Phase Solubility Changes during Hydration of Monocalciumaluminate and Calcite—The Influence of Alkali Accumulation

The reaction of CA (monocalcium aluminate) with calcite was closely monitored with regard to phase development, pore water ion content and heat flow. Calcite acts as filler and reactant, finally leading to thermodynamically stable products after hydration at ambient conditions. For better understanding the mechanism taking place, a CA-cement and a commercial calcite mix were compared to a pure CA and pure calcite mix. Both reaction paths were compared. Thermodynamic modeling with PhreeqC gave insight about factors that can influence the course of the hydration reaction. Alkali ions in pore solution of the CA-cement relocate solubility curves of hydration products. Taking into account as many of the alkaline ions as possible, resulted in the closest representation of the measured phase content, confirming thermodynamic modeling. The high dynamics that develop during reaction could only be addressed if a concentration of alkalis in the pore solution at later points in time was respected, thus leading to a shift of solubility curves over time. This was not observed with the pure CA in absence of alkalis.

Materials 2020, 13 Focusing on the processes in the liquid phase there are calculated thermodynamic data available on the equilibrium surfaces of the system CaO-Al 2 O 3 -CaCO 3 -H 2 O, as well as the influence of the ion Na + on the solubility curves of such a system. The calculation of equilibrium surfaces and their shifts when Na is added, as well as its influence on the phase formation were in focus of Damidot et al. [13].
Subsequently our paper provides a new approach, adding more information by combining computed modeling with measured data. Data of a pure and a technical system (including some impurities) determined from experiments where set in relation to thermodynamic calculations of solubility curves. For our calculations not only Na but also K and S were taken into account.
Some points that are specifically covered and discussed in this paper: • The pore water ion content of the samples during early hydration and its relation to the respective solubility curves and hence the formation of hydration products is compared to measured crystalline phase content.

•
Comparison of a pure CA and calcite mixture to a mix of industrial grade materials is provided (some of the data of the latter was already published in a different context [2]).

•
Insights are given into expected effects of impurities during the application of industrial grade high alumina CAC in general.
The obtained knowledge is also contributing to a deepening of our general understanding of CA hydration in the presence of calcite.

Materials and Methods
The starting materials were pure synthesized CA as well as a commercial CA-cement (with only CA phase) with two calcite sources, pure synthesized and industrial grade calcite, were examined by means of XRF (X-ray fluorescence), QXRD (quantitative X-ray diffraction) and BET (Brunauer-Emmett-Teller method) (data collection from three independent experiments each). All hydration measurements were done at 23 • C with a w/s value of 1.0 (heat-flow calorimetry, in-situ X-ray diffraction). Pore solution data were obtained by ICP-MS (inductively coupled plasma mass spectrometry). For comparison to QXRD thermodynamic modeling of predicted stable phase contents was performed.

Commercial Calcite
Due to the paste properties and enhancing reproducibility a commercial calcium carbonate (high specific surface area of 21.7 ± 0.1 m 2 /g BET) was used for experiments with the CA-cement. Acting as a fine filler, leading to a very reproducible start of the main reaction, it also functions as a carbonate source. CA-cement and calcite ratio was 70/30 (wt%) during experiments. Table 1 shows the chemical analysis by XRF and QXRD data obtained by G-factor method [14] can be found in Table 2. P 2 O 5 contained in the calcite is a modification residue to achieve higher BET surface. No crystalline phase was detected by XRD but instead amorphous content.  Table 2. QXRD results from G-factor method (* data as published in [2]).

High Purity Calcite
The pure calcite was used in combination with the synthesized CA in order to avoid introduction of foreign ions. The producer gives a purity of >99.5 % with a sulfate content of ≤0.01 %. The high purity calcite has a BET surface area of 0.8 ± 0.2 m 2 /g. XRF analysis data are given in Table 1 and QXRD results obtained by G-factor method can be found in Table 2.

Characterization of the CA-Cement
The commercial CA-cement used contains mainly CA. G-factor method [14] shows a CA 2 content of 4.7 ± 0.2 wt%. and 8.5 wt% amorphous phase. Chemical composition by XRF is shown in Table 1, Table 2 provides the data of QXRD phase analysis. A value of 2.51 ± 0.05 m 2 /g for BET surface area was determined (measured with Gemini 2360 Micromeritics, Micromeritics Instrument Corp., Norcross, GA, USA). Initial Na release was tested by shaking the pure CA-cement with deionized H 2 O for ten minutes (w/c = 0.7-similar to used mixtures). A readily soluble Na concentration of 2.3 ±0.1 mmol/L was measured.

Characterization of Pure CA
The material was synthesized in platinum crucibles from a stoichiometric mix of CaCO 3 and Al 2 O 3 by milling followed by calcination at 1000 • C, then the mix was re-milled and afterwards sintered at 1400 • C. A BET surface of area of 0.6 ± 0.1 m 2 /g was measured. A test of Na release (CA + H 2 O w/c = 0.7, shaken for 10 min.) revealed a readily soluble Na concentration ≤0.1 mmol/L. XRF analysis data was obtained (Table 1) and QXRD results by G-factor method can be found in Table 2.

Sample Composition for Hydration Experiments
The compositions of the CA-cement mixture and the pure mixture are given in Table 3.

Heat Flow Calorimetry
In order to obtain comparable data, isothermal heat flow calorimetry was performed for all mixes at 23.0 • C to determine hydration behaviour. All tests were accomplished with a TAM Air calorimeter (TA Instruments, New Castle, DE, USA). The device was equipped with InMixEr tools (a customized solution for sample equilibration and mixing of water and cement inside a calorimeter [15]). To ensure sufficient equilibration the base line was observed. Before the measurement was started the base line was set to zero. Water injection into the mixture was chosen as starting point for data recording. Homogeneity of the samples was achieved by manual stirring for 30 s and then stirring with an external motor (constant stirring rate of 715 rpm) for 1 min. Three independent measurements were performed.

Pore Solution Analysis and Description of Thermodynamic Calculations
Pore solution extraction started from 0.25 h until end of main period of hydration for each system examined. Containers of the dry mixes of CA-source plus CaCO 3 -source (each ca. 50 g) and sealed containers containing H 2 O were temperature-equilibrated for 10 h at 23.0 ± 0.1 • C. H 2 O was stirred with the dry powder for 1 min. Pore solution extractions were performed for the mixtures after following intervals considering the heat flow curve (0.25, 0.5, 1, 2.5, 3, 3.5, 4, 5, 6, 8 h for the CA-cement mix and 0. 25,1,5,10,15,17,20,22,25 h with the pure CA mix) at 23 ± 0.1 • C. Different extraction methods were required due to paste workability. Pastes still showing workability were centrifuged for 8 min (3220 g). For samples after setting, pore solution extraction by pressing (pressure ≤ 47.6 N/mm 2 ) was deployed. A hydraulic press (Stürmer Metallkraft wpp30, Stürmer Maschinen, Hallstadt, Germany) equipped with an extraction cell (Remt Industries, Grena, France) was used. Resulting pore solution was filtered (0.2 µm syringe filter) and HNO 3 was added in defined amounts in order to prevent precipitation. Ion concentrations of Ca, Al, Na, K and S were investigated with a Thermo iCapQ ICP-MS (ThermoFisher, Langenselbold, Germany). To determine inorganic carbon content CO 2 concentration exetainer vials containing defined amounts of pore solution and HNO 3 were used. Analyses were carried out on an IRMS Delta plus XP with a Gasbench II (ThermoFisher, Langenselbold, Germany). Blank value for inorganic carbon in deionized H 2 O was <0.05 mmol/L.
The thermodynamic database Cemdata 18 [16] was used in order to calculate the solubility curves in the system Na/K-Ca-Al-C. The advantage of such calculations and plots arises from the possibility to plot the evolution of the pore solution. The diagrams allow to visualize the equilibrium of the pore solution with the precipitated hydrate phases which helps discussing the results gained from pore solution experiments. With the GEM-Selector 3.3 [17][18][19] and the CEMDATA 18 database [19] the activities of different ion species where calculated considering the measured ion concentrations. Activity coefficients where computed using the extended Debye-Hückel equation. This data was used further, to calculate the solubility curves of hydrates (line where the saturation index is zero) within the system Ca-Al. For further insights into the computation of such curves and the thermodynamic background the authors like to refer to recently published articles [20][21][22].
The pH was calculated with the program PhreeqC Interactive 3.4.0-12927 [23] from the pore water ion content with the Cemdata 18 database [16].

QXRD and in-Situ QXRD Procedure and Set-up of Thermodynamic Calculations
A Bruker D8 (Bruker Corporation, Billerica, MA, USA) diffractometer operated at 23 ± 0.1 • C XRD was used to for all XRD experiments. Raw sources of CA and CaCO 3 were analysed by powder XRD (instrument data Table A1-column dry samples). The measurement parameters (angular range up to 70 • 2θ, small step size and counting time/step) were chosen to determine the purity of the material in a precise way. In-situ measurements parameters are listed in Table A1-column in-situ XRD/pastes. The goal was an efficient way to determine the quantitative phase development of crystalline phases during a rapid reaction. Therefore step size and counting time/step where chosen in a way to keep the measurement time for each diffractogram short (10 min per diffractogram). The angular range is also short. It starts at 6 • 2θ to avoid missing any hydrate phase peaks. The preparation of the pastes was done by stirring equilibrated dry mixes and H 2 O for 1 min. Then they were applied to a temperature-controlled sample holder. Kapton ® X-ray film (RESOGOO GmbH & Co. KG, Westerrönfeld, Germany) was stretched above the samples to reduce evaporation of mixing water. Phase development over time was determined by in-situ measurements up to 40 h. Structure data for Rietveld refinement is presented in Table A2. For quantitative phase content G-Factor method [14,24] was applied with exception of C 2 AH 8 (specification as scale factor). C 2 AH x phases are shown in arbitrary units of their scale factors in the following figures.

Heat Flow Calorimetry
Heat flow and heat of hydration of the systems examined are shown in Figure 1

191
During dormant period Ca content drops from a maximum of 30 mmol/L to 11 mmol/L. Al stays Particle sizes and surface areas of the reactants are known to have significant influence on the start of cement main reaction, due to the filler effect of the fine [25,26]. In the used samples the different heat flow behavior is mainly a result of the different surface areas and thus both the reactivity of the compounds as well as the filler effect. This is to be expected regarding the use of an industrial product Materials 2020, 13, 1406 6 of 16 line and a self-synthesized compound. The heat flow was measured to get an overview of the different reaction times, to create a reproducible experimental setting for planning of water extraction.
Pore solution composition during CAC hydration was found to be dominated by initial CA dissolution [27], which is already happening during mixing in both systems (see first hour in Figure 1) and no influence of particle size distribution can be expected at this time.

215
The pure CA with pure CaCO3 shows a hydration behavior comparable to the commercial CAC 216 with commercial CaCO3 as shown in the chapters before. However, there are some significant 217 differences that will be discussed.

218
The highest Ca content measured for the pure CA mix was 31 mmol/L while for CAC it was 24

227
The arrows in Figure 3 mark the average path along the Ca and Al ions follow during the 228 reaction (red for the commercial CAC system, blue for the pure CA system). The pure CA mix shows 229 a comparable amount of dissolved Ca ions at early hydration times compared to the CA-cement.

230
During the whole reaction of the pure CA the Al amount dissolved is lower than in the CA-cement.

231
In conclusion it can be noted that the pore solution evolves differently in both systems, although 232 it can be seen that the hydration is resulting in the formation of similar hydrate phases.  [2]. ICP-MS error is usually in ng/L range while measured data is in the range of 10 to 600 mg/L, therefore error was not depicted.
During dormant period Ca content drops from a maximum of 30 mmol/L to 11 mmol/L. Al stays at a concentration of 37 mmol/L until 10 h before it drops to 8 mmol/L. Inorganic carbon is constantly below 0.1 mmol/L during the whole time observed. Na concentration in the pore solution stays at 0.1 mmol/L during the dormant period and increases to 3 mmol/L after the main reaction starts. This is caused by hydration of CA, releasing Na into the pore water ( Figure 2a). Additionally, the analysis also shows a sulphur (S) content of 0.9 ± 0.3 mmol/L in the pore water during the whole measurement, which might be a residue from the used calcite.

Pore Solution Data for CA-Cement with Commercial CaCO 3
For the CA-cement with commercial CaCO 3 the calculated pH ranged from 11.8 to 12 during the dormant period (Figure 2b

Comparison of the Pore Solution Evolution in the Systems Examined
The pure CA with pure CaCO 3 shows a hydration behavior comparable to the commercial CAC with commercial CaCO 3 as shown in the chapters before. However, there are some significant differences that will be discussed.
The highest Ca content measured for the pure CA mix was 31 mmol/L while for CAC it was 24 mmol/L. While the reactions took place the Ca ion contents of pure CA with pure calcite declined ending at 12 mmol/L and for the commercial CAC with commercial calcite 1 mmol/L.
Aluminum concentrations in the pore solution also dropped during hydration of both systems examined. For CA the starting Al concentration was 38 mmol/L and dropped to 8 mmol/L. In the commercial CAC system Al concentration starts at 47 mmol/L declining to 36 mmol/L at its lowest.
For the pure CA the Ca to Al ratio at 0.25 h is close to the ratio in dry CA, namely 0.5. This suggests that at the start of the hydration, the Ca and Al ion content is determined by the dissolution of CA. Later the Ca ion content rises.
The arrows in Figure 3 mark the average path along the Ca and Al ions follow during the reaction (red for the commercial CAC system, blue for the pure CA system). The pure CA mix shows a comparable amount of dissolved Ca ions at early hydration times compared to the CA-cement. During the whole reaction of the pure CA the Al amount dissolved is lower than in the CA-cement.    In conclusion it can be noted that the pore solution evolves differently in both systems, although it can be seen that the hydration is resulting in the formation of similar hydrate phases. Theoretical crystalline CA amount in the mixed paste at hydration start should be 35 wt%. The detected amount of CA in the paste before the main reaction is the result of a thin water film on top of the XRD sample leading to a small underquantification of CA. After 25 h initial amounts have decreased to 7 wt%. During hydration (see Figure 4) C 2 AH x phases were detected from 11 h on, increasing until 15 h and dissolving again until end of measurement. CAH 10 and AH 3 formation starts after 12 h. There is a slight deviation to the start of the heat flow at 14 h observed in Figure 1.

272
As it can be seen from modelling the initially formed C2AHx phase is not stable in the system.

273
Hence, the dissolution of this phase as measured is in accordance with the equilibrium in the system.

274
CAH10 is also formed and is still present at end of our measurement, although it is not 275 thermodynamically stable. However, since our examination duration is only one day it is possible 276 that the destabilization and dissolution of the initially formed CAH10 phase is much slower compared 277 to the dissolution of the C2AHx phase. Therefore, it does not take place within our examination period.

278
There is a surplus of CaCO3 which will not react and hence, acts as a filler. In both systems the

Phase Development of CA-Cement with Commercial CaCO 3
Initial crystalline CA amount in the mixed paste at hydration start was calculated and should be 30 wt%. A small underquantification can be seen. However, after 15 h initial amounts have decreased to 3 wt%. Phase content of CA 2 does not change significantly within the 20 h measured in our experiments (see Figure 4a). During hydration (see Figure 4b) C 2 AH x phases were detected from 3.5 h onwards. C 2 AH x amount increases until 6 h followed by a drop to zero after 15 h. CAH 10 and AH 3 formation starts slightly before 5 h, correlating well with the observed heat flow in Figure 1. After 5 h precipitation of hemicarbonate (HC) and monocarbonate (MC) was observed. Hemicarbonate content reaches a maximum at 7 h, then decreases reaching zero after 15 h (see Figure 4c).

Summary and Discussion
Although both samples mainly contain CaAl 2 O 4 (CA) as a reactive phase, the CAC shows a number of impurities, especially alkalis (determined by solubility test). Those are a result of the production of the contained Al 2 O 3 in the cement from the Bayer process and do influence the hydration reaction.
Especially the solubilities of the hydrate phases formed can be influenced through the presence of alkalis. Additionally, dissolution of the oxides influences the pH of the pore solution of a cement. Figure 5 shows the phase development regarding the thermodynamic stable outcome that was predicted by GEMS calculation. The main stable phases that should develop are monocarbonate (MC) and AH 3 . Both phases are formed, as predicted by the thermodynamic model.

288
The Ca to Al ratio passes different stability fields during hydration (see Figure 6). The solution 289 generally is supersaturated in AH3 but passes from possible precipitation of amorphous AH3 to the 290 microcrystalline modification. Different specimen of AH3 were described and discussed by

291
Lothenbach et al. [28]. Precipitation of AH3 with a low crystallinity was observed by QXRD from 12 As it can be seen from modelling the initially formed C 2 AH x phase is not stable in the system. Hence, the dissolution of this phase as measured is in accordance with the equilibrium in the system. CAH 10 is also formed and is still present at end of our measurement, although it is not thermodynamically stable. However, since our examination duration is only one day it is possible that the destabilization and dissolution of the initially formed CAH 10 phase is much slower compared to the dissolution of the C 2 AH x phase. Therefore, it does not take place within our examination period. There is a surplus of CaCO 3 which will not react and hence, acts as a filler. In both systems the predicted phase composition at equilibrium conditions could not be reached within the first day of hydration.

Solubility Curves and the Pure CA Mix
The first calculations of solubility curves with PhreeqC took into consideration the following ions: Ca, Al, Na, K. The solubility curves for the phases of interest in the system Ca Al at low alkali concentrations (0.05 mmol/L K, 0.5 mmol/L Na) are shown in Figure 6.
The Ca to Al ratio passes different stability fields during hydration (see Figure 6). The solution generally is supersaturated in AH 3 but passes from possible precipitation of amorphous AH 3 to the microcrystalline modification. Different specimen of AH 3 were described and discussed by Lothenbach et al. [28]. Precipitation of AH 3 with a low crystallinity was observed by QXRD from 12 h onwards. Nevertheless, the calculated solubility curves and the measured ionic concentrations do not fit very well if the XRD-measured phase assemblages are considered. E.g., C 2 AH x is always calculated as supersaturated although a clear decline was observed by XRD in reality.
For a proper modelling which is in accordance with the experimental findings even minor amounts of S (<1 mmol/L) in the pore solution have to be taken into account [28]. This need was addressed in further calculations and the actual progress of the reaction during the different stability fields can be observed in Figure 7.
generally is supersaturated in AH3 but passes from possible precipitation of amorphous AH3 to the 290 microcrystalline modification. Different specimen of AH3 were described and discussed by Lothenbach et al. [28]. Precipitation of AH3 with a low crystallinity was observed by QXRD from 12 h onwards. Nevertheless, the calculated solubility curves and the measured ionic concentrations do 293 not fit very well if the XRD-measured phase assemblages are considered. E.g., C2AHx is always 294 calculated as supersaturated although a clear decline was observed by XRD in reality. For a proper modelling which is in accordance with the experimental findings even minor 301 amounts of S (<1 mmol/L) in the pore solution have to be taken into account [28]. This need was 302 addressed in further calculations and the actual progress of the reaction during the different stability 303 fields can be observed in Figure 7.

304
The first pore solution composition is mainly driven by an initial dissolution of the cement

320
A logarithmic presentation of (a), giving insight in the effect of small scale changes at low Ca content.

321
The circled zone marks the area where the system is heading.

322
The presence of S showed the necessity of including all ions in the calculations, shifting the 323 solubility curves and leading to a fitting model of the reaction, able to describe the measured phase

326
There is a noticeable increase of Na and K during hydration of the commercial CAC. Therefore, 327 solubility curves for high and low alkali contents were plotted in the same diagram, in order to show 328 the development of the pore solution in the commercial CAC system. The evolution of the pore 329 solution during this reaction and the different solubility curves can be observed in Figure 8.

330
It has to be considered that the reaction is never static but an ongoing process, leading to an 331 increase of the concentration of alkalis in the pore water over time (no alkali phases can precipitate).

332
Thus, the solubility curves shift during hydration time to lower calcium concentrations, as indicated 333 with arrows in Figure 8. The first pore solution composition is mainly driven by an initial dissolution of the cement phases and the calcite. The composition of the first pore solution is in a region where the solubilities of the phases formed during hydration show almost equal solubility. Hence, C 2 AH x (here presented as C 2 AH 7.5 ), HC, MC and AH 3 all show supersaturation and as a consequence all phases are formed during the early hydration.
However, it can be seen that the pore solution follows the solubility curve of MC during the whole hydration. The reason for this is, that the final composition of the pore solution will be in equilibrium with the thermodynamically stable phases MC and AH 3 , which is indicated as a grey circle in Figure 7b. As a consequence, the pore solution drifts in the direction of equilibrium following the solubility of MC. At around 18 h the pore solution crosses the solubility curve of C 2 AH 7.5 .
At low Al values solubilities of CAH 10 , C 2 AH 7,5 , C and HC differ significantly. Hence, following the ion content compared to its solubility curve C 2 AH 7,5 is dissolved again.
The presence of S showed the necessity of including all ions in the calculations, shifting the solubility curves and leading to a fitting model of the reaction, able to describe the measured phase content.

Solubility Curves and the Influence of Alkalis on Pore Solution Evolution in the Commercial CA-Cement
There is a noticeable increase of Na and K during hydration of the commercial CAC. Therefore, solubility curves for high and low alkali contents were plotted in the same diagram, in order to show the development of the pore solution in the commercial CAC system. The evolution of the pore solution during this reaction and the different solubility curves can be observed in Figure 8.  It has to be considered that the reaction is never static but an ongoing process, leading to an increase of the concentration of alkalis in the pore water over time (no alkali phases can precipitate). Thus, the solubility curves shift during hydration time to lower calcium concentrations, as indicated with arrows in Figure 8.
During the complete reaction the solution is supersaturated with respect to microcrystalline AH 3 . A low crystalline form of AH 3 is the first phase that can be observed by QXRD slightly before 5 h. Predicted precipitation of other phases like MC, HC, C 2 AH 7,5 and CAH 10 can be correlated to phases observed by XRD starting from 5 h onwards.
Again initial dissolution leads to supersaturation of all phases during the first 5 h of reaction. Pore solution at times later than 5 h evolves alongside with equilibrium to the stable phases (MC, AH), in accordance to GEMS modelling shown in chapter 4.1.
At low Ca values solubilites of C 2 AH 7,5 , MC and HC differ, as can be observed in the enlarged Figure 8b. The solubility curve of C 2 AH 7,5 is crossed and formed C 2 AH x dissolves from 6 h onwards.
To show the immense impact of an increase of alkali content on the solubility of a hydrate phase, curves for two examples are given in Figure 9. It can be seen that there is a significant shift of the solubility curves of C 2 AH 7.5 ( Figure 9a) and MC (Figure 9b) with increasing alkalis. Hence, it can be observed why the pore solution composition of the commercial CAC + CaCO 3 system drifts to lower Ca values. The pore solution will be in equilibrium with the thermodynamically stable phases at the end of hydration, which are MC and AH 3 in our system. During hydration the pore solution always follows the solubility curve of the stable phase MC. Hence, the drift of the pore solution composition is in accordance with the drift of the solubility curves as indicated in Figure 9b.
To show the immense impact of an increase of alkali content on the solubility of a hydrate phase, 347 curves for two examples are given in Figure 9. It can be seen that there is a significant shift of the observed why the pore solution composition of the commercial CAC + CaCO3 system drifts to lower

Impact of ion content on the path of reaction:
The comparison of a pure CA & calcite mix to a CA-cement & industrial calcite mix enabled insight in the influence of alkali-release into pore solution on the possible paths of reaction. A similar direction of the Ca to Al ratio was observed, although the total ratios varied significantly in pure CA and CA-cement mix. The solubility curves are shifting over time. This effect is connected to the increase of alkalis due to increasing concentration of the pore solution.
Impact of alkalis on the solubility of hydrate phases: With increasing Na & K content the solubilities curves of C 2 AH 7.5 and MC in the pore water shift. Supersaturation occurs already at a lower Ca content. Precipitation is possible at very low Ca concentrations in the end of the hydration reaction because of the increased alkali concentrations.
Pore solution always follows the solubility of the stable phase: In our system the thermodynamic stable phase is monocarbonate. During the hydration the pore solution always follows the solubility curve of this phase. This development is fueled by the ongoing precipitation of monocarbonate and further dissolution of the reactive CA, thus following an equilibrium along the solubility curve. At the end, the zone where both, monocarbonate and AH 3 coexist must be reached. This is the case where the respective solubility curves intersect.

Future outlook:
The understanding of the reaction paths and resulting phase formations can be greatly enhanced by thermodynamic calculations. This knowledge might be of use for creating future experiment set ups that utilize a certain artificial ion content to influence the reaction path into a pre calculated hydration regime.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results. Table A1. Settings for powder and in-situ XRD (Bruker D8 Advance, Lynx Eye detector).