescence spectroscopy are given in Table 1. The alkali activator solutions of sodium hydroxide (NaOH) and sodium silicate (Na 2SiO 3) solutions in the ratio of 2.5 were chosen for this work [ 29]. Ratio of SiO 2/N a 2O by mass of 2.0 sodium silicate solution and 8 M concentration of sodium hydroxide solution was used. Combined crushed granite coarse aggregate were used with max- imum sizes of 8, 12 and 20 mm. M-sand was collected fromKMC blue metals, Theni to be used as fine aggregate. The con- stituent of M-sand is given in Table 2. The aggregates were used in saturated surface dry (SSD) condition [ 30]. To achieve the workability of geopolymer concrete, super plasticizer (naphtha- lene based) was used.Table 1 Fly ash and GGBS Chemical compositions Composition (%) SiO 2 CaO MgO Al 2O3 Na2OK 2OF e 2O3 SO4 LOIa Fly ash 63.32 2.49 0.29 26.76 0.0004 0.0002 5.55 0.36 0.97 GGBFS 35.05 34.64 6.34 12.5 0.9 0.6 0.3 0.38 0.26 Table 2 Constituents of M-sand Constituents (%) CaO SiO 2 MgO SO 4 Cl Al 2O3 Fe2O3 Na2OK 2OP H M-Sand 6 63.86 0.7 0.07 0.07 22.93 4.25 0.0001 Nil 8.74508 Silicon (2021) 13:507–516 i ii Spherical Shape Irregular Shape iii Angular Shape Fig. 1 SEM analysis of i) Fly ash ii) GGBS iii) M-Sand 500 1000 1500 2000 2500 3000 3500 400065707580859095100Si-O-Si or Al-O-SiecnattimsnarT% Wave numbers (cm-1) Wave numbers (cm-1)2950 500 1000 1500 2000 2500 3000 3500 400060708090100 920ecnattimsnarT%Si-O-Si or Al-O-Si1200 (i). FT-IR spectra of Flyash (ii). FT-IR spectra of GGBS Fig. 2 (i). FT-IR spectra of Fly ash, (ii). FT-IR spectra of GGBS509 Silicon (2021) 13:507–516 2.2 Characterization of Materials 2.2.1 Scanning Electron Microscopy (SEM) with Energy-Dispersive X-Ray Spectroscopy (EDX) Analysis Scanning Electron Microscopy (SEM) was used to find surface morphology which was conducted usingEVO 18 research microscope, LaB 6filaments electron source. The resolution used to take this image was 8 kV . The samples were evaluated in system vacuum technique. For observingthe microstructure, SEM analysis was done on fly ash and GGBS as shown in Figs. 1(i)and1(ii). The SEM analysis of M-sand as shown in Fig. 1(iii). The Scanning Electron Microscope was equipped with Energy-Dispersive X-ray spectrometer used to characterize the micro structure of geopolymer concrete. The resultsshowed that fly ash and GGBS having high amount of silica and alumina. Fly ash particles viewed as spherical, GGBS observed as Granular and M-sand as angular in shape. 2.2.2 Fourier Transform Infrared Spectroscopy (FTIR) FTIR is a method of chemical analysis to search Si-O & Al-O reaction zones and to find the degree of geopolymerization andstructure of reaction products in geopolymer concrete [ 31]. An Attenuated total reflection (ATR) accessory technique was used in FTIR analysis. Absorbance spectra collected from 4000 cm −1 to 400 cm−1at a resolution of 2 cm−1. The infrared spectroscopic r e s u l t so ff l ya s ha n dG G B Sa r es h o w ni nF i g . 2(i),( i i ) . The IR spectrum of fly ash shows transmission bands at 3852, 3750, 3735, 3689, 3675, 3648, 1558, 1540, 1521, 1507,1090 cm −1. The main peaks are at 1540, 1507 and 1090 cm−1. The IR spectrum of GGBS shows the main peak is at 920 cm−1. Si-O-Si or Al-O-Si was observed at 1090 cm−1 for FA and 920 cm−1for GGBS. The wave number of the band in the raw material of GGBS is decreased compared with FA. As calcium content increases in the raw materials, the bandwave number match to low degrees of cross linking of the amorphous phase of raw materials is reduced [ 32]. 2.2.3 X-Ray Diffraction (XRD) Analysis A mineralogical analysis was performed by XRD used to explain the mechanical performances of materials. XRD data were obtained using a Bragg- Brentano geometry powder dif- fractometer with the parameters of 30 mA, 40KV and CuK α radiation. Scanning rate of XRD is one degree per minute from 10 to 90 degrees (2 θ) and steps of 0.05 degrees (2 θ).10 20 30 40 50 60 70 80 900100200300400500600700800 MQ MQ- Quartz M - Mullite MMQ).u.a(ytisnetnI Q MQ MQ 10 20 30 40 50 60 70 80 90020406080100----- GGBS).u.a(ytisnetnI (i). XRD pattern of fly ash (ii).XRD pattern of GGBS Fig. 3 (i) XRD pattern of fly ash, (ii).XRD pattern of GGBS Table 3 Mix Proportions (kg/m3) Mix Id Fly ashGGBS OPC M- SandNatural SandCoarse AggregateAlkali solutionWater F100M100 380 0 – 660 0 1189 171 – F90M100 342 38 – 660 0 1189 171 – F80M100 304 76 – 660 0 1189 171 – F70M100 266 114 – 660 0 1189 171 – CR100 –– 380 0 660 1189 – 171510 Silicon (2021) 13:507–516 Wave length selected for XRD was 0.154 nm (Cu). X-ray diffraction (XRD) patterns of fly ash and GGBS are shown in Fig. 3(i)and3(ii). The XRD pattern of raw fly ash illustrates with single letter for easier representation. The crystalline phases namely Q- quartz (SiO 2; JCPDS File card # 00 –046-1045) and M- mullite (Al 6Si2O13; JCPDS File card # 00 –015-0776) were determined in the diffractogram. Crystalline band of fly ash are attributed to the peaks at a 2 θabout 26°. The XRD pattern of GGBS is more amorphous and illustrates broad spectrum diffuse band in the range of 2 θabout 20 –40° which has better reactivity when compared with crystalline phases of fly ashwhich contains crystalline phases of silica and alumina. 2.3 Techniques Used Fly ash based geopolymer concrete samples with partial re-placement of GGBS (0%, 10%, 20% & 30%) were cast with the ratio of alkali solution to total alkali binders as 0.45. The alkali solution prepared by mixing of NaOH and Na 2SiO 3emits heat in large quantities. Hence, the solution was mixed 24 h before making the concrete specimen [ 22]. Some studies showed that the solutions were directly mixed to the dry mix-ture of other materials [ 33]. NaOH solution is prepared 24 h before the casting of concrete, to avoid the extra heat in geopolymer. Before adding the alkalis, the Saturated SurfaceDry (SSD) aggregates were mixed with the binders in the pan mixture for 5 min. To maintain the workability of concrete, 1% of super plasticizer was added to the total binder. The mixproportion has been adopted based on IS 10262 for attaining M30 grade of concrete. The same mix proportion has been taken for G30 geopolymer concrete since no mix design isavailable for geopolymer concrete [ 34]. The mix proportions are given in Table 3. Geopolymer concrete was designated as FxMy, where F indicates fly ash, x indicates the percentage of fly ash used when replaced by GGBS, M indicates M-sand and y indicates replacing percentage of natural sand by M-sand. For example,F80M100 indicates that 20% GGBS replaced the fly ash and 100% river sand was fully replaced by M-sand. Geopolymer concrete was prepared and cast in moulds of size150x150x150mm. To find the effect of GGBS and M-sand on compressive strength, the specimens were tested after 28 days under ambient curing conditions. For making ordi- nary portland cement concrete, cement was used as binder and river sand was used as fine aggregate for comparison purpose.CR100 indicates cement with natural sand. The fresh mix condition of geopolymer concrete is shown in Fig. 4. After finding the compressive strength of geopolymer con- crete mixtures, the micro structure of concrete specimens were examined by microscopic analysis. 3 Results and Discussions 3.1 Compressive Strength Four geopolymer concrete and one conventional concretecube specimens were tested to find the compressive strength of concretes and the test results are indicated in Fig. 5.T h eF100M100 F90M100 F80M100 F70M100 CR100aPM,htgnertSevisserpmoC MixtureFig. 5 Compressive strength Vs percentage variation of GGBS Fig. 4 Geopolymer concrete in fresh mix condition511 Silicon (2021) 13:507–516 average of three specimens is considered for the compressive strength test result. Mix numbers 1(F100M100), 2(F90M100), 3(F80M100), and 4(F70M100) were used to analyze the impact of increas-ing GGBS percentages with fly ash such as 0%, 10%,20%, and30% in geopolymer concrete. It was noted from Fig. 5,t h a t the compressive strength of geopolymer concrete increasesabout 1.18 times, 1.33 times, and 1.44 times while increasing the percentage of GGBS in the mixes of 2(F90M100) as 10%, 3(F80M100) as 20%, and 4(F70M100) respectively comparedto the fly ash alone mix of 1(F100M100). The dissolution of fly ash is not completed in ambient curing. Also the setting time of mix 1(F100M100) increases when comparing with theother mixes of 2(F90M100), 3(F80M100), 4(F70M100) due to low reactivity of fly ash [ 35]. While comparing the Ordinary Portland Cement (OPC) concrete with geopolymerconcrete, the G30 concrete could be achieved in the mixes of 3(F80M100) and 4(F70 M100) i.e. 20% and 30% replacement of fly ash with GGBS. Formulations in terms of SiO 2/Al2O3 ratio and CaO for geopolymer concrete are given in Table 4. The compressive strength and initial setting time of geopolymer concrete increases with increase in replacementpercentage of GGBS. This is due to an increase in the Si/Al ratios in source materials. As –Si-O-Si bonds are stronger than Al-O-Al and –Si-O-Al bond, increasing Si/Al ratio increases the number of –Si-O-Si bonds thereby achieving elevated compressive strength [ 36]. The fly ash/GGBS based geopolymer concrete speed up the initial setting time due to high calcium in GGBS [ 37].The compressive strength of 20% replacement of GGBS and full replacement of natural sand with M-sand showed equal strength with the conventional concrete. Microstructural observations have been conducted to justifythe better mix proportions. 3.2 Scanning Electron Microscopy (SEM) with Energy-Dispersive X-Ray Spectroscopy (EDX) Analysis of Geopolymer Concrete The SEM with EDX of the geopolymer concrete samples is shown in Figs. 6,7,8 and9. These are the micro structural images of the samples F100M100, F90M100, F80M100, and F70M100 of geopolymer concrete at age of 28 days. The geopolymer con- crete having 30% replacement of GGBS (F70M100) is morecompact, has less micro cracks and is less porous than other mixes of such as F100M100, F90M100 and F80M100 of geopolymer concrete. Higher portion of GGBS geopolymerconcrete shows some non-reacted or partly reacted GGBS particles when compared with other mixes. The fly ash alone mix F100M100 geopolymer concrete produces geopolymergel primarily sodium alumin o silicate hydrate. Sodium alumino silicate hydrate (N-A-S-H) is the prime reaction prod- uct of the geopolymer gel which is produced by the low cal-cium fly ash [ 38]. While increasing GGBS, calcium alumino silicate hydrate (C-A-S-H) is reaction product when calcium compound rises in geopolymer concrete [ 39]. By adding the fine particles as an additive in geopolymer concrete, the den- sity and homogeneity could be improved. The elemental More Porous Fig. 6 SEM & EDX analysis of mix 1 F100M100Table 4 Formulations of geopolymer concrete in terms of SiO 2/Al2O3ratio and CaOMixture Fly ash (wt%) GGBS (wt%) SiO 2/Al2O3ratio CaO (wt%) F100M100 100 0 2.36 2.49 F90M100 90 10 2.38 5.71F80M100 80 20 2.41 8.92F70M100 70 30 2.43 12.14512 Silicon (2021) 13:507–516 percentage by EDX analysis (Atomic %) is given in Table 5. The 3D networked polysialate- siloxo, and polysialate- disiloxo polymers are created by higher Si/Al i.e. Si/Al > 2.5 . Silicon, oxygen and aluminium along with lowconcentrations of calcium and sodium are qualified to the formation of 3D network (Ca,K)-polysialate-siloxo [ 41]. X-Unreacted or partially reacted GGBS particles. Y-Unreacted or partially reacted fly ash particles.Z-Geopolymer gel. Y Fig. 7 SEM & EDX analysis of mix 2 F90M100 X Y Z Fig. 8 SEM & EDX analysis of mix 3 F80M100 More Compact and less cracks Fig. 9 SEM & EDX analysis of mix 4 F70M100513 Silicon (2021) 13:507–516 3.3 Fourier Transform Infrared Spectroscopy (FTIR) Analysis Figure 10shows the FTIR results of 28 days for replacing the fly ash with GGBS as 0%, 10%, 20% & 30%. The stretching vibration of fly ash occurred at 1062 cm−1which is shift to- wards low wave number like 1011 cm−1,9 9 5c m−1,9 8 6c m−1 and 979 cm−1for replacing percentage of 0%, 10%, 20% and 30% respectively in geopolymer concrete. The shift is approx-imately 51 cm −1,6 7c m−1,7 6c m−1,8 3c m−1. From the results, it reveals that due to increase amount of GGBS in the geopolymer concrete forming the C-S-H gel with N-A-S-Htype gel with a reduction of Al. This leads to changes in geopolymerization and improve the strength properties. Large bands around 3648 –3852 cm −1are H-O-H stretching vibrations. Bands like 1507 –1540 cm−1are -OH group of bending vibration of products of hydrated reaction of water. In geopolymer paste, these bands shows the reaction of alka-line activation products and water [ 42].3.4 X-Ray Diffraction (XRD) Analysis The XRD pattern of geopolymer concrete with percentages of G G B Si ss h o w ni nF i g . 11. It reveals that there was a change from the raw materials chemistry since the reaction of alumino silicate materials with alkali activator solutions. The peaksaround 27 –30° corresponding to C-S-H gel are present found in GGBS increasing geopolymer concrete and also the intensity of crystalline phases decreases when increasing GGBS percent-age. A-S-H and C-S-H gel are formed around 30° and 50° in all GGBS samples and crystalline phases decreases [ 43]. During first step of geopolymerization the amorphous compounds dis-solve easier than crystalline compounds (i.e dissolution of spe- cies) which yield higher amounts of SiO 2and Al 2O3to combine geopolymerization reaction product. Due to this, the results givehigh degree of geopolymerization and higher mechanical strength when the replacement percentage of GGBS is higher in geopolymer concrete. Calcium aluminosilicate structure morereactive than siliceous structure. 4 Conclusion The present study examined the compressive strength and micro structural studies of G30 grade geopolymer concrete at ambient curing conditions. The following conclusions are drawn from the results. &G30 grade of geopolymer concrete is achieved by the combination of 80% fly ash and 20% GGBS aluminoTable 5 Elements percentage by EDX analysis (Atomic %) Element O K Si K Na K Ca K Al K Si/ Al F100M100 65.80 20.46 5.75 0.63 5.81 3.52 F90M100 65.49 16.91 5.65 1.57 4.62 3.66F80M100 63.06 14.33 4.69 2.87 3.85 3.72F70M100 55.86 14.13 3.54 4.78 3.68 3.83 600 900 1200 1500 1800 2100 2400 2700 3000 3300 3600 39006080100120140160 F70M100 F80M100 F90M100ecnattimsnarT% Wave numbers (cm-1)F100M100Fig. 10 FT-IR spectra of geopolymer concrete514 Silicon (2021) 13:507–516 silicate source material with the ratio of Na 2SiO 3to NaOH as 2.5. &When increasing GGBS percentage in fly ash based geopolymer concrete as 10%, 20%, and 30%, the com- pressive strength has been increased as 17%, 31%, and41% respectively when compared with fly ash alone in geopolymer concrete. &Geopolymer concrete is a replacement concrete for con- ventional concrete, since the compressive strength of GGBS-Fly ash based geopolymer concrete is mostly equal with the conventional concrete. &SEM / EDX analysis showed a less dense structure with low content of GGBS or without GGBS while the com- pactness of geopolymer concrete increases when the con-tent of GGBS increases. &FTIR results showed the changes in geopolymerization and formation of C-S-H gel when increasing the percent-age of GGBS, which in turn increases the strength properties. &XRD analysis revealed that the intensity of crystalline phases decreased when increasing the percentage of GGBS and forming the C-S-H gel which has better reac- tivity when compared with fly ash. Acknowledgements Authors would like to acknowledge the Management and Dean-School of Civil Engineering, Vellore Institute ofTechnology, Chennai, India for providing the necessary support to carry out this research.

Construction and Building Materials 302 (2021) 124170 Available online 13 July 2021 0950-0618/© 2021 Elsevier Ltd. All rights reserved.Review A review on the utilization of red mud for the production of geopolymer and alkali activated concrete Aman Kumara, T. Jothi Saravanana,*, Kunal Bishtb, K.I. Syed Ahmed Kabeerc : Alkali-activated binder Fly ash Ground granulate blast furnace slag Rice husk ash Metakaolin Silica fume ABSTRACT The production of ordinary Portland cement releases many greenhouse gases, which have led to adverse envi- ronmental changes such as global warming and climate change. These phenomena have invigorated the utili- zation of various industrial waste products to synthesize geopolymer materials and alkali-activated binders (AAB) to alleviate OPC ’s use in construction and building materials. Red Mud (RM) is one such harmful radioactive waste material generated as a by-product in aluminum production by Bayer ’s process. This article highlights RM’s utilization of materials such as fly ash, ground granulated blast furnace slag, silica fume, met- akaolin, refuse mudstone, and rice husk ash to synthesize geopolymer and AAB samples. The heavy metal ions present in the raw materials in their weak acid-soluble and reducible fraction were in their residual and oxidizable fractions in the geopolymer, providing better stability and curing a severe threat to the environment. The freeze –thaw results showed that the mass-loss rate of the geopolymer was less than that of concrete. The geopolymer samples exhibited activity concentration levels well within the prescribed limit. Studies also include the analysis of Fe’s role in the ettringite structure using the Mossbauer spectra analysis. Besides, an overview of the recent advances in the use of RM for geopolymer synthesis has been presented in terms of hardened, fresh, and thermal properties and geopolymer and AAB ’s microstructure analysis. The effect of ambient curing, heat treatment, elevated temperature curing, and autoclave curing has been briefly outlined, along with the impact of the RM fraction in the geopolymer samples. The research findings revealed that the RM-based geopolymer and AAB manifested similar properties to OPC concrete for various civil engineering applications. 1.Introduction An enormous amount of waste generated by industries, some even toxic and radioactive, and if not congruously disposed of, could lead to grave environmental issues. The growing quantities of waste materials, landfill area deficit, and depletion of natural earth materials highlight the gravity of finding innovative recycling techniques and reusing waste materials . Recently, these waste materials have proven to be fit for use in the construction industry, majorly by recycling (i.e., processing waste materials for their application in construction materials) and reusing the components . Red mud (RM) is the by-product of bauxite ores after digestion by caustic soda utilizing Bayer ’s process to extract alumina. Globally, around 58 million tonnes of alumina is produced annually, of which India accounts for 2.7 million tonnes . Current disposal exercises involve dry stacking, storage in huge ponds, waste lakes, or landfills, as shown in Fig. 1 [4–6]. The environmental concerns linked with RM’s disposal are related to its high pH (10.5 –12.5) value, the uncertainty of storage, seepage of alkali into groundwater, vast areas of land consumed . Besides, RM’s chemical properties, such as the deficiency of plant nutrients, high salinity and alkalinity, and low organic matter content, hinder plant growth. Moreover, the high fineness of RM particles leads to alkaline dust forming that contaminates air (dry disposal), thus jeopardizing the surrounding vegetation . Traces of heavy metals have been found in RM, which, when seeped into the water and soil, degrades their quality and leaves them unfit for consumption and farming, respectively . There has been consistent development and research on the usage of RM as source material for a range of products. RM has a high quantity of metallic oxides, especially iron. It has good particle dispersion and a *Corresponding author. E-mail address: tjs@iitbbs.ac.in (T.J. Saravanan). Contents lists available at ScienceDirect Construction and Building Materials u{�~zkw! s{yo| kro>! ÐÐÐ1ow �o�to~1m{y2w {mk�o2m{zl� twnyk�! Received 30 April 2021; Received in revised form 25 June 2021; Accepted 5 July 2021 Construction and Building Materials 302 (2021) 124170 2high specific surface area. It has excellent stability in the solution, meaning that it does not easily coagulate. It has other characteristics comparable to porous materials, such as low density and a range of properties in the thermal, physical, and mechanical fields. All these properties make RM suitable for its use in numerous applications. Generally, three approaches are followed for the utilization of RM, namely, for recovery of valuable metals, use as construction material (brick, cement, and ceramic), and for safeguarding the environment (harmful gases, wastewater, and potentially harmful elements contam - inated soil rehabilitation) . There is vast literature available on the various disposal techniques and applications in which red mud is utilized. Correspondingly, at- tempts have been made to review the progress and the recent de- velopments in handling the bauxite residue. For example, Zhang et al. (2021) and Mukiza et al. (2019) reviewed the articles focused upon the use of RM as a road material in pavements, road bases, and asphalt mixtures. While Reddy et al. (2020) did a comprehensive review on the use of RM in geotechnical engineering applications, the author also highlighted the need to study the performance of RM in synergy with other materials. Several other research papers also provide an overview of the various applications in which RM can be used, such as construction and chemical applications, environmental and agro- nomic applications, and metallurgical applications [12–17]. Power et al. (2011) discussed the current management, disposal, and storage practices in use for RM, where the remediation of RM is given in detail. While Lima et al. (2017) also pointed out that the neutralization of RM is not enough to ensure a safe and sustainable application. Instead, it should be used in applications where it acts as an inert material and avoids leaching. Kanyal et al. (2021) provided a review on the effect of RM utilization in concrete as cement replace - ment. They observed that the mechanical properties of concrete enhanced up to a 15% addition of RM. While, Putrevu et al. (2021) reviewed the use of RM as a partial replacement of cement, fly ash, and filler materials in the cementitious materials. The studies base on the effect of RM addition as an admixture in self-compacting concrete was analyzed by Topli ¯ci˘c-˘Cur¯ci˘c et al. (2017) , where the authors also suggested ways in which it can be used in the production of construction and building materials. Pontikes and Angelopoulos (2013) reported using bauxite residue as a pozzolanic material and as a raw material in cement production. Even though the prevailing reviews provide an overview of the various applications RM can use, there is a scarcity of review papers discussing the use of RM in construction materials in synergy with other waste materials. This paper attempts to highlight RM’s valorization as a construction material in the form of geopolymer composites by reviewing the previous studies conducted on the subject. It tries to point out the optimum conditions for utilization of RM in different civil en- gineering applications, which might be useful for the end-users of RM, the policy makers, and researchers by providing them comprehensive practical and fundamental information for employing RM in various civil engineering projects. 2.Materials characterisation and experimental Methodology: This article reviews results from papers aiming towards the use of RM in combination with other materials [Fly Ash (FA), Ground Granulate Blast Furnace Slag (GGBFS), Silica Fume (SF), Rice Husk Ash (RHA), Metakaolin (MK), Refuse Mudstone (RFM)] for the synthesis of the Fig. 1.(a) Disposal of RM by NALCO, effectuating the formation of a lake in the vicinity; (b) RM at NALCO (dry disposal); (c) Red mud at BALCO; (d) Red mud pond at INDAL (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) A. Kumar et al. Construction and Building Materials 302 (2021) 124170 3geopolymer using different techniques. Fig. 2 shows countries where studies have been done corresponding to the different materials used along with RM. The present studies on the utilization of RM-based geopolymers and AABs are very disjointed. To date, there is a lack of extensive review or compilation of the developments on the use of RM-based geopolymers and AABs in construction and building materials. Considering this research gap, authors have tried to analyze the most recently published articles from various reputed journals. Among the papers published related to the term ‘red mud, ’ the papers that have dealt with red mud- based geopolymers and AAB in the construction and building domain were considered. Then the selected articles were further categorized according to the material they were blended with, the properties stud- ied, and the type of sample (binder, mortar, concrete, controlled low strength material, and grout), and then the in-depth review was carried out. Several parameters were varied in the reported studies, and their effect on the geopolymer properties was observed. He et al. (2011) and Li et al. (2018) examined the influence of the addition of RM up to 80% in the mix, while the concentration of alkaline activator NaOH was varied from 0 to 15 M with heat curing in work done by Hoang et al. (2020) . The effect of synthesizing the geopolymer using different gradations of FA was investigated by Choo et al. (2016) , while the influence of changing the water to binder ratio was examined by Kim et al. (2017) . Sulphur and freeze –thaw durability were studied by Zhang et al. (2016) and Zhao et. (2019) , respectively. More - over, geopolymer behavior at elevated temperatures was appraised by Yang et al. (2019) . The effect of pretreatment on the RM, such as pulverization, calcination, washing, and geopolymer synthesis after going through flue-gas desulfurization (FGD), was also studied. The ef- fect of alkali thermal activation was studied in the research conducted by Hu et al. (2018) . Table A1. lists the variation in the studies considered on RM and FA mixture in this article. Like the studies based on RM-FA geopolymer, the influence of RM’s quantity in the RM-GGBS based geopolymer was also investigated. Zhang et al. (2020) studied the effect of RM gradation by utilizing five different sizes of RM for geopolymer synthesis. At the same time, both Ye et al. (2014) and Bayat et al. (2018a) noted the changes in the geopolymer and alkali-activated binder (AAB) synthe - sized, respectively, using calcined RM. Where Krivenko et al. (2017) utilized three different alkaline components (soda ash, sodium meta- silicate, soluble sodium silicate) for the synthesis of the AAB, Chen et al. (2018) examined the influence of alkaline content present in RM by utilizing washed RM for binder and concrete formation. RM was washed by immersing in distilled water for 24 h, and pH was measured and repeated until the pH was around seven. The thermal conductivity and efflorescence in the AABs were examined by Bayat et al. (2018) and Jung et al. (2018) , respectively. Studies on the use of GGBS and RM as source materials are depicted in Table A2. The use of FA and GGBS together with RM was also promising, studies on which are shown in Table A3. Singh et al. (2018) pre- pared geopolymer samples with both pulverized and unprocessed RM in different proportions with FA to study the effect of mechanical activa - tion. Lin et al. (2020) appraised the geopolymer synthesized at different RM, slag, FA, and alkali activator content, by varying all three systematically. Apart from FA and GGBS, attempts have also been made to use other materials combined with RM to attain a geopolymer. While Ye et al. Fig. 2.World map showing the countries where studies have been done showing materials used in combination with red mud. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) A. Kumar et al. Construction and Building Materials 302 (2021) 124170 4Table A1 Studies done on utilizing RM and FA for the synthesis of geopolymer and AAB [4,5,23,25 –30,43,45,46] Source of RM Country RM combined with Parameters studied Role of RM Reference Noranda Alumina, LLC USA FA (Class C) ≡RM was blended with FA in three different RM/FA weight ratios (i.e., 80:20, 50:50, 20:80) to form the geopolymer with a 1.5 M sodium trisilicate solution activator. The alkaline solution was not used as such. Source material for geopolymer He et al. (2011) ALCOA Australia FA (Class F) ≡The incorporation of RM was varied from 0 to 40 wt% in the RM- FA mix, which used 6 M NaOH solution as an activator. Source material for geopolymer and formation of paving blocks Kumar et al. (2012) Alcoa WorldAlumina L. L.C. USA FA (Class F) ≡Geopolymer was synthesized using RM & FA in a mass ratio of 4:1. ≡50 wt% Sodium hydroxide, 2 M sodium trisilicate, and deionized water were used to make the alkaline activator with a proportion of 3:7:3. ≡The durability of this geopolymer was then tested in sulfuric acid solutions and deionized water. Source material for geopolymer Zhang et al. (2016) KC Corporation South Korea FA ≡RM was used as an alkali activator for the synthesis of RM-FA AAB where FA of five different median sizes (D50 ˆ12, 21, 22, 15, 20) was mixed with RM (RM varying from 0% to 60% by mass) for the synthesis of the AAB. ≡The properties of the AAB so formed using the different FA and varying RM content were then studied. CLSM Choo et al. (2016) Xinfa Group, Shandong China FA (class C) ≡Effect of varying the NaOH concentration (0, 1.5, 2.5, 3.5, 4.5, 5.5, 6.5 Mol) in geopolymers made with FA & RM before and after FGD was studied. Source material for geopolymer Nie et al. (2016) Samho-eup, Jeollanam-do South Korea CNFClass F FA ≡RM was mixed with CNF powder (Ca(OH) 2, Na2CO3 & FA) to prepare the geopolymer, where RM was varied from 0 to 30 wt% and composition of C, N and F were varied. ≡RM incorporation was studied by varying RM content (RM ˆ 10%, 20%, 30%). ≡Influence of water to binder ratio was studied (w/b ˆ0.4, 0.44, 0.48). ≡The concentration of NaOH was varied (1, 3 & 5 wt%) to study its influence. Brick production Kim et al. (2017) Xinfa Group, Shandong China FA (American class F & class C, Chinese class F) ≡Three types of FA were used. ≡Variation in NaOH (mol) was studied (2.5, 5, 7.5, 10) with and without the addition of a hybrid solution of sodium silicate and sodium hydroxide. ≡Curing was done at two temperatures (20 •C & 60 •C). Source material for geopolymer Hu et al. (2018) Bayer factory in southern China China MSWIFA ≡The ratio of MSWIFA to RM was varied (2:8, 3:7, 4:6, 5:5). ≡Proportion of sodium silicate was varied (8%, 10%, 12%, 14%, 16%). ≡The effect of Mechanical activation provided was also studied (30 min). Source material for geopolymer Li et al. (2018) Zibo City, Shandong Province China FA ≡Geopolymer was synthesized using alkali thermal pre-treated RM (calcined at 800 0C with different dosages of Na2O, i.e., 0, 2.5, 5 & 7.5 wt%) was mixed with FA in different mass ratios. ≡RM: FA was fixed to 50:50, and the influence of Na2O dosage was studied. ≡Na2O was fixed at 5%, and the influence of FA content was observed (FA content was taken to be 0, 25, 50, 75, and 100 wt %). Source material for geopolymer Hu et al. (2018) Alcoa World Alumina, LLC USA FA (Class F) ≡In this study, geopolymer was synthesized from RM slurry and FA (1:4 mass ratio) with 50% sodium silicate and 50% sodium hydroxide as the alkaline activator. ≡The behavior of the geopolymer so formed was studied at elevated temperatures (room temperature, 160, 400, 600, 800 & 1000 0C). Source material for geopolymer Yang et al. (2019) AlcoaWorld Alumina, LLC USA FA (Class F) ≡RM slurry was mixed with class F FA to synthesize the geopolymer cured at three different temperatures (room temperature, 50 •C & 80 •C). ≡Freeze-thaw durability was studied along with the effect of curing duration. Source material for geopolymer Zhao et. (2019) Shanxi province China FA, PC ≡RM was used as an alkali source in the geopolymer synthesized using RM, FA, and PC. ≡Mass % of PC was fixed at 10% for all the specimens, while the effect of RM incorporation was studied by varying the RM and FA mass fraction (RM ˆ0, 20, 40, 60 wt%). CLSM Yuan et al. (2019) Tan Rai Alumina plant Vietnam FA (Class F) ≡Effects of heat curing and autoclave curing were examined in this study. ≡For heat curing the temperature (50, 80, 100, 150 and 200 •C), dwell time (10 and 24 h) and NaOH concentration (0, 1, 3, 5, 7, 9, 11, 13 and 15 M) were varied. ≡For autoclave curing, additionally, pressure (0.4, 0.8, 1.2, and 1.6) was also varied apart from temperature (144, 170, 188, and Masonry units Hoang et al. (2020) (continued on next page) A. Kumar et al. Construction and Building Materials 302 (2021) 124170 5(2016) synthesized one-part geopolymer by blending alkali thermal activated RM with Silica Fume (SF) in different proportions, Kaya et al. (2016) examined the behavior of geopolymer upon varying the amount of calcined RM used with Metakaolin (MK). The influence of the quantity of Refuse Mudstone (RFM) and alkaline activator in the geo- polymer prepared using RFM and RM was observed by Zhou et al. (2020) . Both He et al. (2012) and S˘ekou et al. (2017) studied the effect of curing duration, Rice Husk Ash (RHA) to RM mass ratio, and concentration of NaOH solution in the geopolymer synthe - sized using RHA and RM. For geopolymer solely based on RM, the effect of different Na2O dosages in the alkali thermal activation was discerned. Table A4 depicts the various parameters studied while using RM with these materials. In the documented studies, different tests (Table A5 to A7) were performed to understand better the properties of the geopolymer formed. Compressive strength was the most important factor of them all and was tested in all of the studies. Additionally, freeze –thaw durability of the RM slurry-FA-based geopolymer , durability to sulphate attack resistance , thermal conductivity, and specific heat were also evaluated. Fe phases in the geopolymer product were mentioned in studies conducted by several authors [6,8,24,38,42,43] . At the same time, Hu et al. (2018) investigated Fe’s role in geopolymerization using Mossbauer spectra analysis. Besides, Kang et al. (2016) and Jung et al. (2018) also observed the occurrence of efflorescence in the samples using Paint.NET software. The raw materials ’ chemical composition was determined using X- ray fluorescence shown in Tables A8 to A9. Chemical analysis revealed that RM is rich in SiO2 and Al2O3, and its high alkalinity was beneficial for the dissolution of the raw materials to give rise to reactive Si and Al . The presence of CaO in ordinary Portland cement (OPC) assisted the formation of CSH gel in the geopolymer [8,35] . Hu et al. (2018) also attributed the higher strength of geopolymer and a denser structure Table A1 (continued ) Source of RM Country RM combined with Parameters studied Role of RM Reference 201 •C), dwell time (4, 8, 10, 12, and 16 h), and NaOH concentration (0 and 1 M). Zibo, Shangdong China FA, CG ≡Apart from FA and RM, CG was also utilized for synthesizing the geopolymer, the ratio being 2:6:2 (with uncalcined CG), 4:4:2 (with Calcined CG) and 2:6:2 (with Calcined CG). Source material for geopolymer Koshy et al. (2019) Table A2 Studies done on utilizing RM and GGBFS for the synthesis of geopolymer and AAB [8,32 –37,44,47,48] Source of RM Country Material of study Variation studied Role of RM Reference Chalco Co., Zhengzhou China GGBFS ≡The effects of modulus SiO2/Na 2O molar ratio (1.3, 1.4, 1.5, 1.6, 1.7, 1.8, and 1.9). ≡Dosage of sodium silicate (3.45, 4.6, 5.75, 6.9, 8.05, 9.2, and 10.35% by weight). ≡RM/GGBFS mass ratio (10:0, 9:1, 8:2, 7:3, 6:4, 5:5, 4:6, 3:7, 2:8, 1:9, and 0:10). ≡Calcination temperature for RM (200, 300, 400, 500, 600, 700, 800, 900, and 1000 •C). Source material for geopolymer Ye et al. (2014) – South Korea OPC, AASC ≡Two AAB composites were produced using Alkali Activated Slag Cement (AASC) and OPC as binders with RM as a replacement (RM ˆ0% to 30%). ≡Efflorescence characteristics were also studied along with other properties of the geopolymers. Source material for AAB and a replacement of OPC Kang et al. (2016) – Ukraine GGBS, OPC ≡RM, GGBS, and OPC content were varied along with a variation in the alkaline component (soda ash, sodium metasilicate, soluble sodium silicate) for the synthesis of the AAB. ≡Radiological properties were also evaluated. Aggregate in cement and concrete Krivenko et al. (2017) Guangxi Province China GGBFS ≡Slag was substituted at 25, 50, and 75% by RM in the system. ≡Sodium silicate solutions of modulus ranging from 1.6 to 2.2 were used. ≡Samples were cured at 25, 40, and 60 •C. Source material for geopolymer Lemougna et al. (2017) Zhaofeng Aluminum Company, Shanxi China GGBFS ≡The influence of the addition of washed RM was also studied and the variation in RM’s incorporation. ≡RM/GGBS (wt./wt.) was taken to be 3:7 and 5:5, and the content of NaOH, water, and Na2SiO3 was varied in the alkali activator. Pervious concrete good at adsorbing heavy metals Chen et al. (2018) Jajarm Alumina Plant Iran AAS ≡AAB concrete was prepared using RM & Alkali Activated Slag (AAS) (with RM content ˆ0%, 10%, 20%, 30% & 40%). ≡Samples with RM calcined at 550 •C and 750 0C were also prepared with the incorporation of 40%. Source material for AAB Bayat et al. (2018) – Iran GGBS ≡RM content was varied (0, 10, 20, 30, and 40% by weight), and its effect on the properties of AAB was studied, including the thermal conductivity for its application in JPCP. JPCP Bayat et al. (2018) – South Korea AASC ≡Alkali Activated Slag Cement (AAS) was used with RM for pavement design in this study. ≡The quantity of RM was varied (0, 10, 20, 30% by weight), and the influence of RM was studied along with the efflorescence results in the formed geopolymer. Pavement Hyeok-Jung et al. (2018) Xinfa Plant, Shandong China GGBFS ≡In this study, RM of different particle size fractions (bulk, 39, 63, 104, 268, 392 µm as mean particle size) was used to study the influence of particle size of RM in the geopolymer synthesized by RM and slag with a mass ratio of 6:4, activated by 2 M NaOH solution. Source material for geopolymer Zhang et al. (2020) – China BFS ≡Chloride ion permeability test was carried out along with the corrosion test of the steel bar. Source material for geopolymer Liang et al. (2021) A. Kumar et al. Construction and Building Materials 302 (2021) 124170 6to the formation of hydration products like calcium silicate hydrate (CSH) from the Ca present in the class C fly ash. The oxide and elemental composition analyses of RM particles used in the studies are given in Table 1. The physical properties of the RM particles also influenced the properties of the corresponding geopolymer. Figs. 3–8 shows the RM particles ’ physical properties used in the studies, while Table A10 de- picts the other materials used with RM. The variability in the RM par- ticles ’ surface area was due to the difference in processing methods used. Also, the adsorption-based Brunner –Emmet –Teller method (BET) method gave higher surface area values than the Blaine method for the similar fineness of particles. Additionally, the lime reactivity test to study RM’s pozzolanic behavior by Singh et al. (2018) . 2.1. Pretreatment It is seen that pulverization of RM (Fig. 9) made the sample more homogeneous, consequently minimizing the effect of compositional variation on the end product, which made it easier to control the syn- thesis process and better facilitate the geopolymerization . Singh et al. (2018) observed an increase in the SiO2 / Na2O ratio and a fall in the Na2O / Al2O3 molar ratio after pulverization of RM, indicating a reduction in the optimum content of alkali needed for geopolymerization. Likewise, calcination of RM was also seen to improve the properties of the geopolymer gel. Calcination of RM improved the solubility of aluminosilicate materials . Bayat et al. (2018) chose 550 •C and 750 •C for the calcination of RM based on the two endothermic peaks seen in the thermogravimetric differential thermal analysis thermogram of RM, whereas Hu et al. (2018) calcined the RM at different temperatures ranging from 200 •C to 1000 •C and found a temperature of 800 •C to be suitable for calcination based on the alkali leaching test. Muraleedharan and Nadir (2021) reported that ambient curing was sufficient for RM geopolymers in applications requiring lower strength, while heat curing at 60–80 OC was suggested for higher strength and durability geopolymer samples. 3.Properties of red mud in combination with fly ash 3.1. Setting time and hydration The initial and final setting times for FA-based geopolymers were 210 and 310 min, as reported by Kumar and Kumar (2013) . The setting time of the geopolymer decreased with the addition of RM up to 15% by weight. Further addition of RM deaccelerated the setting time of the geopolymer. In the study conducted by Hu et al. (2018) , it was observed that when only NaOH was used as the activator, the geo- polymer synthesized using RM and FC in equal quantity had the highest cumulative heat in the first 72 h. Kumar and Kumar (2013) studied the effect on the calorimetry of geopolymer with RM’s addition, and the highest peak of heat evolution was observed in the sample with 10% RM content. Furthermore, the rise in RM content leads to a decrease in peak in- tensity. This decrease resulted from iron oxide ’s receding solubility, which is more soluble in the acidic region than in the basic region. Yuan et al. (2019) studied the reaction kinetics of the RM-activated binder (RM ˆ0 to 60 wt%). The time to reach the main peak was the longest for samples with 20% RM and the shortest for samples with 60% RM incorporation. The former also had the highest amount of cumula - tive heat released. Additionally, the height of the main peak increased consistently with the rise in RM content. 3.2. Compressive strength 3.2.1. Binders Hu et al. (2018) synthesized binders using various FA types and reported a maximum strength of 36.1 MPa, registered for samples with Chinese Class F FA and 5 mol NaOH at ambient temperature curing for two months. It was also noted that samples with composite activator solution reached a higher strength than the samples using just NaOH at ambient and elevated temperatures. According to the study, an increase in the curing temperature gave rise to the concentration of NaOH so- lution required for optimum strength. Nie et al. (2016) proposed that Class C FA is used to prepare the geopolymer as Class F FA could not provide enough strength at ambient temperature. The optimal NaOH concentration for RM before flue gas desulfurization (FGD) was 2.5 M, which gave a strength of 15.2 MPa at 28 days. The optimal NaOH concentration for geopolymer utilizing RM after FGD was 3.5 M and granted it a strength of 20.3 MPa, which shows that the optimum strength increased upon the use of RM obtained after FGD. Hence, providing a way to utilize the waste material in a better manner. Li et al. (2019) discerned an increase in the compressive strength of the binder with an increase in municipal solid waste incin - eration fly ash (MSWIFA) up to a maximum substitution of 30% RM. UCS of the binder was higher when the modulus was 2 (Na2O/SiO 2 was 0.54), and it was maximum when the activator added was 14% (SiO 2/ Al2O3 ratio was 2.08). While the study also pointed out that mechanical activation was more efficient in the mix with 14% activator content, resulting in a 28-day binder strength of 12.75 MPa. In the compressive strength results of geopolymer binders synthesized using RM and FA with 6 M NaOH concentration used as an activator, Kumar and Kumar (2013) observed samples with 10% RM by wt. (RM10N6) had the Table A3 Use of FA & GGBS in combination with RM for the synthesis of the geopolymer and AAB [38,39,59] Source of RM Country Material of study Variation studied Role of RM Reference Ms. Hindalco, Belgaum India FA (Class F), GGBS, MS ≡Mechanical activation was provided to the RM in the form of pulverized/processed RM. ≡Both unprocessed and processed RM were then used to synthesize the geopolymer at varying RM and FA content (RM ˆ30%, 50%, 70%, 90%). ≡The influence of variation in NaOH concentration (6, 8 &10 M) was also studied. Source material for geopolymer, pavers, high strength concrete Singh et al. (2018) Ms. Hindalco refinery, Belgaum, Karnataka India FA, GGBFS ≡Variation in RM content (10% to 50%) and mortar proportion (RM: FA: sand: GGBFS) was studied along with the variation in a binder: fine aggregate ratio in the geopolymer brick formed (1:1 and 1:2). Masonry brick Singh et al. (2020) Shandong Xinfa aluminum company China GGBFS, FA ≡RM, FA, GGBFS, and NaOH were used for the synthesis of the geopolymer. NaOH was used as the alkali activator. ≡The study was divided into two phases: the RM: slag ratio was varied (7:3,6:4,5:5,4:6 and 3:7) and the alkaline activator content (4,6,8,10 and 12 M) keeping the FA content constant at 0%. ≡The amount of FA in the mix was varied from 0 to 25% with an RM: slag ratio of 5:5 and an alkali activator concentration of 8%. Grout Lin et al. (2020) A. Kumar et al. Construction and Building Materials 302 (2021) 124170 7maximum compressive strength. Hu et al. (2019) examined the effect of the addition of an alkali activator in the RM-FA (50:50) geo- polymer along with the effect of FA dosage. The RM used for the geo- polymer synthesis was alkali thermal pre-treated by calcination of RM with NaOH pellets at different dosages of Na2O (0, 2.5, 5, and 7.5 wt%), done at a temperature of 800 •C. Maximum compressive strength was shown by sample with 5 wt% of Na2O. As Na2O dosage increased further, there was a decline in the strength of the geopolymer. Next, the effect of FA dosage was studied by keeping the Na2O dosage at 5 wt%. The geopolymer sample with 50 wt% of FA had the maximum compressive strength of 23.8 MPa at the curing age of 28 days. As the mass fraction of FA was increased further, the strength depreciated. The geopolymer sample with 100 wt% of FA had a strength of 5 MPa at the curing age of 28 days. Thus, pretreated RM and FA synergistically contributed to the strength of the geopolymer. In another experiment, Kim et al. (2017) examined the strength of the geopolymers made from RM, FA, Ca(OH) 2, and Na2CO3, while NaOH was used as the alkaline additive. With the increase of RM incorporation, the authors discerned a decrease in the samples ’ compressive strength. Also, increasing the water binder ratio (w/b) in pastes with low flowability reduced the geopolymer ’s 28-day compres - sive strength. The geopolymer prepared in the study done by He et al. (2012) constituted of RM and FA, and a 1.5 M sodium trisilicate solution was added to the mix as an activator. The authors documented that the samples ’ compressive strength manifested an increase in its value with an increment in the Si to Al molar ratio. Hoang et al. (2020) Table A4 Studies done on utilizing RM and other materials for the synthesis of geopolymer and AAB [6,23,40 –42,50,56,57,60] Source of RM Country Material of study Variation studied Role of RM Reference Noranda Alumina LLC (Slurry form) USA RHA ≡Curing duration was varied (14, 21, 28, 35, 42, and 49 days). ≡RHA/RM ratio was varied (0.3, 0.4, 0.5 and 0.6). ≡The concentration of the NaOH solution was varied (2, 4, and 6 M). ≡The effect of change in RHA gradation was observed. Source material for geopolymer He et al. (2012) CHALCO Henan Branch, Zhengzhou China RM/NaOH ≡Alkali thermal pre-activation was performed on the RM at different dosages of Na2O (0, 5, 10, and 15% by weight). Source material for geopolymer Ke et al. (2014) – Turkey MK ≡The synergistic use of metakaolin and calcined RM was done to form the RM-MK-based geopolymer. The effect of incorporation of different amounts of RM used was studied. ≡Content of RM varied from 0 to 40% by weight, in steps of 10%. Source material for geopolymer Kaya et al. (2016) Chalco Co., Zhengzhou China SF ≡One-part geopolymer was synthesized, blending alkali thermal activated RM with SF in different proportions (RM was varied from 70 to 100% by weight). ≡Different water to solid ratio (0.45, 0.5, 0.55, 0.6, 0.65). ≡Sodium lignosulphonate was used to achieve a higher strength without any issues in workability and flowability. Source material for geopolymer Ye et al. (2016) ACG Alumina Plant in Fria (Slurry form) Guinea RHA ≡Curing duration was varied (15, 21, 28, 33, and 43 days). ≡RHA/RM ratio was varied (0.3, 0.4, 0.5 and 0.6). ≡Concentration of NaOH solution was varied (2, 2.5, 4, 4.5, 6, 6.5 and 8 M). Source material for geopolymer S˘ekou et al. (2017) Shangdong Aluminium Industry Co. LTD, Zibo China Ref. Mudstone, GGBS ≡Refused mudstone, GGBS, and RM were the precursor materials used to synthesize the geopolymer with standard sand as aggregate in the polymer composites. ≡Sodium silicate utilized as the alkaline activator, and the effect of variation in refused mudstone content and amount of alkali activator used was studied. ≡The quantity of refuse mudstone varied from 30 to 70%, while the alkali activator was varied from 10 to 30%. Source material for geopolymer Zhou et al. (2020) – Romania WG ≡RM was blended with waste glass in a 1:4 ratio and the effect of thermal treatment at varying temperature was studied. Source material for foamed geopolymer Badanoiu et al. (2015) Gardanne France WG ≡The soda lime glass to RM ratio as varied as 40/60, 45/55, 50/50. Source material for geopolymer Toniolo et al. (2018) Shandong branch Aluminum Corporation of China Limited, Zibo China CG ≡The effect of thermal activation and mechanical grinding preactivation on the geopolymer incorporating RM and Coal Gangue was studies. ≡The RM/CG ratio was varied (0.5, 0.6, 0.7, 0.8, 0.9). Source material for geopolymer Geng et al. (2017) Eti Seydis ¸ ehir Aluminum Factory, Konya Turkey MK ≡The author examined the use of RM-MK geopolymer, having sufficient strength at the same time to be used as a self-cleaning construction material. ≡The effect of anatase addition in the RM-MK geopolymer was investigated. Self-Cleaning construction materials Kaya- Ozkiper et al. (2021) Hejin city, Shanxi China MK ≡The authors have examined the effect of Na/Al ratio keeping everything else constant (Na/Al ˆ0.8, 0.9, 1.0, 1.1, 1.2, 1.3). ≡≡:RM/MK was kept constant at 7:3 while Si/Al was kept constant at 1.2. Source material for geopolymer Liu et al. (2020) – – MK ≡Microstructure and compressive strength variation in samples with different RM/MK ratio. ≡MK was substituted by 1/4, 1/6, 1/8, 1/10 and 1/12 of RM. Source material for geopolymer Hajjaji et al. (2013) – Turkey RM ≡RM as used as a substitute in metazeolite (MZ) with the substitution being 25%, 50%, 75% and 100% by wt. ≡The optimum amount of quartz sand was found by keeping RM and MZ ratio constant at 1:4. Source material for geopolymer Aygormez (2021) Trussing mining Australia RM, QS ≡A phase change material was synthesised in the form of geopolymer utilizing RM and quartz sand (QS). ≡Characterization of the PCM encapsulated red-mud geopolymer composites is conducted to determine the thermophysical and compressive strength. Application in building sector Afolabi et al. (2019) A. Kumar et al. Construction and Building Materials 302 (2021) 124170 8 noted that the samples prepared with the only RM cured at high temperature under atmospheric pressure were not hardened and lost strength upon soaking in water. In contrast, when autoclave treatment was provided, it attained a compressive strength value of 10.6 MPa. The compressive strength of the RM-FA-based AAB was seen to increase with the dwell time and temperature for both the heat-cured samples and samples provided with autoclave curing. Moreover, the increase in pressure under autoclave curing also enhanced the compressive strength and attained a maximum value of 20.1 MPa. It was also found that the compressive strength was closely correlated to the amount of active SiO2 available. Koshy et al. (2019) reported a strength of 5.7 MPa for the geopolymer syn- thesized using RM, FA, and coal gangue, which are suitable for use as soil stabilizers and road bases. From all the studies mentioned above, the strength of samples having adequate RM content and having the highest strength in each study is plotted in Fig. 10. The highest strength observed was that of a sample from the study Hu et al. (2018) , which had a strength of 43.1 MPa. The strength of all samples can be observed from Fig. 10, and the respective reference of the samples referred to can be found in Table 2. 3.2.2. Controlled low strength materials (CLSM) Choo et al. (2016) propounded the use of RM as an activator and discerned a linear increase in UCS of the samples with an increase in the RM content up to 60% RM by mass. The samples with 60% RM content attained a higher strength than pure RM and were suitable for use in construction works requiring low strength. Compared to the FA geo- polymer samples activated using NaOH, the samples activated with RM did not show much difference. The UCS of the sample with RM content of 40–60% was comparable to samples activated with 3–5% NaOH. This similarity inferred that RM could be used as a solid alkali activator to synthesize one-part alkali-activated FA. Yuan et al. (2019) also examined the use of RM as an alkali activator in the binary binder composed of FA and Portland Cement (PC). The substitution of FA with RM discerned an increase in the compressive strength values of the CLSM at all ages, which was evident from the fact that the strength of the sample with 60% RM attained a value of 5.3 MPa at the curing age of 28 days. Although both the studies utilized RM for the same purpose of alkali activation in the CLSM, the strength discerned by Yuan et al. (2019) was higher than that of the corresponding samples having the same RM content in work done by Choo et al. (2016) . This was because Yuan et al. (2019) used 10% Portland Cement (PC) in all the samples, apart from RM and FA, which gave it greater strength. Additionally, a polycarboxylate superplasticizer was also used to prepare the sample to have a slurry with a consistent spread flow. Table A5 Experiments conducted in the studies included in this report (Fresh, hardened, and Physical properties) [4–8,23 –57,59,60] CS FS BT ST WA FT SE FC AR DS UPV RV OM BSA VR PC AT SC BET LRT EC Reference ● Hu et al. (2018) ● ● ● Li et al. (2018) ● ● ● Krivenko et al. (2017) ● ● ● ● Chen et al. (2018) ● Ye et al. (2016) ● ● Zhou et al. (2020) ● ● Kumar et al. (2012) ● Kaya et al. (2016) ● ● ● ● ● Bayat et al. (2018) ● Nie et al. (2016) ● He et al. (2012) ● ● ● Singh et al. (2018) ● ● Singh et al. (2020) ● Hu et al. (2018) ● ● Kim et al. (2017) ● Yang et al. (2019) ● ● Kang et al. (2016) ● ● Zhang et al. (2016) ● Choo et al. (2016) ● ● ● Zhao et. (2019) ● ● ● ● ● ● ● Bayat et al. (2018) ● He et al. (2011) ● ● ● Hyeok-Jung et al. (2018) ● ● ● Zhang et al. (2020) ● ● Yuan et al. (2019) ● ● ● Lin et al. (2020) ● Hoang et al. (2020) ● Ke et al. (2014) ● S˘ekou et al. (2017) ● Ye et al. (2014) ● ● ● ● Lemougna et al. (2017) ● Badanoiu et al. (2015) ● Geng et al. (2017) ● Kaya- Ozkiper et al. (2021) ● ● Toniolo et al. (2018) Liang et al. (2021) ● ● Koshy et al. (2019) ● Liu et al. (2020) ● Hajjaji et al. (2013) ● ● ● ● Aygormez (2021) ● Afolabi et al. (2019) CS: Compressive Strength, FS: Flexural Strength, BT: Brazil Tensile Strength, ST: Slump test, WA: Water absorption test/Absorption coefficient, FT: Freeze-Thaw cycle test, SE: Setting time, FC: Flow cone method, AR: Abrasion Resistance, DS: Drying Shrinkage, UPV: ultrasonic pulse velocity, RV: Rebound value, OM: Optimum moisture content, BSA: Blaine Specific Surface Area, VR: Void ratio, PC: Permeability coefficient, AT: Adsorption test, SC: softening coefficient, BET: Brun- ner–Emmet –Teller method, LRT: Lime Reactivity Test, EC: Electrical Conductance A. Kumar et al. Construction and Building Materials 302 (2021) 124170 93.3. Water absorption Li et al. (2019) conducted the water absorption test and found that the water absorption of the MSWIFA-RM geopolymer was 16.02%, against water absorption of FA bricks, which generally ranges between 14.29% and 16.70%, thus highlighting that the durability of MSWIFA- RM geopolymer was comparable to that of the FA bricks. However, in work done by Kim et al. (2017) , water absorption of 18.50 wt% was observed in the trial bricks test. 3.4. Durability against Freeze-Thaw and elevated temperature Li et al. (2019) showed that the geopolymer ’s mass loss rate with MSWIFA and RM was 1.78%, which is quite less than the mass-loss rate of concrete, which is generally around 5%. Zhou et al. (2019) noted that the deterioration of strength was more pronounced in sam- ples cured for 28 days, particularly from 5 to 20F-T cycles, irrespective of the curing temperature. Despite the higher deterioration of strength in samples cured for 28 days, the samples cured at elevated temperatures after 50F-T cycles were still high (about 9 MPa). The study conducted by Yang et al. (2019) showed that on exposing the samples to a tem- perature of 160 •C, there was a significant increase in the UCS of the sample, which reduced when the temperature further increased. The unreacted FA particles contributed to strength development as they initially acted as fillers for pores and cracks in the material. 3.5. Flexural strength Zhang et al. (2016) noted that the sample ’s flexural strength decreased from about 7 MPa to about 4 MPa after soaking for one day and hardly declined further. The deterioration of the flexural strength of specimens was ascribed to the partial dissolution of geopolymer gels and the samples ’ significant moisture adsorption while soaking. 3.6. Durability against acid attack Zhang et al. (2016) examined the effect of sulfuric acid exposure on the geopolymer samples with RM and Class F FA. Control samples were also prepared using OPC. Regardless of the type of leachate used, UCS of the geopolymer samples after soaking for 28, 56, and 120 days were statistically maintained around 6–7 MPa. The geopolymer sample soaked in acid for 120 days had a statistical reduction of 30%, while that of the OPC sample was 14% (from 9.3 ±2.8 MPa to 7.3 ±1.2 MPa). It is noteworthy that although the strength reduction was higher for the geopolymer samples, the strength was still similar to the strength of the OPC sample. These results showed that the geopolymer samples had Table A6 Experiments conducted in the studies included in this report (Microstructural and other properties) [4–8,23 –57,59,60] XRD SEM XPS EDS HR TE SA FTIR MIP BJH TGA LT TC RE ET MSA IC SH RH Reference ● ● ● ● Hu et al. (2018) ● ● ● ● Li et al. (2018) ● ● ● ● Krivenko et al. (2017) ● ● ● Chen et al. (2018) ● ● ● ● Ye et al. (2016) ● Zhou et al. (2020) ● ● ● ● ● ● Kumar et al. (2012) ● ● ● Kaya et al. (2016) ● ● ● ● ● ● Bayat et al. (2018) ● ● ● Nie et al. (2016) ● ● ● He et al. (2012) ● ● ● ● ● ● Singh et al. (2018) ● Singh et al. (2020) ● ● ● ● ● ● Hu et al. (2018) ● ● ● ● ● ● Soon Yong et al. (2017) ● ● ● ● Yang et al. (2019) ● ● ● ● ● ● ● Kang et al. (2016) ● ● ● ● Zhang et al. (2016) ● ● Choo et al. (2016) ● ● Zhao et. (2019) ● ● ● ● Bayat et al. (2018) ● ● ● He et al. (2011) ● ● ● Hyeok-Jung et al. (2018) ● ● ● ● ● ● Zhang et al. (2020) ● ● ● ● ● Yuan et al. (2019) ● ● ● ● ● ● Lin et al. (2020) ● ● ● Ke et al. (2014) ● ● S˘ekou et al. (2017) ● ● ● ● ● Ye et al. (2014) ● ● ● ● Lemougna et al. (2017) ● ● ● Badanoiu et al. (2015) ● ● ● ● Geng et al. (2017) ● ● ● ● ● ● Kaya- Ozkiper et al. (2021) ● ● ● Toniolo et al. (2018) ● ● Liang et al. (2021) ● ● ● ● ● Koshy et al. (2019) ● ● ● ● ● Liu et al. (2020) ● ● ● ● Hajjaji et al. (2013) ● Aygormez (2021) ● Afolabi et al. (2019) XRD: X-ray Diffraction Analysis, SEM: Scanning Electron Microscopy, XPS: X-ray photoelectron spectroscopy, EDS: Energy Dispersive X-ray Spectroscopy/Electro- Magnetic Pulse, HR: High-Resolution Transmission Electron Microscopy, TE: Transmission Electron Microscopy, SA: Selected Area Electron Diffraction, FTIR: Fourier-transform infrared spectroscopy, MIP: Mercury Intrusion porosimetry, BJH: Barrett-Joyner-Hallenda method, TGA: Thermo Gravimetric Analysis/Differential Thermal Analysis, LT: Leaching test, TC: Thermal Conductivity, RE: Radiological Evaluation, ET: Efflorescence test, MSA: Mossbauer spectra analysis, IC: Isothermal Calorimeter, SH: Specific Heat, RH: Rheological properties A. Kumar et al. Construction and Building Materials 302 (2021) 124170 10similar durability to the control OPC samples after soaking them in the leachant. 3.7. Leaching test Leaching heavy metals from the geopolymer samples was measured using the toxicity leaching procedure (TCLP) [24,27,43] . The heavy metal leaching concentration from the geopolymer samples was less than the limits specified by USEPA in all three studies. Additionally, the heavy metal leaching discerned in the sample was much lower than that of the source materials. This decline in the heavy metal leaching indi- cated that the formation of hydration products in the geopolymer helped solidify the heavy metals . Hu et al. (2019) evaluated the ac- tivity of SiO2 and Al2O3 in RM, where the dissolution efficiencies of both Al2O3 and SiO2 were seen to increase with the temperature rise and had a peak at 800 •C. Correspondingly, the increase in the addition of Na2O was seen to effectuate the dissolution efficiency of Al2O3 and SiO2. In another work, Zhang et al. (2016) studied the leaching behavior of Al present in the geopolymer samples soaked in sulfuric acid and deionized water. It was found that the synthesized geopolymer showed better immobilization of As than the other heavy metals. Thus, the geopolymer also acts as a filter for heavy metals and prevents them from entering the groundwater. 3.8. Microstructural analysis 3.8.1. Thermogravimetric analysis (TGA)/Differential thermal analysis (DTA) In all the samples prepared in the study by Kim et al. (2017) , the results showed that the control sample had a larger amount of C-S-H than that of the sample with 20% RM incorporation, resulting in its higher strength. It was also observed that the use of 5% Na2O increased the DTA peak of C-S-H, and peaks of Ca(OH) 2 were reduced, which were in harmony with the compressive strength results. On the other hand, Yuan et al. (2019) reported an increase in the mass loss between 200 •C and 1000 •C, with an increase in the quantity of RM from 0 to 60%. Table A7 List of experiments along with the code used [1,8,23,26 –28,32,34 –40,44,45,47,48,56,59] Test Code Reference Binder Concrete Compressive Strength Test DSTU B.V 2.7–181:2009 DSTU – N B. V.2.7 –304:2015 Krivenko et al. (2017) ASTM C109/ C109Me11b BS EN 206–1:2000 Chen et al. (2018) ASTM C39 – Zhou et al. (2020) ASTM C109 – Bayat et al. (2018) IS: 4031 (Part 6) 1988 – Singh et al. (2018) – IS 3495: 1992, part 1 Singh et al. (2020) KS F 2405 – Kang et al. (2016) ASTM C109 – Choo et al. (2016) – ASTM C39 (2018) Bayat et al. (2018) ASTM C39/C39M – He et al. (2011) – KS F 2405 Hyeok-Jung et al. (2018) Flexural Strength ASTM C78 – Zhang et al. (2016) – ASTM C78 (2018) Bayat et al. (2018) Brazil Tensile Strength ASTM D 3976 – Zhou et al. (2020) Mini Slump Flow test ASTM C939 – Bayat et al. (2018) ASTM C1437 – Kim et al. (2017) – ASTM C143 (2010) (Slump flow test) Bayat et al. (2018) GB/T8077-2000 – Zhang et al. (2020) Flow cone method ASTM C939 – Bayat et al. (2018) GB/T8077-2000 – Lin et al. (2020) Spread flow GB/T 50,448 (Chinese Standard 2008a) – Yuan et al. (2019) Drying shrinkage test ASTM C596 – Bayat et al. (2018) – ASTM C596 (2009) Bayat et al. (2018) Setting time ASTM C191 – Bayat et al. (2018) ASTM Standard C191 – Lin et al. (2020) CNS786 & ASTM C191-01 – Lemougna et al. (2017) Void ratio – ASTM C1754 (2012) Chen et al. (2018) Permeability Coefficient – ACI 552R-08 Chen et al. (2018) Chloride ion permeability ASTM C1202 Liang et al. (2021) TCLP EPA 1311 – Kumar et al. (2012) – US EPA 1311 Kim et al. (2017) US EPA – Zhang et al. (2016) GB16889 (Chinese Standard 2008b) – Yuan et al. (2019) EN 12,457 Toniolo et al. (2018) Specific surface GB/T 8074 –2008 – Table A7 (continued ) Test Code Reference Binder Concrete Zhang et al. (2020) Lime Reactivity Test IS 1727: 1967 – Singh et al. (2018) Water absorption test – IS 3495 Singh et al. (2020) – ASTM C642 (2006) Bayat et al. (2018) Water absorption coefficient KS 2609 – Kang et al. (2016) Initial Rate of absorption test – ASTM C-67 –12 Singh et al. (2020) Radioactivity – ISO11929:2010 Krivenko et al. (2017) GB 6566 –2010 – Zhang et al. (2020) Optimum moisture content – KS F 2312 A Hyeok-Jung et al. (2018) Efflorescence test As/NZS 6656.6 – Kang et al. (2016) – As/NZS 6656.6 Hyeok-Jung et al. (2018) Fall cone method British Standard 1377 – Choo et al. (2016) A. Kumar et al. Construction and Building Materials 302 (2021) 124170 113.8.2. X-ray diffraction analysis The increase of RM incorporation in the RM-FA geopolymer resulted in the broad, amorphous peak ’s reduced intensity indicating amorphous gel formation . These amorphous gel peaks in the geopolymer were ranged between 200-360 2θ [4,5,24] . This was indicative that the crys- talline phases in the raw material did not participate in the geo- polymerization reaction but acted as inactive fillers in the geopolymer binder [5,23] . Table A8 Oxide composition analysis of materials (% by wt.) used in combination with RM for the synthesis of the geopolymer and AAB samples [4–6,8,23,25 –39,40,42 –49,51,52 –57] Material SiO 2 Al2O3 Fe2O3 CaO MgO SO3 Na2O K2O MnO TiO 2 LOI P2O5 Reference FA C 47.5 15.3 4.2 24 4.8 – – 0.6 – – – – He et al. (2011) F 60.48 28.15 4.52 1.71 0.47 0 0.14 1.41 – – 1.59 – Kumar et al. (2012) F 57.91 27.44 6.32 1.3 – 0.21 – – – – 2.15 – Zhang et al. (2016) FC 44.41 18.79 10.01 18.59 3.03 1.81 0.89 1.44 – – – – Nie et al. (2016) FF 43.16 22.8 23.6 3.29 0.73 1.2 0.75 1.62 – – – – 49.5 31.9 5.9 2.9 0.9 0.5 – – – – 2.5 – Kang et al. (2016) FA1 42.05 20.06 7.27 4.73 1.17 – 1.52 1.13 – 1.27 19.2 0.6 Choo et al. (2016) FA2 48.39 20.33 5.19 1.41 0.48 – 0.38 1.56 – 1.18 19.7 0.5 FA3 58.6 17.37 5.3 3.94 0.97 – 1.99 1.22 – 1.05 8 0.4 FA4 50.61 22.37 6.37 3.78 1.28 – 0.98 1.66 – 1.02 9.6 1.44 FA5 60.04 18.43 4.05 2.86 1.06 – 1.95 1.59 – 0.97 7.5 0.4 F 52.3 22.6 9.1 6.2 1.8 – 1.8 – – 1.3 1.7 1.9 Kim et al. (2017) FC 44.4 18.7 10 18.5 3 1.8 0.8 1.4 – – – – Hu et al. (2018) FF 43.1 22.8 23.6 3.2 0.7 1.2 0.7 1.6 – – – – F1 49.1 38.7 6 1.8 0.4 1.1 0.2 0.8 – – – – 59.65 27.64 3.3 1.84 0.64 – 0.4 1.46 – 1.42 2.56 – Hu et al. (2018) 61.54 25.37 6.73 3.1 0.73 0.62 0.97 – – – 0.39 – Singh et al. (2018) F 59.7 27.51 92.16 1.45 1.18 0.16 0.82 2.39 – – 2.66 – Yang et al. (2019) F 59.74 27.51 4.91 1.45 1.18 0.16 0.82 2.39 – – 2.66 – Zhao et. (2019) 51.7 28 3.7 3.94 – 0.54 0.33 1.54 – 1.4 7.36 – Yuan et al. (2019) 45.5 38.4 3.2 4.7 0.6 0.8 0.4 0.6 0 1.6 3.2 0.5 Koshy et al. (2019) F 47.74 35.36 7.02 4.2 – – 0.69 0.41 – 0.43 3.85 – Hoang et al. (2020) 49.26 28.64 7.89 2.63 1.56 – – 2.65 – – 6.03 – Lin et al. (2020) GGBS 32.85 14.96 0.55 37.64 9.27 – 0.41 0.57 – 0.98 0.62 – Ye et al. (2014) 33.6 14.5 0.7 43.5 5.2 1.4 – – – – 0.23 – Kang et al. (2016) 37.9 6.85 – 44.6 5.21 – – – 0.106 0.35 – – Krivenko et al. (2017) 34.21 15.32 0.63 37.15 9.34 – 0.39 0.41 0.59 0.8 – – Lemougna et al. (2017) 34.58 15.01 0.27 37.51 9.02 2.03 0.68 – – – 1.1 – Chen et al. (2018) 34.25 10.75 0.45 38.57 7.75 3.16 0.59 0.96 1.49 1.75 0.02 – Bayat et al. (2018) 34.25 10.75 0.45 38.57 7.75 3.16 0.59 0.96 1.49 1.75 0.02 – Bayat et al. (2018) 37.73 14.42 1.11 37.34 8.71 0.39 – – 1.41 – Singh et al. (2018) 20.5 12.1 0.55 57.2 5.05 0.83 0.36 0.58 – 1.6 2.83 – Zhang et al. (2020) 21.86 9.18 0.49 58.26 5.17 – – 0.63 3.86 – Lin et al. (2020) 34.35 17.54 0.77 32.74 8.79 – – – 0.52 0.67 – – Zhou et al. (2020) 31.35 18.65 0.57 34.65 9.31 – – – – – 0.7 – Liang et al. (2021) OPC 19.8 4.8 3 61.9 3.8 3 0.61 – – – 2.24 – Zhang et al. (2016) 21.7 5.7 3.2 63.1 2.8 2.2 – – – – 1.32 – Kang et al. (2016) 23.4 5.17 4.12 64.13 0.88 – – – – – – – Krivenko et al. (2017) 21.91 6.31 2.95 55.56 2.92 1.8 0.85 – – – 6.34 – Chen et al. (2018) 19.98 3.5 4.11 64.73 2.07 3.79 0.15 0.63 0.2 0.27 0.35 – Bayat et al. (2018) 19.98 3.5 4.11 64.73 2.07 3.79 0.15 0.63 0.2 0.27 0.35 – Bayat et al. (2018) 19.4 3.87 3.74 68.5 0 0.75 0.08 0.47 – – 1.12 – Yuan et al. (2019) RHA 91.5 – – – – – – 2.3 – – – – He et al. (2012) SF 94.43 0.27 0.14 0.28 0.29 0.24 0.25 0.3 – – 3.61 – Ye et al. (2016) MK 56.21 41.04 0.36 0.09 0.07 – – 0.46 – 1.15 – 0.06 Kaya et al. (2016) AASC 22.1 8.9 1.4 54.9 3.3 5.2 – – – – 2.23 – Kang et al. (2016) & Hyeok-Jung et al. (2018) Microsilica 99.3 0.2 0.09 0.1 0.04 – – 0.02 0.01 – Singh et al. (2018) RFM 60.62 15.92 5.39 – – – 1.96 2.35 – – – – Zhou et al. (2020) Waste Glass (WG) 70.5 3.2 0.42 10 – – 12 1 – – – – Toniolo et al. (2018) Coal Gangue (CG) 45.69 22.21 5.49 0.98 0.4 3.06 0.33 – – 0.54 19.82 0.12 Geng et al. (2017) MK 56.21 41.04 0.36 0.09 0.07 – – 0.46 – 1.15 – 0.06 Kaya- Ozkiper et al. (2021) MK 52.62 45.42 0.45 0.17 0.11 – 0.25 0.13 – 0.85 – – Liu et al. (2020) MK 54.4 39.4 1.75 0.1 0.14 – – 1.03 0.01 1.55 2.66 – Hajjaji et al. (2013) MZ 76.9 13.5 1.4 2 1.1 – 0.3 3.5 0.1 0.1 1.1 – Aygormez (2021) Quartz Sand (QS) 46.07 49.46 1.05 0.18 0.03 – 0.14 0.09 – 2.24 – 0.04 Afolabi et al. (2019) Table A9 Element composition analysis (% by wt.) of Municipal Solid Waste Incineration Fly Ash Elements Ca Si Al Fe Na Cl Mg K Zn Pb Cu Cr As Ni Ti LOI* Reference MSWIFA 43.4 1.71 0.45 0.98 4.43 20.5 1.64 4.14 0.59 0.17 0.1 0.08 0 0 – 17.9 Li et al. (2018) A. Kumar et al. Construction and Building Materials 302 (2021) 1241700123 4Kumar and Ku mar (2012)Chen et al . (2018)Jung et al . (2018)Yuan et al . (2019)Zhou et al. (2020)Hoang et al . (2020) Density (gm/ cm3)0 500 1000 1500 2000Bayat e t al. (2018a)Bayat e t al. (2018a) [RM- 550]Bayat e t al. (2018a) [RM- 750]Bayat e t al. (2018b) Fineness Modu lus (m2/Kg)Fig. 3.Density and fineness modulus of RM used in the respective studies [25,34 –37,40,43,45] 0 0.5 1 1.52 2.5 3 3.5 4 4.5He et al . (2012)Kang et al . (2016)Choo et al. (2016)Bayat e t al. (2018a)Bayat e t al. (2018a) [RM-550]Bayat e t al. (2018a) [RM-750]Singh et al . (2018) [URM]Singh et al . (2018) [PRM]Bayat e t al. (2018b)Singh et al . (2018) [PRM] Specific Gravity Fig. 4.Specific gravity of the RM particles used in the studies [23,26,34,36,38,44] 0 5000 10000 15000 20000 25000 30000 35000 40000 45000Kang et al . (2016)Choo et al. (2016)Krivenko et al . (2017)Lemougna et al . (2017)Chen et al . (2018)Singh et al . (2018) [URM]Singh et al . (2018) [PRM]Jung et al . (2018)Yuan et al . (2019)Zhou et al. (2020)Lin et al . (2020)Zhan g et al . (2020) [RM-392]Zhan g et al . (2020) [RM-268]Zhan g et al . (2020) [RM-104]Zhan g et al . (2020) [RM-39]Zhan g et al . (2020) [RM-Bul k] Surface area (m2/Kg) Fig. 5.Surface area of the RM particles used in the studies [8,24,26,32,35,37,38,40,44,45,48] A. Kumar et al. Construction and Building Materials 302 (2021) 124170 13Li et al. (2019) observed that mechanical activation enhanced the dissolution of the aluminite and katoite, and the activators in the geopolymer assisted in the formation of ettringite in the sample. Furthermore, calcium hydroxide was seen to either react with the so- dium silicate or enter the new aluminosilicate structure to form ettrin - gite. The only difference between the two geopolymers synthesized in the study conducted by Nie et al. (2016) , using RM before and after FGD, was the appearance of a few weak peaks between 240 and 350 (2θ) in the geopolymer synthesized using RM after FGD, that marked the presence of calcium sulphate. Kim et al. (2017) found that most of the detected phases were derived from raw materials. While the study conducted by Hu et al. (2019) depicted that at different dosages of Na2O, phases of nepheline and peralkaline aluminosilicate phases in RM weakened after hydration. Yuan et al. (2019) discerned that with an increase in the RM quantity, the intensities of hydration products like ettringite decreased, while the concentration of monocarbonate increased. The typical hydration product of PC, Portlandite, was not observed in the geopolymer attributable to the carbonation or addition of FA, utilizing which the pozzolanic reaction eventuates. 3.8.3. Fourier transform infrared spectroscopy Li et al. (2019) marked a reduction in the bands at 562 cm1 and 997 cm1 (asymmetric Al-O and Si-O-Si) when the geopolymer was provided mechanical activation. Si-O-Si (Al) bands in samples with no activator solution shifted from 997 cm1 to 668 cm1 and 682 cm1 after undergoing mechanical activation, having 14% sodium silicate. Kumar and Kumar (2013) discerned a similar absorption spectrum for all the samples with different RM content. However, the intensity of the band was maximum in samples with 10% RM. Hu et al. (2019) marked a shift of the asymmetric stretching vibrations of Si-O-Si and Si- O-Al from 995 cm1 (in pretreated RM) to 1002 cm1 in the geopolymer. This shifting of the asymmetric stretching vibrations of Si-O-Si and Si-O- Al was also reported by Yang et al. (2019) . Higher the intensity of the broad peak of Si-O-Si (Al), better the degree of geopolymerisation. Likewise, the stretching Fe-O band observed in raw and pretreated RM (at 442 cm1) disappeared, speculating Fe ions’ participation into so- dium aluminosilicate hydrate network structure. 3.8.4. Scanning Electron Microscopy/Electron Microprobe/Back-scatter Electron Compared to the geopolymers activated with NaOH, those activated with the composite solution discerned a denser structure, hence a higher compressive strength . Moreover, the geopolymers synthesized using Class C FA exhibited a more compact structure than those utilizing Class URM - Unproces sed RM, PRM – Pulverised RM 00.511.522.533.54 URM PRMLime r eactivity (MPa) 0 10 20 30 40 50 60Choo et al. (2016)Bayat e t al. (2018a)Bayat e t al. (2018a) [RM-550]Bayat e t al. (2018a) [RM-750]Bayat e t al. (2018b) LL (% )Fig. 6.Lime reactivity and Liquid limit of the RM particles used in the studies [26,34,36,38,53] 0 5 10 15He et al . (2011)He et al . (2012)Kang et al . (2016)Bayat e t al. (2018a)Bayat e t al. (2018a) [RM-550]Bayat e t al. (2018a) [RM-750]Singh et al . (2018) [URM]Singh et al . (2018) [PRM]Li et al. (2018)Yuan et al . (2019) pH Fig. 7.pH of the RM particles used in the studies [23,24,34,38,44,45,49] 0 50 100 150 200 250 300 350 400 450He et al . (2011)Ye e t al. (2016)Choo et al. (2016)Lemougna et al . (2017)Jung et al . (2018)Yuan et al . (2019)Hoang et al . (2020)Zhan g et al . (2020) [RM-268]Zhan g et al . (2020) [RM-39] d50 (μm) Fig. 8.Mean diameter of the RM particles used in the corresponding studies [23,25,26,32,37,42,45,48] A. Kumar et al. Construction and Building Materials 302 (2021) 124170 14F FA. According to Kumar and Kumar (2013) , the geopolymer samples having 10% RM content exhibited an increase in A-S-H gel formation. Simultaneously, the samples with 30% RM depicted the presence of small-sized microspheres, which were reported to be mainly hematite particles scattered all over the matrix. The binder with 50% FA and no alkali activator used exhibited a porous microstructure, ac- cording to Hu et al. (2019) , indicating insufficient geo- polymerization that was in harmony with its lower compressive strength values. A porous structure of the RM-FA geopolymer was also witnessed by He et al. (2012) . The elemental mapping of raw RM and RM ‡NaOH carried out on the BSE showed that the morphology of RM did not change with the addition of NaOH, although the elements Ca, Ti, and Fe was well dispersed when NaOH was added as compared to the sample without NaOH, where they were found to be agglomerated . The SEM image of inorganic polymers activated with 10% NaOH discerned the disso - lution of FA particles and the formation of the inorganic polymer gel. While the inorganic polymer activated with 20%, RM had an excess of unreacted FA particles surrounded by a limited polymer gel. However, there was an increase in the formation of inorganic polymer gel with an increase in RM content, suggesting that the polycondensation process enhanced as RM content increased . Li et al. (2019) noted that the microstructure of geopolymer was dense, and some needle-shaped particles were reported when both me- chanical activation and alkali activator were provided to the geo- polymer. The microstructure of the geopolymer with both mechanical and alkali activation was similar to the needle-shaped particles found in the ettringite. Thus, it was concluded that mechanical activation and addition of alkali activator lead to the formation of ettringite and other complex hydration products, consequently increasing the UCS of the geopolymer. Yuan et al. (2019) discerned a column-like ettringite in samples having no RM content with a relatively large diameter. With the increase in RM content, the needles became thinner while the crystals ’ length did not change. There was an increase in the formation of gel-like products when the addition of RM exceeded 40%. The RM in the geopolymer matrix influenced the mechanical properties of the CLSM majorly in two ways; it chemically accelerated the hydration of the binary binders and aided in the densification of the matrix. RM also helped raise the pH of the pore solution as it was rich in NaOH and Na2CO3, which further accelerated the dissolution of FA, resulting in a higher degree of polycondensation and subsequently enhanced strength. 3.8.5. Mercury Intrusion porosimetry (MIP) When 20% RM was incorporated in the reference sample (RM ˆ0 wt Table A10 Physical properties of the materials used in combination with RM for the synthesis of the geopolymer and AAB samples [8,23,24 –26,34 –39,40,43 –45,48,49] Material Density (g/ cm3) Specific Gravity Surface area (m2/Kg) Fineness Modulus (m2/Kg) Lime reactivity (MPa) LL (%) pH d50 (µm) Reference FA – – – – – – – 9.5 He et al. (2011) FA 1.89 – – – – – – – Kumar et al. (2012) FA1 – 2.25 5350 – – 37 – 12 Choo et al. (2016) FA2 – 2.13 2570 – – 33.8 – 21 FA3 – 2.28 3310 – – 36.5 – 22 FA4 – 2.34 4720 – – 28.6 – 15 FA5 – 2.22 2000 – – 35.4 – 20 MSWIFA – – – – – – 9.86 – Li et al. (2018) FA – 2.25 419 – 2.63 – 8.03 – Singh et al. (2018) FA 2.31 – 2030 – – – – 18.7 Yuan et al. (2019) FA – – 459 – – – – – Lin et al. (2020) FA 2.2 – – – – – – 48.2 Hoang et al. (2020) GGBS 2.91 – 425.4 – – – – – Kang et al. (2016) GGBS – – 1440 – – – – 17.08 Lemougna et al. (2017) GGBS – – 350–450 – – – – – Krivenko et al. (2017) GGBS – 2.88 – 430 – – – 14.1 Bayat et al. (2018) GGBFS 2.931 – 4245 – – – – – Chen et al. (2018) GGBS – 2.88 – 430 – – – – Bayat et al. (2018) GGBS 2.67 – 4210 – – – 12.94 Zhou et al. (2020) GGBFS – – 396 – – – – – Lin et al. (2020) AASC 2.83 – 405.8 – – – – – Kang et al. (2016) AAS 2.83 – 405.8 – – – – – Hyeok-Jung et al. (2018) Slag – – 423.75 – – – – – Zhang et al. (2020) OPC 3.15 – 314.4 – – – – – Kang et al. (2016) OPC – – 350–450 – – – – – Krivenko et al. (2017) OPC – 3.15 – 320 – – – – Bayat et al. (2018) OPC 3.15 – 314.4 – – – – – Hyeok-Jung et al. (2018) PC 3.101 – 3264 – – – – – Chen et al. (2018) PC – 3.15 – 320 – – – – Bayat et al. (2018) PC 3.1 – 1190 – – – – 10.1 Yuan et al. (2019) RHA – 2.06 – – – – – 32 He et al. (2012) RHA (f) – – – – – – – 25 Refuse mudstone 2.23 – 4280 – – – – 11.66 Wei Zhou et al. (2020) MSWIFA : Municipal Solid Waste Incineration Fly Ash, RHA: Rice Husk Ash, RHA (f): Rice Husk Ash of with gradation, OPC: Ordinary Portland Cement, PC: Portland Cement, AAS: Alkali Activated Slag, AASC : Alkali Activated Slag Cement, GGBS : Ground Granulated Blast Furnace Slag, GGBFS : Granulated Blast Furnace Slag, FA: Fly Ash A. Kumar et al. Construction and Building Materials 302 (2021) 124170 15%) synthesized by Kim et al. (2017) , the pores in the range of 0.1 to 3 µm were eliminated by the filling effect caused by the fine particles present in RM at three days. Meanwhile, the sample ’s total porosity was greater than that of the reference sample at 3 and 28 days of curing. Additionally, the introduction of NaOH to the RM incorporated sample resulted in a considerable decrease in the pore sizes. Yuan et al. (2019) showed that after a curing age of 28 days, a decrease in the sample ’s porosity was apparent, increasing RM content up to 40%. Still, a further increase resulted in increased porosity. 4.Red mud incorporated with ground Granulated Blast Furnace slag 4.1. Fresh properties Bayat et al. (2018a) and Lemougna et al. (2017) observed an increase in the setting time of the AAB and geopolymer as the quantity of RM increased, i.e., the RM acted as a retardant in the geo- polymer mix. The lowest initial and final setting times were displayed by samples containing no RM. Bayat et al. (2018a) noted that all the samples had a significantly larger spreading diameter higher than that for the OPC sample, except for the samples containing RM calcined at 750 •C. The sample with RM calcined at 750 •C was the only sample having a smaller spreading diameter than the OPC sample, which was hardly influenced by time. As the RM content increased in the mix, the rapid reduction in consistency decreased due to the reduction in the free Ca from slag. The consistency loss in the sample with 40% RM was the least. Thus, the addition of RM mitigated the slump loss and made the paste more stable during the initial mixing hours. The sample-based on AAS had the highest slump value that decreased with RM while the cohesiveness increased. Zhang et al. (2020) reported a decline in the fluidity of the RM- Table 1 Oxide composition analysis of RM (% by Wt.) used with materials considered in the study [4–6,8,23,25 –28,31 –40,42 –57]. RM with SiO 2 Al2O3 Fe2O3 CaO MgO SO3 Na2O K2O MnO TiO 2 LOI P2O5 NaAlO 2 FA 25.5 26.4 23.2 1.3 0.1 0.8 14.9 0.2 – – – – – Hu et al. (2018) 29.2 15.2 31.5 4.5 0.2 – 3.1 – – – 10.2 – – Kumar and Kumar (2012) 25.58 26.4 23.26 1.33 0.15 0.8 14.98 0.21 – – – – – Nie et al. (2016) [RM1] 26.15 27.23 23.98 1.13 0.17 2.73 11.45 0.23 – – – – – Nie et al. (2016) [RM2] 19.43 17.89 33.88 2.66 0.1 – 12.18 0.37 – 0.72 12.84 – – Hu et al. (2018) 18.99 25.5 31.2 2.39 – 0.35 13.85 – – 6.79 11.17 – – Kim et al. (2017) 22.82 15.06 17.34 12.24 0.27 – 4.37 1.19 0.36 3.43 15.75 2.43 – Zhang et al. (2016) 18.3 21.6 29.45 2.76 – – 12.02 0.03 – 6.26 9.1 0.13 – Choo et al. (2016) 1.2 14 30.9 2.5 – – – – 1.7 4.5 – – 23 He et al. (2011) 21.4 22.7 8.31 16.5 – 0.39 11.5 0.42 – 3.99 13.4 – – Yuan et al. (2019) 7.4 13.65 56.05 3.1 – – 3.63 0.25 – 0.15 12.5 – – Hoang et al. (2020) 11 21.8 41.0 1.6 0.1 0.5 8.0 0.1 0.0 7.0 8.7 0.2 – Koshy et al. (2019) GGBS 4.8 12.9 48.6 10.1 – – – – – 5.3 – – – Krivenko et al. (2017) 21.43 22.72 9.98 16.49 – – 11.51 0.42 – – 13.41 – – Chen et al. (2018) 13.26 15.41 20.54 19.87 1.5 0.54 5.87 0.73 – 4.97 16.32 – – Bayat et al. (2018a) 38.3 16.1 22.8 3.4 0.2 – 10 0.4 – – – – – Kang et al. (2016) 13.26 15.41 20.54 19.87 1.5 0.54 5.87 0.73 – 4.97 16.32 – – Bayat et al. (2018b) 38.8 16.1 22.8 3.4 0.2 – 10 – – – – – – Jung et al. (2018) 20.15 21 24.9 7.3 0.17 0.28 9.7 0.11 – 6.19 10.2 – – Zhang et al. (2020) [RM-Bulk] 17.6 29.1 30.1 3.6 – – 19.6 – – – – – – Zhang et al. (2020) [RM-392] 15.4 27.4 30.1 3.3 – – 23.7 – – – – – – Zhang et al. (2020) [RM-268] 16.98 13.35 7.43 30.29 1.5 – 2.82 0.38 – 2.29 24.96 – – Liang and Ji (2021) GGBS 14.2 28.4 26.8 1.2 – – 29.4 – – – – – – Zhang et al. (2020) [RM-104] 13.9 28.2 27 1 – – 30.1 – – – – – – Zhang et al. (2020) [RM-63] 12.6 26.6 30.5 1.3 – – 29 – – – – – – Zhang et al. (2020) [RM-39] 20.38 24.5 9.48 12.86 1 – 11.46 0.88 – 2.92 15.4 – – Ye et al. (2014) 9.39 18.47 33.99 14.19 0.32 – 5.11 0.1 0.091 5.42 – – – Lemougna et al. (2017) FA � GGBS 9.93 18.1 42.9 2.3 – – 5.58 – – – 10.5 – – Singh et al. (2018) 6.86 17.38 13.2 40.32 0.74 – – 0.26 – – 18.14 – – Lin et al. (2020) RHA 1.2 14 30.9 2.5 – – – – 1.7 4.5 – – – He et al. (2012) NaOH 20.4 24.5 9.5 12.9 – 0.7 11.5 0.9 – – 15.4 – – Ke et al. (2014) SF 20.38 24.5 9.48 12.86 1 0.67 11.46 0.88 – 2.92 15.4 – – Ye et al. (2016) MK 11.67 14.02 37.1 1.1 0.23 – 9.39 0.31 – 5.78 – 0.03 – Kaya et al. (2016) [Raw] MK 12.59 16.85 37.45 1.44 0.2 – 10.55 0.31 – 5.94 – – – Kaya et al. (2016) [Calcined] MK 5.54 18.8 51.8 3.27 – 11.2 6.84 0.08 0.04 0.23 1.90 – – Hajjaji et al. (2013) MK 21.05 27.38 6.42 14.91 0.53 – 11.86 0.77 – 4.04 8.78 – – Liu et al. (2020) MK 11.67 14.02 37.1 1.1 0.23 – 9.39 0.31 – 5.78 – 0.03 – Kaya- Ozkiper et al. (2021) MK 12.59 16.85 37.45 1.44 0.2 – 10.55 0.31 – 5.94 – – – Kaya- Ozkiper et al. (2021) [Calcined] RFM 14.48 18.85 28.37 15.43 1.6 – 0.63 – 0.41 – – – – Zhou et al. (2020) MZ 14.77 25.18 34.90 1.81 0.26 – 9.13 0.30 0.07 0.02 8.34 – – Aygormez (2021) QS 12.52 12.70 49.72 8.81 0.54 – 2.53 0.34 – 9.56 – 1.06 – Afolabi et al. (2019) WG 5.21 15.21 52.94 2.95 – – 2.40 0.63 – – – – – Toniolo et al. (2018) CG 12.83 20.26 33.39 0.87 0 0.6 10.85 – – 7.35 12.28 0.17 – Geng et al. (2017) 122211Dried Pulverised Calcined Alkali-therm al activatio n Washed RM after FGD Fig. 9.Number of studies with different pretreatment provided to RM. A. Kumar et al. Construction and Building Materials 302 (2021) 124170 16slag paste with an increase in the mean particle size of RM. The maximum fluidity was that of the samples having a bulk particle size ascribed to the good particle size distribution in the geopolymer paste. Bayat et al. (2018a) discerned that the samples ’ efflux time increased with increment in RM addition, leading to a reduction in fluidity. 4.2. Compressive strength 4.2.1. Binders Chen et al. (2019) discerned a decrease in strength with an increase in RM content for both the binder and concrete samples (Fig. 11). For mix designations used in the study , refer to Table 3. The difference in the strength of samples with washed RM (A1-2 or A2- 2) and the samples with untreated RM (A1-1 or A2-1) validated RM’s role as an activator. Ye et al. (2014) pointed out that the geopolymer binders with 10% RM had a strength higher than the binders based on GGBFS at the same curing age. The mix having an RM/GGBS ratio of 5:5 rendered a 28 days strength of 49.2 MPa, which was sufficient for its potential application in construction materials. Lemougna et al. (2017) noticed a reduction in strength with an increase in the quantity of RM in the system. However, the decrement in strength was not so significant to an addition of 50% RM, which suggested that a geopolymer could be prepared using 50% RM to have sufficient strength for practical appli - cations. The optimum curing temperature for the sample with 50% RM was 40 •C, which exhibited the highest strength. 4.2.2. Mortar The 7-day compressive strength of the mortar sample prepared by Bayat et al. (2018) was found to be the maximum among the other papers studied, as shown in Fig. 12. The sample with raw-RM content of 40% (AASR40) and sample with RM thermally treated at 750 •C (AASR40-750) had a 19% higher strength than the OPC mortar. The final compressive strength and strength development rate was higher in AASR40-550 than in AASR40, which suggests that the thermal treat- ment of RM is an effective way for yielding geopolymers with good strength . Krivenko et al. (2017) synthesized the AAB using 60% RM, 30% GGBS, and 10% OPC, all by mass. After curing for two days, the cement-sand mortar strength (1:3) reached a value of 6.25 MPa which further increased to 30 MPa and 60 MPa at the end of 7 days and 28 days, respectively, which can be used in places requiring high strength materials. The AAB samples synthesized in work done by Jung et al. (2018) used alkali-activated slag (AAS) and RM. The 28-day strength of the AAS samples (18.9 –27.0 MPa) was similar to that of the OPC samples (18.4 –28.8 MPa), as shown in Fig. 13. The 28-day compressive strength of the samples prepared by Kang & Kwon (2017) decreased as the RM content increased. The reduction in strength of the samples was ascribed to reducing binder content due to an increase in RM. Lemougna et al. (2017) studied sand addition on the geo- polymer ’s dry and wet compressive strength, where slag was replaced by sand keeping the RM quantity constant at 50%. The greater value of the geopolymers ’ dry compressive strength than its counterpart was ascribed to converting a few Si-O bonds into Si-OH bonds after being immersed in water, thus weakening the structure. The addition of sand resulted in a decrease in the geopolymer ’s compressive strength, which ought to be due to the absence of solid interfacial bonding between sand particles and the geopolymer matrix. To produce lightweight materials, 05101520253035404550Compressive strength (MPa )3d 7d 14d 28d 16hFig. 10.Compressive strength of the samples utilizing RM and FA [4,5,23 –25,27,31,43] . Table 2

materials Article Mechanical and Fracture Properties of Fly Ash Geopolymer Concrete Addictive with Calcium Aluminate Cement Yamin Wang1,2, Shaowei Hu1,2,* and Zhen He1 1School of Water Resources and Hydropower Engineering, Wuhan University, Wuhan 430072, China; wangyam@whu.edu.cn (Y.W.); hezhen@whu.edu.cn (Z.H.) 2Department of Materials and Structural Engineering, Nanjing Hydraulic Research Institute, Nanjing 210024, China *Correspondence: hushaowei@nhri.cn; Tel.: +86-025-85829601 Received: 7 August 2019; Accepted: 11 September 2019; Published: 15 September 2019 /gid00030/gid00035/gid00032/gid00030/gid00038/gid00001/gid00033/gid00042/gid00045 /gid00001 /gid00048/gid00043/gid00031/gid00028/gid00047/gid00032/gid00046 Abstract: In this paper, the mechanical and fracture properties of fly ash geopolymer concrete (FAGC) mixed with calcium aluminate cement (CAC) were explored. Fly ash was partially replaced by CAC with 2.5%, 5% and 7.5%. The results exhibit that the mechanical and fracture behaviors of FAGC are significantly influenced by CAC content. Based on the formation of more aluminum-rich gels, C-(A)-S-H and C-S-H gels, with the increase of CAC content, the compressive strength, splitting tensile strength and elastic modulus improved. Meanwhile, the peak load and e ective fracture toughness show a monotone increasing trend. In addition, because C-S-H gels absorbed more energy, the fracture energy of FAGC increases. The maximal peak load, double-K fracture toughness and fracture energy reached up to1.79 kN, 4.27 MPam0.5, 10.1 MPam0.5and 85.8 N /m with CAC content of 7.5%, respectively. Keywords: fly ash geopolymer concrete; calcium aluminate cement; mechanical properties; fracture properties 1. Introduction Globally, concrete is the most widely used building material; However, the production of one ton of cement will emit 600–800 kg of CO 2, which accelerates global warming [ 1]. It is necessary for concrete to switch over from Portland cement to a greener and environmentally friendly alternative binder with desirable mechanical and durability properties [ 2]. As one of the novel types of aluminosilicate inorganic polymer materials, geopolymer is produced by the reaction of solid aluminosilicate source materials and high concentration of alkali activator, which was described by Davidovits in the 1970s [ 3]. Through a large number of studies, fly ash geopolymer was reported to have many superior mechanical performances, such as high compressive strength, negligible shrinkage, good resistance to acid, and thermal stability, etc. [ 4–8]. In order to expand the application range of fly ash geopolymer, a large number of studies have been conducted to improve the performance of geopolymer by adjusting the chemical composition of the aluminosilicate raw materials. The preferred methods are to incorporate calcium-rich or silica-rich source materials, such as Portland cement, blast furnace slag, silica fume, rice husk ash, metakaolin, and nano-particles etc. [9–12]. In fact, not only silicon-rich materials and calcium-rich materials can improve the properties of geopolymers, but also aluminum-rich materials can achieve this goal. For alkali activator materials, Criado et al. [ 13] claimed that the amount of Al(OH)4tetrahedral groups significantly a ected the number of aluminosilicate gels, owing to Al(OH)4tetrahedral groups can attract positive charges. Therefore, under the environment of alkali activator, increasing the content of reactive aluminum can Materials 2019 ,12, 2982; doi:10.3390 /ma12182982 www.mdpi.com /journal /materials Materials 2019 ,12, 2982 2 of 13 increase the content of Al(OH)4in geopolymer mixes. As a kind of special cement, calcium aluminate cement (CAC) consists of various aluminates with the content ranging from 40% to 90% [ 14]. CAC has many excellent properties, such as high early strength, high temperature resistance and wide corrosion resistance [ 15,16]. The main components of CAC contain CaO
Al2O3(CA), CaO
2Al 2O3(CA 2) and a portion of 12CaO
7Al 2O3(C12A7) . The hydration products vary with curing temperature and humidity. Under low temperature (<20C) conditions, the main hydration product is CaO
Al2O3
10H 2O (CAH 10) , once the temperature is greater than 20C, the main hydration products are 2CaO
Al2O3
8H2O (C 2AH 8) and Al(OH) 3(AH 3) . According to the literature [ 20], when the temperature and humidity are increased, CAH 10and C 2AH 8, which are metastable phases, will transform into stable 3CaO
Al2O3
6H2O (C 3AH 6) and AH 3. In an alkaline environment, AH 3will react with OHto form Al(OH)4. Building on these results, the CAC can be used as an aluminum-rich material to improve the properties of fly ash geopolymer. Very few studies have investigated the e ect of CAC as the aluminum-rich material on the alkali activator materials. Arbi et al. [ 22] investigated the blast furnace slag and natural rock mixed with CAC. The results showed that the former system obtained C-(A)-S-H gel and sulfoaluminate in the medium alkaline medium. Vafaei and Allahverdi [ 14] studied the influence of CAC on natural pozzolan. They found that CAC increased the aluminum content of the binder and promoted the geopolymerization to form more aluminosilicate gels. Reig et al. [ 23] explored the e ect of red clay brick waste with CAC. The results identified that CAC accelerated the activation process of red bricks and the compressive strength achieved 50 MPa. Tao et al. [ 24] explored the mechanical properties of fly ash geopolymer concrete (FAGC) with CAC. The results identified that the optimal replacement rate of CAC is 10% on the compressive strength of 7 days and 28 days. So far, fly ash geopolymer concrete as the most widely explored alkali-activated material, there are not enough studies on the properties of FAGC combined with CAC under the high temperature curing condition. The combination of FAGC and CAC has been, for obvious reasons, of particular interest. The CAC, which will refer to as the aluminum-rich material, contains reactive aluminum and calcium ions, both of which can be used to improve the properties of FAGC. As a kind of new composite material, appropriate researches on the basic properties of FAGC are required to provide support for the engineering applications, among which mechanical properties and the fracture characteristics are particularly important. Above all, the aim of this study is to explore the influence of CAC content on the mechanical and fracture properties of FAGC cured at 75C, including a portion of CAC replacement percentage (2.5%–7.5%). For mechanical properties, the compressive strength, splitting tensile strength and elastic modulus were selected for testing. In addition, for fracture properties, by means of three-point bending test, the load-crack mouth opening displacement curve, the peak load, the fracture energy and the double-K fracture toughness were obtained. 2. Materials and Methods 2.1. Material In this study, the fly ash obtained from the Ningdong power plant in China were used. CAC purchased in Jianai Special Aluminates Co., Ltd. (Zhengzhou, China). Fly ash and CAC were analyzed by XRF to determine their chemical compositions, which are presented in Table 1. The main compositions of fly ash are SiO 2and Al 2O3, and the main compositions of CAC are Al 2O3and CaO. The morphology of fly ash and CAC were observed by scanning electron microscope (SEM) (Quanta, Philips corporation, Eindhoven, Holland), as shown in Figure 1. The alkali activator consists of sodium silicate (Na 2SiO 3) (composed of 25.89% SiO 2and 8.11% Na 2O by mass), with a SiO 2/Na2O molar ratio of 3.3, as well as sodium hydroxide (NaOH) pellets (96%). The alkaline activator is prepared by mixing sodium hydroxide and sodium silicate [ 25]. River sand and gravel were used as fine aggregate Materials 2019 ,12, 2982 3 of 13 and coarse aggregate in accordance with the Chinese standard JGJ 52-2006 [ 26]. For river sand, the density was 2645 kg /m3, the sand was 2645 kg /m3, absorption was 2.9%, and fineness modulus was 2.63, respectively. For gravel, the bulk density was 2530 kg /m3, and water absorption was 1.83%, respectively . Table 1. Chemical composition of materials. Materials SiO 2 Al2O3 Fe2O3 CaO MgO K 2O Na 2O SO 3 FA 49.37 32.14 5.20 4.77 1.80 1.59 1.30 1.03 CAC 7.09 49.67 1.99 36.69 0.367 0.56 0.12 0.69 Materials 2019 , 12, x FOR PEER REVIEW 3 of 13 river sand, the density was 2645 kg/m3, the sand was 2645 kg/m3, absorption was 2.9%, and fineness modulus was 2.63, respectively. For gravel, the bulk density was 2530 kg/m3, and water absorption was 1.83%, respectively . Figure 1. SEM images of ( a) fly ash and ( b) calcium aluminate cement (CAC) particles. Table 1. Chemical composition of materials. Materials SiO 2 Al 2O3 Fe 2O3 CaO MgO K 2O Na 2O SO 3 FA 49.37 32.14 5.20 4.77 1.80 1.59 1.30 1.03 CAC 7.09 49.67 1.99 36.69 0.367 0.56 0.12 0.69 2.2. Specimen Preparation T h e m i x t u r e s w e r e m i x e d i n a l a b o r a t o r y m i x e r . T o p r o d u c e a l k a l i a c t i v a t o r , N a O H w a s dissolved in distilled water and stirre d uniformly, and then mixed with Na 2SiO 3 solution. Firstly, gravel, river sand and fly ash was poured into a laboratory mixer, and stirred for 4 min. After the dry mixing, the alkaline activator, extra water and sodium gluconate (sg) were then added in the mix gradually, and the wet mixed for 2 min. The fresh concrete mixture was cast in the molds of cubes and beams, without any compaction to fill spaces of molds by its own weight. The molds were then stored in a curing box at the temperature of 75 °C for 16 h. Before testing, the samples were removed from molds after curing and left in the room with the temperature varying between 18 and 23 °C. The mixes description and the proportions of ingredients are as per Table 2. Table 2. Details of mix proportion (kg/m3). Mix Fly Ash NaOH Na 2SiO 3 Fine Aggregate Coarse Aggregate CAC Water sg M0 563.54 44.51 124.55 732.60 599.40 0 132.43 8.45 M1 549.45 44.51 124.55 732.60 599.40 14.09 132.43 8.45 M2 524.09 44.51 124.55 732.60 599.40 28.18 132.43 8.45 M3 521.27 44.51 124.55 732.60 599.40 42.26 132.43 8.45 2.3. Compressive Strength, Splitting Tensile Strength and Elastic Modulus To investigate the influence of CAC content on the mechanical properties of FAGC, the compressive strength, splitting tensile strength and elastic modulus were determined. The test specimen size and test procedure were carried out in accordance with Chinese standard GB/T 50081- 2002 . The 100 × 100 × 100 mm3 cube specimen was selected for the compressive strength test, the 150 × 150 × 150 mm3 prism specimen was used for the splitting tensile strength test, and the 150 × 150 × 300 mm3 prism specimen was used for the elastic mo dulus test. The loading rates of compressive strength and splitting tensile strength were set to 2.4 kN/s and 50 N/s, respectively. Figure 1. SEM images of ( a) fly ash and ( b) calcium aluminate cement (CAC) particles. 2.2. Specimen Preparation The mixtures were mixed in a laboratory mixer. To produce alkali activator, NaOH was dissolved in distilled water and stirred uniformly, and then mixed with Na 2SiO 3solution. Firstly, gravel, river sand and fly ash was poured into a laboratory mixer, and stirred for 4 min. After the dry mixing, the alkaline activator, extra water and sodium gluconate (sg) were then added in the mix gradually, and the wet mixed for 2 min. The fresh concrete mixture was cast in the molds of cubes and beams, without any compaction to fill spaces of molds by its own weight. The molds were then stored in a curing box at the temperature of 75C for 16 h. Before testing, the samples were removed from molds after curing and left in the room with the temperature varying between 18 and 23C. The mixes description and the proportions of ingredients are as per Table 2. Table 2. Details of mix proportion (kg /m3). Mix Fly Ash NaOH Na 2SiO 3 Fine Aggregate Coarse Aggregate CAC Water sg M0 563.54 44.51 124.55 732.60 599.40 0 132.43 8.45 M1 549.45 44.51 124.55 732.60 599.40 14.09 132.43 8.45 M2 524.09 44.51 124.55 732.60 599.40 28.18 132.43 8.45 M3 521.27 44.51 124.55 732.60 599.40 42.26 132.43 8.45 2.3. Compressive Strength, Splitting Tensile Strength and Elastic Modulus To investigate the influence of CAC content on the mechanical properties of FAGC, the compressive strength, splitting tensile strength and elastic modulus were determined. The test specimen size and test procedure were carried out in accordance with Chinese standard GB /T 50081-2002 [ 28]. The 100 100100 mm3cube specimen was selected for the compressive strength test, the 150 150150 Materials 2019 ,12, 2982 4 of 13 mm3prism specimen was used for the splitting tensile strength test, and the 150 150300 mm3 prism specimen was used for the elastic modulus test. The loading rates of compressive strength and splitting tensile strength were set to 2.4 kN /s and 50 N /s, respectively. 2.4. Three Point Bending Test To determine fracture properties of the FAGC specimens, the three-point bending test was conducted reference to the RILEM guidelines [ 29,30]. To explore the fracture behaviors of FAGC, the size of the notched beam used in three-point bending test is 80 mm 80 mm400 mm. In the forming process of the beam, the notch is made of steel plate. The parameters of the precast crack are set as follows: The thickness is 3 mm, depth is 32 mm, and the tip angle is 15. Four groups of specimens contain sixteen specimens. The beam is placed on the ball bearing support in the form of notched face down, with a span of 320 mm (Figure 2). The test was carried out on a HUALONG 200C electronic universal testing machine (Shanghai, China) with a load capacity of 20 tons. During the three-point bending test, the loading rate is set to 0.5 mm /min, and a clip gauge was installed at the mouth of the notch to record the data of crack mouth opening displacement (CMOD). Materials 2019 , 12, x FOR PEER REVIEW 4 of 13 2.4. Three Point Bending Test To determine fracture properties of the FAGC specimens, the three-point bending test was conducted reference to the RILEM guidelines [29,30]. To explore the fracture behaviors of FAGC, the size of the notched beam used in three-point bend ing test is 80 mm × 80 mm × 400 mm. In the forming process of the beam, the notch is made of steel plat e. The parameters of the precast crack are set as follows: The thickness is 3 mm, depth is 32 mm, and the tip angle is 15 °. Four groups of specimens contain sixteen specimens. The beam is placed on th e ball bearing support in the form of notched face down, with a span of 320 mm (Figure 2). The test was carried out on a HUALONG 200C electronic universal testing machine (Shanghai, China) with a load capacity of 20 tons. During the three-point bending test, the loading rate is set to 0.5 mm/min, and a clip gauge was installed at the mouth of the notch to record the data of crack mouth opening displacement (CMOD). Figure 2. Schematic diagram of the three-point bending test. 3. Results and Discussion 3.1. Compressive Strength Figure 3 and Table 3 express the effect of CA C content on compressive strength developments of FAGC. Obviously, the CAC content plays a major role in the compressive strength of FAGC. It can be seen from Figure 3 that the compressive strength of FAGC improves from 33.45 MPa to 41.02 MPa, when the CAC content grows from 0 to 7.5%. In addi tion, the value of compressive strength is 36.79 MPa with 2.5% and 38.53 MPa with 5%, respective ly. The value of compressive strength with CAC content of 2.5%, 5% and 7.5% is 9.99%, 15.19% and 22.63% higher than that with plain geopolymer concrete, respectively. Taking into account the CAC content, increasing CAC content always resulted in the enhancement of compressive strength. Based on previous lite rature [31,32], the polycondensation reaction of geopolymer wi ll produce much more Si-O-Al bonds, which significantly affect the development of compressive strength. The aluminum plays an important role in the polycondensation of geopolymer; what is more, the geopolymer mechanical properties are affected by the calcium content. As a good source of reactive aluminum and additional calcium, the CAC can be taken up into the geopolymerization pr ocess, the high amount of aluminum promotes the formation of more aluminum-rich gels, the addi tional calcium favors the formation of aluminum- substitute calcium silicate hydrate (C-(A)-S-H) and C-S-H gels , which may further promote the properties of geopolymers. In su mmary, the addition of CAC has positive influence on the strength development of FAGC. Figure 2. Schematic diagram of the three-point bending test. 3. Results and Discussion 3.1. Compressive Strength Figure 3 and Table 3 express the e ect of CAC content on compressive strength developments of FAGC. Obviously, the CAC content plays a major role in the compressive strength of FAGC. It can be seen from Figure 3 that the compressive strength of FAGC improves from 33.45 MPa to 41.02 MPa, when the CAC content grows from 0 to 7.5%. In addition, the value of compressive strength is 36.79 MPa with 2.5% and 38.53 MPa with 5%, respectively. The value of compressive strength with CAC content of 2.5%, 5% and 7.5% is 9.99%, 15.19% and 22.63% higher than that with plain geopolymer concrete, respectively. Taking into account the CAC content, increasing CAC content always resulted in the enhancement of compressive strength. Based on previous literature [ 31,32], the polycondensation reaction of geopolymer will produce much more Si-O-Al bonds, which significantly a ect the development of compressive strength. The aluminum plays an important role in the polycondensation of geopolymer; what is more, the geopolymer mechanical properties are a ected by the calcium content. As a good source of reactive aluminum and additional calcium, the CAC can be taken up into the geopolymerization process, the high amount of aluminum promotes the formation of more aluminum-rich gels, the additional calcium favors the formation of aluminum-substitute calcium silicate hydrate (C-(A)-S-H) and C-S-H gels [ 30], which may further promote the properties of geopolymers. In summary, the addition of CAC has positive influence on the strength development of FAGC. Materials 2019 ,12, 2982 5 of 13 Materials 2019 , 12, x FOR PEER REVIEW 5 of 13 Figure 3. The compressive strength of fly ash geopol ymer concrete (FAGC) with different CAC content. Table 3. Results of mechanical properties of FAGC. Mix CAC Content (%) Compressive Strength (MPa) Splitting Tensile Strength (MPa) Elastic Modulus (GPa) M-0 0 33.45 2.47 11.84 M-1 2.5 36.79 2.59 14.79 M-2 5 38.53 2.84 15.44 M-3 7.5 41.02 2.91 16.93 To observe the microstructure of FAGC with different CAC contents, SEM and EDS analysis were performed. The results are shown in Figure 4 and Table 4. In Figure 4a, for FAGC with 0% CAC, there are many unreacted FA particles, what is more, the improper bonding of FA particles with binders can be observed, which result in the generation of weak points and decrease compressive strength . With the increase of CAC conten t, the unreacted FA particles to be less, the aluminosilicate gels to be more compact and the am ount of C-A-S-H gel increase, as shown in Figure 4b–d. Based on the results of EDS, the presence of calcium, sodium, silicon and aluminum confirm the C-A-S-H gel in coexistence with N-A-S-H gel. It can also be noticed that Ca/Na increased and Si/Al ratio decreased with an increase in the CAC content from 0% to 7.5%, which favor the amount of C-A-S-H gel increase. Figure 3. The compressive strength of fly ash geopolymer concrete (FAGC) with di erent CAC content. Table 3. Results of mechanical properties of FAGC. MixCAC Content (%)Compressive Strength (MPa)Splitting Tensile Strength (MPa)Elastic Modulus (GPa) M-0 0 33.45 2.47 11.84 M-1 2.5 36.79 2.59 14.79 M-2 5 38.53 2.84 15.44 M-3 7.5 41.02 2.91 16.93 To observe the microstructure of FAGC with di erent CAC contents, SEM and EDS analysis were performed. The results are shown in Figure 4 and Table 4. In Figure 4a, for FAGC with 0% CAC, there are many unreacted FA particles, what is more, the improper bonding of FA particles with binders can be observed, which result in the generation of weak points and decrease compressive strength [ 33]. With the increase of CAC content, the unreacted FA particles to be less, the aluminosilicate gels to be more compact and the amount of C-A-S-H gel increase, as shown in Figure 4b–d. Based on the results of EDS, the presence of calcium, sodium, silicon and aluminum confirm the C-A-S-H gel in coexistence with N-A-S-H gel. It can also be noticed that Ca /Na increased and Si /Al ratio decreased with an increase in the CAC content from 0% to 7.5%, which favor the amount of C-A-S-H gel increase. Table 4. Results of EDS of FAGC (atom percent %). Spectrum Ca Na Al Si O Ca /Na Si /Al S1 0.32 4.69 7.64 14.35 60.08 0.07 1.88 S2 1.07 4.08 9.05 14.75 61.71. 0.26 1.62 S3 1.42 3.47 13.51 15.44 60.48 0.41 1.14 S4 2.08 3.42 11.37 11.34 60.49 0.61 1.00 Materials 2019 ,12, 2982 6 of 13 Materials 2019 , 12, x FOR PEER REVIEW 6 of 13 Figure 4. SEM Images of FAGC: ( a) CAC-0; ( b) CAC-2.5%; ( c) CAC-5%; ( d) CAC-7.5%. Table 4. Results of EDS of FAGC (atom percent %). Spectrum Ca Na Al Si O Ca/Na Si/Al S1 0.32 4.69 7.64 14.35 60.08 0.07 1.88 S2 1.07 4.08 9.05 14.75 61.71. 0.26 1.62 S3 1.42 3.47 13.51 15.44 60.48 0.41 1.14 S4 2.08 3.42 11.37 11.34 60.49 0.61 1.00 3.2. Splitting Tensile Strength As summarized in Figure 5a and Table 3, it can be seen that the CAC has a significant effect on the splitting tensile strength of FAGC. With the incr ease of CAC content, the splitting tensile strength of FAGC improved. Typically, for the different CAC contents, splitting tensile strength is 2.47 MPa with plain geopolymer concrete, 2.59 MPa with 2.5%, 2.84 MPa with 5%, and 2.91 MPa with 7.5%. The value of splitting tensile strength with CAC content of 2.5%, 5% and 7.5% is 4.86%, 14.98% and 17.81% higher than that with plain geopolymer concrete, respectively. Reference to the literature , some recommend ed equations and empiri cal equations can be used to predict the splitting tensile strength fr om compressive strength of FAGC. The empirical equations recommended in codes of practice were used to predict splitting tensile strength. For instance, American concrete institute (ACI) Building Code 318 and Ding et al. can be used to predict the splitting tensile strength, wh ich is expressed by Equations (1) and (2). 𝑓௧=0 . 5 6 ඥ𝑓௖ (1) where 𝑓௧ is splitting tensile strength (MPa) and 𝑓௖ is compressive strength (MPa), 𝑓௧= 0.527 ඥ𝑓௖ (2) The splitting tensile strength of FAGC with di fferent CAC contents obtained from tests, and those obtained by Equations (1) and (2) are expressed in Figure 5b. It is clear that the splitting tensile strength of FAGC is overestimated by ACI 318 Mode l and the equation proposed by Ding et al. . For the given compressive strength of 41.02 MPa, the splitting tensile strength of FAGC is 2.91 MPa. The results obtained by Equations (1) and (2) are 3.59 MPa and 3.38 MPa, respectively, which are 23.37% and 16.15% higher compared with the experimental result, respectively. Figure 4. SEM Images of FAGC: ( a) CAC-0; ( b) CAC-2.5%; ( c) CAC-5%; ( d) CAC-7.5%. 3.2. Splitting Tensile Strength As summarized in Figure 5a and Table 3, it can be seen that the CAC has a significant e ect on the splitting tensile strength of FAGC. With the increase of CAC content, the splitting tensile strength of FAGC improved. Typically, for the di erent CAC contents, splitting tensile strength is 2.47 MPa with plain geopolymer concrete, 2.59 MPa with 2.5%, 2.84 MPa with 5%, and 2.91 MPa with 7.5%. The value of splitting tensile strength with CAC content of 2.5%, 5% and 7.5% is 4.86%, 14.98% and 17.81% higher than that with plain geopolymer concrete, respectively. Materials 2019 , 12, x FOR PEER REVIEW 7 of 13 (a) ( b) Figure 5. (a) Relationship between splitting tensile strength and compressive strength; ( b) relationship of splitting tensile strength between experiment and proposed equations. 3.3. Elastic Modulus As one of the important mechanical properties of concrete, the value of elastic modulus varies with the compressive strength. The mean value of elastic modulus for FAGC with different CAC contents were obtained from tests, and the Equati ons are shown in Figure 6a and Table 3. What is more, Figure 6b shows the variation of elastic modu lus with respect to comp ressive strength. Elastic modulus increases with the improvement of comp ressive strength. The elastic modulus of FAGC without CAC is 11.84 GPa corresponding to the compressive strength of 33.45 MPa. With the increasing of CAC content, an increase in the el astic modulus of FAGC was observed. The values of elastic modulus are up to 14.79 GPa and 15.44 GPa with CAC content of 2.5% and 5% reference to 36.79 MPa and 38.53 MPa, respectively. The highest elastic modulus achieves up to 16.93 GPa with the CAC content of 7.5% for compressive strength of 41.02 MPa. Relative to the FAGC without CAC, the improvements of elastic modulus of FAGC with 2.5%, 5% and 7.5% reach up to 24.92%, 30.41% and 42.99%, respectively. These results can be attr ibuted to the Young’s modulus of C-S-H gel which is equal to around 16–44 GPa, the Young’s modulu s of N-A-S-H gel is about 17–18 GPa [36–38], which is significantly lower than that of C-S-H gel. Therefore, the increasing of C-S-H content resulted in the improvement of elastic modulu s with different CAC contents. (a) ( b) Figure 6. (a) Relationship between elastic modulus and compressive strength; ( b) relationship of elastic modulus between experime nt and proposed equations. Generally, the elastic modulus of concrete is believed to be related to compressive strength. Thus, some empirical equations can be used to predict the elastic modulus from compressive strength. The Equations (3) and (4) proposed by Hardjito et al. and Lee and Lee based on test results of geopolymer concrete. Figure 5. (a) Relationship between splitting tensile strength and compressive strength; ( b) relationship of splitting tensile strength between experiment and proposed equations. Reference to the literature [ 34], some recommended equations and empirical equations can be used to predict the splitting tensile strength from compressive strength of FAGC. The empirical equations recommended in codes of practice were used to predict splitting tensile strength. For instance, Materials 2019 ,12, 2982 7 of 13 American concrete institute (ACI) Building Code 318 [ 35] and Ding et al. [ 34] can be used to predict the splitting tensile strength, which is expressed by Equations (1) and (2). ft=0.56p fc (1) where ftis splitting tensile strength (MPa) and fcis compressive strength (MPa), ft=0.527p fc (2) The splitting tensile strength of FAGC with di erent CAC contents obtained from tests, and those obtained by Equations (1) and (2) are expressed in Figure 5b. It is clear that the splitting tensile strength of FAGC is overestimated by ACI 318 Model and the equation proposed by Ding et al. [ 34]. For the given compressive strength of 41.02 MPa, the splitting tensile strength of FAGC is 2.91 MPa. The results obtained by Equations (1) and (2) are 3.59 MPa and 3.38 MPa, respectively, which are 23.37% and 16.15% higher compared with the experimental result, respectively. 3.3. Elastic Modulus As one of the important mechanical properties of concrete, the value of elastic modulus varies with the compressive strength. The mean value of elastic modulus for FAGC with di erent CAC contents were obtained from tests, and the Equations are shown in Figure 6a and Table 3. What is more, Figure 6b shows the variation of elastic modulus with respect to compressive strength. Elastic modulus increases with the improvement of compressive strength. The elastic modulus of FAGC without CAC is 11.84 GPa corresponding to the compressive strength of 33.45 MPa. With the increasing of CAC content, an increase in the elastic modulus of FAGC was observed. The values of elastic modulus are up to 14.79 GPa and 15.44 GPa with CAC content of 2.5% and 5% reference to 36.79 MPa and 38.53 MPa, respectively. The highest elastic modulus achieves up to 16.93 GPa with the CAC content of 7.5% for compressive strength of 41.02 MPa. Relative to the FAGC without CAC, the improvements of elastic modulus of FAGC with 2.5%, 5% and 7.5% reach up to 24.92%, 30.41% and 42.99%, respectively. These results can be attributed to the Young’s modulus of C-S-H gel which is equal to around 16–44 GPa, the Young’s modulus of N-A-S-H gel is about 17–18 GPa [ 36–38], which is significantly lower than that of C-S-H gel. Therefore, the increasing of C-S-H content resulted in the improvement of elastic modulus with di erent CAC contents. Materials 2019 , 12, x FOR PEER REVIEW 7 of 13 (a) ( b) Figure 5. (a) Relationship between splitting tensile strength and compressive strength; ( b) relationship of splitting tensile strength between experiment and proposed equations. 3.3. Elastic Modulus As one of the important mechanical properties of concrete, the value of elastic modulus varies with the compressive strength. The mean value of elastic modulus for FAGC with different CAC contents were obtained from tests, and the Equati ons are shown in Figure 6a and Table 3. What is more, Figure 6b shows the variation of elastic modu lus with respect to comp ressive strength. Elastic modulus increases with the improvement of comp ressive strength. The elastic modulus of FAGC without CAC is 11.84 GPa corresponding to the compressive strength of 33.45 MPa. With the increasing of CAC content, an increase in the el astic modulus of FAGC was observed. The values of elastic modulus are up to 14.79 GPa and 15.44 GPa with CAC content of 2.5% and 5% reference to 36.79 MPa and 38.53 MPa, respectively. The highest elastic modulus achieves up to 16.93 GPa with the CAC content of 7.5% for compressive strength of 41.02 MPa. Relative to the FAGC without CAC, the improvements of elastic modulus of FAGC with 2.5%, 5% and 7.5% reach up to 24.92%, 30.41% and 42.99%, respectively. These results can be attr ibuted to the Young’s modulus of C-S-H gel which is equal to around 16–44 GPa, the Young’s modulu s of N-A-S-H gel is about 17–18 GPa [36–38], which is significantly lower than that of C-S-H gel. Therefore, the increasing of C-S-H content resulted in the improvement of elastic modulu s with different CAC contents. (a) ( b) Figure 6. (a) Relationship between elastic modulus and compressive strength; ( b) relationship of elastic modulus between experime nt and proposed equations. Generally, the elastic modulus of concrete is believed to be related to compressive strength. Thus, some empirical equations can be used to predict the elastic modulus from compressive strength. The Equations (3) and (4) proposed by Hardjito et al. and Lee and Lee based on test results of geopolymer concrete. Figure 6. (a) Relationship between elastic modulus and compressive strength; ( b) relationship of elastic modulus between experiment and proposed equations. Generally, the elastic modulus of concrete is believed to be related to compressive strength. Thus, some empirical equations can be used to predict the elastic modulus from compressive strength. The Materials 2019 ,12, 2982 8 of 13 Equations (3) and (4) proposed by Hardjito et al. [ 39] and Lee and Lee [ 40] based on test results of geopolymer concrete. E=2070p fc+5300, (3) where Eis the elastic modulus (GPa), E=53003p fc (4) The comparison between the elastic modulus obtained by experiment and the predicted by the above equations are plotted in Figure 6b. The experimental values are lower than those calculated reference to empirical equations of Equations (3) and (4). 3.4. Fracture Properties 3.4.1. Load-CMOD Curves The fracture behavior of concrete can be expressed by means of the complete load-CMOD curve. The typical load-CMOD curves of FAGC exposed to four di erent CAC contents are shown in Figure 7. Based on these curves, it can be seen that FAGC is almost in a state of linear elastic deformation at the start of the loading. Before the load reached up to the initial load, there is not an obvious observation on nonlinear deformation. For the ascending branches of the load-CMOD curve, the slope increase with the increasing of CAC content, owing to the compressive strength and elastic modulus of FAGC increased with the improvement of CAC content. In addition, the crack initiated at the moment of reaching the peak load, then the peak curve shows a decreasing trend. Like the ascending branch, the descending branch can also express the fracture property of the cracked specimen, which is the ductility of FAGC. With the increase of the compressive strength of FAGC, the slope presents a tendency to decrease. This shows that the ductility of FAGC reduced when CAC content increased. Before complete failure, the FAGC with CAC of 7.5%, which has the highest strength, showed greater stretch of the descending branch. Materials 2019 , 12, x FOR PEER REVIEW 8 of 13 𝐸 = 2070 ඥ𝑓௖+ 5300 , (3) where 𝐸 is the elastic modulus (GPa), 𝐸 = 5300 ඥ𝑓௖య (4) The comparison between the elastic modulus obta ined by experiment an d the predicted by the above equations are plotted in Figure 6b. The expe rimental values are lower than those calculated reference to empirical equation s of Equations (3) and (4). 3.4. Fracture Properties 3.4.1. Load-CMOD Curves The fracture behavior of concrete can be expre ssed by means of the complete load-CMOD curve. The typical load-CMOD curves of FAGC exposed to four different CAC contents are shown in Figure 7. Based on these curves, it can be seen that FAGC is almost in a state of linear elastic deformation at the start of the loading. Before the load reached up to the initial load, there is not an obvious observation on nonlinear deformation. For the as cending branches of the load-CMOD curve, the slope increase with the increasing of CAC content, owing to the compressive strength and elastic modulus of FAGC increased with the improvement of CAC content. In addition, the crack initiated at the moment of reaching the peak load, then the peak curve shows a decreasing trend. Like the ascending branch, the descending branch can also express the fracture property of the cracked specimen, which is the ductility of FAGC. With the increase of the compressive strength of FAGC, the slope presents a tendency to decrease. This sh ows that the ductility of FAGC reduced when CAC content increased. Before complete failure, the FAGC with CAC of 7.5%, which has the highest strength, showed greater stretch of the descending branch. Figure 7. Load- crack mouth opening displacement (CMOD) curves of FAGC with different CAC contents. 3.4.2. Peak Load Table 5 presents the peak loads of FAGC wi th different CAC contents. The corresponding relationship between peak load and compressive st rength of FAGC are presented in Figure 8. The peak load increases with the increase of compressive strength. The peak load of FAGC varies in the range from 1.08 to 1.79 kN. The FAGC mixed with 7.5% CAC with the maximal peak load and compressive strength. Obviously, when the CAC content grows up to 2.5%, 5% and 7.5%, the peak loads of FAGC are 1.24 kN, 1.35 kN and 1.79 kN, which are 14.81%, 25% and 65.74% larger than that of FAGC without CAC, respective ly. The CAC content greatly influences the compressive strength, which leads to the variation of peak load at the same time. As expected, the peak loads of FAGC increase with the increase of CAC content, that is to say, improve with compressive strength. Figure 7. Load- crack mouth opening displacement (CMOD) curves of FAGC with di erent CAC contents. 3.4.2. Peak Load Table 5 presents the peak loads of FAGC with di erent CAC contents. The corresponding relationship between peak load and compressive strength of FAGC are presented in Figure 8. The peak load increases with the increase of compressive strength. The peak load of FAGC varies in the range from 1.08 to 1.79 kN. The FAGC mixed with 7.5% CAC with the maximal peak load and compressive strength. Obviously, when the CAC content grows up to 2.5%, 5% and 7.5%, the peak loads of FAGC are 1.24 kN, 1.35 kN and 1.79 kN, which are 14.81%, 25% and 65.74% larger than that of FAGC without CAC, respectively. The CAC content greatly influences the compressive strength, which leads to the Materials 2019 ,12, 2982 9 of 13 variation of peak load at the same time. As expected, the peak loads of FAGC increase with the increase of CAC content, that is to say, improve with compressive strength. Table 5. Experimental results of fracture parameters. Mix CAC Content (%) Pini(N) Pmax(N) CMOD c(mm) ac(mm) M-0 0 621.1 1079 0.0695 36.5 M-1 2.5 817.6 1236 0.06824 42.3 M-2 5 862.4 1353.6 0.04853 44.1 M-3 7.5 1259.6 1788.4 0.0377 44.4 Materials 2019 , 12, x; doi: FOR PEER REVIEW www.mdpi.com/journ al/materials Table 5. Experimental results of fracture parameters. Mix CAC Content (%) 𝑷𝒊𝒏𝒊 (N) 𝑷𝒎𝒂𝒙 (N) 𝑪𝑴𝑶𝑫 𝒄 (mm) 𝒂𝒄 (mm) M-0 0 621.1 1079 0.0695 36.5 M-1 2.5 817.6 1236 0.06824 42.3 M-2 5 862.4 1353.6 0.04853 44.1 M-3 7.5 1259.6 1788.4 0.0377 44.4 Figure 8. Peak loads of FAGC with different CAC contents. 3.4.3. Fracture Energy Fracture energy refers to the energy needed to generate cracks per unit area. As shown in Equation (5), it is a method to obtain the fractu re energy of the three-point bending specimen by calculating the area surrounded by the measured lo ad-CMOD curve divided by the ligament area. In order to simplify the test, when the measured crack wo rk is close to the actual fracture energy, the end point is close to the point of complete failure. According to the literature , the constant value of the far tail can be used to calculate the true frac ture energy. In this study, the test stops when the descending branch of the load-CMOD curve is full-tailed. Moreover, in the calculation of fracture energy, the self-weight of the sample is not taken in to account owing to the si ze of the specimen and the supporting form of the specimen. 𝐺 ௙=𝑊଴ 𝐴௟௜௚ (5) where W଴ is equal to the area of load-CMOD curve and 𝐴௟௜௚ is the ligament area (m2). In Figure 9, the fracture energy of FAGC is distinctly affected by CAC content. The change in the fracture energy of FAGC has the same trend as that in CAC content. At the beginning, the value of 𝐺௙ of FAGC without CAC is 67.3 N/m; however, when the CAC content grows from 0 to 2.5% and 5%, the 𝐺௙ rapidly increases up to the value of 72.3 N/m and 76.7 N/m. On prolonging the CAC content up to 7.5%, a similar phenomenon, 𝐺௙ tends to increase at a homologous rate, has been observed. The value of 𝐺௙ eventually reaches up to 85.8 N/m. Summarized the above results, it can be concluded that the fracture energy continuously increased with the increasing of CAC content. Consequently, when the CAC content grows up to 2.5%, 5% and 7.5%, the 𝐺௙ of FAGC are 7.4%, 13.91% and 27.49% higher than that of FAGC without CAC, respectively. The reason why increasing CAC content increases the 𝐺௙ is that the CAC provided more calcium and additional reactive aluminum to participate in geopolymerization, whic h resulted in the Ca/Si to be larger and yield to more C-A-S-H and C-S-H gels. The C-A-S-H and C-S-H gels with more initial micro-cracks resulted Figure 8. Peak loads of FAGC with di erent CAC contents. 3.4.3. Fracture Energy Fracture energy refers to the energy needed to generate cracks per unit area. As shown in Equation (5), it is a method to obtain the fracture energy of the three-point bending specimen by calculating the area surrounded by the measured load-CMOD curve divided by the ligament area. In order to simplify the test, when the measured crack work is close to the actual fracture energy, the end point is close to the point of complete failure. According to the literature [ 30], the constant value of the far tail can be used to calculate the true fracture energy. In this study, the test stops when the descending branch of the load-CMOD curve is full-tailed. Moreover, in the calculation of fracture energy, the self-weight of the sample is not taken into account owing to the size of the specimen and the supporting form of the specimen. Gf=W0 Alig(5) where W0is equal to the area of load-CMOD curve and Aligis the ligament area (m2). In Figure 9, the fracture energy of FAGC is distinctly a ected by CAC content. The change in the fracture energy of FAGC has the same trend as that in CAC content. At the beginning, the value of Gf of FAGC without CAC is 67.3 N /m; however, when the CAC content grows from 0 to 2.5% and 5%, theGfrapidly increases up to the value of 72.3 N /m and 76.7 N /m. On prolonging the CAC content up to 7.5%, a similar phenomenon, Gftends to increase at a homologous rate, has been observed. The value of Gfeventually reaches up to 85.8 N /m. Summarized the above results, it can be concluded that the fracture energy continuously increased with the increasing of CAC content. Consequently, when the CAC content grows up to 2.5%, 5% and 7.5%, the Gfof FAGC are 7.4%, 13.91% and 27.49% higher than that of FAGC without CAC, respectively. The reason why increasing CAC content increases theGfis that the CAC provided more calcium and additional reactive aluminum to participate in geopolymerization, which resulted in the Ca /Si to be larger and yield to more C-A-S-H and C-S-H gels. The C-A-S-H and C-S-H gels with more initial micro-cracks resulted in more ductility matrix to absorb Materials 2019 ,12, 2982 10 of 13 more energy during crack propagation. Hence, the increasing of CAC content gives rise to the increase in the Gf. Materials 2019 , 12, x FOR PEER REVIEW 2 of 13 in more ductility matrix to absorb more energy during crack propagation. Hence, the increasing of CAC content gives rise to the increase in the 𝐺௙. Figure 9. Fracture energy of FAGC with different CAC contents. 3.4.4. Effective Fracture Toughness Reference to the double-K fracture model , the 𝑃௜௡௜ and 𝑎଴, which are initial cracking load and initial notch depth, can be substituted into Equation (6) to calculated the initial cracking fracture toughness 𝐾ூ௖௜௡௜. 𝐾ூ௖௜௡௜=3𝑃௜௡௜𝑆 2𝐷ଶ𝐵ඥ𝑎଴𝐹ଶቀ𝑎଴ 𝐷ቁ, 𝑉ଶቀ𝑎 𝐷ቁ = 0.76 − 2.28𝑉 ଶቀ𝑎 𝐷ቁ + 3.87𝑉 ଶቀ𝑎 𝐷ቁଶ −2 . 0 4ቀ𝑎 𝐷ቁଷ +0.66 ቀ1 −𝑎 𝐷ቁଶ, (6) where 𝑃௜௡௜ is the initial cracking load (kN), S is the span of TPB beam (mm), D and B is the height and thickness of TPB beam (mm), 𝑎 and 𝑎଴ is the effective crack length and initial crack length (mm). Based on the parameters of peak load 𝑃௠௔௫ and critical notch length 𝑎௖ for 𝑃௜௡௜ and 𝑎଴, the unstable fracture toughness 𝐾ூ௖௨௡can be calculated by the above formulas with the substituting method. The 𝑃௜௡௜ was obtained by the initial point of non-linearity in the P-CMOD curve . These parameters mentioned above are summarized in Table 5. Based on Tada et al. , for the TPB specimen of S/D = 4, the relation between load and CMOD expresses as follows: 𝐶𝑀𝑂𝐷 =24𝑃𝑎 𝐵𝐷𝐸𝑉ଶ(𝛼), 𝑉ଶ(𝛼)= 0.76 − 2.28𝑉 ଶ(𝛼)+ 3.87𝑉 ଶ(𝛼)ଶ−2 . 0 4 (𝛼)ଷ+0.66 (1−𝛼 )ଶ, (7) where 𝛼 = ( 𝑎+ℎ ଴𝐷+ℎ ଴ ⁄ ); ℎ଴ is equal to 3 mm; 𝑎௖ can be determined by Equation (7) when 𝑃 reached up to 𝑃௠௔௫ and CMOD is replaced by 𝐶𝑀𝑂𝐷 ௖; and E can be obtained by Equation (8). 𝐸=24𝑃𝑎 ଴ 𝐶௜𝐵𝐷𝑉ଶ(𝛼଴), (8) in which, 𝛼଴ = (𝑎଴+ℎ ଴𝐷+ℎ ଴ ⁄ ); 𝐶௜ = 𝐶𝑀𝑂𝐷 ௜/𝑃௜. The effective fracture toughness of FAGC with di fferent CAC contents is shown in Figure 10. As shown in Figure 10, it can be seen that 𝐾ூ௖ is a monotonically increasing function of the CAC content, and the similar development trend of 𝐺௙ with different CAC contents. For example, when the CAC content of 0%, the 𝐾ூ௖௜௡௜ is 1.06 MPam0.5; however, once the content increases from 0 to 2.5%, 5% and 7.5%, the values of 𝐾ூ௖௜௡௜ improves up to 1.70, 1.79 and 4.27 MPam0.5, respectively. The value of 𝐾ூ௖௜௡௜ with CAC content of 2.5%, 5% and 7.5% is 60.38%, 68.87% and 302.83% larger than that with plain Figure 9. Fracture energy of FAGC with di erent CAC contents. 3.4.4. E ective Fracture Toughness Reference to the double-K fracture model [ 41], the Piniand a0, which are initial cracking load and initial notch depth, can be substituted into Equation (6) to calculated the initial cracking fracture toughness Kini Ic. Kini Ic=3PiniS 2D2Bpa0F2a0 D , V2a D =0.762.28V2a D +3.87V2a D22.04a D3+0.66 (1a D)2,(6) where Piniis the initial cracking load (kN), S is the span of TPB beam (mm), D and B is the height and thickness of TPB beam (mm), aanda0is the e ective crack length and initial crack length (mm). Based on the parameters of peak load Pmaxand critical notch length acforPinianda0, the unstable fracture toughness Kun Iccan be calculated by the above formulas with the substituting method. The Pini was obtained by the initial point of non-linearity in the P-CMOD curve [ 42]. These parameters mentioned above are summarized in Table 5. Based on Tada et al. [ 43], for the TPB specimen of S /D=4, the relation between load and CMOD expresses as follows: CMOD =24Pa BDEV2(), V2()=0.762.28V2()+3.87V2()22.04()3+0.66 (1)2,(7) where =(a+h0/D+h0);h0is equal to 3 mm; accan be determined by Equation (7) when Preached up to Pmaxand CMOD is replaced by CMOD c; and Ecan be obtained by Equation (8). E=24Pa0 CiBDV2(0), (8) in which, 0=(a0+h0/D+h0);Ci=CMOD i/Pi. The e ective fracture toughness of FAGC with di erent CAC contents is shown in Figure 10. As shown in Figure 10, it can be seen that KIcis a monotonically increasing function of the CAC content, and the similar development trend of Gfwith di erent CAC contents. For example, when the CAC content of 0%, the Kini Icis 1.06 MPam0.5; however, once the content increases from 0 to 2.5%, 5% and 7.5%, the values of Kini Icimproves up to 1.70, 1.79 and 4.27 MPam0.5, respectively. The value of Kini Icwith CAC content of 2.5%, 5% and 7.5% is 60.38%, 68.87% and 302.83% larger than that with plain geopolymer concrete. Afterwards, there is a positive correction between the CAC content and Kun Ic. Typically, for the Materials 2019 ,12, 2982 11 of 13 dierent CAC contents, Kun Icis 4.44 MPam0.5with plain geopolymer concrete, 6.4 MPam0.5with 2.5%, 7.54 MPam0.5with 5%, and 10.1 MPam0.5with 7.5%. The values of Kun Icwith CAC content of 2.5%, 5% and 7.5% are 44.14%, 69.82% and 127.48%, higher than that with plain geopolymer concrete. These results are owing to that geopolymerization process not only produced the aluminum-rich gels, such as N-A-S-H gels, but also yield C-A-S-H and C-S-H gels, when the FAGC added CAC. According to these results, the e ect of CAC on KIcis getting more and more. Therefore, we can again confidently conclude that CAC content is a decisive factor in determining the properties of FAGC. Materials 2019 , 12, x FOR PEER REVIEW 3 of 13 geopolymer concrete. Afterwards, there is a po sitive correction between the CAC content and 𝐾ூ௖௨௡. Typically, for the different CAC contents, 𝐾ூ௖௨௡ is 4.44 MPam0.5 with plain geopolymer concrete, 6.4 MPam0.5 with 2.5%, 7.54 MPam0.5 with 5%, and 10.1 MPam0.5 with 7.5%. The values of 𝐾ூ௖௨௡ with CAC content of 2.5%, 5% and 7.5% are 44.14%, 69.82% and 127.48%, higher than that with plain geopolymer concrete. These results are owing to that geop olymerization process not only produced the aluminum-rich gels, such as N-A-S-H gels, but also yield C-A-S-H and C-S-H gels, when the FAGC added CAC. According to these re sults, the effect of CAC on 𝐾 ூ௖ is getting more and more. Therefore, we can again confidently conclude that CAC content is a decisive factor in determining the properties of FAGC. Figure 10. Fracture toughness of FAGC with different CAC contents. 4. Conclusion In this study, the effect of CAC content on th e mechanical and fracture properties of fly ash geopolymer concrete was investigated. The CAC co ntent ranges from 0 to 7.5%. The compressive strength, splitting tensile strength and elastic mo dulus of FAGC with different CAC contents were measured. The key fracture parameters, such as load-CMOD curves, peak load, 𝐺௙, 𝐾ூ௖௜௡௜ and 𝐾ூ௖௨௡, were also evaluated from the experimental da ta for different CAC contents. The following conclusions can be drawn: (1) The compressive strength, splitting tensile st rength and elastic modulus of FAGC improved with the increase of CAC content, the maximal valu es of them are 41.02 MPa, 2.91 MPa and 16.93 GPa with the growth of 22.63%, 17.81% and 42.99%, respectively, which are corresponding to the CAC content was equal to 7.5%. (2) The effect of CAC content on 𝐺 ௙ was significant, and it caused a favorable effect on 𝐺௙. When the CAC content was equal to 7.5%, the maximal value of 𝐺௙ is almost 85.8 N/m, which is owing to the initial micro-crack in the C-S-H gels absorbed much more energy. (3) The value of 𝐾ூ௖ became higher with increasing of CAC content. Typically, for CAC content of 7.5%, the 𝐾ூ௖௜௡௜ of FAGC reaches up to a maximal value of 4.27 MPam0.5. Meanwhile, the biggest value of 𝐾ூ௖௨௡ achieves up to 10.1 MPam0.5. Author Contributions: Conceptualization, S.H. and Y.W.; methodolo gy, S.H. and Y.W.; software, Y.W. and Z.H.; formal analysis, Y.W. and S.H.; data curation, Y.W.; writing—original draf t preparation, Y.W.; writing—review and editing, S.H., Z.H. and Y.W.; funding acquisition, S.H. Funding: This research was funded by the National Natura l Science Foundation of China, grant number is 2016YFC0401902, 51739008 and 51527811. Figure 10. Fracture toughness of FAGC with di erent CAC contents. 4. Conclusions In this study, the e ect of CAC content on the mechanical and fracture properties of fly ash geopolymer concrete was investigated. The CAC content ranges from 0 to 7.5%. The compressive strength, splitting tensile strength and elastic modulus of FAGC with di erent CAC contents were measured. The key fracture parameters, such as load-CMOD curves, peak load, Gf,Kini IcandKun Ic, were also evaluated from the experimental data for di erent CAC contents. The following conclusions can be drawn: (1) The compressive strength, splitting tensile strength and elastic modulus of FAGC improved with the increase of CAC content, the maximal values of them are 41.02 MPa, 2.91 MPa and 16.93 GPa with the growth of 22.63%, 17.81% and 42.99%, respectively, which are corresponding to the CAC content was equal to 7.5%. (2) The e ect of CAC content on Gfwas significant, and it caused a favorable e ect on Gf. When the CAC content was equal to 7.5%, the maximal value of Gfis almost 85.8 N /m, which is owing to the initial micro-crack in the C-S-H gels absorbed much more energy. (3) The value of KIcbecame higher with increasing of CAC content. Typically, for CAC content of 7.5%, the Kini Icof FAGC reaches up to a maximal value of 4.27 MPam0.5. Meanwhile, the biggest value of Kun Icachieves up to 10.1 MPam0.5. Author Contributions: Conceptualization, S.H. and Y.W.; methodology, S.H. and Y.W.; software, Y.W. and Z.H.; formal analysis, Y.W. and S.H.; data curation, Y.W.; writing—original draft preparation, Y.W.; writing—review and editing, S.H., Z.H. and Y.W.; funding acquisition, S.H. Funding: This research was funded by the National Natural Science Foundation of China, grant number is 2016YFC0401902, 51739008 and 51527811. Acknowledgments: The authors are grateful to the National Key Research and Development Program of China (2016YFC0401902), the State Key Program of National Natural Science of China (51739008) and the National Major Scientific Instruments Development Project of China (51527811), and Nanjing Hydraulic Research Institute (Y17017, Y17015) for the provision of financial support to undertake this research. Conflicts of Interest: The authors declare no conflict of interest. Materials 2019 ,12, 2982 12 of 13

Comparison of long term performance between alkali activated slag and fly ash geopolymer concretes(Universitas Negeri Surabaya), Kampus Unesa Ketintang, Surabaya 60231, Indonesia bCivil and Infrastructure Engineering, School of Engineering, RMIT University, 124, La Trobe Street, Melbourne, Victoria 3000, Australia highlights /C15Engineering properties of FAGP concrete improve from 28 to 540 days from casting. /C15Continuing gel production of FAGP concrete densify microstrucre over time. /C15Mechanical properties of AAS concrete decrease between 90 and 540 days from casting. /C15Disjoining pressure & self-desiccation effect propagate cracks in AAS in long term. /C15FAGP concrete is behaving in a similar manor to PC concrete. article info Article history: Received 29 December 2016Received in revised form 1 March 2017Accepted 14 March 2017 Keywords: Fly ash geopolymerAlkali activated slagMechanical properties Durability Long term performanceabstract This paper reports the comparison of engineering properties of alkali activated slag (AAS) and low cal- cium fly ash geopolymer (FAGP) concretes up to 540 days. The results showed that the AAS concrete had higher compressive and tensile strength, elastic modulus and lower permeation characteristics thanFAGP concrete in the initial 90 days. However, a reduction in AAS concrete performance was observed between 90 and 540 days, while an increase was noted in FAGP concrete over the same time period. The microscopy revealed that both reactions progressed beyond 90 days with the slag–alkali producingexcess C–S–H gel which was observed to increase the crack propagation and crack width at latter ages, attributed to the combined effect of disjoining pressure and self-desiccation. The fly ash geopolymeriza- tion also continued following an initial 24 h heat curing resulting in a crack-free dense microstructure at540 days. Overall the discrepancy in microstructural development beyond 90 days in the two concreteswould explain the contradictory performance over the longer time frame. /C2112017 Elsevier Ltd. All rights reserved. 1. Introduction Concrete is the most widely used construction material in soci- ety today. Concrete is conventionally produced by using Portlandcement (PC) as the primary binder with the ratio of PC in tradi- tional concrete being approximately 10–15% by the mass of con- crete. However, the production of PC has led to environmental concerns over the production of CO 2. Cement production has been estimated as contributing between 5 and 7% of the current anthro- pogenic CO 2emissions worldwide [1,2] , with the production of 1 ton of cement producing from 0.6 up to 1 ton of CO 2, depending on the power plant [3–5] . This had led to the adoption of wastematerials, such as fly ash (FA) and ground granulated blast- furnace slag (GGBS), as a replacement for PC due to their ability to enhance the physical, chemical and mechanical properties of cements and concretes. More recently research has shown that itis possible to develop geopolymer concretes based solely on waste materials activated directly, without the presence of PC, utilizing an alkaline activator [6–12] . A major benefit of geopolymer con- crete is that the reduction of CO 2emission by 26–45% with the replacement of PC with no adverse economic effects [13–15] . In the geopolymerization process, alumina and silica species in FA rapidly react with highly alkaline activator solution and pro- duce a three-dimensional polymeric chain and ring structure con- sisting of Si–O–Al–O bonds. The schematic formation of the final geopolymer product is sodium-aluminosilicate (N–A–S–H) gel, which governs the properties of low calcium fly ash geopolymer (FAGP) concrete . Conversely, in AAS concrete, the calcium sil- icate hydrates (C–S–H) gel is the main resultant product of 0950-0618/ /C2112017 Elsevier Ltd. All rights reserved.⇑Corresponding author. E-mail addresses: ariewardhono@unesa.ac.id (A. Wardhono), chamila.gunase- kara@rmit.edu.au (C. Gunasekara), david.law@rmit.edu.au (D.W. Law), sujeeva. setunge@rmit.edu.au (S. Setunge).Construction and Building Materials 143 (2017) 272–279 Contents lists available at ScienceDirect Construction and Building Materials journal homepage: www. elsevier.com/loc ate/conbuildmat geopolymerisation, which is similar to the primary binding phase of PC and blended cement concretes . Hardjito & Rangan and Fernandez-Jimenez et al. stud- ied the mechanical properties of FAGP concrete up to 90 days and observed that it has a comparable compressive strength, higher flexural and splitting tensile strength, but a lower elastic modulus to that of PC concrete. Ryu et al. showed that the splitting ten- sile strength to compressive strength ratio at 28 days ranged between 7.8 and 8.2%, similar to that of PC concrete. Neupane et al. and Loya et al. also found that the relationship between elastic modulus and compressive strength of FAGP con- crete is similar to that of PC concrete. Research has also demon- strated similar mechanical properties for AAS concrete to PC concrete for periods up to 90 day [17,23,24] , though a reduction of compressive strength with time has been reported by Collinsand Sanjayan , while Bernal et al. found that AAS concrete has a comparable compressive strength, but higher flexural strength than PC concrete. Considering the permeation characteristics, Bernal et al. showed that the binder content of the concretes has a particularly strong effect on the water absorption properties of AAS concrete. Collins and Sanjayan reported that AAS concrete has a lower water absorption due to the presence of very refined, tortuous and closed porosity in the concrete. Moreover, Olivia et al. sta- ted that fly ash geopolymer concrete exhibits low water absorption and sorptivity compared to the PC concrete. The water/binder ratio and well-graded aggregate influence were noted to influence the permeation characteristics. However, these studies were only con- ducted up to 90 days, and there is no comparison between AAS and fly ash geopolymer concretes over the long term. In order to function as a construction material, it is imperative that both AAS and FAGP concretes maintain their performance over the design life of a structure. This paper reports the details of an experimental research program that has been undertaken to inves- tigate a range of mechanical and durability properties of AAS and FAGP concrete up to 540 days. The properties assessed were com- pressive strength, flexural and splitting tensile strength, elastic modulus, water absorption and water permeability. 2. Significance of research Published research to date on AAS and FAGP concrete has been reported their performance only up to 90 days (short term), in each study using a mixing process unique to that study, with no com- parison of long term performance between them. This research reports the performance of AAS and FAGP concretes up to one and half year while applying the same mixing process, providing a systematic long term comparison study of the engineering prop- erties between them. Research data presented here thus will be extremely useful to comprehend the long term behavior of AAS and FAGP concretes. 3. Experimental procedure 3.1. Materials used The GGBS was a construction grade slag conforming to Australian Standard, AS 3582.2 , with the basicity coefficient of 0.81 and the hydration modulus of 1.5. The low calcium, class F FA conforming to Australian standard, AS 3582.1 was obtained from Tarong power station in Australia. The chemical composition, parti-cle size distribution and mineralogical composition of fly ash and GGBS, determined by X-ray fluorescence (XRF), Malvern particle size analyzer instruments and X-ray diffraction (XRD), respectively are shown in Table 1 and 2 . Brunauer Emmett Teller (BET) method by N 2absorption was used to determine the fly ash surface area. The alkaline activator used in AAS and FAGP concretes consisted of a mixture of Commercially available sodium silicate solution with a specific gravity of 1.53 andan alkaline modulus ratio (Ms) equal to 2 (where Ms = SiO 2/Na 2O, Na 2O = 14.7%,SiO 2= 29.4% and 55.9% H 2O by mass), and sodium hydroxide solution. A 15 M NaOH solution was used for the manufacture of the FAGP and a 10 M NaOH solution usedfor the AAS. The selection of two different molarity in sodium hydroxide solution isdependent on the mix optimization based on 28-day compressive strength Bothcoarse and fine aggregate were prepared in accordance with AS 1141.5 . The aggregate was in a saturated surface dry condition. The fine aggregate was river sand in uncrushed form with a specific gravity of 2.5 and a fineness modulus of3.0. The coarse aggregate was crushed granite aggregate of two-grain sizes:7 mm, 2.58 specific gravity and 1.60% water absorption, and 10 mm, 2.62 specificgravity and 0.74% water absorption. Demineralized water was used throughoutthe experiment. 3.2. Mix proportions and specimen preparations Mix proportions used in AAS and FAGP concretes were based on a previous study, which is summarized in Table 3 . The activator modulus (SiO 2/Na 2Oi n alkaline activator) is fixed at 1.0 for both concretes while Na 2O dosage (Na 2Oi n alkaline activator/FA) is fixed at 5% and 15% in the AAS and FAGP concretes, respec-tively. The ratio of components, such as binder (GGBS or FA), alkaline activator,aggregate and water, was calculated based on the absolute volume method . The total aggregate in the concrete was kept to 64% of the entire mixture by volumefor all mixes. A water solid ratio (w/s) of 0.44 and 0.37 was used to prepare the AASand FAGP concrete, which gave a consistent workability in the mixing process. The total liquid and solid content is shown in Table 3 . The mass of water in the mix was taken as the sum of mass of water contained in the sodium silicate, sodium hydrox-ide and added water. The mass of solid is taken as the sum of binder (GGBS or FA),the solids in the sodium silicate and the sodium hydroxide solution. The mixing of concretes was carried out using a 120 liter concrete mixer. The dry materials (GGBS or FA, fine aggregates and coarse aggregates) were mixed firstfor 4 min. Then activator and water were added to the dry mix and mixed contin-uously for another 8 min until the mixture was glossy and well combined. The mix- ture was then poured into moulds and vibrated using a vibration table for 1 min to remove air bubbles. After vibration both AAS and FAGP concrete specimens werekept at room temperature (23 /C176C) for 1 day. The AAS specimens were demoulded, water-cured (23 /C176C) for 6 days and kept at room temperature until being tested. The FAGP specimens were heat-cured (80 /C176C) using dry oven for 24 h, the moulds were removed from the oven and left to cool to room temperature before demould-ing, and the samples were kept at room temperature until being tested. 3.3. Testing The compressive strength test was performed by MTS machine with a loading rate of 20 MPa/min according to AS 1012.9 . The flexural and splitting tensile strength tests were conducted to determine the tensile strength of concretes inaccordance with AS 1012.11 and AS 1012.10 respectively. The flexural ten- sile strength test was carried out on a MTS machine with additional testing appara-tus under a four point bending test with a loading rate of 1 MPa/min. The splittingtensile strength test was performed on MTS machine equipped with splitting ten-sile strength test equipment under a loading rate of 1.5 MPa/min. The elastic mod- ulus was determined using Tecnotest concrete testing machine coupled with the compressometer/extensometer with a loading rate of 0.25 MPa/s in accordancewith AS 1012.17 , and dry density was measured accordance with AS 1012.12.2 . The ultrasonic pulse velocity test was conducted in accordance with ASTM C597 standard using a portable ultrasonic non-destructive digital indicating tester with a 54 kHz transducer. The water permeability tests were performed using the Autoclam Permeability System. Water is admitted into the test area through a prim- ing pump and the pressure inside is increased to 0.5 bar at the end of the priming.The quantity of water flowing into the concrete is recorded every minute for dura-tion of 15 min. The water absorption test was carried out in accordance with AS1012.21 standard to determine the immersed absorption. Immersed absorp- tion ( Ai) is the ratio (%) of the mass of water contained in a concrete specimen, and was used to determine the water absorption of concrete specimens. The appar-ent volume of permeable void (AVPV) percentage is also measured in accordance with AS 1012.21 standard . The specimens of 100 mm diameter /C2200 mm long cylinders were cut into four equal slices for both experiments and the resultreported is the average of the results for the four slices. All tests were conductedat 28, 56, 90, 180, 360 and 540 days of casting. The reported test results in eachspecific test for each concrete are an average of three samples. The microstructure development was observed using scanning electron micro- scopy (SEM) imaging employing backscatter electron detector with 15 eV of energy.Energy dispersive X-ray spectroscopy (EDS) analysis was performed using Oxford instruments nano-analysis software (AZtec 2.1 )to determine the chemical compo- sition of the reacted geopolymer. Specimens were cut using a diamond saw to a sizeof 2–4 mm in height and 5–10 mm in diameter. The samples were subsequentlycarbon coated and then mounted on the SEM sample stage with conductive,double-sided carbon tape. A total of three samples were investigated for eachgeopolymer concrete.A. Wardhono et al. / Construction and Building Materials 143 (2017) 272–279 273 4. Experimental results 4.1. Mechanical properties 4.1.1. Density and compressive strength The dry density and compressive strength development of AAS and FAGP con- cretes between 28 and 540 days are displayed in Fig. 1 (a and b). The dry density of AAS and FAGP concretes ranged between 2453–2460 and 2302–2326 kg/m3, respectively from 28 to 540 days. While the density of FAGP concrete is slightlylower than PC concrete, which is characteristically cited as 2400 kg/m 3, the AAS concrete was marginally higher compared to PC concrete. It is noted that both concretes dsplay a less than 1% of densitly increase during the 28 and 540 days period. The AAS concrete had a higher compressive strength than FAGP concrete throughout. The AAS achieved 39.5 MPa at 28 days, compared to an initial strengthof 22.4 MPa for the FAGP. However, the FAGP concrete gave an increase in strengthwith time, achieving 33.2 MPa after 360 days, which then remains constant to540 days. However, while the AAS concrete does display an increase in strengthto 41.6 MPa at 56 days no further increase is observed. Indeed a slight reduction in strength is noted, 40.2 MPa at 180 days, though a strength above 40 MPa is main- tained throughout the remaining period. Overall, whilst FAGP concrete displayed a48.2% (10.2 MPa) compressive strength increase from 28 to 540 days, only a 2.3%(0.9 MPa) strength development is observed for the AAS concrete in same period.This indicates an ongoing geopolymerisation reaction in FAGP concrete followingthe initial heat curing [42,43] , while the same is not observed for the AAS. 4.1.2. Tensile strength The flexural strength and splitting tensile strength development of AAS and FAGP concretes between 28 and 540 days are shown in Fig. 2 (a and b). Similar to compressive strength development, the flexural strength of FAGP concrete tendedto increase with time. It ranged from 4.7 to 7.2 MPa between 28 and 365 days,i.e. a 53.2% of flexural strength increase during this period. The AAS achieves a6 MPa flexural strength at 28 days, compared to an initial flexural strength of4.7 MPa for the FAGP. However, the AAS concrete shows a decrease in flexuralstrength with time, achieving 5.2 MPa at 540 days, which is a 13.3% fall in strength. It is noted that both FAGP and AAS concretes obtained similar flexural strength (5.8 MPa) at 90 days. Subsequently, FAGP increased in flexural strength, whileAAS decreased, during the 90–540 day period. Overall the flexural strength of theFAGP and AAS concrete ranged between 20–22% and 14–16% of the compressivestrength respectively, compared to a range of 9–12% typically cited for PC concrete[11,24] .On the other hand, both AAS and FAGP concretes show a lower splitting tensile strength than corresponding flexural strengths as shown in Fig. 2 . The splitting ten- sile strength of FGAP concrete increased with time, from 2.1 to 4.1 MPa between 28and 540 days, and varied from 9 to 12% of the compressive strength. The AASshowed higher splitting tensile strength than FAGP concrete up to 90 days, but afterthat FAGP showed a significant improvement compared to the AAS concrete andachieved a 24.2% increase in splitting tensile strength at 540 days. 4.1.3. Elastic modulus The modulus of elasticity development of AAS and FAGP concretes between 28 and 540 days are shown in Fig. 3 . This property of concrete expresses the ratio between a certain range of unit stress and unit elongation within the elastic limit.A higher elastic modulus will represent a better quality of concrete specimen. Theelastic modulus of FAGP and AAS concretes ranged between 8022–15942 and26768–15279 MPa, respectively over the 28 to 540 day period. The AAS achieves a significantly high elastic modulus (26768 MPa) at 28 days, compared to an initial elastic modulus of 8022 MPa for the FAGP concrete. It is worth noting that the dataagain shows contrasting trends, with the FAGP concrete displaying an increase withtime, while the AAS concrete shows a decrease with time, such that by 540 days theFAGP has a higher elastic modulus than the AAS concrete. The FAGP concrete has atwofold increase of elastic modulus from 28 to 540 days, but AAS concrete shows a43% decrease during this time interval. 4.2. Permeation properties 4.2.1. Water absorption and AVPV The water absorption and AVPV of AAS and FAGP concretes are shown in Fig. 4 (a and b). The FAGP concrete had higher water absorption (7.75%) than AAS concrete(4.79%) at 28 days. However, the long term data displays a reduction of water absorption in FAGP to 6.82% at 360 days and to 6.74% at 540 days with an overall decrease of 13% observed from 28 to 540 days. In PC concrete, a water absorptiongreater than 5% is classified as high permeable concrete, while less than 3% is clas-sified as low permeable concrete . The FAGP concrete exceeded this upper limit at all ages and behaved as high permeable concrete, which indicates a highly porousexternal surface. The AAS concrete shows water absorption less than 5% in the first90 days, but then exceeded this upper limit achieving 5.36% at 540 days. Thesetrends are consistent with those observed for the flexural strength and elastic mod- ulus, indicating an improvement in density of the pore-structure for the FAGP over the longer term, while also suggesting a reduction in density within the pore-structure for the AAS concrete during this period. The AVPV is a percentage of pore space such as capillary pores, gel pores and air voids within the concrete. The trends observed were similar to water absorption forboth concretes, AVPV decreasing with time in FAGP, while AAS increases. In PC con-crete an AVPV less than 13% is classified as good quality concrete, while greater than 18% is classified as poor quality concrete . It is noted that AAS concrete at all ages were below this lower limit though it has an increase of AVPV with time,indicating limited pore interconnectivity in their pore-structure. However, theFAGP concrete had AVPV percentage between 13% and 15% at all ages. 4.2.2. Water permeability & UPV The variation of water permeability index with time is shown in Fig. 5 . In the water permeability test, both capillary absorption and the applied pressure con- tribute to the rate of water flow. The slope of the linear regression curve betweenwater flow and square root of time provides the corresponding water permeabilityindex , given in Fig. 6 . The AAS concrete showed an increase of water permeabil- ity index with age, but is classified as low water permeable concrete at all ages asthe WPI did not exceed 1.3 /C210 /C07m3/pmin . The FAGP concrete had a signif-Table 1 Chemical composition. Material By weight (%) Loss on Ignition (%) SiO 2 Al2O3 Fe2O3 CaO P 2O5 TiO 2 MgO K 2OS O 3 MnO Na 2O GGBS 36.9 14.2 0.3 36.0 0.4 0.6 5.1 0.1 6.1 0.4 0 0.3 FA 70.3 23.1 1.4 0.2 0.2 2.6 0.6 0.9 0.2 0 0.4 2.0 Table 2 Physical and mineralogical properties. Properties investigated GGBS FA Specific Gravity 2.85 2.12 BET Surface Area, (m2/kg) 3852 1876 Fineness (%) at 5 mm 20.9 22.7 at 10 mm 43.5 43.0 at 20 mm 71.9 63.0 at 45 mm 96.9 81.8 at 75 mm 100.0 91.2 Amorphous content (%) 71.7 66.3Crystalline content (%) 28.3 33.7 Table 3 Mix design details (kg/m3). Concrete GGBS (kg)FA (kg)Aggregates (kg) Alkali Activator (kg) Added Water (kg)Total Water (kg)Total Solid (kg) Sand 7 mm 10 mm Na 2SiO 3 (Water)Na2SiO 3 (Solid)NaOH (Water)NaOH (Solid) AAS 415 – 784 346 693 40 31 29 17 136 205 463 FAGP – 409 686 303 606 114 90 80 49 10 204 548274 A. Wardhono et al. / Construction and Building Materials 143 (2017) 272–279 icantly higher water permeability index than AAS up to 90 days and was well above the minimum limit of low water permeable concrete . However, it dramatically decreased by 180 days and further reduced at later ages, having a lower value thanAAS concrete at 360 and 540 days. This is again consistent with on-going geopoly-merization and in agreement with the corresponding UPV and strength data. Fig. 6 shows the ultrasonic pulse velocity (UPV) changes with the age of AAS and FAGP concrete. Generally, the pulse velocity values represents the uniformity and the presence of defects in the microstructure and pore-structure, such as voidsand cracks , which directly influence to the permeation properties of concrete. The FAGP concrete displays an increase of UPV with age while AAS has a fall from3.91 to 3.62 between 28 and 540 days. The standard pulse velocity of PC concretegenerally falls in the range 3.5 to 4.5 km/s , which can be categorized as being in good condition which implies that the concrete is free from any large voids or cracks that may affect the long term structural reliability. It is interesting to note that while AAS shows a decrease of UPV with time, all the data points are wellabove the 3.5 km/s. The FAGP concrete is identified as poor quality concrete withUPV values below 3 km/s at 56 days, however it obtained a value of 3.5 km/s of UPV by 540 days. Overall the data from the permeability properties would further suggest ongo- ing geopolymerization and concurrent gel formation resulting in a densermicrostructure and pore-structure for FAGP concrete over the 540 days. However, the permeation properties would suggest that there is no improvement in the AAS concrete beyond 90 days, with even a slight deterioration observed. 5. Discussion 5.1. Microstrucre The microstructural development of FAGP concrete at 28, 56, 90, 180, 360 and 540 days is displayed in the SEM images in Fig. 7 . A non-uniform, heterogeneous aluminosilicate gel matrix with a numbers of unreacted/partially reacted FA particles was observed in the microstructure at 28 and 56 days. This was com- prised of unreacted fly ash particles that were separated from the geopolymeric binder, indicating weak adhesion between the gel and the particles. In addition there were also partially dissolved particles embedded in the precipitated gel. The unreacted/partially2308232123262453245724582000210022002300240025002600 28 56 90 180 360 540Dry density (kg/m3) Duration (days)(a) FAGP AAS25.131.133.239.541.340.4 40.428 56 90 180 360 540Compressive Strength (MPa) Duration (days)(b) FAGP AAS Fig. 1. (a) Density and (b) Compressive strength development vs time.5.36.67.26.05.85.4 5.228 56 90 180 360 540Flexural Strength (MPa) Duration (days)(a) FAGP AAS2.43.44.13.33.33.200.511.522.533.544.55 28 56 90 180 360 540Spliting Tensile Strength (MPa) Duration (days)(b) FAGP AAS Fig. 2. (a) Flexural strength and (b) Splitting tensile strength development vs time.98541446215942267682236117446050001000015000200002500030000 28 56 90 180 360 540Elastic Modulus (MPa) Duration (days)FAGP AAS Fig. 3. Elastic modulus development vs time.A. Wardhono et al. / Construction and Building Materials 143 (2017) 272–279 275 reacted FA particles behave as composites. These composites and the interface between them and the geopolymer matrix is hypoth- esized as an area of weakness and thus has a significant bearing on the overall strength of the concrete . Moreover micro cracks were distributed throughout the gel matrix. It was noted that more cracks and greater crack widths were observed in the 28 and 56 day specimens compared to those at latter ages. The occurrenceof these micro cracks was most likely due to evaporation of the water and self-desiccation during the heat curing stage. Further- more, the raw FA contains 2% unburnt carbon, Table 1 . The unburnt carbon acts as an inert particulate in the gel matrix, which can sup- port the crack propagation. It is hypothesized that a combination of the unreacted FA particles coupled with the presence of the micro cracks resulted in the initial lower compressive strength of FAGP. It is understood that the Si/Al (atomic) ratio determines the structure of the geopolymer backbone [52,53] . In this study, Si/Al ratio of FAGP concrete ranged between 4.27 and 3.41 over the 28–540 days period. As such, the geopolymer structure was inferred to be polysialate–disiloxo (Si–O–Al–O–Si–O–Si–O). In FAGP concrete, the Si/Al ratio decreased with age indicating an on-going geopolymerization process beyond the 90 day time per- iod, with continuous gel formation along with incorporation of alu- mina into the silicate backbone. The FAGP microstructure shows a decrease in the number of unreacted particles at 360 and 540 days, further substantiating the indication that there has been some additional geopolymeriza- tion and gel formation. The SEM indicated the gel had diffused through the surface covering and coalescing the remaining par- tially reacted FA spheres together. The gel was also observed to fill the interior voids, resulting in the formation of a semi- homogeneous, but highly compacted dense microstructure at 540 days. The decrease in the Si/Al ratio coupled with this semi- homogeneous and compact microstructure observed is hypothe- sized as the reason that the FAGP had a significantly improved per-formance at a later age, compared to the initial 90 days period. The microstructural development of AAS concrete is displayed inFig. 8 . A fairly uniform, but heterogeneous gel matrix can be seen at 28 days. Most of the slag grains have been partially dissolved by the alkali solution, forming a C-S-H gel with the silica from the activator solution. Moreover, several small micro cracks had been formed on the surface of the unreacted/partially reacted slag grains. This is attributed to a rapid reaction between slag and alka- line activator in the initial period [23,25] . The microstructural development observed shows that in the period from 28 to 90 days additional C-S-H gel, due to the dissolution of remaining unre- acted/partially reacted slag grains, with significantly less unre- acted/partially reacted grains being observed at 90 days. This resulted in formation of densely packed microstrucre at 90 days. Whilst this matrix contains fewer micro cracks compared to those observed at 28 days, the width of the cracks at the interface of C-S- H gel and partially reacted slag particles is wider. The formation of micro-cracks at the slag grain/aggregate interface has also been reported by Collins and Sanjayan .14.3914.2913.978.159.059.77 10.728 56 90 180 360 540AVPV (%) Duration (days)(b) FAGP AAS7.757.226.82 6.744.794.955.3 5.3628 56 90 180 360 540Water Absorption (%) Duration (days)(a) FAGP AAS Fig. 4. (a) Water absorption and (b) AVPV vs. time.1.4130.8370.5740.5410.6450.8020.00.20.40.60.81.01.21.41.6 28 56 90 180 360 540Water Permeability Index x 10-7(m3/√√min) Duration (days)FAGP AAS Fig. 5. Water permeability index vs time.2.93.13.53.913.753.6900.511.522.533.544.5 28 56 90 180 360 540Ultrasonic Pulse Velocity (km/s) Duration (days)FAGP AAS Fig. 6. Ultrasonic pulse velocity vs time.276 A. Wardhono et al. / Construction and Building Materials 143 (2017) 272–279 It is noted that there is a significant discrepancy of AAS microstructure before and after 90 days. The uniformity and den- sity of the gel matrix drastically reduced between 90 and 180 days, and a less compacted, loosely packed gel matrix was formed at 180 days. It is hypothesized that as the reaction continued it pro- duces additional C-S-H gel which led to the formation of wider cracks at 180 days. These cracks were seen to further increase at 360 and 540 days. This is attributed to the combined effect of dis- joining pressure and the self-desiccation effect. This resulted in for- mation of less dense small gel units, as observed at 360 and 540,days rather than an interconnected dense gel matrix, as produced in the initial 90 days. Data demonstrated in this study is consistent with those reported by literature [19,20,54] up to first 90 days. This improve- ment continues in similar manner up to 540 days. In contrast, the AAS concrete shows better initial performance than FAGP concrete, with some increase in performance, other than elastic modulus, up to 90 days, which is in agreement with literature [23,32] . However, it is noted that over the longer term data (360 and 540 days) the AAS concrete shows a reduction in engineering properties, which 50µm 50µm 50µm 50µm 50µm 50µm28d 540d 360d90d 180d56d Si/Al=4.27 Si/Al=4.05 Si/Al=3.82 Si/Al=3.76 Si/Al=3.51 Si/Al=3.41Unreacted FAN-A-S-H gelMicro-cracks Dense gel matrix Fig. 7. Microstructural development of FAGP concrete. 28d 56d 90d 180d 360d 540d50µm 50µm 50µm 50µm 50µm 50µmUnreacted Slag grains C-S-H gel Micro-cracks Expanded cracks Fig. 8. Microstructural development of AAS concrete.A. Wardhono et al. / Construction and Building Materials 143 (2017) 272–279 277 commenced post 90 days. The contradictory behaviour is reflected in the changes observed in their microstructure between 90 and 540 days, Figs. 7and8. 5.2. Tensile strength and elastic modulus The tensile strength (i.e. flexural and splitting strength) of FAGP and AAS concrete is strongly dependent on the gel-aggregate bond strength. The gel-aggregate zone is critical because it is known to have a different microstructure from the bulk of hardened gel paste and the interface is also considered as the specific location of early cracking. This is primarily caused by incomplete packing of unre- acted fly ash/slag grains in the transition between the gel paste and coarse aggregates as the gel formation is strongly dependent on the degree of alkali reactivity of fly ash/slag particles. In this study, both fly ash and slag contain approximately similar amor- phous content, Table 2 , which is an important factor for alkali dis- solution and gel precipitation [55,56] . However, recent research [54,57,58] has shown that it is not only the total amorphous con- tent but also the distribution of this in the fly ash/slag grains that influences the dissolution process and variances in microstructure formation. On the other hand, the elastic modulus of FAGP and AAS concrete is affected by the modulus of aggregate and gel paste. FAGP concrete had a less compacted heterogeneous microstructure up to 90 days. However, beyond 90 days the microstructure was significantly denser, consistent with continuing dissolution and gel formation increasing the physical solidity. The improvement in the gel-microstructure, as well the gel-aggregate interfacial zone is hypothesized as the reason for the observed improved elastic modulus and tensile strength development in FAGP concrete between 28 and 540 days. Conversely, the AAS concrete had a dense gel matrix with less micro cracks in the first 90 days, butthese were observed to increase over time. This propagation of the cracks and the associated reduction of packing density of the gel matrix is hypothesized as the cause the lack of development in the compressive and tensile strength, as well the significant fall of elastic modulus over time. 5.3. Transport properties The reaction products and the packing density are crucial in determining the water absorption. Water absorption is primarily governed by the capillary suction. Capillary suction is governed by the connectivity of the concrete surface to the bulk concrete via the pores in the gel paste. The FAGP concrete however dis- played a high AVPV percentage, which is also dependent on the interconnectivity of the capillaries in the concrete. The FAGP con-crete also displayed high water permeability characteristics at 90 days, but it significantly reduced with time and behaved as low permeable concrete at 360 and 540 days. In the water perme- ability test, the applied pressure is the principal driver of water ingress rather than capillary suction, and this would give an indi- cation about the overall image of pore-structure throughout the concrete specimen. This reduction in water permeability with time is attributed to the development of sodium-aluminosilicate gel during the on-going geopolymerisation process. The additional gel fills the interface between the geopolymer pastes and the aggregates, and reduces the volume of the pore-structure leading to a denser the microstructure, as reflected in the increase of UPV values over time. It is noted that AAS concrete had lower permeation characteris- tics than FAGP up to 90 days. However, a considerable increase of water absorption, interconnectivity of pore-structure and water permeability in AAS concrete is observed after 90 days, which cor- relates with its microstructural changes. The increase in crackwidths observed in the gel matrix, Fig. 8 , would lead to an increase in the permeation characteristics at later ages. 5.4. Comparison The overall trends observed for the mechanical and permeation properties show that the FAGP concrete is behaving in a similar manor to PC concrete, where the engineering properties improve with time. The AAS concrete, however, does not follow these trends with no increase in performance beyond 90 days, and a slight decrease actually observed in a number of these properties. In addition to the mechanical properties the long term durability of AAS and FAGP concrete is dependent upon the permeability char- acteristics of concrete which is associated with the ability of the surface layer to resist the penetration of water, carbon dioxideand water-borne chlorides into the concrete and initiate reinforce- ment corrosion. The rate of this is a function of the packing density of C-S-H/aluminosilicate gel matrix, the porosity and the connec- tivity of the pore structure. The data obtained up to 540 days sug- gests that FAGP concrete improves resistance to water permeation with age due to on-going geopolymerization and give a perfor- mance comparable with PC and blended cement concretes. Thus while the AAS concrete shows significantly greater compressive, tensile and flexural strength, coupled with a higher modulus of elasticity and lower water absorption, AVPV and water permeabil- ity in the initial 90 days and while the overall performance com- pared with similar strength PC concrete remains good, by 540 days only the compressive strength remains above the FAGP concrete, and the water absorption and AVPV lower, with the mod- ulus of elasticity similar and the flexural strength less and the water permeability higher than that of the FAGP. 6. Summary and conclusions The compressive strength, flexural strength, splitting tensile strength, elastic modulus, water absorption, pore- interconnectivity and water permeability for AAS and FAGP con-cretes were studied experimentally up to 540 days. The principal conclusions from the work presented are: /C15Compressive strength of FAGP and AAS concretes ranged between 22.2–33.2 MPa and 39.5–40.4 MPa from 28 to 540 days, respectively. The order of /C2448% and /C242% strength increase is observed in two concretes, respectively during this period. /C15FAGP concrete showed /C2453% flexural strength increase between 28 and 540 days, compared to the /C2413% decrease in AAS con- crete. Moreover, FAGP concrete achieved a twofold splitting tensile strength evolution during this time while AAS remained constant throughout. /C15AAS showed significantly higher early stiffness (i.e. elastic mod- ulus) than FAGP concrete in the first 90 days. However, the elas- tic modulus of AAS concrete drastically reduced with time and displayed /C2443% drop from 28 to 540 days, opposed to that of FAGP which had a /C2498% increase in the same period. /C15FAGP concrete had a higher water permeability index than AAS in the first 90 days, but significantly reduced with the age. The water permeability of the AAS increased with time, but was classified as low permeable concrete over the entire testing period. /C15The combined effect of disjoining pressure and self-desiccation due to the on-going reaction and C–S–H gel formation causing the propagation of wider cracks in microstructure, is hypothe- sized as the reason for the reduction in engineering perfor-mance observed for the AAS concrete over the long term.278 A. Wardhono et al. / Construction and Building Materials 143 (2017) 272–279 /C15The increase of in the packing density of the aluminosilicate gel matrix is hypothesized as positively influencing the elastic modulus and strength development in FAGP concrete observed between 90 and 540 days. Acknowledgments Materials support from the Independence Cement Pty. Ltd. Aus- tralia and PQ Australia for carrying out this research project is gratefully acknowledged. The authors also wish to acknowledge the X-ray facility and Microscopy & Microanalysis facility provided by RMIT University and the scientific and technical assistance.

Characteristic compressive strength of a geo polymer concrete M. Padmakar⇑, B. Barhmaiah, M. Leela Priyanka Department of Civil Engineering, Vignan’s Institute of Information Technology (A), Duvvada, Visakhapatnam-530049, AP, India article info Article history: Received 11 July 2020Accepted 27 July 2020Available online xxxx Keywords: Geo-polymer concreteIndustrial by productsNaOH & Na2SiO3abstract The principle issue the world confronting today is natural contamination. Because of industrialization, there is tremendous discharge of ozone harming substances (for example CO2) into condition. We can diminish the impact of contamination on the earth by expanding the use of modern results. There comesthe idea ‘‘GEO POLYMERS CONCERTE”. In this entire cement content is swapped with engineering by products. We are replacing with Ground Granulated Blast-furnace Slag (GGBS), silica flumes & gypsum. Alkaline liquid like NaOH & Na2SiO3 are used for binding of materials. On an experimental basis weselected the proportions 1:1.5:3 and 1:1:2 and we are considering 9 M and 13 M of NaOH and 20% and40% concentration of Na2SiO3. This examination researches the trademark compressive quality of geopolymer concrete by throwing solid shapes and chambers and discovering the trademark compressive qualities at 28 days utilizingencompassing relieving. We are replacing entire cement content with GGBS (70%) and silica fumes (30%). From the outcomes we see that trademark compressive superiority of Geopolymer dense additions with the development in sodium silicate fixation and most extreme happened at 40% of Na2SiO3 and theannouncement ‘‘compressive strength of cylinder is 0.8 times the compressive strength of cube” is not valid for geopolymer concrete. We also found that geo-polymerization process is sensitive with temperature./C2112020 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific committee of the International Conference on Newer Trends and Innovation in Mechanical Engineering: Materials Science. 1. Introduction The term geopolymer concrete is first begat by Davidovits in 1978 to speak to an expansive scope of materials described by chains or system of inorganic atoms. These are chains or system of mineral atoms connected with covalent bonds. These are deliv-ered by polymeric response of basic arrangements with source materials of land birthplace or modern results. The polymerization technique includes a bigheartedly quick blend response under sol- uble complaint (i.e, NaOH, Na2SiO3) on Si-Al raw materials that carry about a three dimensional polymeric hawser and circle con- struction encompassing of Si-O-Al-O promises. Geopolymers like- wise show comparative or fairly better designing properties thought about than customary Portland concrete Tables 1-9 . This paper compares the compressive asset of Geo polymer material having definite proportions of GGBS and silica fume under selected proportions 1:1.5:3 and 1:1:2 and considering 9 M and13 M of Na1OH and 20% and 40% concentration of Na2SiO3 Figs. 1-4 . 2. Review of studies In the wake of doing writing review we come to realize that, Water retention property of a geopolymer concrete is lesser than the ostensible cement, the compressive excellence and split springiness, flexural superiority of geopolymer material higher than the typical cement . Slag as a piece of glide debris folio is viable to quicken situation period of geopolymer material in encompassing ailment . In the event that mass proportion of sodium silicate to sodium hydroxide is 2.50, ideal compressivesuperiority of geo polymer material is watched . Increment of GGBS in the glide debris based geopolymer blend decreases the usefulness and setting period . Blends having antacid activator arrangement with sodium silicate to sodium hydroxide proportion of 2.5 demonstrated less droop and setting period than those with 1.5 and 2.0 . Higher extent of sand, lower solid thickness and 2214-7853/ /C2112020 Elsevier Ltd. All rights reserved. Selection and peer-review under responsibility of the scientific committee of the International Conference on Newer Trends and Innovation in Mechan ical Engineering: Materials Science.⇑Corresponding author. E-mail address: padmakarmaddala@gmail.com (M. Padmakar).Materials Today: Proceedings xxx (xxxx) xxx Contents lists available at ScienceDirect Materials Today: Proceedings journal homepage: www.else vier.com/locate/matpr Please cite this article as: M. Padmakar, B. Barhmaiah and M. Leela Priyanka, Characteristic compressive strength of a geo polymer concrete, Materia ls Today: Proceedings, higher volume of cover glue in solid blend are the explanations behind the lower modulus of versatility of super-serviceable cement . Geopolymer concrete containing silica smolder invigo- rated higher compressive . 3. Experimental details 3.1. Materials We are utilizing GGBS, Silica smoke and Gypsum as our geopolymer solid covers. Ground-granulated influence heater slag (GGBS or GGBFS) is learned by quenching liquid iron slag, silica fury which is a side-effect of silicon composites fabricating ven- tures and it is created when SiO gas, radiates as the quartz is decreased, blends in with oxygen. Here SiO is oxidized to SiO2 con- solidating into the round particles of silica smolder framing the sig-nificant piece of the smoke or smoke from the heater and gypsum which goes about as activator for polymerization process. Soluble base arrangements like sodium silicate and sodium hydroxide (drops) for planning of geopolymer concrete. We can likewise uti- lize potassium silicate and potassium hydroxide however they are costly. 4. Methodology Collection of materials which are required for preparation of geopolymer concrete and conduct different tests for materials. Cal- culation of mix proportion for binder materials, fine aggregate and coarse aggregate and estimate the quantities of materials required. We are using 9 M and 13 M molarities of NaOH and considered 20% and 40% of Na 2SiO 3for preparation of geopolymer concrete. The alkali binder ratio is assumed to be 0.8.Based on the alkali binder ratio estimate the quantities of Na 2SiO 3and NaOH. Do the morality calculations required for estimation of water and NaOH flakes quantity using the following table. The alkali solution required to be mix with estimated water quantity is prepared one hour before the preparation of geopoly- mer concrete. Grease should be applied to the moulds and required proportion geopolymer concrete should be placed in to these moulds. The moulds are left for ambient curing and demoulded after 48 hrs. The moulds are tested after 28 days using Universal Testing Machine. 5. Results 5.1. Conclusions From the results of our experimental study which we conducted on two different mix proportion i.e. 1:1.5:3 and 1:1:2, we got higher compressive strength values for 1:1:2. From the compressive strength values of cubes & cylinders we come to know that 13 M of NaOH and 40% of Na2SiO3 gave better results than all other considered mixes. The proclamation ‘‘compressive quality of chamber is 0.8 occa- sions the compressive quality of 3D shape” isn’t substantial for geopolymer concrete. Compressive superiority of Geopolymer solid additions with the growth in sodium silicate fixation and most extreme occurred at 40% of Na2SiO3Table 1 Physical Properties of GGBS (Sri Vishnu Sai Saravana Enterprises, Auto Nagar, which istaken from Rajiv Nagar area, Visakhapatnam.) Bulk Density 1.35 g/Cm3 Compressive Strength 31 N/mm2 Slump value 70 mmSpecific Surface Area 275 Cm2/gConsistency 33%Permeability 320 mm2/Kg Table 2 Physical Properties of Silica Fume (BOB Trading Company, Auto Nagar,Visakhapatnam). Density (g/cm3) 2.25 Specific External Area (m2/g) 0.1–0.2 Particle Size (Mm) 17–20 Table 3 Chemical Properties of Silica Fume (BOB Trading Com-pany, Auto Nagar, Visakhapatnam) Compounds % Of Composition SiO 2 92.57 Al2O3 0.33 Fe2O3 0.047 CaO 0.22SO 2 0.26 K2O 0.658 Table 4 Properties of sodium hydroxide (Gupta Chemicals, Barracks, Visakhapatnam) Appearance Flakes NaOH (%mass) 99.51 Na2CO3(%mass) 0.35 Cl (%mass) 0.05 SO4 0.005 SiO 4 0.004 Iron 8 ppm Table 5 Physical properties of sodium silicate (Gupta Chemicals barracks, Visakhapatnam) Appearance Color Less Viscous Liquid Specific Gravity 2.69 MgO 9%SiO 2 28% Solids 35–40%Table 6 Preparation of 1 kg SHS (in grams) Molarity M mole/I SH Solids Water Sodium Hydroxide Solution (SHS) 1 39 961 1000 2 74 926 10003 108 892 10004 140 860 10005 171 829 10006 200 800 10007 228 772 1000 8 255 745 1000 9 281 719 100010 306 694 100011 331 669 100012 354 646 100013 377 623 100014 400 600 1000 15 422 578 1000 16 443 557 100017 464 536 100018 485 515 100019 505 495 1000M. Padmakar et al. Materials Today: Proceedings xxx (xxxx) xxxTable 7 Estimated quantities of various materials: Mix proportion NaOH Na2SiO3 GGBS Silica Fume GYPSUM COARSE AGGREGATEFine Aggregate MASS OF NaOH MASS OF Na2SiO3 Molarity % 2O mm 10 mm W shf W water W ss W water 1:1.5:3 9 M 20% 7.86 3.366 0.591 21.29 14.19 17.74 0.252 0.647 0.45 1.8 40% 7.86 3.366 0.591 21.29 14.19 17.74 0.252 0.647 0.90 1.35 13 M 20% 7.86 3.366 0.591 21.29 14.19 17.74 0.339 0.56 0.45 1.8 40% 7.86 3.366 0.591 21.29 14.19 17.74 0.339 0.56 0.90 1.35 1:1:2 9 M 20% 11.01 4.788 0.81 19.5 12.99 16.26 0.40 1.039 0.72 2.89 40% 11.01 4.788 0.81 19.5 19.5 16.26 0.40 1.039 1.44 2.16 13 M 20% 11.01 4.788 0.81 19.5 19.5 16.26 0.549 0.901 0.72 2.89 40% 11.01 4.788 0.81 19.5 19.5 16.26 0.549 0.901 1.44 2.16 Table 8 Compressive strength of cubes after 28 days For M25 Concrete9M 13 M 20% Na2siO3and 80% water 24.4 25.5 26.6 20% Na2siO3 and 80% water 25.3 27.2 27.5 40% Na2siO3and 60% water 25.7 27.1 28.8 40% Na2siO3 and 60% water 30.0 26.6 40.6 For M20 Concrete9M 13 M 20% Na2siO3 and 80% water 20 21.7 22.0 20% Na2siO3 and 80% water 21.7 22.1 22.2 40% Na2siO3 and 60% water 20.4 20.8 22.2 40% Na2siO3 and 60% water 22.6 23.1 24.2 Table 9 Compressive strength of cylinders after 28 days For M25 Concrete9M 13 M 20% Na2siO3and 80% water 12.8 13.4 14.0 20% Na2siO3 and 80% water 13.3 14.3 14.4 40% Na2siO3and 60% water 13.5 14.3 15.1 40% Na2siO3 and 60% water 15.8 14 21.3 For M20 Concrete 9M 13 M 20% Na2siO3 and 80% water 10.5 11.4 11.6 20% Na2siO3 and 80% water 11.4 11.6 11.640% Na2siO3 and 60% water 10.7 10.9 11.6 40% Na2siO3 and 60% water 11.9 12.1 12.7 Fig. 3. Strength vs Na 2SiO 3and Water percentage for M-20 grade concrete. Fig. 1. Strength vs Na 2SiO 3and Water percentage for M-25 grade concrete. Fig. 2. Strength vs Na 2SiO 3and Water percentage for M-25 grade concrete. Fig. 4. Strength vs Na 2SiO 3and Water percentage for M-20 grade concrete.M. Padmakar et al. Materials Today: Proceedings xxx (xxxx) xxxBased on the visual perception usefulness of Geopolymer solid increases with the growth in sodium silicate focus We found that geopolymerization process is sensitive with temperature. Declaration of Competing Interest The authors declare that they have no known competing finan- cial interests or personal relationships that could have appeared to influence the work reported in this paper.

Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 6 | June -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 5720 Mechanical Properties and Flexural Performance of Geopolymer Concrete R.Prakash 1Assistant Professor, Department of Civil Engineering, A. C. College of Engineering and Technology, Karaikudi - 630 003, Tamilnadu, India ----------------------- -------------------------------------------- --***-------------------------------------- ------------------------------- Abstract - The Ordinary Portland Cement (OPC), which is widely used material not only consumes significant amount of natural resources an d energy but also pollutes the atmosphere by the emission of CO2, So reduce this ill effect, the search for alternative result is geopolymer concrete. In this work, low calcium class F fly ash is used as the base material. This paper presents the results o f an experimental investigation to determine the performance characteristics of geopolymer reinforced concrete. Two kinds of systems are considered in this study using 100% replacement of cement by ASTM class Fly ash. The beams were made with Geopolymer co ncrete having compressive strength in the range of M20 - M35 by heat curing. The ratio between sodium hydroxide to sodium silicate solution is 1:2.5. The specimen was cured at 60˚C for 24 hrs. The compressive strength test was performed after the curing pe riod and strain was also measured using LVDT. An empirical formula is derived for fly ash based Geopolymer concrete using the results from experimental work. Key Words : Geopolymer concrete, Class F Fly ash, Compressive strength, Flexural strength, Elasti c Modulus . 1. INTRODUCTION After wood, concrete is the most often used material by the community. Concrete is conventionally produced by using the Ordinary Portland cement (OPC) as the primary binder. The environmental issues associated with the productio n of OPC are well known. The amount of the carbon dioxide released during the manufacture of OPC due to the calcination of limestone and combustion of fossil fuel is in the order of one ton for every ton of OPC produced. In addition, the amount of energy r equired to produce OPC is only next to steel and aluminium. On the other side, the abundance and availability of fly ash worldwide create opportunity to utilise this by - product of burning coal, as partial replacement or as performance enhancer for OPC. Fl y ash is itself does not possess the binding properties, except for the high calcium or ASTM Class C fly ash. However, in the presence of water and in ambient temperature, fly ash reacts with the calcium hydroxide during the hydration process of OPC to for m the calcium silicate hydrate (C -S-H) gel. This pozzolanic action happens when fly ash is added to OPC as a partial replacement or as an admixture The binder produced in this case is due to polymerisation. Davidovits (1999) in 1978 named the later as Geop olymers, and stated that these binders can be produced by a polymeric synthesis of the alkali activated material from geological origin or by -product materials such as fly ash and rice husk ash. However, not a great deal was known regarding using the geopo lymer technology to make fly ash - based geopolymer concrete. The research reported in this thesis was dedicated to investigate the process of making fly ash -based geopolymer concrete and the short -term engineering properties of the fresh and hardened concre te. 2. MATERIALS AND MIX PROPORTION S The materials used to making geopolymer concrete were Fly Ash, Sand, Coarse aggregate, and alkaline solution such as sodium hydroxide solution and sodium silicate solution as binders and water as workability measure. 2.1 Fly Ash Fly Ash obtained from Mettur power plant was used as 100% replacement of cement. Table -1: Physical Properties of fly ash Properties Value Finess Modulus 7.86 Sp.Gravity 2.30 Table -2: Physical Properties of fly ash Chemical Prope rties minimum % by mass As per IS 3812 - 1981 Fly Ash Mettur Power Plant SiO 2+Al 2O3+FeO 3 70 90.5 SiO 2 35 58 CaO 5 3.6 SO 3 2.75 1.8 Na 2O 1.5 2 L.O.I 12 2 MgO 5 1.91 Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 6 | June -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 5721 2.2 Fine aggregate Natural river sand with fineness modulus of 2.64. Its gradation meets zone II of IS 383 (1970) requirements. Specific gravity of sand is 2.65. 2.3 Coarse Aggregate Crushed blue granite stones were passed through 20mm sieve and meets graduation requirements of IS 2386(1963). The apparent specific gravity is 2.83 and f ineness modulus is 6.40. 2.4 Sodium hydroxide (NaOH) Since geopolymer concrete used in this study is homogeneous material and its main process to activate the sodium silicate Pellet Sodium Hydroxide is recommended to use the lowest cost i.e. up to 94% to 96 % purity. Fig -1: Sodium hydroxide 2.5 Sodium silicate In present investigation sodium silicate 2.0 (ratio between Na2O to Sio2) is used. A per the manufacture, silicate were supplied to the detergent company and textile industry as bonding agent. Same sodium silicate is used for the making of geopolymer concrete. The chemical properties and the physical of the sodium silicates were given below. Table -3: Physical and Chemical properties of Na 2Sio 3 Properties Value Na 2O 15.9 % SiO 2 31.4 % H2O 52.70 % Appearance Liquid (gel) Colour Light Yellow Liquid (gel) Boiling Point 102oC for 45% aqueous solution Molecular Weight 184.04 Specific Gravity 1.60 2.6 Mix Proportion s The mix design in the case of g eopolymer concrete is based on convenient concrete with some modification. In the case of conventional concrete, the materials proportion can be found out of required strength using the code. Table -4: Mix Proportion for geopolymer Concrete Mix Na 2Siokg/m3 NaOH kg/m3 Exta wate r ml Fly ash kg/m Fine Aggr. r kg/m3 Coars e Aggr kg/m3 MS 1 168.00 67.20 12.6 420 710 1061 MS 2 217.85 87.14 15 500 621.8 5 928.5 MS 3 239.64 95.85 16.5 550 535.7 5 800.0 0 MS 4 274.28 109.7 1 12.8 640 434.8 1 773.02.7 Preparation of alkaline liquid Sodium Hydroxide Solution Sodium Hydroxide pellets are taken and dissolved in water at the rate of 16 molar concentrations. It is strongly recommended that the sodium hydroxide, solution must be prepared 24 hours prior to u se and also if it exceeds 36 hours it terminate to semi solid liquid state. Hence the prepared solution should be used within the time period. 2.8 Molarity Calculation The solids must be dissolved in water to make a solution with the required concentratio n. The concentration of sodium hydroxide solution can vary in different molar. The mass of NaOH solids in a solution varies depending on the concentration of the solution For instance, NaOH solution with a concentration of 16 molar consists of 16 x 40 = 640grams of NaOH solids per litre of water, were 40 is the molecular weight of NaOH. Note that the mass of water is the major component on both the alkaline solutions. The mass of NaOH solids was measured as 444 grams per Kg of NaOH solution with Concentrati on of 16 molar . 2.9 Alkaline Liquid Generally alkaline liquids are prepared by mixing of the sodium hydroxide solution and sodium silicate solution at the room temperature. When the solution mixed together the both solution start to react that is polymeri zation take place. It liberate large amount of heat so it is recommended to leave it for about 20 minutes thus the alkaline liquid is ready as binding agent. 2.10 Casting and Curing It was found that the fresh fly ash based geopolymer concrete was dark in colour (due to the dark colour of the fly ash). The amount of water in the mixture played an important role on the behavior of fresh concrete when the mixing time was long. The sodium hydroxide available in pellets form it is dissolved in water. Morality to be used in the concrete is 16 molar in which 444 grams of NaOH solids dissolved in 556 grams of water. Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 6 | June -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 5722 Mix sodium hydroxide solution and sodium silicate solution together at least one day prior to adding to the dry materials. Mix all dry materials in t he pan mixer for amount three minutes. Add the liquid compound of the mixture at the end of dry mixing and continue the wet mixing for another 2 minutes. The Geopolymer specimen were cast and placed inside a jute canvas or tarpaulin. The entire specimen wa s kept inside the heat curing chamber at 60oC and a temperature indicator was also placed outside the set up. The canvas should be so tight such that the heat can’t come out of the heat curing set up. The beams were cured for 24 hours. Geopo lymer spec imens shoul d be cured at elev ated tempera ture in a dry enviro nment to prev ent excessive evaporation Geopo lymer concrete did not harden immediately at room temperature. When the room temperature was less than 30oC the hardening did not occur at least for 24 hour s. Also the handling time is a more appropriate parameter (rather than setting time used in the case of OPC Concrete) for fly ash based geopolymer concrete. 3. Test Results and Discussions 3.1 Compressive strength The compressive strength of GPC after 24 hours heats curing at 60oC. The average compressive strength values observed after 24 hours are given in table 5. 3.2 Flexural Strength Tests were carried out conforming to IS 516 (1959) to obtain the flexural strength of various concrete mixtures. Eigh t beams of beam 100mmx100mmx500mm were cast and the beams are shown in fig.2 The beams were tested by two points loading method in UTM. The experimental results of flexural strength are shown in table 5 Fig -2: Specimen for f lexural strength testing Table -5: Compressive and Flexural Strength of GPC Sl. No Grade Comp.strength N/mm2 Flex. Strength N/mm2 MS1 M20 25 3.5 MS 2 M25 29 3.66 MS 3 M30 37 4.13 MS 4 M35 39 4.45 3.3 Modulus of Elasticity Young’s modulus E fo r the geopolymer concrete investigated was determined at 24 hours. Tests were carried out in accordance with the Indian Standard. For each Mixture, four 100x300 mm concrete cylinders were made. Four of These cylinders were used to determine the elastic mod ulus and Poisson’s ratio. Four other cylinders were tested to determine the average compressive strength. All the specimens were capped in accordance with the Indian Standard. The range of poisons ratio falls between 0.19 and 0.22. For Portland cement conc rete, the Poisson’s ratio is usually between 0.11 and 0.21, with the most common value taken as 0.15 (Warner et al. 1998) or 0.15 for high strength concrete and 0.22 for low strength concrete (Neville 2000). These ranges are similar to those measured for t he geopolymer concrete. Table 6 shows the modulus of elasticity of concrete. Table -6: Modulus of Elasticity Sl. No Grade Modulus of Elasticity N/mm2 MS1 M20 19677 MS 2 M25 22000 MS 3 M30 24099 MS 4 M35 26030 3.4 Flexural behaviour of RCC bea m Beams of size 100mmx200mmx1800mm were tested under two point loading. The First crack load, Ultimate load and deflection at ultimate loads are tabulated in table no. 7. The reinforcement details and test set up of the beams are shown in figures 3 and 4. Figs 5 to 9 show the load versus deflection curve of the beams at mid span at all stages of loading up to failure. Fig 10 shows the crack patterns of the beams tested in the present work. All the cracks appeared between the point loads, showing that they w ere flexural ones. Fig -3: Reinforcement details Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 6 | June -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 5723 Fig -4: Test set up of GPC beam Table -7: First crack load, Ultimate load and Deflection Sl. No Grade First crack load(kN) Ultimate load (kN) Deflection at ultimate load (mm) MS1 M20 44 56 6.52 MS 2 M25 40 48 5.89 MS 3 M30 46 59 7.05 MS 4 M35 42 53 6.21 Fig -5: Load Vs deflection , M20 Fig -6: Load Vs deflection , M25 Fig -7: Load Vs deflection , M30 Fig -9: Load Vs deflection , M20 Fig -10: Crack failure patterns of the beams 3.5 Deformations at first crack The deformation at this stage o f loading is only a fraction of those occurring at the design service loads. The visible first crack loads of the beams varied from 30 -35% of the Inte rnational Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395 -0056 Volume: 0 4 Issue: 0 6 | June -201 7 www.irjet.net p-ISSN: 2395 -0072 © 2017, IRJET | Impact Factor value: 5.181 | ISO 9001:2008 Certified Journal | Page 5724 experimental failure loads. The flexure rigidity EI at the point of maximum moment was calculated by dividing the bending moment by the curvature obtained from the concrete strain reading in the compression zone. 3.6 Deformation at failure All the beams failed by yielding of tension steel. After yielding the beams exhibited significant in elastic deformations before the ultimate load was reached. 4. CONCLUSIONS 1. For all the mixes considered in this investigation, there was increase in load carrying capacity of beam for increase in grades. 2. The measured values of the modulus elasticity of fly ash - based ge opolymer concrete with compressive strength in the range of 20 to 35 MPa were similar to those of OPC concrete. The measured values are at the lower end of the values calculated using the current design Standards due to the type of coarse aggregate used in the manufacture of the geopolymer concrete. 3. The Poisson’s ratio of fly ash -based geopolymer concrete with compressive strength in the range of 20 to 35 MPa falls between 0.19 and 0.22 . These values are similar to those of OPC concrete. 4. The stress -strain relations of fly ash -based geopolymer concrete in compression fits well with the expression developed for OPC concrete as per IS 456 – 2000. 5. The compressive strength of fly ash -based geopolymer concrete is high, as in the case of Portland cemen t concrete. The measured values are higher than those recommended by the relevant Indian Standard. 5. RECOMMENDATIONS FOR FUTURE RESEARCH To date, the reaction mechanism of geopolymerisation is still not clear. Fundamental research in this area would inc rease the potential of the material. For example a study is needed to identify the scientific reason for increase in strength after a longer resting period, and to investigate the role of water in geopolymerisation. Although the present work identified man y salient parameters that influence the properties of fresh and harden fly-ash based geopolymer concrete, a large database should be built on the engineering properties of various mixtures using fly ash from different sources. Such a database may identify additional parameters, and lead to familiarise the utilisation of this material in many applications. Further research should identify possible applications of geopolymer technology. This would lead to research areas that are specifically oriented towards applications. The geopolymer technology has the potential to go beyond making concrete, there could be possibilities in other areas of infrastructure needed by the community. REFERENCES ASTM C 642, (1997), “Standards test method for density, absorption a nd voids in hardened concrete”, Annual book of American Society for Testing Materials Standards. BIS: 516 -1959 (reaffirmed 1997) “Methods of Tests for Strength of Concrete, Bureau of Indian Standards”, New Delhi. BIS: 383 -1970 (reaffirmed 1997) “Specificat ion for Coarse and Fine Aggregates from Natural Source for Concrete”, New Delhi. BIS: 12269 -1987 (reaffirmed 1999) “Specification for 53 grade Ordinary Portland Cement”, New Delhi. BIS: 456 -2000 (reaffirmed 2005) “Plain and Reinforced Concrete – Code of Pr actice”, Fourth Revision, pp.14. Djwantoro hardjito - 2004 – ACI Materials Journal - on the Development of fly ash -Based Geopolymer, Title No.101 -M52 – Technical Paper Delsye C. L. Teo1, Md. Abdul Mannan and John V. Kurian, Flexural Behaviour of Reinforced Lightweight Concrete Beams Made with Oil Palm Shell (OPS) Journal of Advanced Concrete Technology Vol. 4, No. 3, 1 -10, 2006, October 2006 / Copyright © 2006 Japan Concrete Institute Mark Anderson, Ph.D., P.E. and Maj Dov Dover, P.E - A as published in Proc eedings, NBC Defense Collective Protection Conference (COLPRO O2), Orlando, Florida, Oct 2002. Low Energy replacement for Portland cement concrete to improve resistance to chem. – bio Intrusion. Mohd Mustafa Al Bakri1*, H. Mohammed, H. Kamarudin1, Khairul Niza and Y. Zarina1, Review on fly ash-based geopolymer concrete without Portland Cement Journal of Engineering and Technology Research Vol. 3(1), pp. 1 -4, January 2011.

Workability and mechanical properties of alkali-activated fly ash-slag concrete cured at ambient temperature Guohao Fang, Wing Kei Ho, Wenlin Tu, Mingzhong Zhang⇑ Advanced and Innovative Materials (AIM) Group, Department of Civil, Environmental and Geomatic Engineering, University College London, London WC 1E 6BT, UK highlights /C15Workability of alkali-activated fly ash-slag (AAFS) concrete measured. /C15Mechanical properties of AAFS concrete measured. /C15Effects of slag content, molarity of SH, AL/B ratio and SS/SH ratio estimated. /C15Prediction equations for splittingtensile strength and flexural strengthproposed. /C15Optimal mixtures of AAFS concretefor engineering application obtained.graphical abstract article info Article history: Received 17 October 2017Received in revised form 28 February 2018Accepted 1 April 2018Available online 4 April 2018 Keywords: Alkali-activated concreteWorkabilitySetting timeStrengthOptimal mixturesabstract Alkali-activated fly ash-slag (AAFS) concrete is a new blended alkali-activated concrete that has been increasingly studied over the past decades because of its environmental benefits and superior engineer-ing properties. However, there is still a lack of comprehensive studies on the effect of different factors on the fresh and hardened properties of AAFS concrete. This paper aims to provide a thorough understanding of workability and mechanical properties of AAFS concrete cured at ambient temperature and to obtainthe optimal mixtures for engineering application. A series of experiments were carried out to measureworkability, setting time, compressive strength, splitting tensile strength, flexural strength and dynamic elastic modulus of AAFS concrete. The results showed that workability and setting time decreased with the increase of slag content and molarity of sodium hydroxide solution (SH). Compressive strengthincreased with the increase of slag content and molarity of SH as well as the decrease of alkaline activator to binder (AL/B) ratio, but it did not have significant relationship with sodium silicate to sodium hydrox- ide (SS/SH) ratio. In addition, equations provided by ACI code, Eurocode and previous researchers for ordi-nary Portland cement concrete overestimated the values of splitting tensile strength, flexural strength and dynamic elastic modulus of AAFS concrete. The optimal mixtures of AAFS concrete were set as slag content of 20–30%, AL/B ratio of 0.4, 10 M of SH, and SS/SH ratio of 1.5–2.5 considering the performancecriteria of workability, setting time and compressive strength./C2112018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( http:// creativecommons.org/licenses/by/4.0/ ). 1. Introduction Alkali-activated materials (AAM) is an inorganic binder derived by the reaction of an alkali metal source (solid or dissolved) with asolid silicate powder such as fly ash (FA) and slag . To date, AAM has been recognized as a promising alternative binder to ordinary Portland cement (OPC) because of its environmental benefits and superior engineering properties [2–5] . The manufacture of OPC is known as a significant contributor to greenhouse gas emissions accounting for around 5% of global CO 2emissions [6,7] . In compar- ison, there are about 55–75% less greenhouse gas emissions in the 0950-0618/ /C2112018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license ( ).⇑Corresponding author. E-mail address: mingzhong.zhang@ucl.ac.uk (M. Zhang).Construction and Building Materials 172 (2018) 476–487 Contents lists available at ScienceDirect Construction and Building Materials journal homepage: www. elsevier.com/loc ate/conbuildmat production of FA and slag . Thus, the application of AAM as a binder can significantly reduce the CO 2emissions of concrete production. FA has been increasingly considered as a suitable raw material for alkali-activated concrete (AAC) due to its wide availability and adequate composition of silica and alumina. Previous studies [9–19] reported that alkali-activated fly ash (AAF) concrete has excellent mechanical and durability properties when it is cured at elevated temperature. Normally, the curing temperature of 60–85 /C176C is required to activate FA as the reactivity of FA at ambient temperature is too low to be activated by alkali activators [20–22] . Such curing condition may be suitable for manufacturing precast concrete members, but it is not suitable for cast-in-situ concrete in practice. Therefore, it is vital to develop a new type of AAC without curing at elevated temperature, which will widen the practical appli-cation of AAC. In addition, the cost and energy consumption associ- ated with the heat curing pro cess will also be reduced. In order to achieve ambient curing, some researchers attempted to improve the reactivity of FA in alkaline environment . In par- ticular, one of the acceptable attempts is to add some calcium con- taining materials such as slag in AAC . The addition of slag would accelerate FA dissolution and enhance reaction products formation in room curing condition . Both the early and later age properties of AAF concrete are also significantly affected by the additional slag. Until now, an increasing number of studies have been undertaken to investigate the effect of slag on the engi- neering properties of AAF [2,3,26–30] . Nath and Sarker [2,3,31] studied the influencing factors on the fresh and hardened proper- ties of alkali-activated fly ash-slag (AAFS) concrete. It was found that the dominant influencing factors are the slag replacement level for FA along with the type and content of alkaline activator. One main limitation of this research is that the different activating conditions were not fully considered. For example, the effect of slag content on the properties of AAFS concrete may be affected by the activator with different molarity. Lee [29,32] also explored the mechanical properties of AAFS concrete and suggested a proper slag content of 15–20% of total binder considering the setting time and compressive strength of AAFS concrete. However, it should be noted that the workability of AAFS concrete was not considered in the selection of slag content. In addition, an optimal mixture of AAFS concrete should not only include the slag replacement level but also the alkaline activator to binder (AL/B) ratio, molarity of sodium hydroxide (SH) solution and sodium silicate to sodium hydroxide (SS/SH) ratio, etc. Thus, it is of importance to conduct a comprehensive research focusing on the effects of different fac- tors on the fresh and hardened properties of AAFS concrete and to evaluate the optimal mixtures by taking into account the basic performance criteria of workability, setting time and compressivestrength. The main purpose of this study is to provide a thorough under- standing of workability, setting time and mechanical properties of AAFS concrete cured at ambient temperature. Low calcium FA and ground granulated blast-furnace slag (GGBS) were used as binder materials. Alkaline activator was prepared by mixing SH and SS solution. Special attention was paid to the main influencing factors, including FA/GGBS ratio, AL/B ratio, molarity of SH and SS/SH ratio on the workability, setting time, compressive strength, splitting tensile strength, flexural strength and dynamic elastic modulusdevelopment of AAFS concrete. Splitting tensile strength, flexural strength and dynamic elastic modulus were further analysed using existing standards and codes in order to propose prediction equa- tions suitable for AAFS concrete. Finally, the optimal mixtures of AAFS concrete were obtained based on the performance criteria of workability, setting time and compressive strength. 2. Experimental program 2.1. Materials In this study, low calcium FA and GGBS were used as binder. The chemical com- positions of FA and GGBS are listed in Table 1 . The mean particle size of FA and GGBS is 26.81 and 14.77 mm, respectively. Alkaline activator (AL) was mixed by sodium hydroxide (SH) with distilled water and sodium silicate solution (SS). TheSiO 2to Na 2O ratio of SS was 2.0 with chemical composition of 30.71 wt% SiO 2, 15.36 wt% Na 2O and 53.93 wt% H 2O. Since the modified polycarboxylate-based superplasticizers (SPs) have a significant effect on the workability of AAFS ,i t was used to improve the workability of AAFS in this work. The properties of this SPs are given in Table 2 . Natural sand with a nominal maximum size of 2 mm was used as the fine aggregate. Coarse aggregates (CA) were prepared by mixingcrushed granite with nominal maximum sizes of 10 and 20 mm. Fine aggregatesand coarse aggregates were used in saturated surface dry (SSD) condition accordingto ASTM C128-15 and ASTM C127-15 , respectively. 2.2. Mixture proportions AAFS specimens with different FA/GGBS ratio, AL/B ratio, molarity of SH and SS/SH ratio were prepared and tested in this work. The optimal scope of mixtureproportions was selected according to the relevant studies [2,3,6,28,29,32] . The mix proportions of AAFS concrete are listed in Table 3 and labelled with specific codes. The labels ‘A’, ‘B’, ‘C’, and ‘D’ represent different specimen series, while the numbers, ‘10 0, ‘150, ‘200, ‘250and ‘300, stand for the percentages of GGBS replacement for FA by weight, respectively. In Series A, mixture 1 (A10) to mixture 5 (A30) referto those with GGBS content of 10%, 15%, 20%, 25% and 30% of total binder, respec-tively. The AL/B ratio in these mixtures was kept constant at 0.4 with molarity of SHand SS/SH ratio of 10 M and 2.0, respectively. In Series B, i.e., mixture 6 (B15) tomixture 8 (B25), the molarity of SH was changed from 10 M to 12 M while the AL/B ratio and SS/SH ratio were kept constant at 0.4 and 2.0, respectively. In Series C, i.e., mixture 9 (C15) to mixture 11 (C25), the AL/B ratio was changed from 0.4 to0.35 while the molarity of SH and SS/SH ratio were kept as 10 M and 2.0, respec-tively. In Series D, the SS/SH ratios for mixture 12 (D15) and mixture 13 (D25) were1.5 and 2.5, respectively. The SPs content was kept constant at 1% of the total binderfor all mixtures. As such, the effect of slag content on the engineering properties ofAAFS can be studied through Series A containing various slag content ranging from10% to 30% of binder by weight. The effect of SH molarity can be investigated using Series A and Series B containing SH molarity of 10 and 12, respectively. Series A and Series C with AL/B ratios of 0.4 and 0.35 respectively were also designed to estimatethe effect of AL/B ratio. Furthermore, Series A and Series D were conducted to eval-uate the effect of SS/SH ratio on the engineering properties of AAFS with various SS/SH ratios ranging from 1.5 to 2.5. The concrete mixtures were proportioned based on the unit volume of 1 mwhile the total binder content was kept constant as 400 kg/m3. The ingredient con- tents of binder were calculated based on their weight ratio. The total volume of aggregates was the residual volume except binder volume. The aggregates were mixed by the volume of 10 mm, 20 mm and fine aggregates with 22%, 43% and35%, respectively. In addition, the mix proportions of AAFS pastes were similar tothose of concrete mixtures excluding aggregates (see Table 3 ). Table 1 Chemical compositions (wt%) of FA and GGBS. Oxide SiO 2 Al2O3 CaO MgO K 2OF e 2O3 TiO 2 Na2OS O 3 FA 53.24 26.42 3.65 9.55 2.57 1.65 0.86 0.76 0.56 GGBS 36.77 13.56 37.60 7.45 0.55 0.41 0.79 0.25 1.82Table 2 Properties of superplasticizers. Specific gravity(25/C176C)pH (25/C176C)Content of chloride ion (%)Content of alkaline (%) 1.08 4–5 /C200.1 /C200.4G. Fang et al. / Construction and Building Materials 172 (2018) 476–487 477 2.3. Specimen preparation The mixing method of AAFS concrete is presented as follows. FA, GGBS, fine aggregate and coarse aggregate were dry-mixed for 2 min to ensure homogeneityof the mixture. Then, AL and SPs were added into mixture and mixed for another3 min. The fresh AAFS concrete was then immediately cast into three differentmoulds (cube with size of 100 mm for compressive test, cylinder with size of 100/C2200 mm for splitting tensile test and prism with size of 100 /C2100/C2500 mm for four-point bending test and dynamic elastic modulus test). For each testingage, three specimens were prepared for each mixture. All specimens were thenstored in a curing room (20 ± 2 /C176C, 60% ± 5% RH). After 24 h, the specimens were de-moulded and placed in curing room under the same environmental conditionuntil the day of testing. 2.4. Testing methods The workability of AAFS pastes was investigated using flow table test according to ASTM C230-14 , where the diameter of the paste spread in two directions at right angles was measured to calculate the flow value. Slump value as described inASTM C143-15a was conducted to determine workability of AAFS concrete, where vertical difference between the top of the mould and the displaced originalcentre of the top surface of the specimen was measured as the slump value. The ini-tial and final setting time of AAFS pastes was determined by the Vicat setting test according to ASTM C191-08 . AAFS paste was proportioned and mixed to nor- mal consistency for the setting time test according to previous research . Peri- odic penetration tests were performed by allowing a 1-mm Vicat needle to settleinto this paste. The Vicat initial setting time was the time elapsed between the ini-tial contact of raw materials and activator and the time when penetration was mea-sured to be 25 mm. The Vicat final setting time was the time elapsed between initialcontact of raw materials and activator and the time when the needle did not leave acomplete circular impression in the paste surface. According to BS EN 12390- 3:2009 , a universal testing machine was used to test the compressive strength of AAFS concrete at 1, 7, 14, 28 and 56 d, where the constant rate of loading was setas 5 kN/s. The splitting tensile strength of AAFS concrete at 7 and 28 d was testedaccording to ASTM C496-11 at a loading rate of 0.6 kN/s. The flexural strength of AAFS concrete at 28 d was determined using four-point bending test according toASTM C78-16 at a loading rate of 40 N/s. The dynamic elastic modulus of AAFS concrete was determined according to ASTM C215-14 at 7, 14, 21 and 28 d, where the fundamental longitudinal frequency along with the dimensions and mass of the specimen were used to calculate the dynamic elastic modulus. 3. Results and discussion 3.1. Workability Generally, the workability of AAC is lower than that of OPC con- crete because the presence of silicate in AAC would bring a sticky characteristic. Nevertheless, AAC can compact well on a vibrating table even for relatively low slump value. Therefore, the workabil- ity of AAC is classified based on the condition of compaction as shown below . When AAC achieves a slump value of 90 mm and over, it is regarded as a highly workable concrete. AAC with the slump values in the range of 50 mm and 89 mm is classifiedas medium workability, while AAC with slump values below 50 mm is considered as low workability due to the significant vibra- tion of compaction. Thus, in this study, this criterion was applied to identify the optimal mixture of AAFS concrete in terms of workability. 3.1.1. Effect of fly ash/slag ratio Fig. 1 shows the flow value of AAFS pastes and slump value of AAFS concrete with various slag content. Both the flow value of AAFS pastes and the slump value of AAFS concrete were influenced by the replacement level of slag in binder. The slump and flow val- ues decreased with the increase of slag content in the mixture, which is consistent with previous studies [2,3] . This can be attrib- uted to the accelerated reaction of calcium and the angular shape of slag comparing with the spherical shape of fly ash particles . However, the effect varied with the amount of slag. For Series A, the slump value of A15, A20, A25 and A30 decreased 5.40%, 17.50%, 24.62% and 25.91% respectively as compared to A10. For Series B, the slump decreasing value of B20 and B25 was 51.03% and 64.48% respectively as compared to B15. In Series C, the slump value of C20 and C25 was 6.96% and 14.55% lower than that of C15, respectively. The effect of slag at 20% replacement level appeared to be more pronounced. Moreover, it seems that the effect of slag content was more significant at higher molarity of SH (12 M) and higher AL/B ratio (0.4). 3.1.2. Effect of molarity of sodium hydroxide solution Comparing the flow and slump values of Series A and Series B in Fig. 1 , it can be found that the workability of AAFS decreased with the increase of SH molarity. All the slump value of 10 M specimens in Series A was higher than 171 mm, while the slump value of 12 M specimens in Series B was lower than 145 mm. This is mainly because that the increase of SH molarity increased the viscosityof the solution . However, the effect of SH molarity varied with slag replacement ratio. The slump value of B15 (145 mm) was decreased by 74 mm compared with that of A15 (219 mm). When the replacement of slag reached 20% and 25%, the slump values of B20 (71 mm) and B25 (51.5 mm) were decreased by 120 mm and 123 mm respectively compared with the slump value of corre- sponding 10 M specimens in Series A. The effect of SH molarity was more obvious for the mixtures with a higher slag replacement level. 3.1.3. Effect of alkaline activator/binder ratio As shown in Fig. 1 , the mixtures with AL/B ratio of 0.35 (Series C) exhibited relatively low flow value and slump value as com-Table 3 Mixtures of alkali-activated fly ash-slag concrete. Mix No. Labels Mixture proportions Concrete mixture quantity (kg/m3) FA/GGBS AL/B ratio Molarity of SH (M) SS/SH ratio FA GGBS SH SS SPs Sand CA 10 mm CA 20 mm A 1 A10 90/10 0.4 10 2 360 40 53 107 4 644 399 798 2 A15 85/15 0.4 10 2 340 60 53 107 4 646 400 8003 A20 80/20 0.4 10 2 320 80 53 107 4 648 401 8024 A25 75/25 0.4 10 2 300 100 53 107 4 650 402 8055 A30 70/30 0.4 10 2 280 120 53 107 4 652 403 807 B 6 B15 85/15 0.4 12 2 340 60 53 107 4 646 400 800 7 B20 80/20 0.4 12 2 320 80 53 107 4 648 401 802 8 B25 75/25 0.4 12 2 300 100 53 107 4 658 407 815 C 9 C15 85/15 0.35 10 2 340 60 47 93 4 651 403 806 10 C20 80/20 0.35 10 2 320 80 47 93 4 661 409 81811 C25 75/25 0.35 10 2 300 100 47 93 4 671 415 831 D 12 D15 85/15 0.4 10 1.5 340 60 64 96 4 637 395 789 13 D25 75/25 0.4 10 2.5 300 100 46 114 4 659 408 815 Note : FA (Fly Ash); GGBS (Ground Granulated Blast-furnace Slag); SH (Sodium Hydroxide); SS (Sodium Silicate); SPs (Superplasticizers); CA (Coarse Aggr egates).478 G. Fang et al. / Construction and Building Materials 172 (2018) 476–487 pared to the mixtures in Series A with AL/B ratio of 0.4. The slump value of the specimens in Series C was in the range between 67 mm and 79 mm, while the slump value of Series A was much higher than that of Series C with a range from 171 mm to 231 mm. More specifically, the effect of AL/B ratio can be observed in the mixtures with different slag replacement levels. The slump value of C15 decreased 63.95% as compared to A15, while those of B20 and B25 decreased 61.51% and 61.31% as compared to A20 and A25, respectively. It is indicated that the alkaline activator content in the mixtures plays a dominant role in workability of AAFS concrete, which is consistent with other research [2,3] . 3.1.4. Effect of sodium silicate/sodium hydroxide ratio As shown in Fig. 1 , the slump value of D15 (SS/SH ratio of 1.5) kept the highest level (228 mm). When the SS/SH ratio increased to 2.0, the slump value decreased to 219 mm (A15). Similarly, when the SS/SH ratio increased from 2.0 to 2.5, the slump value (174.5 mm) of A25 decreased to 170 mm (D25). It indicates that the workability of AAFS concrete decreased with the increase of SS/SH ratio, which is in good agreement with previous research [2,3] . Since SS is the most viscous solution in the alkaline liquid, with the increase of SS/SH ratio from 1.5 to 2.5 the viscosity of AAFS mixture tends to increase resulting in a lower workability[2,3,45] . 3.1.5. Optimal mixtures in terms of workability According to the classification of AAC workability as mentioned above, the mixtures were divided into three different categories (see Fig. 1 ). For Series A, all AAFS mixtures can be classified as highly workable concrete as the slump value of these specimens was higher than 90 mm. For Series B, only mixture 6 (B15) was defined as highly workable concrete, while other two mixtures were classified as medium workable concrete. Furthermore, all the mixtures in Series C were classified as medium workable con- crete because the slump value of these mixtures were in the range between 50 and 90 mm. For Series D, all the mixtures were consid- ered as highly workable concrete according to the classification. Therefore, the mixtures with a slag replacement level from 10% to 30%, AL/B ratio of 0.4, 10 M of SH, and SS/SH ratio in the range of 1.5–2.5 can be suggested as optimal mixtures for the sake of workability.3.2. Setting time Setting time is one of the important properties of concrete. It can be divided into initial setting time and final setting time based on the degree of rigidity. According to ASTM C403-08 , the ini- tial and final setting time can be determined as the time when the penetration resistance equals to 3.5 MPa and 27.6 MPa, respec- tively. According to BS EN 197-1:2011 , the initial setting time of OPC with strength class of 42.5 should be longer than 60 min. In this study, this criterion was used to verify the feasibility of AAFS concrete in terms of setting time. 3.2.1. Effect of fly ash/slag ratio Fig. 2 shows the variation of setting time of AAFS pastes with different slag content. Both the initial and final setting time of AAFS pastes decreased with the increase of slag replacement level. High initial and final setting time of specimens with 10% slag replace- ment can be found. When more slag was added into the mixtures the initial setting time was reduced significantly from 350 min to 77–118 min, while the final setting time was decreased from 470 min to 107–128 min. The initial setting time of A15, A20, A25 and A30 was found to be reduced by 18.57%, 62.85%, 63.14% and 65.71% and the final setting time was reduced by 31.91%, 63.83%, 64.04% and 70.21% compared to the setting time of A10, respectively. The reducing percentage of setting time was higher for the specimens with higher slag content. This is mainly ascribed to the corresponding higher amount of reactive slag, which con- tributes to the formation of C-A-S-H gel along with N-A-S-H gel at early duration, as such the reaction process is accelerated . In addition, the effect of slag content seemed more significant at higher molarity of SH (12 M) and higher AL/B ratio (0.4). 3.2.2. Effect of molarity of sodium hydroxide solution As shown in Fig. 2 , the 10 M specimen with 15% slag (A15) had an initial setting time of 285 min and final setting time of 320 min, while the 12 M specimen (B15) had a relatively higher initial set- ting time of 320 min and final setting time of 400 min. Similarly, for the specimens with 20% slag (A20 and B20), the initial setting time was found to be increased from 130 min to 183 min and the final setting time was increased from 170 min to 213 min with the increase of SH molarity from 10 M to 12 M. However, increas- ing the SH molarity in specimens with 25% slag (A25 and B25)0306090120150180 D C BFlow value (%) Flow value Slump value A 1 2 3 4 5 6 7 8 9 10 11 12 13 A10 A15 A20 A25 A30 B15 B20 B25 C15 C20 C25 D15 D25 050100150200250300 Slump value (mm)Fig. 1. Flow and slump values of alkali-activated fly ash-slag pastes and concrete.G. Fang et al. / Construction and Building Materials 172 (2018) 476–487 479 reduced the initial setting time from 129 min to 118 min and the final setting time from 169 min to 128 min. It indicated that the setting time increased with the increase of molarity of SH except for the specimens with 25% slag. Generally, with the increase of SH molarity in the mixtures the setting time is decreased if only SH is used as alkaline activator. This is because it would increase the hydroxide ion concentration and accelerate the dissolution of raw materials . Nevertheless, a mix solution composed of SS and SH was used as activator. It means that increasing the SH molarity would also affect the silica modulus (molar ratio of SiO 2/Na 2O) of the alkaline liquid, which would influence the alka- line activation process. In this study, the modulus of mixtures in Series A was decreased from 1.13 to 1.04 when the SH molarity was increased from 10 M to 12 M. Normally, a higher modulus of alkaline solution would accelerate the alkaline activation process and reduce the setting time . Thus, this may be the reason why increasing the SH molarity would increase the setting time. 3.2.3. Effect of alkaline activator/binder ratio Series A and Series C in Fig. 2 represents the setting time of specimens with AL/B ratio of 0.4 and 0.35, respectively. When the AL/B ratio was reduced from 0.4 to 0.35, the range of initial set- ting time was decreased from 129 to 285 min to 102–189 min and the final setting time was reduced from 169 to 320 min to 112– 239 min. It indicated that a lower AL/B ratio led to a shorter setting time because of the decrease of total liquid content. According to previous research, the reduction of AL/B ratio would decrease the consistency of AAFS concrete, which would result in the acceler- ated reaction of raw materials . 3.2.4. Effect of sodium silicate/sodium hydroxide ratio As the SS/SH ratio increased from 1.5 to 2.5, the setting time of specimens was significantly affected (see Series D in Fig. 2 ). For the mixtures with 15% slag (i.e. A15 and D15), the initial setting time of AAFS pastes was found to be reduced from 285 min to 132 min and the final setting time was reduced from 320 min to 172 min when the SS/SH ratio was decreased from 2.0 to 1.5. However, the initial and final setting time was decreased with the increase of SS/SH ratio from 2.0 to 2.5 (see A25 and D25 in Fig. 2 ). The different effect of SS/SH ratio on setting time may be attributed to the interactionbetween SS and SH . Thus, more information (e.g. the silica modulus of alkaline solution) is required to discuss this interaction mechanism. According to the mix proportion, the solution moduli of mixtures with three different SS/SH ratios, including 1.5 (D15), 2.0 (A15 and A25) and 2.5 (D25) were calculated to be 0.99, 1.13 and 1.9, respectively. There is not much difference between the moduli of A15 and D15, which means that the dissolute silica con- tent may be not the main reason to explain why the setting time decreased with the decrease of SS/SH ratio from 2.0 to 1.5. Thus, the main reason may be attributed to the increase of relative amount of SH. Increasing the amount of SH increases the hydroxide ion concentration in mixtures, which would accelerate dissolution of raw materials and thus reduce setting time . On the other hand, the modulus of D25 (1.9) is much higher than that of A25 (1.13), which indicates that the dissolute silica content was increased with the increase of SS/SH ratio from 2.0 to 2.5. Thus, the decrease of setting time with the increase of SS/SH ratio from 2.0 to 2.5 may be ascribed to the dissolute silica content. A higher content of dissolute silica would enhance the alkali activation pro- cess and reduce the time to complete the dissolution reaction resulting in the decrease of setting time . 3.2.5. Optimal mixtures in terms of setting time As shown in Fig. 2 , the initial setting time of all mixtures is longer than 60 min, which can fulfil the setting time requirement of BS EN 197-1:2011 . Thus, the chosen parameters in this work (i.e., slag replacement level from 10% to 30%, AL/B ratio of 0.4 and 0.35, 10 M and 12 M of SH, and SS/SH ratio in the range of 1.5–2.5) are suitable for the AAFS in terms of setting time. How- ever, it should be mentioned that the performance criteria would be varied according to different engineering application. 3.3. Compressive strength Compressive strength is one of the most important mechanical properties of concrete. According to ACI 318 M-05 , the 28-d compressive strength of concrete need to achieve at least 28 MPa for the basic engineering application. For the corrosion protection of reinforcement in concrete, the minimum compressive strength of concrete is 35 MPa. In this study, these criteria were used to0100200300400500600 1 2 3 4 5 6 7 8 9 10 11 12 13 A10 A15 A20 A25 A30 B15 B20 B25 C15 C20 C25 D15 D25 Setting time (min) Initial setting time Final setting time 60A B C D Fig. 2. Setting time of alkali-activated fly ash-slag pastes.480 G. Fang et al. / Construction and Building Materials 172 (2018) 476–487 identify the optimal mixtures of AAFS concrete in terms of com- pressive strength. 3.3.1. Effect of fly ash/slag ratio Fig. 3 shows the compressive strength of AAFS specimens with different slag replacement levels. It can be seen that the develop- ment of compressive strength follows the similar trends but differ- ent magnitudes. The compressive strength of AAFS concrete increased dramatically at early 28 d but after that the increasing rate of compressive strength became slower. In addition, the com- pressive strength of AAFS concrete increased with the amount of slag. For instance, when the slag replacement level was increased from 10% to 30%, the 28-d compressive strength was increased from 21.90 to 56.43 MPa. This can be attributed to the formation of C-A-S-H gels, which would reduce the porosity and condense the microstructure of AAFS matrix [24,52–55] . Furthermore, the effect of slag at 20% replacement level on the increase of compres- sive strength appeared to be more pronounced. For example, the 28-d compressive strength of specimens was increased by 34.29%, 90.70%, 113.44% and 157.66% with every 5% increment of slag content from 15% to 30% when compared to the specimen with 10% slag (A10). 3.3.2. Effect of molarity of sodium hydroxide solution Fig. 4 shows the effect of molarity of SH on compressive strength of AAFS concrete. It can be noted that increasing themolarity of SH from 10 M (series A) to 12 M (series B) gradually increased the compressive strength of AAFS concrete, which can be explained by the reaction of the internal Si, Al and Ca compo- nents caused by the increased breakage of the T-O-T bonds (T: Si or Al) in FA and Ca-O and Si-O bonds in GGBS provoked by the high alkalinity resulting from the increasing molarity of SH [56,57] . The 28-d compressive strength of AAFS concrete with 15% slag (B15) was increased by 23.08% as compared to that of A15. The 28-d compressive strength of B25 was increased by 23.19% as compared to that of A25, while the effect was less pronounced for B20, the 28-d compressive strength of which was increased by 10.96% com- pared to that of A20. 3.3.3. Effect of alkaline activator/binder ratio Fig. 5 shows the compressive strength of AAFS concrete with different alkaline activator to binder ratios. It was found thatdecreasing the AL/B ratio from 0.4 (Series A) to 0.35 (Series C) increased the compressive strength of concrete. The difference of compressive strength between specimens with different AL/B ratios became larger with the increase of curing age from 1 to 14 d. However, after that the difference become smaller and the compressive strength at 28 d was almost the same. Thus, it can be said that the amount of AL would strongly affect the early-age (<14 d) compressive strength of AAFS concrete, but no significant effect on the 28-d compressive strength. According to previous research, the alkaline activation process of AAFS would be acceler- ated with the decrease of AL/B ratio due to the decrease of consis- tency of mixtures . In this case, the reaction products such as C-A-S-H gel and N-A-S-H gel can be produced quickly in the mix- tures with low AL/B ratio and contribute to the development ofearly-age compressive strength (from 1 to 14 d in this study) [24,52] . Nevertheless, the reaction rate became slow after 14 d because most of the raw materials have been reacted. 3.3.4. Effect of sodium silicate/sodium hydroxide ratio Fig. 6 shows the compressive strength of AAFS concrete with different SS/SH ratios of 1.5 (D15), 2.0 (A15 and A25) and 2.5 0 1 02 03 04 05 06 001020304050607080 35Compressive strength (MPa) Age (d) A10 A15 A20 A25 A30Fig. 3. Compressive strength of alkali-activated fly ash-slag concrete with different slag content.0 5 10 15 20 25 3010203040506070 35Compressive strength (MPa) Age (d) B15 B20 B25 A15 A20 A25Fig. 4. Compressive strength of alkali-activated fly ash-slag concrete with different molarity of sodium hydroxide solution. 0 5 10 15 20 25 30102030405060 35Compressive strength (MPa) Age (d) C15 C20 C25 A15 A20 A25Fig. 5. Compressive strength of alkali-activated fly ash-slag concrete with different alkaline activator to binder ratios.G. Fang et al. / Construction and Building Materials 172 (2018) 476–487 481 (D25). Comparing the corresponding graphs for same slag replace- ment level in Series A and Series D, it can be seen that the effect of SS/SH ratio on the compressive strength of AAFS concrete was not obvious. The 1-d compressive strength of AAFS concrete was increased slightly with the increase of SS/SH ratio, while the com- pressive strength of AAFS concrete with low SS/SH ratio was slightly higher than that of specimens with high SS/SH ratio at later age. This phenomenon was also found by Nath and Sarker . 3.3.5. Optimal mixtures in terms of compressive strength As shown in Figs. 3–6 , the 28-d compressive strengths of all mixtures except the mixtures with 10% slag are higher than 28 MPa and satisfy the basic requirement of normal concrete. How- ever, for the application of reinforced concrete, only mixtures with relatively high slag content ( =20%) can meet the performance cri- teria because they can achieve a 28-d compressive strength of 35 MPa and more. Therefore, the mixtures with slag replacement level from 20% to 30%, AL/B ratio from 0.35 to 0.4, 10 M and 12 M of SH, and SS/SH ratio in the range of 1.5–2.5 can be suggested as optimal mixtures for the sake of compressive strength. 3.4. Splitting tensile strength Splitting tensile strength of concrete is an important mechani- cal property, which is related to some aspects of concrete struc- tures such as initiation and propagation of cracks, shear and anchorage of reinforcing steel in concrete. Generally, the splitting tensile strength of concrete can be predicted based on its compres- sive strength. ACI 318-05 and Eurocode 2 are commonly used to predict the splitting tensile strength of OPC concrete, according to which the splitting tensile strength of concrete can be calculated from the compressive strength using Eqs. (1) and (2), respectively. fct¼0:56ffiffiffiffi f0 cq ð1Þ fct¼1 3/C18/C19 ðfcÞ2=3for fc<50MPa ð2Þ where fctis the splitting tensile strength (MPa), f0 cis the specified compressive strength (MPa) and fcis the average compressive strength (MPa).For AAFS concrete, it was found that most of the measured split- ting tensile strengths were lower than those predicted by the ACI 318-08 and Eurocode 2 [32,59] . Lee and Lee observed that the splitting tensile strength of AAFS concrete had a linear relation- ship with the square root of the compressive strength, as shown in Eq.(3). Similarly, Sofi et al. found that the linear relationship between splitting tensile strength and square root of compressive with the constant of 0.48 (Eq. (4)), which was slightly higher than the fitting results proposed by Lee and Lee . fct¼0:45ffiffiffiffi f0 cq ð3Þ fct¼0:48ffiffiffiffi f0 cq ð4Þ As shown in Fig. 7 , however, the measured splitting tensile strength in this study was lower than the predictions by the ACI 318-05, Eurocode 2, Lee and Sofi et al. [32,59] . However, it is worth pointing out that the predicted relationship between splitting ten- sile strength and compressive strength of AAFS concrete is strongly affected by many influencing factors, such as chemical and physi- cal properties of raw materials and type of alkaline activators.The splitting tensile strengths of AAFS concrete at 7 and 28 d are plot- ted in Fig. 8 . The splitting tensile strength increased with the increase of curing age for all mixtures. As seen in Series A, the split- ting tensile strength of AAFS concrete increased with the increase of slag content and was also affected by the molarity of SH in mix- tures. When the molarity of SH was increased from 10 M (Series A) to 12 M (Series B), the splitting tensile strength of AAFS concrete was improved. Furthermore, comparing the corresponding graphs for same slag replacement level in Series A and Series C, it can be observed that the 7-d splitting tensile strength increased with the decrease of AL/B ratio from 0.4 to 0.35. However, the 28-d split-ting tensile strengths of mixtures with AL/B ratio of 0.35 (Series C) were smaller than those of mixtures with AL/B ratio of 0.4 (Series A). It is indicated that AL/B ratio would strongly affect the early-age splitting tensile strength development of AAFS concrete, but less significantly at the development of later-age splitting tensile strength. The influences of FA/GGBS ratio, SH molarity and AL/B ratio on the splitting tensile strength are similar to those on com- pressive strength. However, for the effect of SS/SH ratio on splitting tensile strength, the developing trend was different as compared to0 5 10 15 20 25 301020304050Compressive strength (MPa) Age (d) D15 D25 A15 A2528 Fig. 6. Compressive strength of alkali-activated fly ash-slag concrete with different sodium silicate to sodium hydroxide ratios. 01 0 2 0 3 0 4 0 5 0 6 0012345 Lee and Lee ACI 318-05 Eurcode 2 Splitting tensile strength (MPa) Compressive strength (MPa)Experimental results Sofi et al. Fig. 7. Comparison of experimental and predicted splitting tensile strength of alkali-activated fly ash-slag concrete at 28 d.482 G. Fang et al. / Construction and Building Materials 172 (2018) 476–487 that on compressive strength. The splitting tensile strength increased with the increase of SS/SH ratio from 1.5 (D15) to 2.0 (A15), while the splitting tensile strength decreased when the SS/SH ratio increased from 2.0 (A25) to 2.5 (D25). 3.5. Flexural strength Flexural strength of concrete is another key mechanical prop- erty, which represents the ability of a beam or slab to resist failure in bending. Normally, the flexural strength of concrete has a strong relationship with its compressive strength. The flexural strength of OPC concrete, fct:f, is commonly predicted using the ACI 318-05 as fct:f¼0:62ffiffiffiffi f0 cq ð5Þ Similarly, some equations were proposed to predict the flexural strength of AAFS concrete. The relationship between flexural strength ( fct:fÞand compressive strength ( fc) of AAF concrete can be expressed as fct:f¼0:69ffiffiffiffi fcp as suggested by Diaz-Loya et al. . Nath and Sarker also proposed an equation (fct:f¼0:93ffiffiffiffi fcp ) to predict the flexural strength of AAFS concrete. The experimental and predicted value are plotted in Fig. 9 . The flexural strength of AAFS concrete calculated using the equations proposed by Diaz-Loya et al. and Nath and Sarker was higher than measured values in this study, while the flexural strength of concrete predicted by ACI 318-05 was closer to mea- sured values. Nevertheless, it should be highlighted that a general relationship between flexural strength and compressive strength of AAFS concrete is required due to limited available data and vari- ability of mixture composition of AAFS concrete. It can be seen from Fig. 10 that the 28-d flexural strength of AAFS concrete increased with the increase of slag content. Compar- ing the corresponding graphs for same slag content in Series A and Series B, the flexural strength of AAFS concrete increased signifi- cantly when the molarity of SH was increased from 10 M (Series A) to 12 M (Series B). As seen in Series A and Series C, the flexural strength of AAFS concrete increased dramatically with the decrease of AL/B ratio from 0.4 (Series A) to 0.35 (Series C). The effect of SS/SH ratio on flexural strength of AAFS concrete was not significant, as seen in Series A and Series D. These phenomena are similar to the results shown in the development of compressive strength. 3.6. Dynamic elastic modulus Dynamic elastic modulus is a mechanical property of visco- elastic material, which is defined as the ratio of stress to strain when the material is undertaking dynamic loading. According to the CEB-FIP model code and the British testing standard BS8100 Part 2 , the dynamic elastic modulus of OPC concrete can be calculated by Eqs. (6) and (7) . Ec¼22ðfc=10Þ0:3ð6Þ Ed¼Ecþ19 1:25ð7Þ0.00.51.01.52.02.53.03.54.04.5 D 1 2 3 4 5 6 7 8 9 10 11 12 13 A10 A15 A20 A25 A30 B15 B20 B25 C15 C20 C25 D15 D25Splitting tensile strength (MPa) 7 d 28 d A B C Fig. 8. Splitting tensile strength of alkali-activated fly ash-slag concrete at 7 d and 28 d. 0 1 02 03 04 05 06 0012345678 Experimental results ACI 318-05 Diaz-Loya et al. Nath and Sarker Flexural strength (MPa) Compressive strength (MPa) Fig. 9. Comparison of experimental and predicted flexural strength of alkali- activated fly ash-slag concrete at 28 d.G. Fang et al. / Construction and Building Materials 172 (2018) 476–487 483 where Ecis the static elastic modulus (GPa), Eddenotes the dynamic elastic modulus (GPa) and fcis the compressive strength (MPa). An empirical relationship between dynamic elastic modulus and compressive strength of OPC concrete, Ed¼14:72ðfcÞ0:3/C09:56, was also proposed by Zhou et al. .I n addition, Noguchi et al. proposed an equation, Ed¼12:66ðfcÞ0:27, to predict the dynamic elastic modulus of OPC concrete. The experimental and measured values are plotted in Fig. 11 , which shows that the predicted dynamic elastic modulus using equations proposed by Zhou et al. and Noguchi et al. was obviously higher than the measured values. The predicted data using Eqs. (6) and (7) were also higher than the measured data forAAFS concrete. Thus, a more suitable equation, Ed¼7:64ðfcÞ0:35/C03:75, was proposed here to predict the dynamic elastic modulus of AAFS concrete. Nevertheless, more experimen- tal data are required to accurately predict the relationship between dynamic elastic modulus and compressive strength of AAFS concrete. The dynamic elastic modulus of AAFS concrete with different mix proportions is plotted in Fig. 12 . As seen in Fig. 12 a, the dynamic elastic modulus of AAFS concrete steadily increased with the increase of slag replacement level. As the SH molarity increased from 10 M (Series A) to 12 M (Series B), the dynamic elastic mod- ulus increased dramatically (see Fig. 12 b). Furthermore, the dynamic elastic modulus of AAFS concrete would also be affected by AL/B ratio (see Fig. 12 c). With the decrease of AL/B ratio from 0.4 (Series A) to 0.35 (Series C), the dynamic elastic modulus increased significantly. These phenomena can also be found in the developing trend of compressive strength. However, the effect of SS/SH ratio on dynamic elastic modulus is different from that on compressive strength. As shown in Fig. 12 d, the dynamic elastic modulus of AAFS concrete decreased with the increase of SS/SH ratio from 1.5 (A15) to 2.0 (D15, and from 2.0 (A25) 2.5 (D25), respectively. 3.7. Optimal mixtures Fig. 13 shows the obtained optimal mixtures of AAFS concrete according to the performance criteria of workability, setting time and compressive strength. Based on the discussion mentioned above, the optimal AAFS mixtures should have high workability (i.e., achieving a slump value of 90 mm or over) , suitable set- ting time (i.e., minimum initial setting time of 60 min) and high compressive strength (i.e., minimum 28-d compressive strength of 35 MPa) . Therefore, the mixtures with slag replacement level from 20% to 30%, AL/B ratio of 0.4, 10 M of SH, and SS/SH ratio in the range of 1.5 to 2.5 were suggested as optimal mixtures.01234 1 2 3 4 5 6 7 8 9 10 11 12 13 A10 A15 A20 A25 A30 B15 B20 B25 C15 C20 C25 D15 D25 Flexural strength (MPa)A B C D Fig. 10. Flexural strength of alkali-activated fly ash-slag concrete at 28 d. 10 20 30 40 50 6015202530354045 Experimental results Fitting equation CEB-FIP model code & BS8100 [61, 62] Zhou et al. Noguchi et al. Dynamic elastic modulus (Gpa) Compressive strength (MPa) Fig. 11. Comparison of experimental and predicted dynamic elastic modulus of alkali-activated fly ash-slag concrete.484 G. Fang et al. / Construction and Building Materials 172 (2018) 476–487 4. Conclusions In this study, the workability, setting time and mechanical properties of AAFS with different slag content, AL/B ratio, molarity of SH and SS/SH ratio were investigated. Afterwards, the optimal mixtures of AAFS were proposed. Based on the experimental results, the main conclusions can be drawn as follows: /C15Workability of AAFS decreased with the increase of slag content and molarity of SH, as well as the decrease of AL/B ratio. The influence of slag content seemed more significant at higher molarity of SH and higher AL/B ratio. The specimens with higher molarity of SH at higher slag replacement level exhibited a more significant loss of workability than specimens with lower slag content. /C15Setting time of AAFS pastes decreased with increasing slag con- tent and decreasing AL/B ratio. The effect of slag at 20% replace- ment level on setting time appeared to be more pronounced. /C15Compressive strength of AAFS increased significantly with the increase of slag content and molarity of SH as well as the decrease of AL/B ratio. The effect of slag content at 20% was5 1 01 52 02 53 01416182022242628303234Dynamic elastic modulus (GPa) Age (d) A10 A15 A20 A25 A30(a) 51 0 1 5 2 0 2 5 3 01618202224262830Dynamic elastic modulus (GPa) Age (d) B15 B20 B25 A15 A20 A25(b) 5 1 01 52 02 53 016182022242628Dynamic elastic modulus (GPa) Age (d) C15 C20 C25 A15 A20 A25(c) 5 1 01 52 02 53 016171819202122232425Dynamic elastic modulus (GPa) Age (d) D15 D25 A15 A25(d) Fig. 12. Dynamic elastic modulus of alkali-activated fly ash-slag concrete with (a) different slag content, (b) different molarity of sodium hydroxide solut ion, (c) different alkaline activator to binder ratios, and (d) different sodium silicate to sodium hydroxide ratios. Fig. 13. Schematic illustration of optimal mixtures of alkali-activated fly ash-slag concrete in terms of workability, setting time and compressive strength.G. Fang et al. / Construction and Building Materials 172 (2018) 476–487 485 more significant, while the effect of SH molarity was less pro- nounced at slag replacement level of 20%. In addition, the amount of AL would significantly affect the compressive strength development at early age (<14 d), but the effect became less significant for the specimens at older age (28 d). /C15Existing equations provided by ACI code, Eurocode and other researchers for OPC concrete overestimated the values of split- ting tensile strength, flexural strength and dynamic elastic modulus of AAFS concrete. These mechanical properties of AAFS concrete cured at ambient temperature mostly followed a sim- ilar development trend as compressive strength. /C15The mixtures of AAFS with slag replacement level from 20% to 30%, AL/B ratio of 0.4, 10 M of SH, and SS/SH ratio in the range of 1.5 to 2.5 can be suggested as optimal mixtures regarding the performance criteria of workability, setting time and compres-sive strength. Conflict of interest There is no conflict of interest. Acknowledgements The authors gratefully acknowledge the financial support of the Royal Society (IE150587) and EPSRC (EP/R041504/1). The financial support provided by University College London (UCL) and China Scholarship Council (CSC) to the first author is gratefully acknowl- edged. The authors would like to thank Mr Chun Wai Goh, Mr Hui Zhong, and Miss Yi Wang for their support throughout this research. The authors would also like to thank Mr Warren Gaynor and Dr Shi Shi from UCL Laboratory of Advanced Materials and Dr Judith Zhou from UCL Environmental Engineering Laboratory for their help with experiments.