Influence of Al on Evolution of the Inclusions in Ti-Bearing Steel with Ca Treatment

Abstract

Experimental simulations of steelmaking with different amounts of aluminum were achieved in the tube furnace at 1873 K and field scanning electron microscopy and energy dispersive X-ray spectroscopy (FE-SEM and EDX) were employed to explore the characteristics of the inclusions in Ti-bearing steel during the calcium treatment process. It was found that morphologies, chemical compositions, and the size distribution of the inclusions were obviously different before and after calcium treatment. The calcium addition need be carefully considered regarding the mass fraction of aluminum with the purpose of modifying the solid inclusions to liquid phases. The thermodynamic analysis of inclusion formation in the Al–Ti–Ca–O system at 1873 K was conducted, as well as transformation behaviors of inclusions including all types of solid inclusions and liquid phases during solidification. The thermodynamic equilibrium calculations are in good agreement with experimental data, which can be used to estimate inclusion formation in Ti-bearing steel.

1. Introduction

The quality of steel products can be effectively increased after being treated with the element titanium [1,2,3,4]. Titanium oxides and titanium carbides can be generated during the deoxidation process, which can promote the mechanical properties around the region of welded steels by improving the nucleation ratio of intragranular bainite [5,6,7,8,9]. However, TiO_x inclusions have a high probability of being agglomerate and subsequently form clusters, which results in the serious clogging issue of submerged entry nozzle and, consequently, lower the productivity of the continuous casting process as well [9,10,11]. Owing to the importance of Ti-containing inclusions, scholars have conducted a number of studies on inclusion control in Ti-bearing Al-killed steel, such as on thermodynamic computation and analysis [12,13,14,15,16,17,18,19,20,21,22], size distribution statistics [23,24,25,26], precipitation mechanisms [10,11,16,20], reaction kinetics, and evolution trajectory [14], mostly within the alumina and titanate system. In other words, it is usually difficult to keep the Al–Ti–O system inclusions as liquid phases during the production process at the present period.

It is generally the technique of calcium treatment that has been introduced to get liquid phases of calcium aluminate at casting temperature, which relieves the nozzle clogging issue during continuous process of steels, especially of grades of Al-killed steel, and also has benefits for the mechanical performance of final steel products [27,28,29,30,31,32,33,34,35,36,37]. However, the research involving the calcium treatment process of Ti-bearing Al-killed steel have been limited, until now [38,39,40,41,42]. In addition, the types of inclusions at the solidification temperature are directly related to the properties of the inclusions in the final product.

In this work, the laboratory-scale melts, with different mass fractions of calcium and aluminum, were prepared in order to clarify the influence of aluminum element on inclusion properties during the calcium treatment process of Ti-bearing Al-killed steels. Then, FE-SEM and EDX were employed to observe and analyze morphologies, chemical compositions, and number and size of inclusions in steel cylinder sampled from high-temperature melts before and after calcium treatment. In addition, equilibrium phases of inclusions at smelting temperature were calculated, and the transformation behaviors of inclusions during the process of solidification were estimated by the commercial software FactSage. Present works will lay the experimental and thermodynamic foundation on expanding the combined treatment of titanium and calcium to industrial-scale production, and suggest an alternative way to eliminate the nozzle clogging issue during the continuous casting process of Ti-bearing Al-killed steel.

2. Experimental Methods

Three sets of experiments were conducted in the furnace according to the schematic diagram shown in Figure 1. A 350 gram plate of iron was loaded in the Al₂O₃ crucible surrounded by the graphite crucible, and then the graphite crucible was placed into the furnace which was heated to 1873 K in the protective atmosphere of high-purity argon gas with a constant flow rate. Thereafter, the furnace temperature was maintained for 0.5 h after the raw materials were completely melted at 1873 K to reach the full homogenization of solutes. At the time node of before and after deoxidant (Al and Ti) addition, the activity of dissolve oxygen was determined by the oxygen probe with a resolution of ±10⁻⁶. The cylinder steel sample was extracted by the quartz tube with the inner diameter of 3 millimeters, and then immediately quenched into water at ambient temperature. At last, Ca treatment was carried out by adding Ca–Fe alloy. Another steel sample was obtained in the above way just 10 min after Ca treatment. The experimental details are shown in Figure 2, and the compositions of the raw materials in the work involving ARMCO iron, Ca–Fe alloy, and Al wire are listed in

Table 1. The compositions of experimental materials (mass %).

Type	Fe	C	Ti	Mn	Si	Ca	Al	P	S	Others
Ca–Fe alloy	69.82	-	-	-	-	30.10	-	-	-	0.08
Ti–Fe alloy	26.513	0.130	69.594	0.241	0.030	-	3.192	0.025	0.011	0.284
Al wire	-	-	-	-	-	-	99.99	-	-	0.01
Pure iron	99.944	0.002	-	0.03	0.01	-	0.001	0.007	0.007	0.043

.

The obtained steel samples were machined to two cylinders. One was adopted to analyze the chemical compositions by inductively coupled plasma-optical emission spectroscopy (ICP-OES) (Thermo Fisher Scientific ICAP6300, Waltham, MA, USA) with the resolution of ±5 × 10⁻⁷. The total oxygen levels (T.[O]) of the steel samples were measured by inert gas fusion-infrared absorptiometry with the resolution of ±10⁻⁶. The chemical compositions of all steel samples are given in

Table 2. The compositions of the obtained sample (mass %).

No.	[Al]	[Ca]	[Ti]	a[O]	T.[O]
1-A	0.0042	-	0.0105	0.0005	0.0053
1-B	0.0036	0.0005	0.0094	-	0.0049
2-A	0.0053	-	0.0118	0.0006	0.0045
2-B	0.0051	0.0038	0.0115	-	0.0038
3-A	0.0430	-	0.0112	0.0002	0.0038
3-B	0.0413	0.0025	0.0106	-	0.0032

. Another was polished for monitoring the inclusion properties, such as morphologies, size, and chemical compositions by scanning electron microscopy and energy dispersive spectroscopy (Zeiss Ultra-Plus, ZEISS, Jena, Germany). The size distribution of precipitates was surveyed in the observed regions of 19.86 mm², magnified 500× in the cross-section.

3. Results and Discussion

3.1. Chemical Composition, Size Distribution, and Morphologies of Inclusions

Titanium in the steel production process shows many valence states, such as Ti²⁺, Ti³⁺, and Ti⁴⁺ [43,44,45]. In addition, the types of titanium oxides are determined by combining the partial pressure of oxygen with mass fraction of titanium in steel [46]. Although a number of Ti-containing oxides (TiO, TiO₂, Ti₃O₅, Ti₂O₃, etc.) can exist as products of deoxidation reactions by titanium, when mass % Ti was between 0.0004 to 0.36, Ti₃O₅ was the only stable equilibrium oxide in steel, as demonstrated by the electron backscatter diffraction technique [43,44,45,46]. Based on the compositions of steel in the current research, the isothermal section of the Al₂O₃–Ti₃O₅–CaO ternary phase diagram at 1873 K with p(O₂) = 10⁻¹⁴ atm (computed by FactSage 7.0, THERMFACT LTD, Quebec, Canada), as shown in Figure 3, was introduced to analyze the composition of inclusions. It can be easily seen from Figure 3 that solid phases (CaO, Ti₃O₅, Al₂O₃, perovskite, calcium aluminates, and titanium aluminates) coexist with liquid phases in the two- or three-phase zones. It is noticeable that two liquid phases emerge at 1873 K, which are locate in the regions with a small amount of CaO and Ti₃O₅.

The chemical elements, as well as mole ratios in each observed precipitate, were determined by SEM-EDS (Zeiss Ultra-Plus, ZEISS, Jena, Germany), and then the data were converted to mass fraction of CaO, Ti₃O₅, and Al₂O₃. As presented in Figure 4, each plot in the phase diagram represents an individual inclusion to assess inclusion behavior and transformation, and the thick red lines are liquiduses at 1873 K. As shown in Figure 4a, the inclusions in No.1 melt after Ti addition are mainly spherical titanium aluminates. There is no significant change in the morphologies of the inclusions after calcium treatment. Only a small amount of compositional changes occur in these inclusions which are still in liquid phases. By contrast, a mass of irregular inclusions of Al–Ti–Ca–O system are generated in No. 2 melt as result of the increased mass fraction of 40 ppm [Ca] by calcium treatment. As the mass fraction of CaO in these inclusions is significantly increased, the locations of inclusion compositions in the isothermal section diagram are beyond the liquid region. From the view of the phase types, they mostly locate in perovskite and (CaO)₃·(TiO_x)₂. When the mass fraction of [Al] was increased from 40 ppm to 400 ppm, almost no liquid inclusions are found in No. 3 melt. Compared with No. 1 and No. 2 melts, the locations of inclusion compositions in No. 3 melt are outside of the liquidus in the isothermal section before calcium treatment, and bring on the corresponding irregular appearance. However, inclusions are evolved to calcium aluminates of liquid state after calcium treatment.

Figure 5 gives the size distributions of inclusions in the melts. The results indicate that more than half of the inclusions in sample 1-A are larger than 5 μm, and about 20% of the inclusions are smaller than 2 μm. There is no significant change in the size of inclusions after calcium treatment. There is just a little increase in the proportions of the larger inclusions (>10 μm) in sample 1-B. This may be due to collision and aggregation of the inclusions. The situation of No. 2 melt before calcium treatment is nearly identical to the former one. Nevertheless, the number of inclusions that are smaller than 5 μm in sample 2-B is more than 65%. This illustrates that the solid inclusions of Al₂O₃–CaO–TiO_x system trend to be fine and well-dispersed in the melts of Ti-bearing steel as treated by calcium. Apparently, the inclusions in sample 3-A also have a relatively small size. After calcium treatment, the proportion of the smaller inclusion in sample 3-B is obviously decreased and the numbers of inclusions are present as liquid state, combined with the results in Figure 4.

3.2. Thermodynamic Analysis of Al–Ti–Ca–O System

A series of generating reactions of inclusions, as listed in Table 2, were considered to understand the transformation process of Al–Ti–Ca–O system. The intermediate products, titanium oxides, CaO and calcium titanates in

Table 3. Standard Gibbs energies of inclusions formation.

No.	Reactions	ΔGθ = A + B × T /J·mol−1		References
		A	B
1	Al2O3 = 2[Al] + 3[O]	867,500	−222.5	[27]
2	CaO = [Ca] + [O]	138,227	63.0	[47]
3	Al2TiO5 = [Ti] + 2[Al] + 5[O]	1,435,000	−40.5	[23]
4	CaO·6Al2O3 = CaO(s) + 6Al2O3(s)	16,380	37.58	[48]
5	CaO·2Al2O3 = CaO(s) + 2Al2O3(s)	15,650	25.82	[48]
6	CaO·Al2O3 = CaO(s) + Al2O3(s)	17,910	17.38	[48]
7	12CaO·7Al2O3 = 12CaO(s) + 7Al2O3(s)	−618,000	612.1	[49]
8	Ti2O3 =2[Ti] + 3[O]	822,000	−247.7	[50]
9	Ti3O5 =3[Ti] + 5[O]	1,307,000	−381.8	[50]
10	TiO2 = [Ti] + 2[O]	−675,600	234	[51]
11	3CaO·Ti2O3 = 3CaO(s) + Ti2O3(s)	192,745 (1873K)		[52]
12	3CaO·2TiO2 = 3CaO(s) + 2TiO2(s)	148,365	24.14	[52]
13	CaO·TiO2 = CaO(s) + TiO2(s)	74,392	10.13	[53]

, could be firstly generated according to the very negative Gibbs energies of reactions between titanium/calcium and oxygen at 1873 K. As 12CaO·7Al₂O₃ is the only compound located inner of the liquidus in the phase diagrams of Al₂O₃–TiO_x and Al₂O₃–CaO–TiO_x systems at 1873 K, it is introduced as liquid calcium aluminate to calculate the equilibrium state with solid inclusions.

The phase equilibrium calculations are based on the minimum ΔG theory, since the elements (such as Ca, Ti, Al, etc.) involved in the present research are of low concentration in molten steel. Therefore, the molten steel can be assumed to be an ideal solution and to follow Henry’s law. The component activity coefficient was calculated by using 1% (mass) extremely dilute solution as the standard state. The activity coefficient between the contents of all elements and the interaction coefficient (shown in

Table 4. The interaction coefficients of Fe–Al–Ti–Ca–O system molten steel at 1873 K [54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69].

i	j	k	${\mathit{e}}_{\mathit{i}}^{\mathit{j}}$	${\mathit{r}}_{\mathit{i}}^{\mathit{j}}$	${\mathit{r}}_{\mathit{i}}^{(\mathit{j},\mathit{K})}$	i	j	k	${\mathit{e}}_{\mathit{i}}^{\mathit{j}}$	${\mathit{r}}_{\mathit{i}}^{\mathit{j}}$	${\mathit{r}}_{\mathit{i}}^{(\mathit{j},\mathit{K})}$
O	Al	Ca	−3.9	−0.01	0	Al	Al	O	0.043	−0.001	−0.028
		Ti			-		Ca	O	−0.047	0	0
		O			47.45		Ti	-	0.004	-	-
	Ca	Al	−310	−17,984	0		O	Al	−1.98	39.82	−0.028
		O			519,903			Ca			0
	Ti	O	−0.34	0.031	0.026	Ca	Al	-	−0.072	0.0007	-
	O	Al	−0.20	0	47.45		Ca	O	−0.002	-	−90,227
		Ca			520,000		Ti	-	−0.13	-	-
Ti	Al	-	0.0037	-	-		O	Ca	−580	650,129	−90,056
	Ti	O	0.042	−0.001	0.20	-
	O	Ti	−3.4	−0.0355	0.20

) was expressed by Wagner’s formula, as shown in Formula (1).

$$\log f_i={\displaystyle \sum }_{j=2}^n{e}_i^j\left[\%j\right]+{\displaystyle \sum }_{j=2}^n{r}_i^j\left[\%j\right]+{\displaystyle \sum }_{j=2}^n{\displaystyle \sum }_{k=2}^n{r}_i^{j,k}\left[\%j\right][\%k]$$

(1)

According to the activity of all elements and the free energy change for stable oxides’ region transformation, the phase diagrams can be calculated, and the details of the calculation methods are mentioned in other work and my former work [70,71]. The phase diagrams involving Al–Ti–Ca–O system at 1873 K were worked out to estimate potential oxides in molten steel, and are shown in Figure 6. It can be seen from Figure 6a that CaO and CT (3CaO·2TiO₂, 3CaO·Ti₂O₃, CaO·TiO₂) are the main inclusions in the molten steel, as the mass fraction of calcium and titanium increase from 10⁻⁵% to 10⁻²%, and 10⁻⁴% to 1%, respectively. When 0.005% aluminum is added into the Ca–Ti–O system as shown in Figure 6b, LCA (liquid calcium aluminate) precipitates, which squeezes the region of CT. It is obvious that aluminum content in steel has an important influence on stable phases of inclusions. As the mass fraction of aluminum increases to 0.05%, SCA (solid calcium aluminate) is present, which results in the further reduction of CT region, as shown in Figure 6c. It is noticeable that the stability region of liquid calcium aluminate is located in a common calcium content range. As the compositions of the steel samples in the current work (based on the content of the calcium and titanium) and experimental results introduced from Seo and Zheng [38,39] were marked in the phase diagrams, a good consistency in the above data is conspicuous, as shown in Figure 6b,c. Consequently, the calculated phase diagrams of Al–Ca–Ti–O system are reliable for effectively estimating the evolution behavior of inclusions at steelmaking temperature.

Some transformation behaviors of precipitates are overlooked in this work, due to the extreme cooling speed during the sampling process. For this reason, the phase transformation of inclusions during solidification of melts was computed by FactSage 7.0 as the FSstel, FactPS, and FToxid databases were employed, and the relevant results are present in Figure 7. The mass fraction of aluminum and calcium varies in the Fe–Al–Ca–0.01Ti–0.005O systems when temperature decreases from 1873 K to 1473 K. It can be seen from Figure 7a that liquid inclusions are present in the steel at a wide temperature range, from 1873 K to about 1623 K, when the mass fractions of aluminum and calcium are both small, around 0.005%, and the transformation process follows liquid inclusions → Al₂O₃ → 2CaO·Ti₂O₃ → Ti₂O₃ during solidification. Nevertheless, the liquid inclusions only exist around 1873 K as the mass fraction of calcium increases to 0.003% and the formation of solid calcium titanates is favorable, as shown in Figure 7b. As the mass fraction of aluminum increases to 0.03%, while that of calcium is 0.005%, only alumina and calcium aluminates precipitate at 1873 K and no other inclusions are formed during the cooling process, as shown in Figure 7c. However, the liquid phase appears again at a higher temperature as the mass fraction of calcium increases 0.003% in Figure 7d. The liquid inclusions are modified into solid calcium aluminates during the solidification process. When the mass fraction of calcium increases further to 0.007%, the main stable phases become CaO and solid calcium titanates in steel, replacing the calcium aluminates as shown in Figure 7e.

From the found inclusions of Al–Ti–Ca–O system and the above thermodynamic analysis, it is suggested that the calcium treatment technique associated with the right aluminum addition is available to get the liquid phase inclusions at the casting temperature. The calcium addition needs to be reconsidered as the amount of aluminum varies in the steelmaking process.

4. Conclusions

The characteristics and transformation behaviors of the inclusions in Ti-containing steel after calcium addition with different aluminum amount have been discussed by physical simulations and thermodynamic analysis at 1873 K (1600 °C), as well as during solidification. The main results are summarized as follows.

The morphologies, chemical compositions, and size distribution of the inclusions are dramatically different before and after calcium treatment, and the calcium addition should be reconsidered according to the mass fraction of aluminum in order to get liquid phase inclusions. The generation of liquid inclusions is more favorable as less calcium addition is needed at the lower amount of aluminum, and as more calcium is appropriate for a higher amount of aluminum. That is, 0.0005% calcium for 0.0036% aluminum, and 0.0025% calcium for 0.0413% aluminum in this study. The inappropriate calcium treatment level can induce the generating trend of solid inclusions in melts. The inclusion-oriented diagrams of Al–Ti–Ca–O system in melts at 1873 K, and the transformation behaviors of inclusions during solidification of steel, were systemically computed involving all types of solid inclusions and liquid phases. The thermodynamic equilibrium calculations are in good agreement with experimental data, and the liquid inclusions can exist during the whole cooling process, as formed at steelmaking temperature.