Hydration of C4AF in the presence of other phases: A synchrotron X-ray powder diffraction study

Manuscript submitted to Construction and Building Materials 1 2 Hydration of C4AF in the presence of other phases: a synchrotron X-ray powder 3 diffraction study 4 A. Cuestaa, I. Santacruza, S. G. Sanfélixb, F. Fauthc, M. A. G. Arandaa,c, A. G. De la Torrea,*. 5 6 a Departamento de Química Inorgánica, Universidad de Málaga, Campus Teatinos S/N. 29071-Málaga, 7 Spain. 8 b Unidad Técnica de Investigación de Materiales, AIDICO, Avda. Benjamín Franklin, 17 Paterna, Valencia, 9 Spain. 10 c ALBA-CELLS synchrotron, Carretera BP 1413, Km. 3.3, E-08290 Cerdanyola, Barcelona, Spain. 11 12 * Corresponding author. Tel.: +34952131877; fax: +34952132000. 13 E-mail address: [email protected] (A.G. De la Torre) 14 15 1 Abstract 1 Hydration behaviour of C4AF in selected experimental conditions has been determined. C4AF has 2 been hydrated in the absence and presence of gypsum, two polymorphs of ye’elimite and different 3 water/solid ratios. C4AF in the presence of water hydrates to form mainly a hydrogarnet-type phase. 4 The crystal structure of C3A0.845F0.155H6 is reported from the Rietveld analysis of its synchrotron 5 X-ray powder diffraction pattern. The hydration of C4AF in the presence of gypsum gives AFt. 6 However, the mixture tetracalcium aluminoferrite/gypsum/ye’elimite gives both AFt and AFm 7 phases. C4AF hydrated with ye’elimite in the absence of gypsum gives only AFm. Ye’elimite has 8 inhibited tetracalcium aluminoferrite hydration. 9 Keywords: Hydration mechanism, Rietveld method, dissolution and crystallization kinetics, 10 tetracalcium aluminoferrite. 11 12 13 2 1. Introduction 1 X-ray powder diffraction (XRPD) is very well suited for in-situ studies of chemical processes 2 involving crystalline materials [1,2]. During the last years, it has been reported quantitative phase 3 analysis of cements, clinkers and supplementary cementitious materials by combining XRPD and 4 Rietveld methodology [3-6]. More recently, this procedure has been expanded to hydrated 5 cementitious systems [1,7] and in some of these studies, the non diffracting fraction was also 6 determined [8-10], although a more precise term was coined: Amorphous and Crystalline not-7 quantified content (ACn) [11]. Furthermore, the use of intense monochromatic X-rays, such as 8 synchrotron X-rays, coupled with a fast X-ray detection system permits high-resolution time-9 resolved diffraction experiments allowing in-situ measurements during the hydration process of 10 cements [1,12,13]. 11 The tetracalcium aluminoferrite phase, C4AF in cement nomenclature, also known as 12 brownmillerite, is the major iron-containing phase in Ordinary Portland Cement (OPC) and is also 13 present in iron rich belite calcium sulfoaluminate cements [14-15]. This phase has been deeply 14 studied [16–19] and for this reason, their crystal structures and chemical compositions are currently 15 known, including the structural variations which have placed in the Ca2(Fe2-yAly)O5 series [20], 16 where y can oscillate from 0 to about 1.33 [19] in high iron content cements. In Portland cements 17 this phase is supposed to have an ideal composition with y=1. However, it has been reported that 18 other elements could be also present in the composition of the C4AF phase and for this reason the 19 A/F ratio is not exactly unity [21]. 20 In the absence of any other phases, the hydration of C4AF is similar to the hydration of C3A, in 21 which a C-A-H gel first coats the C3A grains. This gel presents metastable hexagonal C-A-H plates 22 that finally convert to the stable cubic hydrate C3AH6 [22-24]. However, the hydration products are 23 (quite often) assumed to incorporate some iron in the case of the tetracalcium aluminoferrite phase 24 [25-29]. Firstly, in the hydration of tetracalcium aluminoferrite with water a metastable type-gel C-25 (A,F)-H [28] is formed and with the time this gel converts to a hydrogarnet phase, also known as 26 3 katoite, C3(A,F)H6 [28], with an Al/(Al + Fe) ratio of about 0.4. However, other authors [25,30,31] 1 stated that solid solution between C3AH6 and C3FH6 is not formed. 2 The exact Al/Fe ratios of the hydrogarnets are as yet under debate but it is generally accepted that 3 the Al/Fe ratio of the crystalline products is greater than in C4AF itself [32]. 4 The hydration of C4AF [28] could be written as: 5 C4AF + 10H → C3AH6 + FH3 + CH (1) 6 where FH3 indicates a hydrated amorphous Fe-containing gel. If the case that Fe is incorporated 7 into the hydrogarnet product, then the reaction could be expressed as: 8 C4AF + 10H → 4 3C3(A0.75F0.25)H6 + 2 3FH3 (2) 9 When calcium sulfates are added, the direct hydration of C4AF to C3(A,F)H6 is inhibited. 10 Consequently, ettringite is the common hydration product observed in this case. There are some 11 theories about the mechanism governing the retardation process. However, the formation of 12 hydroxy-AFm gel surrounding C4AF particles which yields to ettringite crystallization centers is the 13 most likely mechanism [29]. 14 In the presence of a source of sulfate the reaction could be formulated as [28]: 15 C4AF + 3CS �H2+ 30H → C6AS � R3H32 + CH + FH3 (3) 16 where C6AS � R3H32 denotes ettringite (also known as AFt). Again, iron may be incorporated in the 17 ettringite structure as detailed by [26] according to reaction (3.1): 18 C4AF + 4CS �H2+ 106 3H → 4 3C6(A0.75F0.25)S � R3H31 + 2 3FH3 (3.1) 19 where C6(A0.75F0.25)S � R3H32 stands for iron-bearing AFt. 20 Successively, ettringite can decompose to form an AFm monosulfoaluminate hydrate in the 21 presence of C4AF as stated next: 22 C6AS � R3H32 + C4AF → 2C4AS �Hn + CS �H2 + (26 - 2n)H + FH3 + CH (4) 23 4 If there is any amount of C4AF, this could react with gypsum to form AFm (C4AS �Hn), then the 1 reaction could be expressed as: 2 C4AF + CS �H2 + (n + 2)H → C4AS �Hn + FH3 + CH (5) 3 Moreover, it has been reported [28] that a partial series of solid solutions can be formed between 4 C3AH6 and C3FH6; accordingly the previous equations could be also expressed in terms of 5 Cx(A,F)yHz or Cx(A,F)yS �wHz. 6 The main purpose of the present work has been to study the hydration of C4AF phase in different 7 environments: in the presence and absence of gypsum and in the presence of both polymorphs of 8 ye’elimite, stoichiometric (orthorhombic) and solid solution (pseudo-cubic). The interest of 9 studying these systems as simple mixtures is to better understand the hydration mechanisms of new 10 eco-cements based on calcium sulfoaluminate, which contains different polymorphs of ye’elimite 11 with C4AF [12, 14]. In order to do so, synchrotron XRPD (SXRPD) and Rietveld methodology are 12 employed. The advantage of using high energy radiation is mainly the minimization of 13 microabsorption effects. Kinetics of hydration have been established and correlated to calorimetric 14 data. Scanning electron microscopy (SEM) has been also done to corroborate the results. Moreover, 15 high-resolution SXRPD and transmission electron microscopy (TEM) were used to refine the 16 crystal structure of one of the hydrated crystalline phases, C3(A,F)H6. 17 2. Experimental Section 18 2.1. Sample preparation. 19 Tetracalcium aluminoferrite (C4AF) was prepared by mixing suitable amounts of CaCO3 (99.95%, 20 Alfa Aesar), Al2O3 (99.997%, Alfa Aesar) and Fe2O3 (99.945%, Alfa Aesar), to obtain 21 approximately 50 g of sample with targeted chemical formula of Ca2AlFeO5, i.e. x=1. The raw 22 mixture was ground for 5 minutes and heated at 1000ºC for 4 hours (heating rate of 10 ºC/min). 23 After that, the powder was ground for 45 min in a Micro-deval machine with a cylinder container 24 and steel balls and was pelletized (600 mm diameter and 1000 MPa). Finally, the pellets were 25 5 heated at 1350ºC for 4 hours (heating rate of 10 ºC/min) followed by quenching from high 1 temperature with an air flow. 2 Moreover, stoichiometric (st-) Ca4[Al6O12]SO4 and solid solution (ss-) 3 Ca3.8Na0.2Al5.6Fe0.2Si0.2O12SO4 ye’elimites has been used for this study. Ye’elimite polymorphs 4 were prepared as previously reported [33, 34]. 5 C4AF was mixed, in some cases, with gypsum (g), stoichiometric (st-) or solid solution (ss-) 6 ye’elimite. Table 1 reports paste mix proportions, including water/solid (w/s) ratios. The gypsum 7 used for the hydration studies was that marketed by BELITH S.P.R.L. (Belgium). 8 For this study, in-situ SXRPD experimental set up was employed. All the anhydrous mixtures were 9 mixed with 15 wt% of SiO2 (99.56%, ABCR) as an internal standard [13, 35] and SXRPD data 10 were collected to obtain the initial phase assemblage (t0). This standard presents an amorphous 11 content of 12.7(1) wt% which was determined by the external standard method [36]. Moreover, it is 12 important to bear in mind that in the reported water/solid ratio, the amount of internal standard is 13 not taken into account. Pastes were ex-situ prepared and immediately loaded into glass capillaries of 14 0.5 mm of diameter with a syringe. The capillaries were sealed with grease to avoid any water loss. 15 2.2. Synchrotron X-ray powder diffraction (SXRPD). 16 SXRPD patterns were collected at room temperature in Debye-Scherrer (transmission) mode using 17 the high-resolution X-ray powder diffraction beamline of ALBA synchrotron (Barcelona, Spain) 18 [37]. The wavelength, 0.61975(1) Å, was selected with a double-crystal Si (111) monochromator 19 and determined from Si640d NIST standard (a=5.43123 Å). The diffractometer is equipped with a 20 MYTHEN detector especially suited for time-resolved experiments. The capillaries were rotated 21 during data collection to improve diffracting particle statistics and the synchrotron beam was 22 focused in the detector to improve the powder diffraction peak shapes. The data acquisition time 23 was ∼15 min per pattern to attain very good signal-to-noise ratio over the angular range 1-35º (2θ). 24 6 The diffractometer is also equipped with a detector system based on crystal analyzers in the 1 diffracted beam especially suited for extremely high-resolution experiments giving also a very flat 2 background. One of the samples was measured with this detector in a capillary of 0.5 mm of 3 diameter that was rotated at 400 rpm during data collection to improve diffracting particle statistics. 4 The data acquisition time was very long, ∼4 hours, to attain a very good signal-to-noise ratio over 5 the recorded wide angular range 1-45º (2θ), to perform the structural study of a crystalline hydrate 6 sample, see below. 7 2.3. SXRPD data analysis. 8 Raw SXRPD patterns were normalized taking into account the decay of X-ray beam flux with time. 9 Patterns were analysed by using the Rietveld methodology as implemented in the GSAS software 10 package [38], in order to obtain Rietveld quantitative phase analysis (RQPA). The refined overall 11 parameters were background coefficients, cell parameters, zero-shift error, peak shape parameters, 12 and phase scale factors. Powder diffraction peak shapes were fitted by using the pseudo-Voigt 13 function [39]. The ACn contents were determined by internal standard methodology [35] from 14 SXRPD data as detailed previously [13]. For the structural analysis, atomic parameters, in addition 15 to the overall parameters, were also optimised. 16 2.4. Isothermal calorimetry. 17 The isothermal calorimetric study was performed in an eight channel Thermal Activity Monitor 18 (TAM) instrument using glass ampoules. Some pastes were selected and were prepared ex-situ by 19 mixing ~6 g of each sample with the appropriated water amount. Then, the pastes were immediately 20 introduced in the calorimeter. A stabilization period of 45 minutes was needed to start the 21 measurements. The heat flow was collected up to 2 days at 20ºC. 22 2.5. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) 23 studies. 24 7 Prior to SEM observation, the hydration of selected samples was stopped by immersing them in 1 isopropanol for 3 days and then heated at 40 ºC for 24 h. Samples suitable for the SEM 2 characterization were prepared in cylinders following the methodology previously reported [13]. 3 Microscopic characterization of samples was performed in a JEOL JSM-6490LV. Samples were 4 impregnated with low viscosity resin and polished down to 1 µm using diamond spray and further 5 sputtered with graphite. Energy dispersive spectroscopy (EDS) measurements were carried out with 6 the OXFORD INCA Energy 350 attachment. The voltage used was 20 kV and the working distance 7 was around 10 mm. 8 Moreover, the fracture cross-sections of some pastes after 48 hours of hydration were also observed 9 by SEM in a JEOL-JSM-840. These samples were gold sputtered. 10 Finally, TEM studies were done in a Philips CM-200 for some particles of a hydrated sample 11 (without any thermal/chemical treatment to stop hydration), with microanalysis performed with the 12 EDAX Génesis-4000 attachment. 13 2.6. Thermal analysis. 14 Differential thermal analysis (DTA) and thermogravimetric (TGA) measurements were performed 15 in a SDT-Q600 analyzer from TA instruments (New Castle, DE) in non-stopped hydration samples. 16 The temperature was varied from RT to 1000ºC at a heating rate of 10 ºC/min. Measurements were 17 carried out in open platinum crucibles under air flow. 18 3. Results and discussions. 19 Initial C4AF characterization. The C4AF sample was studied by SXRPD and the Rietveld method. 20 For this synthesis, the only additional phase was C3A, 1.94(6) wt%. The derived unit cell 21 parameters for C4AF, from SXRPD, were: a=5.56638(4)Å, b=14.5227(1)Å, c=5.34835(4)Å and 22 V=432.354(8) Å3. As the quality of the SXRPD was very high, we could optimize the Fe/Al ratio in 23 the octahedral and tetrahedral sites (constrained to an overall Fe/Al ratio of 1.00). The resulting 24 value was 0.746(2) of Fe occupancy at the octahedral site and therefore 0.254(2) of Fe occupancy at 25 8 the tetrahedral site. We are aware that different cooling rates could change Fe/Al distributions at 1 these sites but to study the possible consequences of this variation on the hydration properties of 2 C4AF is out of the scope of the present work. 3 3.1. Hydration of C4AF. 4 Initially, the hydration behaviour of just C4AF was studied. The w/s ratio used was 0.8, see Table 1. 5 C4AF_0.8 sample has been characterized up to 5.5 hours by in-situ SXRPD with internal standard 6 methodology for determining the ACn content. Time-resolved SXRPD was employed to track the 7 dissolution of the anhydrous phases followed by the crystallization of the different hydrated phases. 8 Table 2 shows the phase assemblages at different ages. In agreement with reaction (2), C3(A,F)H6 9 crystallized, and during the first hours some C2AH8 was also formed. It has been reported that 10 C2AH8 phase is formed after 5 minutes from hydration and then it abruptly converts into 11 C3(A,F)H6 [28]. In this study, the first pattern was taken at 1 hour, consequently big amounts of 12 C2AH8 phase were not expected and C3(A,F)H6 is the main phase after 5.5 hours of hydration. 13 Moreover, the small amount of C3A in this sample does not affect the hydration products as only 14 hydrogarnet with iron was observed at all the hydration times. The hydrogarnet type phase present 15 in all the patterns presented systematic peak shifts, indicating the formation of the solid solution 16 C3(A,F)H6 in agreement with [28] and in contrary with results presented in [25,30,31]. 17 The first column in Table 2 gives t0 values obtained from the SXRPD pattern of the anhydrous 18 sample renormalized with theoretical free water, FW. Remaining values obtained from internal 19 standard method encompass not only ACn but also FW (not chemically bound water) and are 20 expressed as a single value in Table 2. 21 ACn and FW are expressed as a single value due to the inability of the internal standard 22 methodology to distinguish between different not-diffracting phases. It can be observed that these 23 values slightly diminished with time. This is mainly due to the precipitation of a crystalline phase, 24 C3(A,F)H6, which is consuming free water. 25 9 observed. The lattice parameters of AFm and AFt were also checked. The average lattice 1 parameters obtained for the AFm formed in these samples are a=5.758(4) Å and c=28.66(3) Å and 2 the volume of AFt was 2348.2(4) Å3. These values are quite similar to that obtained in the hydration 3 of pure ye'elimite without C4AF [13] given above. This may be due to the fact that the incorporation 4 of iron is not taking place in these samples. 5 3.3.2. Role of w/s ratio. The sample which contains st-ye’elimite, tetracalcium aluminoferrite and 6 gypsum was hydrated with 0.7 and 1.3 w/s ratios up to 12 hours. Tables 5 and 6 show the phase 7 assemblages with time for both mixtures. It is well known that higher amounts of water enhance 8 ye’elimite reactivity [13, 41] as well as the other cementitious systems (OPC, Calcium Aluminate 9 Cements, etc). Figure S5, provided as supplementary material, confirms this behaviour showing the 10 degree of reaction of ye’elimite for these samples. Moreover, in the case of C4AF_st-C4A3𝑆_g_0.7, 11 the main hydration product was AFt, within 12 hours in contrast with C4AF_st-C4A3𝑆_g_1.0 and 12 C4AF_st-C4A3𝑆_g_1.3 where there was a competition between AFm and AFt. That indicates that 13 higher amounts of water favour the formation of AFm, in agreement with a previous report [13]. 14 Figure 6a and 6b shows Rietveld plots for C4AF_st-C4A3𝑆_g_0.7 and C4AF_st-C4A3𝑆_g_1.3, 15 respectively, at 12 hours of hydration. This Figure illustrates the small amount of AFm with respect 16 to AFt for C4AF_st-C4A3𝑆_g_0.7, see Figure 6a. However, for C4AF_st-C4A3𝑆_g_1.3, the presence 17 of both hydrates phases, AFt and AFm, is clearly seen (Figure 6b). 18 3.3.3. Hydration of C4AF and st-ye’elimite in the absence of gypsum. Finally, C4AF_st-19 C4A3𝑆_1.0 sample, see Table 1 was studied. The w/s ratio was 1.0. Table 7 shows the phase 20 assemblage as a function of time up to 20 h. The main difference when compared to the sample 21 with gypsum, C4AF_st-C4A3𝑆_g_1.0, is that AFm was the only hydration product and AFt was not 22 observed at any time. Moreover, ACn and FW contents were more or less constant with time due to 23 that the consumption of water was being compensated by the formation of AFm, which also 24 develops a high amorphous content. 25 16 Finally, we would like to highlight that we have shown in a previous work that st-ye’elimite 1 hydration (without gypsum) gives AFt (and also some amounts of AFm which depends upon the 2 water/ye’elimite ratio) [13]. AFt is formed because ye’elimite dissolution is also releasing sulfate 3 anions to the water in the porous microstructure [13]. Now, when we compared these results with 4 those shown in Table 7, it is clear that the presence of C4AF has inhibited the formation of AFt 5 from ye’elimite. Moreover, the hydration of tetracalcium aluminoferrite has also been inhibited by 6 the presence of ye’elimite, see Table 7, as it has not significantly reacted after 20 hours. 7 4. Conclusions 8 This work reports a comprehensive hydration study of C4AF. Several parameters have been tested 9 to unravel their effects on tetracalcium aluminoferrite hydration behaviour, such as the presence of 10 different sulfate sources (gypsum or calcium sulfoaluminate phase) and the water/solid ratio. 11 C4AF in the presence of water hydrates to form mainly a hydrogarnet-type phase C3(A,F)H6. The 12 hydration of C4AF with w/s ratio of 1.0 yielded to C3A0.84F0.16H6 as single crystalline phase. Its 13 crystal structure has been analyzed by the Rietveld method and reported here. 14 The presence of sulfates strongly modifies tetracalcium aluminoferrite hydration behavior. The 15 hydration of C4AF in the presence of gypsum gives AFt and amorphous aluminum hydroxide and 16 once gypsum is completely dissolved crystalline AFm starts to precipitate jointly with more 17 amorphous phase(s). SEM-EDS studies showed that both AFt and AFm phases contain iron. 18 The hydration of two different polymorphs of ye’elimite with tetracalcium aluminoferrite in the 19 presence of gypsum has also been studied. The mixture of tetracalcium aluminoferrite, gypsum and 20 stoichiometric ye’elimite (which is orthorhombic) gives a mixture of AFt and AFm phases. 21 However, C4AF hydration is slightly slowed down and moreover, these hydration products did not 22 contain iron in their structures as determined by SEM-EDS. The retarder effect of ye’elimite over 23 C4AF hydration was much stronger with solid solution ye’elimite (which is pseudocubic), and 24 consequently AFt was the main hydration phase. 25 17 The effect of w/s ratio has also been studied in the sample containing tetracalcium aluminoferrite, 1 gypsum and stoichiometric ye’elimite. Results indicate that higher amounts of water favour the 2 formation of AFm. 3 Finally, C4AF hydrated with ye’elimite in the absence of gypsum gives AFm as the unique 4 hydration product. Consequently, it has been demonstrated that gypsum (which is a high soluble 5 sulfate) is the main responsible for AFt formation in these conditions. Moreover, tetracalcium 6 aluminoferrite has modified ye’elimite hydration as AFt was not precipitated, and ye’elimite has 7 inhibited tetracalcium aluminoferrite hydration as it did not react after 20 hours of hydration. 8 Acknowledgements 9 This work has been supported by Junta de Andalucía through P11-FQM-7517 research grant, by 10 Spanish MINECO through MAT2010-16213 research grant, which is co-funded by FEDER and 11 FEDER/University of Málaga (FC14-MAT-23) grant. We also thank CELLS-ALBA (Barcelona, 12 Spain) for providing synchrotron beam time at BL04-MSPD beamline. 13 References 14 [1] Merlini M, Artioli G, Meneghini C, Cerulli T, Bravo A, Cella F. The early hydration and the set 15 of Portland cements: In situ X-ray powder diffraction studies. Powder Diffr 2007; 22:201-208. 16 [2] Snellings R, Mertens G, Adriaens R, Elsen J. In situ synchrotron X-ray powder diffraction study 17 of the early age hydration of cements blended with zeolite and quartzite fines and water-reducing 18 agent. J Appl Clay Sci 2013; 72:124-131. 19 [3] Álvarez-Pinazo G, Cuesta A, García-Maté M, Santacruz I, Losilla ER, De la Torre AG, León-20 Reina L, Aranda MAG. Rietveld quantitative phase analysis of Yeelimite-containing cements. Cem 21 Concr Res 2012; 42:960-971. 22 [4] De la Torre AG, Cabeza A, Calvente A, Bruque S, Aranda MAG. Full phase analysis of 23 Portland clinker by penetrating synchrotron powder diffraction. Anal Chem 2001; 73:151-156. 24 [5] De la Torre AG, Aranda MAG. Accuracy in Rietveld quantitative phase analysis of Portland 25 cements. J Appl Cryst 2013; 36:1169-1176. 26 [6] León-Reina L, De la Torre AG, Porras-Vázquez JM, Cruz M, Ordonez LM, Alcobe X, Gispert-27 Guirado F, Larrañaga-Varga A, Paul M, Fuellmann T, Schmidt R, Aranda MAG. Round robin on 28 Rietveld quantitative phase analysis of Portland cements. J Appl Cryst 2009; 42:906-916. 29 [7] Scrivener KL, Füllmann T, Gallucci E, Walenta G, Bermejo E. Quantitative study of Portland 30 cement hydration by X-ray diffraction/Rietveld analysis and independent methods. Cem Concr Res 31 2004; 34:1541-1547. 32 [8] Klaus SR, Neubauer J, Goetz-Neunhoeffer F. Hydration kinetics of CA2 and CA - 33 Investigations performed on a synthetic calcium aluminate cement. Cem Concr Res 2013; 43:62-69. 34 18 [9] Jansen D, Goetz-Neunhoeffer F, Stabler C, Neubauer J. A remastered external standard method 1 applied to the quantification of early OPC hydration. Cem Concr Res 2011; 41:602-608. 2 [10] Clark SM, Barnes P. A comparison of laboratory, synchrotron and neutron diffraction for the 3 real time study of cement hydration. Cem Concr Res 1995; 25:639-646. 4 [11] Aranda MAG, De la Torre AG, León-Reina L. Rietveld Quantitative Phase Analysis of OPC 5 Clinkers, Cements and Hydration Products. Rev Mineral Geochem 2012; 74:169-209. 6 [12] Álvarez-Pinazo G, Cuesta A, García-Maté M, Santacruz I, Losilla ER, Ordóñez LM, Sanfélix S 7 G, Fauth F, Aranda MAG, De la Torre AG. In-situ early-age hydration study of sulfobelite cements 8 by synchrotron powder diffraction. Cem Concr Res 2014; 56:12-19. 9 [13] Cuesta A, Álvarez-Pinazo G, Sanfélix SG, Peral I, Aranda MAG, De la Torre AG. Hydration 10 mechanisms of two polymorphs of synthetic ye'elimite. Cem Concr Res 2014; 63:127-136. 11 [14] Álvarez-Pinazo G, Santacruz I, León-Reina L, Aranda MAG, De la Torre AG. Hydration 12 Reactions and Mechanical Strenght Developmenrts of Iron Rich Sulfobelite Eco-cements. Ind Eng 13 Chem Res 2013; 52:16606-16614. 14 [15] Senff L, Castela A, Hajjaji W, Hotza D, Labrincha JA. Formulations of sulfobelite cement 15 through design of experiments. Constr Build Mater 2011; 25:3410-3416. 16 [16] Bertaut EF, Blum P, Sagniers A. Structure du ferrite bicalcique et de la brownmillerite, Acta 17 Crystallogr 1959; B12:149-159. 18 [17] Colville AA, Geller S. The crystal structure of brownmillerite, CA2FeAlO5. Acta Crystallogr 19 1971; B27:2311-2315. 20 [18] Colville AA, Geller S. The structures of Ca2Fe1.43Al0.57O5 and Ca2Fe1.28Al0.72O5. Acta 21 Crystallogr 1972; B28:3196-3200. 22 [19] Touzo B, Scrivener KL, Glasser FP. Phase compositions and equilibria in the CaO-Al2O3-23 Fe2O3-SO3 system, for assemblages containing ye’elimite and ferrite Ca2(Al,Fe)O5. Cem Concr 24 Res 2013; 54:77-86. 25 [20] Redhammer GJ, Tippelt G, Roth G, Amthauer G. Structural variations in the brownmillerite 26 series Ca2(Fe2-xAlx)O5: Single-crystal X-ray diffraction at 25ºC and high-temperature X-ray 27 powder diffraction (25ºC ≤ T ≤ 1000ºC). Am Mineral 2004; 89:405-420. 28 [21] Taylor HFW. Cement Chemistry. London: Telford; 1997. 29 [22] Jupe AC, Turrillas X, Barnes P, Colston SL, Hall C, Hausermann D, Handfland M. Fast in situ 30 x-ray-diffraction studies of chemical reactions: A synchrotron view of the hydration of tricalcium 31 aluminate. Phys Rev 1996; B 53:14697-14700. 32 [23] Meredith P, Donald AM, Meller N, Hall C. Tricalcium aluminate hydration: Microstructural 33 observations by in-situ electron microscopy. J Mater Sci 2004; 39:997-1005. 34 [24] Zivica V, Palou M, Bagel L, Krizma M. Low-Porosity tricalcium aluminate hardened paste. 35 Constr Build Mater 2013; 38:1191-1198. 36 [25] Rogers DE, Aldridge LP. Hydrates of calcium ferrites and calcium aluminoferrites. Cem Concr 37 Res 1977;7:399-410. 38 [26] Fukuhara M, Goto S, Asage K, Daimon M, Kondo R. Mechanisms and kinetics of C4AF 39 hydration with gypsum. Cem Concr Res 1981; 11:407–414. 40 [27] Odler I. In: Lea’s chemistry of cement & concrete, ed. P. C. Hewlett, Butterworth-Heinemann, 41 Oxford; 1998, p. 195–240. 42 19 [28] Meller N, Hall C, Jupe AC, Colston SL, Jacques SDM, Barnes P, Phipps J. The paste hydration 1 of brownmillertire with and without gypsum: a time resolved synchrotron diffraction study at 30, 2 70, 100 and 150 ºC. J Mat Chem 2004; 14:428-435. 3 [29] Ectors D, Neubauer J, Goetz-Neunhoeffer F. The hydration of synthetic brownmillerite in 4 presence of low Ca-sulfate content and calcite monitored by quantitative in-situ-XRD and heat flow 5 calorimetry. Cem Concr Res 2013; 54:61-68. 6 [30] Dilnesa BZ, Lothenbach B, Renaudin G, Wichser A, Kulik D. Synthesis and characterization 7 of hydrogarnet Ca3(AlxFe1−x)2(SiO4)y(OH)4(3−y). Cem Concr Res 2014; 59:96-111. 8 [31] Flinp EP, McMurdie HF, Wells LS. Hydrothermal and X-ray studies of the garnet-hydrogarnet 9 series and relationship of the series to hydration products of Portland cement. J Res Nat Bur Stand 10 1941; 26:13-33. 11 [32] Meller N, Hall C, Crawshaw J. ESEM evidence for through-solution transport during 12 brownmillerite hydration. J Mat Sci 2004; 39:6611-6614. 13 [33] Cuesta A, De la Torre AG, Losilla ER, Peterson VK, Rejmak P, Ayuela A, Frontera C, Aranda 14 MAG. Structure, atomistic simulations, and phase transition of stoichiometric ye`elimite. Chem 15 Mater 2013; 25:1680-1687. 16 [34] Cuesta A, De la Torre AG, Losilla ER, Santacruz I, Aranda MAG. Pseudocubic Crystal 17 Structure and Phase Transition in Doped Ye’elimite. Cryst Growth Des 2014; 14:5158-5163. 18 [35] De la Torre AG, Bruque S, Aranda MAG. Rietveld quantitative amorphous content analysis. J 19 Appl Crystallogr 2001; 34:196-202. 20 [36] Jansen D, Stabler CH, Goetz-Neunhoeffer F, Dittrich S, Neubauer J. Does ordinary Portland 21 cement contain amorphous phase? A quantitative study using an external standard method. Powder 22 Diffrn 2011; 26:31-38. 23 [37] Fauth F, Peral I, Popescu C, Knapp M. The new Material Science Powder Diffraction beamline 24 at ALBA Synchrotron. Powder Diffr 2013; 28:S360-S370. 25 [38] Larson AC, Von Dreele RB. General Structure Analysis System (GSAS). Los Alamos National 26 Laboratory Report LAUR 2000; pp 86-748. 27 [39] Thompson P, Cox DE, Hasting JB. Rietveld refinement of Debye-Scherrer synchrotron X-ray 28 data from Al2O3. J Appl Cryst 1987; 20:79-83. 29 [40] Larger GA, Armbruster TH, Faber J. Neutron and X-ray diffraction study of hydrogarnet 30 Ca3Al2(O4H4)3. A Miner 1987; 72:756-765. 31 [41] Winnefeld F, Barlag S. Calorimetric and thermogravimetric study on the influence of calcium 32 sulfate on the hydration of ye’elimite. J Therm Anal Calorim 2010; 101:949-957. 33 [42] Allmann R. Refinement of the hybrid layer structure [Ca2Al(OH)6]+·[1/2SO4·3H2O]-. Neues 34 Jahrb Mineral Monatsh 1977; 136-144. 35 [43] Song F, Yu Z, Yang F, Lu Y, Liu Y. Microstructure of amorphous aluminium hydroxide in 36 belite-calcium sulfoaluminate cement. Cem Concr Res 2015; 71:1-6. 37 38 20 Figure Captions 1 Figure 1. SXRPD Rietveld plot for C3A0.845F0.155H6 (λ~0.62Å). Inset details the high-angle region. 2 The tic marks are the allowed Bragg reflections: C3A0.845F0.155H6 lower row; Quartz (internal 3 standard) upper row. 4 5 Figure 2. Selected range of the SXRPD raw patterns for C4AF_g_1.0 recorded at different time of 6 hydration, with the main peaks due to a given phase labelled; AFt: circle, AFm: star, Qz: triangle; 7 CS �H2: rhombus and C4AF: inverted triangle. 8 9 Figure 3. Fe/Ca atomic ratio vs. Al/Ca atomic ratio for SEM-EDS study of a) C4AF_g_1.0, b) 10 C4AF_st-C4A3𝑆_g_1.0 (plus symbol) and C4AF_ss-C4A3𝑆_g_1.0 (crosses) at 48 hours. Solid 11 symbols represent the theoretical composition of the phases: AFt: circle, AFm: star, C4AF: inverted 12 triangle and C4A3𝑆: square. SEM photograph of C4AF_g_1.0 (top) and C4AF_st-C4A3𝑆_g_1.0 13 (bottom). 14 15 Figure 4. Selected range of the SXRPD raw patterns for a) C4AF_st-C4A3𝑆_g_1.0 and b) C4AF_ss-16 C4A3𝑆_g_1.0 recorded at different hours of hydration, with the main peaks due to a given phase 17 labelled as in Figure 2. The long arrow highlights one peak of C4AF. 18 19 Figure 5. Full quantitative phase analysis results with time for a) C4AF_st-C4A3𝑆_g_1.0 and b) 20 C4AF_ss-C4A3𝑆_g_1.0 (C4A3𝑆: square, C𝑆H2: rhombus, C4AF: inverted triangle, AFt: circle, 21 AFm: star and ACn+FW: open square). Calorimetric heat flow curves (dashed line) are also 22 displayed. 23 24 Figure 6. SXRPD Rietveld plots for a) C4AF_st-C4A3𝑆_g_0.7 and b) C4AF_st-C4A3𝑆_g_1.3, at 12 25 hours of hydration, with the main peaks due to a given phase labelled as in Figure 2. Insets detail 26 the low-angle regions. 27 28 21 Table 1. Paste mix proportions in weight percentages (wt%). The total heat evolved at 2 days of hydration is also given. Mixture C 4 AF wt% st-C 4 A 3 𝑺 � wt% ss-C 4 A 3 𝑺 � wt% gypsum wt% water/solid Total heat (J/g) C4AF_0.8 100 - - - 0.8 - C4AF_g_1.0 65.3 - - 34.7 1.0 382.4 C4AF_st-C4A3𝑺 �_g_1.0 32.7 49.1 - 18.2 1.0 382.2 C4AF_ss-C4A3𝑺 �_g_1.0 32.7 - 49.1 18.2 1.0 409.1 C4AF_st-C4A3𝑺 �_g_0.7 32.7 49.1 - 18.2 0.7 - C4AF_st-C4A3𝑺 �_g_1.3 32.7 49.1 - 18.2 1.3 - C4AF_st-C4A3𝑺 �_1.0 66.6 33.4 - - 1.0 - Table 2. Quantitative phase analysis results (wt%) for C4AF_0.8 paste, as a function of hydration time obtained by SXRPD. RWP (%) disagreement factors are also given. Wt% to 1 h 2 h 3.5 h 5.5 h C4AF 46.5(1) 34.9(1) 32.1(1) 28.7(1) 24.9(1) C2AH8 0 2.4(2) 1.5(3) 1.2(4) 1.4(5) C3(A,F)H6 0 9.2(2) 14.7(2) 19.6(1) 24.8(1) ACn+FW 9.1(1)+44.4*=53.4 53.5(1) 51.7(1) 50.5(1) 48.9(1) RWP (%) 9.2 7.0 7.4 7.3 7.3 *Theoretical free water content Table 3. Elemental chemical composition (atomic wt%) of C3(A,F)H6 obtained by Rietveld methodology and transmission electron microscopy. Element Rietveld (wt%) TEM (wt%) Ca 65.7(-) 62(2) Al 24.9(1) 21(2) Fe 9.4(3) 17(3) Table 4. Quantitative phase analysis results (wt%) for C4AF_g_1.0 paste, as a function of hydration time obtained by SXRPD. RWP (%) disagreement factors are also given. Wt% to 5.5 h 12 h 16 h 18 h 21 h 29 h 48 h C4AF 29.7(1) 21.4(1) 19.1(1) 10.4(2) 7.7(3) 4.3(3) 2.6(4) 1.3(3) C𝑺 �H2 14.5(1) 5.6(1) 0.9(1) 0 0 0 0 0 AFt 0 23.6(1) 31.4(1) 25.5(2) 22.1(3) 19.1(4) 15.9(5) 15.5(5) AFm 0 0 0 10.5(3) 14.4(4) 20.5(5) 20.0(6) 21.2(6) ACn+FW 5.8(1)+50*=55.8 49.4(1) 48.5(1) 53.6(2) 55.8(2) 56.1(3) 61.4(3) 62.0(3) RWP (%) 7.0 5.7 5.4 10.8 9.7 13.8 14.3 13.2 *Theoretical free water content 22 Table 5. Quantitative phase analysis results (wt%) for C4AF_st-C4A3𝑆_g_0.7 paste, as a function of hydration time obtained by SXRPD. RWP (%) disagreement factors are also given. to 1 h 4 h 6.5 h 12 h C4A3𝑺 � 25.7(1) 23.7(1) 8.9(1) 7.9(1) 7.7(1) C𝑺 �H2 9.4(1) 9.8(1) 1.3(1) 1.2(1) 1.4(1) C4AF 16.4(1) 16.9(1) 15.2(1) 12.6(1) 11.3(1) AFt 0 4.6(1) 28.9(1) 29.1(1) 27.6(1) AFm 0 0 2.9(1) 3.7(2) 5.2(2) ACn+FW 7.3(1)+41.2*=48.5 45.0(1) 42.7(1) 45.5(1) 46.8(1) RWP (%) 6.0 5.3 4.5 4.5 5.4 *Theoretical free water content Table 6. Quantitative phase analysis results (wt%) for C4AF_st-C4A3𝑆_g_1.3 paste, as a function of hydration time obtained by SXRPD. RWP (%) disagreement factors are also given. to 1 h 3.5 h 6.5 h 8.5 h 12 h C4A3𝑺 � 19.1(1) 18.2(1) 5.2(1) 4.8(2) 5.0(2) 4.6(2) C𝑺 �H2 7.0(1) 6.3(1) 0.4(1) 0 0 0 C4AF 12.2(1) 12.8(1) 9.1(1) 6.6(2) 5.4(2) 3.8(2) AFt 0 3.2(1) 21.9(1) 19.5(2) 18.1(2) 15.2(2) AFm 0 0 2.9(1) 6.4(2) 8.8(2) 11.8(2) ACn+FW 5.3(1)+56.5*=61.8 59.6(1) 60.6(1) 62.7(1) 62.7(1) 64.6(2) RWP (%) 6.0 3.6 4.9 5.6 6.6 8.4 *Theoretical free water content Table 7. Quantitative phase analysis results (wt %) for C4AF_st-C4A3𝑆_1.0 paste, as a function of hydration time obtained by SXRPD. RWP (%) disagreement factors are also given. to 5 h 18 h 20 h C4A3𝑺 � 31.5(1) 29.5(1) 10.9(2) 10.7(3) C4AF 14.9(1) 14.0(1) 15.8(2) 14.8(2) AFm 0 0 17.9(3) 18.1(3) ACn+FW 3.6(1)+50.0*=53.6 56.5(1) 55.4(2) 56.5(2) RWP (%) 5.4 4.2 10.6 10.7 *Theoretical free water content 23 y 10.0 20.0 30.0 40.0 0.0 1.0 2.0 3.0 y 30.0 32.0 34.0 36.0 38.0 40.0 42.0 44.0 -0.5 0.0 0.5 1.0 I (a.u.) 2θ (0) Figure 1 4 6 8 10 12 I (a.u.) 2θ (0) t0 5.5h 48h 29h 18h 12h Figure 2 100 200 300 400 500 600 60 70 80 90 100 -2 -1 0 * T(ºC) Weight (%) Heat Flow (W/g) Figure S2 0 5 10 15 20 25 30 35 40 45 0.00 0.01 0.02 0.03 t(h) Normalized heat flow (W/g) C4AF_st-C4A3S_g_1.0 C4AF_ss-C4A3S_g_1.0 C4AF_g_1.0 Figure S3 Figure S4 (a) (b) 0 5 10 15 20 45 50 0 10 20 30 40 50 60 70 80 90 100 t (h) α (%) Figure S5