Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript contains interesting ideas and appears potentially publishable, but several aspects require clarification and improvement before further consideration. Main comments and questions:

1. In the Introduction (around lines 61–65), the manuscript attributes a destructive β→γ phase transformation to wollastonite (CaSiO₃). This transformation is more commonly associated with β→γ C₂S. Could the authors please clarify or correct this statement and provide an appropriate reference?
2. The manuscript states that flotation increases the CaO/SiO₂ ratio from 0.68 to 2.85. Could the authors clarify whether these values refer to raw ore vs flotation concentrate, or to fused magnesia products (FM-1 vs FFM-3)? Please specify CaO/SiO₂ ratios for the raw ore and the concentrate.
3. The reported bulk densities of FM-1 and FFM-3 differ by only 0.01 g/cm³. How many specimens were measured for each sample, and what are the mean values with standard deviations? Is this difference statistically significant?
4. In Fig. 7, C₂S is identified, but the polymorphic form is not specified. Were β- and/or γ-C₂S phases distinguished after cooling? If so, how was this confirmed (XRD peak assignment, other evidence)?
5. Microcracks are attributed to the β→γ transformation of C₂S. Could the authors provide quantitative crack analysis for fused samples (e.g., crack density or area fraction), similar to the analysis shown for raw ores in Fig. 6?
6. Flotation is described as an industrial process, but key operating parameters are missing. Could the authors provide reagent dosages, conditioning time, pulp density, flotation time, and grinding details to improve reproducibility?
7. Were flotation reagents removed or washed from the concentrate prior to fusion? Could residual reagents influence grain boundary phases or densification?
8. Why was fused magnesia not prepared from the unfloated low-grade ore (DSQLM-3) as a baseline for comparison? Please clarify the rationale.
9. Table 4 summarizes phase formation as a function of CaO/SiO₂ ratio. Does this assume equilibrium at a specific temperature and impurity level? How sensitive are these phase relations to Al₂O₃ and Fe₂O₃ impurities?

Author Response

Comment 1:In the Introduction (around lines 61–65), the manuscript attributes a destructive β→γ phase transformation to wollastonite (CaSiO₃). This transformation is more commonly associated with β→γ C₂S. Could the authors please clarify or correct this statement and provide an appropriate reference?

Response1: Thank you for pointing this out. We confirm that this was a typographical error. The destructive polymorphic transformation discussed here should refer to β-dicalcium silicate (β-C₂S, Ca₂SiO₄) transforming to γ-C₂S during cooling, rather than wollastonite (CaSiO₃). We have corrected the statement accordingly and added appropriate references. The β→γ transformation of C₂S is well documented and is associated with a significant unit-cell/volume increase (≈10–12%), which can lead to cracking/pulverization (“dusting”).

We thank the reviewer for pointing out this issue. We have reviewed the manuscript and found that the attribution of the β→γ phase transformation to wollastonite (CaSiO₃) was indeed an error. This transformation is more accurately associated with β-C₂S (dicalcium silicate). We have corrected this statement in the manuscript and revised the relevant section of the Introduction to reflect this clarification. The corrected statement now reads: "SiO₂ tends to form low melting point silicate phases, and concurrently, CaO reacts with SiO₂ to form β-C₂S (dicalcium silicate), which undergoes a destructive β→γ phase transformation upon cooling, resulting in volumetric expansion and disintegration of refractory products."We have made the revision on Pages 2–3, Lines 86–90. We also added an appropriate reference to support this correction:

[1] Chan, C.J.; Kriven, W.M.; Young, J.F. Physical stabilization of the β→γ transformation in dicalcium silicate. J. Am. Ceram. Soc. 1992, 75(6), 1621–1627. https://doi.org/10.1111/j.1151-2916.1992.tb04234.x

Comment 2:The manuscript states that flotation increases the CaO/SiO₂ ratio from 0.68 to 2.85. Could the authors clarify whether these values refer to raw ore vs flotation concentrate, or to fused magnesia products (FM-1 vs FFM-3)? Please specify CaO/SiO₂ ratios for the raw ore and the concentrate.

Response2: We thank the reviewer for this valuable observation. To clarify, the CaO/SiO₂ ratio of 0.68 to 2.85 refers to the fused magnesia products, specifically FM-1 and FFM-3, rather than the raw ore and flotation concentrate. We have revised the manuscript to clearly specify this and provide additional details on the CaO/SiO₂ ratios for both the raw ores and the flotation concentrate.

The CaO/SiO₂ ratios for the raw ores and the flotation concentrate are as follows:

DSQLM-1 (high-grade ore): CaO/SiO₂ ratio = 0.79

DSQLM-2 (medium-grade ore): CaO/SiO₂ ratio = 0.24

DSQLM-3 (low-grade ore): CaO/SiO₂ ratio = 0.15

F-DSQLM-3 (flotation concentrate): CaO/SiO₂ ratio = 3.57

The corrected text now includes these values and specifies that the CaO/SiO₂ ratio shift from 0.68 to 2.85 refers to the fused magnesia products (FM-1 and FFM-3). This information has been added to Table 1 for clarity.

We hope this resolves the reviewer’s concern, and we appreciate the suggestion for further clarification.

The changes have been made in Table 1 and the Results section (Page 5, Line 207).

Comment 3: The reported bulk densities of FM-1 and FFM-3 differ by only 0.01 g/cm³. How many specimens were measured for each sample, and what are the mean values with standard deviations? Is this difference statistically significant?

Response3: We have updated the manuscript with detailed information about the number of specimens measured, the mean values, standard deviations, and statistical analysis. Specifically, we measured the bulk densities of five specimens from each of the two samples (FM-1 and FFM-3). The updated information has been added to the Results section, and the relevant details have been incorporated into Table 3 to provide clarity on the measured values and the statistical significance of the difference in bulk density.

The Statistical Significance column confirms that the difference between FM-1 and FFM-3 is not statistically significant based on the t-test analysis at the 95% confidence level.This additional information has been included in Table 3 and mentioned in the Results section of the manuscript.

We believe these changes clarify the measurement procedure and the statistical treatment of the bulk density data.These modifications have been reflected in Table 3 and the updated Results section (Page 10, Line 357).

Comment 4: In Fig. 7, C₂S is identified, but the polymorphic form is not specified. Were β- and/or γ-C₂S phases distinguished after cooling? If so, how was this confirmed (XRD peak assignment, other evidence)?

Response4: We appreciate the reviewer’s valuable comment. In response, we carefully reviewed the X-ray diffraction (XRD) data and confirmed that γ-C₂S (gamma dicalcium silicate) is present in the sample, with a characteristic peak observed at 2θ = 34.2°. This peak corresponds to the γ-C₂S phase, confirming the polymorphic transformation from β-C₂S to γ-C₂S during cooling.We have updated Fig. 7 to reflect this distinction and added a discussion in the manuscript regarding the identification of both β- and γ-C₂S phases. Specifically, we have now clarified that the β-C₂S phase was observed initially, and upon cooling, it transformed into γ-C₂S, which is evidenced by the appearance of the peak at 34.2° in the XRD pattern.

The revised manuscript now includes this additional information and the relevant peak assignment, providing further insight into the polymorphic transformation of C₂S and its effects on the material's microstructure.

Changes made: The XRD peak assignment for γ-C₂S (2θ = 34.2°) has been added to Fig. 7, and the discussion about the β → γ phase transformation has been incorporated into the Results and Discussion sections of the manuscript (Page 12, Line 440).

We thank the reviewer for pointing out the need for further clarification regarding the polymorphic form of C₂S identified in Fig. 7. In response, we would like to confirm that both β-C₂S and γ-C₂S phases were indeed distinguished after cooling in the fused magnesia sample FFM-3.

Comment 5: Microcracks are attributed to the β→γ transformation of C₂S. Could the authors provide quantitative crack analysis for fused samples (e.g., crack density or area fraction), similar to the analysis shown for raw ores in Fig. 6?

Response5: We thank the reviewer for the valuable suggestion. In response, we have conducted a quantitative crack analysis for the fused magnesia sample (FFM-3), similar to the analysis shown for the raw ores in Fig. 6. Using ImageJ software, we measured the crack area fraction in FFM-3.

We have added Fig. 8 to the manuscript, which shows the crack area fraction in FFM-3. The analysis revealed that the crack area fraction in FFM-3 is 0.31%, indicating a significantly higher microcrack density compared to FM-1（0.07%）. This higher crack density is attributed to the β→γ transformation of C₂S, which induces internal stresses and promotes microcrack formation.

We hope this addition addresses the reviewer’s comment and enhances the clarity of the manuscript regarding the impact of microcracks on the densification and performance of fused magnesia samples.

Changes Made:Added Fig. 9, showing the quantitative analysis of crack area fraction in FFM-3.

Included a description of the crack area fraction in the Results section (Page 13, Line 449).

Comment 6: Flotation is described as an industrial process, but key operating parameters are missing. Could the authors provide reagent dosages, conditioning time, pulp density, flotation time, and grinding details to improve reproducibility?

Response6: Thank you for this important comment. We agree that the original description of the industrial flotation procedure was too brief for reproducibility. In the revised manuscript (Section 2.1), we have added the key operating parameters and grinding details used during the plant campaign.

Specifically, we now report that the DSQLM-3 ore was crushed and ground to a flotation feed of 100% passing 0.088 mm ; flotation was performed at pH 8.5 ± 0.2 and ambient temperature (≈20–25 °C) with a pulp density of ~25–30 wt.% solids. We clarified the reagent scheme and addition sequence: sodium silicate was used as a combined depressant/dispersant and added before the collector, followed by sodium oleate as the collector. To improve reproducibility while reflecting plant practice, reagent dosages are provided as operating windows: Na₂SiO₃ ≈ 0.2–0.8 kg·t⁻¹ and NaOl ≈ 0.2–0.4 kg·t⁻¹, adjusted within these ranges according to feed variability. We also added the main time parameters: conditioning was ~3 min after each reagent addition, and the flotation residence time was ~3 min under plant-set aeration/agitation conditions (equipment-fixed settings). The concentrate washing and drying procedures before fusion are also stated explicitly.

Revisions made: Section 2.1 Raw materials (Pages 3-4, Lines 135–167).

Comment 7: Were flotation reagents removed or washed from the concentrate prior to fusion? Could residual reagents influence grain boundary phases or densification?

Response7: Thank you for this important question. Yes. As described in Section 2.1, the flotation concentrate was thoroughly washed with clean water after flotation to remove residual reagents, and then dried at 110 °C for 24 h prior to the fusion experiments. This washing step was implemented specifically to minimize any potential carryover of reagents into the high-temperature process.

Regarding potential effects, we agree that residual organics or soluble salts could, in principle, affect high-temperature reactions or intergranular phases. However, in our case the concentrate underwent intensive water washing followed by subsequent calcination (900 °C) and arc melting (2800 °C), under which any trace organic collector would be expected to decompose/volatilize well before fusion. Moreover, the key microstructural differences observed (CMS in FM-1 vs. C₂S in FFM-3, and the associated microcracking) correlate directly with the measured oxide balance (C/S ratio) and the corresponding phase evolution, rather than with reagent-related species. Nevertheless, we added an explicit note in the manuscript acknowledging that reagent carryover cannot be completely excluded and that future work will further verify this aspect (e.g., by monitoring residual carbon/surfactant markers if needed).

Revisions made: A clarification on concentrate washing/drying and a brief note on potential reagent carryover were added/strengthened in Section 2.1, and Discussion as appropriate (Page 4, Lines 154–156).

Comment 8: Why was fused magnesia not prepared from the unfloated low-grade ore (DSQLM-3) as a baseline for comparison? Please clarify the rationale.

Response8: Thank you for your valuable comment. The reason fused magnesia was not prepared from the unfloated low-grade ore (DSQLM-3) as a baseline for comparison is primarily due to the high impurity content, particularly the elevated SiO₂ content, in the unfloated ore. After fusion, the high silica content would lead to the formation of low melting point phases, such as M2S (calcium magnesium silicate), as the predominant bonding phase. This results in very low purity of the fused magnesia, which is not suitable for high-performance applications.

In contrast, the flotation process significantly reduces the impurity levels, especially the silica content, resulting in a concentrate (FDSQLM-3) with MgO content that is close to that of high-grade ores. The purity of the flotation concentrate is much higher, making it a more appropriate material for comparison with high-grade ores. The flotation concentrate provides a more meaningful comparison for evaluating the effects of impurities and beneficiation on the properties of fused magnesia, as it more closely resembles the purity levels of higher-grade ores.

We hope this explanation clarifies the rationale for not using the unfloated low-grade ore (DSQLM-3) as a baseline in the study.

Comment 9:Table 4 summarizes phase formation as a function of CaO/SiO₂ ratio. Does this assume equilibrium at a specific temperature and impurity level? How sensitive are these phase relations to Al₂O₃ and Fe₂O₃ impurities?

Response9: Thank you for this thoughtful question. Table 4 is intended as a simplified thermodynamic guideline based on the MgO–CaO–SiO₂ ternary equilibrium tendency (with the listed “initial melting temperatures” referring to the corresponding primary Ca–silicate/Ca–Mg–silicate phases). It does not imply that the industrial arc-melting and subsequent natural cooling strictly reach equilibrium at a single fixed temperature; rather, it summarizes the expected dominant silicate assemblage as a function of the C/S ratio for low-impurity fused magnesia, which helps rationalize why a low C/S favors CMS-type phases whereas a high C/S shifts the tendency toward Ca-rich silicates (C₂S/C₃S).

Regarding the sensitivity to minor oxides, we agree that Al₂O₃ and Fe₂O₃ can modify liquid formation and phase relations (e.g., by forming aluminosilicate glassy phases, substituting into silicates, or affecting the liquidus/solidus). However, in the present study these oxides are low and comparable between FM-1 and FFM-3 (Al₂O₃ ≈ 0.18–0.20 wt.% and Fe₂O₃ ≈ 0.53–0.56 wt.% in Table 3). Therefore, their influence on the CMS vs. C₂S-dominated phase selection is expected to be secondary relative to the much larger change in the CaO/SiO₂ ratio. We have clarified this point in the revised manuscript and explicitly stated that Table 4 provides an equilibrium-based compositional tendency for the primary silicate phases, while the actual products are also subject to non-equilibrium cooling kinetics.

Revisions made: We added clarification around Table 4, Section 3.3 stating the equilibrium-assumption nature of the table and noting that Al₂O₃ and Fe₂O₃ are minor and similar in both samples, so the C/S ratio remains the dominant driver  (Page 10, Lines 345–348). 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

• Table 2 requires editing as some data are overlapping, preventing accurate reading.
• The flotation process description is incomplete, hindering reproducibility. The following are not specified: exact dosages of collector and depressant/dispersant, conditioning time, stirring speed, solid/liquid ratio, or temperature.
• The study attributes the microstructural changes primarily to the increased C/S ratio, but an experiment was not designed to independently control the C/S ratio. Only two different raw materials are compared, introducing multiple simultaneous variables (thermal history, impurity distribution, grain morphology, residual chemistry). This prevents isolating the true effect of the C/S ratio.
• The “natural” cooling from 2800 °C is not quantified (rate, atmosphere, thermal gradients). Since the β→γ-C₂S transformation is highly dependent on thermal kinetics, this omission is critical and compromises the interpretation of microcrack formation.
• Although 22 EDS points are reported, their representativeness is not justified, nor are selection criteria, repeatability, or standard deviation indicated. The analysis remains essentially qualitative, despite the quantitative conclusions drawn regarding Ca and Si enrichment in microcracks.
• The interpretation based on the MgO–CaO–SiO₂ ternary system assumes thermodynamic equilibrium, when the industrial melting and cooling process is clearly non-equilibrium. The possible formation of metastable phases and the discrepancy between equilibrium and actual kinetics are not discussed.
• Unlike the detailed gangue analysis in the raw materials, the following are not quantified: C₂S volume fraction vs. CMS, microcrack density and length, and the statistical distribution of grain size in FM-1 and FFM-3. Without this data, the claims regarding structural deterioration remain at a descriptive level.
• A possible degradation of mechanical strength and thermal shock is inferred, but no mechanical or thermomechanical properties are measured. This weakens the applied validity of the results and leaves the conclusions in the speculative realm.

Author Response

Comment 1: table 2 requires editing as some data are overlapping, preventing accurate reading.

Response1: Thank you for pointing this out. We agree that the original formatting of Table 2 caused overlapping text and hindered accurate reading. In the revised manuscript, we reformatted Table 2 to eliminate overlap by (i) adjusting the column widths/row spacing to provide sufficient space for long phase names and multi-element entries, (ii) standardizing and slightly reducing the font size for consistent readability across all cells, (iii) unifying the line spacing/paragraph spacing within table cells to prevent text compression, and (iv) setting the table text wrapping to “None” (Table Properties → Text wrapping: None) to avoid layout conflicts. These changes remove the overlapping entries and improve the overall clarity of Table 2.

Revisions made: Table 2 was reformatted accordingly (Page 9, Line 319).

 

Comment 2:The flotation process description is incomplete, hindering reproducibility. The following are not specified: exact dosages of collector and depressant/dispersant, conditioning time, stirring speed, solid/liquid ratio, or temperature.

Response2: Thank you for this important comment. We agree that the original description of the industrial reverse-flotation operation was too concise to be reproducible. In the revised manuscript, we expanded Section 2.1 to report the key technological parameters that control flotation outcomes, including (i) the flotation feed size (100% passing 0.088 mm), (ii) pulp density (solid/liquid ratio), (iii) reagent addition sequence and dosage ranges for sodium oleate and sodium silicate, (iv) conditioning and flotation times, and (v) operating temperature.

Because the flotation was conducted in an operating plant, reagent dosages are routinely adjusted according to feed variability and plant control strategy; however, to improve reproducibility we now provide the operating windows used during our campaign, which are also consistent with ranges reported for NaOl/Na₂SiO₃ magnesite flotation in recent studies. Specifically, the pulp density was maintained at ~25–30 wt.% solids; pH was controlled at 8.5 ± 0.2; sodium silicate (depressant/dispersant) and sodium oleate (collector) were added sequentially with conditioning of ~3 min after each addition, followed by a flotation time of ~3 min at ambient temperature (≈20–25 °C). Reagent dosages are now reported as ranges (Na₂SiO₃: ~0.2–0.8 kg·t⁻¹; NaOl: ~0.2–0.4 kg·t⁻¹, depending on pulp density), consistent with published practice

Revisions made: Section 2.1 (Raw materials / flotation procedure) (Pages 3-4, Lines 135–167).

 

Comment 3: The study attributes the microstructural changes primarily to the increased C/S ratio, but an experiment was not designed to independently control the C/S ratio. Only two different raw materials are compared, introducing multiple simultaneous variables (thermal history, impurity distribution, grain morphology, residual chemistry). This prevents isolating the true effect of the C/S ratio.

Response3: Thank you for your insightful comment. We recognize that comparing only two different raw materials, DSQLM-1 and FDSQLM-3, introduces several variables, such as thermal history, impurity distribution, and grain morphology, which complicate isolating the effect of the CaO/SiO2 (C/S) ratio on the microstructure of the fused magnesia.

We agree that independently controlling the C/S ratio in a more systematic experiment would be ideal for confirming its direct influence on the material's microstructure and densification. While the current study provides valuable insight into the relationship between the C/S ratio and the formation of dicalcium silicate (C₂S), we acknowledge the need for further research to eliminate these confounding factors.

To address this, we propose two possible avenues for future work:

(1)Experimental Design with Controlled C/S Ratios: A set of experiments where the CaO/SiO₂ ratio is precisely varied while keeping all other factors, including thermal history, impurity distribution, and grain morphology, constant. This would allow for a direct investigation into how the C/S ratio influences the microstructure and densification of fused magnesia.

• Use of Synthetic Samples: In future studies, synthetic magnesite samples with controlled compositions could be used to isolate the effect of the C/S ratio. By carefully controlling the impurity contents and the distribution of phases, we can better understand the impact of the CaO/SiO₂ratio without the influence of other variables.

We hope that these suggestions can help address the concern about isolating the true effect of the C/S ratio in the fused magnesia microstructure and densification. We will clarify this limitation in the revised manuscript and indicate the necessity of future studies to better isolate the C/S ratio's effect in a controlled environment.

Thank you again for your valuable feedback.

Comment 4: The “natural” cooling from 2800 °C is not quantified (rate, atmosphere, thermal gradients). Since the β→γ-C₂S transformation is highly dependent on thermal kinetics, this omission is critical and compromises the interpretation of microcrack formation.

Response4: Thank you for your valuable comment. We agree that the cooling process is critical in determining the thermal kinetics of the β→γ-C₂S transformation, and the absence of detailed quantification of cooling rate, atmosphere, and thermal gradients is indeed a significant limitation of the current study. The transformation from β-C₂S to γ-C₂S is known to be highly sensitive to cooling conditions, as it is associated with a substantial volume expansion, which can lead to microcrack formation and, consequently, impact the densification of fused magnesia.

In the context of fused magnesia production, it is important to note that the electrically fused magnesia production process still primarily uses the arc furnace method, where the cooling rate and thermal gradients are largely determined by empirical operational practices. The cooling rate, shell removal, and insulation time during this process significantly influence the occurrence and extent of the β→γ-C₂S transformation. Variations in these parameters can lead to considerable differences in the microstructure of the final product, including the formation and distribution of microcracks.

Currently, the cooling process is not strictly controlled in a standardized manner, and each manufacturer may adopt slightly different cooling protocols, resulting in variations in the microstructure of fused magnesia from different producers. This could explain why products from different manufacturers show significant differences in the degree of microcracking and densification, even if they are produced from similar raw materials.

To address this issue, we suggest that future work should explore the impact of controlled cooling rates, atmosphere, and thermal gradients on the β→γ-C₂S transformation. Such studies would allow for a clearer understanding of the role of thermal kinetics in microcrack formation. We acknowledge that incorporating detailed cooling parameters in future experiments would significantly improve the interpretation of the microstructure and densification behavior of fused magnesia.

Thank you again for pointing out this critical aspect of the study.

Comment 5: Although 22 EDS points are reported, their representativeness is not justified, nor are selection criteria, repeatability, or standard deviation indicated. The analysis remains essentially qualitative, despite the quantitative conclusions drawn regarding Ca and Si enrichment in microcracks.

Response5: Thank you for this constructive comment. We agree that point EDS results are inherently local and semi-quantitative, and that the representativeness and intended use of the 22 points should be clarified. In the revised manuscript, we therefore made two adjustments.

First, we clarified that the 22 EDS points were used primarily for phase/occurrence identification, not for statistical quantification of bulk compositions. The points were selected from representative BSE micrographs to cover the key impurity occurrence modes relevant to beneficiation (intergranular vs. intragranular dolomite, fracture-filling quartz, chlorite/talc in pores/fractures, and minor accessory phases). We added explicit selection criteria (BSE contrast + morphology + location at grains/fractures/triple junctions) and stated that all EDS measurements were collected under identical instrument conditions to ensure internal comparability.

Second, we revised the wording in Section 3.3 to avoid implying a statistical/quantitative conclusion from EDS. The statement “Ca and Si enrichment in microcracks” is now explicitly framed as localized compositional evidence (semi-quantitative EDS indicating Ca–Si–rich secondary phase association at crack/grain-boundary regions), while the quantitative support for microstructural degradation is provided by ImageJ-based crack area fraction analysis (FFM-3: 0.31 vs. FM-1: 0.07) and by the Rietveld-QPA phase identification/quantification. We also acknowledged as a limitation that standard deviations for EDS spot compositions were not established in this study due to the strong spatial heterogeneity of microcrack/secondary-phase regions, and we stated that future work will include replicated spot/line-scan or mapping statistics to quantify variability more rigorously.

Revisions made: Clarifications and toned-down wording were added in Section 3.2 (EDS point selection and purpose) (Pages 6, Lines 244–248).

Comment 6: The interpretation based on the MgO–CaO–SiO₂ ternary system assumes thermodynamic equilibrium, when the industrial melting and cooling process is clearly non-equilibrium. The possible formation of metastable phases and the discrepancy between equilibrium and actual kinetics are not discussed.

Response6: Thank you for this important comment. We agree that the industrial arc-melting and natural cooling of fused magnesia are inherently non-equilibrium processes, and that interpreting phase assemblages solely by equilibrium phase relations can be oversimplified. In the revised manuscript, we clarified that the MgO–CaO–SiO₂ ternary system is used as a thermodynamic guideline to rationalize the tendency of impurity-phase selection with changing CaO/SiO₂ ratio (C/S), rather than as a strict equilibrium prediction of the final phase assemblage.

We also expanded the discussion to explicitly address kinetic and metastability effects relevant to this system: (i) high C/S compositions may promote formation of metastable Ca–silicate phases during cooling (e.g., transient C₃S and its decomposition to C₂S + free CaO), and (ii) the retention and polymorphic evolution of β-C₂S (and its β→γ transformation) are sensitive to thermal history; (iii) rapid solidification can also favor glassy/intergranular amorphous phases and non-equilibrium solubility (e.g., Ca incorporation in MgO), which may not be captured by equilibrium phase diagrams. We now explicitly state that the fused blocks were naturally cooled and the cooling rate was not independently controlled, so the extent of metastable phase formation and transformation-induced microcracking may vary with cooling kinetics.

Accordingly, we added this equilibrium-vs-kinetics point as a limitation and outlined future work using controlled cooling schedules and complementary characterization (e.g., full quantitative phase analysis and, where possible, thermal analysis/in situ approaches) to systematically evaluate kinetic effects and validate the thermodynamic trends across broader processing windows.

Revisions made: Section 3.3 (thermodynamic/kinetic discussion) and Conclusions/Limitations were revised accordingly (Pages 10, Lines 340–344).

 

Comment 7: Unlike the detailed gangue analysis in the raw materials, the following are not quantified: C₂S volume fraction vs. CMS, microcrack density and length, and the statistical distribution of grain size in FM-1 and FFM-3. Without this data, the claims regarding structural deterioration remain at a descriptive level.

Response7: Thank you for this constructive comment. We agree that the Discussion should be supported by quantitative metrics rather than purely descriptive observations. In the revised manuscript, we have strengthened the quantitative support in two key aspects:

CMS vs. C₂S quantification: We performed Rietveld quantitative phase analysis (QPA) of the XRD patterns using HighScore. The refinements show good fitting quality (FM-1: wRp ≈ 5.75; FFM-3: wRp ≈ 6.25) and yield quantitative phase fractions (Table X): FM-1 contains MgO 95.83 wt.%, CMS 3.12 wt.%, and Ca₃(PO₄)₂ 1.05 wt.%; FFM-3 contains MgO 98.24 wt.%, β-C₂S 1.51 wt.%, and γ-C₂S 0.25 wt.%. This directly substantiates the CMS-dominant versus C₂S-dominant impurity-phase distinction.

Microcrack quantification: We added ImageJ-based quantitative image analysis to quantify microstructural degradation. The crack area fraction increases from 0.07 (FM-1) to 0.31 (FFM-3) (Fig. 9), providing a quantitative metric linking high C/S chemistry and C₂S formation with increased cracking.

Regarding microcrack density/length and the statistical grain-size distribution, we acknowledge that the present work reports periclase grain size mainly as ranges estimated from representative micrographs (FM-1: ~300–400 μm; FFM-3: ~400–600 μm), and that a full stereological/statistical distribution analysis (e.g., crack density/length distributions and grain-size histograms over multiple fields of view) was not conducted. We have therefore toned down any overextended statements and explicitly stated this as a limitation, and we will quantify these descriptors systematically in future work to establish a more complete microstructure–property framework.

Revisions made: QPA results were added (Table 5) and discussed in Section 3.3; quantitative crack analysis was explicitly reported (Fig. 9); limitations were clarified in the Conclusions and Limitations and scope. (Pages 14, Lines 344–348).

 

Comment 8:  A possible degradation of mechanical strength and thermal shock is inferred, but no mechanical or thermomechanical properties are measured. This weakens the applied validity of the results and leaves the conclusions in the speculative realm.

Response8: We appreciate the reviewer’s insightful comment. We agree that mechanical and thermomechanical properties (e.g., strength, fracture toughness, thermal-shock resistance) were not directly measured in the present study, and therefore any statement regarding strength/thermal-shock degradation should be expressed cautiously.

In this work, our primary objective is to clarify how impurity occurrence and the CaO/SiO₂ratio (C/S) affect phase formation (CMS vs. C₂S), microcrack generation (β→γC₂S transformation), and densification behavior of fused magnesia prepared from flotation concentrate versus natural high-grade ore. The inference on potential performance deterioration was derived from direct microstructural evidence, including:

formation of C₂S in the high C/S sample (FFM-3) confirmed by XRD/SEM-EDS;abundant intergranular microcracks observed in FFM-3; and quantitative ImageJ analysis showing a much higher crack area fraction in FFM-3 than in FM-1, while the bulk densities remain comparable (statistically insignificant difference).

We have revised the manuscript to:

explicitly state that mechanical/thermomechanical properties were not measured in this study;

tone down the claims regarding mechanical strength and thermal-shock resistance, and reframe them as a plausible implication based on microcracking rather than a demonstrated property change;focus the conclusions on our validated findings (phase evolution, microcracking mechanism, and densification), and add that mechanical/thermal-shock testing will be performed in future work to fully establish applied performance.

Revisions made: Highlights, Abstract, Section 3.3 (Microstructure and Properties of Fused Magnesia), and Conclusions (statement on thermomechanical properties) were adjusted accordingly (Pages 10, Line 320, Pages 104, Lines 509–511).

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript addresses an important and practically relevant topic concerning the utilization of flotation-processed magnesite for fused magnesia production.

1. The novelty of the study is only moderately articulated and should be stated more explicitly.
2. The conclusions rely on a limited comparative dataset (one high-grade ore versus one flotation-derived concentrate), which restricts the general applicability of the findings.
3. Although microstructural evidence of cracking is convincingly presented, the lack of direct measurements of mechanical or thermomechanical properties weakens the claim that densification and service performance are significantly impaired.
4. The discussion would benefit from a more critical comparison with recent international studies and from acknowledging the absence of quantitative phase fraction analysis and controlled cooling-rate effects. Despite these limitations, the manuscript provides useful insights for industrial  strategies.

 

Author Response

Comment 1: The novelty of the study is only moderately articulated and should be stated more explicitly.

Response1: Thank you for this valuable suggestion. We agree that the novelty of the study should be stated more explicitly. The key contributions of this work are:

Impurity Occurrence and Flotation Performance: We provide a detailed analysis of how different impurity occurrence states in magnesite ores (such as dolomite, quartz, talc, and chlorite) influence flotation performance. By combining FE-SEM/EDS analysis, we elucidate how the physical and chemical properties of these impurities affect their liberation and removal during flotation, which is essential for understanding flotation efficiency in low-grade ores.

C/S Ratio and Phase Evolution: We demonstrate that the CaO/SiO₂ (C/S) ratio is a critical factor that governs the phase evolution in fused magnesia. The study shows that an increase in C/S ratio can lead to the formation of dicalcium silicate (C₂S), which significantly impacts the microstructure and densification of fused magnesia. This mechanistic link between C/S ratio and phase evolution is a novel insight that extends previous work on magnesite flotation.

Microstructure–Property Link: The work introduces the first quantitative evidence linking the formation of C₂S with microcracking in fused magnesia. Through image analysis (ImageJ), we quantify microcrack formation, showing that high C/S ratios contribute to an increase in microcrack area fraction. This connection between phase formation and structural deterioration has not been adequately explored in previous studies.

Practical Implications for Industrial Processes: We propose practical strategies for optimizing the flotation process to control the C/S ratio, ensuring the stability of non-transformable phases like CMS and avoiding the undesirable formation of metastable C₂S. This approach provides insights into how to achieve high-quality fused magnesia from low-grade magnesite ores, offering valuable guidance for industrial applications.

Comment 2: The conclusions rely on a limited comparative dataset (one high-grade ore versus one flotation-derived concentrate), which restricts the general applicability of the findings.

Response2: Thank you for this important comment. We acknowledge that the fused magnesia comparison in this study is based on one natural high-grade ore (DSQLM-1 → FM-1) and one flotation-derived concentrate from a low-grade ore (FDSQLM-3 → FFM-3), which may limit broad statistical generalization across different deposits and beneficiation routes. To address this concern and avoid over-generalization, we strengthened the scope/limitation statements in both the Section 3.3 discussion and the Conclusions, and we explicitly stated that broader validation using additional ore sources and multiple concentrates will be conducted in future work.

Added/strengthened limitation and scope in Section 3.3 (Discussion): We clarified that the study is designed as a controlled, mechanism-oriented comparison under identical fusion conditions, and framed the CMS-to-C2S shift and transformation-induced microcracking as mechanistic outcomes associated with a pronounced change in CaO/SiO₂ (C/S) ratio, while noting that broader generality requires further validation.

Added limitation statement and future validation plan in Conclusions: We added a dedicated statement emphasizing that the proposed C/S-control guideline should be validated with a broader dataset (multiple ore sources and flotation concentrates with varied impurity occurrence and C/S ratios) under comparable processing conditions (Pages 12, Lines 436–439).

Comment 3: Although microstructural evidence of cracking is convincingly presented, the lack of direct measurements of mechanical or thermomechanical properties weakens the claim that densification and service performance are significantly impaired.

Response3: We appreciate the reviewer’s insightful comment. We agree that mechanical and thermomechanical properties (e.g., strength, fracture toughness, thermal-shock resistance) were not directly measured in the present study, and therefore any statement regarding strength/thermal-shock degradation should be expressed cautiously.

In this work, our primary objective is to clarify how impurity occurrence and the CaO/SiO₂ ratio (C/S) affect phase formation (CMS vs. C₂S), microcrack generation (β→γ C₂S transformation), and densification behavior of fused magnesia prepared from flotation concentrate versus natural high-grade ore. The inference on potential performance deterioration was derived from direct microstructural evidence, including:

formation of C₂S in the high C/S sample (FFM-3) confirmed by XRD/SEM-EDS; abundant intergranular microcracks observed in FFM-3; and quantitative ImageJ analysis showing a much higher crack area fraction in FFM-3 than in FM-1, while the bulk densities remain comparable (statistically insignificant difference).

We have revised the manuscript to: explicitly state that mechanical/thermomechanical properties were not measured in this study; tone down the claims regarding mechanical strength and thermal-shock resistance, and reframe them as a plausible implication based on microcracking rather than a demonstrated property change;focus the conclusions on our validated findings (phase evolution, microcracking mechanism, and densification), and add that mechanical/thermal-shock testing will be performed in future work to fully establish applied performance.

Revisions made: Highlights,Abstract, Section 3.3 (Microstructure and Properties of Fused Magnesia), and Conclusions (statement on thermomechanical properties) were adjusted accordingly (Pages 14, Lines 507–511).

 

Comment 4: The discussion would benefit from a more critical comparison with recent international studies and from acknowledging the absence of quantitative phase fraction analysis and controlled cooling-rate effects. Despite these limitations, the manuscript provides useful insights for industrial  strategies.

Response4: Thank you for your valuable comments. We understand that the discussion section would benefit from a more in-depth comparison with recent international studies and further acknowledgment of the limitations in our study, particularly regarding quantitative phase fraction analysis and controlled cooling-rate effects.

We note that while research on magnesia ceramics and magnesia-based refractories has increased in recent years, systematic studies on the impact of impurities on phase evolution and microstructure during rapid densification are still relatively limited, especially in terms of the control of grain growth under varying cooling rates. Therefore, we have tried to integrate the existing research and point out the importance of future studies in these areas.

In the revised manuscript, we have added a comparison with current studies and clearly acknowledged the absence of quantitative phase fraction analysis and controlled cooling-rate experiments, suggesting that future research could expand on these aspects to complement the current experimental data and validate our findings.

This response acknowledges the limitations, highlights the challenges in the field, and provides a reasonable explanation for future research directions while still emphasizing the value of the current study.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The materials of the article submitted for review are devoted to a very relevant and interesting problem of increasing the efficiency of processing magnesium raw materials. The authors have extensively studied the features affecting the stage of chemical processing. However, the reviewer recommends that the authors pay attention to the following shortcomings.
1) The authors should describe the flotation conditions in more detail, namely, specify the mass of the sample, the reagents consumption value, the solid content, and what type of flotation machine was used.
2) It is also necessary to specify in more detail the characteristics of the flotation feed, namely its granulometric composition and technological parameters of the sample grinding process to a given size.
3) In line 216, the authors should clarify the term concentrate floatability in the context of the description of slime formation. Perhaps the authors meant that the chlorite passes into slurries during re-grinding and adheres to the particles of magnesite, which leads to a decrease in the content of the valuable component in the resulting concentrate of reverse flotation.
4) The authors should also clarify the sentence in line 237 regarding mechanical entrainment. If the authors use reverse flotation, then the mechanical entrainment of minerals from the host rock, on the contrary, will increase the selectivity of separation, or does the authors have another mechanism in mind?
5) The authors are also encouraged to clarify their conclusion more clearly regarding the limitations in the applicability of flotation for the pre-concentration of magnesite. Did the authors mean that flotation, in principle, cannot ensure compliance with the requirements set by the authors for the phase composition of the concentrate, or is it simply necessary to select the correct separation conditions?

Author Response

Comment 1: The authors should describe the flotation conditions in more detail, namely, specify the mass of the sample, the reagents consumption value, the solid content, and what type of flotation machine was used.

Response1: Thank you for your valuable feedback. In response to your suggestion, we have provided a more detailed description of the flotation conditions in the revised manuscript, including the sample mass, reagent consumption, solid content, and the type of flotation machine used.

Specifically, we used industrial reverse flotation equipment, with a sample mass of 100% passing 0.088 mm. The solid content during flotation was maintained at 25–30 wt.%, and the flotation was conducted under weak alkaline conditions (pH 8.5 ± 0.2). Sodium silicate was used as a depressant/dispersant with a consumption rate of 0.2–0.8 kg·t⁻¹, and sodium oleate was used as the collector with a consumption rate of 0.2–0.4 kg·t⁻¹. The flotation process was performed under mechanical agitation, with a flotation residence time of approximately 3 minutes.

These details have been added to the Materials and Methods section to clarify the flotation process conditions (Pages 3-4, Lines 135–167).

Comment 2: It is also necessary to specify in more detail the characteristics of the flotation feed, namely its granulometric composition and technological parameters of the sample grinding process to a given size.

Response2: Thank you for your valuable feedback. In response to your suggestion, we have provided additional details regarding the flotation feed characteristics and the technological parameters of the sample grinding process in the revised manuscript.

In this study, the flotation feed was selected by screening through a 180 mesh sieve, and the undersized material was used for flotation. Due to some inherent errors in the screening process, the exact granulometric range may not be very precise. Specifically, the feed mainly consisted of particles passing through a 180 mesh sieve (approximately 88 μm), but due to the lack of particle size analysis data, we cannot provide a more precise granulometric distribution. We have clarified this process in the Materials and Methods section and included a description of the screening process to help readers understand the sample preparation method.

Regarding the grinding process, the sample was ground using standard ball milling methods, ensuring that the sample achieved the desired particle size range (Pages 3-4, Lines 141–147).

Comment 3: In line 216, the authors should clarify the term concentrate floatability in the context of the description of slime formation. Perhaps the authors meant that the chlorite passes into slurries during re-grinding and adheres to the particles of magnesite, which leads to a decrease in the content of the valuable component in the resulting concentrate of reverse flotation.

Response3: Thank you for this helpful comment. We agree that the term “concentrate floatability” is ambiguous in our original wording. What we intended to convey is that chlorite is prone to generating ultrafine slimes during grinding, and these slimes may (i) adhere to/coat magnesite surfaces (slime coating) and (ii) cause mechanical entrainment of fine gangue into the froth, both of which reduce flotation selectivity. As a result, the reverse-flotation performance deteriorates, typically reflected by a lower concentrate quality (i.e., reduced MgO grade and/or increased gangue content) and potentially reduced recovery. We have revised the manuscript accordingly by replacing “concentrate floatability” with a clearer description focusing on slime coating/entrainment and their impact on flotation selectivity and concentrate grade/recovery.

Revisions made: The sentence around in Section 3.2 was revised to clarify the mechanism and to avoid the ambiguous term (Page 7, Lines 276–281).
Comment 4: The authors should also clarify the sentence in line 237 regarding mechanical entrainment. If the authors use reverse flotation, then the mechanical entrainment of minerals from the host rock, on the contrary, will increase the selectivity of separation, or does the authors have another mechanism in mind?

Response4: Thank you for this comment. We agree that our original wording may cause confusion. In flotation terminology, “mechanical entrainment” refers to the non-selective physical carryover of fine particles with water into the froth, rather than true bubble–particle attachment. In our reverse flotation circuit, the froth product is the gangue-rich reject, while the underflow/remaining slurry is the magnesite concentrate. Therefore, increased entrainment does not improve selectivity; instead, it can (i) carry fine magnesite (or Mg-bearing composite particles) into the froth, resulting in valuable-mineral loss and reduced recovery, and/or (ii) increase water recovery and fine particle carryover, complicating froth cleaning. In other words, entrainment is detrimental because it introduces a non-selective transport pathway, reducing separation efficiency even in reverse flotation.

We revised the sentence around line 237 to explicitly state that the microcrack-rich, porous microstructure promotes entrainment of fine particles into the froth, which may increase magnesite losses to the froth and thereby reduce reverse-flotation selectivity and efficiency.

Revisions made: Section 3.2, sentence around.
Comment 5: The authors are also encouraged to clarify their conclusion more clearly regarding the limitations in the applicability of flotation for the pre-concentration of magnesite. Did the authors mean that flotation, in principle, cannot ensure compliance with the requirements set by the authors for the phase composition of the concentrate, or is it simply necessary to select the correct separation conditions?

Response5: Thank you for this important clarification request. We would like to emphasize that we do not conclude that flotation is fundamentally incapable of producing a suitable magnesite concentrate. On the contrary, flotation is a mature and effective industrial technology for upgrading low-grade magnesite and is one of the most practical routes for its high-efficiency utilization. In this study, reverse flotation successfully removed >97% SiO₂ and significantly increased the MgO grade of the concentrate, demonstrating its strong beneficiation capability.

The “limitation” we intend to highlight is more specific: flotation is typically designed to maximize impurity removal and concentrate grade/recovery, but selective removal of a specific impurity (e.g., SiO₂) can unintentionally alter the relative oxide balance of the remaining impurities (e.g., CaO/SiO₂ ratio). This compositional shift can influence the downstream phase assemblage during fusion (CMS vs. transformation-prone C₂S) and thus affect microstructural integrity and densification. Therefore, rather than stating that flotation cannot meet phase-composition requirements in principle, we argue that flotation pre-concentration should be coupled with composition/oxide-ratio control and application-oriented optimization (e.g., adjusting separation schemes, blending, or other corrective measures) when the concentrate is intended for fused magnesia/refractory applications.

We revised the Conclusions to express this point more clearly and to state that, although flotation is a powerful and mature method for low-grade magnesite upgrading, the application of flotation concentrates (powder) to fused magnesia still requires further theoretical understanding and practical process optimization to consistently achieve the desired impurity-phase outcomes (Page 7, Lines 276–281).

Author Response File: Author Response.pdf

Reviewer 5 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for submitting your manuscript. Your work presents a valuable and well-documented investigation into the role of impurity chemistry on fused magnesia performance. The integration of flotation beneficiation, microstructural characterization, and phase evolution analysis is particularly commendable.

However, before the manuscript can be considered for publication, several important points require clarification or further development:

1. Quantitative phase analysis is needed (e.g., Rietveld refinement) to support conclusions regarding CMS and C₂S dominance.
2. Mechanical or thermomechanical properties should be included or the absence explicitly justified.
3. Quantitative crack analysis should be provided to strengthen the link between β-C₂S formation and microstructural degradation.
4. Thermodynamic discussion (C/S ratio, phase stability, cooling effects) should be expanded and better supported.
5. Industrial relevance should be reinforced by proposing practical strategies for C/S ratio control during processing.
6. Minor improvements are needed regarding figure clarity, terminology consistency, and statistical significance of density differences.

With these revisions, the manuscript has strong potential to make a meaningful contribution to the field of refractory materials and magnesite beneficiation.

I encourage you to address these points thoroughly and look forward to reviewing a revised version.

Kind regards,

Author Response

Comment 1: Quantitative phase analysis is needed (e.g., Rietveld refinement) to support conclusions regarding CMS and C₂S dominance.

Response2: Thank you for this important suggestion. We agree that quantitative phase analysis (QPA) via Rietveld refinement provides stronger support for assessing the relative abundance of impurity phases. Following the reviewer’s recommendation, we performed Rietveld refinement and QPA of the XRD patterns using HighScore, with MgO (periclase) explicitly included in the refinement for both samples. The refinements show good fit quality (FM-1: wRp ≈ 5.75; FFM-3: wRp ≈ 6.25). The quantified crystalline phase fractions are as follows: for FM-1, MgO 95.83 wt.%, CMS 3.12 wt.%, and Ca₃(PO₄)₂ 1.05 wt.%; for FFM-3, MgO 98.24 wt.%, β-C₂S 1.51 wt.%, and γ-C₂S 0.25 wt.%. These results are summarized in Table X and are consistent with our qualitative XRD/SEM–EDS observations, confirming that CMS is the main Ca–Mg–silicate impurity phase in FM-1, whereas C₂S (β/γ) is present in FFM-3.

To avoid over-interpretation, we also clarified in the manuscript that the reported values refer to the crystalline phases included in the refinement; since no internal standard was used, any amorphous contribution was not quantified.

Revisions made: The Rietveld-QPA results were added in Section 3.3 and Table 5, and related statements in the Abstract/Discussion/Conclusions were updated accordingly (Page 12, Line 444).

Comment 2: Mechanical or thermomechanical properties should be included or the absence explicitly justified.

Response2: Thank you for this important comment. We agree that direct mechanical/thermomechanical testing (e.g., strength, fracture toughness, elastic modulus, and thermal-shock resistance) would further strengthen the service-performance relevance. In the present work, however, our primary objective was to elucidate the mechanistic link between impurity occurrence/CaO–SiO₂ balance and the resulting phase evolution (CMS vs. C₂S), microcrack formation (β→γ C₂S transformation), and densification behavior of fused magnesia produced from a flotation concentrate versus a natural high-grade ore under identical fusion conditions. Accordingly, the experimental program focused on compositional/phase characterization (XRF, XRD including Rietveld-QPA, SEM–EDS) and densification metrics (bulk density and ImageJ-based crack/feature quantification).

We have explicitly stated in the revised manuscript that mechanical and thermomechanical properties were not measured in this study and therefore any performance-related statements must be treated as plausible implications based on the observed microcracking, rather than demonstrated property changes. We have toned down the wording in the Abstract, Section 3.3 (Discussion), and the Conclusions, and we added a clear limitation statement. We also indicate that systematic mechanical and thermal-shock testing will be carried out in future work to establish a complete structure–property relationship for fused magnesia derived from flotation concentrates.

Revisions made: Explicit limitation and revised wording were added/updated in the Abstract, Section 3.3, and Conclusions (Page 14, Lines 505–511).

Comment 3: Quantitative crack analysis should be provided to strengthen the link between β-C₂S formation and microstructural degradation.

Response3: Thank you for this helpful suggestion. We agree that quantitative crack analysis strengthens the mechanistic link between C₂S formation/transformation and microstructural degradation. In the revised manuscript, we have provided quantitative crack evaluation using ImageJ-based image analysis on representative BSE micrographs. The results show that the crack area fraction in FFM-3 is 0.31, which is substantially higher than 0.07 in FM-1 (Fig. 9). This quantitative difference is consistent with our phase/microchemical evidence: XRD/Rietveld-QPA identifies β/γ-C₂S in FFM-3, and SEM–EDS confirms Ca–Si enrichment associated with the intergranular crack regions. Together, these results support the proposed mechanism that β→γ C₂S transformation contributes to crack formation and microstructural deterioration.

Revisions made: Quantitative crack analysis and the corresponding discussion were added/expanded in Section 3.3 and presented in Fig. 9.

Comment 4: Thermodynamic discussion (C/S ratio, phase stability, cooling effects) should be expanded and better supported.

Response4: Thank you for this valuable suggestion. We agree that the thermodynamic basis linking the CaO/SiO₂ ratio (C/S), phase stability/selection (CMS vs. C₂S/C₃S), and cooling-related damage should be articulated more clearly and supported with stronger evidence. In the revised manuscript, we expanded the discussion in Section 3.3 in the following ways:

C/S-controlled phase selection in the MgO–CaO–SiO₂ system. We strengthened the phase-equilibrium rationale by explicitly linking the measured C/S ratios (FM-1: 0.68; FFM-3: 2.85) to the expected impurity-phase fields in the MgO–CaO–SiO₂ system (Table 4). Consistent with the phase diagram-based guidance, low C/S favors CMS-type phases, whereas high C/S promotes C₂S (and potentially C₃S at elevated C/S), providing a thermodynamic explanation for the CMS→C₂S shift after selective silica removal by reverse flotation [54,55].

Phase stability during cooling and the β→γ C₂S transformation. We expanded the cooling-related discussion to emphasize that C₂S is particularly detrimental not because of its high melting point, but because of the polymorphic β→γ transformation during cooling (≈725 °C) accompanied by ~12% volume expansion, which generates internal stresses and microcracking [56]. We also clarified that this is a thermodynamic/structural transformation whose damage severity can be influenced by thermal history.

Quantitative support by Rietveld-QPA and microcrack quantification. To better support the phase-selection argument, we added HighScore Rietveld refinement and quantitative phase analysis (QPA) for both samples with MgO included. The refinements yielded good fits (FM-1: wRp ≈ 5.75; FFM-3: wRp ≈ 6.25) and quantified crystalline phase fractions (FM-1: MgO 95.83 wt.%, CMS 3.12 wt.%, Ca₃(PO₄)₂ 1.05 wt.%; FFM-3: MgO 98.24 wt.%, β-C₂S 1.51 wt.%, γ-C₂S 0.25 wt.%), summarized in Table X. In addition, ImageJ analysis provides quantitative crack area fractions (FFM-3: 0.31 vs. FM-1: 0.07; Fig. 9), supporting the proposed cooling-transformation damage mechanism.

Controlled cooling-rate effects as a limitation and outlook. We explicitly acknowledged that the fused blocks were naturally cooled under industrial conditions and that cooling rate was not independently controlled; therefore, the extent of β→γ transformation damage may vary with cooling kinetics. We added this as a limitation and outlined future controlled cooling experiments to further validate the phase–microstructure linkage and broaden applicability.

Revisions made: The thermodynamic discussion was expanded in Section 3.3 (Discussion), supported by Table 4, the added Rietveld-QPA results (Table 5), and the quantitative crack analysis (Fig. 9), and the cooling-rate limitation was explicitly stated (Page 12, Line 443).

.

Comment 5: Industrial relevance should be reinforced by proposing practical strategies for C/S ratio control during processing.

Response5: Thank you for this valuable suggestion. We agree that translating the C/S–phase–microcracking mechanism into actionable processing guidance is important for industrial relevance. In the revised manuscript, we strengthened the industrial implications by explicitly proposing practical strategies to control the CaO/SiO₂ ratio (C/S) during the integrated “flotation–fusion” route. Specifically, we added a processing-oriented discussion emphasizing that, beyond lowering total impurities, concentrate design should include oxide-ratio management to suppress transformation-prone C₂S formation and favor stable intergranular phases (e.g., CMS). The proposed practical approaches include: (i) avoiding excessive selective desilication that over-enriches CaO relative to SiO₂ (i.e., optimizing flotation flowsheet/reagent scheme to balance SiO₂ removal and C/S stability); (ii) implementing blending/ore mixing between concentrates (or between concentrate and selected high-silica but low-Ca streams) to target an appropriate C/S window; (iii) where applicable, adding a Ca-reduction step prior to fusion; and (iv) applying composition correction by controlled addition of Si-bearing modifiers during pelletization to adjust C/S when blending is not feasible. We also added a simple quality-control recommendation: routine monitoring of CaO and SiO₂ in concentrates (XRF) and using C/S as a control index to guide downstream fusion.

Revisions made: Practical C/S-control strategies and quality-control suggestions were summarized in the Conclusions (Page 14, Lines 502-505) 

Comment 6: Minor improvements are needed regarding figure clarity, terminology consistency, and statistical significance of density differences.

Response6: We appreciate the reviewer’s valuable feedback. In response to your suggestions, we have made the following improvements to the manuscript:

Figure Clarity: We have improved the clarity of all figures by increasing the font size and enhancing the resolution to ensure that the figures are more readable and maintain their quality when zoomed in. This ensures that all details are clearly visible in the final manuscript (Figure 1,2. Page 6, Line 232, 236, Table 2., Line 318).

Terminology Consistency: We have reviewed the manuscript to ensure consistency in the use of technical terms. Specifically, we have standardized the terms "bulk density" and "relative density", making sure that they are used consistently throughout the text to avoid any confusion or ambiguity.

Statistical Significance of Density Differences: A t-test was performed to analyze the bulk density differences between FM-1 and FFM-3. The results show that the difference in bulk density (0.01 g/cm³) is not statistically significant (p-value > 0.05). We have included these results in the revised manuscript to support the conclusion that the bulk densities of both samples are effectively comparable (Page 10, Line 356).

Thank you again for your constructive comments. We believe these changes enhance the clarity, consistency, and scientific rigor of the manuscript, and we hope the revisions meet your expectations.

 

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

After a successful revision, this paper can be recommended for publication.

Author Response

Dear Reviewer,

Thank you for taking the time to review my manuscript again and for your positive feedback. I am very pleased to know that there are no further comments from your side. Your recognition is greatly appreciated. Once again, thank you for your valuable time and support!

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript shows clear improvements after revision, particularly in the methodological description, the incorporation of quantitative data on phases and microcracks, and the moderation of some claims. However, a fundamental problem remains in the interpretation of the results. The study does not allow for isolating the effect of the C/S ratio or controlling for a critical parameter such as cooling, even though both are central to the proposed mechanism of microcrack formation associated with the β→γ-C₂S transformation. Although these limitations are acknowledged in the text, they are not yet sufficiently reflected in the overall approach of the work. For the manuscript to be acceptable, the study must be explicitly presented as a comparison of industrial materials, avoiding direct causal inferences and adjusting the title, abstract, and conclusions so that relationships are expressed as data-supported associations, rather than as proven mechanisms.

Author Response

Comment:The manuscript shows clear improvements after revision, particularly in the methodological description, the incorporation of quantitative data on phases and microcracks, and the moderation of some claims. However, a fundamental problem remains in the interpretation of the results. The study does not allow for isolating the effect of the C/S ratio or controlling for a critical parameter such as cooling, even though both are central to the proposed mechanism of microcrack formation associated with the β→γ-C₂S transformation. Although these limitations are acknowledged in the text, they are not yet sufficiently reflected in the overall approach of the work. For the manuscript to be acceptable, the study must be explicitly presented as a comparison of industrial materials, avoiding direct causal inferences and adjusting the title, abstract, and conclusions so that relationships are expressed as data-supported associations, rather than as proven mechanisms.

Response: We sincerely thank the reviewer for re-evaluating our manuscript and for this crucial comment. We fully agree with the core concern: because this study is based on industrial raw materials and industrial processing/cooling conditions, it does not allow us to independently isolate the effect of the CaO/SiO₂ (C/S) ratio or to control a critical parameter such as the cooling regime. As a result, the present work cannot support definitive causal attribution of the observed microcracking to a single mechanism.

Following your recommendations, we have made fundamental adjustments to the manuscript. The core revision is to explicitly shift the study's positioning from "mechanism verification" to a "comparative analysis and investigation of associations for industrial materials." The specific modifications are as follows, all key modifications have been highlighted in green in the manuscript.

Repositioning the Nature of the Study: In the title, abstract, introduction, and conclusions, we have explicitly framed this research as a comparative analysis of fused magnesia from two industrial sources (natural high-grade ore and flotation concentrate). We emphasize that it aims to reveal associations among composition, processing, and properties.

Systematically Weakening Causal Language: We have systematically reviewed the entire text, replacing vocabulary that implies direct causation (e.g., "promotes," "induces," "leads to," "controls") with terms that express association, correlation, or data-supported trends (e.g., "is associated with," "correlates with," "is linked to," "favors the formation of").

Strengthening Limitations and Reframing Conclusions: In the discussion and conclusions, we have more prominently highlighted the complexity of industrial materials (coupling of multiple variables). We clearly state the observed associations between the C/S ratio, C₂S phase transformation, and microcracking, and explicitly list the rigorous verification of causal mechanisms as a direction for future controlled experiments.

(1) Title

Before revision: Effect of Impurity Occurrence and CaO/SiO₂ Ratio in Flotation-Processed Magnesite on the Microstructure and Densification of Fused Magnesia

After revision: Comparative Analysis of Industrial Fused Magnesia from Natural and Flotation-Processed Magnesite: Associations Among CaO/SiO₂ Ratio, Silicate Phase Formation, and Microcracking(Page 1, Lines 2–5).

Reason: The new title explicitly states "comparative analysis" and "associations," toning down the causal implication of a singular "effect."

• Highlights

Before revision: Flotation enriches CaO/SiO₂ ratio in magnesite concentrate, shifting from 0.68 to 2.85.

After revision: Reverse flotation coincided with higher CaO/SiO₂ ratio (0.68 to 2.85).

Before revision: High CaO/SiO₂ ratio promotes β-dicalcium silicate (C₂S) formation over calcium magnesium silicate (CMS).

After revision: Higher CaO/SiO₂ coincided with C₂S dominance; lower ratio with CMS.

Before revision: The β→γ C₂S transformation (12% vol. expansion) induces microcracks, impairing densification.

After revision: More C₂S coincided with more microcracks and slightly lower densification.

Before revision: For beneficiation, controlling impurity composition (particularly the CaO/SiO₂ ratio) appears as critical as reducing total impurity content.

After revision: Beneficiation may benefit from controlling oxide balance, not only impurities.

Before revision: Tailoring the CaO/SiO₂ ratio in the concentrate is suggested as a strategy to favor the formation of stable intergranular phases and mitigate microcracking.

After revision: Tailoring CaO/SiO₂ may help avoid conditions linked to C₂S and microcracking.

Before revision: This strategy enables high-value utilization of low-grade magnesite for refractory applications with improved microstructural stability.

After revision: Findings are from industrial comparison; causality needs controlled experiments.

(Page 1, Lines 16–22).

Reason: The revisions were made to shift the focus from causal claims to data-supported associations, as per the reviewer’s recommendation. We removed strong causal language such as "promotes" and "induces" and replaced it with terms like "coincided with" and "associated with" to emphasize correlations rather than definitive cause-and-effect relationships. The final highlight emphasizes the need for controlled experiments to validate the causality of the observed associations, ensuring that the study is presented as a comparative analysis, not a mechanism verification.

(3) Abstract

Revised the abstract to better reflect the scope and findings of the study, addressing the reviewer's comments and ensuring clarity in the interpretation of the results. Specifically, we have made the following adjustments:

Clarified the Study Design: We have clearly stated that the study involves a comparative analysis of two industrial fused magnesia products, one produced from high-grade natural ore and the other from a flotation-concentrated low-grade ore. This helps position the study as an industrial comparison rather than a mechanism-focused study.

Refined the Explanation of Key Results: We have improved the description of the results, especially regarding the differences in the CaO/SiO₂ (C/S) ratio, phase composition (CMS vs β/γ-C₂S), and the microstructural implications (higher microcrack metrics in FFM-3). These observations are now presented as data-supported associations rather than causally inferred.

Addressed the Limitations: We explicitly mentioned that cooling history and composition were not independently controlled in this industrial comparison, and that the relationships between the C/S ratio, phase formation, and microcracking should be interpreted cautiously as correlations rather than direct causal links.

Suggested Future Work: We acknowledged that mechanical and thermomechanical properties were not measured in this study and suggested that they should be evaluated in future research to fully understand the performance implications of the findings.

 (Pages 1–2, Lines 24–43).

(4) Introduction

Before revision: CaO reacts with SiO₂ to form β-C₂S… undergoes a destructive β→γ phase transformation upon cooling, resulting in volumetric expansion and disintegration.
After revision: CaO can react with SiO₂ to form β-C₂S, which is reported to undergo a β→γ transformation upon cooling, involving volumetric expansion that may contribute to cracking/disintegration in Ca-bearing systems [14,15] (Page 2, Lines 59–62).

In the research objectives and methodology section, we adjusted the sentence structure to make the research goal clearer, explicitly stating that this study is a comparative analysis of two industrial materials, emphasizing that it is an association analysis rather than a verification of causal mechanisms.

We highlighted the limitations of the study: at the end of the paragraph, we added a note regarding the lack of independent control over cooling history and composition, emphasizing that the observed relationships are data-supported associations, not direct causal inferences.

We revised the language: "causal mechanisms" was changed to "data-supported associations" to better align with the reviewer's requirement to avoid causal relationships. We avoided absolute causal inferences and rephrased it as an analysis based on data-supported correlations.

We also emphasized future research directions: finally, we highlighted that future work needs to verify these relationships through controlled experiments, ensuring the scientific rigor of the study. (Pages 2–3, Lines 81–112).

• Results

Before revision: This suggests that selective silica removal during flotation resulted in relative enrichment of residual CaO in the final product.

After revision: This is consistent with selective silica removal being accompanied by a higher C/S ratio (i.e., relative enrichment of CaO with respect to SiO₂) in the final product (Page 9, Lines 307–309).

Before revision: “Due to the polymorphic transformation… induces substantial internal stresses… These stresses promote… ultimately compromising…”

After revision: The β→γ transformation of C₂S is reported to occur near ~725 °C and involves ~12% volume expansion [56], which could plausibly generate stresses during cooling. In this industrial comparison, the co-occurrence of Ca–Si-enriched regions (consistent with β/γ-C₂S) and grain-boundary microcracking is therefore interpreted as an association consistent with transformation-related cracking scenarios, rather than definitive causal proof (Page 11, Lines 381–386).

Before revision: This increased crack content in FFM-3 aligns with the β→γ transformation and the internal stresses generated, which are responsible for the formation of microcracks.
After revision: This higher crack area fraction in FFM-3 is consistent with increased icrostructural damage in the C₂S-containing sample and may be related to transformation-associated stresses reported for β→γ C₂S during cooling (Page 11, Lines 396–398).

Before revision: The findings above highlight a critical… leading to… thereby facilitating… significantly compromises…

After revision:  These observations highlight a potential trade-off in the flotation–fusion route: selective silica removal may reshape the residual oxide balance (including C/S ratio), coinciding with the occurrence of Ca-rich silicates such as C2S. In the present industrial comparison, C2S occurrence coincided with higher microcrack metrics and slightly lower densification, suggesting that concentrate design may benefit from considering oxide balance in addition to total impurity removal (Page 11, Lines 399–404).

• Discussion

We have made the following adjustments:

We replaced phrases such as "provides a mechanistic basis" and "is critical" with more cautious language like "associated with" and "may generate," which reflects the observed correlations without implying direct causality.

We also emphasized that the C/S ratio and phase transformation are linked to microcrack formation through associations rather than direct causal mechanisms.

Additionally, we clarified that while phase diagrams provide equilibrium tendencies, the industrial processes used in this study are non-equilibrium, and thus metastable phases and cooling-kinetic effects may play a role.

These changes were made to ensure that the study is presented as a comparative analysis of industrial materials, as you recommended. The revised manuscript now reflects a more cautious and data-supported interpretation of the relationships observed, in line with your feedback (Page 13, Lines 436–457).

• Conclusions

Original: The elevated C/S ratio promotes the discontinuous precipitation of β - dicalcium silicate (β - C₂S)... which induces microcrack formation...

Revised: In the fused magnesia derived from the flotation concentrate, the elevated C/S ratio was correlated with the precipitation of β-C₂S. Microcrack formation was observed, which is consistent with the volumetric expansion associated with the β→γ C₂S transformation during cooling (Page 14, Lines475–478).

Original: “...should manage not only total impurities but also the CaO/SiO₂ (C/S) ratio to suppress transformation-prone Ca–silicates and improve microstructural stability.”

Revised: The results suggest that for the flotation–fusion route, managing the CaO/SiO₂ (C/S) ratio in addition to total impurities could be a key strategy to avoid conditions favoring transformation-prone Ca–silicates and thereby potentially improve microstructural stability (Page 14, Lines 479–482).

(7) Limitations and Scope

Strengthening the existing statement:

In the "Limitations and scope" section, the statement regarding "broader generality should be validated..." is further strengthened to:

" it introduces a fundamental limitation of this industrial comparison is the inability to independently control the CaO/SiO₂ ratio and the cooling regime. Therefore, while a strong correlation between a high C/S ratio, C₂S formation, and microcracking is established, the study cannot definitively isolate the causative role of each parameter within the proposed mechanism. The findings should be interpreted as identifying critical interlinked factors in an industrial process, with the mechanistic pathway requiring validation through controlled experiments (Page 14, Lines 487–495).

Author Response File: Author Response.pdf

Reviewer 5 Report

Comments and Suggestions for Authors

Dear Authors, 

All right, many thanks

Cheers

Author Response

Dear Reviewer,

Thank you for taking the time to review my manuscript again and for your positive feedback. I am very pleased to know that there are no further comments from your side. Your recognition is greatly appreciated. Once again, thank you for your valuable time and support!