Inclusion Characteristic in Tinplate Steel in RH Refining and Kinetics Limitation of Calcium Transfer by Refining Slag

Abstract

The characteristics of inclusions including composition, morphology, number, and size in tinplate steel were studied by industrial experiments and thermodynamic calculations during the RH refining process. The results indicated that two types of Al₂O₃ inclusions including cluster and single-particle are generated at first after Al addition. With the slag-metal and refractory-metal reactions, Al₂O₃ inclusions, CaO·Al₂O₃ inclusions, MgO·Al₂O₃ spinel inclusions, and CaO–MgO–Al₂O₃ ternary system inclusions are found in the middle of RH refining. Only single-particle Al₂O₃, CaO·Al₂O₃ inclusions with high melting point, and CaO–MgO–Al₂O₃ ternary system inclusions are found at the end of RH refining. From Al addition to the end of RH refining, the total number of inclusions showed a decreasing trend and the proportion of the number density decreased by 70%. About 62% of inclusions are smaller than 10 µm at the end of RH refining, which are difficult to be removed from the liquid steel. The mass transfer of Ca from the refining slag to the liquid steel has a significant effect on the content of [Ca] in liquid steel. Al₂O₃ inclusions generated in liquid steel can only be modified to CaO·Al₂O₃ inclusions in the present RH refining time. Aiming to generate 12CaO·7Al₂O₃ inclusions quickly, moderate calcium treatment as a supplementary measure for refining slag is recommended to modify inclusions during the RH refining process.

1. Introduction

Non-metallic inclusions have a great effect on tinplate steel properties. Al₂O₃ inclusions, MgO·Al₂O₃ spinel inclusions with high hardness, and CaO–Al₂O₃ inclusions with high melting point are harmful to the castability of tinplate steel and the product quality.^1,2,3,4) Besides, inclusions with high melting point are especially negative to the improvement of productivity and a typical problem caused by high-melting-point inclusions is nozzle clogging,^5,6,7) which was initiated by the agglomeration of those inclusions on the inner wall of the submerged entry nozzle of continuous casting. For the production of high-quality tinplate steel, as usual recognition, one should not only improve the cleanliness of liquid steel but also target the low-melting-point inclusions. The cleanliness of tinplate is seriously affected by non-metallic inclusions. Hence, the control of composition, morphology, number, and size of the inclusions is likely to become an effective method to reduce the harmful influence of the inclusions.

Metallic aluminum is a popular deoxidizer for its strong affinity to oxygen, yet the produced high-melting-point Al₂O₃ inclusions in liquid steel tend to agglomerate into large clusters and frequently are observed to initiate submerged entry nozzle clogging.⁸⁾ Therefore, calcium treatment is usually used to solve the nozzle clogging and blockage caused by alumina inclusions in Al-killed steel. However, although calcium treatment can modify Al₂O₃ inclusions and form low-melting-point inclusions, the relevant problems also exist. As the effect of calcium treatment is greatly affected by the composition of liquid steel and temperature, the yield of calcium is relatively low. Therefore, an excessive amount of calcium is often added to achieve the desired modification effect, but the application of calcium treatment has some harmful effects on the properties of steel if the amount of calcium is not properly controlled. Higher dissolved [Ca] may react with dissolved [S] to form solid CaS, which may also cause serious nozzle blockage during continuous casting.^8,9)

In China, some steel enterprises also adopt high-basicity refining slag to modify inclusions, and the degree of inclusion modification is closely correlated with the CaO-rich refining slag. Ca can be transferred from refining slag into liquid steel and modify Al₂O₃ into liquid CaO–Al₂O₃ system inclusions when the CaO in the refining slag is reduced by metallic aluminum. Ren et al.¹⁰⁾ pointed out that a high-basicity slag improved the cleanness of steel, while a low basicity slag gave aid to lower acid-soluble [Al] in stainless steel, which was beneficial to avoid the formation of Al₂O₃-rich inclusions and improve the deformability of inclusions. This result was confirmed by Yan et al.¹¹⁾ and Sakata.¹²⁾ So the refining slag has a great effect on the chemical composition of inclusions. In addition, MgO·Al₂O₃ spinel inclusion is also found because MgO-based refractory materials are widely used in steelmaking.^13,14) Kang et al.¹⁵⁾ indicated that MgO·Al₂O₃ spinel was not stable when the dissolved [Ca] existed in liquid steel and would transform into CaO–MgO–Al₂O₃ system.

In order to make a more accurate evaluation on the effect of inclusion modification by refining slag in the RH refining process, in this research work, the characteristics of inclusions including composition, morphology, number, and size are systematically investigated based on industrial experiments and sampling analysis. Meanwhile, according to the variations of acid-soluble [Al], [Ca] and [Mg] contents in the liquid steel, the formations of CaO–Al₂O₃ and CaO–MgO–Al₂O₃ system inclusions at different refining stages are studied by thermodynamic calculations. In addition, the controlling step of Ca transferring from refining slag to liquid steel is discussed, and then the kinetic limitation of inclusions modification by refining slag is also evaluated.

2. Experiments

2.1. Industrial Experiments and Sampling

The industrial experiments were carried out in a steel mill in China, in which the tinplate steel was produced through the steelmaking route of “hot metal pre-desulphurization → 180 t BOF steelmaking → 180 t RH refining → continuous casting”. The [C] content was 0.05–0.07% with the content of total oxygen 428 ppm at the blowing end-point of BOF, and no deoxidant agent was added into the ladle for pre-deoxidation during the tapping process. 200 kg RH refining slag is added into steel ladle at the beginning of RH vacuum treatment. During the RH refining process, liquid steel was recycling degassing for 3 min firstly. After that, metallic aluminum was added into the liquid steel for deoxidation. The additive amount of deoxidizer metallic aluminum was determined according to the [O] content in the liquid steel. In the middle of the RH refining process (refining time was about 12–14 min), MnFe alloy was added to adjust the compositions of the liquid steel. The chemical compositions of MnFe alloy are shown in Table 1. In the present work, the RH vacuum treatment time was about 20 min with vacuum pressure less than 133 Pa and the liquid steel was gently stirred for 5 min by argon gas. The chemical compositions of the refining slag used in the industrial experiment are shown in Table 2. Three samples of liquid steel were taken at different refining stages, i.e., S1 stage (after Al addition), S2 stage (middle of the RH refining process), and S3 stage (end of the RH refining process), as shown in Fig. 1.

Table 1. Chemical compositions of MnFe alloy (mass%).

C	Si	Mn	P	Fe
0.296	0.475	79.440	0.344	19.445

Table 2. Chemical compositions of RH refining slag (mass%).

CaO	Al2O3	MgO	SiO2
54.46	31.66	5.01	8.87

Fig. 1.

Illustration of sampling during the RH refining process. (Online version in color.)

2.2. Analysis of Samples

Steel samples were prepared for chemical analysis. The contents of acid-soluble [Al], [Ca], and [Mg] in the steel samples were analyzed by the ICP-AES. Total oxygen and nitrogen contents of the steel samples were determined by fusion and infrared absorption method. The inclusions in the specimens were analyzed by SEM-EDS to obtain the information of morphology, size, and chemical composition. A quantitative analysis of the inclusions was performed using the INCA software of the scanning electron microscope. To ensure the accuracy of the automated EDS analysis, the size of the detected inclusions was larger than 2 μm. In addition, the scanning region of each specimen was more than 100 mm².

3. Results and Discussion

3.1. Chemical Compositions of Steel Sample

The chemical compositions of steel samples at three different refining stages (S1, S2, S3) are given in Table 3. It can be found that the content of [N] decreases from 0.0018% to 0.0013% after RH vacuum treatment. Besides, T.O content is 428 ppm at the blowing end-point of BOF, while T.O content is 37 ppm at the end of RH refining. Meanwhile, the content of [Al] decreases 0.016% from S1 to S3 stage and Al₂O₃ inclusions would generate in liquid steel, indicating that RH vacuum treatment has a strong ability in removing nitrogen and inclusions.¹⁶⁾

Table 3. Chemical compositions of steel samples at different RH refining stages (mass%).

Sample	[C]	[Si]	[Mn]	[P]	[S]	[Al]	[Ca]	[Mg]	T.O	[N]
S1	0.0420	0.0010	0.1370	0.0170	0.0080	0.0520	0.0008	0.0006	0.0048	0.0018
S2	0.0300	0.0026	0.1500	0.0180	0.0080	0.0480	0.0018	0.0008	0.0043	0.0015
S3	0.0300	0.0026	0.1600	0.0186	0.0080	0.0360	0.0024	0.0010	0.0037	0.0013

Before Al addition during the RH refining, the oxygen content in liquid steel is so high that the contents of dissolved [Ca] and [Mg] can be considered as zero. However, under the condition of high-basicity refining slag, CaO in the RH refining slag can be reduced by dissolved [Al] and generate dissolved [Ca] in liquid steel. With the progress of the slag-metal reaction, the [Ca] content at S3 stage increases 0.0016% compared with that at S1 stage. Dissolved [Ca] through a reduction reaction would react with the deoxidation product Al₂O₃ to form CaO–Al₂O₃ system inclusions. At the same time, the presence of magnesium is also detected in liquid steel. The content of [Mg] is 0.0010% at S3 stage due to the reduction of MgO in slag by dissolved [Al], expressed by Eqs. (1) and (2).²⁾ In the calculation, the activities of MgO and Al₂O₃ are 1, and the activity coefficients of element Al and Mg are 1.024 and 0.719 respectively. When the content of [Mg] and [Al] are 0.0008% and 0.052% at S1 stage, the actual change in Gibbs free energy for Eq. (1) is −3.06 × 10⁴ J/mol. Therefore, it can be concluded that MgO in the refractory can be reduced by dissolved [Al] after Al addition.   

$$\begin{array}{l}3\mathrm{M}\mathrm{g}\mathrm{O}(\mathrm{s})+2[\mathrm{A}\mathrm{l}]=3[\mathrm{M}\mathrm{g}]+\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{0.5em}\mathrm{\Delta }{G}_{\text{1}}^{\theta }=-\text{937}\mathrm{\hspace{0.17em} }\text{860}+\text{639}\text{.47}\mathrm{\hspace{0.17em} }T\end{array}$$(1)

  

$$\mathrm{\Delta }{G}_{\text{1}}=\mathrm{\Delta }{G}_{\text{1}}^{\theta }+RTln{\displaystyle \frac{a_{{\text{Al}}_{\text{2}}{\text{O}}_{\text{3}}}\cdot a_{[\text{Mg}]}^3}{a_{\text{MgO}}^3\cdot a_{[\text{Al}]}^2}}$$(2)

Neither P₂O₅ nor MnO is found in the refining slag. It can be noted that the contents of Mn and P in the MnFe alloy are 79.440 mass% and 0.344 mass% respectively. Therefore, the increments of [Mn] and [P] contents in the liquid steel during S2 and S3 stages can be considered due to the addition of MnFe alloy, rather than through the reduction reactions of P₂O₅ and MnO in the refining slag by [Al] in liquid steel.

3.2. Characteristics of Inclusions during RH Refining Process

3.2.1. Morphology and Chemical Composition of Typical Inclusions

Typical inclusion morphology and chemical composition during the RH refining process are presented in Figs. 2, 3, 4, 5, 6. There are four types of inclusions, namely Al₂O₃ inclusions shown in Fig. 2, CaO–Al₂O₃ inclusions shown in Fig. 3, MgO·Al₂O₃ inclusions shown in Fig. 4, and CaO–MgO–Al₂O₃ system inclusions shown in Fig. 5.

Fig. 2.

Quantitative analysis result of typical Al₂O₃ inclusions in steel sample.

Fig. 3.

Quantitative analysis result of typical CaO–Al₂O₃ inclusions in steel sample.

Fig. 4.

Quantitative analysis result of typical MgO·Al₂O₃ spinel inclusions in steel sample.

Fig. 5.

Quantitative analysis result of typical CaO–MgO–Al₂O₃ system inclusions in steel sample.

Fig. 6.

Elemental mapping of typical CaO–MgO–Al₂O₃ system inclusions. (Online version in color.)

Typical Al₂O₃ inclusions observed during the three refining stages show two types of appearances. One is completely Al₂O₃ inclusion in cluster with size larger than 50 μm observed at S1 stage, as shown in Fig. 2(a), and the other is single-particle Al₂O₃ inclusion observed at both S2 and S3 stages, as shown in Fig. 2(b).

Typical CaO–Al₂O₃ system inclusions shown in Fig. 3 are globular or blocky and mainly observed at S2 and S3 stages. Based on the EDS results of Ca and Al elements, the molar ratios of Ca/Al shown in Figs. 3(a) and 3(b) are 0.55 and 0.48 respectively. As a stoichiometric amount of oxygen is supposed, CaO/Al₂O₃ molar ratio in the inclusion is approximated to that in CaO·Al₂O₃. Therefore, it can be concluded that the generated CaO–Al₂O₃ system inclusions in tinplate steel are mainly CaO·Al₂O₃ during the RH refining process.

The typical MgO·Al₂O₃ spinel inclusion shown in Fig. 4 is blocky and mainly observed in the middle of the RH refining process, but no MgO·Al₂O₃ inclusion is found at the end of RH refining.

The quantitative analysis result and elemental mappings of typical CaO–MgO–Al₂O₃ system inclusions are shown in Figs. 5 and 6 respectively. These inclusions are mainly observed at S2 and S3 refining stages. As can be found in Fig. 6(a), the distribution of Mg element inside the inclusion is different from those of Al and Ca. Mg element is mainly distributed in the inner part of the inclusion with a MgO·Al₂O₃ core, and Ca and Mg elements present complementary as MgO·Al₂O₃ spinel inclusions surrounded by an outer CaO–Al₂O₃ layer. The distribution of Mg, Ca and Al elements are uniform in Fig. 6(b). Based on Figs. 6(a) and 6(b), it can be concluded that CaO–MgO–Al₂O₃ ternary system inclusions are formed by Ca element substitution for Mg element in MgO·Al₂O₃ inclusions. After RH refining treatment, the CaO–MgO–Al₂O₃ system inclusions are mainly spherical.

The compositions of inclusions in samples taken after Al addition, in the middle of RH refining and the end of RH refining respectively, are projected into CaO–MgO–Al₂O₃ ternary, as shown in Fig. 7. With the slag-metal reaction, CaO content increases with CaO·Al₂O₃ inclusion generated in the middle of the RH refining process, as shown in Fig. 7(b). At the end of RH refining, most of CaO·Al₂O₃ inclusions change to CaO–MgO–Al₂O₃ system inclusions, and some of these inclusions are in the compositions of the liquid region, as shown in Fig. 7(c). Some MgO·Al₂O₃ inclusions are transformed into CaO–MgO–Al₂O₃ system inclusions with a relatively lower melting point at S2 and S3 stages. CaO–MgO–Al₂O₃ system inclusions are liquid under the steel-making temperature and easily float upward to be removed through collision and coalescence. As can be seen in Fig. 7(c), Al₂O₃ inclusions can be hardly fully modified into CaO–Al₂O₃ inclusions with low melting point or CaO–MgO–Al₂O₃ system inclusions. The average compositions of inclusions during the RH refining process are given in Fig. 7(d). The contents of CaO and MgO increase from 2.8 and 4.8 mass% to 12.1 and 18.2 mass% respectively, while the content of Al₂O₃ decreases from 92.3 mass% to 70.0 mass%.

Fig. 7.

Ternary phase diagrams of CaO–MgO–Al₂O₃ inclusions from S1 to S3 stage. (a) S1, (b) S2, (c) S3. (Online version in color.)

3.2.2. Size Distribution of Inclusions

The inclusions with the sizes range of 2 μm < D ≤ 5 μm, 5 μm < D ≤ 10 μm, 10 μm < D ≤ 15 μm, and D > 15 μm are analyzed, respectively. The black line is calculated by FactSage 7.0 within which is the liquid region below 1873 K. The inclusions in this region are fully liquid, while those outside this region are non-liquid ones including fully solid or liquid + solid inclusions. The number density of inclusions per unit area is defined by Eq. (3).   

$$\text{ND}=\frac{n}{A_{\text{total}}}$$(3)

where ND is the number density of inclusions, A_total is the total area scanned (mm²), and n is the number of detected inclusions with size ≥ 2 μm on A_total of steel sample.

Figure 8(a) shows the change of ND during the RH refining process. It can be observed that ND decreases from 10.32 per mm² to 3.09 per mm², indicating that a lot of Al₂O₃ inclusions are generated with Al addition. Moreover, most inclusions are removed after RH vacuum treatment. Figure 8(b) shows that the size of inclusions varies mainly smaller than 10 μm. The proportion of inclusions with sizes smaller than 5 μm gradually decreases by 10% and that of inclusions with sizes range of 5–10 μm increases. Therefore, it can be concluded that these inclusions with sizes smaller than 10 μm are difficult to be removed from liquid steel through floatation.

Fig. 8.

Size distribution of inclusions from S1 to S3 stage.

3.3. Inclusion Formation by Thermodynamics Analysis

The standard states in this study are chosen for both the activities of oxides in the inclusions and slag and the elements in liquid steel, the former is taken relative to the pure solid taking the unit as molar fraction while the latter to a dilute solution taking the unit as mass percent. The activity of element i can be calculated based on Eqs. (4) and (5).   

$$log{f}_i=\sum e_i^j[\text{mass}\%j]$$(4)

  

$$a_i=f_i\cdot [\text{mass}\%i]$$(5)

In Eq. (4), $e_i^j$ is the interaction coefficient of element j to i and f_i is the activity coefficient of element i, while a_[i] is the activity of constituent [i] in liquid steel in Eq. (5).

Thermodynamic calculations are carried out to discuss the formation of inclusions in tinplate steel during the RH refining process. In the calculations, the interaction coefficients used in thermodynamic calculations are listed in Table 4. The activity coefficients of element i are listed in Table 5.

Table 4. Interaction coefficients used in thermodynamic calculations at 1873 K.^2,17)

j i	C	Si	Mn	S	P	Al	O
Ca	−0.34	−0.097	−0.0156	−336	−0.097	−0.072	−250020)
Mg	−0.1518)	−0.0919)	/	−1.38	/	−0.12	−430
Al	0.091	0.0056	0.035	0.030	0.033	0.045	−1.98
O	−0.436	−0.131	−0.021	−0.133	0.07	−1.1720)	−0.20

Table 5. Activity coefficients of element i.

fC	fCa	fMg	fAl	fO
0.973	0.0012	0.720	1.024	0.696

3.3.1. Formation of CaO·Al₂O₃ Inclusion

The dissolved oxygen content in liquid steel is an important factor in thermodynamic calculations. The dissolved oxygen content is either determined by vacuum decarburization or by aluminum deoxidation. The equilibrium constant K of this decarburization reaction is calculated according to Eqs. (6) and (7).   

$$[\mathrm{C}]+[\mathrm{O}]=\mathrm{C}\mathrm{O}(\mathrm{g})\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\mathrm{l}\mathrm{o}\mathrm{g}{K}_2=2.003+1\mathrm{\hspace{0.17em} }160/T$$(6)

  

$$K_{\text{2}}=\frac{P_{\text{CO}}}{a_{[\text{C}]}\cdot a_{[\text{O}]}}=\frac{P_{\text{CO}}}{f_{\text{C}}\cdot w_{[\text{C}]}\cdot f_{\text{O}}\cdot w_{[\text{O}]}}$$(7)

where P_CO is the partial pressure of CO in gas phase, Pa. a_[i] is the activity of constituent [i] in liquid steel. w_i is the mass fraction of component i. When P_CO is 133 Pa and w_[C] is 0.042%, the activity coefficients of elements C and O are 0.973 and 0.696 respectively, the content of dissolved oxygen is calculated to be 111 ppm at 1873 K. On the other hand, when Al is added to the liquid steel after recycling degassing for 3 min during the RH refining process, the Al–O deoxidization reaction occurs so fast that the content of dissolved oxygen would be very low. In the calculation, the [Al] content is 0.052% at S1 stage based on Table 3 and the activity coefficients of elements Al and O are 1.024 and 0.696 based on Table 5 respectively. In addition, the activity of Al₂O₃ is considered as unity. So the content of dissolved oxygen in liquid steel is calculated to be about 3.2 ppm according to Eqs. (8) and (9). It is found that the dissolved oxygen content of Al–O equilibrium is lower through the above thermodynamic calculations. Therefore, the dissolved oxygen content is considered as 3.2 ppm in the following calculations.

Dissolved [Ca] in liquid steel can react with [Al] and [O] to produce various calcium aluminates. For the formation of complex CaO–Al₂O₃ system inclusions, the related reactions involved are listed in Eqs. (10), (11), (12), (13), (14), (15), (16), (17).²¹⁾   

$$2[\mathrm{A}\mathrm{l}]+3[\mathrm{O}]=\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\mathrm{l}\mathrm{o}\mathrm{g}{K}_3=-20.57+64\mathrm{\hspace{0.17em} }000/T$$(8)

  

$$K_{\text{3}}=\frac{a_{{\text{Al}}_{\text{2}}{\text{O}}_{\text{3}}}}{a_{[\text{Al}]}^2\cdot a_{[\text{O}]}^3}$$(9)

  

$$[\mathrm{C}\mathrm{a}]+[\mathrm{O}]=\mathrm{C}\mathrm{a}\mathrm{O}(\mathrm{s})\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\mathrm{l}\mathrm{o}\mathrm{g}{K}_4=3.29+7\mathrm{\hspace{0.17em} }220/T$$(10)

  

$$K_{\text{4}}=\frac{a_{\text{CaO}}}{a_{[\mathrm{C}\mathrm{a}]}\cdot a_{[\mathrm{O}]}}$$(11)

  

$$\begin{array}{l}[\mathrm{C}\mathrm{a}]+4[\mathrm{A}\mathrm{l}]+7[\mathrm{O}]=\mathrm{C}\mathrm{a}\mathrm{O}\cdot 2\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{l}\mathrm{o}\mathrm{g}{K}_5=-46.643+160\mathrm{\hspace{0.17em} }138/T\end{array}$$(12)

  

$$K_{\text{5}}=\frac{a_{\text{CaO}\cdot {\text{2Al}}_{\text{2}}{\text{O}}_{\text{3}}}}{a_{[\text{Ca}]}\cdot a_{[\text{Al}]}^4\cdot a_{[\text{O}]}^7}$$(13)

  

$$\begin{array}{c}[\mathrm{C}\mathrm{a}]+2[\mathrm{A}\mathrm{l}]+4[\mathrm{O}]=\mathrm{C}\mathrm{a}\mathrm{O}\cdot \mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{l}\mathrm{o}\mathrm{g}{K}_6=-26.817+97\mathrm{\hspace{0.17em} }426/T\end{array}$$(14)

  

$$K_{\text{6}}=\frac{a_{\text{CaO}\cdot {\text{Al}}_{\text{2}}{\text{O}}_{\text{3}}}}{a_{[\text{Ca}]}\cdot a_{[\text{Al}]}^{\text{2}}\cdot a_{[\text{O}]}^{\text{4}}}$$(15)

  

$$\begin{array}{l}12[\mathrm{C}\mathrm{a}]+14[\mathrm{A}\mathrm{l}]+33[\mathrm{O}]=12\mathrm{C}\mathrm{a}\mathrm{O}\cdot 7\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{l}\mathrm{o}\mathrm{g}{K}_7=-223.4+849\mathrm{\hspace{0.17em} }615/T\end{array}$$(16)

  

$$K_{\text{7}}=\frac{a_{\text{12CaO}\cdot {\text{7Al}}_{\text{2}}{\text{O}}_{\text{3}}}}{a_{[\text{Ca}]}^{\text{12}}\cdot a_{[\text{Al}]}^{\text{1}4}\cdot a_{[\text{O}]}^{\text{33}}}$$(17)

In the above calculations of Eqs. (10), (11), (12), (13), (14), (15), (16), (17), the composition of the liquid steel sample is taken at S1 stage. The dissolved oxygen content is 3.2 ppm and the aluminum content is 0.052%. The phase stability diagram of calcium aluminate system is calculated, as shown in Fig. 9. Based on the measured contents of [Ca] and [Al] in the steel sample, it is also verified that the CaO·Al₂O₃ inclusions were formed at S2 and S3 refining stages. Thus, thermodynamic analysis agrees well with the industrial experimental results.

Fig. 9.

Phase stability diagram of CaO–Al₂O₃ system. (Online version in color.)

3.3.2. Formation of CaO–MgO–Al₂O₃ Ternary System Inclusion

In Fig. 7, the inclusions of the CaO–MgO–Al₂O₃ ternary system could be observed at S2 and S3 refining stages due to the reaction between [Ca] and MgO·Al₂O₃ inclusions, expressed by Eqs. (18) and (19), indicating that the MgO·Al₂O₃ inclusions are unstable and would change into CaO–MgO–Al₂O₃ ternary system inclusions. Finally, the CaO–MgO–Al₂O₃ ternary system inclusions tend to form calcium alumina inclusions. In this study, CaO·Al₂O₃ is chosen for the calculations of phase stability between MgO·Al₂O₃ and CaO·Al₂O₃.   

$$\begin{array}{l}[\mathrm{C}\mathrm{a}]+\mathrm{M}\mathrm{g}\mathrm{O}\cdot \mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})=\mathrm{C}\mathrm{a}\mathrm{O}\cdot \mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})+[\mathrm{M}\mathrm{g}]\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{l}\mathrm{o}\mathrm{g}{K}_8=-0.40+2\mathrm{\hspace{0.17em} }476/T\end{array}$$(18)

  

$$K_{\text{8}}=\frac{a_{\text{CaO}\cdot {\text{Al}}_{\text{2}}{\text{O}}_{\text{3}}}\cdot a_{[\text{Mg}]}}{a_{\text{MgO}\cdot {\text{Al}}_{\text{2}}{\text{O}}_{\text{3}}}\cdot a_{[\text{Ca}]}}$$(19)

The boundary between Al₂O₃ and CaO·Al₂O₃ is calculated by Eqs. (20) and (21).²²⁾   

$$\begin{array}{l}3[\mathrm{C}\mathrm{a}]+4\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})=3\mathrm{C}\mathrm{a}\mathrm{O}\cdot \mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})+2[\mathrm{A}\mathrm{l}]\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{0.5em}\mathrm{l}\mathrm{o}\mathrm{g}{K}_9=-33.17-39\mathrm{\hspace{0.17em} }532/T\end{array}$$(20)

  

$$K_{\text{9}}=\frac{a_{\text{3 CaO}\cdot {\text{Al}}_{\text{2}}{\text{O}}_{\text{3}}}\cdot a_{[\text{Al}]}^2}{a_{{\text{Al}}_{\text{2}}{\text{O}}_{\text{3}}}^4\cdot a_{[\text{Ca}]}^3}$$(21)

The boundary between Al₂O₃ and MgO·Al₂O₃ is calculated by Eqs. (22) and (23).²³⁾   

$$\begin{array}{l}3[\mathrm{M}\mathrm{g}]+4\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})=3\mathrm{M}\mathrm{g}\mathrm{O}\cdot \mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3(\mathrm{s})+2[\mathrm{A}\mathrm{l}]\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\mathrm{l}\mathrm{o}\mathrm{g}{K}_{10}=34.37-46\mathrm{\hspace{0.17em} }950/T\end{array}$$(22)

  

$$K_{\text{10}}=\frac{a_{\text{3 MgO}\cdot {\text{Al}}_{\text{2}}{\text{O}}_{\text{3}}}\cdot a_{[\text{Al}]}^2}{a_{{\text{Al}}_{\text{2}}{\text{O}}_{\text{3}}}^4\cdot a_{[\text{Mg}]}^3}$$(23)

In the above calculations, the activities of MgO·Al₂O₃ and CaO·Al₂O₃ in Eq. (18) and Al₂O₃ in Eqs. (20) and (22) are considered as unity, and the other thermodynamic parameters are calculated based on Table 5 and Eqs. (4) and (5). The phase stability diagram of MgO·Al₂O₃–CaO·Al₂O₃–Al₂O₃ is shown in Fig. 10. It can be clearly noted in Fig. 10 that the MgO·Al₂O₃ phase is stable after Al addition. When the [Ca] content at S3 stage reaches 24.0 ppm, CaO·Al₂O₃ inclusion would be generated. Meanwhile, it also implies that the calculation shown in Fig. 9 is reasonable.

Fig. 10.

Phase stability diagram of MgO·Al₂O₃–CaO·Al₂O₃–Al₂O₃.

In order to ensure the inclusion composition in the liquid region as possible, it is necessary to control a proper content of [Ca] in liquid steel. In the present study, if all the Al₂O₃ inclusions in the steel can be modified to 12CaO·7Al₂O₃ inclusions, the content of [Ca] in liquid steel should be maintained above 112 ppm based on Eqs. (14), (15), (16), (17). However, although the equilibrium [Ca] content in the liquid steel is calculated to be 398 ppm based on the composition of the initial RH refining slag shown in Table 2, the measured [Ca] content at the end of RH refining is merely 24.0 ppm, indicating that the slag-liquid steel reaction is far from equilibrium. That is to say, in the present industrial experiment, the RH refining slag with high Ca/Al ratio can meet the requirement of Al₂O₃ inclusion modification from the viewpoint of thermodynamics, however, the transport of calcium component from the refining slag to the liquid steel is possibly restrained by the kinetics condition, and thus the CaO·Al₂O₃ inclusions with high melting point mainly present in the liquid steel. In the following, the mass transfer of Ca from the refining slag to the liquid steel should be discussed and the modification limitation of the inclusions by refining slag would be evaluated.

3.4. Kinetic Limitation of Calcium Transfer from Refining Slag to Liquid Steel

Here, the kinetics model of slag/metal reaction established by Okuyama et al.²⁵⁾ is used to discuss the mass transfer of Ca between refining slag and liquid steel. As described in section 3.1 above, the increments of [Mn] and [P] contents in the liquid steel during S2 and S3 stages can be considered due to the addition of MnFe alloy, rather than through the reduction reactions of P₂O₅ and MnO in the refining slag by [Al] in liquid steel. So P₂O₅ and MnO are not considered in the establishment of the kinetic model. In this model, it is assumed that the concentrations of Al₂O₃, SiO₂, MgO and CaO in the slag side and of [Al], [Si], [Ca], [Mg] and [O] in the liquid steel show respective concentration gradients in the double film boundary layer at the slag-metal interface, and the slag-metal interface reactions have reached a thermodynamic equilibrium. If the rates of chemical reactions involving the respective components at the slag-metal interface are big enough, the mass transfer is considered to be the rate-determining step and the reaction rates can be expressed by Eqs. (24), (25), (26), (27), (28), (29), (30).^24,25,26)   

$$-\frac{d[\text{Al}]}{dt}=\left(\frac{A\cdot k_{\text{Al}}}{W_{\text{m}}}\right)\cdot \left\{[\text{Al}]-\frac{({\text{AlO}}_{1.5})}{B_{\text{Al}}\cdot a_{\text{O}}^{*1.5}}\right\}$$(24)

  

$$-\frac{d[\text{Si}]}{dt}=\left(\frac{A\cdot k_{\text{Si}}}{W_{\text{m}}}\right)\cdot \left\{[\text{Si}]-\frac{({\text{SiO}}_2)}{B_{\text{Si}}\cdot a_{\text{O}}^{*2}}\right\}$$(25)

  

$$-\frac{d[\text{Ca}]}{dt}=\left(\frac{A\cdot k_{\text{Ca}}}{W_{\text{m}}}\right)\cdot \left\{[\text{Ca}]-\frac{(\text{CaO})}{B_{\text{Ca}}\cdot a_{\text{O}}^{*}}\right\}$$(26)

  

$$-\frac{d[\text{Mg}]}{dt}=\left(\frac{A\cdot k_{\text{Mg}}}{W_{\text{m}}}\right)\cdot \left\{[\text{Mg}]-\frac{(\text{MgO})}{B_{\text{Mg}}\cdot a_{\text{O}}^{*}}\right\}$$(27)

  

$$-\frac{d[\text{O}]}{dt}=\left(\frac{A\cdot k_{\text{O}}}{W_{\text{m}}}\right)\cdot \left\{[\text{O}]-\frac{a_{\text{O}}^{*}}{f_{\text{O}}^{*}}\right\}$$(28)

  

$$B_i={\displaystyle \frac{K_i\cdot f_i\cdot M_{i{\text{O}}_n}\cdot \sum \left({\displaystyle \frac{(i{\text{O}}_n)}{M_{i{\text{O}}_n}}}\right)}{{\gamma }_{i{\text{O}}_n}}}$$(29)

  

$$\frac{1}{k_i}=\frac{1}{{\rho }_{\text{m}}\cdot k_{i\text{m}}}+\frac{M_{i{\text{O}}_n}}{B_i\cdot M_i\cdot {\rho }_{\text{s}}\cdot a_{\text{O}}^{*n}\cdot k_{i\text{s}}}$$(30)

Therefore, oxygen activity, $a_{\text{O}}^{*}$, at the slag-metal interface can be obtained from the above equations and the mass balance equation for oxygen at the interface, Eq. (31).   

$$\begin{array}{l}{\displaystyle \frac{1.5}{M_{\text{Al}}}}\cdot {\displaystyle \frac{d[\text{Al}]}{dt}}+{\displaystyle \frac{2}{M_{\text{Si}}}}\cdot {\displaystyle \frac{d[\text{Si}]}{dt}}+{\displaystyle \frac{1}{M_{\text{Ca}}}}\cdot {\displaystyle \frac{d[\text{Ca}]}{dt}}\\ +{\displaystyle \frac{1}{M_{\text{Mg}}}}\cdot {\displaystyle \frac{d[\text{Mg}]}{dt}}-{\displaystyle \frac{1}{M_{\text{O}}}}\cdot {\displaystyle \frac{d[\text{O}]}{dt}}=0\end{array}$$(31)

It is assumed that the mass transfer coefficient, k_im, in the metal side has an identical value of 2.0 × 10⁻⁴ m/s for all the components, including [Ca], [Mg] and [Al]. On the other hand, it is assumed that the mass transfer coefficient for SiO₂, in the slag side, k_is, has a different value from the other components. The mass transfer coefficient of the components other than SiO₂ in the slag side, k_s, is 1.0 × 10⁻⁵ m/s, and the mass transfer coefficient of SiO₂ in the slag side is 2.0 × 10⁻⁶ m/s.²⁷⁾

Figure 11 presents the measured [Ca] content in the liquid steel at three refining stages (S1, S2, and S3), in comparison with the results calculated by Eq. (26). The calculated results of [Ca] content in liquid steel show good agreement with the measured ones. Moreover, for the formation of 12CaO·7Al₂O₃ inclusion in the liquid steel, the required time to reach the [Ca] content of 112 ppm in the liquid steel is about 163 min, as shown in Fig. 11, which is far more than the refining time of 25 min in the present industrial experiment. Therefore, the inclusion modification by refining slag in the present work is limited.

Fig. 11.

Comparison between the measured [Ca] contents and the calculated ones in liquid steel. (Online version in color.)

In the present industrial experiment, the mass transfer resistances of Ca component in both the metal and slag sides are obtained based on Eq. (30), as shown in Table 6. The density of metal ρ_m is assumed to be 7000 kg/m³ and the density of slag ρ_s is 2700 kg/m³. $a_{\text{O}}^{*}$ is calculated as 8.41 × 10⁻² based on Eq. (31). It can be concluded that the content of [Ca] in liquid steel is controlled by the mass transfer of calcium in the boundary layer of liquid steel.

Table 6. Mass transfer resistance of Ca component in liquid steel and refining slag (m²·s/kg).

mass transfer resistance in liquid steel	mass transfer resistance of CaO in refining slag
${\displaystyle \frac{1}{{\rho }_{\text{m}}\cdot k_{i\text{m}}}}$	${\displaystyle \frac{M_{\text{CaO}}}{B_{\text{Ca}}\cdot M_{\text{Ca}}\cdot {\rho }_{\text{s}}\cdot a_{\text{O}}^{*}\cdot k_{\text{s,Ca}}}}$
0.71	0.0411

Therefore, from the viewpoint of thermodynamics, the RH refining slag with high Ca/Al ratio in the present industrial experiment can meet the requirement of Al₂O₃ inclusion modification. However, CaO·Al₂O₃ inclusions with high melting point are mainly present in the steel samples, indicating that the content of calcium in liquid steel supplied by the RH refining slag is not sufficient to modify Al₂O₃ inclusions to 12CaO·7Al₂O₃ inclusions with a lower melting point. The content of [Ca] in liquid steel is controlled by the mass transfer inside the boundary layer of liquid steel based on the kinetics model established by Okuyama et al.²⁴⁾ In order to generate 12CaO·7Al₂O₃ inclusions with lower melting point, appropriate calcium treatment is considered as an auxiliary measures to modify inclusions together with refining slag during the RH refining process.

4. Conclusions

The characteristics of inclusions including composition, morphology, number, and size in tinplate steel during the RH refining process are studied by industrial sampling analysis and thermodynamic calculation. Furthermore, the mass transfer of Ca between refining slag and liquid steel is discussed to evaluate the degree of inclusions modification by refining slag. The following conclusions are obtained.

(1) Two types of Al₂O₃ inclusions including cluster and single-particle are generated after Al addition in the RH refining process. With the slag-metal and refractory-metal reactions, Al₂O₃ inclusions, CaO·Al₂O₃ inclusions, MgO·Al₂O₃ spinel inclusions, and CaO–MgO–Al₂O₃ ternary system inclusions are found in the middle of RH refining. Only single-particle Al₂O₃, CaO·Al₂O₃ inclusions, and CaO–MgO–Al₂O₃ ternary system inclusions with high melting point are found at the end of RH refining.

(2) In the present industrial experiment, RH refining has a strong ability in removing inclusions, and T.O is decreased from 428 ppm to 37 ppm. From Al addition to the end of RH refining, the total number of inclusions shows a decreasing trend and the proportion of the inclusion number density is reduced by 70%. The sizes of about 62% inclusions are smaller than 10 μm at the end of RH refining.

(3) The mass transfer of Ca from the refining slag to the liquid steel has a significant effect on the content of [Ca] in liquid steel. Al₂O₃ inclusions generated in the liquid steel can only be modified to CaO·Al₂O₃ inclusions within the present refining time. Aiming to generate 12CaO·7Al₂O₃ inclusions quickly, moderate calcium treatment as a supplementary measure for refining slag is recommended to modify inclusions during the RH refining process.

Acknowledgments

This work was supported by the National Key R&D Program of China [Grant number 2016YFB0300602, 2017YFB0304201 and 2017YFB0304203] and National Natural Science Foundation of China [Grant numbers 51774072, 51774073 and 51974080].

Nomenclature

A: Area of reaction interface (m²)

a_i: Activity of component i (liquid steel component a_i = f_i × [%i], slag component a_i = γ_i × N_i)

f_i: Activity coefficient of component i in liquid steel

γ_i: Activity coefficient of component i in molten slag

N_i: Mole fraction of component i

K_i: Equilibrium constant of i + nO = iO_n

k_i: Total mass transfer coefficient of component i (g/m²·s)

k_m: Mass transfer coefficient of the metal side (g/m²·s)

k_s: Mass transfer coefficient of the slag side (g/m²·s)

M_i: Molecular weight of component i

W_m: Mass of liquid steel (kg)

ρ_m: Density of metal

ρ_s: Density of slag

References

• 1)   P.  Kaushik,  M.  Lowry,  H.  Yin and  H.  Pielet: Ironmaking Steelmaking, 39 (2012), 284.
• 2)   Z. Y.  Deng and  M. Y.  Zhu: ISIJ Int., 53 (2013), 450.
• 3)   S. J.  Li,  G. G.  Cheng,  Z. Q.  Miao,  W. X.  Dai,  L.  Chen and  Z. Q.  Liu: ISIJ Int., 58 (2018), 1781.
• 4)   J. H.  Park and  H.  Todoroki: ISIJ Int., 50 (2010), 1333.
• 5)   M.  Jiang,  X. H.  Wang,  B.  Chen and  W. J.  Wang: ISIJ Int., 48 (2008), 885.
• 6)   M.  Jiang,  X. H.  Wang and  W. J.  Wang: Steel Res. Int., 81 (2010), 759.
• 7)   M.  Jiang,  X. H.  Wang and  W. J.  Wang: Ironmaking Steelmaking, 39 (2012), 20.
• 8)   S. P.  He,  G. J.  Chen,  Y. T.  Guo,  B. Y.  Shen and  Q.  Wang: Metall. Mater. Trans. B, 46 (2015), 585.
• 9)   W. V.  Bielefeldt and  A. C. F.  Vilela: Steel Res. Int., 86 (2015), 375.
• 10)   Y.  Ren,  L. F.  Zhang,  W.  Fang,  S. J.  Shao,  J.  Yang and  W. D.  Mao: Metall. Mater. Trans. B, 47 (2016), 1024.
• 11)   P. C.  Yan,  S. G.  Huang,  L.  Pandelaers,  J.  Van Dyck,  M. X.  Guo and  B.  Blanpain: Metall. Mater. Trans. B, 44 (2013), 1105.
• 12)   K.  Sakata: ISIJ Int., 46 (2006), 1795.
• 13)   M.  Jiang,  X. H.  Wang,  B.  Chen and  W. J.  Wang: ISIJ Int., 50 (2010), 95.
• 14)   J. H.  Park,  S. B.  Lee and  H. R.  Gaye: Metall. Mater. Trans. B, 39 (2008), 853.
• 15)   Y. J.  Kang,  F.  Li,  K.  Morita and  D.  Sichen: Steel Res. Int., 77 (2006), 785.
• 16)   G. W.  Yang,  X. H.  Wang,  F. X.  Huang,  W. J.  Wang and  Y. Q.  Yin: Steel Res. Int., 85 (2014), 26.
• 17)   T.  Murai,  H.  Matsuno,  E.  Sakurai and  H.  Kawashima: Tetsu-to-Hagané, 84 (1998), 13 (in Japanese).
• 18)   H.  Suito and  R.  Inoue: ISIJ Int., 36 (1996), 528.
• 19)   G. K.  Sigworth and  J. F.  Elliott: Met. Sci., 8 (1974), 298.
• 20)   H.  Ohta and  H.  Suito: ISIJ Int., 36 (1996), 983.
• 21)   T. S.  Zhang,  Y.  Min,  C. J.  Liu and  M. F.  Jiang: ISIJ Int., 55 (2015), 1541.
• 22)   H.  Itoh,  M.  Hino and  S.  Ban-ya: Tetsu-to-Hagané, 83 (1997), 773 (in Japanese).
• 23)   D. W.  Zhao,  H. B.  Li,  C. L.  Bao and  J.  Yang: ISIJ Int., 55 (2015), 2115.
• 24)   G.  Okuyama,  K.  Yamaguchi,  S.  Takeuchi and  K.  Sorimachi: ISIJ Int., 40 (2000), 121.
• 25)   S.  Ohguchi and  D. G. C.  Robertson: Ironmaking Steelmaking, 11 (1984), 262.
• 26)   H.  Matsuno and  Y.  Kikuchi: CAMP-ISIJ, 7 (1994), 1126 (in Japanese).
• 27)   S.  Shinozaki,  K.  Mori and  Y.  Kawai: Tetsu-to-Hagané, 68 (1982), 72 (in Japanese).