Deoxidation of Electroslag Remelting (ESR) – A Review

Abstract

Deoxidation of liquid steel and alloy during electroslag remelting (ESR) is always an ongoing concern for producing advanced clean steel and alloy. The increasing demands for more excellent performance of steel urge metallurgists to further improve the steel cleanliness. Lowering the oxygen content and non-metallic inclusions amount during the ESR process is of great importance as ESR is the last processing procedure for refining liquid steel in the steel product manufacturing process. Deoxidation of ESR is dependent on one or more aspects including the initial oxygen content and oxide inclusions in the electrode, remelting atmosphere, deoxidation schemes, slag compositions, reoxidation degree, melting rate and filling ratio. This paper reviews the state of the art in the deoxidation of ESR and deoxidation-related oxide inclusions. The oxygen transfer behavior during ESR process is described first. Deoxidation of liquid steel during ESR is discussed based on the thermodynamic and kinetic considerations. The dependence of the oxygen on the processing parameters of ESR is reviewed and discussed. The influence of these parameters on the oxide inclusions associated with the deoxidation of ESR is also assessed. The suggestions for the future work are proposed in this article.

1. Introduction

Electroslag remelting (ESR) is widely used to produce some varieties of special steels and alloys, mainly because of its ability to provide high cleanliness level and excellent homogeneity of solidified ingot structure (reducing segregation, shrink holes, etc.) simultaneously.^1,2,3) Deoxidation of liquid steel and alloy is always an ongoing concern for eliminating the detriments of oxide inclusions to the processing and mechanical properties of the steel and alloy. Great efforts have been put forward to minimize the amount of oxide inclusions by decreasing the oxygen content of the steel during ESR process.^4,5,6,7,8,9)

Deoxidation of liquid steel during ESR is basically different from that in other steelmaking process operations. The iron oxide activity of the slag is a measure of its oxygen potential during ESR.^4,5,9,10) The intention of deoxidizing agent addition in ESR process is to deoxidize the molten slag.^{4,5,6,7,8,9,11,12)} In this process, the oxygen level of liquid steel is determined by the interactions of atmosphere-slag-metal-inclusion. Deoxidation during ESR is therefore affected by multiple factors, such as alloy compositions, deoxidizing agents, absorption ability of slag to oxide inclusions, remelting atmosphere, slag compositions, and oxide inclusion evolution. Some of these factors could only be varied in a limited range in the ESR practical production. As for a specific ESR practice, the contribution ratio of each of these factors is different, namely major, minor or negligible role. Lowering the oxygen potential of slag could not only reduce the oxygen introduction from the slag to liquid metal, but also suppress the loss of alloying elements, such as Ti, Al, Si and Mn, in liquid metal during the ESR.^6,11,13)

Deoxidation of ESR comes down to the removal degree of oxide inclusions in liquid metal.¹⁴⁾ The effects of the abovementioned factors, which influence the deoxidation of ESR, on the oxide inclusions therefore should also be addressed because different types of inclusions basically experience various evolution trajectories during the ESR, as demonstrated in the previous studies regarding Al₂O₃, MgO·Al₂O₃, nitride, sulfide, and calcium aluminate inclusions.^{14,15,16,17,18,19)} However, only a few works regarding the correlation of deoxidation and oxide inclusions have been reported. In view of the significance of decreasing oxygen level, the complexity of multiple factors affecting the deoxidation, and infeasibility in on-line monitor during on-going ESR, it is quite important to summarize the existing works on the deoxidation and associated oxide inclusions during ESR for the purpose of assisting future work. Electroslag remelting is generally operated using high frequency alternating current in production practices around the world so far. Although low frequencies of alternating current (up to <5 Hz) and direct current of ESR exert apparent influences on the contents of oxygen and non-metallic inclusions in steel according to laboratory-scale ESR trials,²⁰⁾ the effect of power supply modes on the deoxidation of ESR is not summarized in this article.

The present article reviews the deoxidation findings of ESR over the past decades, which have been accomplished by the research groups around the world. The oxygen transfer behavior, and thermodynamic and kinetics considerations on the deoxidation of ESR are described and the underlying mechanisms are discussed. Next, the crucial aspects, which are commonly focused on the deoxidation of ESR, are presented and discussed. In view of the indivisible correlation of deoxidation with oxide inclusions, the role of deoxidation operations of ESR on the oxide inclusions is also evaluated. Finally, a general concluding remark and perspective for future work are present.

2. Oxygen Transfer during ESR Process

The oxygen in the ESR process arises from the following sources: (1) The original oxygen in the consumable electrode. (2) The oxide scale on the electrode steel surface. (3) The oxygen in the air atmosphere, which affects the oxygen content of liquid steel in two different ways: (i) atmospheric oxygen permeates directly through molten slag into liquid metal pool by diffusion of physically dissolved oxygen. It should be mentioned on the basis of experimentally measured permeability that the permeation of oxygen as O₂, which physically dissolves in molten slag, plays a minor role in oxygen transfer,^21,22,23) (ii) atmospheric oxygen reacts with the steel electrode at the electrode-atmosphere surface to form iron oxide as the electrode is heated at elevated temperatures. (4) Reducible oxides in the slag. (5) The moisture in the atmosphere and in the slag reacts with O²⁻ in the slag to form (OH⁻), which thereafter introduces oxygen into liquid steel.²⁴⁾ (6) The dissolved oxygen as the products of desulfurization reactions in the ESR process. A schematic illustration of the oxygen transfer in the ESR process is shown in Fig. 1.

Fig. 1.

Schematic illustration of oxygen transfer in the ESR process. Reproduced from Ref. [8]. (Online version in color.)

The relative contribution of each of these sources is dependent on the specific ESR conditions. In the case of ESR for producing heavy ingot, the oxide scale that originates from the reaction between atmospheric oxygen and the steel electrode at the electrode surface during ESR is the major source of oxygen.¹²⁾ As for the ESR operation that is performed in open air atmosphere, the oxygen in air atmosphere is responsible for the increase in the oxygen content of liquid metal in the abovementioned third way. The ferrous oxide formed on the electrode surface and the original oxide scale on the electrode surface enter into the slag pool during ESR, resulting in an increase in the oxygen potential of the molten slag. The oxygen potential of the ESR-type slag is typically determined by the FeO activity in the slag (in some cases, MnO is present in the slag, but in a quite small fraction). FeO in molten slag pool introduces oxygen into the liquid metal pool as expressed in Eq. (1).   

$$(\mathrm{F}\mathrm{e}\mathrm{O})=[\mathrm{O}]+[\mathrm{F}\mathrm{e}]$$(1)

In this article, the square bracket [ ] indicates the component in liquid metal, and the bracket ( ) indicates the component in molten slag unless specially stated.

In addition, FeO could also introduce oxygen into the liquid metal film during the liquid metal film formation and collection into metal droplets at the electrode tip. Ferrous oxide in slag thus plays an important role in introducing oxygen to the metal phase.

In the case where inert gas atmosphere is employed in the ESR process, it is helpful to (i) prevent (at least substantially reduce) the formation of FeO, resulting from the chemical reaction between atmospheric oxygen and the steel electrode taking place on the electrode surface, which can indirectly lead to oxygen pickup in liquid steel, (ii) meanwhile, prevent the transfer of oxygen from the air atmosphere into liquid steel through the molten slag.

Furthermore, the soluble oxygen [O] as the product of desulfurization reactions in the ESR process could also result in oxygen pickup in liquid metal as expressed in Eq. (2).   

$$[\mathrm{S}]+({\mathrm{O}}^{2-})=({\mathrm{S}}^{2-})+[\mathrm{O}]$$(2)

3. Thermodynamic Considerations on Deoxidation of ESR

In conventional steelmaking, the deoxidation of liquid steel is generally realized through adding deoxidizing agents into liquid steel to combine the dissolved oxygen as oxide inclusions first, and then removing oxide inclusions for obtaining lower oxygen content of liquid steel. Different kinds of deoxidizing agents, such as Al, Si–Mn, Si–Fe, Al–Ti and SiAlBa, are widely used for liquid steel deoxidation according to the requirements of individual steel grade. As for the deoxidation of ESR, Al and Ca–Si are widely used for lowering the oxygen potential of molten slag and the oxygen content of liquid metal.

In the ESR refining practice, the presence of FeO in molten slag could hardly be prevented. The measured FeO content in the slag collected both during ESR and after ESR is lower than 1 mass% in the production practice.^8,25,26,27) As a reducible oxide in molten slag, FeO in molten slag could introduce oxygen into liquid steel according to the reaction (FeO)=[O]+[Fe] at a low oxygen level of liquid steel. On the contrary, the oxygen in liquid steel is transported into the molten slag according to the reaction [O]+[Fe]=(FeO) in the case of a low oxygen potential of molten slag and high oxygen level of liquid steel. Lowering the oxygen potential of molten slag could suppress the oxygen supply into liquid steel by the molten slag during ESR.

Taking electroslag remelting of H13 tool steel as an example, the oxygen content of liquid steel estimated from the reaction [Fe]+[O]=(FeO) is plotted against the FeO content in slag as shown in Fig. 2. The estimated oxygen content in liquid steel according to the reaction 3[O]+2[Al]=(Al₂O₃) is also present in Fig. 2 for comparison. A detailed description of the thermodynamic calculation has been present elsewhere.¹⁸⁾ The used slag is composed of 30.4 mass% CaF₂, 28.7 mass% CaO, 30.7 mass% Al₂O₃, 2.5 mass% MgO, 6.7 mass% SiO₂, and others (FeO<0.4 mass%, S<0.02 mass%, TiO₂<0.03 mass%). It can be learned from Fig. 2 that the measured total oxygen content of the remelted ingot, which includes both the dissolved oxygen (free oxygen) and the oxygen bonded as oxide inclusions, is much lower than the oxygen level determined by (FeO)–[O] equilibrium. The dissolved oxygen content of liquid steel determined by [Al]–[O] equilibrium is far lower than that by (FeO)–[O] equilibrium, as shown in Fig. 2. The dissolved oxygen of liquid steel is few parts per million, which is much lower than the total oxygen level (8 ppm in steel electrode and 14 ppm in remelted ingots).¹⁸⁾ The dissolved oxygen level of liquid steel is determined by [Al]–[O] equilibrium. Although the reaction equilibrium between liquid steel and molten slag with respect to the oxygen could hardly be reached in the ESR refining practice, the FeO content in the molten slag is indeed far higher than the equilibrium value. FeO in the slag consequently introduces the oxygen into liquid steel during the ESR process, resulting in the dissolved oxygen pickup in liquid steel. Even a small amount of FeO can lead to high oxygen potential of the slag for introducing oxygen into liquid steel.

Fig. 2.

Relationship between the dissolved oxygen concentration of liquid steel equilibrated with FeO in the slag and the FeO content as well as the oxygen level determined by [Al]–[O] equilibrium and the measured total oxygen content (T.O) in the remelted ingot.¹⁸⁾ (Online version in color.)

Thoroughly removing the oxide scale on the steel electrode surface prior to ESR and lowering the oxygen potential of molten slag during protective inert atmosphere ESR are suggested to prevent (at least suppress) the oxygen pickup in liquid steel. It is a common operation for deoxidation of ESR through decreasing the oxygen potential of molten slag by adding deoxidizing agents to molten slag pool. It has been verified from the plant trials when using the purified pre-melted slag (finally the total oxygen content <10 ppm in the remelted ingots), in which the FeO content is less than 0.1 mass%, combining with deoxidizing agent aluminum particles addition to slag pool during the ongoing protective argon gas atmosphere ESR process.¹⁸⁾

In the case of deoxidizing agent addition during ESR process, due to the large temperature difference and heat transfer between molten slag and deoxidizing agents, solid deoxidizing agents melt immediately after their addition. Taking Al-containing deoxidizing agent for an example, aluminum dissolved in molten slag pool will lower the oxygen potential of molten slag to restrain the transport of oxygen by FeO in molten slag into liquid steel as expressed in Eq. (3).   

$$2{[\mathrm{A}\mathrm{l}]}_{\mathrm{i}\mathrm{n}\mathrm{\hspace{0.17em} }\mathrm{ }\mathrm{s}\mathrm{l}\mathrm{a}\mathrm{g}}+3(\mathrm{F}\mathrm{e}\mathrm{O})=(\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3)+3[\mathrm{F}\mathrm{e}]$$(3)

According to reaction (3), the concentration of FeO in molten slag is reduced substantially, which destroys the oxygen transport by FeO into liquid steel. Meanwhile, the aluminum in molten slag brought by deoxidizing agents addition can also directly react with the oxygen in liquid metal film and liquid metal pool, as reaction 2[Al]+3[O]=(Al₂O₃), to reduce the oxygen content of liquid steel. In the case of a very low dissolved oxygen content (see Fig. 2 for example, the dissolved oxygen content in liquid steel is calculated to be only a few ppm based on (Al₂O₃)–[Al]–[O] equilibrium), deoxidizing agent Al could hardly lower the dissolved oxygen in liquid steel directly. A schematic illustration of the deoxidation mechanism by deoxidizing agent Al in the ESR process is present in Fig. 3.

Fig. 3.

Schematic illustration of oxygen transfer and deoxidation in the ESR process. The solid black point represents the dissolved oxygen in the liquid metal film. The open arrow indicates the oxygen transfer path, except for the representation of the path from [Al]_{in slag} to [Al] in liquid metal pool. The solid arrow indicates the direction of deoxidation.²⁸⁾ (Online version in color.)

The removal of oxide inclusions, which are the products of liquid steel deoxidation of ESR, is affected by various factors in the ESR process. If these oxide inclusions are not removed during the ESR, the intention of deoxidizing agent addition could not be realized. The aim of deceasing the oxygen content of the steel is to reduce the oxide inclusion amount. Therefore, the deoxidation operations of ESR should also consider their effects on the oxide inclusions for low oxygen steel production, such as inclusion chemistry and amount.

4. Deoxidation Kinetics of ESR

The kinetics of chemical reactions taking place in the ESR process has been studied in literatures. However, the works on this topic are still very limited. The kinetics studies that have been conducted up to date include the thermodynamic equilibrium model proposed by Hawkins,²⁹⁾ the single-stage reactor model developed by Etienne and Mitchell³⁰⁾ for predicting the changes in the concentrations of alloying elements experiencing metal-slag reactions, the mass transfer model of slag-metal reactions developed by Fraser and Mitchell,^31,32) the model for chemical reactions and mass transfer between slag and liquid steel developed by Wei and Mitchell,^25,26,33,34) the mass transfer model of desulfurization,³⁵⁾ and the deoxidation model developed by the present author.³⁶⁾

The prediction of the mass transfer model developed by Fraser and Mitchell^31,32) for calculating the oxidation loss of Mn in liquid steel by FeO in the slag agrees well with the experimental results in both the steady and the unsteady states. The model developed by Wei and Mitchell³³⁾ has been successfully applied to predict the changes in the concentrations of Mn, Si and Al in liquid steel, as well as the concentrations of MnO, SiO₂, Al₂O₃ and FeO in slag with the remelting time during ESR of low-alloy steel SAE 1020,^33,37) and the concentrations of these components during ESR of Cr–Mo–V turbine rotor steel on an industrial unit.²⁵⁾ The changes in the concentrations of Mn, Si, Al, Ti and Cr, as well as the concentrations of MnO, SiO₂, Al₂O₃, Cr₂O₃ and FeO in slag during ESR of stainless steel 1Cr18Ni9 are also precisely predicted with this model.^26,34)

In this article, the deoxidation kinetics of ESR, which is taken from the results of the present author’s research group, is presented. Based on the penetration theory, a kinetic model for oxygen transfer between molten slag and liquid steel during protective argon gas atmosphere ESR process was established.³⁶⁾ When the oxygen potential of slag is higher than that of liquid steel, FeO in molten slag introduces oxygen to liquid steel. On the contrary, the oxygen in liquid steel is transferred to the molten slag. According to the penetration theory, the oxygen potential of fluid microelement in liquid metal film at the electrode tip varies as the fluid microelement flows down along the electrode tip. A schematic illustration of microelement flowing of fluid at electrode tip is shown in Fig. 4. The interval time between starting of the fluid microelement formation and detachment of the fluid microelement from electrode tip is defined as residence time t_e.

Fig. 4.

Schematic illustration of microelement flowing of fluid at electrode tip.³⁶⁾

Considering that the rate of chemical reactions at the slag–metal interface is sufficiently large at elevated temperatures, mass transfer is the rate-controlling step of oxygen transfer in ESR process. The diffusion flux of oxygen transfer from molten slag to liquid steel introduced by FeO is expressed as follows³⁶⁾   

$$J=\frac{1}{t_{\text{e}}}{\int }_0^{t_{\text{e}}}\sqrt{\frac{D}{\pi t}}(c_{\text{s}}-c_{\text{b}})dt=2\sqrt{\frac{D_{\text{FeO}}}{\pi t_{\text{e}}}}(c_{\text{s, FeO}}-c_{\text{b, FeO}})$$(4)

where J is diffusion flux (mol/(m²·s)), c_{s, FeO} is the concentration of FeO at slag-liquid steel interface (mol/m³), c_{b, FeO} is the concentration of FeO in slag bulk (mol/m³), D_FeO is diffusion coefficient of FeO (m²/s), t_e is residence time (s).

The developed kinetic model is applied to protective argon gas atmosphere ESR of S136 tool steel. A description of the experimental details of protective atmosphere ESR and the results have been presented elsewhere.⁸⁾ The diffusion coefficient of FeO in 60 mass% CaF₂–20 mass% CaO–20 mass% Al₂O₃ molten slag D_FeO is taken as 1.5×10⁻⁸ m²/s.³³⁾ The model parameters are listed as follows:³⁶⁾ melting rate of protective atmosphere ESR is 1.59×10⁻⁶ m³/s, the measured cone angle θ on the electrode tip is 40°, the viscosity of liquid steel μ_m is 0.005 Pa·s, the density of liquid steel is 7.0×10³ kg/m³, the density of the molten slag ρ_s is 2.81×10³ kg/m³, the radius of the steel electrode R_e is 0.04 m, the residence time of fluid microelement at the electrode tip t_e is calculated to be 1.2 s. Figure 5 shows the relationship between the diffusion flux of oxygen transfer between molten slag and liquid steel phases, and FeO concentration in the molten slag. The remelting trials T1 and T2 were conducted in argon gas atmosphere throughout the ESR process, whereas the trials T3 and T4 were performed in open air atmosphere. For three heats (Exp. T1, T3, and T4), the deoxidizing agent was added continually into the slag pool during the ESR process for slag deoxidation. The addition rate of deoxidizing agent is about 20 kg/t in the trials T1 and T4, as well as about 40 kg/t in the trial T3.

Fig. 5.

Relationship between the diffusion flux of oxygen transfer between molten slag and liquid steel phases and FeO concentration in the slag.³⁶⁾ (Online version in color.)

It is learned from Fig. 5 that there is a critical value of FeO content in the slag. In the case where the FeO content of the slag is lower than this critical value, the oxygen transfer is from liquid steel to molten slag. Otherwise, the oxygen transfer is from molten slag to liquid steel. The critical value of the FeO concentration in the slag calculated using the developed kinetic model is 0.26 mass%, 0.45 mass%, 0.42 mass% and 0.70 mass%, respectively, as shown in Fig. 5.

The diffusion flux of oxygen transfer from liquid steel to molten slag decreases with increasing the concentration of FeO in the slag up to the critical value as shown in Fig. 5, whereas the diffusion flux of oxygen transfer from molten slag to liquid steel increases with increasing the FeO concentration in the slag. It indicates that ferrous oxide plays an important role in transporting oxygen between molten slag and liquid metal phases. It is crucial for reducing the oxygen content of liquid steel to an extra-low level through lowering FeO concentration in the slag before and during ESR process.

5. Evaluation of the Dependence of Oxygen on the Processing Parameters of ESR

5.1. Initial Oxygen Content of Steel Electrode

It has been verified by many laboratory-scale experiments and plant trials that the oxygen content of the steel could be greatly reduced after ESR refining. For electroslag remelting of the steel with high oxygen contents, ESR lowers the oxygen content invariably even if the protective atmosphere and/or slag deoxidation are not employed in the ESR operation. However, there are controversial findings on whether the oxygen content in the remelted ingot is dependent on the original oxygen content of the steel electrode or not.^{1,14,16,18,20,38,39,40,41,42,43,44)} Plöckinger¹⁾ claimed that the oxygen content level was dependent only on the reactions between slag and liquid steel, and not concerned with the initial oxygen content of steel electrode. In recent decades, the oxygen content in the steel used as the electrode for ESR could be reduced to a low level because of the improvement of steelmaking technologies. Several previous studies demonstrated that the oxygen content of the steel increased after ESR of the steel electrode with a low oxygen content, such as the work by Paar et al.,²⁰⁾ Medina and Cores,³⁸⁾ Wang et al.,³⁹⁾ Chang et al.,⁴⁰⁾ and Li et al.⁴¹⁾ Moreover, Zhou et al.⁴²⁾ reported that the oxygen content of the bearing steel was in the range of 15 to 30 ppm after ESR in open air atmosphere irrespective of the oxygen level (ranging from 5 ppm to 40 ppm) of the consumable steel electrode.

More and more efforts have been devoted to improve the cleanliness of the steel electrode. It is expected that a low initial oxygen content of the steel electrode gives a lower oxygen level in the remelted ingot. In the face of such situation, oxygen pickup is the key focus, and should also be suppressed during the ESR. The degree of oxygen pickup during ESR is dependent on the operation parameters of ESR, such as slag chemistry, steel composition and remelting atmosphere, especially the oxygen potential of the slag. The plant trials show that the oxygen content of martensitic stainless steel 8Cr17MoV is decreased from 58 ppm in the consumable steel electrode to 40 ppm after protective atmosphere ESR,¹⁶⁾ and the oxygen content of Si–Mn-killed steel is decreased from 74 ppm in the steel electrode to 34 ppm–38 ppm after protective atmosphere ESR.⁴⁵⁾ On the contrary, the oxygen content of H13 tool steel increases from 18 ppm in the consumable steel electrode to 21 ppm–34 ppm in remelted ingots,¹⁴⁾ and from 8 ppm in the consumable steel electrode to 14 ppm–17 ppm after protective atmosphere ESR.¹⁸⁾ However, this is not always the case. A low initial oxygen content of the steel electrode (13 ppm) was reduced to 8 ppm–12 ppm after open air atmosphere ESR in the case where the FeO content in the slag was at a low level (0.1–0.2 mass%).⁴³⁾ It is really a tough case to lower the oxygen content to a lower level after ESR of low oxygen steel. In order to further lower the oxygen during ESR of the steel with low oxygen content, the strategic point is to prevent the oxygen pickup during the ESR process, in which the oxygen potential of slag is the key source.

5.2. Oxide Inclusions in Steel Electrode

Deoxidation of liquid steel is directly related to the removal of oxide inclusions during ESR. In general, a majority of the original oxide inclusions in the steel electrode are removed during the ESR process, especially a preferential elimination of large inclusions.^6,43,44) Apart from some types of original oxide inclusions in the steel electrode (for example, MnO–SiO₂–Al₂O₃ inclusions⁴⁵⁾), other types of the original oxide inclusions could partially survive from the steel electrode to ESR ingot.^46,47) The factors, which affect the oxide inclusion evolution during ESR, greatly influence the oxygen content of ESR ingot.

Different deoxidation schemes (deoxidizing agent type, addition amount, and addition time, etc.) are generally adopted for liquid steel deoxidation according to the requirements of individual steel grade when producing the steel electrode for ESR, which consequently generate different types of oxide inclusions in the steel electrode, such as Al₂O₃, MgO·Al₂O₃ and various calcium aluminate inclusions. The removal degree of these oxide inclusions during ESR largely determines the total oxygen content of ESR ingots. Different types of original oxide inclusions basically experience various evolution trajectories during ESR, as demonstrated in the previous studies regarding Al₂O₃, MgO·Al₂O₃, MnO–SiO₂–Al₂O₃ and calcium aluminate inclusions,^{8,14,15,16,18,19,45)} consequently giving different removal degrees of oxide inclusions during the ESR of the steel. The types, compositions and size of the original oxide inclusions in the steel electrode exert a significant influence on the refining consequence of ESR in terms of the oxygen level of the remelted ingots.⁴⁸⁾

The cleanliness and oxide inclusions in the bearing steel ZGCr15 after electroslag remelting of the steel electrode deoxidized by different amounts of Al, Si–Ca, Si–Fe, Si–Mn–Ca or Al–Mn–Si in induction furnace melting which caused the generation of different types of original oxide inclusions were compared.⁴⁸⁾ It shows that the deoxidization of the steel electrode using Si–Ca gives the most effective refining in terms of the oxygen level and oxide inclusions in remelted ingots, whereas the deoxidization of the steel electrode using Al leads to the least effective refining, which is attributed to easier removal of low-melting-point silicate inclusions during the ESR process.

The deoxidation of 316LC stainless steel during open air atmosphere ESR was investigated by Ahmadi et al.,⁴⁹⁾ through deoxidizing the steel electrodes by aluminum or by aluminum and calcium-silicon wires. The results show that it is more effective for removal of calcium aluminate inclusions than alumina inclusions during the ESR, which is because the removal of alumina in the slag pool by chemical reactions is almost impossible, whereas calcium aluminate inclusions can float up easily in the liquid metal pool, consequently leading to a lower oxygen content of remelted ingot. Unfortunately, the comparison by Ahmadi et al.,⁴⁹⁾ was not conducted at the same oxygen level of the steel electrodes. Meanwhile, the differences in the elimination degrees of calcium aluminate inclusions and alumina inclusions during ESR were attributed to floatation and adherence by Ahmadi et al.⁴⁹⁾ However, the removal of inclusions through their floatation in liquid metal pool is possible, but it contributes in a very small manner with regard to the removal of inclusions in the ESR process.⁵⁰⁾ It was reported in some studies that removal of inclusions by floatation in liquid metal pool was unlikely because of their fine size.⁵¹⁾

For Si–Mn deoxidized steel, the oxide inclusions are manganese silicates which are liquid at steelmaking temperatures in many cases.^52,53,54,55) The present author investigated the evolution of oxide inclusions in Si–Mn-killed steel during protective atmosphere ESR.⁴⁵⁾ The results show that the oxide inclusions in the steel electrode are ternary (7.1–26.4 mass%) MnO-(53.3–82.8 mass%) SiO₂-(8.3–23.1 mass%) Al₂O₃ without exception, which are fully removed during the protective argon gas atmosphere ESR in two ways: a portion of these inclusions are dissociated in its individual chemical species into liquid steel as the liquid metal films form on the downside of the electrode tip and collect into liquid metal droplets subsequently during ESR, which consequently causes soluble oxygen pickup in liquid steel, whereas the others are removed by absorbing them into molten slag before liquid metal droplets collect in liquid metal pool during ESR. MgAl₂O₄ and Al₂O₃ inclusions readily form in the liquid metal pool as a result of the reaction between alloying elements and the dissolved oxygen in liquid steel that dissociates from MnO–SiO₂–Al₂O₃ inclusions.⁴⁵⁾ Consequently, in this case, the oxygen content of the remelted ingot is dependent on the ratio of MnO–SiO₂–Al₂O₃ inclusions dissociation to absorption of MnO–SiO₂–Al₂O₃ inclusions into molten slag during the ESR process. It is different from the removal of Al₂O₃ and MgAl₂O₄ inclusions, which are removed through absorbing them into molten slag.^51,56) The deoxidation degree of liquid steel virtually is different in these cases.

The change in the tantalum content during vacuum induction melting (VIM) and subsequent ESR of martensitic steel CPJ7 was investigated by Detrois et al.⁵⁷⁾ Their results demonstrate that the tantalum content of CPJ7 steel decreases by 25% during ESR, which is attributed to the formation of Ta₂O₅ inclusions during VIM and subsequent 95% reduction in the number density of Ta₂O₅ inclusions (be transferred to the slag) during ESR. A significant decrease in the oxygen content is observed for all ESR trials with an oxygen concentration of around 28 ppm in the ESR ingots. In their work, ESR of the CPJ7 steel electrodes results in, on average, 49% decrease in the oxygen content, which correlates to the removal of Ta₂O₅ inclusions in the steel electrode during ESR.

5.3. Remelting Atmosphere

Electroslag remelting is accompanied with chemical reactions among gas-slag-metal-inclusion phases. The chemical reactions at slag-metal interface are closely related to the oxygen in air atmosphere. The oxygen transfer from air atmosphere to slag-metal phases is schematically described in Fig. 1. In the case of the remelting leaving the slag surface and steel electrode open to the air atmosphere, atmospheric oxygen reacts with steel electrode at the electrode-atmosphere surface to form iron oxide as the electrode tip is heated at elevated temperatures during ESR. The formed iron oxide introduces oxygen into the liquid metal, as illustrated in Section 2. Apart from this trajectory of oxygen transfer, oxygen gas can be physically dissolved in molten slag as diatomic molecules,^21,58,59,60) oxygen molecules thereafter dissolve into liquid steel below the slag pool after permeating through the molten slag.²³⁾

The physical dissolution of oxygen gas and its permeability through the molten slag have been measured by several researchers.^{21,22,23,60,61)} The oxygen transfer rate through CaO–SiO₂–Al₂O₃ melts was measured to be in the range of 3×10⁻¹⁹ to 6×10⁻¹⁸ mol/(cm·s), and 5×10⁻¹² to 5×10⁻⁸ mol/(cm·s) through CaF₂–CaO–SiO₂–Al₂O₃ melts.^22,61) The physical solubility of oxygen in molten slag is largely dependent on the slag composition but not on the temperature.²¹⁾ The physical permeability of oxygen through CaF₂–Al₂O₃ and CaF₂–Al₂O₃–CaO molten slag for ESR was measured using an oxygen concentration cell with ZrO₂ solid electrolyte by Wei and Liu.²³⁾ The measured permeability of oxygen is 1×10⁻²⁰~6×10⁻¹⁹ mol/(cm·s) and 1×10⁻²¹~5×10⁻¹⁸ mol/(cm·s) in the temperature range of 1673–1873 K (1400–1600°C) under pure oxygen gas atmosphere at 0.1 MPa. These results show that the oxygen in air atmosphere could penetrate through molten slag as oxygen molecules into liquid steel, but the oxygen transfer in this trajectory makes a very small contribution. The oxygen pickup in this way is virtually prevented when protective inert gas atmosphere is employed in the ESR process. Protective inert gas atmosphere is increasingly used in ESR practice around the world, in order to prevent the oxygen introduction into liquid metal and the loss of alloying elements in steel and alloy.

The relationship between the oxygen content of electroslag remelted 304 stainless steel and the partial pressure of oxygen in gas atmosphere during ESR is shown in Fig. 6. The oxygen contents of the remelted ingots increase with increasing the partial pressure of oxygen in gas atmosphere irrespective of the slag systems. The oxygen content of the steel is reduced to a lower level by using the gas atmosphere with sufficiently low oxygen content in the ESR process. The oxygen contents of the ingots produced by using different slag compositions are apparently different because of different oxygen gas permeability of these slag systems at the same oxygen partial pressure.

Fig. 6.

Relationship between the oxygen content of remelted AISI304 stainless steel ingots and the partial pressure of oxygen in the gas atmosphere.⁶²⁾ (Online version in color.)

The ESR practice conducted in open air atmosphere normally leads to higher FeO content in the slag than the remelting in protective atmosphere,⁴³⁾ which results in a higher oxygen content of liquid steel. For protective atmosphere ESR, inert gas is introduced into the gas protective cap of protective atmosphere ESR equipment to protect the remelting atmosphere against the surrounding air. Protective inert gas atmosphere could suppress the oxygen pickup during the ESR process. In comparison with the remelting in protective argon gas atmosphere, it was reported by Wang et al.³⁹⁾ that the oxygen content of bearing steel GCr15 increased by 6 ppm–12 ppm in the case of the remelting in open air atmosphere. Chang et al.⁶³⁾ compared the remelting atmosphere of ESR, and found that the oxygen content increased from 20 ppm in the steel electrode to 36 ppm accompanied by a significant loss of Al and Si in the steel after ESR in open air atmosphere, whereas the oxygen content increased from 20 ppm to 24 ppm with an accompanying slight loss of Al and Si in steel in the case of the remelting in protective argon gas atmosphere.

Increasing the argon gas flow rate from 2.5 NL/min to 4 NL/min in the protective atmosphere ESR process was verified to decrease the oxygen content from 16 ppm to 12 ppm in the remelted FGH95 alloy ingot of 160 mm in diameter.⁶⁴⁾ Unfortunately, the chemical compositions and FeO content of the slag are not present and evaluated in the studies by Wang et al.,³⁹⁾ Chang et al.,⁶³⁾ and Chen et al.,⁶⁴⁾ which could be the main source of oxygen pickup in these results.

Targeting a low FeO content in the slag throughout the ESR process is indispensable for producing low oxygen steel and alloy. Although the slag with a low FeO content (FeO+TiO₂ ≤ 0.2 mass%) is used in the ESR practices, the oxygen content of the steel still increases by more than three times after ESR, accompanying with an increase in the amount of oxide inclusions.²⁰⁾ The oxygen pickup in these trials is attributed to the absence of the protective atmosphere for ESR operation, leaving the molten slag and the hot part of the electrode in permanent contact with air.²⁰⁾

Even though protective argon gas atmosphere is employed in many ESR experiments and production practices, the oxygen content in the gas atmosphere is seldom monitored and controlled. The ESR production practices demonstrate that the oxygen content is decreased by ESR in protective argon gas atmosphere (the oxygen content is lower than 100 ppm in the remelting atmosphere) from 17 ppm in the steel electrode to 8 ppm in the ingot of 600 mm in diameter.⁶⁵⁾ The application of protective argon gas atmosphere, in which the oxygen content could be monitored and lowered to a low level, plays an important role in producing ultralow oxygen steel.

Vacuum electroslag remelting has been developed to further improve the cleanliness of the steel in terms of the oxygen and inclusion contents.⁶⁶⁾ The work by Huang et al.⁶⁷⁾ shows that the oxygen content increases from 13 ppm in the steel electrode to 24 ppm and 18 ppm after ESR process in argon atmosphere (0.1 MPa) and vacuum atmosphere (0.01 MPa), respectively. The reduction in the oxygen increase when using vacuum atmosphere is attributed to carbon deoxidization, which only occurs in vacuum atmosphere. The development of vacuum electroslag remelting is still in its infancy, and more studies on the mechanism of deoxidation are still needed.

Pressurized electroslag remelting is recognized as a promising technology for producing high nitrogen steel.^68,69,70) In the research fields of pressurized electroslag remelting, the precious studies are mainly focused on the microstructure and mechanical properties of high nitrogen steel produced by pressurized electroslag remelting.^69,70,71) There is a lack of studies on the physicochemical phenomena of pressurized electroslag remelting, such as thermodynamic conditions and kinetics of chemical reactions. The transfer behavior of nitrogen at different nitrogen partial pressures and electrode immersion depths during pressurized electroslag remelting has been first ascertained by Yu et al.⁷²⁾ The role of pressurized atmosphere of ESR on the deoxidation of ESR and oxide inclusions has not been studied in published literatures yet. Thus future work is quite needed on this topic.

5.4. Deoxidation Schemes of ESR

Iron oxide in molten slag is the source of oxygen potential for oxygen pickup and the main driving force for the oxidation reactions in the ESR process, as illustrated in Section 3, which has also been manifested by other researchers according to their ESR trials.^4,5,73) It is demonstrated that the chemical reactions between FeO in the slag and deoxidizing agent predominate at high oxygen potential (FeO > 0.7 mass% in slag), whereas exchange reactions between deoxidizing agents and the slag components alter the inclusion compositions at low oxygen potential (FeO < 0.2 mass% in slag).⁵⁾ However, the presence of iron oxide in the molten slag during the ESR refining of steel is unavoidable. The sources of iron oxide include the component in the initial slag and the oxide scale formed on the electrode surface before and during ESR as a result of the reaction between the steel electrode and the oxygen in air atmosphere. Even a small amount of FeO in slag can result in a high level of oxygen in liquid steel, as illustrated in Section 3. The laboratory-scale ESR trials with a frequency of 4.5 Hz show that even though the ESR trials of the low oxygen steel electrode are performed in open air atmosphere, the oxygen content is reduced, instead increased, from 13 ppm in the steel electrode to 8 ppm–12 ppm in the ingots of 165 mm in diameter because of a low FeO content in the slag (0.1–0.2 mass%).⁴³⁾

Continuous or periodic addition of deoxidizing agents into the slag pool is widely used for deoxidation of ESR.^{4,5,7,8,12,38,74,75,76)} The main function of deoxidizing agents addition during the ESR process is to reduce the concentration of FeO in the slag.^{4,5,7,12,38,74,75,76)} The added deoxidizing agents, more or less, could hardly be avoided to enter into the liquid steel in the ESR process.^8,12,77) Deoxidizing agents are normally in the form of small pellets or grains for slag deoxidation in the ESR process. The high efficiency of this technique has been verified by many ESR experiments, plant trials and production practices. In many ESR practices, inert atmosphere and deoxidizing agent addition are employed simultaneously for reducing the oxygen content of the metal to an extremely low level.

Deoxidation schemes of ESR play an important role not only in the contents of oxygen and sulfur, the composition of liquid steel, but also in the oxide inclusions. In addition, the slag composition is also sensitive to deoxidation because of the accumulation of deoxidation products in slag, which causes the change in the chemistry of the slag.⁵⁾ This variation of slag compositions could cause the change in liquid steel composition through slag-metal reactions.

Aluminum is the most commonly used deoxidizing agent not only in conventional steelmaking process, but also in the ESR refining. Yoshinao et al.⁷⁸⁾ lowered the oxygen content of high nitrogen stainless steel to 20 ppm–30 ppm after ESR when using aluminum deoxidation for CaF₂–CaO–Al₂O₃ slag, and a higher deoxidation efficiency was obtained for the case of the remelting when using the slag with a lower Al₂O₃ content and higher CaO content. Shao et al.⁷⁹⁾ successfully kept the homogeneity of aluminum and silicon contents in the remelted ingots because of the decreased FeO and SiO₂ in the slag through adding aluminum powders into the slag pool during ESR of high speed steel M2Al.

The deoxidation ability of Al, Ca–Si and Al–Si during laboratory-scale ESR of SAE4340 steel in protective argon gas atmosphere using 50 mass% CaF₂-20 mass% CaO-30 mass% Al₂O₃ slag was compared by Mitchell et al.⁷⁾ The results show that FeO in slag plays a key role in the chemical reactions in the ESR process. In the case of a low oxygen potential (FeO < 0.2 mass%), the aluminum content of the steel increases because of the reaction between calcium and Al₂O₃ in slag when calcium-silicon is used for slag deoxidation. At the medium oxygen potential of the slag (0.4 to 0.6 mass% FeO), the deoxidation effectiveness of aluminum and calcium is comparable.

For designing proper schemes of deoxidizing agent addition, different addition rates of Al-based deoxidizing agents were attempted for the deoxidation of protective argon gas atmosphere ESR by the present author.^8,15,80) The oxygen content of high-Al steel was lowered from 34 ppm to 10 ppm after protective argon gas atmosphere ESR with Al-based deoxidizing agent addition (30 mass% Al+15 mass% Al₂O₃+15 mass% CaF₂+40 mass% iron powders) of 1.2 kg per ton steel.^17,77) The mechanical mixtures of 20 mass% Al, 20 mass% Al₂O₃, 20 mass% CaF₂ and 40 mass% iron powders were added continually into the slag pool for slag deoxidation of protective argon gas atmosphere ESR of S136 tool steel. The oxygen content was decreased from 89 ppm in the steel electrode to 12 ppm in the ESR ingot in the case of a deoxidizing agent addition rate of about 20 kg/t.⁸⁾ The oxygen content was decreased from 29 ppm in the steel electrode to 8 ppm in the 5 tonne-scale ingot produced by protective argon gas atmosphere ESR with Al-based deoxidizing agent addition for slag deoxidation.⁸⁰⁾ The deoxidizing agents not only lower the oxygen potential of the slag, but also react with the soluble oxygen in liquid steel.^8,77) It is schematically described in Fig. 3.

Deoxidizing agent Ca–Si is generally used for the deoxidation of ESR, in which the contents of aluminum, calcium, titanium and/or Al₂O₃-containing inclusions in the steel should be strictly controlled.⁷⁶⁾ Plöckinger¹⁾ claimed that Si deoxidation equilibrium was depended on the activity of SiO₂ in the slag during ESR, and the oxygen content of the steel could be reduced to a low level only at a low SiO₂ content in the slag.

Medina and Cores³⁸⁾ added Ca–Si (70 mass%) into the slag pool to reduce the FeO content in the slag during ESR for producing the steel with a good homogeneity of alloying elements. Their results show that the oxygen content of the steel increases after ESR in all cases (from initial 24–42 ppm to 42–83 ppm), the oxygen contents of the ingots microalloyed with Ti are determined by A1 and Ti contents. However, the reason for the oxygen pickup is not included in the study by Medina and Cores.³⁸⁾ The present author claim that the chemical reactions between microalloyed elements Ti, Si, Al and the slag should be responsible for the oxygen increase in their study.

Medina et al.⁷⁶⁾ lowered the FeO content of the slag by using Ca–Si during electroslag remelting of microalloyed steel. It is found that the oxygen content of the ingots is determined by the contents of calcium and aluminum in the steel. A thermodynamic equilibrium is established between the molten slag and liquid steel, and not between liquid steel and inclusions in the case of deoxidation with Ca–Si, which contributes to the absence of the loss of alloying elements.

The present author compared the effect of different calcium addition rates on the oxygen content of steel refined by protective argon gas atmosphere ESR. The results show that calcium addition makes no contribution to further lowering the oxygen content of the steel in comparison with the absence of calcium addition in the protective argon gas atmosphere ESR.^15,16)

For electroslag remelting of the steel with a low oxygen content, slag deoxidization could suppress the oxygen pickup in the steel. The electroslag remelting of GCr15 bearing steel with the oxygen content lower than 6 ppm in open air atmosphere was conducted by periodic addition of Ce–La for slag deoxidization. The degree of the oxygen pickup in the steel after ESR was reduced by nearly a half in comparison with the ESR trials without Ce–La addition.³⁹⁾

The addition rate of deoxidizing agents (termed deoxidation rate in some previous articles) is an important index in affecting the deoxidation of ESR. The addition rate is dependent on one or more of the FeO content in slag, slag composition, steel composition, oxygen content of the electrode, and yield of deoxidizing agents. Improper types and addition rates of deoxidizing agents will cause oxygen pickup in liquid steel as the deoxidation products and/or chemical reaction products between deoxidizing agent and slag components. In the case where the activity of FeO in slag is reduced to a low level after slag deoxidation, the oxide components of the slag could be reduced by the alloying elements or by excessive deoxidizing agents, resulting in an increase in the contents of some alloying elements in the ingot.^38,81) The present author carried out a serious of ESR deoxidation trials, and found that aluminum pickup took place in the steel during ESR when using Al-based deoxidizing agent for slag deoxidation.^8,15,17) In some cases, aluminum contents even roughly doubled in the steel after protective argon gas atmosphere ESR with Al-based deoxidizing agent addition.¹⁵⁾ The experimental work by Reyes-Carmona and Mitchell⁴⁾ shows that the aluminum content of the steel increases with increasing the addition rate of deoxidizing agent aluminum, which also results in the increase in the Al₂O₃ content in the complex oxide inclusions.

For deoxidation of ESR with aluminum when using different slag systems, Kajioka et al.⁸²⁾ found that the deoxidation rates between 0.05 to 0.1% Al produced an almost steady distribution of aluminum in the ESR ingots. Holzgruber⁸³⁾ proposed a proper deoxidation rate of 0.2% in the ESR operation. The experimental work by Wang et al.⁸⁴⁾ demonstrated that the deoxidation of ESR using aluminum powders, Ca–Si powders or RE–Mg–Si caused the pickup of aluminum in CrNiMoV steel. Ca–Si is the relatively proper deoxidizing agent, and its proper addition rate is 1.5 kg per ton steel in order to control the oxygen potential of the molten slag and prevent aluminum pickup in ESR of CrNiMoV steel. For electroslag remelting of T8MnA steel, the fluctuation of Si and Mn contents in the remelted ingots is originated from the reactions between SiO₂ and MnO in the slag and deoxidizing agent Al for slag deoxidation during the ESR process.⁸⁵⁾

Slag deoxidation is also widely employed to prevent the loss of concerned alloying elements in steel and alloy in the ESR process. This operation has been validated for successfully keeping constant Ti content in the stainless steel.^86,87) Decreasing the oxygen potential of the slag by periodic addition of aluminum shots (amounting to 0.1% of the ingot weight) into the slag pool during ESR in open air atmosphere was conducted by Chatterjee et al.⁶⁾ This operation eliminated the loss of silicon, manganese and chromium in 15CDV6 steel, otherwise 5–8% losses took place in the case without slag deoxidation during ESR. The chemical composition was uniform over the whole ingot in the longitudinal and transverse directions. The amount of inclusions was lowered from the index 0.48 to 0.12–0.17. The inclusions larger than 6 µm decreased from 40% of the total amount to 5%. The loss of Si and Mn was prevented and low Al content (<0.01 mass%) in steel was kept during ESR for 180 tonne-scale 26Cr2Ni4MoV steel production when using Ca–Al deoxidation for the earlier stage and Al deoxidation for the later stage.¹¹⁾ Wang et al.⁸⁸⁾ confirmed that the loss of Si and Mg in CrNiMoV steel was prevented in the case of ESR using Al for slag deoxidation, meanwhile homogeneous distribution of Al content in the ESR ingot was obtained.

Many efforts have been made to investigate deoxidation of ESR, but only a few studies have carried out to ascertain the effect of deoxidation operations of ESR on the characteristics of inclusions. There is a critical deoxidation rate for a specific ESR operation, which could result in a change in the chemical reactions between deoxidizing agent and FeO in the slag as well as other slag components. It was revealed in the previous study that excessive Al-based deoxidizing agent addition resulted in Al₂O₃ inclusions formation during protective atmosphere ESR of H13 tool steel, which survived in the remelted ingots ultimately.⁸⁰⁾ In the work by Schurman et al.,⁸⁹⁾ it was also noted that excessive addition rate of deoxidizing agent for deoxidation of ESR led to the generation of fresh alumina inclusions. A similar finding was reported in the experimental study by Mehrabi et al.⁹⁰⁾ Reyes-Carmona and Mitchell⁴⁾ recognized that below the critical addition rate of Ca–Si (10 kg per ton steel), the ESR refining system was ruled by the deoxidation reactions which were caused by a high activity of FeO in the slag. Above the critical addition rate, there is a low FeO activity of the slag, and deoxidizing agents react with the slag components CaO, SiO₂ and Al₂O₃, which lead to higher Al contents in remelted ingots and formation of aluminate inclusions. Higher deoxidation rates using Al during ESR were observed to cause higher Al₂O₃ content in aluminate inclusions.⁴⁾ The present author claims that the influence of deoxidizing agents on inclusions is not only dependent on the composition and addition rate of the deoxidizing agent, but also on the compositions of the original oxide inclusions in electrode.

Deoxidation of ESR for producing heavy ingot is still a challenging issue. For heavy ingot production, the types and addition rate of deoxidizing agents for slag deoxidation are mainly dependent on the variation of slag compositions during each ESR heat in dozens of hours. The complex deoxidation of ESR with aluminum at an addition rate of 0.35 kg to 0.25 kg every five minutes and Ca–Si with an addition rate of 0.10 kg to 0.05 kg every five minutes has been confirmed to be a proper deoxidization technology for producing 2.25Cr1Mo steel of 56 tonne ~79 tonne-scale of each ingot, in which the oxygen of the steel is decreased from 89 ppm to 26 ppm.¹²⁾ In the future work, the types and the addition rates of deoxidizing agents should be designed for the deoxidation of more steel and alloy grades in the different scale ESR production. For doing this, the online monitor of the change in the slag chemistry during ESR is indispensable. The process modelling of deoxidation for heavy remelted ingot will be a promising work.

5.5. Role of Slag Compositions

ESR slags generally are CaF₂–CaO–Al₂O₃–based system with minor additions of MgO, TiO₂ and/or SiO₂ to tailor the slag for the specific remelting requirements. The functions of the slag in the ESR process have been summarized elsewhere.⁹¹⁾ Deoxidation of ESR is strongly dependent not only on both the types and addition rates of deoxidizing agents, but also on the slag compositions. Different compositions of the ESR-type slags basically have different capacities for adsorption and dissolution of oxide inclusions,⁹²⁾ which cause different deoxidation degrees of liquid steel.^14,92)

FeO in slag determines the oxygen content of the remelted ingots in many cases.^42,93) The deoxidation operations in the ESR process are mainly to lower the FeO content in the slag through adding deoxidizing agents to the slag pool, as summarized above. Machining the steel electrode to a metallic bright surface could prevent the introduction of FeO into the slag to a higher level. It has been verified by Wang et al.³⁹⁾ and Liu et al.⁹⁴⁾ that this operation decreases the oxygen content of liquid steel effectively to a lower level for ESR of 1.2 tonne-scale bearing steel in comparison with the ESR without the operation for removing oxide scale on the steel electrode surface. The essence of removing oxide scale on steel electrode surface is to lower the increase of the FeO content in the slag so as to prevent the oxygen pickup in liquid steel.

The conventional slag used for electroslag remelting of superalloy is required to avoid the presence of SiO₂ in order to prevent the chemical reactions between SiO₂ and strong oxidizing alloying elements in the alloy such as Al and Ti. It is very difficult and costly to keep extremely low SiO₂ content in ESR slags during practical slag manufacturing process due to the impurity in raw materials. For electroslag remelting of most steels, however, SiO₂ is a permissible constituent in the ESR slags.⁹⁵⁾ In fact, it is suggested that a certain amount of SiO₂ addition in the CaF₂–CaO–Al₂O₃ slags can meet several requirements of drawing-ingot-type electroslag remelting of steel.⁹¹⁾ In the case of same types of deoxidizing agents (Al or Ca–Si) and addition rate, it has been verified by Reyes-Carmona and Mitchell⁴⁾ that the non-metallic inclusions change from only calcium aluminate inclusions to calcium aluminate and Ca–Al silicate inclusions when using 55mass% CaF₂-15mass% CaO-15mass% Al₂O₃-15mass% SiO₂ slag for the ESR instead of 50mass% CaF₂-20mass% CaO-30mass% Al₂O₃ slag.

The slag with different SiO₂ contents makes a difference to the steel cleanliness by affecting the liquid steel compositions, and indirectly changes the oxide inclusions. Mitchell et al.⁵⁾ conducted ESR trials using the slag with different SiO₂ contents (from 5 mass% to 23 mass%), and found that the slag with low SiO₂ activity gave rise to an inclusion population which is predominantly alumina or low-calcium aluminates in low-alloy steel, and the SiO₂ content of the inclusions increased as the SiO₂ activity of the slag increased, leading to the generation of aluminosilicate inclusions.

For the ESR refining process in which the reoxidation of liquid steel takes place, the oxygen content of the steel increases appreciably, but keeps roughly constant in the remelted ingots even though the slag with different SiO₂ contents is used in the ESR.¹⁴⁾ Increasing SiO₂ contents in the slag preferentially results in a decreasing pickup of aluminum in liquid steel, consequently contributes to a decreasing reduction of SiO₂ in CaO–Al₂O₃–SiO₂–MgO inclusions during the ESR.¹⁴⁾

Al₂O₃ is an indispensable component in almost all commercial ESR-type slag (mostly 20–30 mass%). It greatly affects the viscosity and electrical conductivity of the slag.⁹⁶⁾ The effect of varying Al₂O₃ content in slag on the steel cleanliness has been scarcely studied. For ESR of the steel with the oxygen content of 13 ppm, Schneider et al.⁴³⁾ recognized that increasing the activity of Al₂O₃ in the slag caused the increase in the oxygen content of the steel, in which the oxygen content almost doubled to 24 ppm in some cases, even though the FeO content in the slag was kept at a very low level (0.1 mass%). Chang et al.⁹⁷⁾ proposed that Al₂O₃ in the slag was an important source of oxygen pickup during ESR of the steel with low oxygen (18 ppm), in which the oxygen pickup was originated from the decomposition of Al₂O₃ from the slag. In their study, FeO and SiO₂ (4–6 mass%) in the slag was not considered for evaluating the factors affecting the oxygen content of liquid steel.

Commercial ESR-type slag contains a large amount of CaF₂ (typically 40–70 mass%), aiming to reduce the melting temperature and viscosity of the slag.⁹⁸⁾ Although CaF₂ plays an important role in the ESR-type slag, the evaporation of fluoride from the slag melts during ESR process has always been an extremely serious issue because it poses serious contamination of environment,^99,100,101) health hazard to plant operators,^101,102) corrosion of plant equipment,¹⁰³⁾ as well as the change in the slag chemistry. The variation of slag chemistry generally causes the fluctuation of the viscosity and other thermo-physical properties of the slag, which thereby could change the contents of some elements in liquid steel and degrade the reliability of ESR operating practice and the quality of remelted product, especially for large-scale ESR.¹⁰⁴⁾ However, the development of fluoride-free or low-fluoride slag for ESR is still in its infancy. Radwitz et al.¹⁰⁵⁾ compared the oxygen contents of the steel remelted using CaF₂–CaO–Al₂O₃ slag with varying CaF₂ contents (mass%CaO/mass%Al₂O₃=1). The results show that a lower oxygen content is obtained by protective atmosphere ESR of the steel with 25 ppm oxygen in the case of a coupled increase in the CaO and Al₂O₃ contents and decrease in the CaF₂ content in the slag. In the case of protective atmosphere ESR using high CaF₂ slag, the oxygen content of the ingot is slightly higher than that of the steel electrode. The total amount of inclusions in the steel decreases after protective atmosphere ESR, but increases with increasing the CaF₂ content in the slag and decreasing the contents of CaO and Al₂O₃ accordingly.

The commercial ESR-type slag generally contains about 3 mass% MgO. To evaluate the refining characteristics of ESR when using the slag with four different contents of MgO, a slight overpressure (0.12 MPa) argon gas atmosphere ESR of 21CrMoV5-7 steel was performed.⁹⁾ The chemical compositions of the slags are shown in Table 1.

Table 1. Chemical Compositions of the slag used in the ESR (mass%).

Slag No.	CaF2	CaO	Al2O3	MgO	SiO2	FeO
S1	59.34	18.98	18.64	2.03	0.22	0.046
S2	59.34	17.48	17.13	5.07	0.01	0.04
S3	59.34	15.05	14.69	9.98	0.03	0.04
S4	59.34	12.62	12.25	14.89	0.04	0.03

Figure 7 shows the total oxygen contents of remelted ingots over the ingot height when using different slag compositions. In Fig. 7, the average amount of total oxygen in the electrodes is highlighted in black, whereas the range of measured oxygen values in the electrodes is illustrated in grey. There are no changes in the total oxygen (T.O) contents over the ingot length, despite the variation of aluminum and silicon contents in the steel and different deoxidizing abilities of aluminum and silicon. It is further recognizable that the oxygen levels in the remelted ingots are in the range of that in the electrode or even slightly higher and could not be significantly reduced.⁹⁾ The comparison of the average total oxygen contents depending on the content of MgO shows that the oxygen level is not strongly influenced by the variation in the slag composition (see Fig. 8), only slightly higher contents are detected at 10 mass% MgO.⁹⁾

Fig. 7.

Total oxygen content after remelting using various slags over the ingot height.⁹⁾ (Online version in color.)

Fig. 8.

Comparison of average total oxygen content of remelted ingots depending on the MgO contents of the slag for ESR.⁹⁾ (Online version in color.)

With regard to the inclusions in the ingots, the amount of the inclusions smaller than 4 μm increases with increasing the MgO contents in the slag and even exceeds the amount of the inclusions in the steel electrode.⁹⁾ With the increase in the MgO contents in the slag, the total amount of inclusions increases, which increases up to similar values as that in the steel electrode due to the presence of a large amount of small inclusions in the case of higher MgO content in the slag. More studies are still needed to further clarify the mechanism as well as the effect of MgO on the inclusion chemistry.

5.6. Reoxidation of Liquid Steel

The difference in the oxygen potential between liquid steel and molten slag, as well as gas phase could contribute to the reoxidation of liquid steel during protective atmosphere ESR.¹⁴⁾ The oxygen content of steel nearly doubles (21 to 34 ppm) after protective atmosphere ESR of the steel with low oxygen content (18 ppm), indicating the occurrence of the reoxidation of liquid steel during the protective atmosphere ESR.¹⁴⁾ As part of the original oxide inclusions is removed during the ESR refining, the oxygen level contributed by the reoxidation of liquid steel virtually is higher than the difference in the measured oxygen contents between consumable electrodes and remelted ingots. The introduction of the oxygen from atmosphere and FeO (the newly-formed during on-going ESR, and oxide scale that not being removed from steel electrode surface) could hardly be prevented during ESR, and these aspects are the sources of oxygen pickup in the steel. The reoxidation of liquid steel during protective atmosphere ESR leads to considerable pickup of soluble oxygen in liquid steel, which provides a driving force for the generation of fresh oxide inclusions. In the author’s previous study, the formation of fresh CaO–Al₂O₃–MgO, Al₂O₃, MgAl₂O₄ and CaO–Al₂O₃–SiO₂–MgO inclusions during protective atmosphere ESR of tool steel takes place, as a result of the chemical reactions occurring inside liquid steel in the liquid metal pool caused by reoxidation of liquid steel.¹⁴⁾

Electroslag remelting of 2.4 tonne-scale bearing steel G20CrNi2Mo in protective argon gas atmosphere show that the FeO content in the slag increases from 0.20 mass% at the beginning of ESR to 0.45 mass% at the end of ESR, and the oxygen content increases from 12 ppm in the electrode to 16 ppm–21 ppm in the ingots, whereas aluminum decreases from 0.040 mass% to 0.031 mass%–0.019 mass%.¹⁰⁶⁾ A kinetic model for predicting the variation of oxygen and aluminum contents in bearing steel G20CrNi2Mo was developed based on the penetration and film theories by Li et al.¹⁰⁶⁾ The model reveals that the increase in the soluble oxygen in liquid steel mainly occurs during the metal droplets formation and falling. The rate-determining step of the reoxidation of liquid steel lies in the mass transfer of FeO at the slag side of the slag-steel interface. With the increase in the FeO content from 0.20 mass% to 0.45 mass%, the mass transfer resistance of FeO decreases obviously, thus resulting in an increase in the oxygen content and aluminum oxidation.

The slag deoxidation with Al–Mg alloy during protective atmosphere ESR of the steel with an oxygen content of 15 ppm has been validated to successfully suppress the reoxidation of liquid steel, but this operation fails to reduce the oxygen content of the steel.¹⁰⁷⁾ Although these authors¹⁰⁷⁾ emphasized the role of FeO in the slag on the oxygen pickup in liquid steel, they did not provide insight about the FeO content of the slag and its correlation with slag deoxidation operation.

5.7. Melting Rate and Filling Ratio of ESR

Liquid metal films form at the electrode tip during the ESR process, and thereafter collect as liquid metal droplets. This stage of ESR plays a predominant role in refining of liquid metal during the ESR process. The melting rate of ESR largely determines the thickness of the metal film at the electrode tip, and the residence time of the liquid metal films and metal droplets at the electrode tip.³¹⁾ In addition, the local solidification time and the advance rate of the solidifying front are strongly affected by the melting rate of ESR,^108,109) which therefore influence the formation and removal of fresh inclusions in the liquid metal pool during the ESR process.¹¹⁰⁾

Electroslag remelting of 316LC stainless steel with an oxygen content of 223 ppm shows that increasing the melting rates of ESR lowers the cleanliness of the steel in terms of oxide inclusions amount and the oxygen content of the steel.⁴⁹⁾ Ahmadi et al.⁴⁹⁾ deduced that the increase in the inclusion amount and oxygen content was originated from a faster pass of the liquid metal droplets through the slag pool with increasing the melting rates of ESR, which therefore led to a decrease in the elimination of oxide inclusions. This issue is open to question. The inclusion removal during ESR takes place predominantly at the stage of liquid metal films formation and their collection into droplets at the electrode tip, whereas the stage when the liquid metal droplets pass through the slag pool and the process in the liquid metal pool contributes in a small manner (does not play an important role).^{48,50,110,111,112,113)} The present author insists that the increase in the inclusions amount and oxygen content of the steel is mainly originated from the decrease in the residence time of the liquid metal films and metal droplets at the electrode tip with the increase in the melting rates of ESR, resulting in the decrease in the effective refining of liquid steel for inclusion removal.

In the case of protective atmosphere ESR of the steel with ultra low oxygen content (8 ppm), the present author’s previous study¹⁸⁾ shows that the oxygen content of the steel nearly doubles (up to 14–17 ppm) after protective atmosphere ESR, indicating the occurrence of the reoxidation of liquid steel during the ESR process. In these cases, the melting rates (350 kg/h, 400 kg/h, 450 kg/h, and 500 kg/h) of ESR make a negligible difference in the steel cleanliness in terms of the oxygen, sulfur, and nitrogen contents in the steel.

The filling ratio (ratio of the cross-sectional area of the electrode to the cross-sectional area of the mold) of ESR could be varied in a limited range according to the specific ESR requirements. The dependence of the oxygen content of the remelted ingot on the filling ratios of ESR has been studied scarcely. The electroslag remelting of bearing steel GCr15 with an oxygen content of 10 ppm was performed by Liu et al.⁹⁴⁾ to compare the effect of two filling ratios on the oxygen content of the steel. The results reveal that the increase in the filling ratio leads to a lower oxygen content of the remelted ingot, as shown in Fig. 9. It is attributed to the decrease in the flank area of the steel electrode with increasing the filling ratio of ESR. The present author claims that the smaller flank area of the steel electrode lowers the amount of the oxide scale generated on the steel electrode surface during the ESR process, which lowers the oxygen pickup of liquid steel resulting from FeO in the slag. Unfortunately, the FeO content of the slag is not provided in the article of Liu et al.⁹⁴⁾

Fig. 9.

Relationship between the filling ratios of ESR and the oxygen content of the remelted ingots when using CaF₂–Al₂O₃ slag in the ESR process.⁹⁴⁾ (Online version in color.)

6. Concluding Remarks

Although there are controversial findings on whether the oxygen content of remelted ingot is dependent on the oxygen content of the steel electrode or not, keeping a low oxygen content of the steel electrode is quite necessary for ultralow oxygen steel production. As for protective atmosphere of ESR, the oxygen concentration in the gas atmosphere should be detected online to keep an effective inert atmosphere. The oxygen potential of the slag has to be minimized to prevent the reoxidation of liquid steel even if protective inert atmosphere is employed for ESR. Thoroughly removing the oxide scale on the electrode surface and adding deoxidizing agents for slag deoxidation are indispensable operations in ESR practice for producing low and ultralow oxygen steel.

It is suggested that the addition rate of deoxidizing agents for successful deoxidation of a particular ESR practice should be based on the change in the oxygen potential of the slag and soluble oxygen content of liquid steel. The melting rates exert a negligible effect on the oxygen content in the case of protective atmosphere ESR of the ultra low oxygen steel, unlike the remelting of high oxygen steel. Filling ratio of ESR has a minor influence on the oxygen content of remelted ingot. Deoxidation of ESR for heavy ingot production is still a challenging issue. Monitoring the change in the slag chemistry during ESR is indispensable for designing proper types and addition rates of deoxidizing agents.

Oxide inclusion removal during the ESR process largely determines the oxygen content of remelted steel. The studies on the evolution trajectories of different types of oxide inclusions and its correlation with soluble oxygen and total oxygen of the steel during ESR process are still in its infancy. For example, the dissolution of oxide inclusions or adsorption of oxide inclusions by molten slag at different stages of ESR process, and the respective contribution to oxide inclusions removal in these ways have to be ascertained. Quantitative analysis of oxide inclusion removal in different trajectories will remain a challenge. It will provide guidance for designing the deoxidation schemes for refining liquid steel in the production of the consumable electrode for ESR.

Acknowledgments

The financial support by the National Natural Science Foundation of China (Grant Nos. 51874026 and 51504019), and the Fundamental Research Funds for the Central Universities (Grant No. FRF-TP-18-004A3) is greatly acknowledged. The author is also thankful for the financial support from the State Key Laboratory of Advanced Metallurgy (Grant No. 41618020).

References

• 1)   E.  Plöckinger: J. Iron Steel Inst., 201 (1973), 533.
• 2)   V.  Weber,  A.  Jardy,  B.  Dussoubs,  D.  Ablitzer,  S.  Rybéron,  V.  Schmitt,  S.  Hans and  H.  Poisson: Metall. Mater. Trans. B, 40 (2009), 271. https://doi.org/10.1007/s11663-008-9208-9
• 3)   S. K.  Maity,  N. B.  Ballal,  G.  Goldhahn and  R.  Kawalla: ISIJ Int., 49 (2009), 902. https://doi.org/10.2355/isijinternational.49.902
• 4)   F.  Reyes-Carmona and  A.  Mitchell: ISIJ Int., 32 (1992), 529. https://doi.org/10.2355/isijinternational.32.529
• 5)   A.  Mitchell,  F.  Reyes-Carmona and  E.  Samuelsson: Trans. Iron Steel Inst. Jpn., 24 (1984), 547. https://doi.org/10.2355/isijinternational1966.24.547
• 6)   M.  Chatterjee,  M. S. N.  Balasubramanian,  K. M.  Gupt and  P. K.  Rao: Ironmaking Steelmaking, 17 (1990), 38.
• 7)   A.  Mitchell,  F. R.  Carmona and  C.  Wei: Iron Steelmaker, 9 (1982), 37.
• 8)   C. B.  Shi,  X. C.  Chen,  H. J.  Guo,  Z. J.  Zhu and  H.  Ren: Steel Res. Int., 83 (2012), 472. https://doi.org/10.1002/srin.201100200
• 9)   S.  Radwitz,  H.  Scholz,  B.  Friedrich and  H.  Franz: Proc. 2015 Int. Symp. on Liquid Metal Processing and Casting, The Minerals, Metals and Materials Society, Warrendale, PA, (2015), 153.
• 10)   D. A. R.  Kay,  A.  Mitchell and  M.  Ram: J. Iron Steel Inst., 208 (1970), 141.
• 11)   D.  Xiang,  X.  Zhu and  K.  Wang: J. Iron Steel Res., 1 (1989), 9 (in Chinese).
• 12)   P.  Wen and  D.  Xiang: Heavy Cast. Forg., 1 (2016), 23 (in Chinese).
• 13)   D.  Hou,  Z.  Jiang,  Y.  Dong,  W.  Gong,  Y.  Cao and  H.  Cao: ISIJ Int., 57 (2017), 1400. https://doi.org/10.2355/isijinternational.ISIJINT-2017-147
• 14)   C. B.  Shi,  H.  Wang and  J.  Li: Metall. Mater. Trans. B, 49 (2018), 1675. https://doi.org/10.1007/s11663-018-1296-6
• 15)   C. B.  Shi,  X. C.  Chen,  H. J.  Guo,  Z. J.  Zhu and  X.  Sun: Metall. Mater. Trans. B, 44 (2013), 378. https://doi.org/10.1007/s11663-012-9780-x
• 16)   C. B.  Shi,  W. T.  Yu,  H.  Wang,  J.  Li and  M.  Jiang: Metall. Mater. Trans. B, 48 (2017), 146. https://doi.org/10.1007/s11663-016-0771-1
• 17)   C. B.  Shi,  X. C.  Chen and  H. J.  Guo: Int. J. Miner. Metall. Mater., 19 (2012), 295. https://doi.org/10.1007/s12613-012-0554-x
• 18)   C. B.  Shi,  D. L.  Zheng,  B. S.  Guo,  J.  Li and  F.  Jiang: Metall. Mater. Trans. B, 49 (2018), 3390. https://doi.org/10.1007/s11663-018-1398-1
• 19)   S.  Li,  G.  Cheng,  Z.  Miao,  W.  Dai,  L.  Chen and  Z.  Liu: ISIJ Int., 58 (2018), 1781. https://doi.org/10.2355/isijinternational.ISIJINT-2018-072
• 20)   A.  Paar,  R.  Schneider,  P.  Zeller,  G.  Reiter,  S.  Paul and  P.  Würzinger: Steel Res. Int., 85 (2014), 570. https://doi.org/10.1002/srin.201300317
• 21)   M.  Sasabe and  K. S.  Goto: Metall. Trans., 5 (1974), 2225. https://doi.org/10.1007/BF02643937
• 22)   M.  Sasabe and  Y.  Kinoshita: Tetsu-to-Hagané, 65 (1979), 1727 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.65.12_1727
• 23)   J.  Wei and  Z.  Liu: Acta Metall. Sin., 30 (1994), 350 (in Chinese).
• 24)   A.  Masui,  Y.  Sasajima,  N.  Sakata and  M.  Yamamura: Tetsu-to-Hagané, 63 (1977), 2181 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.63.13_2181
• 25)   J.  Wei and  A.  Mitchell: Acta Metall. Sin., 23 (1987), 126 (in Chinese).
• 26)   J.  Wei: Chin. J. Met. Sci. Technol., 5 (1989), 235.
• 27)   S.  Li,  G.  Cheng,  Z.  Miao,  L.  Chen and  X.  Jiang: Int. J. Miner. Metall. Mater., 26 (2019), 291. https://doi.org/10.1007/s12613-019-1737-5
• 28)   C. B.  Shi,  X. C.  Chen,  Y. W.  Luo and  H. J.  Guo: Materials Processing Fundamentals, The Minerals, Metals and Materials Society, Warrendale, PA, (2013), 31.
• 29)   R. J.  Hawkins,  D. J.  Swinden and  D. N.  Pocklington: Electroslag Refining: Proc. Conf. on Electroslag Refining, The Iron and Steel Institute, London, (1973), 21.
• 30)   M.  Etienne and  A.  Mitchell: 28th Electric Furnace Conf. Proc., Iron and Steel Society of AIME, Warrendale, PA, (1970), 28.
• 31)   M. E.  Fraser and  A.  Mitchell: Ironmaking Steelmaking, 3 (1976), 279.
• 32)   M. E.  Fraser and  A.  Mitchell: Ironmaking Steelmaking, 3 (1976), 288.
• 33)   J.  Wei and  A.  Mitchell: Acta Metall. Sin., 20 (1984), 261 (in Chinese).
• 34)   C.  Wei: J. Xian Inst. Metall. Constr. Eng., 44 (1985), 69 (in Chinese).
• 35)   D.  Hou,  Z. H.  Jiang,  Y.  Dong,  Y.  Li,  W.  Gong and  F.  Liu: Metall. Mater. Trans. B, 48 (2017), 1885. https://doi.org/10.1007/s11663-017-0921-0
• 36)   C. B.  Shi,  H. J.  Guo,  X. C.  Chen,  X.  Sun and  J.  Fu: Special Steel, 34 (2013), 11 (in Chinese).
• 37)   J.  Wei and  A.  Mitchell: Acta Metall. Sin., 20 (1984), 280 (in Chinese).
• 38)   S. F.  Medina and  A.  Cores: ISIJ Int., 33 (1993), 1244. https://doi.org/10.2355/isijinternational.33.1244
• 39)   C. S.  Wang,  S. G.  Liu,  M. D.  Xu,  F. X.  Liu and  D. S.  Li: Special Steel, 18 (1997), 31 (in Chinese).
• 40)   L. Z.  Chang,  H. S.  Yang and  Z. B.  Li: Steelmaking, 26 (2010), 46 (in Chinese).
• 41)   J.  Li and  J.  Zhang: Shanxi Metall., (2019), No. 2, 52 (in Chinese).
• 42)   D. G.  Zhou,  W. G.  Xu,  P.  Wang,  J.  Fu,  J. H.  Xu,  C. S.  Wang and  M. D.  Xu: Iron Steel, 33 (1998), 13 (in Chinese).
• 43)   R. S. E.  Schneider,  M.  Molnar,  S.  Gelder,  G.  Reiter and  C.  Martinez: Steel Res. Int., 89 (2018), 1800161. https://doi.org/10.1002/srin.201800161
• 44)   G. M.  Padki,  M. S. N.  Balasubramanian,  K. M.  Gupt and  P. K.  Rao: Ironmaking Steelmaking, 10 (1983), 180.
• 45)   C. B.  Shi and  J. H.  Park: Metall. Mater. Trans. B, 50 (2019), 1139. https://doi.org/10.1007/s11663-019-01564-6
• 46)   E. S.  Persson and  A.  Mitchell: Proc. Liquid Metal Processing and Casting Conf. 2017, The Minerals, Metals & Materials Society, Pittsburgh, PA, (2017), 373.
• 47)   E. S.  Persson,  A.  Karasev and  P.  Jönsson: Proc. Liquid Metal Processing and Casting Conf. 2017, The Minerals, Metals & Materials Society, Pittsburgh, PA, (2017), 353.
• 48)   Z.  Li,  J.  Zhang and  X.  Che: J. Iron Steel Res., 9 (1997), 7 (in Chinese).
• 49)   S.  Ahmadi,  H.  Arabi,  A.  Shokuhfar and  A.  Rezaei: J. Mater. Sci. Technol., 25 (2009), 592.
• 50)   A.  Mitchell: Ironmaking Steelmaking, 1 (1974), 172.
• 51)   R. G.  Baligidad,  U.  Prakash,  V. R.  Rao,  P. K.  Rao and  N. B.  Ballal: Ironmaking Steelmaking, 21 (1994), 324.
• 52)   Y. B.  Kang and  H. G.  Lee: ISIJ Int., 44 (2004), 1006. http://doi.org/10.2355/isijinternational.44.1006
• 53)   K.  Wang,  M.  Jiang,  X.  Wang,  Y.  Wang,  H.  Zhao and  Z.  Cao: Metall. Mater. Trans. B, 46 (2015), 2198. https://doi.org/10.1007/s11663-015-0411-1
• 54)   A.  García-Carbajal,  M.  Herrera-Trejo,  E. I.  Castro-Cedeño,  M.  Castro-Román and  A. I.  Martinez-Enriquez: Metall. Mater. Trans. B, 48 (2017), 3364. http://doi.org/10.1007/s11663-017-1066-x
• 55)   J. S.  Park and  J. H.  Park: Metall. Mater. Trans. B, 45 (2014), 953. https://doi.org/10.1007/s11663-013-9998-2
• 56)   J.  Fu: Acta Metall. Sin., 15 (1979), 526 (in Chinese).
• 57)   M.  Detrois,  P. D.  Jablonski and  J. A.  Hawk: Metall. Mater. Trans. B, 50 (2019), 1686. https://doi.org/10.1007/s11663-019-01614-z
• 58)   C. H.  Greene and  R. F.  Gaffney: J. Am. Ceram. Soc., 42 (1959), 271. https://doi.org/10.1111/j.1151-2916.1959.tb12952.x
• 59)   M.  Sasabe and  Y.  Kinoshita: Trans. Iron Steel Inst. Jpn., 20 (1980), 801. https://doi.org/10.2355/isijinternational1966.20.801
• 60)   M.  Sasabe and  S.  Kitamura: ISIJ Int., 33 (1993), 133. https://doi.org/10.2355/isijinternational.33.133
• 61)   M.  Sasabe and  Y.  Kinoshita: Tetsu-to-Hagané, 64 (1978), 1313 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.64.9_1313
• 62)   W.  Holzgruber: BHM Berg-und Hüttenmänn. Monatsh., 161 (2016), 2.
• 63)   L. Z.  Chang,  H. S.  Yang and  Z. B.  Li: Special Steel, 30 (2009), 59 (in Chinese).
• 64)   X. C.  Chen,  D.  Feng and  X. F.  Yang: Proc. 2005 China Special Steel Annual Meeting, The Chinese Society for Metals, Beijing, (2005), 64 (in Chinese).
• 65)   D. L.  Zheng,  J.  Li,  C. B.  Shi,  H. C.  Xu and  B. S.  Guo: Proc. 2019 Annual Meeting for Secondary Refining, The Chinese Society for Metals, Beijing, (2019), 528 (in Chinese).
• 66)   A.  Sekiya,  S.  Nakayama and  T.  Taketsuru: Denki-Seiko, 66 (1995), 47 (in Japanese).
• 67)   X.  Huang,  B.  Li,  Z.  Liu,  T.  Jiang,  Y.  Chai and  X.  Wu: Vacuum, 164 (2019), 114. https://doi.org/10.1016/j.vacuum.2019.03.014
• 68)   W.  Holzgruber: Proc. 7th Int. Conf. on Vacuum Metallurgy, Iron Steel Institute of Japan, Tokyo, (1982), 1452.
• 69)   G.  Stein and  J.  Menzel: Int. J. Mater. Prod. Technol., 10 (1995), 290, https://www.inderscience.com/info/inarticle.php?artid=36456, (accessed 2019-10-24).
• 70)   A. D.  Patel,  J.  Reitz,  J. H.  Magee,  R.  Smith,  G.  Maurer and  B.  Friedrich: Proc. 2009 Int. Symp. on Liquid Metal Processing and Casting, The Minerals, Metals and Materials Society, Warrendale, PA, (2009), 1.
• 71)   P.  Pant,  P.  Dahlmann,  W.  Schlump and  G.  Stein: Steel Res., 58 (1987), 18. https://doi.org/10.1002/srin.198701482
• 72)   J.  Yu,  F. B.  Liu,  H. B.  Li,  Z. H.  Jiang,  Y.  Li,  C. P.  Kang,  A.  Wang,  W. C.  Zhang and  H.  Feng: Metall. Mater. Trans. B, 50 (2019), 3112. https://doi.org/10.1007/s11663-019-01714-w
• 73)   K.  Narita,  T.  Onoye,  T.  Ishii and  T.  Kusamichi: Tetsu-to-Hagané, 64 (1978), 1568 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.64.10_1568
• 74)   A.  McLean and  R. G.  Ward: J. Iron Steel Inst., 204 (1966), 8.
• 75)   W.  Holzgruber and  E.  Plöckinger: Stahl Eisen, 88 (1968), 638 (in German).
• 76)   S. F.  Medina,  F.  López and  A. G.  Coedo: Ironmaking Steelmaking, 24 (1997), 329.
• 77)   C. B.  Shi,  X. C.  Chen and  H. J.  Guo: AISTech 2012 Proc., Association for Iron & Steel Technology, Warrendale, PA, (2012), 947.
• 78)   K.  Yoshinao and  K.  Yasuyuki: Proc. Int. Conf. on High Nitrogen Steels 2006, Metallurgical Industry Press, Beijing, (2006), 360.
• 79)   Q. L.  Shao: Hebei Metall., (1999), No. 5, 48 (in Chinese).
• 80)   C. B.  Shi,  X. C.  Chen,  F.  Wang and  H. J.  Guo: Materials Science & Technology 2012 Conf. and Exhibition Proc., The Minerals, Metals and Materials Society, Warrendale, PA, (2012), 1261.
• 81)   G.  Pateisky,  H.  Biele and  H. J.  Fleischer: J. Vac. Sci. Technol., 9 (1972), 1318.
• 82)   H.  Kajioka,  K.  Yamaguchi,  N.  Sato,  K.  Soejima and  S.  Sakaguchi: Proc. 4th Int. Symp. on Electroslag Remelting, ISIJ, Tokyo, (1973), 102.
• 83)   W.  Holzgruber: Proc. 1st Int. Symp. on ESR and Casting Technology, Mellon Institute, Pittsburgh, PA, (1967), 1.
• 84)   A. R.  Wang,  Y. G.  Yu and  J.  Fu: J. Univ. Sci. Technol. Beijing, 19 (1997), 134 (in Chinese).
• 85)   L. J.  Zhang,  C. T.  Wang,  R. L.  Jiang and  L. F.  Liu: Jiangxi Metall., 15 (1995), 6 (in Chinese).
• 86)   D.  Hou,  Z.  Jiang,  Y.  Dong,  W.  Gong,  Y.  Cao and  H.  Cao: ISIJ Int., 57 (2017), 1410. https://doi.org/10.2355/isijinternational.ISIJINT-2017-148
• 87)   D.  Hou,  F.  Liu,  T.  Qu,  Z.  Jiang,  D.  Wang and  Y.  Dong: ISIJ Int., 58 (2018), 876. https://doi.org/10.2355/isijinternational.ISIJINT-2017-687
• 88)   A. R.  Wang,  Y. G.  Yu and  J.  Fu: J. Univ. Sci. Technol. Beijing, 5 (1998), 87.
• 89)   E.  Schurman,  P.  Funders, and  H.  Littersheidt: Arch. Eisenhuttenwes., 46 (1975), 773.
• 90)   K.  Mehrabi,  M. R.  Rahimipour and  A.  Shokuhfar: Int. J. Iron Steel Soc. Iran, 2 (2005), 37.
• 91)   C. B.  Shi,  J.  Li,  J. W.  Cho,  F.  Jiang and  I.  Jung: Metall. Mater. Trans. B, 46 (2015), 2110. https://doi.org/10.1007/s11663-015-0402-2
• 92)   Y. W.  Dong,  Z.  Jiang,  Y.  Cao,  A.  Yu and  D.  Hou: Metall. Mater. Trans. B, 45 (2014), 1315. https://doi.org/10.1007/s11663-014-0070-7
• 93)   X.  Huang,  B.  Li and  Z.  Liu: Metall. Mater. Trans. B, 49 (2018), 709. https://doi.org/10.1007/s11663-017-1158-7
• 94)   S. G.  Liu,  M. D.  Xu,  F. X.  Liu,  C. S.  Wang and  D. S.  Li: Special Steel, 14 (1993), 45 (in Chinese).
• 95)   M.  Allibert,  J. F.  Wadier and  A.  Mitchell: Ironmaking Steelmaking, 5 (1978), 211.
• 96)   S.  Hara,  H.  Hashimoto and  K.  Ogino: Trans. Iron Steel Inst. Jpn., 23 (1983), 1053.
• 97)   L. Z.  Chang,  X. F.  Shi and  J. Q.  Cong: Ironmaking Steelmaking, 41 (2014), 182.
• 98)   C. B.  Shi,  S. H.  Shin,  D. L.  Zheng,  J. W.  Cho and  J.  Li: Metall. Mater. Trans. B, 47 (2016), 3343. https://doi.org/10.1007/s11663-016-0826-3
• 99)   H.  Nakada and  K.  Nagata: ISIJ Int., 46 (2006), 441. https://doi.org/10.2355/isijinternational.46.441
• 100)   N.  Takahira,  M.  Hanao and  Y.  Tsukaguchi: ISIJ Int., 53 (2013), 818. https://doi.org/10.2355/isijinternational.53.818
• 101)   J. L.  Klug,  R.  Hagemann,  N. C.  Heck,  A. C. F.  Vilela,  H. P.  Heller and  P. R.  Scheller: Steel Res. Int., 83 (2012), 93. https://doi.org/10.1002/srin.201200093
• 102)   M.  Persson,  S.  Seetharaman and  S.  Seetharaman: ISIJ Int., 47 (2007), 1711. https://doi.org/10.2355/isijinternational.47.1711
• 103)   T.  Omoto,  Y.  Iwamoto and  H.  Yamaji: Shinagawa Tech. Rep., 45 (2002), 85.
• 104)   D. L.  Xiang: Heavy Cast. Forg., (2011), No. 1, 26 (in Chinese).
• 105)   S.  Radwitz,  H.  Scholz,  B.  Friedrich and  H.  Franz: Proc. European Metallurgical Conf. 2015, Vol. 2, GDMB Society of Metallurgists and Miners, Düsseldorf, (2015), 887.
• 106)   S. J.  Li,  G. G.  Cheng,  Z. Q.  Miao,  L.  Chen,  C. W.  Li and  X. Y.  Jiang: ISIJ Int., 57 (2017), 2148. https://doi.org/10.2355/isijinternational.ISIJINT-2017-227
• 107)   H.  Wang,  C. M.  Shi,  J.  Li,  C. B.  Shi and  Y. F.  Qi: Ironmaking Steelmaking, 45 (2018), 6. https://doi.org/10.1080/03019233.2016.1235078
• 108)   B.  Hernandez-Morales and  A.  Mitchell: Ironmaking Steelmaking, 26 (1999), 423. https://doi.org/10.1179/030192399677275
• 109)   L.  Rao,  J. H.  Zhao,  Z. X.  Zhao,  G.  Ding and  M. P.  Geng: J. Iron Steel Res. Int., 21 (2014), 644.
• 110)   D. A. R.  Kay and  R. J.  Pomfret: J. Iron Steel Inst., 209 (1971), 962.
• 111)   J.  Fu and  J.  Zhu: Acta Metall. Sin., 7 (1964), 250 (in Chinese).
• 112)   Z. B.  Li,  W. H.  Zhou and  Y. D.  Li: Iron Steel, 15 (1980), 20 (in Chinese).
• 113)   D. G.  Zhou,  X. C.  Chen,  J.  Fu,  P.  Wang,  J.  Li and  M. D.  Xu: J. Univ. Sci. Technol. Beijing, 22 (2000), 26 (in Chinese).