Design of ESR Slag for Remelting 9CrMoCoB Steel under Simple Protective Ar Gas

Round 1

Reviewer 1 Report

Dear Authors,

Your paper Design of ESR slag for remelting 9CrMoCoB steel under simple protective Ar gas discusses about the design of a suitable slag to reduce energy consumption during electro-slag remelting of alloyed steel, maximizing the yield of B.

Even the topic is hot and worthy to be published, the paper has serious flaws. For instance, some of the reactions considered are not thermodynamically possible whereas some of the results are contradictorily.

Experimental section is a bit confusing, but acceptable. Results and discussion section is well organized but the discussion is poor in comparison with the literature. Some details about slag and steel requirements are omitted but are fundamental to understand the choices made by the Authors.

Conclusions are only in part supported by the results

For this reason, I cannot recommend the paper for publication

Abstract, line 11: please replace basic with alkaline

Experimental procedure

Reaction 1. I am not convinced about the feasibility of this reaction. Si-SiO₂ system to act as reductor against Al₂O₃, should have an oxygen potential lower than Al-Al₂O₃ However, in the Ellingham diagram is exactly the opposite. In fact, metallic Al generally acts as reductant of SiO₂ in Ladle steelmaking processing, causing an undesired Si pick-up and decreasing the deoxidation efficiency. Could the Authors support their choice with appropriate literature? Reaction 2. Same issue with the B capacity to reduce SiO₂. B-B₂O₃ oxygen potential is higher than Si-SiO₂. Thus, SiO₂ reduction by B is not thermodynamically possible. Please check and support this hypothesis with appropriate references? In my opinion the sequence of reactions taking place during the slag crossing is

3[Si] + 2(B₂O₃) = 3(SiO₂) + 4[B]

2[Al] + (B₂O₃) = (Al₂O₃) + 2[B]

4[Al] + 3(SiO₂) = 2(Al₂O₃) + 3[Si]

This seems demonstrated by your simulation results. B₂O₃ in slag is reduced both by Si and Al in the steel. This is the reason why B increase when B₂O₃ increase in the slag

Al is consumed for SiO₂ and B₂O3 reduction. In fact, it decreases by increases B₂O₃

Si is consumed for B₂O₃ reduction. However, part of the Si is brought back due to SiO₂ reduction by Al. This can lead to an equilibrium and depends upon the different free energy of the two involved reactions.

Section 2.1. “The liquid phase regions at 1300~1400℃ of the CaF₂–CaO–Al₂O₃ with 2%MgO or 3%MgO addition were both suitable for ESR process”. Why the Authors assessed this conclusion? Which characteristics should posses the slag to act as best on the molten pool? In my opinion slag must liquid to refine as much as possible the steel drops passing though. How it should be the slag in terms of viscosity? Fully liquid, creamy? Probably it would be better choosing a chemical composition at 2% MgO that assure a fully liquid slag at 1250 °C, instead of at 3%MgO that reduce the chemical composition combinations to melts the slag at so low temperature. This depends, of course, by the desired slag viscosity. Please explain first which are the main features of a “perfect” ESR slag and then describe your choice about the MgO concentration. Section 2.2.2, line 115-116: what about the SiO₂ in the synthetic slag formulation? Between experiment for equilibrium reaction and lab-scale experiment, why in the design of equilibrium reaction the oxygen was not set as 0.5%? In this way the two different experimental conditions would be performed in a comparable way. Between experiment for equilibrium reaction and lab-scale experiment, is the slag/steel ratio maintained at the same value? This is important to correctly transfer the results of equilibrium reaction to the lab-scale test Line 154-155; “To simplify the calculation, the mass fraction of the crystalline phase was instead of the volume fraction [23]”. How the equation of Roscoe should be modified? Or it was used a different alfa coefficient? Please detail how to simplify the viscosity calculation

Results and discussion

Line 186: 55%CaF₂–20%CaO–3%MgO–22%Al₂O₃–x%SiO₂–y%B₂O₃ (x ≤ 3, y ≤ 3): why exactly this threshold for SiO₂ and B₂O₃ in the slag. Please add a comment or a reference about this. Figure 6: B, Si and Al are in the electrode or in the molten steel below the slag? How it can be that in the initial conditions (B₂O₃ = 0% and SiO₂ = 0%) the concentration of B, Si and Al (even if the electrode or in the molten pool) are different from the chemical composition of the electrode? In particular, how is possible that Al is 0.018 when the initial value is 0.004? Figure 6: what the blue and red lines represent in the B-B₂O₃ graph? Figure 7: B, Si and Al are in the electrode or in the molten steel below the slag? Why the Al behavior is the opposite with respect the simulation? Again, why the values of Al differ so much between measured and Table 2 at the starting condition? Line 212-219: In my opinion the reaching equilibrium in silicon is not due to the FeO, but when the SiO₂ in the slag is enriched, Si oxidation from the bath is annihilated. I think the reaction takes place when steel pass through the slag is

[Si] + (B₂O₃) = (SiO₂) + [B] where () is slag, [B] is bath

If SiO₂ concentration in slag is 0%, a strong gradient between bath and slag exists. Thus, the most of SiO₂ is moved in the slag. When SiO₂ increase in the slag, the concentration gradient is reduced. Thus, the kinetics is lowered, giving time to [Al] to reduce [SiO₂] (imagine this is the oxidation production formed in the bath but not yet moved to slag) and bring back [Si] to the bath.

If the O was the responsible of Si oxidation, Si would oxidize firstly with respect iron, because Si oxidation protects Fe oxidation.

Please check this part and support your discussion with appropriate literature

Figure 11: please correct the x-axis of (c) figure. It was written B₂O₃ instead of length Line 260-272: Before discussing the results obtained from lab-scale ESR, it must be indicated which are the suitable ranges for alloying elements in the steel under examination. For instance, in Table 1 it would be better to include min, max and average value of the steel chemical composition. In this way it will be easy to understand why the slag #3 and #4 were chosen Conclusions: the conclusions (2) and (3) are not supported by the results, since the Al behavior is opposite in the experimental trials with respect the simulation. Please, adapt these conclusions after the correction of results section

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

In general a good high quality manuscript. The experimental work is very much appreciated. Some suggested minor changes and additions could contribute to an even better manuscript.

1) In some parts of the manuscript the English could be improved, especially the syntax of some sentences require some improvement. This would help to increase readability of the manuscript.

2) Line 83: it should be written 25°C interval and not 50°C

3) Line 93: 1250~1400℃ isothermal section is written, but the figure shows the region 1200 - 1400°C

4) Line 115/116: the powders are first mixed and then premelted. Please check the order of the two sentences.

5) Line 117: it should be graphite crucible and not graphic crucible

8) Was the chemical composition analysed after slag premelting? The slag composition can change quite a lot during the premelting process, though e.g. vaporization of some elements. If so, it could be good to state the composition of the mixed and premelted slags in a table.

9) Line 131: How was the O2 concentration in the furnace measured? How can you be sure that the oxygen content in the furnace was not higher as stated? Maybe some error estimation could be included in the manuscript.

10) Line 144: Some of the viscosity models in FactSage give results far away from reality. Did you compare with other publications and literature, so at least the order of magnitude is the same for the calculated viscosity and the viscosities stated in literature? The stated calculated viscosities in the result section for slag #2 – #4 are quite similar. I have my doubts that FactSage can predict the slag viscosities very accurately. Maybe some slag measurements could be contacted in further work.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

The reviewer has following comments and questions which should be considered in the revised version of manuscript.

line 75: ThermoCalc was used to calculate liquidus was steel. why FactSage wasn't used, while all other calculations are carried out by Factsage? Line 77: what database were use? Line 116: Purity level of SiO2 is not mentioned.  Section 2.2.2 presents slag preparation method and Table 4 gives composition of those slags. However, its not clear whether compositions in Table 4 were measured values or target. Its also not mentioned if slag composition was ever measured. How much was variation in target and measured values?  Line 136 -142: it says several samples were taken, but schematic of furnace (Fig 3) doesn't help to determine sampling position.  Line 137: How was slag added?  Line 207 and Figure 7 (c):  The Al content increased and became stable after 40 minutes. It is important to mention the error in the Al content measurement as the change in Al with time is very small (~ 0.001). Figure 7 (c): The X-axis title is not correct. Figure 8: The colouring scheme is not consistent. What is the purpose to present Equation 4 since it’s not being used to calculate viscosity of slags. Line 234: … ‘‘generally kept steady for the four slags’’. It should be ‘three slags’. Figure 10: It is suggested to add scale bar on the images to give idea about the size of ingots. Figure 11 (c): Title of X-axis is wrong. Figure 11 (a): What is the reason of relatively lower Si levels as compared to those shown in Figure 7? Also, what causes Si loss in slag2 ingot? It has been emphasized that slag compositions were selected based upon thermodynamic calculations.( for instance Line 111 and Line 160) However, it is not very clear that what was the criteria for designing the slag compositions.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Dear Authors,

thank you for your kind reply to my comment.

To improve the quality of your paper, please add some of your answers to my questions directly in the manuscript

About point 2 in the previous review round, please add your response (modified opportunely) to explain why equation (1) occurs in ESR slag: “The reaction 3[Si] + 2(Al2O3) = 3(SiO2) + 4[Al] did occur, though, the reaction seems opposite to the Ellingham diagram. The standard state represented in the Ellingham diagram is different than the actual state. For instance, the Al2O3 content in the slag is high and the Al content in the steel is relatively low, while the SiO2 content in the slag is relatively low and the Si content in the steel is relatively high. Thus, the ΔG of this reaction is <0 during the experiments”. You can include a thermodynamic simulation, showing the equilibrium at the actual conditions. Please, also include the references reported in the response file When discussing the figure 7, please add the second part of the response (opportunely modified) given at point 2 of previous review round: The reaction 4[Al] + 3(SiO2) = 2(Al2O3) + 3[Si] maybe did occur during the experiments. But in Figure 7, the Al content increased with the reaction time in the four experiments. Thus, even the reaction 4[Al] + 3(SiO2) = 2(Al2O3) + 3[Si] may occur in some moments, the reaction 3[Si] + 2(Al2O3) = 3(SiO2) + 4[Al] is still the main reaction which led the Al content to increasing with reaction time. The reaction 2[Al] + (B2O3) = (Al2O3) + 2[B] maybe didn't occur”. Include also the references used in the response to my comments When discussing the results of factsage in Figure 1 and 2, please add your comment, and relative reference, to support your statement: “As reported by Zhengbang Li, the liquidus temperature of the molten slag should be lower than the liquidus temperature of steel by 100~200℃ (for the sake of reducing the energy consumption and the good surface quality of the ingot). The liquidus temperature of the 9CrMoCoB steel is around 1500℃ thus, the proper liquidus temperature of the molten slag was 1300~1400℃.” please, include in the text, when discussing the laboratory ESR results. In my opinion is very important to highlight that the behavior of the refining slag does not depend by the mass but by the equilibrium. Please include also the reference you listed in the response file Table 4. The values are measured or calculated by factsage? Please specify in the caption. What does it mean pre-melted? Are the Authors meaning before and after contact with steel? Please clarify this point. Add also some comments with respect the variation of elements in the slag. Are they in agreement with the variation of elements in the steel? Paragraph 3.1: in order to understand the meaning of the numbers in Figure 6, please include your response at point 9. In particular: “Figure 6 is the calculated results of the equilibrium reaction between the steel and the molten slag of 55%CaF2–20%CaO–3%MgO–22%Al2O3–x%SiO2–y%B2O3 by using Factsage 7.3. Thus, the B, Si, and Al contents in Figure 6 was the equilibrium content in the molten steel below the molten slag, after reaction with the molten slag with varied SiO2 and B2O3 contents. When the SiO2 and B2O3 contents in the basic slag of 55%CaF2–20%CaO–3%MgO–22%Al2O3 were zero, the reaction between the molten slag and steel still occurred. The reaction 3[Si] + 2(Al2O3) = 3(SiO2) + 4[Al] led to the increasement of Al and loss of Si. As the experiments were not conducted in a fully sealed furnace, the FeO existed in the slag, and the reaction 2[B] + 3(FeO) = (B2O3) + 3[Fe] led to the loss of B. Thus, the concentration of B, Si and Al are different from the initial chemical composition of the steel. When the SiO2 content in the molten slag is 0, this trigger the reaction 3[Si] + 2(Al2O3) = 3(SiO2) + 4[Al]. For this reason, the increasement of Al is much (even up to 0.018%), when the SiO2 and B2O3 contents in the basic slag of 55%CaF2–20%CaO–3%MgO–22%Al2O3 were zero”. Include proper reference to support your discussion Please check the number of paragraph 3.4. It is 3.2 paragraph 3.4 (Results of equilibrium reaction experiments): regarding Figure 7, it is better to include in the text: “In figure 7, the B, Si, and Al contents changed with the reaction time, not the equilibrium content of the steel, so the Al behavior is the not the same with respect the simulation.” I still cannot understand why the result shown in Fig 6 and 7 are different in trend if they represent the main phenomenon. I would expect that the evolution of B, Al and Si by equilibrium reaction is the same in both calculated and experimental scenario. On the contrary, there is something wrong in the experimental setup, or the simulation is not in good agreement with the real conditions. Please try to explain why B and Al show opposite trend if the conditions should be the same Figure 7(d). Could the Authors specify at which condition the values refer? Are the final values of curves in Figure 6? Or in Figure 7? Or different source? I cannot understand the source of such numbers In the same paragraph, add also: “in the equilibrium reaction experiments, the molten slag will react with the steel. Because of the reaction 3[Si] + 2(Al2O3) = 3(SiO2) + 4[Al] was not in equilibrium state, the Al content changed compared to the initial content.” Add (modified opportunely) your response to point 12 in the discussion of figure 7: “the B content in the steel kept steady with the reaction time increasing, while the Si content reduced slightly with the reaction time increasing. In our opinion, the reaction [Si] + (B2O3) = (SiO2) + [B] was in equilibrium state during the experiments, because…(add proper explanation). The reaction 3[Si] + 2(Al2O3) = 3(SiO2) + 4[Al] occurred, for the Al content increased with the reaction time increasing, but the consumption of Si caused by the reaction between the Si and Al2O3 was very limited, so the loss of Si was not only caused by the reaction 3[Si] + 2(Al2O3) = 3(SiO2) + 4[Al]. Because of the equilibrium reaction experiment was not conducted in the fully sealed furnace, atmospheric oxygen can enter in the furnace, and the reaction between Si and FeO occurred led to the loss of Si. With the SiO2 content in the slag increasing, the reaction between the Si and FeO was annihilated and began to keep steady.” About the slag viscosity calculation, it would be better to declare that the h₀ calculated by factsage is based on mass fraction and not volume fraction of the liquid phase of slag. In this way, the change from volume to mass of the solid fraction it will be congruent

Best Regards

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Thanks for your detailed response to my comments and making changing in the manuscript accordingly. The updated version reads well. 

Author Response

Dear editor and reviewer:

          Thank you for your kind suggestions. The new upload article has been revised carefully.

           Best wishes for you.

 

Round 3

Reviewer 1 Report

Dear Authors,

thank you for your precise revision

 

After reading your answers, the most of points were clarified.

 

I have just some minor suggestions to include in the final version of your paper. These are due allowing all the readers to fully understand your work

 

1) Equation 1. Please include exactly what you reported in the response file (free energy in not standard state). In this way it will be perfectly clear that the reaction 1 can happen spontaneously because the activity of Si and Al are different from the standard state. Include also the hypothesis: “the Al content in the molten steel is zero, and the Si in the molten steel is high enough, while the Al2O3 content in the slag is high, and the SiO2 content in the slag was zero”

 

After this change your paper can be published in Metals journal

 

Thank you and best regards

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf