Effect of Na₂O on Dissolution Rate of Alumina in CaO–Al₂O₃–MgO–SiO₂ Slag

Abstract

As the last step of inclusion removal in tundish, the dissolution of Al₂O₃ in tundish slag is very important and has received many interests by previous researchers. In the present work, the effect of Na₂O addition on dissolution of Al₂O₃ in CaO–Al₂O₃–MgO–SiO₂ slag at 1823 K was investigated by using rotating finger method. The interface between rod and molten slag was investigated using SEM-EDS technique. The dissolution of Al₂O₃ in CaO–Al₂O₃–MgO–SiO₂–Na₂O system increases with increase of rotating speed and immersing time. A linear relationship between logarithm of dissolution rate and logarithm of periphery velocity was found with a slope of 0.53, which indicates that rate determining step for dissolution of Al₂O₃ is the mass transfer of solute in boundary layer of molten slag. The dissolution rate of alumina in CaO–Al₂O₃–MgO–SiO₂–Na₂O slag increases with increase of Na₂O content, which could be attributed to the kinetic and thermodynamic factors. The decrease of viscosity of molten slag with increasing Na₂O content could lead to increased mass transfer in boundary layer of molten slag. On the other hand, the increase of Na₂O would lead to increase of thermodynamic driving force of Al₂O₃ dissolution. Some intermediate compounds such as MgAl₂O₄ and CaAl₄O₇ were found at interface between alumina rod and slag, indicating indirect dissolution of Al₂O₃ in slags.

1. Introduction

Tundish was originally designed to be reservoir and distributor for liquid steel before continuous casting.¹⁾ The cleanliness of liquid steel has impact on quality of steel and has received many attentions in recent year. The role of tundish in inclusion removal has been focused by many researchers. The flow of liquid steel was optimized to prolong the residence time for steel in tundish in order to promote the removal of inclusion.²⁾ Therefore, functions of tundish have been extended from original “reservoir and distributor” to control of steel temperature, chemistry and cleanliness.

Traditionally, tundish covering slag was designed to keep thermal insulation and protect reoxidation of steel. Many recent efforts have been made to improve capability of tundish covering slag to absorb non-metal inclusion.³⁾ As the last step for inclusion removal in tundish, the dissolution of inclusion, e.g. Al₂O₃, in slag is of vital importance to increase the cleanliness of steel.

The dissolution of alumina in slag has been investigated by rotating finger, Confocal scanning laser microscopy, etc.^{4,5,6,7,8,9,10,11,12,13,14)} The dissolution of Al₂O₃ rod in CaO–Al₂O₃–SiO₂ -based slag under forced convection has been studied by many researchers using rotating finger method.^4,5,6,7,8,9) It was found that diffusion in boundary layer was rate controlling and the dissolution rate increased with increasing rotating speed and temperature. Confocal scanning laser microscopy (CSLM) was employed to investigate the dissolution of Al₂O₃ particle in molten slag under natural convection^11,12,13) and it was reported that Al₂O₃ dissolution process in CaO–Al₂O₃–SiO₂ and CaO–Al₂O₃–MgO–SiO₂ is controlled by boundary layer diffusion inside the slag phase. The Al₂O₃ dissolutions in many slags were found to be indirect and intermediate solid phases were found to form at the interface, which could bring strong influences on the overall dissolution.^7,8,9,12,13)

It was reported that addition of alkaline oxide could decrease the viscosity of slag.¹⁵⁾ Considering that diffusion in boundary layer could be enhanced by decrease of viscosity, alkaline oxides have a potential to increase the dissolution rate of inclusion in slag. However, Yu et al.¹⁶⁾ investigated effect of alkaline oxide in tundish slag on cleanliness of steel and found Na₂O and Li₂O has minimal effect on cleanliness of steel while K₂O could increase the cleanliness of steel. However, to the best knowledge of the present authors, there is no report on dissolution of Al₂O₃ in CaO–Al₂O₃–MgO–SiO₂–Na₂O system in literature. To elucidate the influence of Na₂O on inclusion absorption of tundish slag, the dissolution of Al₂O₃ rod in CaO–Al₂O₃–MgO–SiO₂–Na₂O slag at 1823 K was investigated by employing rotating finger method in this work. The dissolution mechanism was elucidated by investigation with varied rotating speed and Na₂O content in slag.

2. Experimental

2.1. Materials Preparation

Reagent grade powders of CaCO₃, SiO₂, Al₂O₃, MgO and Na₂CO₃ with purity being larger than 99% were used as raw materials. CaCO₃ was calcined in a muffle furnace for 10 hours at 1323 K to obtain CaO. SiO₂, Al₂O₃, MgO and Na₂CO₃ powders were also calcined to remove moisture absorbed. Chemical compositions of slags investigated are shown in Table 1. Alumina rods in cylinder shape with 8×10⁻³ m in diameter, 1 m in length and ~3.95 g/cm³ density were prepared by sintering purity Al₂O₃ (>99.5 mass%).

Table 1. Chemical compositions of investigated slags.

Sample No.	Composition (mass%)
	CaO	SiO2	Al2O3	MgO	Na2O
1	40	20	30	10	0
2	39.04	19.52	29.28	9.76	2.4
3	38.56	19.28	28.92	9.64	3.6

2.2. Experimental Procedure

Detailed description for experimental apparatus has been given in our previous publication.¹⁷⁾ Dissolution experiments were carried out in a molybdenum silicide furnace under argon atmosphere 200 g powder mixture was melted in a graphite crucible (ID: 55×10⁻³ m; H: 100×10⁻³ m) in a molybdenum silicide furnace under argon atmosphere with temperature controlled by a B type thermocouple right under the crucible. Al₂O₃ rod was immersed into the molten flux to prescribed length and rotated using a motor. Rotating speed was adjusted using a speed controller. For the sake of comparison, the dissolution experiments at static condition, i.e. natural convection, were also carried out. The Al₂O₃ rod was dipped into molten slag for a prescribed period in static experiments. After dissolution, Al₂O₃ rod was quickly removed from the furnace and cooled in air, and molten slag was quenched on an iron plate. Slag layer attached to the surface of rod was removed using diluted hydrochloric acid, and then diameter of rod was measured by a micrometer with the accuracy of 0.02 mm. Cross diameters at three levels with 1 cm interval in height were measured. Average value of three times results was adopted as the mean diameter of the cylinder after the experiment. Al₂O₃ rods attached with a thin slag layer were cut longitudinally using diamond saw. The cut sample was mounted in epoxy resin, and cross section of rods attached with slag was ground and polished. Then the cross sections were investigated by using Scanning Electronic Microscopy equipped with Energy Dispersed Spectral (SEM-EDS) analysis in order to examine the interface between molten fluxes and rod. SEM-EDS analysis was performed on FEI-MLA250 with Bruker energy dispersed spectral. The working voltage is 25 kV.

3. Results and Discussion

3.1. Effect of Rotation Speed on Dissolution Rate of Al₂O₃

Figure 1 showed decreases in radius of rod in slag with Na₂O%=2.4% as functions of dissolution time at different rotation speeds. Three rotation speeds of 100, 150 and 200 rpm were employed to elucidate the underlying mechanism for dissolution. It could be seen from the figure that decrease in radius of rod increases linearly with increase of dissolution time at the same rotation speed, and decrease in radius increases with increase of rotation speed at the same dissolution time. The decreases in radius of rod at static condition (with no rotation) were also shown in Fig. 1. It was found that the decrease of radius of rod in static condition is much lower than that in rotating condition, indicating that the dissolution was significantly enhanced by forced convection induced by rod rotation.

Fig. 1.

Decreases in radius of rod as functions of dissolution time at different rotation speeds.

The rate of dissolution could be calculated by fitting decrease in radius with respect to dissolution time. Considering the possible presence of the Taylor’s vortex, decrease in radius data in Fig. 1 was fitted with lines with intercepts. Data in Fig. 1 could be fitted well using line with intercepts and the correlation factors R² were shown in Fig. 1. It could be seen from Fig. 1 that dissolution rate as shown by slope of fitting line increases with increase of rotation speed. Assuming that mass transfer of Al₂O₃ in molten slag is controlling step for dissolution of Al₂O₃ in slag, a relationship between the dissolution rate and the periphery velocity could be established as follows:⁴⁾   

$$V=b\cdot U^s$$(1)

where b represents a constant and U represent the periphery velocity U (cm/s) which could be calculated as U=πdm/60, where d represents the mean diameter of cylinder in mm and m represents number of revolutions per min); Exponent s has been reported^4,5,6) in the range of 0.50 to 0.80. The relationship between lnV and lnU was shown in Fig. 2. It can be seen that the lnV is linearly proportional to lnU, which indicates that the dissolution of alumina into flux is controlled by mass transfer in the flux phase. The calculated value of s as the slopes of the lines is about 0.53 and falls within the range of 0.50 to 0.80 reported before. Thus, it is presumed that the rate-controlling step for the dissolution of alumina is mass transfer within the boundary layer in the present slag. Sridhar et al.¹¹⁾ investigate the dissolution of Al₂O₃ particle in 19.5%Al₂O₃-33.4%CaO-7.3%MgO-39.5%SiO₂ (in mass percentage) slag by employing CSLM technique under natural convection. They found that the dissolution rate from 50–80 mm down to 10 mm could be described by considering boundary layer diffusion as dissolution. The dissolution mechanism proposed in the present work is consistent with theirs in spite of the fact that the present dissolution took place under forced convection.

Fig. 2.

Relationship between lnV and lnU.

3.2. Effect of Na₂O Addition on Dissolution of Al₂O₃

Figure 3 showed that decreases in radius as functions of Na₂O content in slag with Al₂O₃ dissolving for 1200 s at 100 rpm. It could be seen that decrease in radius increases with increase of Na₂O content in slags, which indicates that addition of Na₂O would promote the dissolution of Al₂O₃ in the present slag. The present results indicate that the addition of Na₂O would be beneficial to dissolution of alumina in CaO–Al₂O₃–MgO–SiO₂ system, and thereby help to removal of inclusion in tundish.

Fig. 3.

Decrease in radius of rod for dissolution in slags with different Na₂O content.

It has been speculated that the dissolution of alumina in the present slag is determined by the mass transfer in the molten slag. Mass transfer flux during the dissolution can be represented by the following equation:   

$$J=k({\mathrm{C}}_{\mathrm{s}\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3}-{\mathrm{C}}_{n\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3})$$(2)

where k is the mass transfer coefficient (cm/s), C_bAl2O3 and C_sAl2O3 are the concentration of Al₂O₃ in the bulk slag and at the interface between solid and liquid respectively, C_sAl2O3− C_bAl2O3 value represents the thermodynamic driving force for dissolution.

Mass balance of Al₂O₃ gives,   

$${\rho }_{A{l}_2{O}_3}\cdot A(-dr/dt)=\mathrm{A}\cdot {\mathrm{J}}_{A{l}_2{O}_3}$$(3)

where A is the surface area between solid and liquid (cm²) and ρ_Al2O3 is the mass density of solid (g/cm³). Combining Eqs. (3) and (4), one can get   

$$V=-dr/dt=\mathrm{k}\left(\frac{{\rho }_{\mathrm{f}\mathrm{l}\mathrm{u}\mathrm{x}}}{100{\rho }_{A{l}_2{O}_3}}\right)\mathrm{\Delta }\%A{l}_2{O}_3$$(4)

where Δ(%Al₂O₃)=(%Al₂O₃)_s−(%Al₂O₃)_b represents the chemical driving force for dissolution, (%Al₂O₃)_b and (%Al₂O₃)_s are mass percentage of Al₂O₃ in the bulk slag and at the solid-liquid interface respectively. (%Al₂O₃)_s could be obtained in term of liquidus boundary in phase diagram.

From Eq. (4), the dissolution rate was mainly determined by thermodynamic driving force and mass transfer coefficient. Dissolution rates of Al₂O₃ are accordingly enhanced. (%Al₂O₃)_s value represents the saturated Al₂O₃ content in slag which could be obtained from phase diagram. Due to its low melting point (1132°C), Na₂O has the ability to decrease the melting temperature of slag. It could be shown in phase diagram of CaO–Al₂O₃–SiO₂–Na₂O¹⁸⁾ that addition of Na₂O would decrease liquidus temperature of slags and enlarge the liquidus area in phase diagram, thereby increasing saturated Al₂O₃ content and thermodynamic driving force for dissolution increases. Accordingly, the dissolution rates of Al₂O₃ in slag increases.

On the other hand, the mass transfer coefficient k in the slag phase is presented by levich equation:¹⁹⁾   

$$k=0.62D^{2/3}{\nu }^{-1/6}{\omega }^{1/2}$$(5)

where D is the diffusion coefficient of components in slag, ν is the kinematic viscosity of liquid and ω is the angular velocity of rod. As can be seen in Eq. (5), k is proportional to D^2/3. The experimental data for diffusion coefficient of components in slag is very few. However, diffusion coefficient could be related to viscosity of slag by Stokes-Einstein equation:   

$$D=\frac{k_{B}T}{6\pi r\eta }$$(6)

where k_B is Boltzmann’s constant, r is radius of atoms or ions, η is viscosity of slag. Therefore, the effect of Na₂O on viscosity of slag could be employed to interpret the variation of dissolution rates.

Na₂O was usually regarded as a network modifying oxide in silicate network.²⁰⁾ It can broke complex structural units into simple units, and thereby decrease the viscosity of slags. The effect of Na₂O on viscosity of slags was investigated by many authors. Kim et al.¹⁵⁾ measured viscosity of CaO–SiO₂–MgO–Al₂O₃–Na₂O slag with Na₂O%=0−10%, and found that the slag viscosity at 1773 K decreased with increasing Na₂O contents. One of the present authors presented a model to calculate viscosity of multi-component slag.^21,22) The model has been applied to calculate viscosity of many multi-component slag e.g. CaO–SiO₂–MgO–Al₂O₃, CaO–SiO₂–Al₂O₃–Na₂O systems and a good agreement between calculated and measured viscosity values has been achieved with mean deviation less than 25%. Therefore, the model was employed to calculate viscosity of the present slags. Calculation results for viscosity of the present slags were shown in Fig. 4. It could be seen the figure that viscosity of the present slag is lowered by increase of Na₂O content, which is in consistent with viscosity measurements by Kim et al. According to Stokes-Einstein equation, diffusion coefficient of components in slags is inversely proportional to viscosity of slags. The increase of Na₂O content in slag decreases the viscosity of slags, thereby increase the diffusion coefficient of components in slags. According to Eq., the increasing of diffusion coefficient of components in slag would lead to acceleration of dissolution of Al₂O₃ in slag.

Fig. 4.

Calculated viscosity and kinematic viscosity values of CaO–Al₂O₃–SiO₂–MgO–Na₂O slag as a function of Na₂O content.

It could be also seen from Eq. (5) that k is inversely proportional to ν^1/6. kinematic viscosity could be calculated from viscosity and density of slag as η/ρ. Densities of slags investigated in the present work was calculated by using the equation proposed by Mills et al.²³⁾ The calculated kinematic viscosity values are also shown in Fig. 4. It could be seen from the figure that kinematic viscosity of slag decreases with increase of Na₂O content. Therefore, according to Eq. (6), decrease of kinematic viscosity of slag with increase of Na₂O content in slag would also help to increase of mass transfer coefficient, and thereby increase of Al₂O₃ dissolution rate in slag.

3.3. Interface between Al₂O₃ and Slag

Figure 5 showed micrographs of Al₂O₃-slag interface for slags with different Na₂O content. The alumina rod was rotated in molten slag at the speed of 100 rpm for 1200 s. The Dark grey part in micrographs is Al₂O₃ rod, and the light grey part is slag. It could be seen that there are many dark grey crystals in slag phases. Both equiaxed and columnar crystals could be found in all slags. EDS showed the composition of equiaxed crystals and columnar crystals are MgAl₂O₄ and CaAl₄O₇ respectively. The composition analyses along lines across Al₂O₃-slag interface were also performed for all samples and shown in Fig. 5. Composition gradients could be found in all samples. There are also many composition fluctuations in slag phase due to precipitation of many crystals.

Fig. 5.

SEM micrographs and line composition analyses for interfaces between Al₂O₃ rod and slag with different Na₂O content (A: CaAl₄O₇; B: MgAl₂O₄).

The melting points of MgAl₂O₄ and CaAl₄O₇ are higher than temperature in the present experiments. Therefore, MgAl₂O₄ and CaAl₄O₇ should be in the solid state at experimental temperature and formed by dissolution of Al₂O₃. The formation of many intermediate phases indicates that the dissolution of Al₂O₃ in the present slag is not in a direct manner but proceeds indirectly. The indirect dissolution of Al₂O₃ includes two steps as follows. Firstly, the dissolution of Al₂O₃ in slag leads to the formation of intermediate solid phases at or near Al₂O₃-slag interface. Then, intermediate phases gradually dissolve into slag. The formation of intermediate phases as MgAl₂O₄ and CaAl₄O₇ could be interpreted with the aid of phase diagrams of CaO–Al₂O₃–SiO₂–(MgO).¹⁸⁾ Figure 6 showed the phase diagram of CaO-Al₂O₃-SiO₂-10%MgO. The composition of the present slag was marked on the phase diagram. As Al₂O₃ gradually diffused into slag, the composition would change along with the line slag-(Al₂O₃) in Fig. 6. It could be seen from Fig. 6 that the liquidus temperature of slag increases with increase of Al₂O₃ content. The line slag-(Al₂O₃) is apparently located in the primary crystal field of MgAl₂O₄. Therefore, the enriched Al₂O₃ content in slag would lead to the precipitation of MgAl₂O₄. It could be seen from Fig. 5 that MgAl₂O₄ precipitated in all samples. Sandhage et al.^7,8,9) studied the dissolution of Al₂O₃ in CaO–Al₂O₃–MgO–SiO₂ slags, and observed the formation of MgAl₂O₄ during dissolution. Valdez et al.¹²⁾ Investigate the dissolution of Al₂O₃ in CaO–Al₂O₃–MgO–SiO₂ slags using CSLM technique. They also found the formation of MgAl₂O₄ as an intermediate solid phase. The precipitation of MgAl₂O₄ would gradually lower the concentration of MgO in slag. Therefore, phase diagram of CaO–Al₂O₃–SiO₂ system¹⁸⁾ is more appropriate to interpret the precipitation of secondary crystal. Figure 7 showed the phase diagram of CaO–Al₂O₃–SiO₂ system. The composition the present slag (The MgO and Na₂O contents were neglected) was marked on the diagram. The dissolution of Al₂O₃ in slag would lead to change of composition which could be represented by line slag- Al₂O₃ in Fig. 7. As shown in the figure, the liquidus temperature of slag increases as Al₂O₃ concentration increases. The gradual increase of Al₂O₃ content in slag would cause that composition of slag successively pass through the primary crystal fields of Ca₂Al₂SiO₇, CaAl₄O₇, CaAl₁₂O₁₉ and Al₂O₃. Therefore, CaAl₄O₇ and CaAl₁₂O₁₉ could precipitate from liquid and become possible intermediate solid phases during dissolution of Al₂O₃. Due to possible kinetic factors, only formation of CaAl₄O₇ could be observed in the present work. Park et al.¹³⁾ investigated the dissolution of Al₂O₃ in CaO–Al₂O₃–SiO₂ slag using CSLM technique, and found reaction products such as Ca₂Al₂SiO₇, CaAl₄O₇ and CaAl₁₂O₁₉ formed on the surface of Al₂O₃ particle. The present author reported¹⁷⁾ the kinetics of Al₂O₃ in CaO–Al₂O₃–CaF₂ flux and found the existence of CaAl₄O₇ as an intermediate solid phase during Al₂O₃ dissolution. These previous reports are in line with the present work with regard to intermediate solid phase during dissolution.It could be also found in micrographs that CaAl₄O₇ particles are closer to Al₂O₃ rod than MgAl₂O₄. Due to its small size and equiaxed morphology, MgAl₂O₄ is easier to be thrown out by the centrifugal force induced by rotation. Therefore, the MgAl₂O₄ layer is located at a distance from rod.

Fig. 6.

Phase diagram of CaO-Al₂O₃-SiO₂-10mass%MgO system.

Fig. 7.

Phase diagram of CaO–Al₂O₃–SiO₂ system.

It would be interesting to investigate the interface between rod and slag under static condition. Figure 8 showed the micrograph and the line composition analysis of interface between rod and slag with Na₂O%=2.4% under static condition. It could be seen from the figure that there are many equiaxed and columnar particles existing in slag phase. The equiaxed and columnar particles are MgAl₂O₄ and CaAl₄O₇ respectively. This indicates that the intermediate phases are not changed with introduce of rotation for slag with Na₂O%=2.4%. However, comparing Fig. 8 with Fig. 5(b), it could be found that the sizes and numbers of intermediate phases in slags under static condition are larger than that in slags under forced convection. It has been shown that forced convection by rotating rod accelerates the dissolution of Al₂O₃. Accordingly, force convection should be beneficial to overall dissolution by promoting the dissolution of intermediate solid phases (the second stage). The composition analysis was carried out along a line across slag-Al₂O₃ interface. Compared with samples in forced convection, a mild composition gradient was found in samples under static condition, which is due to the weaker mass transfer under static condition.

Fig. 8.

SEM micrograph and line composition analysis for interface between Al₂O₃ rod and slag with Na₂O=2.4% under static condition.

According to above discussion, the dissolution mechanism for Al₂O₃ could be summarized as bellows. At the first stage, Al₂O₃ was gradually transferred into slag, and the supersaturation of Al₂O₃ would lead to precipitation of the first intermediate solid phase MgAl₂O₄. The second intermediate solid phase CaAl₄O₇ could precipitate from slag with further supersaturation of Al₂O₃. At the second stage, intermediate solid phases gradually dissolved into slag. The steel flow induced by rod rotation would be beneficial to the dissolution of these intermediate solid phases.

It could be inferred from the present work that the dissolution of Al₂O₃ in basic tundish slag could be enhanced with addition of Na₂O. However, the addition of Na₂O leads to more intermediate solid particles which could be obstacles for complete dissolution of Al₂O₃. Yu et al.¹⁶⁾ reported minimal effect of Na₂O in tundish slag on cleanliness of steel. Since removal of inclusion include also separation of inclusion at slag-metal interface other than dissolution of inclusion into slag, the elucidation of effect of Na₂O addition on the cleanliness of steel still needs further studies in the future.

4. Conclusions

Effect of Na₂O addition on dissolution of Al₂O₃ in CaO–Al₂O₃–MgO–SiO₂ slag was investigated by using rotating finger method. The interface between rod and molten slag was investigated using SEM-EDS. Following conclusions have been reached.

(1) The dissolution of Al₂O₃ in CaO–Al₂O₃–MgO–SiO₂–Na₂O system increases with increase of rotating speed and immersing time.

(2) A linear relationship between logarithm of dissolution rate and logarithm of periphery velocity was found with a slope of 0.53, which indicate that rate determining step for dissolution of Al₂O₃ is the mass transfer of solute in boundary layer of molten slag.

(3) The dissolution rate of alumina in CaO–Al₂O₃–MgO–SiO₂–Na₂O system increases with increase of Na₂O content at the same rotating speed. This is due to increase of chemical driving force for dissolution and decrease of slag viscosity.

(4) The dissolution of alumina in CaO–Al₂O₃–MgO–SiO₂–Na₂O system was found to be in a way of indirect dissolution. Some intermediate solid phases such as MgAl₂O₄ and CaAl₂O₄ were found at interface between alumina rod and slag.

Acknowledgement

Financial supports from Natural Science Foundation of China (NSFC contract no. 51174018) and Fundamental Research Funds for the Central Universities (FRF-TP-14-108A2) are gratefully acknowledged.

References

• 1)   D.  Mazumdar and  R. I.  Guthrie: ISIJ Int., 39 (1999), 524.
• 2)   A.  Cwudziński: Steel Res. Int., 85 (2014), 902.
• 3)   L.  Holappa,  M.  Kekkonen,  S.  Louhenkilpi,  R.  Hagemann,  C.  Schröder and  P.  Scheller: Steel Res. Int., 84 (2013), 638.
• 4)   S.  Taira,  K.  Nakashima and  K.  Mori: ISIJ Int., 33 (1993), 116.
• 5)   J.  Choi,  H.  Lee and  J.  Kim: ISIJ Int., 42 (2002), 852.
• 6)   A. H.  Bui,  H. M.  Ha,  Y. B.  Kang,  I. S.  Chung and  H. G.  Lee: Met. Mater. Int., 11 (2005), 183.
• 7)   K. H.  Sandhage and  G. J.  Yurek: J. Am. Ceram. Soc., 71 (1988), 478.
• 8)   K. H.  Sandhage and  G. J.  Yurek: J. Am. Ceram. Soc.,73 (1990), 3633.
• 9)   K. H.  Sandhage and  G. J.  Yurek: J. Am. Ceram. Soc.,73 (1990), 3643.
• 10)   W. D.  Cho and  P.  Fan: ISIJ Int., 44 (2004), 229.
• 11)   S.  Sridhar and  A. W.  Cramb: Metall. Mater. Trans. B, 31 (2000), 406.
• 12)   M. Valdez,  K.  Prapakorn,  A. W.  Cramb and  S.  Sridhar: Ironmaking Steelmaking, 29 (2002), 47.
• 13)   J-H.  Park,  I-H.  Jung and  H-G.  Lee: ISIJ Int., 46 (2006), 1626.
• 14)   X.  Yu,  R. J.  Pomfret and  K. S.  Coley: Metall. Mater. Trans. B, 28 (1997), 275.
• 15)   H.  Kim,  W. H.  Kim,  J. H.  Park and  D. J.  Min: Steel Res. Int., 81 (2010), 17.
• 16)   J. Y.  Yu,  Y.  Kang and  I.  Sohn: Metall. Mater. Trans. B, 45 (2014), 113.
• 17)   J. L.  Li,  Q. F.  Shu,  Y.  Liu and  K. C.  Chou: Ironmaking Steelmaking, 41 (2014), 732.
• 18)   M.  Kowalski,  P. J.  Spencer and  D.  Neutschutz: Slag Atlas, ed. by VDEh, Verlag Stahleisen GmbH, Düsseldorf, Germany, (1995), 105.
• 19)   V. G.  Levich: Physicochemical Hydrodynamic, Prentice-Hall, New York, (1962), 69.
• 20)   F. D.  Richardson: Physical Chemistry of Melts in Metallurgy, Vol. I, Academic Press, London, (1974), 81.
• 21)   Q. F.  Shu,  Z.  Wang and  K. C.  Chou: Steel Res. Int., 82 (2011), 779.
• 22)   Q. F.  Shu: Steel Res. Int., 80 (2009), 107.
• 23)   K. C.  Mills and  B. J.  Keene: Int. Mater. Rev., 32 (1987), 1.