Generation of Heterogeneous Nucleus in Carbon Steel during Solidification by Magnesium Vapor Injection

Abstract

The behavior for injection of magnesium vapor into molten steel and the potential of magnesium vapor for miniaturization of TiN size were evaluated by the experiment in tammann furnace and the experiment in vacuum furnace. From the results of these experiments, the partial pressure of magnesium reacting with molten steel could be increased, and the both high concentration in molten steel and high additive yield of magnesium into molten steel could be done together. It was confirmed that spinel MgO·Al₂O₃ was formed as a heterogeneous nucleus of TiN. The dissolved magnesium concentration governed the heterogeneous nucleus for TiN. The injection of magnesium vapor into molten steel could control the generation of heterogeneous nucleus during solidification.

1. Introduction

High Performance products of carbon steel are demanded recently. Some properties of final product are occasionally affected by the quality of continuous cast slabs. A novel technology to control solidification microstructure has to be developed to improve the internal quality for extremely thick plate products which are difficult to meet the standard of reduction ratio.^1,2,3) With this technology, it is expected to improve the surface quality for direct rolling products which have coarsen γ grains^1,2,3) and to disperse the fine inclusions which are precipitated with microsegregation during solidification. To solve these development subjects, it is necessary to refine the microstructure, reduce microsegregation during solidification and refine the inclusion. Therefore, the study to control both the solidification microstructure and the inclusion in slabs becomes urgent.

The size of the solidification microstructure depends on the growth rate and the temperature gradient, and the solidification microstructure can be made refinement by increasing these values. Besides this, authors have shown that the solidification microstructure became fine by decreasing the solid-liquid interface energy.⁴⁾ This is an idea obtained according to the dendrite growth model, and making the solidification microstructure refinement rise the industrial level.

To refine the inclusion, it is effective to use the method of the heterogeneous nucleation phenomena.^5,6,7) The material which becomes a nucleus is dispersed finely in molten steel and heterogeneous nucleation is generated on this nucleus. It is important to generate the heterogeneous nucleus of which size is micron or nano size, and to disperse that nucleus in high density, without using the expensive rare metals.

In this study, the possibility of refinement of TiN by using heterogeneous nucleus was examined. It is understood that the magnesium oxide is suitable as a heterogeneous nucleus of TiN.^{8,9,10,11,12)} The magnesium oxide is formed as the magnesium added in molten steel reacts with oxygen and aluminum. It is thought that the magnesium oxide becomes small at this reaction if the size of magnesium particle can be refined. As it is clarified that the magnesium vapor is composed of fine particles,¹³⁾ the vaporization of magnesium metal is found to be useful for the refinement of magnesium oxide. By the way, though many studies^14,15,16,17) by which the magnesium vapor are injected into molten steel or molten iron have been done, these studies are limited to the batch process in the crucible or the ladle. However, the behavior of magnesium in the steel slabs injected into molten steel before casting have not been studied. The technique for both addition of cheap alloying element under high concentration and dispersion of fine nucleus is possible to become a revolutionary technique for steelmaking process and materials processing in the future.

Then, the behavior of magnesium vapor injection into molten steel was examined by using the small size tammann furnace and the high frequency induction furnace in the vacuum vessel.

2. Experimental Procedures

2.1. Experiment by Using the Tammann Furnace

To generate the magnesium vapor and to inject the vapor into molten steel safely, it is important to clarify the vaporization behavior of the magnesium metal and to understand the refining characteristics. Then, the experiment has been done by using the tammann furnace with magnesium lump of purity 99.95 mass%.

Figure 1 shows the schematic diagram for experimental apparatus. The steel sample of 0.06 mass%C-1.0 mass%Si-1.3 mass%Mn of 2.5 kg by using the electrolytic iron and the alloying metal was set in an MgO crucible and melted a tammann furnace under an argon gas atmosphere. The temperature of molten steel was controlled by a R type thermocouple and adjusted to 1873 K. The lance made of Mo–ZrO₂ of 1.3×10⁻² m in outside diameter and 0.9×10⁻² m in internal diameter was moved down from the above and immersed into molten steel for 3.0×10⁻² m in depth. An argon gas was supplied in the lance at quantity of 8.3×10⁻⁶ m⁻³s⁻¹ and at pressure of 1.0 MPa through a magnesium folder made of the stainless steel, and then a mixed gas of argon gas and magnesium vapor was injected into molten steel from a hole of 1.0×10⁻³ m in diameter at 1.0×10⁻² m from the tip of lance drilled at the cylindrical face of the lance. An argon gas was supplied in the lance when the tammann furnace was applied by current, and then the lance was maintained with an argon gas atmosphere and hole of lance was prevented clogging with molten steel. At the tip of magnesium holder of 6.0×10⁻³ m in outer diameter and 4.0×10⁻³ m in inner diameter, a pocket was made and magnesium lump of 2.0×10⁻⁴ kg was put in that. This amount of magnesium lump corresponded to 0.008 mas% when the yield was assumed to be 100%. The hole of 4.0×10⁻³ m in diameter for the sidewall of the magnesium folder was drilled, and a magnesium vapor was supplied in the lance. The position of magnesium holder could be changed in the lance and the temperature of magnesium lump could be controlled by using the temperature gradient in lance. The temperature in the holder was measured with the thermocouple set just above the magnesium lump. After moving up the lance, the tammann furnace was powered off immediately and the small ingot of 8.5×10⁻² m in diameter and 5.5×10⁻² m in height was solidified. A granular sample was obtained from the center of ingot and was made an analysis of the composition.

Fig. 1.

Schematic diagram of experimental setup for Mg injection into molten steel in the tammann furnace.

2.2. Experiment by Using the High Frequency Induction Furnace in the Vacuum Vessel

To maintain a magnesium concentration of steel slabs constant, it is necessary to inject a magnesium vapor into molten steel continuously. When the experiment by using the tammann furnace was done, a magnesium lump was added in molten steel. However, it is thought to be best to use the magnesium wire for the continuous feed in molten steel. Therefore, the experiments by using the high frequency induction furnace with molten steel of 200 kg were done to examine possibility of continuous feeding of magnesium wire.

Figure 2 shows the schematic diagram for experimental apparatus. The high frequency induction furnace was set in the vacuum vessel. The vessel was exhausted to the vacuum, and then was charged with argon gas at pressure of 9.9×10⁻² MPa lower an atmospheric pressure. The electrolytic iron and the alloying metal were set in an MgO crucible and melted by induction heating. The chemical composition of the sample is shown in Table 1. The temperature of molten steel adjusted to 1873 K±5 K. After the magnesium vapor was injected, the molten steel was poured the mold of the thickness 0.1×10⁻¹ m, 0.4×10⁻¹ m in width, and 0.5×10⁻¹ m in height.

Fig. 2.

Schematic diagram of experimental setup for Mg injection into molten steel in the vacuum vessel.

Table 1. Chemical composition of sample (mass%).

C	Si	Mn	P	S	Al	Ti	O	N
0.06	0.95	1.30	0.007	0.0004	0.05	0.16	0.0038	0.0057

To inject a magnesium vapor into molten steel which was generated on site, the implement consisted of double pipes was fixed at the flange on the upper part of the vacuum vessel. The immersion lance made of Mo–ZrO₂ cermet of 2.5×10⁻² m in outside diameter, 1.0×10⁻² m in inside diameter and 3.0×10⁻¹ m in length. At the end of the immersion lance, the hole of 3.0×10⁻³ m in diameter was made. The pure magnesium wire of 3.0×10⁻³ m in outside diameter was fixed at the end of the inside pipe. The injection of magnesium vapor into molten steel was done according to the following procedures. The range of 1.0×10⁻¹ m from the end of lance was immersed in the molten steel and was heated. It was found the temperature in lance was reached at 1673 K from the measurement result by using a thermocouple. Because a boiling temperature of pure magnesium is 1363 K, the magnesium wire can be made vapor in the lance. And then, the inner pipe was pushed down in the lance to which was heated at speed 1.0×10⁻¹ ms⁻¹ and a magnesium wire of 0.5–2.0 m in length was fed. This length of wire corresponds to the 6−25×10⁻³ kg of pure magnesium. To inject a magnesium vapor into molten steel, an argon gas was serviced by condition of flow rate 8.3×10⁻⁵ m³s⁻¹ and pressure 1.0×10¹ MPa from the above the outer pipe. After injection of a magnesium vapor, a sample for the analysis of magnesium concentration was obtained from the molten steel by immersion of bomb sampler made of steel. Moreover, for comparison the experiment by which a magnesium lump of 0.5–2.0×10⁻³ kg was added in a molten steel was done in several times. And then the concentration of magnesium in the molten steel added by the lump was compared with that of injected by vapor.

2.3. Advancing Condition of Magnesium Wire

It is necessary to examine the advancing conditions of a magnesium wire to generate a magnesium vapor in the immersion lance. Then, the relationship between temperature and advancing speed of the magnesium wire into the lance was simulated by heat transfer calculation.

As shown in Fig. 3, the length of lance was 3.0×10⁻¹ m and the range of 1.0×10⁻¹ m from the end was heated to 1873 K by molten steel. The magnesium wire which was advanced in the lance was made vapor by heat of radiation.

Fig. 3.

Schematic diagram of generation of magnesium vapor in the immersion lance.

The decrease in temperature by flow of an argon gas was determined by the heat transfer boundary condition from the Nusselt number.¹⁸⁾   

$$Nu=\frac{hD}{k}=1.86\times {\left(Re\times Pr\times \frac{D}{L}\right)}^{\frac{1}{3}}$$(1)

  

$$Re=\frac{u\rho D}{\eta },\hspace{1em}Pr=\frac{\eta C_p}{k}$$

where, D: inner diameter of lance (=1.0×10⁻² m), L: length of lance (=3.0×10⁻¹ m), u: flow rate of an argon gas in lance (=1.1×10⁻² ms⁻¹), ρ: density of an argon gas, k: thermal conductivity of an argon gas, η: viscosity of an argon gas, C_p: specific heat of an argon gas.

High temperature physical properties of argon gas and pure magnesium are listed in Table 2.

Table 2. Physical properties for calculation.

symbol	property	value	Ref.
ρ	density of an argon gas	1.78 (kgm−3)	19)
k	thermal conductivity of an argon gas	5.0×10−2 (Jm−1s−1K−1)	19)
η	viscosity of an argon gas	7.6×10−7 (kgm−1s−1)	20)
Cp	specific heat of an argon gas	5.22×10−4 (Jkg−1K−1)	18)
Tm	melting temperature of magnesium	922 (K)	19)
Tb	boiling temperature of magnesium	1363 (K)	19)
M	molar weight of magnesium	2.43×10−2 (kgmol−1)	19)
	density of magnesium	1.74×103 (kgm−3)	19)
	thermal conductivity of magnesium	1.0×102 (Jm−1s−1K−1)	19)
	specific heat of magnesium	1.36×103 (Jkg−1K−1)	19)
	heat of fusion of magnesium	1.07×102 (Jkg−1)	19)
	heat of evaporation of magnesium	1.54×103 (Jkg−1)	19)
R	gas constant	8.314 (JK−1mol−1)	19)

3. Results and Discussion

3.1. Experiment of Tammann Furnace

Figure 4 shows the measurement result of magnesium temperature when the distance between the end of lance and magnesium holder was 2.4×10⁻¹ m. Holder means Mg holder shown in Fig. 1. After the lance moved down, the temperature of magnesium holder increased and reached the temperature just below boiling point of pure magnesium. Then, the temperature of the holder decreased when the lance moved up. The experiment by which the temperature of magnesium holder was 1243 K was added. In order to examine the vaporization behavior of magnesium, the relationship between amount of evaporation and temperature was predicted by following equation.²¹⁾   

$$\frac{dw}{dt}=S\times p_{Mg}\times \sqrt{\frac{M}{2\pi RT}}$$(2)

where, w: weight of magnesium(kg), t: time (s), S: evaporation area (=1.26×10⁻⁵ m²), M: molar weight of magnesium, R: gas constant.

Fig. 4.

Temperature change of magnesium in the holder with time.

The vapor pressure of magnesium as a function of temperature was predicted by the following equation.²²⁾   

$$log{p}_{Mg}=12.79-1.41\times logT-\frac{7\mathrm{\hspace{0.17em} }550}{T}$$(3)

Figure 5(a) shows the relationship between vapor pressure of magnesium and temperature. The vapor pressure increased with the temperature more than melting temperature. It was understood the magnesium vapor could be generated enough even if the temperature of magnesium did not exceed the boiling point.

Fig. 5.

Relationship between (a) vapor pressure, (b) evaporation rate and temperature. Arrows show experimental temperature.

Figure 5(b) shows the relationship between evaporation velocity of magnesium and temperature. As evaporation velocity depended on the vapor pressure, the higher temperature of magnesium was, the larger evaporation velocity. It was understood that magnesium could evaporate rapidly by increasing the temperature of magnesium to near the boiling temperature. At the temperature in this experiment, evaporation of magnesium would be completed within 1.2 s or 4.0 s. Observing in the holder after the experiments, it was confirmed that magnesium disappeared completely. Moreover, there was no vaporized magnesium in the inner wall of lance, then the magnesium vapor was thought to be injected in molten steel. The analytical results of magnesium concentration in the ingots were 0.0042 mass% and 0.0035 mass%, and the additive yield of magnesium was 53% and 44% respectively. It was confirmed that the magnesium vapor which was generated by using a pure magnesium lump was able to be injected in molten steel. Moreover, when a magnesium vapor was injected in molten steel, there was no splash of molten steel. Therefore, it was also confirmed that large scale experiment was able to be done.

3.2. Experiment of High Frequency Induction Furnace in Vacuum Vessel

The relationship between the magnesium concentration and the additive amount is shown in Fig. 6. The result of magnesium concentration was analyzed with the steel sample which was obtained from a molten steel by using a steel sampler.

Fig. 6.

Relationship between concentration of magnesium in molten steel at 1873 K and additive amount of magnesium.

The magnesium lump was wrapped with pure iron foil and was fixed at the end of steel rod which was held on the movement stage for up and down and was added in molten steel. The addition of a magnesium lump into molten steel was repeated. As the reaction at addition of magnesium lump was severe, it was hard to increase the magnesium concentration in molten steel.

However, when the magnesium vapor was injected into molten steel, there was no splush of molten steel and the magnesium concentration of molten steel was able to increase.

Figure 7 shows the yield of magnesium in the case of addition of magnesium lump into molten steel and in the case of injection of magnesium vapor. When the magnesium lump was added into molten steel, the reaction between molten steel and magnesium was intense and the splush occurred. The reason for this was that the vapor of magnesium was generated at the surface of molten steel. The reason why the yield of magnesium was high in the case of vapor injection was thought that there was no expansion with transformation from solid to gas as in the case of addition of lump, and then the magnesium was supplied in molten steel stably.

Fig. 7.

Comparison of yield of Mg by difference of addition method.

The reaction of magnesium and oxygen was described by the following equation.^23,24,25)   

$$Mg\left(g\right)\hspace{0.17em}\to \hspace{0.17em}\underset{\_}{Mg}$$(4)

  

$$\underset{\_}{Mg}\hspace{0.17em}+\hspace{0.17em}\underset{\_}{O}\hspace{0.17em}\to \hspace{0.17em}MgO$$(5)

  

$$K=\frac{P_{Mg}\times a_O}{a_{MgO}}=P_{Mg}\times a_O,\mathrm{\hspace{0.17em} }{a}_{MgO}=1.0$$(6)

It is thought that either P_Mg or a_o becomes a governing factor in Eq. (6) depending on an addition method of magnesium into molten steel. When the lump of magnesium is added into molten steel, the concentration of magnesium is determined by activity of oxygen a_o and equilibrium constant K. When the vapor of magnesium is injected into molten steel, the vapor pressure P_Mg becomes governing factor and the effect of a_o becomes small. Therefore, the magnesium concentration in molten steel can increase by injection of vapor. The reaction which progresses while balanced is changed and the reaction can be changed arbitrarily by compulsorily changing the variable of steam pressure. It is thought that the equilibrium constant can be changed arbitrarily by changing with vapor pressure. The vapor pressure and the equilibrium constant of magnesium have been studied,²⁶⁾ it was found that the value K was 0.023 at temperature 1873 K. In this study, this value of K was adopted and the concentration of magnesium was examined.

Figure 8 shows the relationship between the concentration and the partial pressure of magnesium. The range of arrow shows the range of experimental results for magnesium concentration in either case of vapor injection or lump addition. As the concentration of magnesium in the case of vapor injection was higher than that of lump addition, it was understood that the partial pressure of magnesium was also higher. As the effect of injection of vapor into molten steel was to be able to increase the partial pressure of magnesium, it was clarified that the process of injection of vapor changed the chemical reaction actively. The yield of magnesium at the injection of vapor into molten steel was higher than the yield at the addition of lump, then the amount of magnesium could decrease and difficulty for operation could decrease. Therefore, it is thought that the injection of magnesium vapor becomes useful method for increasing the concentration of magnesium industrially. Moreover, it is a revolutionary technique in which the addition of high concentration for magnesium is possible.

Fig. 8.

Relationship between concentration and partial pressure of magnesium. Arrows show magnesium concentration range of experimental results for lump addition and vapor injection.

3.3. Advancing Conditions of Magnesium Wire

The temperature change of the magnesium wire advanced in the immersion lance was predicted by heat transfer calculation, and the result of temperature was shown below.

Figure 9 shows the relationship between the distance in lance and the advancing velocity of wire, when the temperature of core of magnesium wire reaches 1373 K which exceeds boiling temperature. The higher the advancing velocity of wire was, the longer the distance for generation of magnesium vapor in the lance. From this result, it is necessary to decrease the advancing velocity of wire below 6.0×10⁻² ms⁻¹. In this experiment, the insertion length of wire into the lance became about less than one third of total length of lance by adjusting the velocity to 1.7×10⁻² ms⁻¹.

Fig. 9.

Relationship between distance from lance entrance where Mg wire reaches 1373 K and Mg wire advancing speed.

Figure 10 shows the change of core temperature of wire with the distance from entrance of lance when the advancing velocity is 1.7×10⁻² ms⁻¹. It was confirmed that the temperature of magnesium wire which was advanced in the lance increased with increasing distance from entrance, and reached the boiling temperature of magnesium at the distance about 3.0×10⁻² m and then the generated the magnesium vapor.

Fig. 10.

Calculated temperature profile in the core of magnesium wire in the lance.

3.4. TiN in the Sample

The magnesium oxide behaves an effective heterogeneous nucleus of TiN. In this study, the experiment of casting was done by using molten steel with high concentration of titanium. From these results, the behavior of heterogeneous nucleation of TiN on the magnesium oxide was examined when the magnesium vapor was injected into molten steel.

The samples of 2.0×10⁻² m on a side were obtained from the ingot at the center of width and the quarter of thickness where the magnesium vapor was injected. The samples was polished like mirror plane by using abrasive diamond compound of 1 μm in diameter. An absolute alcohol was used for a lubricating liquid to polish samples and inclusions were prevented solving to water. And then, the polished face of samples was etched with controlled potential electrolysis and TiN was brought into sight. These TiN were observed by using FE-SEM with EDS. Moreover, some of samples were observed by using FE-SEM with EDS. The solvent²⁷⁾ was made by mixing with triethanolamine 20 mL, methanol 1000 mL, tetramethylammoniumchloride 1.0×10⁻² kg, and barium 1.0×10⁻⁴ kg. The favorable electrolyte conditions were voltage of −150 mV and electric charge 150 C.

Figure 11(a) shows the FE-SEM image of TiN and magnesium oxide when magnesium vapor was injected into molten steel during solidification. The dissolved magnesium concentration was 1.1×10⁻⁴ mass%. From the results of EDS analysis shown in Figs. 11(c) and 11(d), a substance existed in the core of cubic TiN was a spinel MgO·Al₂O₃. It was confirmed that the spinel became a heterogeneous nucleus of TiN the same as previous studies.^{8,9,10,11,12)}

Fig. 11.

Photograph of TiN with spinel by (a) FE-SEM and (b) FE-TEM. Analytical results of spinel by (c) EDS and TiN by (d) EDS.

Figure 11(b) shows the image of TiN by FE-TEM. It was clarified that the spinel whose shape was square of 500 nm on a side existed at the core of TiN and acted as a heterogeneous nucleus.

Figure 12 shows a spinel which exists independently in the sample injected magnesium vapor by observation of FE-TEM. TiN did not nucleate around the spinel which played as a heterogeneous nucleus. It is thought that the frequency of occurrence for heterogeneous nucleation depends on the curvature radius of nucleus besides the disresistry of crystal between inclusion and nucleus. It seems that there is a possibility where a critical radius for generation of heterogeneous nucleation. It seems that the research on both the shape and the curvature radius of the nucleus to frequency of nucleation will become important in the future.

Fig. 12.

Photograph of spinel observed by FE-TEM.

As stated above, when the magnesium vapor is added in molten steel, fine magnesium oxides generate in molten steel. It was clarified to be able to generate TiN heterogeneously on the magnesium oxide.

The injection of metallic vapor into molten steel becomes possible the generation of the inclusion whose size is micron or nano size in high density without using the expensive rare metals. The technique for both addition of cheap alloying element under high concentration and dispersion of fine nucleus is possible to become a revolutionary technique for steelmaking process and materials processing in the future.

Moreover, it is thought that the study on the decrease of the liquid-solid interfacial energy to improve the frequency of heterogeneous nucleation in the future.

4. Conclusions

High Performance products of carbon steel are demanded recently. Some properties of final product are occasionally affected by the quality of steel slabs. Then the control of both the dendrites and the inclusions becomes important. The use of the heterogeneous nucleus has the effect to generate the inclusions and the study to control the generation of heterogeneous nucleation becomes urgent.

In this study, to clarify the influence on the heterogeneous nucleus of the vapor of an additive element, the relationship between the generation behavior of TiN and the magnesium vapor which is constituent for the spinel for generate the heterogeneous nucleus was examined. From the results by using the tammann furnace and the high frequency induction furnace, the following conclusions were obtained.

(1) The technique of the magnesium vapor injection into molten steel enable to increase the partial pressure of magnesium in equilibrium reaction. It is possible to make the high concentration and the high yield of magnesium in molten steel at the same time, which is impossible in the conventional process. From these results of experiment by using the high frequency induction furnace, it was clarified that the magnesium concentration in the molten steel could be controlled by changing the amount of magnesium vapor.

(2) The technique of magnesium vapor injection into molten steel make it possible to generate the heterogeneous nucleus for TiN. Under the experimental conditions, the spinel MgO·Al₂O₃ is formed as a heterogeneous nucleus for TiN. A concentration of dissolved magnesium needed for generating this nucleus is about 1×10⁻⁴ mass%. Moreover, it seems that the lower limit of nucleus size for TiN exist to make effective use of this nucleus for TiN, this limit value is from tens to hundreds nm.

(3) The technique for the metallic vapor injection into molten steel becomes a revolutionary process which generates the heterogeneous nucleus in micron or nano size at high concentration without the expensive rare metal in the casting process.

References

• 1)   T.  Maki: CAMP-ISIJ, 13 (2000), 730.
• 2)   C.  Ouchi: CAMP-ISIJ, 13 (2000), 750.
• 3)   S.  Mizoguchi: Progress in Metallurgy for Ultra Refinement of Grain Size – Advanced Technology for Future Development in New Application of Steel, ISIJ, Tokyo, (2002), 53.
• 4)   H.  Mizukami,  K.  Hayashi,  M.  Numata and  A.  Yamanaka: Tetsu-to-Hagané, 97 (2011), 457.
• 5)   B. L.  Bramfitt: Metall. Trans., 1 (1970), 1987.
• 6)   T.  Ohashi,  T.  Hiromoto,  H.  Fujii,  Y.  Nuri and  K.  Asano: Tetsu-to-Hagané, 62 (1976), 614.
• 7)   O.  Grong: Metallurgical Modelling of Welding 2nd ed., The Institute of Metals, London, (1997), 221.
• 8)   G. V.  Pervashin and  H.  Suito: ISIJ Int., 41 (2001), 748.
• 9)   P.  Misra,  S.  Sridhar and  A. W.  Cramb: Metall. Trans. B, 32B (2001), 963.
• 10)   H.  Fujimura,  S.  Tsuge,  Y.  Komizo and  T.  Nishizawa: Tetsu-to-Hagané, 87 (2001), 707.
• 11)   H.  Ohta,  R.  Inoue and  H.  Suito: ISIJ Int., 48 (2008), 294.
• 12)   K.  Isobe: ISIJ Int., 50 (2010), 1972.
• 13)   N.  Wada: Jpn. J. Appl., 7 (1968), 1287.
• 14)   G. A.  Iron and  R. I. L.  Guthrie: Metall. Trans. B, 12B (1981), 755.
• 15)   J.  Yang,  K.  Okumura,  M.  Kuwabara and  M.  Sano: ISIJ Int., 42 (2002), 685.
• 16)   S.  Mukawa,  Y.  Ueshima,  M.  Sano,  J.  Yang and  M.  Kuwahara: ISIJ Int., 46 (2006), 1778.
• 17)   J.  Yang,  M.  Kuwabara,  T.  Sakai,  N.  Uchida,  Z.  Liu and  M.  Sano: ISIJ Int., 47 (2007), 418.
• 18)   R. B.  Bird,  W. E.  Stewart and  E. N.  Lightfoot: Transport Phenomena, John Wiley & Sons, New York, (1976), 399.
• 19)  National Astronomical Observatory of Japan: Chronological Scientific Tables, Maruzen Publication Tokyo, (2015), 369.
• 20)   G. H.  Geiger and  D. R.  Poirier: Transport Phenomena in Metallurgy, Addison-Wesley Publication, MA, (1980), 7.
• 21)   E.  Kato: Tetsu-to-Hagané, 74 (1988), 764.
• 22)   O.  Kubaschewski and  C. B.  Alcock: Metallurgical Thermochemistry, Pergamon Press, Oxford, (1979), 366.
• 23)   H.  Itoh,  M.  Hino and  S.  Banya: Tetsu-to-Hagané, 83 (1997), 623.
• 24)   G.  Okuyama,  K.  Yamaguchi,  S.  Takeuchi and  K.  Sorimachi: ISIJ Int., 40 (2000), 121.
• 25)   K.  Fujii,  T.  Nagasaka and  M.  Hino: ISIJ Int., 40 (2000), 1059.
• 26)   X.  Zhang, Q. Han and D. Chen: Metall. Trans. B, 22B (1991), 918.
• 27)   R.  Inoue and  H.  Suito: Progress in Research on Ultra Clean Steel, ISIJ, Tokyo, (1988), 348.