Experiments on Distribution of Divalent Metal Cations between Olivine Solid Solutions and Aqueous Chloride Solutions at 700 °C and 100 MPa
_Migdisov.png)
Round 1
Reviewer 1 Report
The authors addressed the concerns that I had in my previous review. I believe that the revised paper can be accepted for publication without any additional revisions or modifications.
Author Response
To the Reviewer 1
Thank you for your positive final decision to our manuscript. We appreciate the constructive comments and suggestions to improve the quality of our manuscript.
Reviewer 2 Report
I am surprised that the authors did not include a response to the concerns raised in earlier reviews, but this revised version does address a number of these, including my main concern regarding mineral-fluid equilibration in the experimental method. BSE images of sectioned run product olivine would have been preferable, but the systematic relationships presented in Uchida et al. 1997 provide confidence that that the experimental method used here did produce the sought results.
The one section of the paper that needs some work is the introduction. It is too brief in its presentation of partitioning and its controls. The authors mention the Matsui et al. 1977 study, which is one of the very early studies into element partitioning, but this field has grown enormously since, and there should be a summary of more recent insights regarding element partitioning (e.g. papers by Jon Blundy, Bernard Wood, Shun-ichiro Karato, Wim van Westrenen, Stephan Klemme, Alex Corgne etc), including mineral-fluid element partitioning (e.g. papers by James Brenan, John Ayers, Andreas Audetat, etc.). The link to skarns is also not presented in sufficient detail. Skarns do indeed represent an appropriate natural setting for the experiments conducted here, but more information needs to be presented to the reader. Also, the implications of the research findings for skarns should then be discussed at the end of the paper.
Author Response
Responses to the comments from the Reviewer 2
Point 1:
The one section of the paper that needs some work is the introduction. It is too brief in its presentation of partitioning and its controls. The authors mention the Matsui et al. 1977 study, which is one of the very early studies into element partitioning, but this field has grown enormously since, and there should be a summary of more recent insights regarding element partitioning (e.g. papers by Jon Blundy, Bernard Wood, Shun-ichiro Karato, Wim van Westrenen, Stephan Klemme, Alex Corgne etc), including mineral-fluid element partitioning (e.g. papers by James Brenan, John Ayers, Andreas Audetat, etc.).
Response:
In "1. Introduction", we newly cited papers of Goldschmidt (1937), Brice (1975), Blundy and Wood (1994, 2003), Wood and Blundy (1997) and Karato (2016) to summarize a history of research on element partitioning between minerals and melts. In addition to them, a paper of Brenan and Watson (1991) was cited in the manuscript as an example of experiments on element partitioning between olivine and aqueous fluids.
Point 2:
The link to skarns is also not presented in sufficient detail. Skarns do indeed represent an appropriate natural setting for the experiments conducted here, but more information needs to be presented to the reader. Also, the implications of the research findings for skarns should then be discussed at the end of the paper.
Response:
The purpose of this experiment is to clarify the partitioning behavior of divalent metal ions between olivine and aqueous chloride solution. It is not intended to clarify the formation conditions of olivine in magnesian skarn formed by the reaction of hydrothermal fluid with dolomite or magnesian limestone. In fact, we have never studied such olivine formed in magnesian skarn.
In addition, even if the chemical composition of olivine is known, it is necessary to know the type and concentration of anions in the hydrothermal fluid as suggested by Brenan and Watson (1991) in order to know the chemical composition of the hydrothermal fluid involved in its formation. This is by no means an easy task.
This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.
Round 1
Reviewer 1 Report
The manuscript presents an interesting study devoted to partitioning of Ni, Mg, Co, Zn, Fe, and Mn between the olivine solid solutions and aqueous phase at 700 C and 100 MPa. I have a few comments, which I would recommend to address in the revised version of this manuscript:
- Table 1 - I would recommend to list the concentrations of metals in aqueous phase explicitly. It is difficult to figure out the real concentrations when it is given as just mix proportions.
- Lines 69-71 "The compositions of the aqueous chloride solutions were adjusted so that the chemical composition analyses of the olivine solid solutions and aqueous chloride solutions after the reaction could be measured as accurately as possible." - I do not understand what this means. I think the authors should elaborate.
- Part 4 - Discussions. I would recommend to explicitly list the formation constants used in the calculations. This will make easier reading of the manuscript and will improve its citation rate.
- Finally - the authors end at derivation of the partitioning coefficients, which are not thermodynamic parameters. I am curious why they did not try to evaluate the mixing parameters of the solid solutions. The obtained partitioning coefficients permit to derive chemical potentials of end members of Ni, Co, Mn, etc in solid solutions, and some conclusions on the binary interaction parameters can be drawn. I am not insisting on doing this, but this step looks logical to me.
Author Response
Responses to the comments from Reviewer 1
Thank you for your insightful comments and suggestions to our manuscript. We revised the manuscript taking them into consideration. Responses to your comments and suggestions are as follows:
Point 1: Table 1 - I would recommend to list the concentrations of metals in aqueous phase explicitly. It is difficult to figure out the real concentrations when it is given as just mix proportions.
Response: Table 2 shows the total concentration of each element and the concentration of the neutral dissolved species. This table reveals that the neutral dissolved species are the predominant dissolved species under experimental conditions. This suggests that in the first approximation, all elements are present as almost neutral aqueous species.
Point 2: Lines 69-71 "The compositions of the aqueous chloride solutions were adjusted so that the chemical composition analyses of the olivine solid solutions and aqueous chloride solutions after the reaction could be measured as accurately as possible." - I do not understand what this means. I think the authors should elaborate.
Response: The sentence was revised as follows; “The composition of the aqueous chloride solution was adjusted so that the concentration of a certain element did not become extremely low in the chemical composition analysis of the olivine solid solution and the aqueous chloride solution after the reaction.”
Point 3: Part 4 - Discussions. I would recommend to explicitly list the formation constants used in the calculations. This will make easier reading of the manuscript and will improve its citation rate.
Response: The dissociation constants of equations (3) to (5) are newly shown in Table 3.
Point 4: Finally - the authors end at derivation of the partitioning coefficients, which are not thermodynamic parameters. I am curious why they did not try to evaluate the mixing parameters of the solid solutions. The obtained partitioning coefficients permit to derive chemical potentials of end members of Ni, Co, Mn, etc in solid solutions, and some conclusions on the binary interaction parameters can be drawn. I am not insisting on doing this, but this step looks logical to me.
Response: To obtain mixing parameters of the solid solution, it is necessary to conduct experiments in binary or ternary system. This is not purpose of this research. However, we already conducted ion exchange experiments for Fe-Mn-Mg olivine solid solution to obtain mixing parameters (Uchida et al., 1997, Mixing properties of Fe-Mn-Mg olivine solid solution determined experimentally by ion exchange method. J. Min. Petr. Econ. Geol 92, 142-153).
Reviewer 2 Report
Please see my comments on the annotated manuscript.
My main comments are that the methodology is not always clear and needs to be more fully described.In particular how the olivine solid solution starting materials were synthesized and how the concentrations of the metals at run conditions were back-calculated from the fluids extracted by rinsing the capsules at the end of the experiment. Secondly, it has not been shown that the experiments represent equilibrium, and this needs to be discussed. A set of experiments at increasing duration would be a good way to show this. Finally, information on the accuracy and precision of results needs to be added.
Comments for author File: 
Comments.pdf
Author Response
Responses to the comments from Reviewer 2
Thank you for your insightful comments and suggestions to our manuscript. We revised the manuscript taking them into consideration. Responses to your comments and suggestions are as follows:
Point 1: How the olivine solid solution starting materials were synthesized.
Response: Experiments were conducted by reacting a mixture of chemicals such as MgO, Fe, and SiO2 corresponding to the composition of the olivine solid solution with an aqueous chloride solution.
Point 2: How the concentrations of the metals at run conditions were back-calculated from the fluids extracted by rinsing the capsules at the end of the experiment.
Response: Since the total amount of metal ions in the reaction solution is 1 mol/L, the concentration of metal ions in the solution after the experiment was calculated using the metal ion ratios obtained by ICP analysis.
Point 3: It has not been shown that the experiments represent equilibrium, and this needs to be discussed.
Response: An equilibrium between the mineral surface and the aqueous chloride solution was assumed in this experiment.
Point 4: Information on the accuracy and precision of results needs to be added.
Response: The accuracy and precision of results were added in the revised manuscript.
Point 5: Provide some details on the experimental methodology: these are re-equilibration experiments in doped solutions.
Response: These are not re-equilibration experiments. An aqueous chloride solution was reacted with a mixture of chemicals with a chemical composition the same as that of olivine.
Point 6: (2.1.) This section is confusing. It reads to me as you conducted a separate synthesis of these starting olivines before the exchange experiments. If so, much more information needs to be provided on this step. How were these synthesized?
Response: As described in the manuscript, a mixture of MgO, Fe, and SiO2 was used as the starting materialｓ of olivine solid solutions.
Point 7: (2.1.) At what P and T, how was their composition verified? Were these homogenous? What was the size of the grains and were they idiomorphic?
Response: These were described in the manuscript.
Point 8: (2.2.) This is not clear. What needed to be optimized and why?
Response: We revised the sentence as follows: The composition of the aqueous chloride solutions was adjusted so that the concentration of certain elements does not become extremely low in the chemical composition analysis of the olivine solid solution and the aqueous chloride solution after experiments.
Point 9: (2.3.) Please explain the reasoning for adding the anthracene. Was it added as a liquid?
Response: The anthracene is a reducing agent to keep iron as Fe2+. The solid anthracene was put into a gold capsule.
Point 10: (2.3.) How was the sample pressurized and heated? Isobarically heated, or heated and pressurized at the same time?
Response: The sample was isobarically heated.
Point 11: (2.3.) What is the variability in P and T over the run
Response: Temperature variation was within ± 3℃, and pressure variation was within ± 1 MPa.
Point 12: (2.3.) Was the pressure controlled during this quench or was it allowed to drop as the T dropped?
Response: The pressure dropped as the temperature dropped.
Point 13: (2.3) How was the amount of fluid collected determined? Was a fixed amount of water used?
Response: A fixed amount of water (50cm3) was used?
Point 14: (2.3.) Why was Si not measured?
Response: Because it was considered that the content of Si in the aqueous solution is very low.
Point 15: (2.3.) Did you only analyse the surface or also a section to see if there was any zoning?
Response: We analyzed only the surface of synthesized olivine because we assume an equilibrium between mineral surface and aqueous chloride solution.
Point 16: (2.3.) Please provide the accuracy and precision of the solid and fluid analyses as determined from duplicates and certified reference materials.
Response: We described the analysis accuracy and reference materials in the revised manuscript.
Point 17: (2.3.) The low standard concentration suggests extensive dilution of the solutions. What were these diluted with? Was a blank dilutant measured as well?
Response: The concentration of cations ranges widely. A blank diluent (distilled water) was also used as the standard solution.
Point 18: The spectra for c and d have a significant XRD-amorphous hump. Although this may well be a grainsize issue, if an amorphous component were to be present, this would affect partitioning significantly, so this should be discussed. Also, there is one unexplained peak at ~28Ëš.
Response: The amorphous pattern is due to the glass plate to which the sample was attached.
Point 19: (3.2.) How were these molarities and ratios calculated from the experimental run product solutions? You mention that you extracted the fluid by rinsing the capsule with water, which would dilute the solution, so how did you determine the concentration at run conditions?
Response: The ratios of each divalent metal cation in the aqueous chloride solutions after the experiment were calculated based on the results of cation analysis by the ICP-AES.
Point 20: (4.1) Before interpreting the data, you should first establish that your experiments have reached equilibrium. A time-series should be conducted to determine whether 5 days was a long enough run duration to reach equilibrium. The run product solids should also be sectioned to make sure that there is no compositional zoning and that the crystals are homogenous. The good correspondence among experiments with different fluids is an indication that equilibrium was likely reached, but is not proof by itself.
Response: Based on the shape and size of the synthesized crystals, we believe that a reaction period of 5 days at 700 ℃ is sufficient for the reaction to reach equilibrium. We already conducted ion exchange experiments for Fe-Mn-Mg olivine solid solution at 600 ℃ and 100 MPa during 5 to 11 days using stating materials with different compositions (Uchida et al., 1997, Mixing properties of Fe-Mn-Mg olivine solid solution determined experimentally by ion exchange method. J. Min. Petr. Econ. Geol 92, 142-153), but we obtained consistent results.
Point 21: (4.1) This is not correct. The Me2SiO4 and Mg2SiO4 mole fractions are to the power 1/2 to be consistent with the exchange reaction in (1).
Response: The equation (2) was corrected as follows:
Point 22: (4.1.) The authors should mention that these speciation calculations are necessary to correct for any differences in speciation among the elements, as not all readers will be familiar with this.
Response: We believe that many readers understand that dissolved species must be considered when handling aqueous chloride solutions.
Point 23: (4.2) Why was a cubic equation chosen do describe the relationship? The Brice equation would likely be a more appropriate equation to fit these data.
Response: Generally, it is approximated using a quadratic curve, so we reconstructed the figure in that way.
Point 24: (4.2) Interesting. This indicates that radius is not the only parameter of importance in determining the partitioning behaviour, because Zn would be expected to enter both the tetrahedral and the octahedral sites if radius alone was the controlling property and therefore plot above the KD curve, rather than below. Instead, you find that the Zn content of the olivine is unexpectedly low, so Zn does not enter the octahedral site despite the fact that the radius would suggest that it could. In other words, what makes the Zn prefer the tetrahedral site? This is likely due to its electronic structure.
Response: Since Zn2+ easily forms a four-fold coordinated bond from the ligand field theory (due to its electronic structure), it is considered to be difficult for Zn2+ to enter sites other than the four-fold coordinated site.
Reviewer 3 Report
My comments to the authors are provided in the attached pdf document.
Comments for author File: Comments.pdf
Author Response
Responses to the comments from Reviewer 3
Thank you for your insightful comments and suggestions to our manuscript. We revised our manuscript taking them into consideration. Responses to your comments and suggestions are as follows:
General Comments:
Point 1: The introduction is missing a large amount of information and references. Information on why olivine is important in skarns is needed, the fact that it is a major-rock forming mineral in igneous rocks is not sufficient. There are no references to the occurrence of olivine in skarns. There are also no references to the involvement of granites in the formation of skarns. The authors need to provide references for all of this information as it is not new.
Response: Olivine is one of major rock-forming minerals in igneous rocks. The next major occurrence of olivine is skarns, which were produced by hydrothermal reaction with dolomite or magnesite. The interaction between hydrothermal solution and olivine is important to estimate the chemical composition of hydrothermal solutions involved in skarn formation. In addition, these experiments are important for investigating the partition behavior of elements between minerals and hydrothermal solutions.
We added one representative reference related to skarn that produces olivine.
Point 2: There is virtually no discussion on the importance of these data. There is no application of the equilibrium constants or discussion of how they might be useful for the community. There is also very little justification for performing these experiments and the choice of conditions. Why is olivine in skarns important? What is the justification for the PT conditions? How can these equilibrium constants be used? There is also no comparison to previous studies on olivine-fluid partitioning, most of which are performed at higher pressures and temperatures.
Response: This experiment was conducted to clarify the factors that govern the partitioning of elements between olivine and hydrothermal solution. A relatively high temperature of 700℃ was selected so that the experiment could easily reach equilibrium. The pressure of 100MPa was selected from the commonly considered skarn-forming pressures.
Line by line comments:
Line 11-12: There needs to be some clarification on the use of the term partitioning. Is this partitioning into the fluid phase relative to the olivine (fluid/crystal) or vice-versa (crystal/fluid)
Response: The sentence was changed as follows: “the tendency of divalent cation partitioning into the olivine solid solutions against the aqueous chloride solutions”.
Lines 15-17: I suggest breaking this into multiple sentences. It is difficult to read as is.
Response: Line breaks are not allowed in “abstract” in “Minerals”.
Lines 17-19: What does “smaller differences in partition coefficients” mean? Is this refereeing to the difference among the elements?
Response: The sentence was changed as follows: “Mg-Fe olivine solid solutions with more fayalite end-members exhibited smaller differences in partition coefficients for the cations.”
Lines 19-21: This sentence is essentially saying that iron is more likely to be involved in a
substitution for other cations compared to magnesium because it has larger ionic ratio, correct?
Response: This sentence was changed as follows: “This trend indicates that fayalite consisting of a large cation (Fe2+) tends to exchange other cations more easily than forsterite consisting of a small cation (Mg2+).”.
Line 26 - 27: Partitioning does not take place between minerals and a groundmass or magma. Partitioning takes place between two phases (mineral/melt, mineral/mineral, fluid/melt etc). The groundmass or a magma is not a phase and therefore partition cannot take place between these things. Matsui simply used the groundmass as an approximation for the melt.
Response: As you suggested, Matsui et al. (1977) simply used the groundmass as an approximation for the melt. However, we don't think this assumption is a big mistake.
Line 34: Does “M/L” refer to mol/L or molarity? M/L is not a proper abbreviation.
Response: M/L was replaced with mol/L.
Lines 41-47: This text is unnecessary. Why is there a discussion on sulfide minerals and the type of bonding? If this is important it needs to be introduced in detail and referenced. Furthermore, there is no transition in the text from one topic to another, there are abrupt changes in topic and the start of the section is also abrupt.
Response: We believe that these texts are necessary to compare the partitioning behavior of minerals with covalent bonds and minerals with ionic bonds such as olivine.
Line 58: The sentence should not start with an abbreviation.
Response: “Fe” was replaced with “Iron (Fe)”.
Line 56-58: I suggest rearranging this sentence to state that “In order to crystallize olivine of varying compositions a mixture of… was used as starting materials”
Response: The sentence was revised as follows: “In order to synthesize olivine of varying compositions, a mixture of MgO (99.99%, Rare Metallic Co., Ltd., Tokyo, Japan), Fe (99.5%, Junsei Chemical Co., Ltd., Tokyo, Japan) and SiO2 (99.99%, Rare Metallic Co., Ltd., Tokyo, Japan) was used as starting materials.”
Line 65: The beginning of the sentence should read “The aqueous solution used as starting
materials…”
Response: The sentence was revised as follows: “The aqueous solutions used as starting materials are mixtures of 1 mol/L NiCl2, MgCl2, CoCl2, ZnCl2, FeCl2, and MnCl2 aqueous solutions.”
Line 68-69: I suggest changing these sentences to read “The composition of the starting solutions is given in Table 1”. But table one also lists the experimental temperatures and weights of starting materials.
Response: The sentence was revised as you suggested.
Line 69 – 70: I do not understand this sentence. Why do the compositions need to be changed so the analyses are more accurate?
Response: The sentence was revised as follows: “The composition of the aqueous chloride solutions was adjusted so that the concentration of certain elements does not become extremely low in the chemical composition analysis of the olivine solid solution and the aqueous chloride solution after the reaction.”
Section 2.3: I suggest combining section 2.1 and 2.2 with section 2.3. It is difficult to follow by mentioning the starting materials first.
Response: The original section 2.3 was moved to the new section 2.1. The original sections 2.1 and 2.2 was moved to the new sections 2.2 and 2.3, respectively.
Line 76-79: This sentence needs to be rewritten. It should be structured so that it read the “The starting materials were contained in gold capsules that were welded shut” with any additional detail included. Additionally, why was anthracene added to the experiments? This is the only mention of this and an explanation is not provided.
Response: The sentence was revised as you suggested. Anthracene was added to keep iron as Fe2+. This explanation was added in the revised manuscript.
Line 82: Stellite should be capitalized
Response: “stellite was replaced with “Stellite”.
Line 81: I believe here and possibly throughout the document that when referring to the capsules or cold-seal vessels the nouns and verbs should be plural (i.e. capsules, vessels, were).
Response: “capsule” and “vessel” were changed to “capsules” and “vessels”.
Line 85: Remove “to raise the temperature”
Response: “to raise the temperature” was removed.
Line 92: Why were 700°C and 100 MPa chosen?
Response: The experiments were conducted at 700 °C and 100 MPa in consideration of the reaction rate and the pressure resistance of the reaction vessels.
Line 97-98: Remove the comma before “and” and after “Then”
Response: We removed the comma before “and” and after “Then”.
Line 105: I suggest “A portion of solid run product was placed on a carbon tape on a glass slide, carbon coated (Quick Carbon Coater SC-701C, Sanyu Denshi Co. Ltd., Tokyo, Japan), and then analyzed by using a scanning electron microscope (JSM-6360, JEOL, Tokyo, Japan). A chemical composition was obtained by using energy-dispersive X-ray analysis (INCA ENERGY, Oxford Instruments, Abingdon, UK).”
Response: We revised the manuscript as you suggested.
Line 111-116: What were the standards for the EDS analyses? Standardless EDS analyses can be very inaccurate. If there are no standards involved in the EDS analyses how do the authors know the concentrations obtained are accurate? At best I would suspect that the ratios of elements would be ok, but the absolute concentrations of the elements will have a large uncertainty without appropriate standards. Perhaps this is ok given only ratios are used, but it still needs to be justified. If these are EDS analyses with standards that standard should be listed and this would support the quality of the analyses.
Response: NiO, MgO, Co, ZnO, Fe2O3, MnO, and SiO2 were used as standards of Ni, Mg, Co, Zn, Fe, Mn and Si analyses, respectively. This was added in the revised manuscript.
Line 117-123: This section brings up a couple of questions about the fluid phase. It is reported (Lines 94-100) that fluid was separated from the solids by filtration, but the fluid recovery process is not clearly described. How much water was used to rinse the capsule? How do the authors know what the weight/volume of the fluid is after the experiment and how much of the fluid was recovered? If there is any exchange between the solid and fluid (presumably there is) the initial mass of the fluid is not the final mass. Further, how much was the solution diluted during the recovery process? This must have been considered in order to calculated the concentration by ICP-MS, but it is not reported.
Response: The filtrated aqueous chloride solution was adjusted by adding distilled water so that the volume was 50 cm3. This sentence was added in the revised manuscript.
Line 126: It would be helpful for the reader to include a reference diffraction pattern for olivine overlain on the experimental ones.
Response: All the detected peaks are due to olivine, and no other peaks have been detected.
Line 129-131: The authors observation that the olivine was pale pink is very surprising, are the authors sure the olive itself is pink or is there some other explanation for a pink powder, perhaps iron oxide?
Response: The pink color may be due to Mn2+ contained in olivine.
Table 2 fluid compositions: For the footnotes listed as 2 and 3 the authors state these concentrations are molality but then state the units for these concentrations are mol/L. Molality (molal concentration) is moles per kilogram and molarity (molar concentration) is moles per liter. The authors need to be clear on the units they are referring to.
Response: We changed molality to molarity.
In the text in section 3.2 the text refers to the liquid compositions as mol%, but this can be confusing. Referring to them as mol% suggests that for example, Mg is 47 mol% of the entire fluid when it is actually 47 mol% of the salts in water, when in actuality the majority of the fluid comprises H2O and Mg is a much lower fraction. The two different liquid compositions listed are very confusing when initially read. It does not become clear what the second compositions means until reading further into the manuscript to
Response: The manuscript was revised to suggest that the proportion of Mg2+ against all divalent metal cations in the aqueous chloride solution is 47 mol%.
Lines 165 to 180. Lines 165 to 180 must come early in the manuscript to avoid confusion. Additionally, the table should indicate the difference in these compositions somehow, perhaps by stating the second compositions are calculated by using equations 3 through 11.
Response: We don't think that it will cause confusion, so leave it as it is.
Table 2: There is no explanation for log KPN in the table and the abbreviation from equation 2 has changed.
Response: P and N are initials for “Partitioning” and “Neutral”, respectively, but we don't think that it is necessary to include this explanation in the manuscript.
Section 4.1: I think equation 1 is written incorrectly for the calculation the authors are trying to perform. As the equation is written now the correct equilibrium constant equation would be ・・・・・・・・・・・
Response: Your suggestion is correct. Therefore, the chemical composition of olivine was set to MeSi0.5O2.
Section 4.2: The use of the term partition coefficient is incorrect. The type of calculation performed by the authors is not a simple partition coefficient, but is the calculation of an equilibrium constant. It is also difficult to comment on this section given the equilibrium constants are incorrect, but I am assuming the general trends will hold. Further, there is no “partitioning “between essential constituents (Mg) of a mineral with any other phase, because that essential constituent is fixed in the given mineral. The mineral itself acts as a buffer, assuming an appropriate assemblage, in the system and describing this as partitioning would be incorrect. The authors are correct in considering an equilibria and an exchange, but need to properly discuss it as so.
Response: Equation (2) is the partition coefficient, not the equilibrium constant. The equilibrium constant is expressed using activity.
Line 195-199: The same comment as noted in the abstract regarding these sentences. Also is there a figure to show this relationship. It is difficult to interpret this sentence without a figure.
Response: We added values so as to make it easier for the reader to understand as follows: “1.18 for Fo100Fa0, 1.00 for Fo75Fa25, 0.92 for Fo50Fa50, 0.82 for Fo25Fa75, and 0.73 for Fo0Fa100 in logarithm.”
Line 207: “According to the results of…”
Response: The manuscript was revised as you suggested.
Line 208: “Thus nickel, which has the optimum ionic radius, exchanges with magnesium more readily than other cations.”
Response: The manuscript was revised as follows: “Thus Ni2+, which has the optimum ionic radius, exchanges readily with other cations.”
Line 215-217: This sentence is irrelevant.
Response: This sentence was deleted.
Line 221: “2 to 2.5”
Response: We revised the manuscript as you suggested.
Line 221-228: This section abruptly changes topics to sulfide. I don’t understand why this is done, there are much better examples of Zn being in four-fold coordination in silicate minerals. For example, willemite is a zinc endmember of the olivine solid solution.
Response: Unfortunately, we did not conducted experiments using willemite. We would like to conduct experiments in the future.
Round 2
Reviewer 2 Report
This revised manuscript and the accompanying response letter address many of the issues raised in the original review, and clarify the experimental approach used in this study. I still have a number of concerns regarding the experimental method, in particular related to characterisation of the run product olivine and attainment of equilibrium. These are outlined in more detail below.
Based on the changes made to the manuscript, and the responses from the authors, it is clear that I had misunderstood the type of experiment conducted. Rather than a re-equilibration experiment between an earlier prepared olivine and a fluid doped with the element(s) of interest, the authors synthesized the olivine at the same time as conducting the partitioning experiment. Although this approach has the benefit of avoiding the slow re-equilibration in silicates, it does suffer from fractional crystallisation changes in mineral and fluid compositions as the experiment progresses. As a result, only the outermost surface of the mineral will be in equilibrium with the run product fluid (see for example Marschall et al. 2009 Contrib Mineral Petrol 158:675-681). The approach of the authors to analyse only this outermost rim is thus the correct approach. However, the assumption that the cations in the fluid will sum to 1 mol/L is not warranted in this experimental approach, as there is no guarantee that the olivine was stoichiometric throughout its growth. Moreover, it is common to find unreacted starting material in the cores of mineral grains synthesized in this fashion, which may well explain the amorphous humps in the XRD patterns. Such unreacted starting materials will not be present in stoichiometric amounts and will thus offset the metal concentrations in the fluid. In summary, assuming that the fluid metal concentrations sum to 1 mol/L is questionable, and needs to be proven.
Synthesis experiments are commonly subject to initial rapid and disequilibrium growth as the starting materials are in strong mutual disequilibrium, which is in fact commonly on purpose to promote reaction (see discussion in Fyfe 1960). It can therefore not be assumed that the results represent equilibrium and this needs to be proven. The authors can achieve this relatively easily by running experiments for different durations, or, even better, to run experiments at different dopant concentration levels. If these experiments give the same distribution coefficients for Ni, Zn, etc, equilibrium has been established.
The authors need to present compositional cross-sections of their synthesized olivine grains to allow the reader to assess their homogeneity, the presence of unreacted starting materials and any potential impact on the metal balance in the fluid. The partition coefficients (D olivine/fluid) vary from ca. 5 for Ni to 0.03 for Mn, in other words highly compatible to highly incompatible. It can thus be expected that the cores of the olivine grains are highly enriched in Ni, but depleted in Mn compared to the rims. Similarly, the experiments show that Mg is more compatible than Fe in the synthesized olivines. This strongly suggests that the first formed olivine will be Fo-enriched, and will therefore display a partitioning behaviour that is different from that of the rim composition, which will be more Fa-rich owing to fractional crystallisation. The presence of such fractional crystallization zoning does not invalidate the approach or the conclusions, but its impact needs to be investigated and presented to the reader. In fact, any trends observed in trace metal contents with compositional zoning provide a verification of the main conclusions, and if indeed the same, strengthen these conclusions.
The authors added information on accuracy, but it appears from the revised manuscript that they determined this from repeat analyses, in which case it would be the precision, not the accuracy. In the response it is stated that reference materials are mentioned, but I could not find this in the text.
Reviewer 3 Report
Attached to this review is a file with comments for the authors regarding the revisions. In the document you will find some of the original comments and the authors responses to those comments, as well as additional comment about the revisions. New comments on the revisions are given in red.
Comments for author File: Comments.pdf
