Effect of Basicity and Al₂O₃ Content in Slag on Cleanliness of Electroslag Ingot

Abstract

In order to clarify the mechanism for the increase in oxygen content during electroslag remelting, this paper analyzes in detail the effect of slag components on the gas content and number, size and composition of inclusions in electroslag ingots. The results show that, when the CaF₂-MgF₂-CaO-MgO-SiO₂ slag system is used for electroslag remelting, the oxygen content in electroslag ingots decreases clearly with the increase in slag basicity. When the basicity is 14~3.3, the oxygen content in electroslag ingots is only 9~11 ppm. When the basicity drops to 2.3, the oxygen content increases to 20 ppm. The slag with high basicity also leads to a decrease in the number and size of inclusions with high Ca content. When the CaF₂-MgF₂-CaO-MgO-SiO2-Al₂O₃ slag system is used for electroslag remelting, with the increase in the Al₂O₃ content in slag with the same basicity, the oxygen content and inclusions in electroslag ingots gradually increase. However, the Al content in inclusions increases and the Ca content decreases. The decomposition of SiO₂ and Al₂O₃ in slag is the main reason for the decrease in the cleanliness of electroslag ingots.

1. Introduction

Non-metallic inclusions in steel are one of the most important factors affecting the fatigue properties of parts. Almost all fatigue cracks originate from non-metallic inclusions [1,2,3]. Therefore, reducing the content and size of inclusions is an important way to improve the performance of parts. Excluding nitride inclusions, non-metallic inclusions in steel mainly include oxides, sulfides and oxygen sulfides. Therefore, decreasing the content of oxygen and sulfur in steel is the premise for reducing the content of inclusions. With the development of modern steelmaking technology, especially secondary refining technology, the oxygen and sulfur content in steel have been effectively decreased. With GCr15 bearing steel as an example, the total oxygen in steel reaches the ultrapure level of 5 ppm [4], whereas reducing the sulfur content to 1 ppm is easily attainable.

However, the decrease in the contents of oxygen and sulfur only means the reduction in the content of inclusions, and not of their size. A large number of studies have shown that not only the total amount of inclusions but also their size have a great impact on the fatigue life of parts [5,6,7]. In the traditional steelmaking processes such as LF (ladle refining), RH (Vacuum circulation degassing refining) and VD (vacuum degassing), or other refining processes, or the continuous casting process, including tundish and mold, all of the molten steel is treated in lined vessels. Although the content of impurity elements in steel can be reduced to a very low level, the molten steel will inevitably be polluted by the furnace lining, including the peeling off of the lining and the involvement of mold flux. These accidental and irregular large inclusions are fatal to the service life of parts.

As a special metallurgical method, electroslag remelting technology has evident advantages in the removal of large inclusions, which is mainly due to the large contact area between steel and slag. According to the literature [8], the contact area between steel and slag at the end of the electrode is 3128 mm²/g, which is considerably higher than that of other metallurgical methods. However, the industrial production and laboratory research have revealed that when the oxygen content in the electrode is low, the oxygen content in the electroslag ingot will increase significantly after electroslag remelting. After electroslag remelting of GCr15 bearing steel, Zhoudeguang et al. [9] observed that the oxygen content increased from 10 ppm in the consumable electrode to 32 ppm in the electroslag ingot. Lin et al. [10] reported that the oxygen content increased from 14 ppm in the consumable electrode to 18 ppm in the electroslag ingot when H13 steel was remelted in a gas shielded electroslag furnace. Zhoulixin et al. [11] also stated that the oxygen content increased from 10 ppm in the consumable electrode to 15 ppm in the electroslag ingot when G20CrNi2MoA bearing steel was remelted. Even if the vacuum electroslag remelting was adopted, the oxygen content in the electroslag ingots still increased [12].

Scientists have searched for the cause of the increased oxygen content after electroslag melting. The literature [13,14] shows that the Al₂O₃ content in slag has a very important effect on the oxygen content in the electroslag ingot. With the increase in the Al₂O₃ content in slag, the oxygen content in the electroslag ingot gradually increases. The oxygen content of the electroslag ingot prepared with Al₂O₃-free slag is the lowest. The authors also observed this phenomenon in their previous research [15,16], and confirmed that there is a clear corresponding relationship between the oxygen content in the electroslag ingot and the Al₂O₃ content in slag. To further analyze the effect of slag composition on the cleanliness of an electroslag ingot, this paper studied in detail the effect of basicity and the Al₂O₃ content in slag on the total oxygen content and inclusion distribution with 304 L stainless steel as the research object.

2. Experimental Section

2.1. Experimental Equipment and Material

The experiment was carried out in an electroslag remelting furnace with gas protection, as shown in Figure 1.

As shown in Figure 1, a 100 kVA Ac supply power was available for electroslag remelting (ESR) ingot production with a high voltage terminal of 380 V, a low voltage terminal ranging from 28 V to 40 V and a maximum current of 2500 A. Based on preestablished electrical parameters, the consumable electrode can be driven up and down at a certain speed. The speed was controlled manually. In this experiment, a mold with a diameter of 100 mm was used for remelting.

The 304 L austenitic stainless steel used in the experiment was produced by 30 t EAF (electric arc furnace)-AOD furnace(argon oxygen decarburization furnace)-LF furnace (ladle refining furnace)-CC(continuous casting), where the metallurgical equipments are made in Xi’an electric furnace institute Co., LTD., Shanxi, China. First, the recycled stainless steel scraps and other metals were melted in an electric arc furnace, and the molten steel was transferred to an AOD furnace for decarburization. After decarburization, ferrosilicon was added for prereduction, and then, after removing the slag and lime, fluorite, and metal aluminum were added into the AOD furnace, where the all raw materials are purchased from Guangde Shunda Mineral Trade Co., Ltd., Xuancheng, China. After the slag in the AOD furnace was fully melted, the molten steel and slag were transferred to the LF furnace for refining. The composition and temperature of molten steel in the LF furnace were further adjusted, and a large amount of inclusions were removed. LF refining time was more than 25 min. After LF treatment, SiCa wire was fed to the molten steel for modification of inclusions. Finally, the molten steel with qualified composition and temperature was cast into a round billet with a diameter of 180 mm, and then forged into the metal consumable electrode with a diameter of 50 mm.

Table 1. Chemical composition of metal consumable electrode.

Elements	C	Si	Mn	P	S	Ni	Cr	O	N	Al
wt%	0.019	0.41	1.18	0.037	0.0025	8.10	18.27	0.0025	0.080	0.010

shows the composition of metal consumable electrode.

To accurately analyze the effect of slag components on the cleanliness of electroslag ingot, we prepared all remelting slags with an analytical reagent made in Sinopharm Chemical Reagent Co., Ltd., Shanghai, China, and the mass percentages of CaO, Al₂O₃, CaF₂, MgF₂, MgO and SiO₂ were greater than 98.0%, 99%, 98.5%, 98%, 99% and 99%, respectively.

2.2. Experimental Schemes and Process

The experimental schemes are shown in

Table 2. Experimental schemes.

Slags	Slag Composition				Basicity(R)	Electrical Parameters
	CaF2 + MgF2/wt%	Al2O3/wt%	CaO + MgO/wt%	SiO2/wt%		Voltage/V	Current/A
1	rest	0	28	2	14	28	2200~2500
2	rest	0	26	4	6.5	28	2200~2500
3	rest	0	23	7	3.3	28	2200~2500
4	rest	0	21	9	2.3	28	2200~2500
5	rest	2	26	4	6.5	28	2200~2500
6	rest	6	26	4	6.5	28	2200~2500
7	70% CaF2	30				28	1500

. For comparison, the traditional slag containing 30 wt%Al₂O_3–70 wt%CaF₂ was also used in the electroslag remelting experiment.

The scale of the electrode surface was removed by grinding prior to ESR. No deoxidizer was added to the slag bath during the ESR process. The experimental process was as follows:

(I) Remelting start. First, the slag was mixed and placed in a graphite crucible. Then, the graphite crucible was placed in a high-temperature furnace with 1600 °C for melting. The slag weighted 1.2 kg. Argon with a flow rate of 1.5 Nm³/h was introduced into the mold through the protective cover. The pressure of the cooling water was between 0.2 MPa and 0.3 MPa. After the slag was completely melted, it was quickly removed and poured into the mold. The consumable electrode dropped and remelting began. (II) Stable remelting. When the remelting process stabilized, the current gradually increased to the set value until the end of remelting. (III) End of remelting. When the electroslag ingot reached the predetermined height, the power and argon supply were turned off. The electroslag ingot was removed after the slag was completely solidified. (IV) Sample test. As shown in Figure 2, the samples of 15 mm × 15 mm × 15 mm and Φ 5 mm × 100 mm were cut off at 40 mm below the upper part of electroslag ingot for inclusion and oxygen/nitrogen content analysis. An ASPEX scanning electron microscope system made in FEI company, Pittsburgh, PA, USA with a step size of 1 μm was used to test the size, morphology and composition of inclusions in the area of 50 mm², and ONH-2000 was used to analyze the oxygen and nitrogen contents. The Ca content was analyzed by inductively coupled plasma atomic emission spectrometry (ICP-AES) and those of other elements were analyzed by electric spark direct reading spectrometry.

3. Experimental Result

3.1. Variation in Gas Content and Chemical Composition in Electroslag Ingots Prepared with Different Slag

Figure 3 shows the oxygen and nitrogen content in electroslag ingots.

It can be seen from Figure 3a that the basicity of remelting slag had minimal effect on the nitrogen content of the electroslag ingot but showed a significant effect on the oxygen content. When the basicity was 14~3.3, the oxygen content of the electroslag ingot was 9~11 mm. However, when the basicity dropped to 2.3, the oxygen content increased to 20 ppm, which was still lower than that of the consumable electrode (25 ppm). When the slag containing 30 wt%Al₂O₃-70 wt%CaF₂ was used for electroslag remelting, the oxygen content of the electroslag ingot increased sharply to 40 ppm, which was higher than that of the consumable electrode.

Figure 3b shows the effect of the Al₂O₃ content in the slag on the gas content of the electroslag ingot when the basicity remained unchanged. The Al₂O₃ content had minimal effect on the nitrogen content, whereas the oxygen content increased rapidly with the increase in Al₂O₃ content. When the basicity was 6.5 and no Al₂O₃ was present in the slag, the oxygen content in the electroslag ingot was 9.2 ppm. When the Al₂O₃ content in the slag was 2%, the oxygen content amounted to 11.2 ppm. When the Al₂O₃ content increased to 6%, the oxygen content rapidly increased to 17.7 ppm. Therefore, the oxygen content in the electroslag ingot was closely related to the slag components.

Table 3. Chemical composition of electroslag ingots.

Experimental Schemes	Chemical Composition wt%
	C	Si	Mn	P	S	Cr	Ni	Al	Ca
1	0.026	0.44	1.14	0.037	0.0011	18.12	8.19	0.008	0.0007
2	0.027	0.45	1.14	0.038	0.0011	18.22	8.15	0.010	0.0008
3	0.020	0.45	1.14	0.039	0.0011	18.25	8.21	0.008	0.0007
4	0.020	0.45	1.15	0.038	0.0024	18.26	8.16	0.009	0.0004
5	0.023	0.44	1.14	0.038	0.0010	18.21	8.21	0.018	0.0005
6	0.020	0.43	1.13	0.037	0.0010	18.18	8.25	0.032	0.0005
7	0.022	0.42	1.13	0.037	0.0028	18.16	8.19	0.036	0.0004

shows the chemical composition of the electroslag ingots. It can be seen from Table 3 that the S content in the electroslag ingots was very low due to the use of the high-basicity slags 1, 2, 3, 5 and 6. The S content was higher after remelting with low-basicity slags 4 and 7. Given that the slags 1, 2, 3 and 4 contained no or very little Al₂O₃, the increase in the Al content in the electroslag ingots was low. The Al content in the electroslag ingots increased with the increase in the Al₂O₃ content in slags 5, 6 and 7. Regardless of the slags used, the Ca content in the electroslag ingots was very low.

3.2. Effect of Basicity on the Number and Size of Inclusions

Figure 4 shows the changes in the number of inclusions with different diameters, in which the equivalent cycle diameter was used.

Figure 4 also reveals that 1244 inclusions were found in the consumable electrode. When slags with the basicities of 14, 6.5 and 3.3 were used for remelting, the numbers of inclusions reduced to 200, 204 and 208, respectively, which was consistent with the low oxygen content in the electroslag ingots. However, when the basicity decreased to 2.3, the number of inclusions increased to 1379. When the slag containing 30 wt%Al₂O₃-70 wt%CaF₂ was used for remelting, the number of inclusions further increased to 1687.

Not only the total amount of inclusions but also the number of inclusions with different diameters changed significantly. No matter the consumable electrode or electroslag ingot, inclusions less than 6 μm in diameter predominated. However, the number of large inclusions in the electroslag ingots changed evidently with the basicity.

A total of 42 inclusions measuring more than 6 μm in diameter were observed in the consumable electrode; 9 of them were larger than 12 μm and the maximum diameter was 27 μm. When the basicities were 14, 6.5 and 3.3, exactly, 7, 4 and 6 inclusions measuring more than 6 μm in diameter were found in the electroslag ingot, respectively. The maximum diameters were 11, 7 and 11 μm, respectively. At 2.3 basicity, 17 inclusions that were more than 6 μm in diameter were detected in the electroslag ingot. However, the maximum diameter was 11 μm. When the slag containing 30 wt%Al₂O₃-70 wt%CaF₂ was used for remelting, 51 inclusions with a diameter of more than 6 µm were observed in the electroslag ingot, and the maximum diameter was 18 µm.

Through the above analysis, after remelting with high-basicity slag, the total number and size of inclusions in the electroslag ingots decreased. With the decrease in basicity, the number of inclusions increased. Compared with the consumable electrode, the diameter of inclusions was still smaller.

3.3. Effect of Basicity on Composition and Morphology of Inclusions

To compare the changes in inclusions in the ingots prepared with different slags, the composition and morphology of inclusions in the consumable electrode were analyzed first. Oxides (676 inclusions), sulfide–oxides (506 inclusions) and a small amount of MnS/CaS (62 inclusions) were found in the consumable electrode. Figure 5 shows the composition and morphology of oxide inclusions in the consumable electrode. In Figure 5b, the energy spectrum analysis is performed to decide the composition distribution of inclusions for the area in the yellow frame.

As can be seen from Figure 5a, the oxides in the consumable electrode were mainly composed of composite inclusions containing Al, Ca and Mg, and the contents of Ca and Mg in the oxides were evidently higher than that of Al. Such a finding was mainly due to deoxidization of the consumable electrode by Al during AOD and LF refining. Then, a SiCa wire was fed into the molten steel to modify the Al₂O₃ inclusions before continuous casting. The inclusions after Ca treatment were liquid in molten steel due to their low melting point. Therefore, they mainly existed in the steel as spherical inclusions, as shown in Figure 5b. Figure 6 shows the composition and morphology of sulfide–oxides in the consumable electrode.

The chemical composition of sulfide–oxides mainly included Mn, Ca, Al, S etc., which interacted to form spherical multi-element composite inclusions. Given the low sulfur content in the electrode, several sulfide inclusions were observed and they are not be analyzed here.

After electroslag remelting with high-basicity slag, the composition of inclusions changed greatly. In particular, slag showed a strong desulfurization capacity, and the sulfur content in the electroslag ingot was very low, which resulted in very few sulfide inclusions in the electroslag ingot. The inclusions were mainly oxides, the composition of which are shown in Figure 7.

Figure 7 shows that when slags with the basicities of 14, 6.5 and 3.3 were adopted, respectively, the Ca content in the inclusion was significantly higher than the Al content. However, after remelting by slag with a basicity of 2.3, the Al content in the inclusions was higher than the Ca content. Regardless of basicity, the inclusions were mainly spherical calcium aluminate, as shown in Figure 8.

Figure 9 shows the composition and morphology of inclusions in the electroslag ingot prepared by 30 wt%Al₂O₃-70 wt%CaF₂ slag. Given that the number of sulfides was 6.5% of the total amount of inclusions, only oxide inclusions were analyzed.

As shown in Figure 9, given the slag containing 30% Al₂O₃, the Al content in the inclusions increased sharply, whereas the Ca content decreased significantly. In addition, numerous pure Al₂O₃ inclusions were observed. Given that Al₂O₃ was a solid inclusion in the electroslag process, it exhibited an irregular morphology, as shown Figure 9b.

3.4. Effect of Al₂O₃ Content on the Number and Diameter of Inclusions

Figure 10 shows the changes in the number of inclusions with different diameters in ingots prepared with the slag containing different Al₂O₃ contents when the basicity remained unchanged.

It can be seen from Figure 10 that 204 inclusions were detected in the electroslag ingot when Al₂O₃-free slag was used for remelting. When the Al₂O₃ content in the slag increased to 2% and 6%, the number of inclusions increased to 402 and 491, respectively. When the Al₂O₃ content was 30%, the number of inclusions increased sharply to 1687, which was consistent with the change in oxygen content shown in Figure 3b.

When Al₂O₃-free slag was used, not only the number of inclusions but also the diameter were the smallest. Four inclusions with a diameter greater than 6 μm were observed, and the maximum diameter was 7 μm. When 2% Al₂O₃ was added to the slag, 13 inclusions with a diameter greater than 6 µm were noticed, and the maximum diameter increased to 14 μm. When the Al₂O₃ content in the slag increased to 6%, the number of inclusions with a diameter greater than 6 μm increased to 26, and the maximum diameter was 17 μm. When the slag with 30 wt% Al₂O₃-70 wt% CaF₂ was adopted, the number of inclusions with a diameter greater than 6 μm increased further to 51. However, the maximum diameter of inclusions barely changed.

From the above analysis, the Al₂O₃ content in slag had a significant effect on the number and diameter of inclusions. With the increase in the Al₂O₃ content, the number and diameter of inclusions increased.

3.5. Effect of Al₂O₃ on Inclusion Compositions in Electroslag Ingot

Figure 11 shows the compositions of inclusions in the electroslag ingots when slags containing 2% and 6% Al₂O₃ were adopted. Given the very low S content in the electroslag ingots and several sulfide inclusions, only the composition of oxide inclusions was analyzed. Combined with Figure 7b and Figure 9a, with the increase in the Al₂O₃ content in slag, the Al content of inclusions gradually increased and the Ca content gradually decreased. When the slag contained 2% Al₂O₃, the Al and Ca contents in the inclusions were almost the same (Figure 11a). When the Al₂O₃ content increased to 6%, the Al content in the inclusions exceeded the Ca content (Figure 11b). When the Al₂O₃ content in the slag was 30%, a large number of pure Al₂O₃ inclusions were observed in the electroslag ingots (Figure 9).

Figure 12 shows the morphology of inclusions in the electroslag ingots. Although the contents of Al and Ca in the inclusions changed, the inclusions retained their spherical shape.

4. Discussion

The experimental results in the previous section showed that when the Al₂O₃-free slag with a high basicity was used for remelting, the oxygen content in the electroslag ingots was low, and the number of inclusions was small. However, the Ca content in the inclusions increased significantly. When the basicity was low, the oxygen content increased significantly, and the Ca content in the inclusions decreased significantly. When the basicity remained constant, the oxygen content in the electroslag ingots increased with the increase in the Al₂O₃ content in the slag. The Al content in inclusions also increased and the Ca content decreased. Therefore, the slag components had evident effects on the cleanliness of the electroslag ingots.

The reaction expressions between the slag and composition of steel are given below together with their equilibrium constants [17].

$$[\mathrm{C}\mathrm{a}]+[\mathrm{O}]=(\mathrm{C}\mathrm{a}\mathrm{O}) \mathrm{Log}{K}_1=\frac{25655}{T}-7.65$$

(1)

$$\frac{1}{2}[\mathrm{S}\mathrm{i}]+[\mathrm{O}]=\frac{1}{2}(\mathrm{S}\mathrm{i}{\mathrm{O}}_2) \mathrm{Log}{K}_2=\frac{15112.5}{T}-5.78$$

(2)

$$\frac{2}{3}[\mathrm{A}\mathrm{l}]+[\mathrm{O}]=\frac{1}{3}(\mathrm{A}{\mathrm{l}}_2{\mathrm{O}}_3) \mathrm{Log}{K}_3=\frac{19440}{T}-6$$

(3)

where T is the temperature, K, and Ki is the equilibrium constant for each reaction.

According to Formulas (1)–(3), the equilibrium relations of [Ca]-[O], [Si]-[O] and [Al]-[O] at different temperatures can be obtained, as shown in Figure 13.

From Figure 13, with the increase in temperature, the oxygen content in equilibrium with Ca, Si or Al was substantially increased, which also indicated that the stability of oxides, such as CaO, SiO₂ and Al₂O₃, deteriorated with the increase in temperature, and the decomposition tendency increased significantly. During the electroslag remelting process, the temperature of the slag pool was higher than that of ordinary steelmaking, which may be as high as 1700~1800 °C [18]. Therefore, the oxides in the slag were partially decomposed, resulting in an increased oxygen content in the ESR ingot. In addition, the instability of CaO, Al₂O₃ and SiO₂ increased successively at high temperatures. Therefore, to reduce the decomposition tendency of oxide components in the slag pool, we reduced the contents of SiO₂, Al₂O₃ and other oxides with poor stability.

The stability of slag (S) was introduced in the literature to further analyze the relationship between the components of slag and oxygen content in steel [19]. The larger the S, the more stable and the more difficult to decompose the slag system. The specific calculation method is as follows.

Metal element X reacts with 1 mol oxygen according to Formula (4).

$$\mathrm{a}\mathrm{X}+{\mathrm{O}}_2={\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2$$

(4)

$\Delta {\mathrm{G}}_{{\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2}^{\mathsf{\Phi }}$ represents the change in standard Gibbs free energy of generating X_aO₂, and $\Delta {\mathrm{G}}_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2}^{\mathsf{\Phi }}$ is used as reference standard. The square of the ratio of $\Delta {\mathrm{G}}_{{\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2}^{\mathsf{\Phi }}$ to $\Delta {\mathrm{G}}_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2}^{\mathsf{\Phi }}$ is called the stability coefficient of X_aO₂, which represents the stability of X_aO₂, as shown in Formula (5).

$${\mathrm{M}}_{{\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2}={\left(\Delta {\mathrm{G}}_{{\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2}^{\mathsf{\Phi }}/\Delta {\mathrm{G}}_{\mathrm{S}\mathrm{i}{\mathrm{O}}_2}^{\mathsf{\Phi }}\right)}^2$$

(5)

The chemical stability of slag (S) is further defined as shown in Formula (6).

$$\mathrm{S}={\displaystyle \sum {\mathrm{M}}_{{\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2}}\cdot {\mathrm{n}}_{{\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2}$$

(6)

where ${\mathrm{n}}_{{\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2}$ is the mole fraction of X_aO₂ in the slag pool.

Given that the influence of CaF₂/MgF₂ on oxygen potential is very complex, consideration of the influence of CaF₂/MgF₂ on S is difficult. Thus, the influence of CaF₂/MgF₂ was ignored.

The calculated values of ${\mathrm{M}}_{{\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2}$ at 2000 K are shown in

Table 4. Slag stability coefficients of various oxides.

Oxides	CaO	Al2O3	MgO	TiO2	SiO2
${\mathrm{M}}_{{\mathrm{X}}_{\mathrm{a}}{\mathrm{O}}_2}$	2.16	1.61	1.34	1.14	1.00

.

As shown in Table 4, the stability of the different slags used in this paper was calculated, as shown in

Table 5. Slag stability of the slags used.

Slags	Slag-1	Slag-2	Slag-3	Slag-4	Slag-5	Slag-6	Slag-7
S	2.09	2.01	1.90	1.83	2.00	1.97	1.61

.

It can be seen from Table 5 that from slag-1 to slag-4, with the decrease in basicity, the stability of slag gradually decreased, and the oxygen content in the electroslag ingots increased, as shown in Figure 3.

When the basicity was high, the oxygen content in the ESR ingots was very low, but the Ca content in the inclusion considerably increased, which indicates that the CaO in the slag pool was slightly decomposed. A small amount of SiO₂ was observed in the slag pool; its activity was very low due to high basicity, and it hardly participated in the reaction during electroslag remelting. The [O] decomposed by CaO entered the metal bath, and [Ca] and [O] further combined to form inclusions with a high Ca content during solidification, as shown in Figure 14a.

With the decrease in basicity, the content of free SiO₂ in the slag pool increased. Given the poor stability of SiO₂ at high temperatures, it decomposed more [O] in the metal bath. Given the stronger binding force of Ca, Al and O, the [Ca], [Al] and [O] in the metal bath formed calcium aluminate inclusions in the solidification process, as shown in Figure 14b.

Under the condition of constant basicity, the stability of slag worsened with the increase in the Al₂O₃ content in the slag pool. When the Al₂O₃ content in the slag pool increased to 2% (slag-5), the oxygen content in the electroslag ingot increased slightly. When the Al₂O₃ content in the slag was 6%, the oxygen and Al contents in the electroslag ingot increased significantly, as shown in Table 4. This finding proves the decomposition of Al₂O₃ at high temperatures, which was also confirmed by the author’s previous research. Correspondingly, the Al content in inclusions also increased significantly with the increase in Al₂O₃ content in the slag, as shown in Figure 14c. The 30%Al₂O₃-70%CaF₂ slag showed the lowest stability, and its decomposition tendency was high at high temperature. When 30%Al₂O₃-70%CaF₂ slag was used for remelting, the Al content in the electroslag ingot increased from 0.01% to 0.036% in the consumable electrode. The Al content in the inclusions was the highest too.

5. Conclusions

• The basicity and Al₂O₃ content in slag have a very important effect on the oxygen content in electroslag ingots. When the basicity was 14~3.3, the oxygen content was 9~11 mm. When the basicity dropped to 2.3, the oxygen content increased to 20 ppm. When no Al₂O₃ was present in slag with the basicity of 6.5, the oxygen content was 9.2 ppm. As the Al₂O₃ content increased to 6%, the oxygen content rapidly increased to 17.7 ppm.
• With the decrease in basicity, the number of inclusions in the electroslag ingots increased, and the inclusions were mainly calcium aluminate. As the basicity decreased from 14 to 2.3, the number of inclusions increased from 200 to 1379. When Al₂O₃-free slag was used for remelting, 204 inclusions were found. When the Al₂O₃ content increased to 6%, the number of inclusions increased to 491. With the increase in the Al₂O₃ content, the Ca content in inclusions decreased, and the Al content increased gradually until pure Al₂O₃ inclusions appeared.
• The decomposition of SiO₂ and Al₂O₃ in the slag pool at a high temperature was the main reason for the increased oxygen content in the electroslag ingots. Therefore, to improve the cleanliness of electroslag ingots, it is advantageous to adopt high-basicity and low-Al₂O₃ slag for remelting.