Effect of Ce₂O₃ on Viscosity and Structure for CaO–SiO₂–Al₂O₃-based Mold Flux

Abstract

The viscosity of 30%CaO-30%SiO₂-15%Al₂O₃-5%MgO-10%Na₂O-10%CaF₂-xCe₂O₃ (mass%, x = 0, 5, 10, 15) were measured by the rotating column method, and then the viscosity and the T_br (break temperature) of the mold flux were analyzed based on the results. Meanwhile, the structural characteristics of the mold flux were investigated using Raman spectroscopy and XRD (X-ray diffraction). The results show that Ce₂O₃ predominantly destroys the silicate network structure at high temperature, reduces the polymerization degree of the mold flux, simplifies the high-temperature structure of the mold flux, and reduces the friction resistance of viscous flow. From the apparent phenomenon, the viscosity of the mold flux decreases at high temperature. In addition, with the increase of Ce₂O₃ content in the mold flux, Ca₄Si₂O₂F changes to Ce_9.33(SiO₄)₆O₂, which enhances the crystallization ability of the mold flux and increases the T_br of the mold flux.

1. Introduction

In the process of continuous casting, mold flux can effectively realize heat transfer and lubrication between slab and mold flux, absorb floating inclusions in molten steel and prevent secondary oxidation of molten steel. Therefore, mold flux has already become the key functional material in continuous casting process.^1,2) Generally, the traditional mold flux for continuous casting is a mixture composed of CaO and SiO₂ as the main components, adding a certain amount of fluxes such as Na₂O and CaF₂, and carbonaceous material for adjusting melting rate. The composition is shown in A region of Fig. 1.^3,4) When casting steel grades containing easily-oxidizable elements, the high content of SiO₂ in the mold flux is easily reduced by the active metal elements (such as Eq. (1)) to form high melting point phases, such as CaAl₄O₇, CaTiO₃, etc., which will increase the melting point and viscosity of the mold flux, and then affect the smooth operation of the continuous casting process.^5,6,7)   

$$2\left[\mathrm{M}\right]+x\left(\mathrm{S}\mathrm{i}{\mathrm{O}}_2\right)=x\left[\mathrm{S}\mathrm{i}\right]+2\left(\mathrm{M}{\mathrm{O}}_x\right)$$(1)

Fig. 1.

The phase diagram of CaO–SiO₂–Al₂O₃. (Online version in color.)

In order to solve the problem of elements oxidation, scholars proposed to use CaO–Al₂O₃-based mold flux,^8,9,10) and its composition is shown in B region of Fig. 1. The composition of the mold flux is easy to form gehlenite with high melting point, which makes the glass performance of the mold flux worse, thus affects the lubrication of mold flux and deteriorates the surface quality of slab.^11,12,13) Therefore, it is necessary to enhance the glass performance of mold flux by adding fluxes such as B₂O₃. However, B₂O₃ reacts more easily than SiO₂, and the reaction described in Eq. (1) will still occur.¹⁴⁾

In view of the problems of element oxidation and lubrication in the continuous casting process, scholars have developed a low-activity mold flux based on CaO–SiO₂–Al₂O₃-based slag system,¹⁵⁾ and its composition is shown in C region of Fig. 1. The main precipitated phases in this region are melilite, CaAl₂Si₂O₈ and CaSiO₃. At this time, SiO₂ can effectively combine with other components, so the activity of SiO₂ in C region is lower than that in A region, and the chemical reaction between slag and steel at the interface is weak. At the same time, the content of SiO₂ in the mold flux is high, which can effectively ensure the lubricating ability of the mold flux.

Adding stable oxide Ce₂O₃ to the CaO–SiO₂–Al₂O₃-based mold flux to form Ce₂Si₂O₇¹⁶⁾ can further control the activity of SiO₂. Anacleto et al.¹⁷⁾ proposed that Ce₂O₃ could reduce the activity of SiO₂ in the CaO–SiO₂–Ce₂O₃ system during the study of the sulfur partition between CaO–SiO₂ slags and carbon-saturated iron at 1500°C. But they did not provide any relevant proof. At present, there are few reports on the effect of Ce₂O₃ on the viscosity and structure of CaO–SiO₂–Al₂O₃-based mold flux. In the study of similar slag systems, Jeong et al.¹⁸⁾ and Kim et al.¹⁹⁾ investigated the network structure of the MnO–SiO₂–Al₂O₃–Ce₂O₃ slag system. When Ce₂O₃ was added to the slag (M/S = 2.2), the [AlO₄]⁻: 0.5Mn²⁺ units were transformed to [AlO₆]³⁻: Ce³⁺ units by employing free oxygen. However, structural changes were not observed despite the addition of Ce₂O₃ to the slag (M/S = 0.9). Qi et al.²⁰⁾ found a similar situation in the study of the CaO–Al₂O₃–Li₂O–Ce₂O₃ slag, where the content of Ce₂O₃ increased from 0 to 10 wt pct, the relative fraction of [AlO₄]-tetrahedron decreased and the relative fraction of [AlO₆]-octahedron increased obviously.

The viscous flow behavior of 30%CaO-30%SiO₂-15%Al₂O₃-5%MgO-10%Na₂O-10%CaF₂-xCe₂O₃ (mass%, x = 0, 5, 10, 15) system was studied in this paper. The effects of Ce₂O₃ content on the viscosity and T_br of the mold flux were discussed in detail. At the same time, its mechanism was clarified by Raman spectroscopy and XRD.

2. Materials and Methods

2.1. Sample Preparation

The mold flux used in the experiment is Ce₂O₃ with different quality added to 30%CaO-30%SiO₂-15%Al₂O₃-5%MgO-10%Na₂O-10%CaF₂. The chemical composition ratio of the mold flux is shown in Table 1. Analytical grade CaO, SiO₂, Al₂O₃, MgO, Na₂CO₃ (instead of Na₂O), CaF₂ and Ce₂C₃O₉ (instead of Ce₂O₃) were used as raw materials (purity > 99%), and CaO, MgO and Ce₂C₃O₉ powders were calcined at 873 K for 2 h to remove water in the experiment.

Table 1. Chemical compositions of the mold fluxes (mass%).

Sample	CaO	SiO2	Al2O3	MgO	Na2O	CaF2	Ce2O3
S0	30.0	30.0	15.0	5.0	10.0	10.0	0.0
S1	28.5	28.5	14.2	4.8	9.5	9.5	5.0
S2	27.0	27.0	13.5	4.5	9.0	9.0	10.0
S3	25.5	25.5	12.7	4.3	8.5	8.5	15.0
S4	24.0	24.0	12.0	4.0	8.0	8.0	20.0

2.2. Viscosity Measurements

The viscosity of mold flux was measured by rotation method.^21,22,23) The experimental equipment is BCT1700, which conforms to the standard methods of American ASTMC 965 and ASTMC 1276. The measurement accuracy of the testing instrument is ±1%, and the reproducibility is within ±0.2%. In the experiment, molybdenum crucible (inner diameter 40 mm) and molybdenum rotor (cylindrical section size 12 mm in diameter and 10 mm in height) were used, and the angles of the top and tail of the rotor were 120°. The schematic diagram of the experimental equipment is shown in Fig. 2.

Fig. 2.

Schematic diagram of experimental equipment. (Online version in color.)

Firstly, the experimental equipment was calibrated with standard silicone oil at room temperature and 100 r·min⁻¹. Secondarily, the sample was heated to 1773 K (1500°C) for 30 mins to obtain a uniform melt. Then the rotor was slowly immersed in the crucible, stopped at 10 mm from the bottom of the crucible, kept concentric with the crucible, and stirred evenly for 10 mins. At last, the viscosity of the mold flux was measured by cooling temperature. The rotor speed was 100 r·min⁻¹, the cooling rate was 10°C·min⁻¹, and the viscosity value reached 10 Pa·s at the end of the measurement. During the whole experiment process, high purity argon was continuously introduced for protection, and the argon flow rate was 0.2 NL·min⁻¹.

2.3. Raman Spectroscopy Measurements

The mold flux was heated to 1773 K (1500°C) in the furnace based on MoSi₂ heating elements, then quenched on water-cooled copper plate, and quenched sample was crushed to less than 74 μm. The XRD results showed that the quenched samples were amorphous without characteristic peaks, as shown in Fig. 3. The structural information of glass samples obtained from quenching was detected by Raman spectroscopy (LabRAM HR evolution, Horiba, jobin, Yvon, France). The experiment was carried out at 25°C, the excitation wavelength was 532 nm, the light source was a semiconductor laser with power of 1 mw, the exposure time was 10 s, and the continuous scanning mode was adopted. The measurement frequency band range was 200–1600 cm⁻¹, the scanning time was 6 times, and the spectral resolution was less than or equal to 0.65 cm⁻¹. Finally, the peak fitting of Raman spectra in the range of 400–1400 cm⁻¹ was performed by using Gaussian function.

Fig. 3.

Water quenching XRD pattern of the mold fluxes at 1773 K. (Online version in color.)

3. Results and Discussion

3.1. Effect of Ce₂O₃ on Viscosity and T_br of the Mold Flux

The average value of the three viscosity tests was used as the viscosity of the mold flux at different temperatures, and the viscosity-temperature curves of the mold flux with different Ce₂O₃ content are shown in Fig. 4.

Fig. 4.

Viscosity-temperature curves of the mold fluxes with different Ce₂O₃ content. (Online version in color.)

It can be seen from Fig. 4 that the viscosity of samples S0–S4 decreases with the increasing temperature. When the temperature is lower than the T_br, the viscosity of the mold flux rises rapidly, reaching 10 Pa·s in a narrow temperature range, the melt changes from Newtonian fluid to non-Newtonian fluid,^24,25) and the fluidity and stability of the mold flux become poor. When T ≥ 1670 K, the content of Ce₂O₃ has less effect on the viscosity of the mold flux. The viscosity of each mold flux changes gently with temperature and the viscosity value shall not exceed 0.03 Pa·s. At this time, the viscosity of the mold flux decreases with the increase of Ce₂O₃ content in the mold flux. In addition, the viscosities at 1573°C are listed in Table 2.

Table 2. Viscosity at 1573 K and T_br of the mold fluxes.

Sample	η (Pa·s)	Tbr (K)
S0	0.29	1418
S1	0.30	1467
S2	0.33	1472
S3	0.35	1486
S4	0.40	1507

The relationship between the activation energy and temperature may be illustrated using the Arrhenius equation, as shown in Eq. (2).²⁶⁾   

$$\eta =Aexp\left(\frac{E_a}{RT}\right)$$(2)

Where, η is the viscosity of the mold flux (Pa·s), A is the pre-exponent factor; E_a is the activation energy for the viscous fluid (J·mol⁻¹), R is the gas constant (8.314 J·mol⁻¹·K⁻¹), and T is the absolute temperature (K). The results are shown in Fig. 5 and the corresponding break temperatures of the mold flux are listed in Table 2.

Fig. 5.

Relationship between lnη and 1/T. (Online version in color.)

It can be seen from Table 2 that the viscosity of the mold flux for samples S0–S4 increases from 0.29 to 0.40 Pa·s with the increase of Ce₂O₃ content at 1573 K, which indicates that when the temperature is lower than the T_br, the viscosity of the mold flux increases with the increase of Ce₂O₃ content in the mold flux. The lowest T_br of mold flux is S0 (1433 K), followed by S1 (1461 K), S2 (1472 K), S3 (1483 K) and S4 (1512 K). With the increase of Ce₂O₃ content in the mold flux, the T_br of the mold flux increases.

T_br is an important parameter of mold flux.^27,28) When the temperature is lower than T_br, a large number of crystalline phases start to precipitate from mold flux, and the viscosity of mold flux will increase rapidly. This indicates that with the increase of Ce₂O₃ content in the mold flux, the crystallization ability of the mold flux increases and the lubricating performance decreases.

The content of Ce₂O₃ in the mold flux should not exceed 15%. On the one hand, when the content of Ce₂O₃ in the mold flux is 15%, the obvious inflection point appears in the viscosity curve of the mold flux, which indicates that the mold flux belongs to the alkaline slag system, as shown in Fig. 4. On the other hand, the T_br of mold flux exceeds 1473 K, which tends to lead to a sharp thinning of liquid slag film and insufficient lubrication of the billet, which is unfavorable for continuous casting production.

3.2. Effect of Ce₂O₃ on High-Temperature Structure of Mold Fluxe

The high-temperature structure of slag directly influences the viscosity of the slag, and the network structure formed by the different oxides act differently at high temperature. Ce₂O₃, as a basic oxide, will become a network modifier like CaO.²⁹⁾ At high temperature, it will break the bridge oxygen bond and form non-bridge oxygen bond, which will lead to the disintegration of slag network structure and the decrease degree of polymerization, thus reducing the viscosity of slag.

In order to understand the polymerization degree of slag at high temperature, NBO/T is often used for characterization, which is expressed as the ratio of the number of non-bridged oxygen atoms to the number of tetrahedral coordination atoms in the slag. The calculation equation is shown in Eq. (3).^30,31)   

$$\begin{array}{l}\mathrm{N}\mathrm{B}\mathrm{O}/\mathrm{T}=\\ {\displaystyle \frac{2x_{\mathrm{C}\mathrm{a}\mathrm{O}}+2x_{\mathrm{C}\mathrm{a}{\mathrm{F}}_2}+2x_{\mathrm{N}{\mathrm{a}}_2\mathrm{O}}+2x_{\mathrm{M}\mathrm{g}\mathrm{O}}-2x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}+2x_{\mathrm{C}{\mathrm{e}}_2{\mathrm{O}}_3}}{x_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2}+2x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}}}\end{array}$$(3)

Where, NBO is the non-bridge oxygen number, T is the number of quaternary coordinated cations, x is the mole fraction of the component. The NBO/T values of the samples S0–S4 are shown in Fig. 6.

Fig. 6.

Change of NBO/T with Ce₂O₃ addition in the mold flux.

It can be seen from Fig. 6 that when Ce₂O₃ component is added to the mold flux with similar mass ratio of other components, NBO/T increases from 2.02 to 2.18, which indicates that the depolymerization degree of the mold flux structure increases. This is due to the fact that Ce₂O₃, as a network modifier, reduces the number of quaternary coordination cations, improves the polymerization degree of the mold flux, and simplifies the high-temperature structure of the mold flux. This will decrease the diffusion resistance of the mold flux components and the energy barrier of crystallization,³²⁾ reduce the viscosity of mold flux and improve the crystallization tendency of the mold flux.

In order to determine the high-temperature structure of mold flux, the structures of samples S0–S4 at 1773 K are analyzed by Raman spectroscopy, as shown in Fig. 7. As can be seen from Fig. 7, the peaks of Raman spectrum are concentrated in three main regions. Among them, the maximum peak value of Raman spectrum curve is located at 800–1100 cm⁻¹, which is the region of symmetrical stretching of [Si–O], and the peak value shifts from 906 to 863 cm⁻¹ with the increase of Ce₂O₃ content in the mold flux. It indicates that with the addition of Ce₂O₃ in the mold flux, the Raman spectrum peaks shift to lower frequency, and the structure of [Si–O] is destroyed to a certain extent, which reflects the role of Ce₂O₃ as a network modifier in the microstructure of the mold flux. In contrast, the characteristic peaks of Raman spectra at 571 cm⁻¹ and 690 cm⁻¹ are not shifted, indicating that Ce did not participate in the [Al–O] structure. This indicates that Ce₂O₃ will preferentially destroy the silicate network structure.

Fig. 7.

Raman spectra of mold fluxes with different Ce₂O₃ content at 1773 K. (Online version in color.)

To further analyze the specific effect of Ce₂O₃ on [Si–O] structure, the signals in the high frequency region 800–1100 cm⁻¹ of Raman in Fig. 7 are fitted, and the results are shown in Fig. 8. It can be seen from the figure that the fitting peaks in the region of 800–1100 cm⁻¹ are 857–852, 903–899, 972–969 and 1039 cm⁻¹ respectively. According to Raman spectrum analysis, the fitting peaks correspond to Q⁰, Q¹, Q² and Q³ structural units in [Si–O] structure respectively, as shown in Table 3. By comparing the shift of fitted peaks, it can be seen that the fitted peak shifts to lower frequency with the increase of Ce₂O₃ content. At this time, the region of these characteristic peaks also changes differently.

Fig. 8.

Raman deconvolution results of (a) sample S0, (b) sample S1, (c) sample S2, (d) sample S3, and (e) sample S4. (Online version in color.)

Table 3. Reference data of Raman active structural units in literature.

Raman bands	Raman shift (cm−1)	Reference
Symmetric stretching vibration $\mathrm{S}{\mathrm{i}}_2{\mathrm{O}}_5^{2-}$, Q3	1040–1060	33,34,35,36,37,38,39
Symmetric stretching vibration $\mathrm{S}{\mathrm{i}}_2{\mathrm{O}}_6^{4-}$, Q2	950–980	33,34,35,36,37,38,39
Symmetric stretching vibration $\mathrm{S}{\mathrm{i}}_2{\mathrm{O}}_7^{6-}$, Q1	900–920	33,34,35,36,37,38,39
Symmetric stretching vibration $\mathrm{S}\mathrm{i}{\mathrm{O}}_4^{4-}$, Q0	850–880	33,34,35,36,37,38,39
Symmetric stretching vibration $\mathrm{A}\mathrm{l}{\mathrm{O}}_4^{5-}$	600–700	40,41,42,43
Symmetric stretching vibration $\mathrm{A}\mathrm{l}{\mathrm{O}}_6^{9-}$	400–600	44,45,46

The relative content of different structural units in the slag can be inferred from the corresponding Raman band area, and then the mole fraction of species in the silicate network ${\mathrm{Q}}_{\mathrm{S}\mathrm{i}}^{\mathrm{n}}$ can be calculated according to Eq. (4).⁴⁷⁾   

$$X_n=\frac{{\theta }_n{A}_n}{\sum _{n=0}^3{\theta }_n{A}_n}$$(4)

Where, X_n is mole fraction (n ranges from 0 to 3), θ_n is Raman scattering coefficient and A_n is Raman band area. Although the exact value of θ_n is still unknown, as an alternative, considering that θ_n is a constant and only determined by the ${\mathrm{Q}}_{\mathrm{S}\mathrm{i}}^{\mathrm{n}}$ species itself, the relative change and conversion of ${\mathrm{Q}}_{\mathrm{S}\mathrm{i}}^{\mathrm{n}}$ are studied by comparing the Raman band area ratio,⁴⁸⁾ and the results are shown in Fig. 9.

Fig. 9.

Relative area fractions of different structural units in the mold flux under different Ce₂O₃ mass fractions. (Online version in color.)

It can be seen from Fig. 9 that when the content of Ce₂O₃ in the mold flux increases from 0 to 20%, the contents of Q⁰ and Q¹ in the [Si–O] structure gradually increase, the content of Q² gradually decreases, and the content of Q³ becomes 0, that is, with the increase of Ce₂O₃ content in the mold flux the relative fraction of [Si–O] tetrahedron (network forming body) in the mold flux decreases.

It is generally believed that the change of relative content and conversion of various structural units will affect the degree of polymerization of slag.⁴⁹⁾ The reason for the depolymerization of network structure may be that Ce₂O₃ is decomposed into Ce³⁺ and O²⁻ in the mold flux, O²⁻ dissociated from Ce₂O₃ participates in [Si–O] structure through Eq. (5).   

$$\begin{array}{l}{\mathrm{O}}^{2-}(\mathrm{f}\mathrm{r}\mathrm{e}\mathrm{e}\mathrm{o}\mathrm{x}\mathrm{y}\mathrm{g}\mathrm{e}\mathrm{n})+{\mathrm{O}}^0(\mathrm{b}\mathrm{r}\mathrm{i}\mathrm{d}\mathrm{g}\mathrm{e}\mathrm{d}\mathrm{o}\mathrm{x}\mathrm{y}\mathrm{g}\mathrm{e}\mathrm{n})\\ =2{\mathrm{O}}^{-}(\mathrm{n}\mathrm{o}\mathrm{n}-\mathrm{b}\mathrm{r}\mathrm{i}\mathrm{d}\mathrm{g}\mathrm{e}\mathrm{d}\mathrm{o}\mathrm{x}\mathrm{y}\mathrm{g}\mathrm{e}\mathrm{n})\end{array}$$(5)

Therefore, Ce₂O₃ can be used as a network modifier, which will destroy the bridge oxygen bond and form non-bridge oxygen bond at high temperature. The disintegration of the network structure simplifies the high-temperature structure of the mold flux and reduces the friction resistance to the viscous flow of the mold flux. From the apparent phenomenon, the viscosity of the mold flux decreases at high temperature.

3.3. Effect of Ce₂O₃ on Crystallization of Mold Flux

The microstructure of melt affects the crystallization behavior of the mold flux and the precipitation of crystalline phase also affects the crystallization temperature of the mold flux. Therefore, in order to explore the effect of Ce₂O₃ on the crystallization behavior of the mold flux, XRD analysis was carried out on the low-temperature water quenched slag of samples S0–S4, as shown in Fig. 10.

Fig. 10.

XRD of different Ce₂O₃ mass fractions. (Online version in color.)

It can be seen from Fig. 10 that the main crystalline phases in the mold flux are Ca₄Si₂O₂F, Na₆Al₄Si₄O₁₇, Ca₂Al₂SiO₇ and Ca₃MgSi₂O₈. When the mass fraction of Ce₂O₃ in the mold flux increases to 10%, Ce_9.33(SiO₄)₆O₂ begins to appear in the mold flux. With the increase of Ce₂O₃ content in mold flux, the peak value of Ce_9.33(SiO₄)₆O₂ increases, and the peak values of Ca₄Si₂O₂F and Na₆Al₄Si₄O₁₇ decrease significantly. This proves that Ce₂O₃ participates in [Si–O] structure, forms cerium silicate, and enhances the crystallization ability of mold flux. In addition, since the melting point of Ce_9.33(SiO₄)₆O₂ (1770°C)¹⁶⁾ is higher than Ca₄Si₂O₂F (1407°C),⁵⁰⁾ the T_br of mold flux will increase with the increase of Ce₂O₃ content in mold flux.

4. Conclusions

In this paper, the viscosity, structure and crystallization of 30%CaO-30%SiO₂-15%Al₂O₃-5%MgO-10%Na₂O-10%CaF₂-xCe₂O₃ (mass%, x = 0, 5, 10, 15) were studied. The conclusions are as follows:

(1) The results of viscosity experiment show that when Ce₂O₃ content increases from 0% to 20% in the mold flux, the high-temperature viscosity of mold flux decreases, and the T_br increases from 1433 K to 1515 K. It is more appropriate to add 15% Ce₂O₃ to mold flux.

(2) The results of NBO/T and Raman spectra show that with the increase of Ce₂O₃ content in the mold flux, the depolymerization degree of the mold flux structure increases and the high-temperature structure becomes simpler. The peak of Raman spectrum shifts from 906 cm⁻¹ to 863 cm⁻¹, and the addition of Ce₂O₃ will preferentially destroy the silicate network structure in the mold flux.

(3) The results of XRD show that when content of Ce₂O₃ in mold flux increases to 10%, Ca₄Si₂O₂F in mold flux changes to Ce_9.33(SiO₄)₆O₂, and the crystallization ability of mold flux increases.

(4) Ce₂O₃ can be used as a network modifier, which will destroy the bridge oxygen bond and form non-bridge oxygen bond at high temperature. The disintegration of the network structure simplifies the high-temperature structure of the mold flux, reduces the friction resistance of the viscous flow of the mold flux. From the apparent phenomenon, the viscosity of the mold flux decreases at high temperature.

Acknowledgements

This work was supported by the National Natural Science Foundation of China [U1760206].

References

• 1)   Q.  Wang,  S. P.  He,  Y. G.  Li,  L. L.  Zhu,  T.  Wu and  Z.  Chen: Contin. Cast., 5 (2014), 1. https://doi.org/10.13228/j.boyuan.issn1005-4006.20140104
• 2)   K. C.  Mills and  A. B.  Fox: ISIJ Int., 43 (2003), 1479. https://doi.org/10.2355/isijinternational.43.1479
• 3)   M.  Kawamoto,  K.  Nakajima,  T.  Kanazawa and  K.  Nakai: ISIJ Int., 34 (1994), 593. https://doi.org/10.2355/isijinternational.34.593
• 4)   W. L.  Wang,  E. Z.  Gao,  L. J.  Zhou,  L.  Zhang and  H.  Li: J. Iron Steel Res. Int., 26 (2019), 355. https://doi.org/10.1007/s42243-018-0207-z
• 5)   Z.  Wang,  Q. F.  Shu and  K.  Chou: Metall. Mater. Trans. B, 44 (2013), 606. https://doi.org/10.1007/s11663-013-9803-2
• 6)   J. Y.  Li,  G. G.  Cheng,  Q.  Ruan,  J. X.  Pan and  X. R.  Chen: Metall. Mater. Trans. B, 50 (2019), 2769. https://doi.org/10.1007/s11663-019-01708-8
• 7)   C. X.  Ji,  Y.  Cui,  Z.  Zeng,  Z. H.  Tian,  C. L.  Zhao and  G. S.  Zhu: J. Iron Steel Res. Int., 22 (2015), 53. https://doi.org/10.1016/S1006-706X(15)30138-2
• 8)   L. J.  Zhou,  H. F.  Wu,  W. L.  Wang,  H.  Luo and  H.  Li: Ceram. Int., 47 (2021), 10940. https://doi.org/10.1016/j.ceramint.2020.12.213
• 9)   J. W.  Cho,  K.  Blazek,  M.  Frazee,  H. B.  Yin,  J. H.  Park and  S. W.  Moon: ISIJ Int., 53 (2013), 62. https://doi.org/10.2355/isijinternational.53.62
• 10)   C. B.  Shi,  M. D.  Seo,  J. W.  Cho and  S. H.  Kim: Metall. Mater. Trans. B, 45 (2014), 1081. https://doi.org/10.1007/s11663-014-0034-y
• 11)   Q.  Wang,  S. T.  Chou,  P.  Zhao,  Z. H.  Liu,  G. L.  Zhu and  K. T.  Hu: Steelmaking, 28 (2012), 74.
• 12)   B. X.  Lu,  K.  Chen,  W. L.  Wang and  B. B.  Jiang: Metall. Mater. Trans. B, 45 (2014), 1496. https://doi.org/10.1007/s11663-014-0063-6
• 13)   W. L.  Wang,  B. X.  Lu and  D.  Xiao: Metall. Mater. Trans. B, 47 (2016), 384. https://doi.org/10.1007/s11663-015-0474-z
• 14)   G. H.  Kim and  I.  Sohn: Metall. Mater. Trans. B, 45 (2014), 86. https://doi.org/10.1007/s11663-013-9953-2
• 15)   J. X.  Gao,  G. H.  Wen,  T.  Huang and  P.  Tang: Proc. 10th Int. Conf. on Molten Slags, Fluxes and Salts, (Seattle), John Wiley and Sons, Hoboken, NJ, (2016), 291. https://doi.org/10.1002/9781119333197.ch31
• 16)   A. C.  Tas and  M.  Akinc: J. Am. Ceram. Soc., 77 (1994), 2953. https://doi.org/10.1111/j.1151-2916.1994.tb04530.x
• 17)   N. M.  Anacleto,  H. G.  Lee and  P. C.  Hayes: ISIJ Int., 33 (1993), 549. https://doi.org/10.2355/isijinternational.33.549
• 18)   S. J.  Jeong,  T. S.  Kim and  J. H.  Park: Metall. Mater. Trans. B, 48 (2017), 545. https://doi.org/10.1007/s11663-016-0828-1
• 19)   T. S.  Kim,  S. J.  Jeong and  J. H.  Park: Met. Mater. Int., 26 (2020), 1872. https://doi.org/10.1007/s12540-019-00474-1
• 20)   J.  Qi,  C. J.  Liu and  M. F.  Jiang: Ceram. Int., 47 (2021), 20850. https://doi.org/10.1016/j.ceramint.2021.04.083
• 21)   G. H.  Kim and  I.  Sohn: Metall. Mater. Trans. B, 42 (2011), 1218. https://doi.org/10.1007/s11663-011-9561-y
• 22)   G. H.  Kim and  I.  Sohn: ISIJ Int., 52 (2012), 68. https://doi.org/10.2355/isijinternational.52.68
• 23)   H.  Park,  J. Y.  Park,  G. H.  Kim and  I.  Sohn: Steel Res. Int., 83 (2012), 150. https://doi.org/10.1002/srin.201100249
• 24)   Y. M.  Gao,  S. B.  Wang,  C.  Hong,  X. J.  Ma and  F.  Yang: Int. J. Miner. Metall. Mater., 21 (2014), 353. https://doi.org/10.1007/s12613-014-0916-7
• 25)   H. S.  Park,  H.  Kim and  I.  Sohn: Metall. Mater. Trans. B, 42 (2011), 324. https://doi.org/10.1007/s11663-011-9474-9
• 26)   Z. L.  Piao,  L. G.  Zhu,  X. J.  Wang,  Z. X.  Liu,  B.  Wang,  P. C.  Xiao and  Z. P.  Yuan: Metall. Mater. Trans. B, 51 (2020), 2119. https://doi.org/10.1007/s11663-020-01915-8
• 27)   T.  Iida,  H.  Sakai,  Y.  Kita and  K.  Murakami: High Temp. Mater. Process., 19 (2000), 153. https://doi.org/10.1515/HTMP.2000.19.3-4.153
• 28)   S.  Sridhar,  K. C.  Mills,  O. D. C.  Afrange,  H. P.  Lörz and  R.  Carli: Ironmaking Steelmaking, 27 (2000), 238. https://doi.org/10.1179/030192300677534
• 29)   Z.  Wang,  Q. F.  Shu and  K.  Chou: Steel Res. Int., 84 (2013), 766. https://doi.org/10.1002/srin.201200256
• 30)   K. C.  Mills and  S.  Sridhar: Ironmaking Steelmaking, 26 (1999), 262. https://doi.org/10.1179/030192399677121
• 31)   J. W.  Cho,  H.  Shibata,  T.  Emi and  M.  Suzuki: ISIJ Int., 38 (1998), 440. https://doi.org/10.2355/isijinternational.38.440
• 32)   C. B.  Shi,  J.  Li,  J. W.  Cho,  F.  Jiang and  I. H.  Jung: Metall. Mater. Trans. B, 46 (2015), 2110. https://doi.org/10.1007/s11663-015-0402-2
• 33)   B. O.  Mysen,  D.  Virgo and  C. M.  Scarfe: Am. Mineral., 65 (1980), 690. https://doi.org/10.1590/1519-6984.14712
• 34)   B. O.  Mysen,  L. W.  Finger,  D.  Virgo and  F. A.  Seifert: Am. Mineral., 67 (1982), 686. https://doi.org/10.1016/0040-1951(83)90053-7
• 35)   B. O.  Mysen and  J. D.  Frantz: Contrib. Mineral. Petrol., 117 (1994), 1. https://doi.org/10.1007/BF00307725
• 36)   P.  González,  J.  Serra,  S.  Liste,  S.  Chiussi,  B.  León and  M.  Pérez-Amor: J. Non-Cryst. Solids, 320 (2003), 92. https://doi.org/10.1016/S0022-3093(03)00013-9
• 37)   P.  McMillan: Am. Mineral., 69 (1984), 622. https://doi.org/10.1007/BF01082965
• 38)   D. W.  Matson,  S. K.  Sharma and  J. A.  Philpotts: J. Non-Cryst. Solids, 58 (1983), 323. https://doi.org/10.1016/0022-3093(83)90032-7
• 39)   K.  Fukumi,  J.  Hayakawa and  T.  Komiyama: J. Non-Cryst. Solids, 119 (1990), 297. https://doi.org/10.1016/0022-3093(90)90302-3
• 40)   S. L.  Lin and  C. S.  Hwang: J. Non-Cryst. Solids, 202 (1996), 61. https://doi.org/10.1016/0022-3093(96)00138-X
• 41)   G.  Leekes,  N.  Nowack and  F.  Schlegelmilch: Steel Res., 59 (1988), 406. https://doi.org/10.1002/srin.198801533
• 42)   P. F.  McMillan,  W. T.  Petuskey,  B.  Coté,  D.  Massiot,  C.  Landron and  J. P.  Coutures: J. Non-Cryst. Solids, 195 (1996), 261. https://doi.org/10.1016/0022-3093(95)00536-6
• 43)   C.  Huang and  E. C.  Behrman: J. Non-Cryst. Solids, 128 (1991), 310. https://doi.org/10.1016/0022-3093(91)90468-L
• 44)   S. S.  Jung and  I.  Sohn: Environ. Sci. Technol., 48 (2014), 1886. https://doi.org/10.1021/es404277w
• 45)   P.  Tarte: Spectrochim. Acta A, 23 (1967), 2127. https://doi.org/10.1016/0584-8539(67)80100-4
• 46)   B. O.  Mysen,  D.  Virgo,  W. J.  Harrison and  C. M.  Scarfe: Am. Mineral., 65 (1980), 900.
• 47)   Y. Q.  Sun,  Z. T.  Zhang,  L. L.  Liu and  X. D.  Wang: J. Non-Cryst. Solids, 420 (2015), 26. https://doi.org/10.1016/j.jnoncrysol.2015.04.017
• 48)   Z. N.  Jin,  H. Y.  Yang,  J. F.  Lv,  L. L.  Tong,  G. B.  Chen and  Q.  Zhang: JOM, 70 (2018), 1430. https://doi.org/10.1007/s11837-017-2660-8
• 49)   Q. H.  Li,  J. T.  Gao,  Y. L.  Zhang, Z. Q. An and Z. C. Guo: Metall. Mater. Trans. B, 48 (2017), 346. https://doi.org/10.1007/s11663-016-0858-8
• 50)   T.  Watanabe,  H.  Fukuyama,  K.  Nagata and  M.  Susa: Metall. Mater. Trans. B, 31 (2000), 1273. https://doi.org/10.1007/s11663-000-0015-1