Analysis on Ancient Bloomery Ironmaking Technology: The Earliest Ironmaking Evidence in the Central Plains of China Was Taken as the Research Object
Round 1
Reviewer 1 Report
The submitted paper deals with the analysis of a sample or several samples from an archaeological site containing remains of ironmaking activities during the Western Zhou dynasty in Central China. The authors basically characterize the samples chemically, morphologically and mineralogically (both from XRD and inferring it from the chemical composition by using an automated TIMA system).
The authors claim to distinguish between the different iron oxidation states in the samples and after stating that the detected Fe3+ was not originally part of the slag composition they use a reference composition to perform a thermodynamic analysis of the sample and its corresponding process of formation.
I have several concerns on the paper:
-the samples that have been analysed are never described (how many are they? what was the sampling strategy?, etc). The paper should contain a section describing and presenting the samples (for instance a “materials and methods” section). Also, the characterization data is never contextualized in the general frame of the analysed sample/s. In particular, it is difficult to judge if the presented SEM photos are really representative of the samples or whether ones are and other not so.
- regarding the obtained data, the distinction between metallic Fe, Fe2+ and Fe3+ seems to be solely based on chemical (EDS and EDX) analyses which are known to be not really reliable in the quantification of O (and oxidation state of the elements is not really seen using these techniques). I would suggest the authors to use for instance Mössbauer spectroscopy to really track the oxidation state of iron.
-on the other hand there seem to be a discrepancy between the XRD results (goethite clearly seen) and the SEM-EDS results where goethite is not identified and in contrast there appers wustite (FeO). The authores never state that wustite has oxidized to goethite.
-The thermodynamic analyses is performed using FactSage7.1 software but the process (what are the input values and why are they so) is not really described neither the obtained potential phase diagram (Figure 6) instead the authors dedicate almost one page (section 4.5) to merely describe results from other authors to similar studies but hardly discussing or connecting these to their own results.
-The paper contains quite a lot of mistakes (grammar, typos and formal register and style), English should be definitely revised.
-In the introduction section (page 2) the authors perform a similar exercise of unconnected sentences describing previous results from other authors that turn to be superfluous and repetitive.
-The paper should be completely rethought and improved if the authors pretend to produce something worthy of publication. Please find attached an annotated version of the paper but I insist that apart from the specific comments the paper should be improved in its whole structure and possibly also in the methods applied.
Comments for author File: 
Comments.pdf
Author Response
Point 1: The samples that have been analysed are never described (how many are they? what was the sampling strategy?, etc). The paper should contain a section describing and presenting the samples (for instance a “materials and methods” section). Also, the characterization data is never contextualized in the general frame of the analysed sample/s. In particular, it is difficult to judge if the presented SEM photos are really representative of the samples or whether ones are and other not so.Response : Thank you for pointing out the defects and shortcomings of the sampling and testing methods.
A total of 10 pieces of slag were recovered from the cultural layer on the western side of the site, small in size, the largest piece being only 2-3 cm in diameter. These slags are black and irregular, magnetic and one of them has rust. The sections are mostly black in crystal-line form. From the morphological presentation of the collected samples, it can be roughly determined that the collected slags has undergone a high-temperature process and that liquid phase flow occurs during smelting. Sampling methods and testing methods, samples I will add in the newly uploaded manuscript (lines 126 to 131).
Point 2: Regarding the obtained data, the distinction between metallic Fe, Fe2+ and Fe3+ seems to be solely based on chemical (EDS and EDX) analyses which are known to be not really reliable in the quantification of O (and oxidation state of the elements is not really seen using these techniques). I would suggest the authors to use for instance Mössbauer spectroscopy to really track the oxidation state of iron.
Response: The titration method and ICP-AES can accurately quantify the composition of the sample. Bloomery ironmaking technology is equivalent in principle to modern direct re-duction iron smelting technology but is limited by low smelting technology and high iron content in the slag, which can be recycled as grade iron ore. In view of this, the analytical detection standard of iron ore and direct reduced iron can be used to detect this batch of ancient slag samples.
The total iron content was determined according to the standard of Iron ores-Determination of total iron content-Titrimetric methods after titanium (III) chloride reduction. The Fe2+ content was determined according to the standard of Iron ores-Determination of iron (II) content-Potassium dichromate titrimetric method. The me-tallic iron was determined according to the standard of Direct reduced iron-Determination of metallic iron content-The potassium dichromate titrimetric method after decomposition of sample by ferric. Mass fraction of carbon or sulfur was determined according to the standard of Iron ores-Determination of carbon and sulfur content-High frequency com-bustion with infrared absorption method. Mass fractions of Mg、Al、Si、P、K、Ca and Cu were determined according to the standard of Iron ores-Determination of aluminum, calcium, magnesium, manganese, phosphorus, silicon, and titanium content-Inductively coupled plasma atomic emission spectrometric method.
Point 3: On the other hand there seem to be a discrepancy between the XRD results (goethite clearly seen) and the SEM-EDS results where goethite is not identified and in contrast there appers wustite (FeO). The authores never state that wustite has oxidized to goethite.
Response : I apologize for not detailing this in Line210 and for the confusion it caused you. These batch samples are ancient slag from about 2600 years ago, buried for a long time in a damp ash pit, and the needle iron ore in the slag is actually a transformation of calcite by water erosion. However, when performing metallurgical thermodynamic analysis, needle iron ore cannot be used. Because needle iron ore is extremely dehydrated by slight heat, high temperature melting environment is needle iron ore cannot exist. In the new manuscript, we will explain this issue in detail to dispel your doubts.
Point 4: The thermodynamic analyses is performed using FactSage7.1 software but the process (what are the input values and why are they so) is not really described neither the obtained potential phase diagram (Figure 6) instead the authors dedicate almost one page (section 4.5) to merely describe results from other authors to similar studies but hardly discussing or connecting these to their own results.
Response: Fe2O3 (FeO) -SiO2-Al2O3 phase diagram or Fe2O3 (FeO) -SiO2 phase diagram is usually used to discuss ancient smelting technology, which is inappropriate. Because the oxygen potential has a great influence on the melting point of slag.
The source of ferric iron can be roughly determined by TIMA and SEM-EDS that it is from oxidized metallic iron. Therefore, we believe that Fe3+ should not be considered as a slag component in slag calculation. After normalizing other components, the components in Table 2 are obtained. This is reflected in lines 280-290.
Point 5: The paper contains quite a lot of mistakes (grammar, typos and formal register and style), English should be definitely revised.
Response 5: We have carried out a comprehensive revision and review of English. Thank you for your reminder.
Point 6: In the introduction section (page 2) the authors perform a similar exercise of unconnected sentences describing previous results from other authors that turn to be superfluous and repetitive.
Response 6: The introduction section cites a large number of previous research results. These researchers have conducted a large number of studies on ancient slags from different regions. However, the characterization methods used by these researchers could not effectively distinguish Fe2+ from Fe3+. Therefore, our study uses quantitative and highly visualized characterization methods to study ancient slag samples in depth. However, the results of previous research cannot be ignored on this basis. The inversion of possible smelting techniques through the characterization results of ancient slags is a study that tests the basic theoretical ability and is of great significance.
Point 7：The paper should be completely rethought and improved if the authors pretend to produce something worthy of publication. Please find attached an annotated version of the paper but I insist that apart from the specific comments the paper should be improved in its whole structure and possibly also in the methods applied.
Response 7: First-hand samples were obtained by conducting field surveys of ancient sites. The samples were characterized using various characterization methods. Based on the characterization results, the thermodynamic phase diagrams were drawn with the aid of the FactSage7.1 method in conjunction with the basic theory of metallurgical thermodynamics. The theoretical smelting temperature is between 1150 ℃ and 1200℃. The fayalite and aluminosilicate in the slag had obvious displacement and inhomogeneity, which pointed to the forging temperature ranging from 1050℃ to 1100℃. Mature methods should not only be promoted and used, but also explore the renewal of research methods, new research materials, and emerging technologies should be introduced. The integration of multidisciplinary methods of investigation and analysis and testing is the guarantee that the discipline of metallurgy can continue to achieve new gains.
Reviewer 2 Report
The paper presented by the authors is interesting for the scientific community.
The authors analyze from all points of view the slag from the manufacture of iron in the furnace with more than 100 years compared to the current period. Slag based on FeO, SiO2, Al2O3 differs slightly from the slag obtained today in the production of cast iron which has CaO to the detriment of Al2O3 being a basic slag.
Based on the results obtained from the slag analysis, the authors present the scenario of the technology used at that time in the production of cast iron.
The results are presented explicitly, the diagrams, figures, tables are sufficient to support the authors' conclusions.
The paper presents 21 bibliographical references, sufficient considering the nature of the research.
The similarity test gave a result of 4%, which confirms the total originality of the research and their presentation in this paper.
I consider that the paper can be published in the form presented.
Author Response
We sincerely thank the reviewer for their valuable feedback that we have used to improve the quality of our manuscript. The reviewer comments are laid out below and specific concerns have been numbered. Our response is given in red text.
Reviewer 3 Report
This is a potentially interesting piece of work. However, the analytical section needs some serious revision before it can be considered for publication.
I cannot sensibly comment on sections 1 and 2. I am a solid state chemist not a metallurgist, I can't comment on iron slags found in different parts of ancient China.
My field of expertise is X-ray Powder Diffraction and there are serious flaws in the XRD section.
Section 3.2 X-ray diffraction (lines 141 to 145) need to be revised an expanded. How many samples were analysed by XRD? How were these samples mounted? What sort of X-ray tube was used for the XRD data collection? If a Cu anode X-ray tube was used on this Fe-rich sample how were the problems of fluorescence overcome?
Section 4.1 line 169 states
The chemical composition of the samples is shown in Table 1.
How many samples, 3.1 talks of a sample (singular) but 4.1 talks of samples (plural), please be consistent on this throughout the text. If there is only one sample then say sample.
4.1 line 186 Wuestite is mentioned but in the text for Figure 5 Wustite is mentioned. Be consistent with spelling of Wuestite, either use Wustite, Wuestite or Wüstite. Give subscripts on numbers in all chemical formulae in the text, Fe2O3 not Fe2O3.
Line 206 Figure 3. Change caption to
XRD pattern for slags from Hengdong slte.
Please recheck the analyses for the powder diffraction pattern data analyses in Figure 3. The abstract mentions
XRD results show that it is mainly composed of goethite [FeO(OH)], aluminosilicate garnet [Fe3Al2 (SiO4)3] and a small amount of iron sulfide (FeS), and there is no obvious metal Fe peak.
However, Figure 3 states that the phases present are Fe2O3, Fe2SiO4, FeS and FeO(OH). None of the quoted PDF pattern numbers match those for these phases in the 2021 edition of the ICDD PDF 4+ database? This Figure needs some serious revision. Were any peaks for FeO Wüstite detected or any aluminosilicate garnet detected in this XRD analysis?
Please redo the XRD data analysis and give the correct PDF numbers for any patterns detected.
Lines 203 and 204 state
The spectral lines of some substances are dispersed, indicating that the degree of crystallization of the material is poor.
Angle-dispersive XRD IS NOT SPECTROSCOPY! The peaks that you see are Bragg reflections corresponding to planes of atoms in the crystal structures of the identified crystalline phases.
Replace spectral lines with XRD Bragg reflections.
Replace
degree of crystallization of the material is poor.
with
some of the phases present in this sample are poorly crystalline.
Section 4.3, could the Figure 4 SEM images a-h be made larger, it is difficult to read some of the text. Give more information on which parts of the sample are imaged in Figure 4.
Line 210 states
The main form of Fe in the slag is FeO.
However, this phase is not detected by XRD? Could the authors please explain this discrepancy?
I am not expert enough to comment on section 4.5
The next section is given as
6. Conclusion and significance of the study
Surely this should be section 5!
Change to
5. Conclusions and significance of the study
Line 335 change Reference to References.
Author Response
We sincerely thank the editor and all reviewers for their valuable feedback that we have used to improve the quality of our manuscript. The reviewer comments are laid out below and specific concerns have been numbered. Our response is given in red text.
Point 1: Section 3.2 X-ray diffraction (lines 141 to 145) need to be revised an expanded. How many samples were analysed by XRD?
Response 1:
A total of 10 pieces of slag were recovered from the cultural layer on the western side of the site, small in size, the largest piece being only 2-3 cm in diameter. These slags are black and irregular, magnetic and one of them has rust. The sections are mostly black in crystal-line form. From the morphological presentation of the collected samples, it can be roughly determined that the collected slags has undergone a high-temperature process and that liquid phase flow occurs during smelting. Sampling methods and testing methods, samples I will add in the newly uploaded manuscript (lines 126 to 131).
Point 2: How were these samples mounted? What sort of X-ray tube was used for the XRD data collection? If a Cu anode X-ray tube was used on this Fe-rich sample how were the problems of fluorescence overcome?
Response 2:
Thank you very much for your professional advice, which has been of great benefit to me. The radiation coming out of a sealed tube or a rotating anode is therefore a superimposition of a continuous spectrum and of characteristic radiations as presented schematically. In general, XRD methods only use the characteristic radiation with the highest intensity, the Kα radiation, and remove most of the remaining radiation by using appropriate filters or monochromators. The filtering is based on the nonlinear absorption of the filter material regarding the wavelength, leading to absorption edges. According to the filter material, the absorption edge is situated at a different wavelength allowing a strong absorption of the continuous spectrum as well as of the Kβ radiation while letting most of the Kα intensity passing through. There are appropriate filter materials for all targets.
The monochromator is a single crystal with a curved surface, adjust its installation direction so that the crystal surface on its surface is exactly diffracted with Kβ radiation, while Kα can pass through. And the radiation of other wavelengths cannot pass because they do not satisfy the single crystal diffraction condition. Therefore, it is possible to remove any wavelengths other than the characteristic rays. The X-ray diffractometer used for our inspection is equipped with a graphite monochromator and a D/teX-Ultra high speed detector, which can be used without changing the sealed tube. In the new manuscript we will detail the specific technical parameters of the X-ray diffractometer. In particular, it should be noted that although the light tube can be replaced to meet the measurement of different samples, changing the sealed tube is a time-consuming task and it is easy to damage the light tube and even the high voltage emitter.
Point 3: How many samples, 3.1 talks of a sample (singular) but 4.1 talks of samples (plural), please be consistent on this throughout the text. If there is only one sample then say sample.
Response 3:
A total of 10 pieces of slag were recovered from the cultural layer on the western side of the site, small in size, the largest piece being only 2-3 cm in diameter. These slags are black and irregular, magnetic and one of them has rust. The sections are mostly black in crystal-line form. From the morphological presentation of the collected samples, it can be roughly determined that the collected slags has undergone a high-temperature process and that liquid phase flow occurs during smelting.
Point 4: 4.1 line 186 Wuestite is mentioned but in the text for Figure 5 Wustite is mentioned. Be consistent with spelling of Wuestite, either use Wustite, Wuestite or Wüstite. Give subscripts on numbers in all chemical formulae in the text, Fe2O3 not Fe2O3.
Response 4:
I apologize for the inconsistency in the spelling of proper nouns. In the new manuscript, Wustite is used consistently, and I will take extra care to correct the numerical subscript of the chemical formula Fe2O3 in the manuscript.
Point 5: Line 206 Figure 3. Change caption to XRD pattern for slags from Hengdong slte. Please recheck the analyses for the powder diffraction pattern data analyses in Figure 3. The abstract mentions XRD results show that it is mainly composed of goethite [FeO(OH)], aluminosilicate garnet [Fe3Al2(SiO4)3] and a small amount of iron sulfide (FeS), and there is no obvious metal Fe peak.
Response 5:
I will revise Figure3 to mark all the substances that can be retrieved and present them in the abstract and the text.
Point 6: However, Figure 3 states that the phases present are Fe2O3, Fe2SiO4, FeS and FeO(OH). None of the quoted PDF pattern numbers match those for these phases in the 2021 edition of the ICDD PDF 4+ database? This Figure needs some serious revision. Were any peaks for FeO Wüstite detected or any aluminosilicate garnet detected in this XRD analysis? Please redo the XRD data analysis and give the correct PDF numbers for any patterns detected.
Response 6:
Thank you very much for your suggestion, the previous database is indeed too old. The PDF-2 2016 version is the most up-to-date and comprehensive database we could find so far. We will definitely take care to use the latest ICDD database in our future studies.
Point 7: Lines 203 and 204 state. The spectral lines of some substances are dispersed, indicating that the degree of crystallization of the material is poor. Angle-dispersive XRD IS NOT SPECTROSCOPY! The peaks that you see are Bragg reflections corresponding to planes of atoms in the crystal structures of the identified crystalline phases. Replace spectral lines with XRD Bragg reflections.Replace degree of crystallization of the material is poor with some of the phases present in this sample are poorly crystalline.
Response 7: The comments you made are very helpful to us. I have used diffraction peaks instead of spectral lines(Lines 233 to Lines 239). In my future research, I will take extra care to use accurate and professional vocabulary.
Point 8: Section 4.3, could the Figure 4 SEM images a-h be made larger, it is difficult to read some of the text. Give more information on which parts of the sample are imaged in Figure 4.
Response 8: Sorry about that. In the new manuscript, the backscatter map will be resized to the right size.
Point 9: Line 210 states The main form of Fe in the slag is FeO. However, this phase is not detected by XRD? Could the authors please explain this discrepancy?
Response 9: These samples are ancient slag from about 2600 years ago, buried for a long time in a damp ash pit, and Goethite ore in the slag is actually a transformation of Wustite by water erosion. However, when performing metallurgical thermodynamic analysis, Goethite cannot be used. Because Goethite is extremely dehydrated by slight heat, high temperature melting environment is Goethite cannot exist. In the new manuscript, we will explain this issue in detail to dispel your doubts.
Round 2
Reviewer 1 Report
The authors have not restructured the paper as suggested, they have just applied minor changes but the main concerns (a better explanation of the production of the phase diagrams and the obtained theoretical values of smelting and forging temperature) should be amended before reconsidering publication.
Other comments:
on point 2: Please add references for the standard methods or the corresponding ISO references. Also, tritation and ICP-AES methods should contitute separate subsections within the correspoding section.
on point 3: I do not understand, goethite is a transformation of calcite by wáter-erosion? What does this mean?
on point 4: My question remains unanswered,the authors present in Figure 6 a T-oxygen potential phase diagram that they claim that corresponds to the Hengdong site but they never show how this was computed, what were the input values and why they are so.
The authors should explain better and show clearly what is the connection between table 1 values and table 2 values. Also, the authors should show the statistical variation of measured values (mean values and standard deviations).
on point 6: my comment was not intended to supress the cites of previous works, it is rather on making a text integrating all this references on a text explaining the similarities and differences in approaches and results obtained in previous works and why it is so crutial the iron II and iron III quantification. Also why this results in less accurate conclusions and how this is differently treated in the present paper.
Instead, the authors cite the different works in unconnected sentences using both present and past verb tenses that makes the text rather awkward. Besides, to me, it is still obscur what is the treatment that the authors apply for iron III, do they convent the measured amount of Fe III into the equivalent quantity of metallic Fe? Or do they simply ignore the amount of measured iron III? They talk about a normalization of the iron values but I do not see what do they actually do.
Please find also a PDF with additional comments on the PDF revised version
Comments for author File: Comments.pdf
Author Response
We sincerely thank the reviewer for thoroughly examining our manuscript and providing very helpful comments to guide our revision. Sincere thanks should be given to the reviewer for the constructive comments and suggestions. The responses to the comments are given below.
1.The authors have not restructured the paper as suggested, they have just applied minor changes but the main concerns (a better explanation of the production of the phase diagrams and the obtained theoretical values of smelting and forging temperature) should be amended before reconsidering publication.
Response
When the high-temperature reduction process occurs, the reduction of Fe2O3 takes place in stages, viz:
Fe2O3→Fe3O4→FeO→Fe.
We can understand the reduction process as a process of iron oxide losing oxygen.
C+O2=CO (1)
3Fe2O3(s)+CO=2Fe3O4(s)+CO2 (2)
Fe3O4(s)+CO=3FeO(s)+CO2 (3)
FeO(s)+CO=Fe+CO2 (4)
reaction (4) Can be regarded as FeO=Fe+0.5O2 (5)
So we choose the oxygen potential(PO2) of the slag as the Abscissa of the calculation. We can judge the direction in which the reaction(5) occurs based on the plotted slag oxygen potential phase diagram. In this way, we can refer to the sulfur-oxygen potential diagram in the copper smelting theory.（https://doi.org/10.1179/cmq.1974.13.3.443）
The starting phase is nearly always hematite (Fe2O3), which very easily reduces to magnetite. Magnetite (Fe3O4) reduces to wustite (FeO), which has a comparatively large phase field before reducing to metallic iron.
The composition of slag samples has a forward reaction trend under high temperatures. Our labeling of the Smelting Area in the figure suggests that within this range the reaction reaches equilibrium or may continue upwards as the oxygen potential decreases, further generating metallic iron. The calculation reference values are in Table 2. Table 2 is obtained by conversion from Table 1, ignoring the influence of P and S, and normalizing after removing Fe3+. Ironmaking is a chemical reaction process with a strong reducing atmosphere. For the characterization of slag in reducing atmosphere, the Fe3+ content should not be considered. In the process of reduction, if iron can be formed, the possibility of the existence of Fe2O3 is low, and the source of trivalent iron is caused by the oxidation of unseparated metal iron in the long-term oxidation process. In the process of calculation, Fe2O3 should not be calculated as part of the slag. We used the Equilib module calculation in Factsage7.1, and selected databases including Ftoxid, Ftmisc, and FTPS.
on point 2: Please add references for the standard methods or the corresponding ISO references. Also, tritation and ICP-AES methods should contitute separate subsections within the correspoding section.
Response
The content detection method strictly complies with National Standard of the People’s Republic of China.
The total iron content was determined according to the standard of Iron ores-Determination of total iron content-Titrimetric methods after titanium(III) chloride reduction. Determination range (mass percentage): 25%-72%.(ICS:73.060.10, GB/T 6730.65-2009)
The Fe2+ content was determined according to the standard of Iron ores-Determination of iron (II) content-Potassium dichromate titrimetric method. Determination range (mass percentage): 0.700%-30.00%.
(ICS: 73.060.10, GB/T 6730.8-2016 )
The metallic iron was determined according to the standard of Direct reduced iron-Determination of metallic iron content-The potassium dichromate titrimetric method after decomposition of the sample by ferric chloride. Determination range (mass percentage):0.150%-3.00%. (ICS:73.060.10, GB/T 6730.6-2016)
Mass fraction of carbon or sulfur was determined according to the standard of Iron ores-Determination of carbon and sulfur content-High frequency combustion with infrared absorption method. Determination range (mass percentage):carbon 0.01%-2.5%, sulfur 0.001%-2.00%. (ICS: 73.060.10, GB/T 6730.61-2005)
Mass fractions of Na, Cu and K: Iron ores-Determination of potassium, sodium, vanadium, copper, zinc, lead, chromium, nickel, cobalt elements-Inductively coupled plasma optical emission spectrometry. Determination range (mass percentage):Na 0.02%-0.4%, Cu 0.001%-0.4%, K 0.003%-0.4%.
(ICS: 73.060.10, GB/T 6730.76-2017)
Mass fractions of Mg, Al, Ca, P, Si: Iron ores-Determination of aluminum, calcium, magnesium, manganese, phosphorus, silicon and titanium content-Inductively coupled plasma atomic emission spectrometric method. Determination range (mass percentage):Mg 0.010%-3.00%, Al 0.020%-5.00%, Ca 0.010%-8.00%, P 0.013%-2.00%, Si 0.10%-8.00%.
(ICS:73.060.10, GB/T 6730.63-2006)
on point 3: I do not understand, goethite is a transformation of calcite by wáter-erosion? What does this mean?
Response
We have added references to illustrate the possible corrosion mechanisms. Thanks for the references, which are now included in the revised manuscript. Specific references are listed as follows:
Iron speciation in blast furnace slag cements (https://doi.org/10.1016/j.cemconres.2020.106287)
Corrosion under oxic conditions can be described by the following equation:
Fe(O)+0.75O2+0.5H2O=FeOOH
on point 4: My question remains unanswered, the authors present in Figure 6 a T-oxygen potential phase diagram that they claim that corresponds to the Hengdong site but they never show how this was computed, what were the input values and why they are so.
Response
The composition of slag samples has a forward reaction trend under high temperatures. Our labeling of the Smelting Area in the figure suggests that within this range the reaction reaches equilibrium or may continue upwards as the oxygen potential decreases, further generating metallic iron. The calculation reference values are in Table 2. Table 2 is obtained by conversion from Table 1, ignoring the influence of P and S, and normalizing after removing Fe3+. Ironmaking is a chemical reaction process with a strong reducing atmosphere. For the characterization of slag in reducing atmosphere, the Fe3+ content should not be considered. In the process of reduction, if iron can be formed, the possibility of the existence of Fe2O3 is low, and the source of trivalent iron is caused by the oxidation of unseparated metal iron in the long-term oxidation process. In the process of calculation, Fe2O3 should not be calculated as part of the slag. We used the Equilib module calculation in Factsage7.1, and selected databases including Ftoxid, Ftmisc, and FTPS.
The authors should explain better and show clearly what is the connection between table 1 values and table 2 values. Also, the authors should show the statistical variation of measured values (mean values and standard deviations).
Response
Subtract the trivalent iron content from the total iron content, select a suitable slag system, and normalize the measured substances.
on point 6: my comment was not intended to supress the cites of previous works, it is rather on making a text integrating all this references on a text explaining the similarities and differences in approaches and results obtained in previous works and why it is so crutial the iron II and iron III quantification. Also why this results in less accurate conclusions and how this is differently treated in the present paper.
Instead, the authors cite the different works in unconnected sentences using both present and past verb tenses that makes the text rather awkward. Besides, to me, it is still obscur what is the treatment that the authors apply for iron III, do they convent the measured amount of Fe III into the equivalent quantity of metallic Fe? Or do they simply ignore the amount of measured iron III? They talk about a normalization of the iron values but I do not see what do they actually do.
Response
For the calculation of thermodynamic properties of slag samples, the composition of slag samples should be studied quantitatively.
Ironmaking is a chemical reaction process with a strong reducing atmosphere. For the characterization of slag in reducing atmosphere, the Fe3+ content should not be considered.
Using charcoal in the reduction of iron oxides, bloomery iron would show different degrees of oxidation during cooling. This is because, bloomery iron is obtained from the direct reduction of solid iron ore by reductive gas, after the reduction reaction is finished, a large amount of honeycomb structure will be formed, and the specific surface area is large, making it active and can easily be secondarily oxidized, forming ferric iron.
（Dutta S K, Chokshi Y B. Basic concepts of Iron and steel making[M]. Springer Nature, 2020.）
In previous studies, without taking this into account, XRF and EDS techniques cannot distinguish the valence state of Fe and quantitative determination of ferric iron is impossible. So the FeO measured before is the total amount of FeO and Fe2O3 and even contains the mass of metallic iron.
In the reduction stage, there is no trivalent iron. However, the existence of trivalent iron is indeed detected in the slag, so according to the characteristics of bloomery iron, we think that it is caused by the re-oxidation of the metal iron left in the slag after long-term stacking in the humid environment. In the analysis of the reduction reaction, the part of re-oxidation should not be considered and the trivalent iron should be removed.
Reviewer 3 Report
Thanks for the comments and revisions.
Response 1: I am happy with the response.
Response 2: None of the questions were answered!
Point 2: How were these samples mounted? What sort of X-ray tube was used for the XRD data collection? If a Cu anode X-ray tube was used on this Fe-rich sample how were the problems of fluorescence overcome?
Please answer the questions from point 2.
Response 3: I am happy with the response
Response 4. Line 224 has Fe2O3 (with subscripts) and Fe2O3 (no subscripts). Please ensure that all chemical formulae have subscripts where appropriate.
Response 5: Thank you for changing the Figure 3 caption. The authors talk of test results and the presence of aluminosilicate garnet, and [Fe3Al2(SiO4)3]. However, these phases are not detected by XRD? Please explain why these phases are not detected by XRD?
Response 6: The quoted PDF numbers for the phases in Figure 3 now match my version of the PDF database.
Response 7: You talk about lines in the spectrum at lines 208 and 214! XRD IS NOT SPECTROSCOPY. Please change this section.
Response 8. You gives Figures 4a-h as SEM photos. Please indicate which parts of the samples correspond to each of the 8 SEM images in Figure 4.
Respone 9.
Point 9: Line 210 states The main form of Fe in the slag is FeO. However, this phase is not detected by XRD? Could the authors please explain this discrepancy?
The authors should make it more clear that FeO transforms to FeOOH and why FeO is not observed by XRD.
Author Response
We sincerely thank the reviewer for thoroughly examining our manuscript and providing very helpful comments to guide our revision. Sincere thanks should be given to the reviewer for the constructive comments and suggestions. The responses to the comments are given below.
Response 1: I am happy with the response.
I'm glad to have your approval.
Point 2: How were these samples mounted? What sort of X-ray tube was used for the XRD data collection? If a Cu anode X-ray tube was used on this Fe-rich sample how were the problems of fluorescence overcome?
Response 2
(1) θ/θ goniometer , where the sample is fixed , while the X-ray source as well as the detector moves.
(2) Ceramic X-ray tube Cu-Kα radiation.
(3) Secondary optics are devices placed between the sample and the detector and are used to define the diffracted beam. Several components described for the primary optics can also be used as secondary optics. Systems of slits are generally used to reduce the divergence of the beam and lead to narrow peaks but with intensity lost. As well, Soller slits can be used to reduce the beam divergence and achieve higher angular resolution.
Secondary monochromators can be used in order to remove undesirable wavelengths such as Kβ radiation or fluorescent radiation (for example, in the case of Cu-Kα radiation used for the investigation of Fe-based material). Filters to reduce the Kβ radiation can generally be placed in front of the detector.
Response 3: I am happy with the response
I'm glad to have your approval.
Response 4. Line 224 has Fe2O3 (with subscripts) and Fe2O3 (no subscripts). Please ensure that all chemical formulae have subscripts where appropriate.
Response 4
I apologize for the inconsistency in the spelling of proper nouns. I will take extra care to correct the numerical subscript of the chemical formula Fe2O3 in the manuscript.
Response 5: Thank you for changing the Figure 3 caption. The authors talk of test results and the presence of aluminosilicate garnet, and [Fe3Al2(SiO4)3]. However, these phases are not detected by XRD? Please explain why these phases are not detected by XRD?
Response 5
The test results show that it is mainly composed of goethite [FeO(OH)], sillimanite (Al2SiO5), aluminum iron (AlFeO3) and a small amount of iron sulfide(FeS).
(line 214 to line 216)
Response 6: The quoted PDF numbers for the phases in Figure 3 now match my version of the PDF database.
I'm glad to have your approval.
Response 7: You talk about lines in the spectrum at lines 208 and 214! XRD IS NOT SPECTROSCOPY. Please change this section.
Response 7
I have made changes in the manuscripts. We will use XRD diffraction pattern instead of Erroneous expression. (line 216 to line 218)
Response 8. You gives Figures 4a-h as SEM photos. Please indicate which parts of the samples correspond to each of the 8 SEM images in Figure 4.
Response 8
It clearly shows a brighter wustite and a darker spinel phase though both of these phases are chemically dominated by Fe and O. It is mainly because the average atomic number of wustite (Fe:O =1:1) is higher than spinel (Fe:O =3:4). We made additional explanations in lines 235 to 246.
Point 9: Line 210 states The main form of Fe in the slag is FeO. However, this phase is not detected by XRD? Could the authors please explain this discrepancy?
Response 9
We have added references to illustrate the possible corrosion mechanisms. The references, which are now included in the revised manuscript. Specific references are listed as follows:
Iron speciation in blast furnace slag cements (https://doi.org/10.1016/j.cemconres.2020.106287)
Corrosion under oxic conditions can be described by the following equation:
Fe(O)+0.75O2+0.5H2O=FeOOH
