Hydration mechanisms of two polymorphs of synthetic ye'elimite

Manuscript submitted to Cement and Concrete Research Hydration mechanisms of two polymorphs of synthetic ye’elimite A. Cuestaa, G. Álvarez-Pinazoa, S. G. Sanfélixb, I. Peralc, M. A. G. Arandaa,c, A. G. De la Torrea,*. a Departamento de Química Inorgánica, Universidad de Málaga, Campus Teatinos S/N. 29071-Málaga, Spain. b Unidad Técnica de Investigación de Materiales, AIDICO, Avda. Benjamín Franklin, 17 Paterna, Valencia, Spain. c ALBA-CELLS synchrotron, Carretera BP 1413, Km. 3.3, E-08290 Cerdanyola, Barcelona, Spain. * Corresponding author. Tel.: +34952131877; fax: +34952132000. E-mail address: [email protected] (A.G. De la Torre) 1 Abstract 1 Ye’elimite is the main phase in calcium sulfoaluminate cements and also a key phase in sulfobelite 2 cements. However, its hydration mechanism is not well understood. Here we reported new data on 3 the hydration behaviour of ye’elimite using synchrotron and laboratory powder diffraction coupled 4 to the Rietveld methodology. Both internal and external standard methodologies have been used to 5 determine the overall amorphous contents. We have addressed the standard variables: water-to-6 ye’elimite ratio and additional sulfate sources of different solubility. Moreover, we report a deep 7 study of the role of the polymorphism of pure ye’elimites. The hydration behaviour of orthorhombic 8 stoichiometric and pseudo-cubic solid-solution ye’elimites is discussed. In the absence of additional 9 sulfate sources, stoichiometric-ye’elimite reacts slower than solid-solution-ye’elimite, and AFm-10 type phases are the main hydrated crystalline phases, as expected. Moreover, solid-solution-11 ye’elimite produces higher amounts of ettringite than stoichiometric-ye’elimite. However, in the 12 presence of additional sulfates, stoichiometric-ye’elimite reacts faster than solid-solution-ye’elimite. 13 14 15 Keywords: Hydration mechanism (A), Rietveld method (B), dissolution and crystallization kinetics 16 (A), calcium sulfoaluminate (D) ye’elimite (D). 17 18 2 1. Introduction 1 Ye’elimite is the most important phase in calcium sulfoaluminate cements (CSA) [1], which are 2 very promising environmentally-friendly materials. These cements address one of the major 3 concerns of the cement industry as they allow decreasing the CO2 footprint of cement production. 4 On average, for every ton of ordinary Portland cement (OPC) produced, 0.97 tons of CO2 are 5 released into the atmosphere, with the cement industry contributing around 6% of all anthropogenic 6 CO2 emissions translating into approximately 4% of the planet’s global warming[2]. By 7 comparison, CSA cements are produced with significantly lower CO2 emissions relative to OPC, 8 achieved through the use of lower amount of carbonated raw-materials (part of calcite is replaced 9 by gypsum) and a reduced clinkering temperature [1]. The overall CO2 emission reduction can 10 amount up to 40% [3]. CSA binders may have quite variable compositions, but all of them contain 11 ye’elimite phase, Ca4Al6O12SO4, also called Klein’s salt or tetracalcium trialuminate sulfate 12 (C4A3𝑆 in cement nomenclature) as their main phase [4,5]. The term CSA cements are usually 13 reserved for those clinkers containing more than 50 wt% of ye’elimite and they may also have 14 minor amount of phases as belite, tetracalcium aluminoferrate, anhydrite, gehlenite or mayenite [6]. 15 Cements with large amounts of ye’elimite may have special applications such as high strength 16 developments at early-ages [7]. Ye’elimite is also present, ~25 wt%, in sulfobelite cements [8,9]. 17 CSA cements were introduced into the Chinese market as a result of the high-performing and 18 dimensionally-stable cementitious matrices develop by China Building Materials Academy [10]. In 19 Europe, the use of CSA cements is limited by the lack of standards concerning special cements 20 derived from non-Portland clinkers. Nevertheless, the manufacture and marketing of CSA cements 21 has recently been started by several European cement companies [4,11-13]. 22 This work is focus on the reactivity of ye’elimite. Stoichiometric ye’elimite has been reported to be 23 orthorhombic at room temperature [(14, 15] and solid solutions of this phase crystallize in a cubic 24 structure [16,17]. There are some studies about the influence of clinkering process on mineral 25 formation of calcium sulfoaluminate based clinkers [18,19]. They have proved that minor elements, 26 such as iron, form solid solution with ye’elimite and stabilize the cubic form. The effect of 27 ye’elimite polymorphism on hydration mechanism has been also addressed. However, these 28 systems are complex and there are too many parameters, f.i. type and amount of sulfate source, 29 belite polymorphism, water/solid ratio, etc., affecting the hydration mechanisms. Consequently, a 30 simplification of the problem by the study of pure phases is advisable. 31 During early age hydration of ye’elimite in the presence of a sulfate source like gypsum, bassanite 32 or anhydrite, ettringite (AFt, C6A𝑆3H32) phase is the main crystalline hydration product, reaction 33 3 (1) [20, 21]. On the other hand, ye’elimite is able to react with water to form AFm (monosulfate) 1 according to reaction (2) [20]. In both reactions amorphous aluminium hydroxide is formed. 2 C4A3𝑆 + 2C𝑆Hx + (38-2x)H → C6A𝑆3H32 + 2AH3 (1) 3 C4A3𝑆 + 18H → C4A𝑆H12 + 2AH3 (2) 4 It is known [20, 21] that for molar ratios of calcium sulfate to ye’elimite larger than 2, only reaction 5 (1) is taking place. However, there are some contradictory results concerning the reactivity of 6 ye’elimite with water in the absence of another sulfate source. Some authors stated that only 7 reaction (2) takes place [20] while others have published that mixtures of AFt and AFm phases are 8 produced [19, 22]. 9 On the other hand, X-ray powder diffraction (XRPD) is very well suited for in-situ studies of 10 chemical processes involving crystalline materials [23, 24]. During the last years, it has been 11 reported quantitative phase analysis of cements, clinkers and supplementary cementitious materials 12 by combining XRPD and Rietveld methodology [13, 25-27]. This combination results in Rietveld 13 quantitative phase analysis (RQPA). More recently, this procedure has been expanded to hydrated 14 cementitious systems [23, 28] and in some of these studies, the non diffracting fraction was 15 determined [29-31], although a more precise term has been coined: Amorphous and Crystalline not-16 quantified, content [32]. Furthermore, the use of an intense X-ray source, such as synchrotron X-17 rays, coupled with a fast X-ray detection system permits time-resolved diffraction experiments 18 allowing in-situ measurements during the hydration process of cements [19, 23]. 19 Here, we report a hydration study of two synthetic ye’elimite samples: stoichiometric ye’elimite 20 that presents an orthorhombic unit cell [15] and solid-solution ye’elimite that crystallizes in a 21 pseudo-cubic unit cell. The final goal is to understand the ye’elimite hydration mechanisms as a 22 function of ye’elimite polymorphism, water content and type and content of sulfate source. In order 23 to do so, laboratory and synchrotron XRPD (LXRPD and SXRPD) and Rietveld methodology are 24 employed. Kinetics of hydration have been established and correlated to calorimetric data. This 25 study is a step forward to better understand the eco-cement CSA performances at early ages. 26 27 2. Experimental Section 28 2.1. Sample Preparation. 29 Stoichiometric ye’elimite (C4A3𝑆), labelled hereafter st-C4A3𝑆, was prepared as previously 30 reported [15]. Solid-solution ye’elimite, labelled ss-C4A3𝑆 hereafter, was prepared as follows 31 (nominal composition: Ca3.8Na0.2Al5.6Fe0.2Si0.2O12SO4). Suitable amounts of CaCO3 (99.95%, 32 Alfa Aesar), Al2O3 (99.997%, Alfa Aesar), Fe2O3 (99.945%, Alfa Aesar), SiO2 (99.56%, ABCR), 33 4 Na2CO3 (99.999%, Sigma Aldrich) and CaSO4·2H2O (ground natural single-crystal from Málaga) 1 were used to obtain approximately 8 g of ss-C4A3𝑆. This composition was chosen following 2 previous studies of active belite CSA [13]. In that work we speculated with the simultaneous 3 presence of Na, Fe and Si within cubic ye’elimite as a consequence of an electron microscopy 4 study. Moreover, it has been previously reported [33] that the cubic symmetry could be restored by 5 substituting larger caged ions for Ca2+ or SO42to expand the framework to a non-collapsed state 6 and by smaller framework cations such as B3+, Si4+ or Fe3+ for Al+3. Consequently, following our 7 previous results and that information reported by others, we were able to obtain ss-C4A3𝑆 sample as 8 a pure single phase with the proposed composition. Further details about crystal structure and 9 thermal behaviour of this sample will be published elsewhere. 10 The mixture was ground for 1 hour in an agate mortar with ethanol. The resulting powder was 11 pelletized (20 mm diameter and 500 MPa) and heated at 1250ºC for 4 hours (heating rate of 5 12 ºC/min) followed by a rapid cooling. The pellets were ground in an agate mortar. Figure 1 shows 13 the LXRPD raw patterns for both anhydrous samples. St-C4A3𝑆 sample presents 1.2(2) wt% of 14 C3A, 4.4(2) wt% of CA and 4.0(2) wt% of C12A7 as impurities [15]. 15 St-C4A3𝑆 and ss-C4A3𝑆 were mixed with gypsum (g) or anhydrite (a), in some cases, according to 16 the stoichiometry of reactions (1) and (2). Table 1 reports pastes mix proportions, including 17 water/solid (w/s) ratios. The gypsum used for the hydration studies was that marketed by BELITH 18 S.P.R.L. (Belgium). Anhydrite was produced by heating that gypsum at 700ºC for 1 hour. Blaine 19 fineness for all the samples was ranged between 4400-5000 cm2/g. 20 Two experimental set ups were employed: i) in-situ SXRPD and ii) ex-situ LXRPD. For the in-situ 21 SXRPD study all the anhydrous mixtures were mixed with 15 wt% SiO2 (99.56%, ABCR) as an 22 internal standard [34] and powder diffraction data were collected to obtain the initial phase 23 assemblage (t0). It is important to bear in mind that in the water/solid ratio the amount of internal 24 standard is not taken into account. Pastes were ex-situ prepared and immediately loaded into glass 25 capillaries of 0.5 mm of diameter with a syringe. The capillaries were sealed with grease to avoid 26 any water loss. Moreover, we were aware that the internal standard could influence the 27 hydration/crystallization processes. We performed an internal study with and without internal 28 standard and check the reproducibility of the Rietveld QPA, observing only minor differences. 29 Consequently, we could trust in the results obtained by this methodology. On the other hand, for the 30 ex-situ LXRPD study, pastes were poured into hermetically closed Teflon® tubes in the form of 31 cylinder until 1 day. Then, the samples were taken out and stored within demineralised water at 32 20ºC. Pieces were taken out at ages ranging between 2 and 7 days. One fraction of the pastes was 33 milled to fine powder in an agate mortar. In order to stop the hydration process, the procedure was 34 5 filtration in a Whatman system (90 mm diameter Whatman filter with a pore size of 2.5 μm on a 1 Teflon support) with acetone twice and finally with ether. These samples were stored in a closed 2 desiccator (without vacuum application) to avoid further hydration and/or carbonation. 3 2.2. Thermal analysis. 4 Differential thermal analysis (DTA) and thermogravimetric (TGA) measurements were performed 5 in a SDT-Q600 analyzer from TA instruments (New Castle, DE) for stopped-hydration pastes. The 6 temperature was varied from RT to 1000ºC at a heating rate of 10 ºC/min. Measurements were 7 carried out in open platinum crucibles under nitrogen flow. The weighed loss from RT to 600ºC 8 was computed to be water (chemically bounded water) and that from 600 to 1000ºC was considered 9 as CO2. Table 2 and Figure S1, S2 and S3 (given as supporting information) report the TGA results. 10 2.3. Laboratory X-Ray powder diffraction (LXRPD) and Synchrotron X-Ray powder 11 diffraction (SXRPD). LXRPD data were recorded on an X’Pert MDP PRO diffractometer 12 (PANalytical) equipped with a Ge (111) primary monochromator, using strictly monochromatic 13 CuKα1 radiation (λ=1.54059 Å) and an X’Celerator detector. An overall measurement time of ∼ 4h 14 per pattern was required for good statistics over the angular range 5.0 - 70.0º (2θ) with a 0.017º step 15 size. 16 SXRPD patters were collected in Debye-Scherrer (transmission) mode using the X-ray powder 17 diffraction station of ALBA synchrotron (Barcelona, Spain) [35]. The wavelength, 0.61975(1) Å, 18 was selected with a double-crystal Si (111) monochromator and determined from Si640d NIST 19 standard (a=5.43123 Å). The diffractometer is equipped with a MYTHEN detector especially suited 20 for time-resolved experiments. The capillaries were rotated during data collection to improve 21 diffracting particle statistics and the synchrotron beam was focused in the detector to improve the 22 diffraction peak shape. The data acquisition time was ∼15 min per pattern to attain very good 23 signal-to-noise ratio over the angular range 1-35º (2θ). The temperature inside the experimental 24 hutch was 26(1) ºC. 25 2.4. XRPD Data Analysis. 26 Raw SXRPD patterns were normalized taking into account the decay of X-ray beam flux with time. 27 SXRPD and LXRPD patterns were analysed by using the Rietveld methodology as implemented in 28 the GSAS software package [36], in order to obtain RQPA. The refined overall parameters were 29 background coefficients, cell parameters, zero-shift error, peak shape parameters, and phase scales. 30 Peak shapes were fitted by using the pseudo-Voigt function [37]. The ACn contents were 31 determined by internal standard methodology [34] from SXRPD data and by external standard 32 method (G-factor) from LXRPD data as detail previously [13, 30] 33 6 2.5. Calorimetry. The isothermal calorimetric study was performed in an eight channel Thermal 1 Activity Monitor (TAM) instrument using glass ampoules. Pastes were prepared ex-situ by mixing 2 ~ 6 g of each sample with the appropriated water and were immediately introduced in the 3 calorimeter. A stabilization period of 45 minutes was needed to start the measurements. The heat 4 flow was collected up to 7 days at 20ºC. 5 3. Results and discussions 6 3.1. Hydration of stoichiometric ye’elimite with variable water/solid ratios and without 7 additional sulfate source. 8 Hydration mechanism of st-C4A3𝑆 was initially studied without additional sulfate source and two 9 different w/s ratios (0.58 and 1.16), see Table 1. These pastes have been studied at early ages, up to 10 30 hours, by in-situ SXRPD with internal standard methodology. Time-resolved SXRPD was 11 employed to track the dissolution of the anhydrous phases followed by the crystallization of the 12 different hydrated phases (AFt and AFm). Moreover, the pastes were also prepared into cylinders 13 and studied with the external standard methodology, G-factor, at 2 days and 7 days by ex-situ 14 LXRPD. 15 The w/s ratio of 0.58 corresponds to the stoichiometric amount of water according to reaction (2) 16 with 10% of excess. The w/s value of 1.16 is the double of the previous value to study the effect of 17 large water excess on hydration mechanisms. Table 3 shows the phase assemblages at different ages 18 for st-C4A3𝑆_0.58 mixture. It is clear that for this w/s ratio the dissolution (and reaction) rate is very 19 slow, since up to 30 hours very small amount of ye’elimite was dissolved and there are no new 20 crystalline hydrates. After 2 days ye’elimite was partially dissolved and the precipitation of small 21 quantities of AFm and AFt were quantified. This observation indicates that both reactions (2) and 22 (1) are taking place. This observation disagrees with the thermodynamic calculation that predicts 23 that only reaction (2) should take place [20]. However, this experimental behaviour has been 24 previously reported [22]. 25 However, to disentangle the extension of both reactions is not an easy task. Interconversion between 26 AFt and AFm may take place depending upon the experimental conditions including the sulfate 27 concentration in the pore water. It has been observed that the crystallization of ettringite reaches a 28 maximum at 2 days of hydration, see Table 3, and then it diminishes. This effect has consequently 29 caused a lack of sulfate ions in the pore solution. For longer hydration times, the formation of AFm 30 is favoured not only from ye’elimite reaction but also likely from AFt dissolution. 31 It is well known that higher amounts of water enhance ye’elimite reactivity [38]. Thus, a paste with 32 w/s of 1.16 was also prepared, st-C4A3𝑆_1.16. Figure 2 shows raw SXRPD and LXRPD patterns as 33 7 a function of time, with peaks due to a given phase labelled. Table 4 gives RQPA results for st-1 C4A3𝑆_1.16 from SXRPD and LXRPD data. Figure 3a shows the degree of reaction of ye’elimite 2 with water. Comparing Tables 3 and 4 and inspecting Figure 3a it is confirmed that reactivity has 3 been enhanced, achieving over 75% of degree of reaction at 7 days with w/s 0.58 and 20 hours, with 4 w/s 1.16. Calorimetric data, see Table 1 and Figure S5, show that more heat is released for w/s ratio 5 of 1.16 than for 0.58, which is in full agreement with a larger reaction degree. In this case, the main 6 hydration product has been AFm, at all hydration times, indicating that higher amounts of water 7 favour reaction (2). This behaviour is in agreement with that previously reported in [20], where only 8 AFm was obtained with a w/s ration of 2.0. However, w/s ratio of 1.16 is not high enough to avoid 9 reaction (1) and a small amount of AFt is also quantified. It is also observed that AFt content is 10 maximum at 2 days, and at later ages partly reacts. 11 The quantification of AFm-type phases presents two important problems: i) broad diffraction peaks 12 due to both poor crystallinity and highly disorder structures and ii) the lack of structural 13 descriptions for some phases. Thus, the crystal structure reported for C4A𝑆H12 [39] has been used 14 to quantify all AFm-type phases by adjusting c-values, as previously reported [19]. Other AFm type 15 phases, such as monoand hemicarbonate AFm phases [40,41], were checked but they did not fit 16 properly the patterns. The AFm contents reported in Tables 3 and 4 are expressed as the total 17 amount of AFm-type phases. Figure 5 shows Rietveld plots of st-C4A3𝑆_1.16 at 2 days of hydration 18 where all the AFm-types phases have been labelled, as an example of a complex sample. Moreover, 19 we have observed that stopping procedure has affected the mineralogical composition, especially 20 AFm-type phases [42]. In the in-situ synchrotron experiment, the hydration was not stopped and 21 higher crystallinity and no modifications of its basal spacing was observed, see Figure 2 up to 31 h 22 of hydration. On the other hand, the ex-situ LXRPD data were collected for stopped samples and 23 broader peaks are observed. Three AFm-type phases with modified c-values are needed to fit the 24 pattern, see broad peaks labelled with a star in Figure 2 at 2 days. 25 Tables 3 and 4 include the ACn values obtained from internal and external standard methodologies. 26 The first column gives t0 values obtained from the SXRPD pattern of the anhydrous samples. Free 27 water, FW, in this column is the theoretical value. Remaining values obtained from internal 28 standard method encompass not only ACn but also FW (not chemically bound water) and are 29 expressed as a single value in Tables 3 and 4. This is due to the inability of the internal standard 30 methodology to distinguish between different not-diffracting phases. On the other hand, data 31 obtained with G-factor methodology from LXRPD, 2 days and 7 days, corresponds only to ACn 32 values, since FW was removed by the stopping hydration procedure. The FW contents were 33 determined by the difference between the ‘theoretical/mixed’ water and the combined water 34 8 determined from TGA study (from RT to 600ºC), Table 2 and Figures S1, S2 and S3. Consequently, 1 in Tables 3 and 4, ACn and FW values are given for data obtained from LXRPD. It is important to 2 highlight that the results obtained by the internal standard method are in agreement with those 3 obtained at later ages showing the consistence of both methodologies. 4 3.2. Hydration of ye’elimite as function of polymorphism and without additional sulfate 5 source. 6 In order to understand the role of polymorphism in the hydration mechanism, a solid-solution 7 ye’elimite paste, with a w/s ratio of 1.16, (ss-C4A3𝑆_1.16), has also been studied. Table 5 shows 8 RQPA results at the measured ages. Comparing Tables 4 and 5, it can be observed that solid-9 solution ye’elimite reacts at a faster pace since after 12 hours the degree of reaction of ye’elimite is 10 ~75%, Figure 3a. In addition to a faster kinetics, solid-solution ye’elimite yields much larger 11 relative amounts of AFt. Figures 4a and 4b show the evolution of sulfate content with time for st-12 C4A3𝑆_1.16 and ss-C4A3𝑆_1.16 pastes. The amount of residual sulfate content was calculated using 13 the quantitative data reported in Tables 4 and 5 for C4A3𝑆, given as solid symbols. In addition, the 14 crystallized sulfate content was calculated from the quantified ettringite in Figure 4a (crossed 15 symbols) and global sulfate content from the amounts of ettringite jointly with AFm-phases in 16 Figure 4b. The dotted and dashed lines in Figures 4a and 4b represents the maximum sulfate group 17 content which can crystallize in each sample. As mentioned before, AFt crystallization in ss-18 C4A3𝑆 Ris higher than in the stoichiometric sample. Moreover, ~17% and ~10% of hydrated sulfate 19 groups for st-C4A3𝑆 and ss-C4A3𝑆, respectively were mainly incorporated into ACn phase(s) and/or 20 in pore solution. Consistently with the former two studies, after two days the amount of AFt starts 21 to decrease. Calorimetric data, see Table 1 and Figure S5 (given as supporting information), show 22 that more heat is released by the solid-solution sample, due to a larger reaction degree of solid-23 solution-ye’elimite and also to a larger relative amount of ettringite at seven days. The higher 24 dissolution rate of ss-C4A3𝑆 may be related by the release of Na+ to the pore solution, provoking a 25 similar effect (although less pronounced) than KOH [20]. 26 3.3. Hydration of ye’elimite with different soluble sulfates and variable water/solid ratios. 27 9 Table 1. Paste mix proportions in weight percentages (wt%). The total heat evolved at 7 days of hydration is also given. Mixture st-C 4 A 3 𝑺 � wt% ss-C 4 A 3 𝑺 � wt% gypsum/anhydrite wt% water/solid Total heat (J/g) st-C4A3𝑺 �_0.58 100 - - 0.58 391 st-C4A3𝑺 �_1.16 100 - - 1.16 555 ss-C4A3𝑺 �_1.16 - 100 - 1.16 577 st-C4A3𝑺 �_g_0.71 64.1 - 35.9/- 0.71 - st-C4A3𝑺 �_g_1.42 64.1 - 35.9/- 1.42 518 ss-C4A3𝑺 �_g_1.42 - 64.1 35.9/- 1.42 488 st-C4A3𝑺 �_a_1.62 69.0 - -/31.0 1.62 566 ss-C4A3𝑺 �_a_1.62 - 69.0 -/31.0 1.62 544 Table 2. Summary of the weight losses from the TGA study for pastes without additional sulfate source. mixture Hydration time / d Theoretical weight loss # / % Weight loss RT-600ºC / % Weight loss 600-1000ºC / % st-C 4 A 3 𝑺 �_0.58 2d 36.7 25.5 0.9 7d 36.7 32.1 1.3 st-C 4 A 3 𝑺 �_1.16 2d 53.7 36.7 1.5 7d 53.7 35.8 1.8 ss-C 4 A 3 𝑺 �_1.16 2d 53.7 38.8 1.8 7d 53.7 39.9 1.8 # Theoretical weight loss: total initial free water. Table 3. Quantitative phase analysis results (wt%) for st-C4A3𝑆_0.58 paste, as a function of hydration time obtained by SXRPD and LXRPD. RWP (%) and χ2 Rietveld agreement factors are also given. SXRPD LXRPD to 8h 24h 30h 2d 7d st-C4A3𝑺 � 61.1(1) 54.1(1) 51.6(1) 46.4(1) 34.3(1) 10.6(2) AFt - - 0.3(1) 0.5(1) 6.3(3) 4.7(1) AFm - - 0.6(1) 3.2(1) 7.1(3) 20.2(4) ACn +FW 2.2(1) +36.7*=38.9 45.9(1) 47.6(1) 49.9(1) 41.0(4) +11.3=52.3 59.9(5) +4.9=64.5 RWP(%) 6.2 5.5 5.2 5.6 8.1 10.3 χ2 29.7 30.7 28.2 25.3 4.6 7.0 *Theoretical free water content 16 Table 4. Quantitative phase analysis results (wt%) for st-C4A3𝑆_1.16 paste, as a function of hydration time obtained by SXRPD and LXRPD. RWP (%) and χ2 Rietveld agreement factors are also given. SXRPD LXRPD to 12h 18h 24h 31h 2d 7d st-C4A3𝑺 � 44.9(1) 43.0(1) 29.4(1) 10.3(3) 9.2(3) 4.6(2) 2.9(2) AFt - - 2.3(1) 1.1(2) 1.0(2) 9.2(3) 6.8(3) AFm - - 7.1(1) 22.0(2) 22.2(2) 25.1(4) 25.8(5) ACn +FW 1.4(1) +53.7*=55.1 57.0(1) 61.2(1) 66.5(2) 67.6(2) 44.1(5) +17=61.1 46.5(6) +17.9=64.4 RWP (%) 6.2 4.9 4.7 11.4 12.8 10.0 10.5 χ2 29.7 20.9 20.3 137.8 199.1 7.1 7.7 *Theoretical free water content 17 Table 5. Quantitative phase analysis results (wt%) for ss-C4A3𝑆_1.16 paste, as a function of hydration time obtained by SXRPD and LXRPD. RWP (%) and χ2 Rietveld agreement factors are also given. SXRPD LXRPD to 6h 8h 12h 15h 18h 24h 2d 7d ss-C4A3𝑺 � 38.2(1) 37.6(1) 37.2(1) 10.6(2) 5.4(2) 2.8(3) 1.7(2) 0.7(1) 0 AFt - 0.5(2) 0.9(2) 5.4(2) 6.3(3) 6.7(3) 7.0(3) 14.7(1) 12.7(1) AFm - - - 13.4(2) 16.3(2) 18.7(2) 18.5(2) 13.4(1) 14.3(1) ACn +FW 8.1(2) +53.7*=61.8 61.9(2) 61.9(2) 70.6(2) 71.9(2) 71.8(2) 72.8(2) 58.0(5) +13.1=71.1 61.0(5) +12=73.0 RWP (%) 9.8 8.6 8.6 7.8 8.9 9.5 10.4 6.6 7.0 χ2 60.3 64.1 55.6 79.3 86.1 84.1 142.9 4.4 4.4 *Theoretical free water content 18 Table 6. Quantitative phase analysis results (wt%) for st-C4A3𝑆_a_1.62 paste, as a function of hydration time obtained by SXRPD. RWP (%) and χ2 Rietveld agreement factors are also given. Wt% to 1h 2h 3h 7h st-C4A3𝑺 � 23.9(1) 25.0(1) 17.2(1) 2.3(2) 0.7(1) C𝐒 � 10.8(2) 11.1(1) 10.0(1) 6.0(1) 0.3(1) AFt - 1.3(1) 4.3(1) 19.0(2) 45.2(3) AFm - - 4.4(1) 10.5(2) 1.4(1) ACn +FW 4.0(1) +61.3*=65.3 62.6(1) 64.2(1) 62.3(1) 52.3(1) RWP (%) 5.5 3.2 3.8 6.2 3.9 χ2 22.2 7.8 12.0 32.0 17.5 *Theoretical free water content Table 7. Quantitative phase analysis results (wt %) for ss-C4A3𝑆_a_1.62 paste, as a function of hydration time obtained by SXRPD. RWP (%) and χ2 Rietveld agreement factors are also given. Wt% to 2h 3h 7h 9h ss-C4A3𝑺 � 20.6(1) 22.4(1) 22.6(1) 20.4(1) 17.8(1) C𝐒 � 10.7(2) 11.2(1) 11.6(1) 9.9(1) 8.7(1) AFt - - 0.2(1) 3.5(2) 8.1(2) ACn +FW 7.4(1) +61.3*=68.7 66.4(1) 65.6(1) 66.2(1) 65.3(1) RWP (%) 7.7 5.5 5.5 5.4 5.0 χ2 63.5 29.8 28.5 21.7 16.6 *Theoretical free water content 19 Figure 1 10 15 20 25 30 35 40 ss-C4A3S st-C4A3S C12A7 CA C3A I (u.a.) 2θ(0) 10 12 14 16 18 20 22 Figure 2 0.04 0.06 0.08 0.10 0.12 0.14 0.16 senθ/λ(Å-1) t0 18 h 31 h 2 d 7 d 0 20 40 60 80 100 110 165 0 20 40 60 80 100 st-C4A3S_0.58 st-C4A3S_g_0.71 st-C4A3S_g_1.42 ss-C4A3S_g_1.42 st-C4A3S_a_1.62 ss-C4A3S_a_1.62 st-C4A3S_1.16 ss-C4A3S_1.16 t (h) α(%) α(%) a) b) Figure 3 110 160 180 0 2 4 6 8 a) b) Sulfate content (wt%) t (h) max ss max st max ss max st Figure 4 y 5.0 10.0 15.0 20.0 25.0 LXRPD Count s X10\ u 3 0. 0 2. 0 4. 0 2\gH, deg Figure 4 RWP=10.0% χ2=7.1 Figure 5 0 20 40 60 80 100 wt (%) 0.000 0.005 0.010 0.015 0.020 0.025 Normalized heat flow (W/g) Ye’elimite Gypsum ACn+FW AFt 0 5 10 15 20 0 20 40 60 80 100 wt (%) t (h) 0.000 0.005 0.010 0.015 0.020 0.025 Normalized heat flow (W/g) Figure 6 a) b) Figure S5 020 40 60 80 100 120 140 160 0 100 200 300 400 500 600 Total heat (J/g) t (h) st-C4A3S_0.58 st-C4A3S_1.16 ss-C4A3S_1.16 st-C4A3S_g_1.42 ss-C4A3S_g_1.42 st-C4A3S_a_1.62 ss-C4A3S_a_1.62 110 0 2 4 6 8 10 12 14 110 0 2 4 6 8 10 12 14 Sulfate content (wt%) max st & ss max ss max st a) b) t (h) Figure S6