Effect of Al Content on the Characteristics of Inclusions in Al–Ti Complex Deoxidized Steel with Calcium Treatment

Abstract

To improve the effect of calcium treatment and the cleanliness of steel and to make use of fine TiO_x to refine the microstructure of steel, the effect of aluminum content on inclusion characteristics of aluminum-titanium complex deoxidized and calcium treated steel is investigated based on the experiment with SEM/EDS, Image Pro-Plus 6.0 and FactSage 6.1 softwares and thermodynamic calculation in the present work. The results show that the inclusions in two steels with different aluminum content are obviously spherical composite inclusions with a two-layer complex structure, consisting of an Al₂O₃–CaO–TiO_X core surrounded by MnS. In low aluminum steel, the oxide core of inclusions contains much TiO_x and CaO, and their composite structure is mosaic compared with bundle in high aluminum steel, the number of inclusions is 2.5 times more than that in high aluminum steel, the thickness of MnS on oxides surface is also thinner. In addition, melting point of the inclusions in low aluminum steel is lower, the cleanliness of the steel is relatively improved because of the inclusions floating up, and the deformation aspect ratio of calcium aluminate inclusions with a certain amount of Ti₂O₃ is effectively improved, which is about 1–2 while the composition of oxide core is x_Al2O3 = 35–55%, x_Ti2O3 = 15–35%, and x_CaO = 10–25%. As a result, less calcium is needed to modify the alumina inclusions to liquid calcium aluminate in the case of lower aluminum deoxidized steel, thus the calcium treatment effect can be improved. The low aluminum in steel is more effective to control the inclusion characteristics to reduce the harm of MnS and to improve the cleanliness of steel.

1. Introduction

Inclusions are unavoidable in the steel and the harmful effect of inclusions on the properties of steel has been widely discussed.^1,2,3,4,5) Inclusions adversely affect the toughness, fatigability and ductility of steel because of their high melting points, high hardness, and brittleness.¹⁾ Oxides inclusions in steel serve as effective sites for the initiation of hydrogen induced cracking (HIC).^2,3) Larger and elongated inclusions such as manganese sulfide (MnS) have a remarkable influence on ductile fractures.^4,5) With the increasing demand of the high performance steels, reducing the size of the inclusions and controlling their characteristics are considered to be important.^3,6,7,8)

Pipeline steels are required to have better properties such as high strength, good toughness, excellent HIC resistance for an inevitable harsh environment in the long distance transportation of oil or natural gas.^2,9,10) It is very important to control inclusions such as oxides and sulfides and obtain acicular ferrite (AF) in the steels for their better properties.^3,11) Therefore, intensive deep desulfurization and deep deoxidation technology are adopted to reduce the generation of harmful inclusions in pipeline steel.^3,4) Accordingly, deep deoxidation by aluminum (Al) in pipeline steel is the thermodynamical prerequisites of deep desulfurization because of its high efficiency and relatively low cost. However, it is inevitable that Al₂O₃ particles are generated in the deoxidation process, which tend to form clusters and remain in the molten steel as inclusions. The inclusions are harmful not only for the mechanical properties of steel product, but also for the manufacturing process such as causing the nozzle clogging problems in continuous casting.^6,7) Calcium (Ca) treatment is a well-established way to modify solid alumina inclusions to liquid or partially liquid calcium aluminates which can be easily removed from the molten steel by coagulation and floating up with the rising argon bubbles.^12,13,14) Meanwhile, the deleterious effect of sulfide inclusions on fracture ductility can be offset to some extent by calcium treatment due to the less deforming of sulfides during hot working.⁴⁾ However, calcium sulfide (CaS) inclusions are solid at molten steel casting temperatures which would be detrimental to castability if they remained in the molten steel after calcium treatment.¹³⁾ Thus, the mass ratios of Ca/S, Ca/O, Ca/Al_S in the molten steel must be controlled accurately in calcium treated steel to obtain the inclusions expected.^4,15,16) But, this conventional process of the deep deoxidation and deep desulfurization process, employed in pipeline steels manufacture, have some limitations and are not economical to some extent, because this conventional process is to generate some inclusions in the molten steel and then modifies or removes them.

Different from the conventional way of utmost removing and strict controlling inclusions, positive utilization of the fine oxide inclusion particles to act as nuclei for ferrite and thus control the grain size of steel to improve the properties of steel, which is called oxide metallurgy, has attracted great increasing attention of the metallurgists all over the world in recent years. Amongst the non-metallic inclusions, Ti-containing inclusions are well known to provide the suitable intragranular nucleation sites for AF, and titanium alloyed steel is developing quickly because of the relative low price of titanium (Ti).^17,18) Fine dispersed Ti-containing inclusions precipitated with MnS were reported to be most effective for intragranular ferrite (IGF) formation, and an AF dominant microstructure could be produced.^17,18,19,20) This intragranular structure has been found to provide a desirable combination of high strength and good toughness because of its small grain size and interlocking microstructure both in weld metals and in the heat-affected zone, meanwhile, the initiation and propagation of cracks can be effectively prevented.^20,21,22) The previous study showed that the best combination of microstructure, impact properties and HIC resistance were obtained in the range of 0.02–0.05 mass% titanium containing steel.²¹⁾ Therefore, Ti deoxidation and alloying has been studied in-depth to improve the properties of the steel.

Extensive studies on C–Mn steels and weld metals have revealed that the nucleation of the fine IGF increases with uniform dispersed fine Ti-containing inclusions and it is very important to control the composition, number, size distribution of fine Ti-containing inclusions.^17,18,19) The influence of the addition order of Al/Ti on the morphology and composition of oxide inclusions was deeply investigated by M. K. Sun.²³⁾ The effect of Ti/Al ratio on the transient behavior of the inclusion composition, shape, and structure in Fe–Al–Ti–O melts was studied by C. Wang.²⁴⁾ A lot of researches on the reaction steps, transient behavior and transformation of inclusions in Al-killed calcium treated steel had been widely conducted.^12,13,14,25) After all, these studies mainly focused on Al-killed Ti-alloyed steel in which acid soluable aluminium (Al_s) concentration is beyond 0.01 mass% and the modifying effect of calcium treatment on Al₂O₃ inclusions. The inclusions characteristic in Al–Ti complex deoxidation calcium treatment steel with low aluminum content is not clear. And the appropriate calcium content, sulfur content limitation and cleanliness of pipeline steel, is also need to be clarified.

In this study, the morphology, composition, number, size distribution, melting point and deformability of inclusions in Al–Ti complex deoxidized, calcium treated steel with different aluminum content were systematically investigated based on the experiment and thermodynamic calculations. This study would be helpful to improve the effect of calcium treatment and the cleanliness of steel and to make use of fine TiO_x to refine the microstructure of steel.

2. Experimental

Raw materials in the present work were pure iron (<0.005 mass% C, 0.03 mass% Si, 0.06 mass% Mn, 0.015 mass% P, 0.005 mass% S, <0.007 mass% Al_s, 0.0156 mass% T. [O], 0.0015 mass% T. [N]), carbon powder (99.95 mass% purity), nickel powder (AR), manganese (>99.7 mass% purity), ferrosilicon (0.15 mass% C, 72.56 mass% Si, 0.43 mass% Mn, 0.017 mass% S, 26.84 mass% Fe), aluminum wire (>99 mass% purity), titanium alloy (40 mass% Ti, 60 mass% Fe), calcium Silicon (1.2 mass% C, 53 mass% Si, 2.4 mass% Al_s, 0.07 mass% S, 0.05 mass% P), micro-carbon ferrochrome (60.13 mass% Cr, 1.96 mass% Si, 0.23 mass% C, 0.045 mass% P, 0.026 mass% S, and balance of Fe) and ferroniobium (65 mass% Nb, 1.5 mass% Si, 0.15 mass% C, 0.2 mass% P, <0.1 mass% S, 0.1 mass% Ta, 1.2 mass% Al, and balance of Fe).

Experiments were conducted in a 25 kg medium frequency induction furnace (Model: ZGJL0.025-100-2.5P, made by Jinzhou Dianlu Co. Ltd.) with an MgO crucible. The weight of an ingot was 15 kg, and the concentration of each element in the steel was controlled according to the yield and composition of X80 pipeline steel. Two experiments with different aluminum content (0.007 mass%, 0.03 mass%) by adding Al or not were carried out at 1600°C. The furnace chamber was evacuated to ultimate vacuum firstly, and then argon was flown (99.999% purity) to keep the inert working atmosphere of the chamber. Carbon, nickel powders, pure iron, micro-carbon ferrochrome and ferroniobium were directly put in the MgO crucible and melted. Other materials were put in the melt in the sequence of carbon powder, nickel powder, ferrosilicon, aluminum wire, titanium alloy, calcium silicon, and manganese. In order to ensure enough Mn-alloying effect and avoid the effect of the volatile manganese on observation from the peephole during the melting process, manganese was added into molten steel at the end of alloying. The experimental procedure diagram is shown in Fig. 1. The cast steels were cooled down to 1200°C in furnace chamber by the rate of 5°C/min in 70 kPa argon atmosphere, and then were cooled to room temperature in the air.

Fig. 1.

Experimental procedure diagram.

The cast steels were forged into cuboid steel blocks (cross section: 60 mm × 60 mm) at 1200°C. After being reheated in a temperature range of 1200–1250°C, the steel blocks with a thickness of 60 mm were hot-rolled into sheets of 10 mm in thickness and 100 mm in width by rolling mill (model: 800) with a finish rolling temperature of 850–860°C.

Total oxygen (T.[O]) and total nitrogen (T.[N]) content in the cast steel were determined by Oxygen and Nitrogen Analyzer (model: LECO-TC500). Al_s, Ti, Mn, Si and Ca content were analyzed by ICP-AES using a Single-channel Scanning Spectrometer (model: DGS-III). Carbon and sulfur content were determined by Carbon and Sulfur Analyzer (model: CS-8800).

The as-cast and hot-rolled steels samples used for analysis were prepared by cutting, grinding and polishing. The position of hot-rolled samples are shown in Fig. 2. The morphology and composition of the inclusions in the samples were analyzed by Field Emission Scanning Electron Microscope (FE-SEM, Model: Nova 400 Nano) at 20.0 kV with Energy Dispersive Spectrometer (EDS, Model: Le350 PentaFETx-3). The number and size distribution of inclusions were counted by Image Pro-Plus 6.0 software with photos from 30 randomly visual fields (area: 300 μm × 250 μm) of the SEM at 1000 magnification.

Fig. 2.

Schematic representation of the hot-rolled samples.

3. Results

3.1. Characteristics of the Inclusions in the As-cast Steel

The chemical composition of the as-cast steel samples is shown in Table 1. The morphology and composition of the typical inclusions in sample 1 (low aluminum content) and 2 (high aluminum content) are shown in Fig. 3. The SEM/EDS results indicate that the inclusions in both steels are obviously spherical composite with a two-layer complex structure, consisting of an Al₂O₃–CaO–TiO_x core surrounded by MnS. The main difference of the oxide core between the two steels is that the oxide core in sample 2 contains a little TiO_x and CaO, and the composite structure of the oxide core in sample 1 is mosaic compared with bundle in sample 2. The element distribution of the mosaic inclusion in sample 1 is presented in Fig. 4. It can be observed clearly that Al₂O₃ and TiO_x precipitate alternately with each other, and then CaO locally precipitated in the places which contain a little TiO_x. The average composition of TiO_x are close to the stoichiometry of Ti₂O₃ as depicted in Figs. 3(a)–3(d) (the numbers beside the inclusion in the figures are atomic percentages, similarly hereinafter) which are consistent with the reported results of similar steels.²⁶⁾ No single MnS is found in the above observed in samples.

Table 1. Chemical composition of experimental steel samples (mass%).

Sample	C	Si	Mn	S	T.[O]	T.[N]	Als	Ti	Ca
1	0.03	0.21	1.89	0.0048	0.0072	0.002	0.0055	0.018	0.0010
2	0.03	0.25	1.87	0.0032	0.0034	0.002	0.026	0.022	0.0011

Fig. 3.

Morphology and composition of typical inclusions in as-cast samples.

Fig. 4.

Morphology and EDS mapping of the typical Al₂O₃–CaO–TiO_x inclusions in cast sample 1.

Figure 5 shows the number and size distribution of the inclusions in the as-cast samples. The number density of the inclusions in sample 1 is 1156/mm² and the average size is 1.00 μm, while the number density of the inclusions in sample 2 is 467/mm² and the average size is 1.15 μm. The percentage of the inclusions smaller than 2 μm was beyond 96%, and no inclusions larger than 8 μm were found during the observation in both samples.

Fig. 5.

Size distribution of inclusions in as-cast steels.

The results indicate that the inclusions number, Ti₂O₃ and CaO content and the composite structure of the composite oxides are different.

3.2. Deformation Behavior of the Inclusions in the Hot-rolled Steel

Figure 6 shows the morphology and composition of the inclusions in the transverse section of both hot-rolled samples. The results illustrate that the inclusions morphology in the transverse section in the samples is almost spherical. Figure 7 shows the inclusion morphology and composition in the longitudinal section of both hot-rolled samples. The deformation of the inclusions in sample 1 is small and the oxide core deforms little as shown in Figs. 7(a)–7(d). The deformation of the inclusions is mainly determined by the deformation of MnS on the surface. As the thickness of the MnS on surface is small, the inclusion deforms little. On the other hand, the inclusion deforms large while the thickness of the MnS on surface is large, as shown in Figs. 7(b) and 7(d). The deformation of the inclusions in sample 2 is larger than sample 1 as shown in Figs. 7(e)–7(h).

Fig. 6.

Morphology and chemistry of typical inclusions in transverse section of hot-rolled samples.

Fig. 7.

Morphology and composition of the typical inclusions in longitudinal section of the hot-rolled samples.

In summary, the inclusions deformation is determined jointly by the deformation of the oxide core and MnS. On one hand, while the oxide core is composed mainly by Al₂O₃ with a little Ti₂O₃ and CaO (mole fraction ratio x_Al2O3/x_CaO >24), the oxide core is almost spherical and deforms little as shown in Figs. 7(e) and 7(f). While the oxide core is composed mainly of Al₂O₃ and CaO (x_Al2O3/x_CaO = 2.5–3.2), the oxide core deforms large and the morphologies are irregular as shown in Figs. 7(g) and 7(h). In addition of above mentioned results, the deformation of inclusions in hot-rolled samples was observed in wider range of x_Al2O3/x_CaO, the results indicate that the deformation of the oxide core and the whole inclusions decrease with the increase of Ti₂O₃ content while x_Al2O3/x_CaO = 1.0–3.2. On the other hand, the inclusions deformation is large while the thickness of the surface MnS is large.

As the deformation of inclusions is counted by aspect ratio (the ratio of length to width of an inclusion) as an indicator, the relationship between the core oxides composition and the deformation aspect ratio of whole inclusions in longitudinal section of hot-rolled steels is shown in Fig. 8, which shows that the deformation aspect ratio of oxides decreases with the increase of Ti₂O₃ content while x_Ti2O3 = 0–50%, increases firstly and decreases later with the increase of Al₂O₃ content while x_Al2O3 = 40–90%, increases with the increase of CaO content while x_CaO = 10–40%. From Fig. 8, one can also find the composition range of the oxide inclusion cores in which the inclusions deformation aspect ratios are in expected value. For example, the inclusions deformation aspect ratio is about 1–2 while the composition of oxide core is x_Al2O3 = 35–55%, x_Ti2O3 = 15–35%, x_CaO = 10–25%.

Fig. 8.

The effect of inclusion composition in the hot-rolled samples with different aluminum content on the deformation aspect ratio of the inclusions.

Figure 9 shows deformation aspect ratio of the inclusions in the transverse and longitudinal section of the hot-rolled samples with different aluminum content. The results indicate that the aspect ratio of the inclusions observed in transverse section is about 1–2 in both sample 1 and 2, while there is a large scatter for the aspect ratio of the inclusions in the longitudinal section in different hot-rolled samples which is 1–2 in sample 1 and 1–7 in sample 2. The inclusions deformation aspect ratio in sample 1 is significantly smaller than that in sample 2.

Fig. 9.

The deformation aspect ratio of inclusions in the transverse and longitudinal section of the hot-rolled samples with different aluminum content.

3.3. Melting Point of the Inclusions in the Hot-rolled Steel

The isothermal ternary phase diagram of Al₂O₃–CaO–Ti₂O₃ system was calculated by software FactSage 6.1 and the oxide core composition in the samples was marked as shown in Fig. 10. Figure 10 shows that the melting point of the oxide core in sample 1 is 1450–1750°C, lower than that in sample 2, which is 1650–1950°C. From the view of the ratio x_Al2O3/x_CaO in the oxide core, variation range of the oxide core composition in sample 1 is smaller than that in sample 2. When Al₂O₃ and CaO formed calcium aluminates, the calcium aluminates of the oxide core in sample 1 is distributed between CaO·2Al₂O₃ and CaO·Al₂O₃ while that in sample 2 is between CaO·6Al₂O₃ and CaO·Al₂O₃ due to the different calcium and aluminum ratio in steel. As can be seen from this figure, the melting point of the inclusions will reduce effectively while about 17.5–33.5 mole percent of Ti₂O₃ is contained when x_Al2O3/x_CaO ≈ 1–3, and the inclusions in sample 2 with higher aluminum content have little titanium oxides.

Fig. 10.

The isothermal phase diagram of Al₂O₃–CaO–Ti₂O₃ system.

4. Discussion

4.1. Effect of Al Content on Inclusion Composition and Morphology

The equilibriums of Al–O and Ti–O reactions in liquid steel at 1600°C were calculated according to the thermodynamic data related with Eqs. (1), (2), (3), (4) in Tables 2 and 3. The equilibrium content of oxygen in liquid steel at 1873 K is plotted against aluminum and titanium, respectively, in Fig. 11(a). Figure 11(a) shows that the priority of oxides to precipitate is in the order of Al₂O₃, Ti₂O₃, Ti₃O₅ and TiO₂. It should be noted that the data points plotted in Fig. 11(a) were total oxygen and acid soluble Al and Ti, we cannot judge the composition of Al and Ti deoxidation products by this figure only even the thermodynamic conditions for the precipitation of Al₂O₃ and Ti₂O₃ are satisfied in both steels according to thermodynamics. As a result of the aluminum and titanium competitive reaction with oxygen, the deoxidization products vary with different Al_S contents. According to the EDS results, the dominant oxide of Al–Ti competitive deoxidation is TiO_x–Al₂O₃ in sample 1, while that is Al₂O₃ with little TiO_x in sample 2.

Table 2. Reactions of Al–Ti complex deoxidation in liquid steel.

Reaction equation	ΔG0
[Ti] + 2[O] = TiO2(s)	–673360 + 227.1 T (J·mol–1)29)	(1)
2[Ti] + 3[O] = Ti2O3(s)	–1100392 + 356.7 T (J·mol–1)30)	(2)
3[Ti] + 5[O] = Ti3O5(s)	–1774016 + 569.9 T (J·mol–1)30)	(3)
2[Al] + 3[O] = Al2O3 (s)	–1224828 + 393.6 T (J·mol–1)27)	(4)

Table 3. Activity interaction parameters of elements in steel at 1600°C.²⁷⁾

i   j	C	Si	Mn	S	Al	Ti	O	Ca
Al	0.091	0.056		0.035	0.043	0.016	–1.98	–0.047
Ti	–0.165	0.05	–0.043		0.024	0.013	–1.8
O	–0.45	–0.066	–0.021	–0.133	–1.17	–0.6	–0.17	–310
Mn	–0.07	–0.017	0	–0.048		–0.05	–0.083
S	0.11	0.063	–0.026	–0.028	0.035	–0.072	–0.27

Fig. 11.

(a) The Ti–O and Al–O equilibriums in molten steel at 1600°C, (b) Equilibrium relationship between Al and Ti contents in liquid iron saturated with Al₂O₃ or Ti₂O₃ ([mass%Mn] = 1.2, [mass%C] = 0.06, [mass%Si] = 0.05, 1600°C).

The stoichiometry of TiO_x is difficult to be determined using EDS. In this study, the EDS results show that TiO_x is close to Ti₂O₃ which agrees well with the observation results²⁶⁾ that the composition of inclusion in 0.06%C-1.2%Mn-0.001%Al-0.014%Ti was Ti₂O₃ rich and the thermodynamic calculations results²⁷⁾ by H. Mitsutaka as shown in Fig. 11(b). Pak and Kim²⁸⁾ reported that Ti₂O₃ could form a solid solution with Al₂O₃ and it is well consistent with the inclusions morphology of sample 1 in the present work as shown in Figs. 3, 6 and 7. Combined with the observation and analysis, TiO_x is considered to be Ti₂O₃ in the steels of this study.

However, the precipitation of the corresponding oxides is inevitable due to the existence of the supersaturated elements in local liquid steel. Thus, the deoxidation product is Al₂O₃–Ti₂O₃ in Ti-killed steel sample 1, while that is Al₂O₃ with a little Ti₂O₃ in Al-killed Ti-alloyed steel sample 2 as shown in Figs. 3 and 4. Figures 3, 6 and 7 show that inclusions with two kinds of composite oxide core are spherical which are Al₂O₃ with a little TiO_x or TiO_x with a little Al₂O₃. Moreover, the core oxides is mosaic while x_Al2O3/x_Ti2O3 = 0.3–3.5.

It is well known that titanium oxides nucleate rapidly while they grow slowly.^17,18) Thus, the size of titanium oxides is small and the spatial distribution is dispersed. Compared with aluminum, titanium deoxidation can increase the number of inclusions with smaller size. Figure 5 shows a substantial increase of inclusions smaller than 1 μm in steel 1 compared to steel 2. Many research works have shown that a suitable size of nonmetallic inclusions between 0.2 μm and 3 μm is conducive to the nucleation of IGF.^17,18,31) Compared to the high aluminum content steel, the number of Ti-containing inclusions in low aluminum content steel is larger, the size is smaller and the spatial distribution is more dispersed. Therefore, the low aluminum content steel would be more conducive to the nucleation and finely diffusion of IGF in steel.

4.2. Effect of Al Content on Calcium Treatment and MnS Precipitation

Calcium was used to modify the inclusions in the steel after Al-Ti complex deoxidation in this study. The calculation of calcium treatment reactions was conducted according to the thermodynamic data shown in Tables 4 and 5. The value of a_CaS is taken as 0.75 while there is a transient CaS mixed with transient MnS.³²⁾ The thermodynamic calculation results are shown in Fig. 12.

Table 4. Reactions in calcium treated molten steel.

Reaction equation	log K
[Ca] + [O] = CaO(s)	25657/T―7.6527)	(5)
[Ca] + [S] = CaS(s)	l9980/T―5.9027)	(6)
3[Ca] + (Al2O3)incl = 2[Al] + 3(CaO)incl	5667/T―2.5827)	(7)
2[Al] + 3[S] + 3(CaO)incl = (Al2O3)incl + 3(CaS)incl	44273/T―15.1227)	(8)
MnS(s) = Mn (l) + S (l)	8628.9/T―4.7433)	(9)
MnS(s) = Mn (δ) + S (δ)	10592.3/T―4.2733)	(10)

Table 5. Activities of CaO and Al₂O₃ in different calcium aluminates at 1600°C.^27,34)

Calcium aluminate	aCaO	aAl2O3	Melting point (°C)
3CaO·Al2O3/Liquid	0.773	0.01	1535
12CaO·7Al2O3	0.340	0.064	1455
Liquid/CaO·Al2O3	0.340	0.275	1605

Fig. 12.

Phase stability diagram of the Al–Ca–O in molten steel at 1600°C.

The equilibrium content of aluminum and calcium equilibrated with different calcium aluminates in liquid steel at 1600°C is plotted in Fig. 12. It shows that the calcium aluminate in both samples is in the stable region of 12CaO·7Al₂O₃ according to the concentration of dissolved calcium and aluminum in liquid steels, while that is closer to the dividing line between 12CaO·7Al₂O₃ and CaO·Al₂O₃ in the sample 2. Suitable calcium content for inclusion modification in low aluminum molten steel (sample 1) is lower and its range is smaller than that in high aluminum molten steel (sample 2). As suggested from above, Ca treatment effect in low aluminum molten steel is better than that in high aluminum molten steel. Since the large liquid inclusions are easy to float up to the surface of the molten steel, even though there are some differences of the calcium aluminate composition between the thermodynamic calculations and the residual inclusions observed in the samples shown in Fig. 10.

Figure 13(a) is the relationship between Al and S in molten steel calculated by the Eq. (8) in Table 4 with the activities of CaO and Al₂O₃ in different calcium aluminate in Table 5. Under the premise that the calcium aluminate in composite inclusions is 12CaO·7Al₂O₃ as shown in Fig. 12, the sulfur content of samples should reach to the equilibrium line of generating 12CaO·7Al₂O₃ as a stable CaS exists in Fig. 13(a). In fact, the sulfur content at certain Al content in this study is much lower than this value. Therefore, CaS is not stable and acts as an intermediary products in liquid steel at 1600°C. However, it can be seen from Eq. (8) and Fig. 13(a) that CaS tends more easily to precipitate in the high aluminum steel than that in the low aluminum steel under a specified sulfur content. Conversely, the sulfur content can be relatively higher in the low aluminum steel than that in the high aluminum steel without CaS precipitation.

Fig. 13.

(a) Relationship between Al and S in molten steel at 1600°C based on the equilibrium of reaction (8), (b) critical sulfur content of MnS precipitation during solidification of the liquid steel.

Precipitation of MnS during solidification of liquid steel was calculated by Eqs. (11), (12), (13), (14) based on the microsegregation model proposed by Clyne-Kurz.³⁵⁾   

$$C_L\text{=}C{}_0\text{[1}-\text{(1}-\text{2}\beta k\text{)}{f}_S]{\hspace{0.17em}}^{\frac{(k-1)}{1-2\beta k}}$$(11)

  

$$\beta \text{=}\alpha \left[1-\text{exp}\left(-\frac{1}{\alpha }\right)\right]-\frac{1}{2}\text{exp}\left(\frac{1}{2\alpha }\right)$$(12)

  

$$\alpha \text{=}\frac{D_S{t}_f}{{(0.5{\lambda }_S)}^2}$$(13)

  

$$t_f\text{=}\frac{T_L-T_S}{C_R}$$(14)

Where T₀, T_L and T_S represent the melting point of pure iron (1536°C), liquidus and solidus temperature of the studied steel, respectively, C_L is the concentration of a given solute element in the liquid at the solid-liquid interface, C₀ is the initial (nominal) liquid concentration, k (= C_S/C_L) is the equilibrium partition coefficient for that element, f_S is the solid fraction, α is a back-diffusion parameter, D_S is the diffusion coefficient of solute in the solid phase in cm²s^–1, t_f is the local solidification time in seconds, λ_S is the secondary dendrite arm spacing in cm, C_R is the cooling rate (°C/s).   

$${\lambda }_S\text{=}(169.1-720.9)C_0{\hspace{0.17em}}^{-0.4935}\hspace{1em}{(0<[\%\mathrm{C}]\le 0.15)}^{}$$(15)³⁶⁾

Where C₀ is the carbon content (in mass percent) of steel and λ_S is the secondary dendrite arm spacing in micron.

The liquidus temperature is calculated using Eq. (16)³⁷⁾ when the solidus temperature is calculated by iterative computation. The temperature at the solid-liquid interface, T, is given by Eq. (17).³⁸⁾ The effect of temperature on interaction paraters between elements is showed by Eq. (18).³⁹⁾ The cooling rate of the solidification is 5°C/s.   

$$\begin{array}{c}{T}_L=1\hspace{0.17em}\hspace{0.17em}809-78w([\mathrm{C}])- 7.6w([\mathrm{S}\mathrm{i}]) \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}-4.9w([\mathrm{M}\mathrm{n}])-34.4w([\mathrm{P}]) -38w([\mathrm{S}])\end{array}$$(16)

  

$$T=T_0-\text{(}{T}_0-T_L)/[1-f_S(T_L-T_S)/(T_0-T_S)]$$(17)

  

$$e_i^{\hspace{0.17em}j}{\hspace{0.17em}}_{(T)}=\left(\frac{2\hspace{0.17em}538}{T}-0.355\right)\times e_i^{\hspace{0.17em}j}{\hspace{0.17em}}_{(1\hspace{0.17em}\hspace{0.17em}873K)}$$(18)

Where $e_i^j$ is interaction parameter between elements.

According to Fe–Fe₃C phase diagram, the solidification of the steels in this experiment with 0.03 mass% carbon is in solid (δ) phase. Data of equilibrium partition coefficients and diffusivity of solute elements are showed in Table 6.

Table 6. The equilibrium partition coefficients (k) and diffusion coefficients (D).³⁷⁾

Element	kδ/L	Dδ/L, (cm2/s)
Mn	0.76	0.76 Exp (–224430/RT)
S	0.05	4.56 Exp (–214639/RT)

Note: R (the gas constant) = 8.314 Joule/mole·K, and T is the temperature in Kelvin

Although the sulfur content can be relatively higher in the low aluminum steel than that in the high aluminum steel without CaS precipitation, it still should lower than the critical sulfur content of MnS precipitation. The critical sulfur contents in molten steels at the beginning of solidification (f_S = 0) were back-calculated with the sulfur contents in the end of solidification (f_S = 1) assuming that MnS would precipitate when f_S = 1 as showed in Fig. 13(b). Figure 13(b) shows that sulfur contents in the steels are lower than the critical content which is 0.0111 mass% in the low aluminum steel and 0.0100 mass% in the high aluminum steel. Combined with Fig. 13(a), the tolerance of sulfur content can be appropriately increased in low aluminum steel while the sulfur content is lower than the critical sulfur content of MnS precipitation in the liquid phase during solidification of steel. Meanwhile, the variations of manganese and sulfur concentration product with solid fraction in both steels are calculated and showed in Fig. 14. The results show that there is no single MnS precipitation during the solidification of steels and it agrees well with the SEM/EDS results.

Fig. 14.

Variations of manganese and sulfur concentration product with solid fraction.

The relationship between the thickness of MnS precipitation on the surface of oxides and the inclusions number in steel with different aluminum content are shown in Fig. 15. The average thickness of MnS precipitation on the surface of oxides in sample 1 is only 0.15 μm, about half of that in sample 2 which is 0.32 μm, and the inclusion number in sample 1 is about 2.5 times of that in sample 2. Several Studies^17,18,19) have shown that MnS heterogenous nucleates easily on some oxides, especially titanium oxide. Thus, well dispersed titanium-containing oxides are beneficial to MnS precipitating to form fine and dispersed MnS inclusions. Compared with sample 2, MnS inclusions in sample 1 are finer and more uniformly precipitated on oxides. Moreover, the possibility of single MnS precipitated at the grain boundaries is reduced.

Fig. 15.

Effect of Al content on inclusion number and MnS thickness.

4.3. Effect of Al Content on the Inclusion Deformation Behavior of the Inclusions

Large and easily deformable inclusions have a harmful influence on ductile fractures.^2,3,4,5) The results show that a large deformation inclusion has two obvious characteristics: i, the thickness of the surface MnS is large as shown in Figs. 9 and 15, ii, the oxide core contains a small amount of titanium oxides, a relatively high calcium content (x_Al2O3/x_CaO = 2.5–3.2) as shown in Figs. 7 and 8. As well as, the deformation of the inclusions is determined by the deformation of the oxide core and the surface MnS layer. Therefore, in order to obtain inclusions with an ideal deformation aspect ratio expected, the composition of inclusions should be appropriately controlled, and the thickness of MnS precipitates on the surface of oxides should be reduced.

As analyzed above, the deformation aspect ratio of the oxide core and the whole inclusions decrease with the increase of Ti₂O₃ content while x_Al2O3/x_CaO = 1.0–3.2. The deformation aspect ratio of inclusions is about 1–2 while the composition of oxide core is x_Al2O3 = 35–55%, x_Ti2O3 = 15–35%, and x_CaO = 10–25%. An appropriate content of aluminum, Ti and Ca will be studied further after the appropriate ratio of Al₂O₃, CaO and Ti₂O₃ is discussed.

Compared with the Al-killed and Ti-alloyed steel (high aluminum content), the Ti-killed steel (low aluminum content) is more conducive to the precipitation of the composite inclusions with a relatively high content of TiO_x, CaO and an appropriate content of Al₂O₃. Thus, inclusions of an ideal deformation aspect ratio can be obtained. Therefore, the low aluminum steel is more conducive to obtain inclusions with smaller deformation aspect ratio.

4.4. Effect of Al Content on the Melting Point of the Inclusions and the Cleanliness of the Steel

The most concise way to evaluate the cleanliness of steel is measuring impurity content such as oxygen and sulfur content, which is restricted in clean steel within a very low level because of the harm of oxide and sulfide inclusions to steel properties. Ultimately, the limitation of impurity in steel depends on the extreme reduction of harmful inclusions.

Oxide metallurgy is employed to form finely dispersed inclusions in the steel to improve the properties of steel in the present experiments. The melting points of the composite inclusions (1400–1750°C) in low aluminum steel are reduced more effectively than that (1650–1950°C) in high aluminum steel due to the formation of TiO_x bearing calcium aluminate inclusions as shown in Fig. 10. Thus, large inclusions with low melting point are removed easily from the molten steel by floating up to the surface of the molten steel. Meanwhile, the number of the fine Ti-containing inclusions in low aluminum steel can be substantially increased than that in high aluminum steel. The results in this study show that the number of inclusions, which are almost all Ti-containing composite inclusions with an average size of 1.0 μm in low Al content steel (sample 1), is 1156/mm² and 2.5 times of that in high Al content steel (sample 2). Previous work⁴⁰⁾ suggested that the non-metallic inclusions would not influence the macroscopic properties of the material while their size was less than 1 μm and the distance between each other was greater than 10 μm. Therefore, the number and size distribution of inclusions in low Al content steel are harmless although its total oxygen content is relatively higher than that in high Al content steel. Moreover, the heterogenous nucleation rate of IGF is increased and the possibility of single MnS precipitated at the ferrite grain boundaries is reduced, owning to a large number of fine dispersed Ti-containing inclusions and fine MnS precipitated uniformly on Ti-containing oxides. As a result, the harmfulness of MnS is reduced and the properties of the steels are expected to be improved.

Refining inclusions and reducing harmful inclusions by controlling the related element content is the key to improving the properties of steel. The evaluation of steel cleanliness only by the element content could not be completely applicable to oxide metallurgy to some extent. According to oxide metallurgy, the critical parameters of the harmful and beneficial inclusions beside the element content must be taken into further consideration in the new evaluation methods of steel cleanliness.

5. Conclusions

In this study, the morphology, composition, number, size distribution, melting point and deformation of inclusions in Al–Ti complex deoxidized followed with calcium treatment steels with different aluminum content have been systematically investigated based on the experiment and thermodynamic calculation. The major findings are as follows:

(1) The inclusions in low aluminum steel appear with a mosaic core of Al₂O₃–CaO–Ti₂O₃ and MnS wrapped, and the inclusions in high aluminum steel appear with a bundle core of Al₂O₃ with a little Ti₂O₃–CaO and MnS wrapped. Inclusion with two kinds of composite oxide core is spherical: i. Al₂O₃ with a little TiO_x, ii. TiO_x with a little Al₂O₃. The oxide core is mosaic while x_Al2O3/x_Ti2O3 = 0.3–3.5.

(2) The number of inclusions in low aluminum steel is 1156/mm² and the average size is 1.00 μm, and those in high aluminum steel are 467/mm² and 1.15 μm respectively. Compared with high aluminum steel, low aluminum steel is more conducive to the nucleation of IGF and grain refinement.

(3) The composition of calcium aluminate in low aluminum steel is more close to 12CaO·7Al₂O₃. Suitable calcium content for inclusion modification in low aluminum molten steel is lower and its range is smaller than that in high aluminum molten steel. Ca treatment effect in low aluminum molten steel is better than that in high aluminum molten steel.

(4) The thickness of MnS layer on the surface of oxides inclusion in low aluminum steel is 0.15 μm, which is smaller than that of 0.32 μm in high aluminum steel. For a certain sulfur content, compared with high aluminum steel, low aluminum steel is more conducive to the finely dispersed MnS precipitating on oxides and thus the single MnS precipitated on grain boundary is possibly avoided to some extent.

(5) The melting point of inclusions in low aluminum steel is about 1450–1750°C, lower than that in high aluminum steel which is 1650–1950°C. The deformation aspect ratio of inclusion in low aluminum steel is about 1–2 smaller than that in high aluminum steel which is 1–7. The melting point and deformation aspect ratio of inclusions can be effectively reduced while a certain amount of Ti₂O₃ is included. The deformation aspect ratio of inclusions is about 1–2 while the composition of oxide core is x_Al2O3 = 35–55%, x_Ti2O3 = 15–35%, and x_CaO = 10–25%.

(6) In the point of view of oxide metallurgy, the critical parameters of the harmful and beneficial inclusions in addition of the element content should be taken into further consideration with the new evaluation method for steel cleanliness.

Acknowledgments

The authors wish to express their appreciation to the National Natural Science Foundation of China (Grant No.51104109, No.51210007), Hubei province Natural Science Fund (2008CDA010) and Wuhan University of Science and Technology research project (010328) for providing financial support which enabled this study to be carried out.

References

• 1)   H.  Ono and  T.  Ibuta: ISIJ Int., 51 (2011), 2012.
• 2)   P.  Fassina,  M. F.  Brunella,  L.  Lazzari,  G.  Re,  L.  Vergani and  A.  Sciuccati: Eng. Fract. Mech., 103 (2013), 10.
• 3)   T. Y.  Jin,  Z. Y.  Liu and  Y. F.  Cheng: Int. J. Hydrogen Energ., 35 (2010), 8014.
• 4)   G. H.  Xiao,  H.  Dong,  M. Q.  Wang and  W. J.  Hui: J. Iron Steel Res. Int., 18 (2011), 58.
• 5)   K. I.  Yamamoto,  H.  Yamamura and  Y.  Suwa: ISIJ Int., 51 (2011), 1987.
• 6)   R.  Takata,  J.  Yang and  M.  Kuwabara: ISIJ Int., 47 (2007), 1379.
• 7)   Z.  Enno,  V. H.  Corrie,  V.  Henk,  W.  Albert and  I. N.  Jung: ISIJ Int., 52 (2012), 52.
• 8)   P.  Kaushik,  J.  Lehmann and  M.  Nadif: Metall. Mater. Trans. B, 43B (2012), 710.
• 9)   H. K.  Sung,  S. Y.  Shin,  W.  Cha,  W.  Cha,  K.  Oh,  S.  Lee and  N. J.  Kim: Mater. Sci. Eng. A, 528 (2011), 3350.
• 10)   C.  Liu and  S. D.  Bhole: Sci. Technol. Weld. Join., 18 (2013), 170.
• 11)   F.  Huang,  J.  Liu,  Z. J.  Deng,  J. H.  Cheng,  Z. H.  Lu and  X. G.  Li: Mater. Sci. Eng. A, 527 (2010), 6997.
• 12)   N.  Verma,  P. C.  Pistorius,  R. J.  Fruehan,  M. S.  Potter,  H. G.  Oltmann and  E. B.  Pretorius: Metall. Mater. Trans. B, 43B (2012), 830.
• 13)   N.  Verma,  P. C.  Pistorius,  R. J.  Fruehan,  M.  Potter,  M.  Lind and  S.  Story: Metall. Mater. Trans. B, 42B (2011), 711.
• 14)   M.  Lind and  L.  Holappa: Metall. Mater. Trans. B, 41B (2010), 359.
• 15)   G. M.  Faulring and  S.  Ramalingam: Metall. Mater. Trans. B, 11B (1980), 125.
• 16)   D. C.  Gatellier: Tetsu-to-Hagané, 70 (1984), 872.
• 17)   M.  Fattahi,  N.  Nabhani,  M.  Hosseini,  N.  Arabian and  E.  Rahimi: Micron, 45 (2013), 107.
• 18)   J. S.  Byun,  J. H.  Shim,  Y. W.  Cho and  D. N.  Lee: Acta Mater., 51 (2003), 1593.
• 19)   H.  Ohta and  H.  Suito: ISIJ Int., 46 (2006), 480.
• 20)   X. J.  Zhuo,  Y. Q.  Wang,  X. H.  Wang and  H. G.  Lee: J. Iron Steel Res. Int., 17 (2010), 10.
• 21)   B.  Beidokhti,  A. H.  Koukabi and  A.  Dolati: J. Mater. Process. Technol., 209 (2009), 4027.
• 22)   D.  Zhang,  H.  Terasaki and  Y. I.  Komizo: Acta Mater., 58 (2010), 1369.
• 23)   M. K.  Sun,  I. H.  Jung and  H. G.  Lee: Metal. Mater. Int., 14 (2008), 791.
• 24)   C.  Wang,  N. T.  Nuhfer and  S.  Sridhar: Metall. Mater. Trans. B, 40B (2009), 1022.
• 25)   R.  Piao,  H. G.  Lee and  Y. B.  Kang: Acta Mater., 61 (2013), 683.
• 26)   H.  Goto,  K. I.  Miyazawa,  W.  Yamada and  K.  Tanaka: ISIJ Int., 35 (1995), 708.
• 27)   H.  Mitsutaka and  I.  Kimihisa: Thermodynamic Data for Steelmaking, Tohoku University Press, Sendai, (2010), 10.
• 28)   J. J.  Pak and  D. S.  Kim: Asia Steel Int. Conf., ISIJ, Tokyo, (2006), 540.
• 29)   G. K.  Sigworth and  J. F.  Elliott: Met. Sci., 8 (1994), 298.
• 30)   M. A.  Van Ende,  M. X.  Guo,  R.  Dekkers,  M.  Burty,  J.  Van Dyck,  P. T.  Jones,  B.  Blanpain and  P.  Wollants: ISIJ Int., 49 (2009), 1133.
• 31)   G.  Thewlis,  J. A.  Whiteman and  D. J.  Senogles: Mater. Sci. Technol., 13 (1997), 257.
• 32)   D. Z.  Lu: Ironmaking Steelmaking, 5 (1991), 342.
• 33)   Y.  Ueshima,  K.  Isobe,  S.  Mizoguchi,  H.  Maede and  H.  Kajioka: Tetsu-to-Hagané, 74 (1988), 465.
• 34)   B.  Korousic: Steel Res., 62 (1991), 285.
• 35)   T. W.  Clyne and  W.  Kurz: Metall. Mater. Trans. A, 12A (1981), 965.
• 36)   Y. M.  Won and  B. G.  Thomas: Metall. Mater. Trans. A, 32A (2001), 1755.
• 37)   Y.  Ueshima,  S.  Mizoguchi,  T.  Matsumiya and  H.  Kajioka: Metall. Mater. Trans. B, 17B (1986), 845.
• 38)   J.  Matsuno: Proc. Symp. Solidification Phenomena, Solidification Comm., ISIJ, Tokyo, (1976), 2.
• 39)   X. H.  Huang: Theory of Iron and Steel Metallurgy, 3rd ed., Metallurgical Industry Press, Beijing, (2002), 111.
• 40)   J. H.  Lowe and  A.  Mitchell: Clean Steel, Superclean Steel Conf. Proc., Institute of Materials, London, (1995), 223.