Ab initio mechanism revealing for tricalcium silicate dissolution

Dissolution of minerals in water is ubiquitous in nature and industry, especially for the calcium silicate species. However, the behavior of such a complex chemical reaction is still unclear at atomic level. Here, we show that the ab initio molecular dynamics and metadynamics simulations enable quantitative analyses of reaction pathways, thermodynamics and kinetics of the calcium ion dissolution from the tricalcium silicate (Ca3SiO5) surface. The calcium sites with different coordination environments lead to different reaction pathways and free energy barriers. The low free energy barriers result in that the detachment of the calcium ion is a ligand exchange and auto-catalytic process. Moreover, the water adsorption, proton exchange and diffusion of water into the surface layer accelerate the leaching of the calcium ion from the surface step by step. The discovery in this work thus would be a landmark for revealing the mechanism of tricalcium silicate hydration.

seen. Those "particles" are way larger than Ca(OH)2, and they might be more likely C-S-H particles precipitated at the surface or less dense C3S region due to partial dissolution. I was extremely surprised when I arrived to line 308 and I realise that TEM is from the authors, not from another work. I don't understand what the TEM is contributing to the paper-Overall, I recommend rejection of the paper. The calculation of the energy barriers are in my opinion interesting, but the paper over shells the relevance. They do not cite all the relevant papers, they make statements that are clearly wrong, and there are methodological problems.
Reviewer #2 (Remarks to the Author): The manuscript NCOMMS-21-44726 « Ab initio mechanism revealing for tricalcium silicate dissolution » presents atomic scale simulations of calcium cations dissolution at the (111) surface of M3 Ca3SiO5 (C3S in cement chemistry notation). The authors ran well-tempered ab initio metadynamics simulations to determine reaction pathways, free energy changes and free energy barriers. The results of this study are originals and of importance for the fundamental understanding of Ca3SiO5 dissolution.
In spite of the originality and the considerable work needed for simulations and analysis, several points should be considered to improve the manuscript: 1) It is not indicated to which oxygen atoms Os refers to. Generally, Os refers to oxygen in silicates, but I supposed that here Os refers to superficial oxygen atoms without discrimination between isolated oxygen anions (Oi) and oxygen in silicates. In both cases, the naming code for chemical species (Ow, Os) should be provided, and more information should be given on Ca-Oi bonds for which the breakage occurs earlier (l. 106-108 for ex.).
2) a) It is also not clear why the (111) surface was chosen for the study. To facilitate the reader's understanding, a top view of this surface before and after protonation occurs, indicating the position of Caɑ and Caβ, should be provided. It would be also pertinent to discuss on the difference in the environment of Ca at this surface compared with other ones. b) Furthermore, the statement: "the results obtained in this work are general and applicable to other kinds of calcium silicate species" (l. 209-210) should be revised. In my opinion, the results are very specific to C3S and arguably to the surface under study. If not so, the applicability of the results to other minerals must be proven. The method however can be applied to other minerals.
c) The formation of stable hydroxyl on the surface before starting the production runs should be discussed because this may influence the dissociation of Ca2+. Moreover, it is not stated when the proton transfer to the Oi of the second layer occurs, and when the water molecule comes to occupy the calcium site. A timeline with the important events could improve the quality of the manuscript.
3) According to l. 176-177, the regions IV, V and VI are similar to the "perfect" Ca3SiO5. Are they similar in width, in intensity and/or position of the peaks? What about the other regions (I, II, III) ? 4) More details should be provided with regard to the following sentence on the spontaneous and unspontaneous nature of Ca dissociation: "The thermodynamic and kinetic analyses show that the detachment of Ca is spontaneous when the Ca is not fully dissolved and unspontaneous when Ca is no more coordinated with Os." (l. 231-233).
5) Other small changes may improve the quality of the manuscript: -The duration of the production and equilibrium runs should be given in a list or a table to improve the readibility (l. [272][273][274][275] -"the M3 type of Ca3SiO5 (obtained from CCSD33), which is most frequently observed in industrial clinkers" → "the M3 polymorph of Ca3SiO5 (obtained from CCSD33), which is the most frequently observed in industrial clinkers together with the M1 form" -The region of the zoom in Fig 4f should be reduced accordingly to the scale bar. The snapshot of the simulation is ~1 nm in width.
-Metadynamics simulations applied to calcium dissociation from Ca3SiO5 surfaces was already performed with ReaxFF. A reference to doi:10.3390/ma12091514 would be pertinent.

Jérôme Claverie
The authors present ab initio molecular dynamics (AIMD) simulations of the dissolution of calcium ions from the (111) surface of calcium silicate. The authors used metadynamics to calculate the free energy change of dissolution of two different calcium ions from the surface. The computational work carried out is impressive. AIMD are very computationally demanding techniques and the authors have combined this with the use of metadynamics. There are a few studies coming out in the literature using this technique, so the work described in this paper is significant. Moreover, the length scale described is quite substantial. I believe that this work is noteworthy as it can provide an atomistic description of the dissolution of this material. The methodology is sound and the results support the conclusions made by the authors.
I have some concerns about the applicability of the research for the wider community especially the experimentalists. The systems described in this paper are very simple and from the pictures provided, seem to be quite small. What is the size of the system being simulated? How can the authors be sure that the free energy change is not an artefact of the small surface area? Would the results be the same if you simulated a system with double the surface area? My concern is that the results could be impacted by the periodic images.
I also have questions surrounding the speciation of the surface. The authors mention that previous studies have shown that there are hydroxyl groups on the surface. How did the authors determine the chemical environment of the surface and therefore justify that the modelled surface is consistent with experimental surface speciation? Moreover, I assume that the pH of the system is 7, therefore, what is the chemical environment of this surface at that pH. Are experiments of dissolution of this material conducted at different pHs.

1
The authors would like to thank the reviewers for their constructive comments and suggestions. Our detailed responses to the specific points raised are given below. The modifications are in red.

Reviewer #1
Comment 1: The paper "NCOMMS-21-44726, Ab initio mechanism revealing for tricalcium silicate dissolution" starts with an interesting premise: compute the dissolution energy barriers of C 3 S using atomistic simulation methods. However, soon in the introduction I have the feeling that the authors are over-selling their research. For instance, starting in line 58: "All these hypotheses and the fit of the calorimetric curve of Ca 3 SiO 5 hydration using thermodynamic calculations may be derived from the ignorance on the interfacial reactions during the Ca 3 SiO 5 dissolution, especially the dissolution behavior of calcium ions at atomic level. Fortunately, atomistic simulations can tackle these problems." Atomistic simulation results, including those presented in this paper are truly far from explaining the calorimetric curve, way further than the "thermodynamic calculations may be derived from the ignorance" that the authors mention. As a matter of fact, we don't even know if dissolution is controlling the hydration of C 3 S, or the nucleation of C-S-H is the main mechanism. In addition, they mention "especially the dissolution behavior of calcium ions at atomic level". Why Ca ions?
Why not SiO 4 ions? there is not reason behind this statement.
Response: Thanks for your comments. We fully agree with your statement that the atomistic simulation results cannot comprehensively explain the calorimetric curve of Ca 3 SiO 5 hydration, and we are sorry for making you confused by our expression. In fact, we just want to introduce the significance of understanding the interfacial reactions during the Ca 3 SiO 5 dissolution. So, we revised the sentence starting in line 58 as follows: Understanding the interfacial reactions at the water/Ca 3 SiO 5 interface using atomistic 2 simulations can provide some supplementary and new insights on the Ca 3 SiO 5 dissolution.

(lines 59-60, page 3)
In addition, As suggested by the reviewer, we deleted the misleading sentence "especially the dissolution behavior of calcium ions at atomic level". As for whether the Ca ion or silicate group dissolves first, it has no solid evidence from experimental observation yet. However,  Chem. Soc., Faraday Trans., 1991,87, 497-505). On the contrary, the Ca ion in solution is generally six-coordinated with the O ion in water molecules (Manzano et al. Langmuir 2012, 28, 9, 4187-4197). The Ca ion on surface is usually coordinated with less than six O ions, thus the water molecules tend to adsorb on the Ca site (Qi et al. Applied Surface Science, 518 (2020) 146255). In addition, the dissolution process is 3 usually described by the ligand-exchange model (Werner et al. Colloids Surf., A 1997, 120, 143−146) and there is few ligands coordinated with the silicate group. So, it is unlikely that the Ca 3 SiO 5 dissolution begins with the dissolution of silicate group. For a more rigorous proof, we will conduct experiments to get solid observation in the future. But undeniable, there are also silicate groups in Ca 3 SiO 5 hydration solution that maybe attributed to the dissolution of Ca making the silicate group isolated with only water molecules surrounding it and more dynamic compared to the initial state. This process is similar to the experimental observation that there is a hydrated layer containing monomeric and subsequently dimeric silicate units around the C 3 S surface (Juilland et al. Cement and Concrete Research, 40 (2010) 831-844) ).

Comment 2:
All the text starting in line 65: "Claverie et al.27 investigated the proton transfer at the water/Ca 3 SiO 5 interface using ab initio molecular dynamics (AIMD) simulations and found that the hydroxides formed on the surface are highly stable. However, they did not observe an obvious vertical displacement of Ca ions relative to the initial position. In fact, it is very hard to probe a complete calcium dissolution process at the atomic level even using the traditional molecular dynamics (MD) simulations28, 29 with large timescale (i.e. nanoseconds). Moreover, the classical MD is not appropriate to simulate the chemical reaction involving the breakage and formation of bonds. Even using the ReaxFF force field is still not preciously enough to present the reaction pathways for a chemical reaction." is written to justify the present work and has no fundament. Cleaverie et al. did not observe Ca dissolution, true, and traditional MD cannot reproduce water dissociation, but then, Why "the ReaxFF force field is still not preciously enough to present the reaction pathways for a chemical reaction"? There is no reasoning or citation behind this statement. In fact, ReaxFF (2019). Reactivity of Different Crystalline Surfaces of C 3 S During Early Hydration by the Atomistic Approach. Materials, 12(9), 1514. The second one is specially relevant, as they use metadynamics to compute dissolution energy barriers. They use ReaxFF instead DFT, but then the methodology is more or less the same as in the present paper. The results are very different, and even if I believe that the present work is more rigorous, the results should be compared and discussed. The simulations of the free energy are interesting, nicely done and nicely presented. As I said, I think that this work is better done than Uddin et al. But I think that there is a mayor methodological problem: why the 111 surface? It is hard to say, but I would say given the structure and symmetry of the MIII polymorph that the created slab is not symmetric. Therefore, how is the plane chosen? Did the authors do any surface reconstruction? are Ca atoms symmetrically distributed in the top and bottom of the slab? The 5 results are compromised by these factors.
Response: Thanks for your pertinent comments. We deleted the statement on "the ReaxFF force field is still not preciously enough to present the reaction pathways for a chemical reaction" and updated the literature as suggested by the reviewer. We rewrote the paragraph as follows: Understanding the interfacial reactions at the water/Ca 3 SiO 5 interface using atomistic simulations can provide some new insights on the Ca 3 SiO 5 dissolution. The adsorption of water on the Ca 3 SiO 5 surface with molecular and dissociative mode 25 is the first step of Ca 3 SiO 5 hydration, which happens even before contacting the bulk water due to the strongly hydrophilic nature of Ca 3 SiO 5 26 . After the surface hydroxylation and the proton hopping into the surface 27 , the Ca ion will dissolve into the solution destroying the initial surface topology and promoting the further water penetration 27 , which is a key step for advancing the Ca 3 SiO 5 hydration. For this process, the density functional theory (DFT) -based geometry optimization calculations 28 indicated the adsorption of water on the Ca ion impairs the bonds strength between the calcium and oxygen ions on the surface. Recently, reactive MD simulations have been widely used to study the Ca 3 SiO 5 dissolution and successfully obtain several new perception on dissolution process. Manzano et al. 27 found the Ca ion desorbs quickly and tends to accumulate as inner-and outer-sphere complexes at the Ca 3 SiO 5 (111) surface. Qi et al. 29 showed a more easier Ca dissolution from the Ca 3 SiO 5 (010) surface than the Ca 2 SiO 4 (100) surface due to the higher surface hydroxylation degree. Sun et al. 30 did not observe dissolution of Ca ions from the (010) surface even after 10 ns at 300 K, but after raising the temperature to ~1000 K, the dissolution rate increases five times than that of room temperature. Claverie et al. 31 first investigated the Ca 3 SiO 5 hydration using ab initio molecular dynamics (AIMD) simulations and found that the hydroxides formed on the surface are highly stable. However, they did not observe an obvious vertical displacement of 6 Ca ions relative to their initial positions. In fact, it is very hard to probe a complete calcium dissolution process at the atomic level using the AIMD simulations 29, 32 with small timescale (i.e. within 100 ps). As for the choice of the surface, as the reviewer said, it is a very essential methodological 7 question needed to be solved. The construction of the slab model for MIII polymorph of Ca 3 SiO 5 is extremely complicated due to its low symmetry. Therefore, before this work, we had done a lot of work to study the Ca 3 SiO 5 surfaces and the corresponding results have been illustrated in detail in our recently published paper and its supplementary information (Li et al.      surfaces were maintained neutral with integer numbers of basis to preclude the polarizing electric field 10, 11 . The vacuum was set as 20 Å with a dipole correction along the z direction to get rid of bogus contributions arousing from asymmetry. The two uppermost layers of atoms were completely relaxed while the rest were fixed 12 . The lattice constants of slab models were fixed 13 . No symmetry was forced on both sides of slabs. The optimization threshold was 0.02 eV Å −1 for ionic relaxations and the k-points mesh is 2 × 2 × 1.
Then we calculated the surface energy using the following expression: Where and are energies of unrelaxed and relaxed surfaces, N is the number of formula units embodied in the slab, is the bulk energy per formula unit and is the in-plane area of slab model.
There is a great discrepancy of surface energies between different terminations even in a same miller index ( Figure 8). It is known that the temperature needed for calcinating the pure Ca 3 SiO 5 is more than 1500 ℃ (1773 K). According to the definition of the Boltzmann constant (k B ), if we describe the relationship between the motion and energy at the molecular level, 1500 ℃ can provide 2.48 J/m2 ( = )) for the system. This thermal energy is not only to make the surface cleaved from the bulk crystals, but also for the decomposition of raw materials and phase transitions of Ca 3 SiO 5 From this perspective, we can approximately say the low-index surfaces with various terminations can be easily formed in the C 3 S calcination process, but it is difficult to exactly identify which surfaces and terminations will form in high-temperature environment and multi-steps reactions. Here, we choose the (111) surface, which has a greater possibility to form and confirmed by the previous DFT calculations ( Durgun et al. J. Phys. Chem. C 2014, 118, 28, 15214-15219) as an example to study the dissolution mechanism. Although the free energy barriers and changes would be various at different surfaces, the nature of dissolution of Ca from C 3 S would be similar. In the future, we will continue to get a more comprehensive understanding of the dissolution process on other surfaces and Ca sites, but now this is out of our scope. In the TEM image that cannot be seen. Atoms, or complexes, cannot be seen. Those "particles" are way larger than Ca(OH) 2 , and they might be more likely C-S-H particles precipitated at the surface or less dense C 3 S region due to partial dissolution. I was extremely surprised when I arrived to line 308 and I realise that TEM is from the authors, not from another work.
I don't understand what the TEM is contributing to the paper-Response: Thanks for your comments. As suggested by the reviewer, we deleted the content about the MSD and the TEM. As for the calculated IR spectra, it has several distinctive advantages compared to the experimental IR spectra. One of the most powerful feature for the calculated IR spectra is that we can compute the IR spectra for an isolated component of a mixture system, which would be very useful for the surface system. In the experimental IR spectra, we cannot extract the spectrum of the adsorbent only and omit the contributions from the surface. In addition, our readers are not only the computational chemists, but also the experimentalists. For computational chemists, they are used to get the bond information through visualizing the trajectory. However, for the experimentalists, they would like to obtain some spectroscopy data that can be mutually confirmed with their experimental results.
Our IR spectra for the isolated Si-OH and Ca-OH cannot be obtained by the experimental data, thus we think our data would be a useful supplementary information for the experimental IR spectra in Ca 3 SiO 5 hydration. If the reviewer still thinks that it is unnecessary to keep this data, we will delete the relevant content later.
Reviewer #2 (Remarks to the Author): The manuscript NCOMMS-21-44726 « Ab initio mechanism revealing for tricalcium silicate dissolution » presents atomic scale simulations of calcium cations dissolution at the (111) surface of M3 Ca 3 SiO 5 (C 3 S in cement chemistry notation). The authors ran well tempered ab initio metadynamics simulations to determine reaction pathways, free energy changes and free energy barriers. The results of this study are originals and of importance for the fundamental understanding of Ca 3 SiO 5 dissolution.
In spite of the originality and the considerable work needed for simulations and analysis, several points should be considered to improve the manuscript: Comment 1: It is not indicated to which oxygen atoms Os refers to. Generally, Os refers to oxygen in silicates, but I supposed that here Os refers to superficial oxygen atoms without discrimination between isolated oxygen anions (Oi) and oxygen in silicates. In both cases, the naming code for chemical species (Ow, Os) should be provided, and more information should be given on Ca-Oi bonds for which the breakage occurs earlier (l. 106-108 for ex.). Response: Thanks for your comments. As said by the reviewer, the situation that the initial configuration of the system influences the outcome of the simulation does exist in the traditional AIMD or MD, due to their insufficient sampling ability. However, metadynamics is an efficient enhancing sampling method for accelerating the MD simulations. Through intermittently introducing an external history-dependent bias potential (in this work are Gaussian kernels) on a few selected degrees of freedom, also called collective variables (CVs), the system is forced to get out of the low energy basin, cross the large energy barrier, and go into more regions on the free energy landscape. Therefore, nearly all the configurations in the designated CVs space can be visited for many times, which are inaccessible in equilibrium MD. In fact, there is no absolute "stable hydroxyl on the surface" in the metadynamics simulations. Through adding a bias potential, the "stable hydroxyl"

Response
formed on the Ca 2+ will dissociate and associate for many times, which guarantees an Response: Thanks for your comments. When the free energy changes between the two states are negative, we think this process is spontaneous and vice versa. We have described the free energy changes in detail in the results section. In addition, As suggested by the reviewer, we explain this conclusion in the discussion section as follows: Besides, when the Ca ion is not fully dissolved, the detachment of the Ca ion is spontaneous, -Metadynamics simulations applied to calcium dissociation from Ca3SiO5 surfaces was 21 already performed with ReaxFF. A reference to doi:10.3390/ma12091514 would be pertinent.

Response:
Thanks for your comments. The duration of the production and equilibrium runs was given in a table as follows: The sentence "M3 type of Ca 3 SiO 5 (obtained from CCSD), which is most frequently observed in industrial clinkers" was revised according to the reviewer.
As suggested by the reviewer 1, the Fig 4f was removed already.
As also suggested by the reviewer 1, we discussed this paper (doi:10.3390/ma12091514) in our manuscript at this time, for brevity, please kindly find the answers in the Response to comment 1 and 2 for the reviewer 1.
"pass through state B first". Why "first"? l. 150: "to exist in reality", what is meant by "in reality"? l. 152-153 "welcoming one more water molecule" l. 158: "processes the greatest possibility" l. 178-179:"the first and second hydration shells … are more than" → "the intensity of the peaks which corresponds to the first and second hydration shells…is greater than" l. 181-183: "a more dynamic property and stronger diffusion ability" l. 190-196: the sentence is long and hard to understand in some parts (e.g. "the proton transfer from the water molecule to the second layer of the interstitial oxygen ion") l. 220: "but easy in reality" l. 23 291-292: " to further investigation of the dissolution process to a larger extent" Response: Thanks for your kind and careful corrections on our manuscript. We revised all the errors and inappropriate expressions as suggested by the reviewer. However, the suggestion " "increase in free energy" should be more appropriate than "increase in free energy changes" " would change what we meant to express originally. The free energy change is ΔA, which is different from the free energy A. Thus, we think the original expression is more appropriate.
Reviewer #3 (Remarks to the Author): There are a few studies coming out in the literature using this technique, so the work described in this paper is significant. Moreover, the length scale described is quite substantial. I believe that this work is noteworthy as it can provide an atomistic description of the dissolution of this material. The methodology is sound, and the results support the conclusions made by the authors. I have some concerns about the applicability of the research for the wider community especially the experimentalists. The systems described in this paper are very simple and from the pictures provided, seem to be quite small. What is the size of the system being simulated? How can the authors be sure that the free energy change is not an artefact of the small surface area? Would the results be the same if you simulated a system with double the surface area? My concern is that the results could be impacted by the periodic images.
Response: Thanks for your comments. The size of the systems for metadynamics simulations is 14.21 Å × 11.72 Å × 36 Å and for equilibrium AIMD is 14.21 Å × 11.72 Å × 48 Å. These systems are not small for the AIMD simulations with such a long simulation time (totally more than 200 ps). Here are some outstanding papers regarding the AIMD simulations, in which the size of the systems can be a reference for the reviewer (Liu et al. ACS Catal. 2021, 11, 19, 12336-12343;Chen et al. Nat Commun 12, 3725 (2021);Li et al. Nat. Mater. 18, 697-701 (2019)). For precluding the periodic images effect, we followed the method suggested by Wang et al. (Wang et al. J. Am. Chem. Soc. 2013, 135, 29, 10673-10683.) that the distance between the targeted object and itself in the next periodic image should be equal to or larger than about 5 Å. Moreover, we did not observe the targeted Ca ion is affected by itself in the next periodic image. As for the surface area, it is well known that the surface coverage of adsorbents will influence the adsorption energy. However, in this work, the surface coverage degree of water molecules is one hundred percent. Thus, under the condition that the periodic images effect is removed, no matter how large the slab, it will not influence this value. As for whether our free energy surface would change as the surface area becomes larger, it really needs to test. In fact, we have followed the reviewer's suggestion and doubled our models to test whether the free energy surface will change. But, when the system becomes double, the time for computation increased more than fourfold. Because it only makes sense when we compare two converged free energy surfaces, we must get a converged free energy surface using this doubled model, which is really computationally demanding. We will continue to run this doubled model and study the surface area effect on free energy changes in detail in the future.  (2021) ). In fact, the surface of Ca 3 SiO 5 can be partially hydroxylated before contacting the bulk water due to the strongly hydrophilic nature of Ca 3 SiO 5 and humidity in the air. Thus, before adding bulk water on the Ca 3 SiO 5 surface models, we first put isolated water molecules to all the possible adsorption sites and did the other newly formed OH on the surface. Thus, it is not very important to set the chemical environment artificially because the chemical environment will evolve as the AIMD proceeds.
What's more, the metadynamics has a strong sampling ability, which can sample nearly all the chemical environment around the targeted Ca ion.
As for the pH value, it can be a microscopic or macroscopic quantity, which is difficult to be considered at atomistic scale. It is unreasonable to compute the pH value only using hundreds of water molecules and several H + or OH -. In fact, the influence of pH on the experimental results of dissolution can be interpreted as the influence of the H + or OHon the targeted Ca ion. In the AIMD, the dissociation of the water and association of the H + and OHare "on-the-fly" as the evolution of the simulations. For example, there is no OHaround the targeted Ca ion initially, but at the end of the AIMD, the stable state of the system is the dissolved Ca ion coordinated with five water molecules and one hydroxyl group. And it is impossible to calculate the pH value at this state due to very limited atoms of the AIMD systems. In addition, the Ca 3 SiO 5 hydration is conducted at pH = 7, and the change of pH value of the Ca 3 SiO 5 hydration is the consequence of the surface reaction, not by the human intervention. The dissolution of the Ca 2+ will make the H + adsorbed on the surface and leave the OHin the solution, thus the pH value will increase from 7 to ~ 12.5 gradually (Kumar, Synthetic calcium silicate hydrates, EPFL, 2017) and our simulations are consistent with this phenomena. However, the reviewer's suggestion inspires us to continue simulate the dissolution with different H + or OHaround the targeted Ca ion to study whether the free energy or mechanism will change with this variable in the future.  Figure 1 (b), they were dissociated from the isolated water molecule through the DFT-based geometry optimization calculations.
As written in the Methods section, we firstly adsorbed isolated water molecules to saturate the dangling bond on the Ca 3 SiO 5 surface considering that the adsorption of water molecule on the Ca 3 SiO 5 surface occurs even before contacting bulk water. During AIMD, the bond formation and breakage were dynamic, which means the initial state we set will change during simulations. Even if we get stable hydroxyl groups on the surface initially through the DFT-based geometry optimization, they will still be possible to dissociate from the surface and reformed on the surface.

Comment 3:
The authors wrote "The breakage of the Ca-O s bond is earlier, but more difficult than with the interstitial oxygen ion (O i )" (l. 136 p.7). I think the terms "earlier" and "more difficult" should be corrected and/or better explained.
Response: Thanks for your comments. We revised the corresponding sentence as follows: 3 The thermodynamic and kinetic analyses show that the detachment of Ca α is spontaneous, while the detachment of Ca β is unspontaneous. (lines 264-265, page 17) Comment 5: The terms "parallel configuration" (l. 201, p.13) and "parallel to the surface" (l 206,p.13 and l.229,p.14) for water molecules could be clearer saying "oriented with their dipole moment perpendicular to the surface". Otherwise, "dissolved Ca ion tunes the direction of the chemisorbed water to a flatter configuration" means that Ca dissolution lead to water molecules oriented with their dipole perpendicular to the surface? The explanation of the regions observed in the atomic density profile was improved. However the authors should explain what is "the destruction area of the layered water structure"? (l. 211, p. 13). In addition, maybe the term "perfect Ca 3 SiO 5 " should be explained as a surface which has not experienced Ca dissolution?
Response: Thanks for your comments. We revised the inappropriate expressions as follows: Additionally This result is different from our previous work on the Ca 3 SiO 5 surface without Ca dissolution.

(lines 201-202, page 13)
All the above three regions are similar to those on the Ca 3 SiO 5 surface without Ca dissolution not only in width but also in intensity. (lines 209-211, page 13) Comment 6: For the sake of clarity, it should be good to explain the notation of the states (e.g. X(CN(Ca-O s ), CN(Ca-O w )) at the beginning of the "Results" section.
Response: Thanks for your comments. We explained the notation of the states at the beginning of the "Results" section as follow: The coordinate of the state is present in form of X(CN (Ca-O  "a less extent of structuration oscillating at 0" (l. 212 p. 13) 6 "surface indexes" (l.245, p. 16) → "Miller indices" "is most likely" (l. 178, p. 11) → "is the most likely" "is most stable" (l.186, p.11) → "is the most stable" "Because the coordination environments of Ca ions may change the dissolution pathways as well as the thermodynamic and kinetic properties. Thus..." → "The coordination environments of Ca ions may change the dissolution pathways as well as the thermodynamic and kinetic properties. Thus…" (l. 144, p. 11) "can be interpreted as a process that" (l. 226, p. 14) → "can be interpreted as a process where" (l. 226, p. 14) "to further investigation" (l. 333, p. 20) → "for further investigation"

Jérôme Claverie
Response: Thanks for your kind and careful corrections on our manuscript. We revised all the errors and inappropriate expressions as suggested by the reviewer.
Reviewer #3: The authors has satisfied all my concerns.