Effect of Ce₂O₃ on the Melt Structure and Properties of CaO–Al₂O₃-based Slag

Abstract

The low-reactive tundish flux is urgently required for the smelting of rare-earth steels. The effect of Ce₂O₃ content on the properties (melting point and viscosity) and structure of CaO–Al₂O₃–Ce₂O₃–MgO–SiO₂ slag were analyzed in the present investigation. The research results indicated that the polymerized Q_Al⁴ and Q_Al³ units were modified to Q_Al² units in aluminate, and the depolymerization of SiO₄-tetrahedrons from Q_Si¹ to Q_Si⁰ units with Ce₂O₃ addition increasing from 5 wt% to 20 wt%. In addition, when the Ce₂O₃ addition increased from 15 wt% to 20 wt%, the increasing area fraction in AlO₆-octahedral units further led to a decrease in the degree of polymerization (DOP) in the melts. When the Ce₂O₃ addition increased from 5 wt% to 20 wt%, the melting temperature decreased from 1320°C to 1301°C. Additionally, the softening temperature and the fluidity temperature also decreased 20°C and 24°C, respectively. The viscosity decreased as the Ce₂O₃ addition increased from 5 wt% to 20 wt%, which was attributed to the decreased DOP and the increased superheat. A reasonable content for Ce₂O₃ should not be excess 10 wt% in the current slag.

1. Introduction

Rapid advancement in the technology of steelmaking created possibilities for rare earth micro-alloying in steel.¹⁾ For instance, the addition of a certain amount of Ce in 310S type austenitic stainless steel led to an increase in high-temperature oxidation resistance.²⁾ However, the addition of Ce in steel also causes problems in steelmaking processes. Due to the low yield rate of cerium, it should be added excessively in the refining process during the smelt of 253MA rare earth heat resistant steel. The soluble Ce in liquid steels is highly active at high temperatures and can easily react with the SiO₂ (cf. Eq. (1)), a major component of traditional CaO–SiO₂-based tundish flux during continuous casting processes. A severe slag-steel reaction would lead to a lower micro-alloying effect of the soluble Ce in liquid steel. Moreover, the reaction would cause the dramatic variation of tundish flux chemical composition and, consequently, its physicochemical properties (e.g., melting temperature and viscosity). To restrain the reaction, the calcium aluminates system has been introduced.³⁾   

$$4\left[\mathrm{C}\mathrm{e}\right]+3\left(\mathrm{S}\mathrm{i}{\mathrm{O}}_2\right)=2\left(\mathrm{C}{\mathrm{e}}_2{\mathrm{O}}_3\right)+3\left[\mathrm{S}\mathrm{i}\right]$$(1)

Many researchers have focused on the properties of the CaO–Al₂O₃-based slag and low-silica calcium aluminate slags. Kim et al.⁴⁾ investigated the structure-viscosity relationship of the low-silica (SiO₂ ≤ 10 wt%) calcium aluminosilicate (CaO–Al₂O₃–SiO₂–MgO–CaF₂) melts. Besides, the relationship between structure and viscosity with different low-silica CaO–Al₂O₃-based slags have also been investigated by Wang et al.,⁵⁾ Gao et al.,⁶⁾ and Shao et al.⁷⁾ So far, there have been several reports on the properties of the slag containing rare earth oxide. Qi et al.⁸⁾ found that the addition of Ce₂O₃ led to the reduction of the viscosity by decreasing the polymerization of the CaO–Al₂O₃–Li₂O–Ce₂O₃ mold flux. Cai et al.⁹⁾ also found the CeO₂ reduced the viscosity of CaO–SiO₂-based mold flux due to the CeO₂ enhanced the depolymerization of the network structure. Similar results were also found in the low-aluminate CaO–SiO₂–Al₂O₃–CeO₂ system.¹⁰⁾ Wu et al.¹¹⁾ investigated the viscosity and melting temperature of the CaO-Al₂O₃-10wt%SiO₂-Ce₂O₃ quaternary refining slag at 1500°C. Tundish flux plays an important role in affecting the quality of the steel. The CaO–Al₂O₃-based tundish fluxes have been successfully applied to high-Al steel.³⁾ However, it is still difficult to ensure long-time use of the tundish flux for rare earth steels. In order to design a low-reactivity tundish flux for rare earth steels, Ce₂O₃ was introduced to the CaO–Al₂O₃-based tundish flux by the present author. Therefore, a reasonable range of Ce₂O₃ content should be determined to ensure suitable properties of the tundish flux.

In the current study, the effect of Ce₂O₃ content on the viscosity, melting properties, and structure of the CaO–Al₂O₃–Ce₂O₃–MgO–SiO₂ systems were investigated by the hemispherical melting point method, the rotating cylinder method, and Raman spectroscopy, respectively. The research results can provide theoretical guidance in the design of low-reactivity tundish flux for rare earth steel.

2. Materials and Methods

2.1. Sample Preparation

Analytical reagent (AR) grade calcium oxide (CaO), aluminum oxide (Al₂O₃), magnesium oxide (MgO), silicon oxide (SiO₂), and cerium oxide (CeO₂) were taken as raw materials and calcined at 1000°C (1273 K) for 4 h to remove retaining water and to decompose any carbonate. Compositions of the samples are listed in Table 1. Note that the CaO–Al₂O₃–SiO₂–MgO system slags were considered as the initial slags, and a certain amount of Ce₂O₃ (5 wt%, 10 wt%, 15 wt%, and 20 wt%) was extra added to these slags. About 150 g samples were fully charged in a graphite crucible and placed in a furnace under Ar atmosphere (purity, > 99%). The oxygen partial pressure in the furnace was less than 10⁻¹² atm, which was measured by a DS oxygen probe produced by Australian Oxytrol System Pty. Ltd. In this condition, cerium is considered as trivalent in the flux. Qi et al.,⁸⁾ Ueda et al.,¹²⁾ and Kim et al.¹³⁾ reported similar results. The samples were melted at 1550°C for 2 h to reach homogenization, and then the samples were quenched with ice water rapidly to obtain the glassy phase. The glassy phase was used to investigate the properties and structure of the slags. X-ray diffraction (XRD, D8 ADVANCE, Bruker, Germany) analysis and scanning electron microscopy (SEM, Phenom Pro X, Netherlands) were used to identify the non-crystalline state of the as-quenched slag. The result shown in Fig. 1 indicates that the as-quenched slag is completely amorphous.

Table 1. Composition of the samples (mass fraction/%).

Exp. Set	CaO/Al2O3	CaO	Al2O3	SiO2	MgO	Ce2O3
01	1	42.86	42.86	4.76	4.76	4.76
02	1	40.91	40.91	4.55	4.55	9.09
03	1	39.13	39.13	4.35	4.35	13.04
04	1	37.50	37.50	4.17	4.17	16.67

Fig. 1.

Typical XRD result and SEM image of the as-quenched slag. (a) XRD result, (b) SEM image. (Online version in color.)

2.2. Raman Analysis of the Melts

The anionic structural units in the aluminosilicate melts are the same before and after quenching have been confirmed by Mysen et al.^14,15) The as-quenched slags were analyzed by Raman spectroscopy to determine the effect of Ce₂O₃ on the structure of the slags. Raman spectra were recorded in the frequency range of 400–1200 cm⁻¹ on a laser confocal with a CCD detector (HR800 UV, HORIBA Jobin Yvon, Paris, France). The excitation source was 633 nm wavelength laser and the resolution of the spectrum was 0.65 cm⁻¹. After obtaining the spectra, a baseline subtraction of the spectra was done by the OriginPro 2018 C software (Northampton, Massachusetts, USA). The peaks were deconvoluted and analyzed to characterize the corresponding structures.

2.3. Melting Properties Measurements

Melting temperature was analyzed by the hemispherical melting point method. The melting temperature is determined as the temperature at which half of the original height, i.e., hemispherical temperature. In addition, the shrinkage values of 25% and 75% were corresponding to the softening temperature, and fluidity temperature, respectively. The pre-melted slags were processed into a cylinder sample (diameter: 3 mm, height: 3 mm) which was heated with a rate of 10°C·min⁻¹ in the electric furnace. Images and temperatures for the sample shrinkage value were 25%, 50%, and 75% were collected by the computer. For each sample, three measurements were performed. If the difference between the three measurements exceeded 3°C, the sample was tested again.

2.4. Viscosity Measurements

The viscosity of the sample was analyzed by the rotating cylinder method. Before the viscosity measurements of slag, the viscometer was calibrated using the standard castor oil (0.986 Pa·s at 20°C, 0.651 Pa·s at 25°C, etc.) with known viscosity at room temperature. The same method has been reported by Feng et al.,¹⁶⁾ and Zhang et al.¹⁷⁾ An amount of 140 g of the pre-melted slag was put into a graphite crucible (inside diameter: 40 mm; height: 80 mm) and heated to 1550°C (1823 K) under an Ar atmosphere (purity, > 99%; flow rate, 1 L·min⁻¹) for 30 minutes to make the samples completely melted. The oxygen partial pressure in the furnace was also lower than 10⁻¹² atm. The experimental apparatus for viscosity measurement is shown in Fig. 2. The Mo spindle (diameter: 10 mm; height: 25 mm) was slowly immersed in the molten slag and kept at a distance of 10 mm above the crucible bottom. The viscosity of the measured sample remained stable with altering rotation speed, indicating that the molten slag had been completely melted and was Newtonian fluid. The viscosity value and temperature were recorded by the computer. The viscosity was measured following YB/T 185-2017 Standard of China.^18,19) For the experiment, the temperature of the furnace was decreased at a slow cooling rate of 3°C·min⁻¹ to ensure more stable experimental conditions and the Mo spindle rotated at a speed of 200 r·min⁻¹.^8,16,17) The experiment ended with a viscosity of 5 Pa·s. To ensure the accuracy of the viscosity value at high temperature (temperature above the breaking temperature), the viscosity of each slag was remeasured under a thermally homogeneous state at 15°C intervals. The slag was held isothermally for 30 minutes before each viscosity measurement.

Fig. 2.

Experimental apparatus for viscosity measurement. (Online version in color.)

3. Results and Discussion

3.1. Theoretical Analysis on Structure

In the current slag systems, the high content of Al₂O₃ would lead to an amount of AlO₄-tetrahedral formed in the molten slag. Unlike Si atoms which formed tetrahedral, Al atoms required charge compensation by alkali metal ions or alkaline earth metal ions because of the only three electrons in the outermost layer.²⁰⁾ As a typical amphoteric ion, apart from forming four-coordination, Al³⁺ ions could form five-, and six- coordination. According to Roy et al.²¹⁾ and Sukenaga et al.,²²⁾ the sequence for charge compensation of AlO₄-tetrahedral units by cations depends on the cationic field strength (Z/r², Z is the valence number of the metal ions, and r is the corresponding cation radius). A Ce³⁺ ion has the radius of 1.01 Å, which is larger than the radius of a Ca²⁺ ion (1.0 Å) and Mg²⁺ ion (0.69 Å).^10,13,23) The cationic field strengths were calculated and listed in Table 2. As can be seen from the table, Ca²⁺ ion had the smallest cationic field strength in the current system. Therefore, AlO₄-tetrahedral units would be preferentially compensated by Ca²⁺ > Ce³⁺ > Mg²⁺. All of the slag compositions listed in Table 1 fulfill $x_{\mathrm{C}\mathrm{a}\mathrm{O}}>x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}$ and $x_{\mathrm{C}\mathrm{a}\mathrm{O}}+x_{\mathrm{M}\mathrm{g}\mathrm{O}}+3x_{\mathrm{C}{\mathrm{e}}_2{\mathrm{O}}_3}-x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}<2(2x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}+x_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2})$ (e.g., xCaO is the molar fraction of CaO calculated based on Table 1), which means there are enough cations for charge compensation in the present slags.²⁴⁾ $x_{\mathrm{C}\mathrm{a}\mathrm{O}}>x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}$ means the component is located in the peralkaline region, which indicated that the melt contains an excess of cations after charge balancing Al³⁺. $x_{\mathrm{C}\mathrm{a}\mathrm{O}}+x_{\mathrm{M}\mathrm{g}\mathrm{O}}+3x_{\mathrm{C}{\mathrm{e}}_2{\mathrm{O}}_3}-x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}<2(2x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}+x_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2})$ means the number of free oxygens less than the number of bridging oxygens in the melt, i.e., there are still bridging oxygens in the melt.^25,26) That is to say, Ce³⁺ ions were considered as a network modifier in the current system. In addition, the effect of Ce₂O₃ on the melt structure was further analyzed via Raman spectroscopy.

Table 2. Cationic field strength of Ca²⁺, Ce³⁺, and Mg²⁺ ion.

Ion	Ca2+	Ce3+	Mg2+
Field strength	2	2.94	4.20

3.2. Raman Analysis on Melt Structure

Raman spectra of the slags with various Ce₂O₃ contents are shown in Fig. 3. The spectra could be divided into two regions: the low-frequency region (LF: 400 to 700 cm⁻¹) and the high-frequency region (HF: 700 to 1200 cm⁻¹). According to the related literature, the spectra pattern near 550 cm⁻¹ is symmetric stretching vibrations of Al–O–Al.²⁷⁾ The peak around 600 cm⁻¹ is the Al–O stretching vibration in AlO₆ units.²⁸⁾ The characteristic peaks within 700–740 cm⁻¹, 740–790 cm⁻¹, and 800–850 cm⁻¹ are symmetric stretching vibrations of AlO₄ with Q_Al² ([AlO₃]³⁻), Q_Al³ ([Al₂O₅]⁴⁻), and Q_Al⁴ ([AlO₂]⁻), respectively.^4,29,30) The structural units of Q_Si⁰ ([SiO₄]⁴⁻), Q_Si¹ ([Si₂O₇]⁶⁻), Q_Si² ([SiO₃]²⁻), and Q_Si³ ([Si₂O₅]²⁻) correspond to the peaks that in the range of 850–880 cm⁻¹, 900–920 cm⁻¹, 950–1000 cm⁻¹, and 1050–1100 cm⁻¹, respectively.^{14,15,20,31,32)} Variations in Raman spectra could lead to the variation of the degree of polymerization (DOP) and thereby affect the viscosity. To quantitatively identify the fraction of characteristic structural units in the melts, the Raman spectra were deconvoluted. The deconvoluted Raman spectra with various Ce₂O₃ contents are illustrated in Fig. 4. The peaks of Raman spectra of structural units were obtained based on the fitting of Gaussian function until the R² (R is the correlation coefficient) value > 0.99. The assignments of Raman bands of the current slags are listed in Table 3.

Fig. 3.

Raman results of the as-quenched slags with various Ce₂O₃ content. (Online version in color.)

Fig. 4.

Deconvoluted Raman spectra obtained for the slags with different Ce₂O₃ content. (a) 5 wt%, (b) 10 wt%, (c) 15 wt%, (d) 20 wt%. (Online version in color.)

Table 3. Peaks analysis of Raman bands of the CaO–Al₂O₃–Ce₂O₃–MgO–SiO₂ slag.

Mass% Ce2O3	5	10	15	20	Raman bonds
Raman shift (cm−1)	552	550	550	550	symmetric stretching vibrations of Al–O–Al
	613	600	598	600	symmetric stretching vibrations of AlO6
	732	735	731	733	symmetric stretching vibrations of AlO4 with NBO/Al = 2, QAl2
	774	779	777	777	symmetric stretching vibrations of AlO4 with NBO/Al = 1, QAl3
	837	843	840	840	symmetric stretching vibrations of AlO4 with NBO/Al = 0, QAl4
	872	876	878	877	symmetric stretching vibrations of SiO4 with NBO/Si = 4, QSi0
	903	905	907	906	symmetric stretching vibrations of SiO4 with NBO/Si = 3, QSi1
R2	0.991	0.992	0.994	0.997

The scattering efficiency had been used to quantify the relative fraction of silicate structural units and the mole fraction of Q_Siⁿ species can be calculated by Eq. (2).³³⁾   

$$x_n=\frac{\frac{A_n}{S_n}}{\sum _{n=0}^3\frac{A_n}{S_n}}$$(2)

where x_n is the mole fraction of Q_Siⁿ, A_n is the area fraction of Q_Siⁿ, S_n is the Raman scattering coefficient. However, the Raman scattering coefficients of various Q_Alⁿ species have not been provided clearly. Therefore, the variation of area fraction of each structural unit was used to represent the variation of structural units in the aluminosilicate system.³⁴⁾

The integrated percentage of the various structural units with different Ce₂O₃ contents are demonstrated in Fig. 5. Figure 5(a) shows the aera fraction of Q_Alⁿ (n=2, 3, and 4) change with the Ce₂O₃ content, and the aera fraction of Al–O–Al linkage is also depicted in Fig. 5(a). The decrease aera fraction of Al–O–Al linkage meant that the number of bridging oxygen (BO) in the aluminate structure decrease with an increase of Ce₂O₃ content. As a result, the aluminate structure was partly broken by the increase of Ce₂O₃ content. The aera fraction of the polymerized aluminate unit Q_Al⁴ decreased dramatically, whereas the aera fraction of Q_Al³ decreased relatively smoothly. The increase of depolymerized Q_Al² aluminate unit in the current melts indicated that the full polymerized Q_Al⁴ aluminate unit was modified to Q_Al³ and Q_Al² units and the Q_Al³ units also turned into Q_Al² units with the addition of Ce₂O₃ content. Nevertheless, the aera fraction of AlO₆ units remained almost constant when the Ce₂O₃ addition increased from 5 wt% to 15 wt%. The depolymerization process of the aluminate structure units is shown in Eqs. (3), (4), (5), (6) and Fig. 6(a) when the Ce₂O₃ content increased from 5 wt% to 15 wt%.   

$$\mathrm{C}{\mathrm{e}}_2{\mathrm{O}}_3=2\mathrm{C}{\mathrm{e}}^{3+}+3{\mathrm{O}}^{2-}$$(3)

  

$${\mathrm{Q}}_{\mathrm{A}\mathrm{l}}^4\to {\mathrm{Q}}_{\mathrm{A}\mathrm{l}}^3:2{[\mathrm{A}\mathrm{l}{\mathrm{O}}_2]}^{-}+{\mathrm{O}}^{2-}={[\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_5]}^{4-}$$(4)

  

$${\mathrm{Q}}_{\mathrm{A}\mathrm{l}}^4\to {\mathrm{Q}}_{\mathrm{A}\mathrm{l}}^2:{[\mathrm{A}\mathrm{l}{\mathrm{O}}_2]}^{-}+{\mathrm{O}}^{2-}={[\mathrm{A}\mathrm{l}{\mathrm{O}}_3]}^{3-}$$(5)

  

$${\mathrm{Q}}_{\mathrm{A}\mathrm{l}}^3\to {\mathrm{Q}}_{\mathrm{A}\mathrm{l}}^2:{[\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_5]}^{4-}+{\mathrm{O}}^{2-}=2{[\mathrm{A}\mathrm{l}{\mathrm{O}}_3]}^{3-}$$(6)

Fig. 5.

Area fraction of the characteristic structural units with various Ce₂O₃. (a) Q_Alⁿ units and Al–O–Al linkages, (b) Q_Siⁿ units. (Online version in color.)

Fig. 6.

Effect of Ce₂O₃ on the low-silica calcium aluminate melts. (a) Depolymerization process of AlO₄ units, (b) Transformations of Al–O groups, (c) Depolymerization process of SiO₄ units. (Online version in color.)

When the content of Ce₂O₃ increased from 15 wt% to 20 wt%, the aera fraction of AlO₆-octahedral units began to increase obviously in the melts (cf. Fig. 5(a)). With the increase of Ce₂O₃ content in the slag, the relative proportions of O²⁻ ions were increased, and O²⁻ ions reacted with the AlO₄-tetrahedral units and formed the AlO₆-octahedral units. Similar results have been reported by Qi et al.⁸⁾ and Park et al.³⁵⁾ Referring to Kim et al.¹³⁾ and Lin et al.’s³⁶⁾ work, cerium ion has a preferential charge balance role to AlO₆-octahedral units, resulting in the formation of the hypothetical [AlO₆]⁹⁻: Ce³⁺ unit. Therefore, the transform of AlO₄ units to AlO₆ units can be represented in Eq. (7), and the schematic diagram is shown in Fig. 6(b). The AlO₆-octahedral units, as a network modifier, would break the network. Increasing the fraction of AlO₆-octahedral units would lead to a reduction in the degree of polymerization in the slag. Therefore, the decrease of DOP was attributed to the depolymerization of AlO₄-tetrahedral units (cf. Eqs. (3), (4), (5), (6)) and the increase of AlO₆-octahedral units when the content of Ce₂O₃ increased from 15 wt% to 20 wt%.   

$${[\mathrm{A}\mathrm{l}{\mathrm{O}}_4]}^{5-}+2{\mathrm{O}}^{2-}={[\mathrm{A}\mathrm{l}{\mathrm{O}}_6]}^{9-}$$(7)

The aera fraction of Q_Si⁰ structural units increases sharply whereas the aera fraction of Q_Si¹ structural units decreases, as shown in Fig. 5(b). This meant the silicate structure was broken with the increasing Ce₂O₃ content. The depolymerization process of the silicate structure unit is shown in Eq. (8) and Fig. 6(c). McMillan et al.²⁶⁾ measured a Raman spectroscopy of CaO–Al₂O₃–SiO₂ glasses and suggested that four peaks at 1140 cm⁻¹, 1000 cm⁻¹, 925 cm⁻¹, and 890 cm⁻¹ were the symmetric stretching vibrations of SiO₄-tetrahedral units in conjunction with one, two, three, and four fully polymerized aluminate, respectively. Thus, it is suggested that the high-frequency vibrations responsible for aluminosilicate bond systems were -Si(OAl)₃ (three fully polymerized aluminate), and Si(OAl)₄ (four fully polymerized aluminate) in the current melts. However, these Si–O–Al bonds vibration regions are overlapped with the symmetric stretching vibrations of SiO₄-tetrahedral units in the Raman spectroscopy.⁴⁾   

$${\mathrm{Q}}_{\mathrm{S}\mathrm{i}}^1\to {\mathrm{Q}}_{\mathrm{S}\mathrm{i}}^0:{[\mathrm{S}{\mathrm{i}}_2{\mathrm{O}}_7]}^{6-}+{\mathrm{O}}^{2-}=2{[\mathrm{S}\mathrm{i}{\mathrm{O}}_4]}^{4-}$$(8)

To evaluate the melt structure evolution of the current slags, the NBO/T and the corrected optical basicity (Λ^corr) are used to describe the DOP of the molten slag.^28,37) NBO/T is the ratio of the molar fraction available network-breaking oxides (where available indicates the total number of cations minus those on charge-balancing duties) divided by the molar fraction of the network-forming oxides and the theoretical NBO/T ratio is defined by the mole fraction of the slag components, as shown in Eq. (9).^38,39) However, the effect of Al₂O₃ as the network modifier has not been considered in Eq. (9). The experimental NBO/T ratio (which represents the non-bridging oxygen of the tetrahedron) is calculated via Eq. (10). The corrected optical basicity (Λ^corr) can be calculated via Eq. (11). The optical basicity of CaO, Al₂O₃, SiO₂, MgO, and Ce₂O₃ is 1, 0.6, 0.48, 0.78, and 0.88, respectively.^39,40)   

$$\frac{\mathrm{N}\mathrm{B}\mathrm{O}}{\mathrm{T}}=\frac{2\left(x_{\mathrm{C}\mathrm{a}\mathrm{O}}+x_{\mathrm{M}\mathrm{g}\mathrm{O}}+3x_{\mathrm{C}{\mathrm{e}}_2{\mathrm{O}}_3}-x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}\right)}{x_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2}+2x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}}$$(9)

  

$$\frac{\mathrm{N}\mathrm{B}\mathrm{O}}{{\mathrm{T}}_{exp}}=\sum x_i\cdot (4-i)$$(10)

  

$$\begin{array}{l}{\mathrm{\Lambda }}^{corr}=\\ \frac{[{\mathrm{\Lambda }}_{\mathrm{C}\mathrm{a}\mathrm{O}}\times (x_{\mathrm{C}\mathrm{a}\mathrm{O}}-x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3})+{\mathrm{\Lambda }}_{\mathrm{M}\mathrm{g}\mathrm{O}}\times x_{\mathrm{M}\mathrm{g}\mathrm{O}}+{\mathrm{\Lambda }}_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}\times 3x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}+{\mathrm{\Lambda }}_{\mathrm{C}{\mathrm{e}}_2{\mathrm{O}}_3}\times 3x_{\mathrm{C}{\mathrm{e}}_2{\mathrm{O}}_3}+{\mathrm{\Lambda }}_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2}\times 2x_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2}]}{[(x_{\mathrm{C}\mathrm{a}\mathrm{O}}-x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3})+x_{\mathrm{M}\mathrm{g}\mathrm{O}}+3x_{\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}+2x_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2}+3x_{\mathrm{C}{\mathrm{e}}_2{\mathrm{O}}_3}]}\end{array}$$(11)

where the x_i is the mole fraction (the area fraction is regarded as the mole fraction for calculation) of the tetrahedral unit in the Raman results, i is the number of bridge oxygen, x_CaO is the mole fraction of CaO in the slag. The theoretical NBO/T ratios were calculated based on the composition listed in Table 1 and experimental NBO/T ratios were calculated based on the deconvoluted Raman spectra. The results are shown in Fig. 7. As the increase of the Ce₂O₃ content, the NBO/T and Λ^corr values increase, which indicated the depolymerization of molten slag.

Fig. 7.

NBO/T and Λ^corr with various Ce₂O₃ contents. (Online version in color.)

3.3. Effect of Ce₂O₃ on the Melting Properties

Figure 8 illustrates the variations for softening temperature, hemispherical temperature (melting temperature), and fluidity temperature of the slag systems with different Ce₂O₃ content. The hemispherical temperatures were 1320°C, 1305°C, 1303°C, and 1301°C when the Ce₂O₃ contents were 5 wt%, 10 wt%, 15 wt%, and 20 wt%, respectively. The melting temperature was relatively stabilized when Ce₂O₃ content was above 10 wt%. The softening temperature and fluidity temperature first decreased and then remained stable with the increase of Ce₂O₃ content. The same results have been reported in the CaO-Al₂O₃-10wt%SiO₂-Ce₂O₃ slag.¹¹⁾ The melting temperature of slag is associated with the substances formed in the glass transition period during the heating process.⁴¹⁾ According to the CaO–Al₂O₃–Ce₂O₃ phase diagram,^12,40) when the temperature decreased from 1600°C to 1500°C, the liquidus line is almost unchanged in the direction of increasing Ce₂O₃ content, whereas the liquidus line is getting closer to w(CaO)/w(Al₂O₃)=1. This indicated that the low-melting temperature compounds would form with the increasing Ce₂O₃ content when the Ce₂O₃ content was less than 30 wt%. Therefore, the melting temperature decreased with the increase of Ce₂O₃ content.

Fig. 8.

Melting properties of the slag with different Ce₂O₃ addition. (Online version in color.)

3.4. Effect of Ce₂O₃ on the Viscosity

Figure 9 shows the original viscosity-temperature curves of the CaO–Al₂O₃–Ce₂O₃–MgO–SiO₂ slags with different Ce₂O₃ content. The critical temperature at which the viscosity changes sharply is defined as the breaking temperature.⁴²⁾ It can be determined by the relationship between lnη (natural logarithm of the viscosity) and temperature (10000/T), as is depicted in Fig. 10. Viscosity measured in the continuous cooling process with a slow rate of 3°C·min⁻¹ could provide sufficient data to analyze the effect of composition on viscosity and solidification behavior. The viscosity of the fully liquid slag was measured at thermal equilibrium. Figure 11 shows the breaking temperature and the viscosity of the fully liquid slag. As can be seen from Fig. 11(a), the viscosity decreased as the addition of Ce₂O₃ increased. Conversely, the breaking temperature increased with Ce₂O₃ content (cf. Fig. 11(b)). The breaking temperature is strongly dependent on the viscosity.⁴¹⁾ With the increase of Ce₂O₃ content, the alkaline of the slag increased. The stronger alkaline the slag is, the easier it is to precipitate crystals during the cooling process, which leads to an increase in breaking temperature. On the other hand, the increase of Ce₂O₃ content would supply more O²⁻ ions to depolymerize the network structure, which would reduce the migration resistance of ions and atom groups, and then the growth process of the crystalline phase was promoted. Therefore, the breaking temperature increased with the increasing Ce₂O₃ content.

Fig. 9.

Viscosity of the slag with different Ce₂O₃ addition. (Online version in color.)

Fig. 10.

The relationship between lnη and 10000/T. (Online version in color.)

Fig. 11.

Breaking temperature and viscosity of the fully liquid slag with different Ce₂O₃ content. (a) Viscosity, (b) Breaking temperature. (Online version in color.)

The superheat of the slag increased due to the reduction of the hemispherical temperature, leading to the decrease of the viscosity at a certain temperature. The other reason was the DOP of the melt decreased with the increasing Ce₂O₃ content. Under present slag systems, the higher breaking temperature meant the slag properties would deteriorate after absorbing Ce₂O₃ inclusion. To ensure the slag has good stability in its viscosity characteristics after absorbing a certain amount of Ce₂O₃, a reasonable content of Ce₂O₃ should not excess 10 wt% in the current slag. Besides, the viscosity and melting point of the designed slags were satisfied with the requirement of tundish flux (viscosity at 1500°C: 0.2–0.5 Pa·s, melting point temperature: above 1300°C^3,43,44)).

Due to the increase of thermal motion, the depolymerization of the melt increased with higher temperatures.⁴⁵⁾ The fully liquid slag which was considered as Newtonian fluid and the viscosity dependence of the temperature can be described by the Arrhenius type equation, which is shown in Eq. (12). For Newtonian fluids, the apparent activation should be a constant value. The apparent activation energy of flux was calculated via the equation.   

$$ln\left(\frac{\eta }{T}\right)=\frac{E_{\eta }}{RT}+ln\mathrm{A}$$(12)

where η is the viscosity (Pa·s), A is a constant, E_η is the apparent activation energy (kJ·mol⁻¹), R is the universal gas constant, T is the absolute temperature (K).

The ln(η/T) versus 1/T for all the slags are plotted in Fig. 12(a). The apparent activation energy for all the slags was calculated via Eq. (12) as the slope of the fitting line, and the activation energy of all the slags are illustrated in Fig. 12(b). It is well known that the melt with higher viscosity would overcome higher resistance to shear, which results in higher apparent activation energy.⁴⁶⁾ Recent studies showed that the activation energy was related to the polymeric unit (T–O–T, T=Si, Al, etc.).⁴⁷⁾ With increasing Ce₂O₃ addition from 5 wt% to 20 wt%, the apparent activation energy decreased from 168.44 ± 6.11 kJ·mol⁻¹ to 79.18 ± 2.59 kJ·mol⁻¹, which indicated that the liquid shearing resistance of the slag reduced by the addition of Ce₂O₃ content.

Fig. 12.

Plots of ln(η/T) versus 1/T and activation energy with the various Ce₂O₃ content. (Online version in color.)

4. Conclusions

The effect of Ce₂O₃ content on the structure and properties of the CaO–Al₂O₃–Ce₂O₃–MgO–SiO₂ system has been investigated in the current study, and the main conclusions were obtained as follows:

(1) With Ce₂O₃ addition rising from 5 wt% to 20 wt%, the polymerized Q_Al⁴ and Q_Al³ units were modified to Q_Al² units in aluminate and the depolymerization of SiO₄-tetrahedrons from Q_Si¹ to Q_Si⁰ units. So, the DOP of the melts decreased gradually. When the Ce₂O₃ addition increased from 15 wt% to 20 wt%, more O²⁻ ions reacted with AlO₄-tetrahedral units to form AlO₆-octahedral units. Meanwhile, the increase of NBO/T and Λ^corr indicated the decrease of DOP in the melt.

(2) When the Ce₂O₃ addition increased from 5 wt% to 20 wt%, the melting temperature decreased from 1320°C to 1301°C. In addition, the softening temperature and the fluidity temperature also decreased 20°C and 24°C, respectively. The viscosity of fully liquid slag decreased as the Ce₂O₃ addition increased from 5 wt% to 20 wt%, which was attributed to the decreased DOP and the increased superheat. A reasonable content for Ce₂O₃ should not be excess 10 wt% in the current slag.

Acknowledgments

The authors gratefully acknowledge supports by the National Natural Science Foundation of China [No. U1908224].

References

• 1)   X.  Xing,  Z.  Han,  H.  Wang and  P.  Lu: J. Rare Earths, 33 (2015), 1122. https://doi.org/10.1016/S1002-0721(14)60535-4
• 2)   M.  Yu and  R.  Sandstrom: Scand. J. Metall., 17 (1988), 156.
• 3)   F.  Li,  B.  Wang,  D.  Zhan,  Z.  Jiang,  S.  He and  Q.  Wang: Adv. Mater. Res., 284–286 (2011), 1284. https://doi.org/10.4028/www.scientific.net/AMR.284-286.1284
• 4)   T. S.  Kim and  J. H.  Park: ISIJ Int., 54 (2014), 2031. https://doi.org/10.2355/isijinternational.54.2031
• 5)   W.  Wang,  H.  Shao,  L.  Zhou,  H.  Luo and  H.  Wu: Ceram. Int., 46 (2020), 26880. https://doi.org/10.1016/j.ceramint.2020.07.164
• 6)   E.  Gao,  W.  Wang and  L.  Zhang: J. Non-Cryst. Solids, 473 (2017), 79. https://doi.org/10.1016/j.jnoncrysol.2017.07.029
• 7)   H.  Shao,  E.  Gao,  W.  Wang and  L.  Zhang: J. Am. Ceram. Soc., 102 (2019), 4440. https://doi.org/10.1111/jace.16322
• 8)   J.  Qi,  C.  Liu,  C.  Zhang and  M.  Jiang: Metall. Mater. Trans. B, 48 (2017), 11. https://doi.org/10.1007/s11663-016-0850-3
• 9)   Z.  Cai,  B.  Song,  Z.  Yang and  L.  Li: ISIJ Int., 59 (2019), 1242. https://doi.org/10.2355/isijinternational.ISIJINT-2018-760
• 10)   W.  Guo,  Z.  Wang,  Z.  Zhao, Z. An and W. Wan: J. Non-Cryst. Solids, 540 (2020), 120085. https://doi.org/10.1016/j.jnoncrysol.2020.120085
• 11)   C.  Wu,  G.  Cheng and  H.  Long: High Temp. Mater. Process., 33 (2014), 77. https://doi.org/10.1515/htmp-2013-0025
• 12)   S.  Ueda,  K.  Morita and  N.  Sano: ISIJ Int., 38 (1998), 1292. https://doi.org/10.2355/isijinternational.38.1292
• 13)   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
• 14)   B. O.  Mysen,  D.  Virgo,  W. J.  Harrison and  C. M.  Scarfe: Am. Mineral., 65 (1980), 900. https://doi.org/10.1007/BF01081857
• 15)   B. O.  Mysen and  D.  Virgo: Am. Mineral., 65 (1980), 885.
• 16)   C.  Feng,  L.  Gao,  J.  Tang,  Z.  Liu and  M.  Chu: Trans. Nonferrous Met. Soc. China, 30 (2020), 800. https://doi.org/10.1016/S1003-6326(20)65255-4
• 17)   S.  Zhang,  X.  Zhang,  W.  Liu,  X.  Lv,  C.  Bai and  L.  Wang: J. Non-Cryst. Solids, 402 (2014), 214. https://doi.org/10.1016/j.jnoncrysol.2014.06.006
• 18)   X.  Yan,  W.  Pan,  X.  Wang,  X.  Zhang,  S.  He and  Q.  Wang: Metall. Mater. Trans. B, 52 (2021), 2526. https://doi.org/10.1007/s11663-021-02200-y
• 19)   S.  Gu,  G.  Wen,  J.  Guo,  Z.  Wang,  P.  Tang and  Z.  Hou: J. Non-Cryst. Solids, 547 (2020), 120312. https://doi.org/10.1016/j.jnoncrysol.2020.120312
• 20)   B. O.  Mysen and  P.  Richet: Silicate Glasses and Melts, Properties and Structure, Elsevier, Amsterdam, (2005), 98.
• 21)   B. N.  Roy and  A.  Navrotsky: J. Am. Ceram. Soc., 67 (1984), 606. https://doi.org/10.1111/j.1151-2916.1984.tb19603.x
• 22)   S.  Sukenaga,  S.  Haruki,  Y.  Nomoto,  N.  Saito and  K.  Nakashima: ISIJ Int., 51 (2011), 1285. https://doi.org/10.2355/isijinternational.51.1285
• 23)   Z.  Wang and  I.  Sohn: J. Am. Ceram. Soc., 101 (2018), 4285. https://doi.org/10.1111/jace.15559
• 24)   P.  McMillan,  B.  Piriou and  A.  Navrotsky: Geochim. Cosmochim. Acta, 46 (1982), 2021.
• 25)   W.  Yan,  G.  Zhang and  J.  Li: Ceram. Int., 46 (2020), 14078. https://doi.org/10.1016/j.ceramint.2020.02.208
• 26)   Y.  Zhen,  G.  Zhang,  X.  Tang and  K.  Chou: Metall. Mater. Trans. B, 45 (2014), 123. https://doi.org/10.1007/s11663-013-0014-7
• 27)   L.  Zhou,  H.  Li,  W.  Wang,  D.  Xiao,  L.  Zhang and  J.  Yu: Metall. Mater. Trans. B, 49 (2018), 2232. https://doi.org/10.1007/s11663-018-1327-3
• 28)   J.  Yang,  Q.  Wang,  J.  Zhang,  O.  Ostrovski,  C.  Zhang and  D.  Cai: Metall. Mater. Trans. B, 50 (2019), 2794. https://doi.org/10.1007/s11663-019-01711-z
• 29)   T. S.  Kim and  J. H.  Park: J. Non-Cryst. Solids, 542 (2020), 120089. https://doi.org/10.1016/j.jnoncrysol.2020.120089
• 30)   D. R.  Neuville,  L.  Cormier and  D.  Massiot: Chem. Geol., 229 (2006), 173. https://doi.org/10.1016/j.chemgeo.2006.01.019
• 31)   P.  McMillan: Am. Mineral., 69 (1984), 622. https://doi.org/10.1007/BF01082965
• 32)   P.  McMillan: Am. Mineral., 69 (1984), 645. https://doi.org/10.1016/0040-1951(85)90292-6
• 33)   T.  Maehara,  T.  Yano,  S.  Shibata and  M.  Yamane: Philos. Mag., 84 (2004), 3085. https://doi.org/10.1080/4786430410001725890
• 34)   T.  Talapaneni,  N.  Yedla,  S.  Pal and  S.  Sarkar: Metall. Mater. Trans. B, 48 (2017), 1450. https://doi.org/10.1007/s11663-017-0963-3
• 35)   J. H.  Park,  D. J.  Min and  H. S.  Song: ISIJ Int., 42 (2002), 38. https://doi.org/10.2355/isijinternational.42.38
• 36)   S. L.  Lin and  C. S.  Hwang: J. Non-Cryst. Solids, 202 (1996), 61. https://doi.org/10.1016/0022-3093(96)00138-X
• 37)   B. O.  Mysen: Earth-Sci. Rev., 27 (1990), 281. https://doi.org/10.1016/0012-8252(90)90055-Z
• 38)   S.  Seetharaman: Treatise on Process Metallurgy, Vol. 1: Process Fundamentals, Elsevier, Oxford, UK, (2014), 156.
• 39)   K. C.  Mills: ISIJ Int., 33 (1993), 148. https://doi.org/10.2355/isijinternational.33.148
• 40)   R.  Kitano,  M.  Ishii,  M.  Uo and  K.  Morita: ISIJ Int., 56 (2016), 1893. https://doi.org/10.2355/isijinternational.ISIJINT-2016-201
• 41)   L.  Zhang,  B.  Zhai,  W.  Wang and  I.  Sohn: J. Sustain. Metall., 7 (2021), 559. https://doi.org/10.1007/s40831-021-00358-y
• 42)   A.  Ilyushechkin,  S.  Hla,  D.  Roberts and  N.  Kinaev: J. Non-Cryst. Solids, 357 (2011), 893. https://doi.org/10.1016/j.jnoncrysol.2010.12.004
• 43)   J.  Gao: Master’s thesis, Chongqing University, (2008), https://d.wanfangdata.com.cn/thesis/D319158, (accessed 2021-10-20).
• 44)   Y.  Liu,  M.  Yang,  J.  Xue,  P.  Zhang and  X.  Zhou: A tundish flux and its application, China Patent, CN201911280453.3, (2020).
• 45)   D.  Yang,  F.  Zhang,  J.  Wang,  Z.  Yan,  G.  Pei,  G.  Qiu and  X.  Lv: J. Mater. Res. Technol., 9 (2020), 14673. https://doi.org/10.1016/j.jmrt.2020.10.061
• 46)   J.  Qi,  C.  Liu and  M.  Jiang: J. Non-Cryst. Solids, 475 (2017), 101. https://doi.org/10.1016/j.jnoncrysol.2017.09.014
• 47)   J. S.  Choi,  T. J.  Park and  D. J.  Min: J. Am. Ceram. Soc., 104 (2021), 140. https://doi.org/10.1111/jace.17432