Effect of Deoxidizing Element on the Hot Ductility of Boron-Containing Steel

Abstract

Improving the surface quality of the casting slab increases the productivity of steel. The slab surface sometimes has transverse cracks along a grain boundary when the slab is bent and/or unbent around the transformation temperature from the austenite to ferrite phases. In particular, for boron-containing steel, the defect is strongly influenced by the precipitation of BN on the grain boundary. In this study, the effect of deoxidizing elements, such as aluminum, calcium, and zirconium, on the hot ductility of boron-containing steel have been investigated fundamentally. The addition of zirconium or calcium improves the hot ductility of boron-containing steel by comparing with that of aluminum. An oxide containing calcium or zirconium acts as a more effective nucleus for BN precipitation than alumina, and the excess precipitation of BN on the grain boundary is suppressed. This improves the hot ductility of boron-containing steel in the region of single-phase austenite. Moreover, the hot ductility of Al+Zr-added steel is the best, even in the region of coexisting austenite and ferrite phases, because the precipitation of ferrite (α-Fe) on the grain boundary is suppressed compared to that of Al+Ca-added steel.

1. Introduction

It is important to improve the surface quality of slabs in the continuous casting process of steel, as doing so results in improved productivity. Surface defects on the slab adversely affect the hot direct rolling process. Suzuki et al. reported that the ductility of steel becomes poor in three temperature ranges, expressed as zones I, II, and III.¹⁾ The surface defects of the casting slab have a strong relation to the poor ductility of steel in zones I and III. Specifically, zone I is the range of temperatures greater than the solidus temperature, where longitudinal cracks and internal cracks occur, whereas zone III is the range around the transformation temperature from the austenite to ferrite phases, where transverse cracks occur on the slab surface.^2,3) This work focuses on the transverse cracks in zone III.

Transverse cracks are often observed on the slab surface in the VB- or S-type continuous caster. When the slab is bent and/or unbent, cracks occur due to the film-like ferrite and/or precipitations, such as carbide and nitride, around the grain boundary, and the fracture mode is the prior austenite intergranular. Many studies on transverse cracks of the casting slab in zone III have been reported already,^{4,5,6,7,8,9,10,11,12,13,14,15,16,17)} and the adverse effect of alloying elements have been discussed. The alloying elements, such as aluminum, boron, niobium, and vanadium, worsen the cracking susceptibility, which has a strong relation to the secondary cooling condition of slab. This is because the cooling condition affects the precipitation behavior of carbides and nitrides consisting of aluminum, boron, niobium, and/or vanadium. In particular, even only a few mass ppm of boron, which greatly improves mechanical properties,¹⁸⁾ such as hardenability and creep property, worsens the susceptibility to transverse cracks.^{8,12,14,15,16)}

The diffusion coefficient of boron in steel is much larger in comparison with those of other elements.^19,20) Moreover, the grain boundary segregation coefficient of boron in steel is also larger than those of other elements.^21,22,23,24) These properties are significantly different from those of other elements worsening the hot ductility of steel. Therefore, a few hundred nanometer-sized of BN prefers to precipitate on the grain boundary. When a tensile stress acts on a continuously casting slab of the boron-containing steel, micro voids generate at the interface between steel and BN precipitating on the grain boundary. They coalesce one another, which causes the intergranular cracking. This is well known as the mechanism of the hot-temperature embrittlement of boron-containing steel,⁸⁾ which is slightly different from that of aluminum, niobium or vanadium-containing steel due to the precipitate-free zone of AlN, Nb(C, N), or V(C, N).^5,25) Accordingly, to improve the hot ductility of boron-containing steel, it is necessary to decrease the precipitation number of BN on the grain boundary by decreasing the nitrogen content in steel or controlling the secondary cooling rate.^8,15,16) However, the application of these approaches is not necessarily fundamental from the perspective of material designs and continuous caster specifications. In this study, a new method to improve the hot ductility of boron-containing steel has been a focus of attention.

If a few micrometer-sized of a deoxidation product, such as alumina, act as a nucleus for BN precipitation, the amount of BN precipitating on the grain boundary may be suppressed. The deoxidation product forms in molten steel. Therefore, the distribution of the deoxidation product does not depend on the solidification structure of steel, and they are seen anywhere in steel. The precipitation amount of BN, which depends on the contents of boron and nitrogen in steel, are thermodynamically determined in accordance with the solubility product of BN. Accordingly, even if the total amount of BN precipitating in steel is the same, it is considered the possibility that a decrease in the amount of BN precipitating on the grain boundary leads to improve the hot ductility of boron-containing steel. Therefore, in this study, the effect of deoxidizing elements, such as aluminum, calcium, and zirconium, on the hot ductility of steel containing boron has been fundamentally investigated and the possibilities of controlling BN morphology and improving the hot ductility have been discussed. In particular, aluminum, calcium, and zirconium were chosen as the deoxidizing elements from the following reason.

The temperature dependences of the Gibbs standard free energies of formation for some oxides and nitrides^26,27) are shown in Fig. 1. Aluminum is commonly used as the deoxidizing agent, but it is demonstrated in Fig. 1(a) that calcium has a stronger affinity with oxygen than aluminum and is a stronger deoxidizing agent. Therefore, it is of great interest to compare the influence of added aluminum or calcium on the hot ductility of boron-containing steel. On the other hand, it is found from Fig. 1(a) that an affinity between zirconium and oxygen is almost the same as that between aluminum and oxygen. However, it is demonstrated in Fig. 1(b) that zirconium has the strongest affinity with nitrogen. It is also interesting to clarify the effect of added zirconium on the hot ductility of boron-containing steel. From the above standpoints, aluminum, calcium, and zirconium were chosen as the deoxidizing elements, and it is determined in this study whether the addition of the deoxidizing element improves the hot ductility of boron-containing steel.

Fig. 1.

Temperature dependence of Gibbs standard free energies of formation for some oxides and nitrides.

2. Experimental

2.1. Hot Tensile Testing of Boron-Containing Steel

The effect of the deoxidizing elements, such as aluminum, calcium, and zirconium, on the hot ductility of boron-containing steel was experimentally investigated. Three kinds of boron-containing steels, called Al-, Al+Ca-, and Al+Zr-added steels, were preliminary melted in a vacuum induction melting furnace, and the ingots were cast as the specimen for hot tensile testing. The chemical compositions of each specimen are tabulated in Table 1. The nitrogen and oxygen contents were analyzed by the inert gas carrier melting-infrared absorption method. The carbon and sulfur contents were analyzed by the infrared absorption method after combustion in an induction furnace. Others contents of each specimen were analyzed by the inductively coupled plasma (ICP) emission spectrometry. The basic components of ingots consist of aluminum, boron, carbon, manganese, nitrogen, phosphorus, silicon, sulfur, and titanium, and the contents were almost the same among the three specimens. In particular, it is noticed that the content of calcium or zirconium in Al+Ca- or Al+Zr-added steel is high to clarify the effect of deoxidizing element. Each ingot was forged and the cylindrical bar (φ10 mm × 190 mm) was prepared as the tensile specimen, and the detail shape is drawn in Fig. 2(a). Both ends of the tensile specimen were threaded to fix it during the tensile testing, and the gauge distance is 30 mm in the hot tensile testing.

Table 1. Chemical compositions of Al-added, Al+Ca-added and Al+Zr-added steels used in the present experiment.

[mass%]
		C	Si	Mn	P	S	Ti	Al	Ca	Zr	B	N	T. O
(a)	Al-added steel	0.091	0.15	1.44	<0.001	0.0030	0.010	0.009	–	–	0.0017	0.0046	0.0027
(b)	Al+Ca-added steel	0.091	0.16	1.44	<0.001	0.0029	0.009	0.009	0.004		0.0016	0.0048	0.0027
(c)	Al+Zr-added steel	0.090	0.15	1.44	<0.001	0.0031	0.010	0.013	–	0.011	0.0017	0.0047	0.0027

Fig. 2.

Schematic illustrations of tensile specimen and the present experimental apparatus. (Online version in color.)

A schematic illustration of the hot tensile testing apparatus used in this study is also shown in Fig. 2(b).¹³⁾ A tensile specimen in a cold crucible was inductively heated and melted before the deformation. The specimen in the liquid state was held by electromagnetic force without contacting the cold crucible. The temperature of the specimen was measured by an optical pyrometer set above the cold crucible. The correlation between the measured temperature and that in the crucible was determined in advance. Accordingly, this apparatus is capable of evaluating the hot ductility of steel in the thermal history condition to simulate the phase-transformation process, including solidification. The thermal and strain history in this experiment is shown in Fig. 3. The tensile specimen was heated to 1823 K in an argon atmosphere and melted. After being held at this temperature for 120 s, the specimen was cooled to 1523 K at a cooling rate of 10 K/s. Moreover, it was continuously cooled to the tensile temperature, which ranged from 973 K to 1273 K, at a cooling rate of 0.40 K/s. After being held at this temperature for 120 s, the sample was deformed and fractured at a strain rate of 3.3×10⁻⁴ s⁻¹, which roughly corresponds to the bending/unbending rate working on the surface of a continuously cast slab. After the specimen was fractured, the specimen was cooled at a cooling rate of about 5 to 7 K/s using an argon gas. During the experiment, the tensile force and elongation of the specimen were monitored, and the diameter of the specimen before and after tensile testing was measured by a Vernier caliper.

Fig. 3.

Schematic illustration of thermal and strain history in this experiment.

2.2. Observation of Specimen after Hot Tensile Testing

The specimen after tensile testing was examined by several methods. First, the fractographic morphology was observed using a scanning electron microscope to distinguish whether the fracture mode was a ductile fracture or intergranular fracture. The precipitations on the intergranular fracture in some cases were also observed. Next, a longitudinal section of the fractured sample was polished and etched in a nital solution consisting of ethanol and 5 vol% nitric acid. The solidification structure of the section was observed with an optical microscope, and the grain size was evaluated. Finally, the precipitation morphology in the as-etched sample was observed using the scanning electron microscope, and was analyzed by means of energy-dispersive X-ray spectroscopic analysis.

3. Results and Discussion

3.1. Phase Equilibrium and Hot Ductility Evaluation of Boron-Containing Steel

As an example, the phase equilibrium of Al-added steel was calculated with Thermo-Calc^28,29) using the TCFE8 database, and the phase-equilibrium diagram is shown in Fig. 4. The phase transformation from austenite (γ-Fe) to ferrite (α-Fe) is correlated with poor ductility of steel in zone III,¹⁾ and it is demonstrated in Fig. 4 that the equilibrium transformation temperature is approximately 1100 K. Moreover, it is found that Ti (C, N), BN, and AlN precipitate in equilibrium at precipitation temperatures of 1700 K, 1355 K, and 1078 K, respectively, which are higher than the equilibrium transformation temperature from austenite (γ-Fe) to ferrite (α-Fe). If the hot ductility of steels used in this study is poor over the equilibrium transformation temperature from austenite (γ-Fe) to ferrite (α-Fe), the poor ductility is considered to arise from precipitation of nitrides such as Ti(C, N) and BN.

Fig. 4.

Phase-equilibrium diagram of Al-added steel calculated by Thermo-Calc using TCFE8 database. (Online version in color.)

The stress–strain curve of Al-added steel under a tensile temperature of 1123 K is exemplified in Fig. 5. It is found that the specimen was deformed elastically in the initial stage of linear relation between strain and stress. Then, it was deformed plastically. It gradually began to neck and finally was fractured. This is a typical stress–strain curve. The accuracy of the hot tensile testing was confirmed by checking the curve of each experiment, and the maximum tensile strength and reduction in area (R. A.), defined as Eq. (1), were determined.   

$$R.A.=\frac{A_0-A_{\text{f}}}{A_0}\times 100\hspace{1em}(\%)$$(1)

Here, A₀ (m²) and A_f (m²) respectively denote the cross-sectional area of the sample before or after hot tensile testing. In particular, higher reduction in area corresponds to better hot ductility of the steel.

Fig. 5.

Stress–strain curve of Al-added steel under a tensile temperature of 1123 K.

3.2. Effect of Deoxidizing Element on Hot Strength and Ductility of Boron-Containing Steel

Figure 6 shows the relationship between the tensile temperature and maximum tensile strength of each sample. The maximum tensile strength of steel depends on the tensile temperature, but is almost same regardless of the kind of steel. Accordingly, it can be understood that the effect of the deoxidizing element on the hot strength of boron-containing steel is very small. On the other hand, the relationship between the tensile temperature and area reduction of each steel is displayed in Fig. 7. The hot ductility of boron-containing steel is observed to be very different among the three kinds of steels. The reduction in area in the case of Al-added steel is less than 60% within the temperature range from 973 K to 1173 K, which suggests that the hot ductility of Al-added steel is poor, even over the equilibrium transformation temperature from austenite (γ-Fe) to ferrite (α-Fe). The hot ductility of Al+Ca-added steel becomes better than that of Al-added steel within the temperature range from 1173 K to 1273 K. Moreover, the reduction in area in the case of Al+Zr-added steel is more than 60% at the higher temperature 1073 K, and the hot ductility of Al+Zr-added steel is greatly improved over those of Al- and Al+Ca-added steels. Accordingly, these results suggest that deoxidizing elements, such as aluminum, calcium, and zirconium, significantly influence the hot ductility of boron-containing steel. In particular, the best hot ductility is that of Al+Zr-added steel, followed by those of Al+Ca- and Al-added steels, although the quantitative effect of improving the hot ductility of steel may depend on the amount of deoxidizing element added.

Fig. 6.

Relationship between the tensile temperature and maximum tensile strength of Al-added, Al+Ca-added, and Al+Zr-added steels.

Fig. 7.

Relationship between tensile temperature and reduction in area of Al-added, Al+Ca-added, and Al+Zr-added steels.

Next, the experimental results at tensile temperatures of 1173 K and 1073 K are discussed in detail by observing the fractographic morphology, solidification structure, and precipitation morphology, such as nitride. The tensile temperature of 1173 K corresponds to the austenite-phase region, whereas that of 1073 K is within the coexistence region of the austenite and ferrite phases. The effect of the deoxidizing element on the hot ductility of boron-containing steel will be discussed from a sequence of observations.

3.3. Effect of Deoxidizing Element on Fractographic Morphology

The fractographic morphology of steel after the hot tensile testing was observed using a scanning electron microscope. The secondary electron images of fractographic morphologies at tensile temperatures of 1173 K and 1073 K are shown in Figs. 8 and 9, respectively. The fractographic morphologies at the same tensile temperature are different among three kinds of steel. The morphologies in Figs. 8(a) and 8(b) are typical intergranular fractures, whereas that in Fig. 8(c) is a ductile fracture. Similarly, the morphologies in Figs. 9(a) and 9(b) are typical intergranular fractures, whereas that in Fig. 9(c) is a ductile fracture. The fractographic morphology is distinguished by open or solid mark in Fig. 7. It can be understood from these results that the fractographic morphology changes from a ductile fracture to an intergranular fracture when the reduction in area is less than approximately 60%. This means that the fracture mode can be predicted from the value of R.A. in the present experiment. Accordingly, it is important to clarify the condition that yields the approximately 60% or higher area reduction to improve the grain-boundary embrittlement of boron-containing steel.

Fig. 8.

Secondary electron images of fractographic morphology at a tensile temperature of 1173 K.

Fig. 9.

Secondary electron images of fractographic morphology at a tensile temperature of 1073 K.

Moreover, the precipitations on the intergranular fracture at the tensile temperature of 1173 K were observed by using an SEM-EDX, and the typical secondary electron images are shown in Figs. 10(a) and 10(b). The main precipitations on the intergranular fracture were a few hundred nanometer-sized of BN in both cases of Al- and Al+Ca-added steels. This seems to agree with the tendency suggested by the mechanism of the hot-temperature embrittlement of born-containing steel reported in the previous works.^8,16)

Fig. 10.

Secondary electorn images of precipitations on the intergranular fractorgraphic in the case of Al- and Al+Ca-added steels at a tensile temperature of 1173 K.

3.4. Effect of Deoxidizing Element on Solidification Structure

The longitudinal section of fractured sample was etched in a nital solution consisting ethanol and 5 vol% nitric acid, and the solidification structure of the section was observed using an optical microscope. Figures 11 and 12 show the optical images of solidification structures at tensile temperatures of 1173 K and 1073 K, respectively. When the hot ductility is good, as shown in Fig. 11(c), the grain boundary of the prior austenite (γ-Fe) is not necessarily seen clearly, and the solidification structures seem to be affected by the dynamic recrystallization. The dynamic recrystallization is likely to occur with increasing the total strain until fracturing, and as a result, the tendency is considered to be observed in the case of good ductility. Nevertheless, it is found from Figs. 11 and 12 that the grain size does not depend on the kind of steel, and the size is approximately 1–3 mm. This means that the difference in the hot ductility of boron-containing steel among these three kinds of steel is not due to the grain size, at least under the present experimental condition.

Fig. 11.

Optical images of solidification structure at a tensile temperature of 1173 K. (Online version in color.)

Fig. 12.

Optical images of solidification structure at a tensile temperature of 1073 K. (Online version in color.)

First, the difference in the hot ductility of steel at 1173 K is of great interest, although the morphologies of the solidification structure at a tensile temperature of 1173 K do not differ among the three kinds of steel. This tensile temperature corresponds to the austenite-phase region, and the hot ductility of steel is considered to be influenced by precipitations on the grain boundary. Next, a tensile temperature of 1073 K is considered within the coexistence region of the austenite and ferrite phases, and the precipitating ferrite phase is observed clearly in Fig. 12. In particular, the thickness of the precipitating ferrite phase on the grain boundary in the case of Al+Ca-added steel is larger than those of Al-added and Al+Zr-added steels, and this is suggested as one of the causes worsening the hot ductility. The hot ductility of steel in the region of austenite phase may be considered to be affected by precipitations on the grain boundary, and that in the coexistence region of the austenite and ferrite phases may be affected by both precipitation and the ferrite phase on the grain boundary. Therefore, the effect of precipitation will be discussed.

3.5. Effect of Deoxidizing Element on Precipitation Morphology

It is suggested from a set of results that the hot ductility of boron-containing steel in the region of austenite phase is significantly affected by the precipitations on the grain boundary. The precipitation in steels at a tensile temperature of 1173 K was observed using a scanning electron microscope and analyzed by energy-dispersive X-ray spectroscopic analysis. All observed images shown here are back scattered electron ones. Figure 13 shows the observation result in the case of Al-added steel, and it is found that there are 1 μm or less voids and a few hundred nanometer-sized of BN on the grain boundary of the prior austenite. Boron nitride precipitating on alumina, which distributes anywhere in steel, was seldom observed. Figure 14 shows the morphology of BN in Al+Ca-added steel. Although BN precipitating on the grain boundary as shown in Fig. 14(b) was sometimes found, BN precipitating on a few micrometers of oxide and oxysulfide, as shown in Fig. 14(a), was mainly observed. Moreover, two morphologies of nitrides in Al+Zr-added steel are shown in Figs. 15(a) and 15(b), and both were observed anywhere regardless of the grain boundary. One is boron nitride precipitating on a few micrometers of zirconium oxide, and the other is zirconium nitride. These results suggest that these oxides containing calcium or zirconium act as a nucleus for BN precipitation in the cases of Al+Ca- and Al+Zr-added steels. Moreover, these oxides can distribute anywhere in solid steel regardless of solidification structure because they form in molten steel as a deoxidation product. Also, the size of each oxide is over a few micrometer, which is much larger than the size of precipitation causing the intergranular cracking.^5,25) It is considered that the morphology of BN rather than that of each oxide strongly affects on the hot ductility of boron-containing steel. Consequently, the hot ductility of Al+Ca- and Al+Zr-added steels in the region of the austenite phase is improved due to the morphology of BN. In addition, zirconium has a strong affinity with nitrogen. In the case of Al+Zr-added steel, the formation of ZrN at higher temperatures suppresses the excess precipitation of BN, which also results in improved hot ductility of boron-containing steel. On the other hand, in the case of Al+Ca-added steel, it is known from Fig. 7 that the hot ductility of Al+Ca-added steel at a tensile temperature of 1073 K is poorer that that at a tensile temperature of 1173 K. This cause is considered to be the thickness of the precipitating ferrite phase on the grain boundary.

Fig. 13.

Morphologies of precipitations in Al-added steel after tensile testing at 1173 K.

Fig. 14.

Morphologies of precipitations in Al+Ca-added steel after tensile testing at 1173 K.

Fig. 15.

Morphologies of precipitations in Al+Zr-added steel after tensile testing at 1173 K.

The nucleation of BN on the oxide in steel is illustrated in Fig. 16, and the relation of Eq. (2) is eastablished.³⁰⁾ In particular, the precipitation of BN on the oxide occurs more easily as the value for θ decreases.   

$$cos\theta =\frac{{\gamma }_{\text{Fe/Oxide}}-{\gamma }_{\text{Oxide/BN}}}{{\gamma }_{\text{Fe/BN}}}$$(2)

Here, γ_Fe/Oxide, γ_Oxide/BN, γ_Fe/BN, and θ respectively denote the interface energy between Fe and oxide, that between oxide and BN, that between Fe and BN, and the contact angle between oxide and BN.

Fig. 16.

Illustration of the nucleation of BN on the oxide in steel. (Online version in color.)

These interface energies are not necessarily available. The interface energy between Fe and BN is considered to depend on the steel composition, however, it is inferred that the effect is little among three kinds of steels. Some of the interface energies between Fe and oxide have been reported,^31,32,33) but the values are scattering widely regardless of the oxide. Accordingly, the magnitude of the interface energy between oxide and BN is considered to influence on the precipitation behavior of BN. Figure 16(a) represents the illustration in the case of larger interface energy between oxide and BN, while Fig. 16(b) is the opposite case. It is predicted from observed BN morphologies that the interface energy between calcium or zirconium oxide and BN is much smaller than that between aluminum oxide and BN. This is considered to result in a decrease of θ and to enhance the precipitation of BN on the oxide by the addition of calcium or zirconium. Nevertheless, it is possible to determine the more effective element in the present approach to improve the hot ductility of boron-containing steel if the interface energies are estimated precisely.

3.6. Mechanism of Improving Hot Ductility of Boron-Containing Steel

A set of findings are schematically illustrated in Fig. 17. First, the addition of zirconium or calcium improves the hot ductility of boron-containing steel in the higher temperature range over 1173 K because the precipitation of BN on the grain boundary is suppressed. Moreover, the added zirconium significantly improves the hot ductility due to the formation of ZrN. However, the hot ductility of Al+Ca-added steel in lower temperatures, under 1123 K, is not greatly improved and is as poor as that of Al-added steel. As shown in Fig. 12, significant precipitation of ferrite (α-Fe) on the grain boundary was observed. Accordingly, the deterioration of the hot ductility of Al+Ca-added steel is considered to arise from the film-like ferrite in the coexistence region of the austenite and ferrite phases, although the precipitation of BN on the grain boundary is suppressed.

Fig. 17.

Schematic illustration on the effect of deoxidizing element on the hot ductility of boron-containing steel. (Online version in color.)

It is found experimentally that deoxidation products containing calcium or zirconium act as a nucleus for BN precipitation, and the precipitation of BN on the grain boundary is suppressed. In particular, the addition of zirconium is the most effective for improving the hot ductility of boron-containing steel.

4. Conclusions

To prevent transverse cracks in boron-containing steel slabs, the effect of deoxidizing elements, such as aluminum, calcium, and zirconium, on the hot ductility of boron-containing steel was fundamentally investigated and discussed. The conclusions obtained are as follows:

(1) The addition of zirconium or calcium improves the hot ductility of boron-containing steel compared with that of aluminum, and the addition of zirconium is the most effective.

(2) The oxide containing calcium or zirconium acts as a more effective nucleus for BN precipitation than alumina. The excess precipitation of BN on the grain boundary is suppressed, which results in improved hot ductility of boron-containing steel in the region of single-phase austenite. Moreover, the added zirconium significantly improves the hot ductility due to the formation of ZrN.

(3) The hot ductility of Al+Ca-added steel is worse than that of Al+Zr-added steel in the coexistence region of the austenite and ferrite phases. This is because the precipitation of ferrite (α-Fe) on the grain boundary is significant.

References

• 1)   H. G.  Suzuki,  S.  Nishimura and  S.  Yamaguchi: Trans. Iron Steel Inst. Jpn., 22 (1982), 48.
• 2)  History of Steel Continuous Casting Technology in Japan, ISIJ, Tokyo, (1996), 577 (in Japanese).
• 3)  Ironmaking and Steelmaking, Handbook of Iron and Steel, 5th ed., Vol. 1, ISIJ, Tokyo, (2014), 421 (in Japanese).
• 4)   B.  Mintz and  J. M.  Arrowsmith: Met. Technol., 6 (1979), 24.
• 5)   Y.  Maehara and  Y.  Ohmori: Mater. Sci. Eng., 62 (1984), 109.
• 6)   Y.  Maehara,  K.  Yasumoto,  Y.  Sugitani and  K.  Gunji: Trans. Iron Steel Inst. Jpn., 25 (1985), 1045.
• 7)   T.  Maki,  T.  Nagamichi,  N.  Abe and  I.  Tamura: Tetsu-to-Hagané, 71 (1985), 1367 (in Japanese).
• 8)   K.  Yamamoto,  H. G.  Suzuki,  Y.  Oono,  N.  Noda and  T.  Inoue: Tetsu-to-Hagané, 73 (1987), 115 (in Japanese).
• 9)   T.  Nagamichi,  Y.  Maehara,  M.  Suzuki and  K.  Gunji: Tetsu-to-Hagané, 75 (1989), 345 (in Japanese).
• 10)   T.  Kato,  N.  Kasai,  M.  Kawamoto,  A.  Yamanaka and  T.  Watanabe: Tetsu-to-Hagané, 84 (1998), 856 (in Japanese).
• 11)   B.  Mintz: ISIJ Int., 39 (1999), 833.
• 12)   S. K.  Kim,  J. S.  Kim and  N. J.  Kim: Metall. Mater. Trans. A, 33 (2002), 701.
• 13)   T.  Kato,  Y.  Ito,  M.  Kawamoto,  A.  Yamanaka and  T.  Watanabe: ISIJ Int., 43 (2003), 1742.
• 14)   K. C.  Cho,  D. J.  Mun,  Y. M.  Koo and  J. S.  Lee: Mater. Sci. Eng. A, 528 (2011), 3556.
• 15)   K.  Taguchi and  T.  Kato: CAMP-ISIJ, 28 (2015), 686 (in Japanese).
• 16)   K.  Taguchi and  T.  Kato: Proc. 9th ECCC European Continuous Casting Conf. 2017, (Vienna), ASMET, Leoben, (2017), 682.
• 17)   K.  Furumai,  X.  Wang,  H.  Zurob and  A.  Phillion: ISIJ Int., 59 (2019), 1064.
• 18)  Steel and Alloying Elements, ISIJ, Tokyo, (2015), 346 (in Japanese).
• 19)   P. E.  Busby,  M. E.  Warga and  C.  Wells: J. Met., 5 (1953), 1463.
• 20)   H.  Oikawa: Tetsu-to-Hagané, 68 (1982), 1489 (in Japanese).
• 21)   M. P.  Seah and  E. D.  Hondros: Proc. R. Soc. Lond. A, 335 (1973), 191.
• 22)   E. D.  Hondros and  M. P.  Seah: Int. Met. Rev., 22 (1977), 262.
• 23)   E. D.  Hondros and  M. P.  Seah: Metall. Trans. A, 8 (1977), 1363.
• 24)  Thermodynamics of Microstructures, JIM, Sendai, (2005), 125 (in Japanese).
• 25)   T.  Maki: Tetsu-to-Hagané, 74 (1988), 1219 (in Japanese).
• 26)   I.  Barin,  O.  Knacke and  O.  Kubaschewski: Thermochemical Properties of Inorganic Substances, Supplement, Springer-Verlag, Berlin/Heidelberg, (1977), 55.
• 27)   D. R.  Stull and  H.  Prophet: JANAF Thermochemical Tables, U. S. Department of Commerce, Washington, D.C., (1985), 99.
• 28)   N.  Saunders and  A. P.  Miodownik: Calphad, A Comprehensive Guide, Pergamon, Oxford, UK, (1998).
• 29)  Thermo-Calc Software: Thermo-Calc Software, https://www.thermocalc.com/, (accessed 2020-03-01).
• 30)   A.  Ostrowski and  E. W.  Langer: Scand. J. Metall., 8 (1979), 153.
• 31)   H.  Sawada and  T.  Ozaki: Nippon Steel Tech. Rep., 392 (2012), 8 (in Japanese).
• 32)   K.  Ogino: Kouon Kaimen Kagaku, Agne Gijutsu Center (The Chemistry of Surfaces and Interfaces at High Temperatures), Tokyo, (2008), 253 (in Japanese).
• 33)  Technology for Control of Nonmetallic Inclusions and Production of Clean Steels, ISIJ, Tokyo, (2004), 27 (in Japanese).