Influence of Top Slag Containing TiO₂ and VO_x on Hot Metal Pre-Desulfurization

Abstract

The desulfurization capacity of top slag in the process of pre-desulfurization of hot metal containing vanadium and titanium was researched. The top slag system of CaO-SiO₂-MgO-Al₂O₃-TiO₂-VO_x that was formed by blast furnace slag and a CaO desulfurization agent reduced the sulfur in hot metal from 0.08 wt.% to 0.02 wt.%. It was found that the resulfurization of the slag happened in the later periods of the desulfurization process. The vanadium–titanium oxides were both acidic in the desulfurization slag. TiO₂ and VO_x reacted with the basic oxides to form CaTiO₃ and MgV₂O₄ at 1623 K, which reduced free CaO and was not conducive to top slag desulfurization. The results of calculation showed that the top slag desulfurization accounted for 15% of the total desulfurization. Using the ionic and molecule coexistence theory of slag structure, it is shown that the desulfurization efficiency could be enhanced by adjusting both the amount of desulfurization agent and the composition of the blast furnace slag before pre-desulfurization.

1. Introduction

Hot metal pre-desulfurization plays an important role in both reducing the burden of blast furnace (BF) desulfurization, and improving the production capacity in the iron and steel smelting process [1,2], where we found the effect of top slag in the desulfurization process is rarely studied. Top slag is mainly composed of the residual BF slag, and the desulfurization agent, during the pre-desulfurization process. The composition of top slag changes with the desulfurizing time during the entire process [2]. A. K. Biswas suggests that the BF slag still has some desulfurization capacity [3]. G. A. Irons and R. I. L. Guthrie found that the role of top slag is to absorb desulfurization products [4]. However, Zhang et al. propose that the top slag has no effect on desulfurization but has a role in resulfurization [5]. The effect of top slag on pre-desulfurization is still controversial.

Due to the use of a vanadium–titanium magnetite, the hot metal pretreatment process in PanGang produces a BF slag that contains 19–27 wt.% of TiO₂ and 1–3 wt.% of VO_x. After desulfurization, the basicity of slag increases from about 1.2 to about 3, and the content of TiO₂ decreases to about wt. 15%. The effect of TiO₂ on the desulfurization capacity of blast furnace slag has been studied in some of the literature and it is suggested that TiO₂ is acidic and not conducive to slag desulfurization [6,7]. However, the role of titanium oxides in the desulfurization slag system has rarely been studied. Moreover, the effect of vanadium oxides in slag on desulfurization is not clear. Vanadium oxides in different valence states, such as +2, +3, +4 and +5, exhibit different acid-base properties [8,9,10,11,12]. Therefore, it was necessary to study the role of a top slag containing oxides of vanadium and titanium in the pre-desulfurization process.

In this paper, the role of top slag on pre-desulfurization is analyzed by using a hot metal containing vanadium–titanium, BF slag and a desulfurization agent. The effect of vanadium and titanium oxides on top slag desulfurization was studied in the laboratory. Based on the experimental results, the pre-desulfurization in PanGang operations is analyzed. The desulfurization effect of a commercial top slag containing vanadium–titanium oxides in the pre-desulfurization process is confirmed.

2. Experimental Details

2.1. Equipment and Materials

The experimental device was a high-temperature furnace as shown in Figure 1. The temperature was controlled with a proportional integral differential (PID) controller using a Pt-6%Rh/Pt-30%Rh thermocouple. The heating element was made of six U-shaped MoSi₂ rods, which were connected in series and distributed evenly along the furnace tube. The furnace tube size was Φ 60 × 750 mm. This constant temperature zone was 100 mm, and the accuracy of temperature control was ±2 K.

The experiments were performed in a graphite crucible, as shown in Figure 2. There was a handle made of molybdenum wire that was tied onto the graphite crucible. The graphite crucible was put in a corundum crucible that was placed in the constant temperature zone. The corundum crucible played a protective role in preventing the hot metal from leaking out. Ar-gas (99.9%) at the bottom entered the furnace tube to keep a non-oxidizing atmosphere. The top firebrick could be taken out when loading and sampling.

The hot metal used in this experiment was pig iron (block) from PanGang and the BF slag (powder) was provided by PanGang too. It should be noted that a central sample of solid molten iron is taken after condensation. The desulfurization agent was compounded with chemicals (powder) with purity greater than 98.0%. Except for ferrous oxalate and vanadium oxalate, oxide materials and slags were placed in a muffle furnace for 8 h at 1173 K before being weighed. The compositions of hot metal, BF slag as well as chemicals are listed in

Table 1. Compositions of the hot metal (wt.%).

C	Si	Mn	P	S	V	Ti
4.23	0.08	0.12	0.07	0.08	0.28	0.19

,

Table 2. Compositions of the BF slag (wt.%).

CaO	SiO2	Al2O3	MgO	TiO2	V2O3	FeO	S
27.96	22.63	11.41	8.55	24.03	0.55	2.17	0.63

and

Table 3. Chemicals used for desulfurization experiment (%).

Name	CaO	SiO2	Al2O3	MgO	TiO2	FeC2O4·2H2O	VOC2O4·5H2O
Purity	98.0	98.5	99.0	98.0	98.0	98.0	99.0

.

2.2. Procedures

In each experiment, 200 ± 5 g of pig iron was weighed, and chemicals were made to a cylinder by using a tablet machine under 0.6 MPa. The alumina tube was flushed with Ar-gas (1.0 L/min) for 6 h, then it was put in the graphite crucible, and then the graphite crucible was inserted into the corundum crucible. Ar-gas was injected throughout the experiment. When the temperature reached 1623 K (1543–1703 K of pre-desulfurization temperature in steel plants), the graphite crucible was put into the corundum crucible. After the experiment, the graphite crucible was quickly taken out and placed into water. During this period, cooling water was prevented from entering the crucible. The sample solidified within 10 s.

To simulate the process where the desulfurization agent enters the top slag, pig iron was put into the graphite crucible. BF slag (5~15 g) was put on the top of the pig iron. In total, 5 g of the desulfurization agent composed of 85%CaO-10%SiO₂-5%Al₂O₃ was placed on top of the BF slag, as shown in Figure 2a. The components of slag are listed in

Table 4. Components of the slag (g).

Sample	BF Slag	CaO Desulfurization Agent
1#	10	0
2#	5	5
3#	10	5
4#	15	5

. The particular ratio of metal/slag/desulfurizing agent was chosen due to the limits of the experiment setup as well as to eliminate the influence of the CaO-based desulfurizer dosage. For Samples 2#–4#, the amount of desulfurization agent was kept unchanged, and the amount of BF slag was changed from 5 g to 15 g. To study the desulfurization capacity of the BF slag in pre-desulfurization, no desulfurization agent was added in Sample 1#. The graphite crucible was kept at 1623 K for 60 min.

To study the effect of TiO₂ and VO_x on pre-desulfurization, 20 ± 0.5 g formulated top slag was put on the top of the pig iron, as shown in Figure 2b. The components of the top slag are given in

Table 5. Components of the top slag (wt.%).

Sample	CaO	SiO2	Al2O3	MgO	TiO2	VO2
a	44.3	14.2	12.3	8.5	20.1	0.6
b	46.3	14.8	12.4	8.6	17.3	0.6
c	48.8	15.6	12.3	8.5	14.2	0.6
d	50.5	16.2	12.5	8.6	11.6	0.6
e	49.4	15.8	12.1	8.5	14.2	-
f	47.6	15.3	12.3	8.4	14.6	1.8
g	47.1	15.1	12.3	8.6	13.9	3.0

. For Samples a–d, the content of TiO₂ was changed from 20% to 12%, and the basicity and other elements were kept unchanged. Similarly, samples c and e–g were used to study the influence of VO_x content (0–3%). All the graphite crucibles were kept at 1623 K for 120 min. It should be noted that the top slag formed during the molten iron pre-desulfurization process is mainly composed of blast furnace slag and desulfurizer. To study the effect of different amounts of top slag on top slag desulfurization, the amount of CaO-based desulfurizer is set as constant and the amount of blast furnace slag is changed. Thus, the blast furnace slag is chosen.

2.3. Analysis

The sulfur concentration in molten iron was analyzed using the combustion-infrared absorption test method (calibrated by metal reference material, ASTM E1019) [13].

The composition of slag was analyzed by an X-ray fluorescence spectrometry (XRF, 1800, Shimadzu, Japan). The phase content of the slag was analyzed by X-ray diffractometry (XRD, TTRIII, Tokyo, Japan).

The melting point of the slag was determined using a hemispheric fully automatic melting point analyzer (RDS-04, Northeastern University, Shenyang, China). Once the pre-desulfurization test ended, the slag was ground and passed through a 200-mesh sieve. Part of the powder was fully blended with a caking agent, then was pressed into a cylindrical sample (Φ 3 × 3 mm). The sample was put in the test tube at 1473 K, then the test began. The slag melting point was determined using a quartation method—the temperature when the sample height reached 1/4 of the initial height. The detailed process schematic is given in Appendix A where Figure A1 shows the process schematic of automatic slag melting point tester.

3. Results

Table 6. Compositions of slag and sulfur content in hot metal after experiment (wt.%).

Sample	[S]	CaO	SiO2	Al2O3	MgO	TiO2	V2O3	FeO	(S)	R(CaO/SiO2)	Melting Point of Slag/K
1#	0.063	27.34	22.81	11.37	8.16	23.94	0.51	2.83	1.02	1.20	1565
2#	0.027	54.57	16.19	8.04	4.31	11.65	0.31	1.48	1.48	3.37	1696
3#	0.023	46.16	18.27	9.16	5.57	15.93	0.42	2.14	1.21	2.53	1687
4#	0.021	41.62	19.05	10.07	6.32	17.86	0.47	2.31	1.06	2.18	1611
a	0.014	43.73	13.43	11.68	8.04	19.73	0.55	1.73	0.67	3.26	1667
b	0.012	45.68	13.62	11.15	7.76	18.04	0.61	2.01	0.70	3.35	1652
c	0.012	47.18	14.54	12.06	8.07	14.71	0.57	1.54	0.68	3.25	1668
d	0.010	48.75	15.87	12.12	7.33	11.92	0.63	1.87	0.69	3.07	1683
e	0.011	48.02	15.28	11.87	7.65	13.97	0.05	2.26	0.71	3.14	1681
f	0.011	46.26	14.63	12.34	8.18	14.06	1.62	1.92	0.68	3.16	1672
g	0.013	46.01	14.17	11.73	7.86	13.81	2.91	2.38	0.69	3.25	1675

presents the experimental results for each sample after the pre-desulfurization test. The sulfur content in the molten iron increased with TiO₂ and VO_x. It should be noted that binary alkalinity (CaO/SiO₂) is the most used evaluation index. Here, it provides a brief description of the experimental slag. It not only illustrates the effect of changes in the amount of blast furnace slag on the binary basicity of top slag but also explains the influence of vanadium–titanium oxide content.

Figure 3 shows the XRD results of some slag samples. The CaTiO₃ content increased and the CaO content decreased with increasing TiO₂ content because of the chemical reaction between CaO and TiO₂. MgO·V₂O₃ appeared in the sample with 3% VO₂ addition.

4. Discussion

4.1. Influence of BF Slag

The relationship between increments of (CaO), (V₂O₃), (TiO₂) and (S) in top slag and BF slag addition is shown in Figure 4a. Vanadium and titanium oxides are mainly derived from the BF slag. The CaO content in the top slag gradually decreased with the amount of BF slag. The actual slag–iron sulfur distribution ratio (L_S) and the sulfur content of the molten iron are both decreased with the amount of BF slag from 54.81 to 50.48 and from 0.027 wt.% to 0.021 wt.%, as shown in Figure 4b. This seems contradictory to some references [14,15,16]. The higher the amount of BF slag, the lower the sulfur concentration in slag. Therefore, the sulfur content in the molten iron also decreases.

The experiment into how the desulfurization agent melts into the BF slag shows that the CaO gradually disperses into the BF slag with time. A desulfurization reaction of [S] + (CaO) = [O] + (CaS) ($\Delta G_{CaS}^o$= 105,784.6 − 28.723T (J·mol⁻¹)) at the slag–iron interface increases with the CaO content in the top slag, as shown in Figure 5. It should be noted that on the left part of Figure 5, it is a schematic diagram of the upper layer desulfurizer entering the blast furnace slag after melting, while on the right part of Figure 5, it is a schematic diagram of some desulfurizers entering the blast furnace slag after a certain test period.

Since the amount of desulfurization agent is constant, the concentration of CaO decreases with the amount of BF slag. The basicity and the melting point of slag decrease with the amount of BF slag. Thus, the sulfide absorption capacity of the slag increases.

4.2. Influence of TiO₂ and VOx

4.2.1. Coexistence Theory Model of the Desulfurization Slag

The results in Table 6 show that the desulfurization slag containing vanadium–titanium oxide is a solid–liquid coexisting phase under the specified experimental temperature, except for Samples 1# and 4#. With regard to the XRD results, Figure 3 shows that the solid phases present in the liquid phase are mainly CaO, MgO and CaTiO₃. Some of the high melting point ingredients, such as CaO and CaS, are dissolved in the liquid phase and the desulfurization reaction occurs at the slag–iron interface. At present, it is difficult to propose the corresponding slag model due to the complexity of its solid–liquid phase components for the solid–liquid coexisting phase slag. Considering that the basicity, Al₂O₃, MgO and other major components of the slag designed for this experiment remain stable, the influence of the solid phase is assumed to be consistent, and the analysis is carried out with an all-liquid-phase slag.

According to the ion and molecule coexistence theory (IMCT) [17,18], the simple molecules in the slag are SiO₂, Al₂O₃, TiO₂ and V₂O₃, and simple ions are Ca²⁺, Mg²⁺, Fe²⁺, O^2– and S^2–. Based on the equilibrium constants of complex oxide reactions, and considering the compositions of slag used in this experiment, the complex molecules are mainly MgO·V₂O₃, 2CaO·SiO₂, CaO·SiO₂, 3CaO·Al₂O₃, 2MgO·SiO₂, CaO·TiO₂, 2MgO·TiO₂, 2FeO·TiO₂, 2CaO·Al₂O₃·SiO₂ and 3CaO·MgO·2SiO₂.

IMCT theory of the slag structure is based on the relative molar concentration of the components during the activity. For simple ions, molecules and complex molecules, the “activity” (N_i) can be calculated using Equation (1). However, the “activity” of ion couples, such as (Ca²⁺ + O^2–), should be calculated using Equation (2) [19,20,21,22,23,24,25]. Components and their activity expressions of the slag and reaction and equilibrium constant expression of molten slag are shown in

Table 7. Components and their activity expressions of the slag.

Component	Mole	“Activity”
Ca2+ + O2–	n1 = nCaO	N1 = 2n1/Σni = NCaO
Mg2+ +O2–	n2 = nMgO	N2 = 2n2/Σni = NMgO
Fe2+ + O2–	n3 = nFeO	N3 = 2n3/Σni = NFeO
SiO2	n4 = nSiO2	N4 = n4/Σni = NSiO2
Al2O3	n5 = nAl2O3	N5 = n5/Σni = NAl2O3
TiO2	n6 = nTiO2	N6 = n6/Σni = NTiO2
V2O3	n7 = nV2O3	N7 = n7/Σni = NV2O3
MgO·V2O3	n8 = nMgO·V2O3	N8 = n8/Σni = NMgO·V2O3
2CaO·SiO2	n9 = n2CaO·SiO2	N9 = n9/Σni = N2CaO·SiO2
CaO·SiO2	n10 = nCaO·SiO2	N10 = n10/Σni = NCaO·SiO2
3CaO·Al2O3	n11 = n3CaO·Al2O3	N11 = n11/Σni = N3CaO·Al2O3
2MgO·SiO2	n12 = n2MgO·SiO2	N12 = n12/Σni = N2MgO·SiO2
CaO·TiO2	n13 = nCaO·TiO2	N13 = n13/Σni = NCaO·TiO2
2MgO·TiO2	n14 = n2MgO·TiO2	N14 = n14/Σni = N2MgO·TiO2
2FeO·TiO2	n15 = n2FeO·TiO2	N15 = n15/Σni = N2FeO·TiO2
2CaO·Al2O3·SiO2	n16 = n2CaO·Al2O3·SiO2	N16 = n16/Σni = N2CaO·Al2O3·SiO2
3CaO·MgO·2SiO2	n17 = n3CaO·MgO·2SiO2	N17 = n17/Σni = N3CaO·MgO·2SiO2

and

Table 8. Reaction and equilibrium constant expression of molten slag.

Reaction	ΔG°/(J·mol−1)	Equilibrium Equation
(Mg2+ + O2–) + (V2O3) = (MgO·V2O3)	–21,416 + 5.54T [25]	K1 = N8/(N2·N7)
2(Ca2+ + O2–) + (SiO2) = (2CaO·SiO2)	–102,090 − 24.27T [23]	K2 = N9/(N12·N4)
(Ca2+ + O2–) + (SiO2) = (CaO·SiO2)	–21,757 − 36.82T [23]	K3 = N10/(N1·N4)
3(Ca2+ + O2–) + (Al2O3) = (3CaO·Al2O3)	–21,771 − 29.31T [20]	K4 = N11/(N13·N5)
2(Mg2+ + O2–) + (SiO2) = (2MgO·SiO2)	–56,902 − 3.35T [21]	K5 = N12/(N22·N4)
(Ca2+ + O2–) + (TiO2) = (CaO·TiO2)	–79,900 − 3.35T [24]	K6 = N13/(N1·N6)
2(Mg2+ + O2–) + (TiO2) = (2MgO·TiO2)	–25,500 + 1.26T [24]	K7 = N14/(N22·N6)
2(Fe2+ + O2–) + (TiO2) = (2FeO·TiO2)	–33,913 + 5.86T [24]	K8 = N15/(N32·N6)
2(Ca2+ + O2–) + (Al2O3) + (SiO2) = (2CaO·Al2O3·SiO2)	–116,315 − 38.91T [21]	K9 = N16/(N12·N5·N4)
3(Ca2+ + O2–) + (Mg2+ + O2–) + 2(SiO2) = (3CaO·MgO·2SiO2)	–205,016 − 31.80T [21]	K10 = N17/(N13·N2·N42)

, respectively.

$$N_i={\displaystyle \frac{n_i}{{\displaystyle \sum n_i}}}$$

(1)

$$N_{(MeO)}=N_{(M{e}^{2+},MeO)}+N_{(O^{2-},MeO)}={\displaystyle \frac{2n_{(MeO)}}{{\displaystyle \sum n_i}}}$$

(2)

where N_i is the “activity” of i; n_i is the moles number of i.

We assume that the amount of each component of the material in the slag is expressed in b_i before the formation of the complex oxide, and the amount of each component is expressed in n_i after the end of the reaction. Thus, the formulas are shown in the following:

$$b_{CaO}=n_{CaO}+n_{CaO·Ti{O}_2}+n_{CaO·\mathrm{S}i{O}_2}+2n_{2CaO·Si{O}_2}+3n_{3CaO·A{l}_2{O}_3}+2n_{2CaO·A{l}_2{O}_3·Si{O}_2}+3n_{3CaO·MgO·2Si{O}_2}$$

(3)

$$b_{Si{O}_2}=n_{Si{O}_2}+n_{CaO·\mathrm{S}i{O}_2}+n_{2CaO·\mathrm{S}i{O}_2}+n_{2MgO·\mathrm{S}i{O}_2}+n_{2CaO·A{l}_2{O}_3·\mathrm{S}i{O}_2}+2n_{3CaO·MgO·2\mathrm{S}i{O}_2}$$

(4)

$$b_{MgO}=n_{MgO}+n_{MgO·V_2{O}_3}+2n_{2MgO·\mathrm{S}i{O}_2}+2n_{2MgO·Ti{O}_2}+n_{3CaO·MgO·2\mathrm{S}i{O}_2}$$

(5)

$$b_{A{l}_2{O}_3}=n_{A{l}_2{O}_3}+n_{3CaO·{\mathrm{Al}}_2{O}_3}+n_{2CaO·{\mathrm{Al}}_2{O}_3·Si{O}_2}$$

(6)

$$b_{Ti{O}_2}=n_{Ti{O}_2}+n_{CaO·Ti{O}_2}+n_{2\mathrm{Mg}O·Ti{O}_2}+n_{2FeO·Ti{O}_2}$$

(7)

$$b_{V_2{O}_3}=n_{V_2{O}_3}+n_{2MgO·V_2{O}_3}$$

(8)

$$b_{FeO}=n_{FeO}+2n_{2FeO·Ti{O}_2}$$

(9)

$${\displaystyle {\sum }_{i=1}^{17}{N}_i}=1$$

(10)

$$K_i=\exp \left(\Delta {{\displaystyle G}}_i^0/RT\right)$$

(11)

The free SiO₂, TiO₂ and V₂O₃ content of the desulfurization slag is extremely low and all of them are in the form of complex oxides such as CaO·SiO₂, MgO·V₂O₃ and CaO·TiO₂. Combining with Table 8 and Equations (1)–(11), N_i is calculated from the composition of the experimental desulfurization slag given in Table 6. According to the calculation for 100 g of slag, the relationship between the “activity” (N_i) and the moles number (n_i) in the slag is shown in Figure 6. There is a good linear relationship between “activity” and the molar number of components after reaction equilibrium.

The relationship between the activity of the main components of slag and the content of TiO₂ and V₂O₃ is shown in Figure 7. With increasing TiO₂ content, the activity of CaO·TiO₂ and 2MgO·TiO₂ also increases, while the activity of free CaO decreases. With the increase in V₂O₃ content, the activity of MgO·V₂O₃ increases, while the activity of free CaO and MgO slightly decreases. TiO₂ reacts with free CaO and MgO to form CaO·TiO₂ and 2MgO·TiO₂ (at 1623 K), which consumes free CaO and MgO, and thus reduces the desulfurization ability of the top slag. V₂O₃ only reacts with a small amount of MgO to produce MgO·V₂O₃, so its effect on the desulfurization ability of the top slag is small.

4.2.2. Effect of TiO₂ and VO_x on L_S

As shown in Figure 3, it can be seen that the desulfurization products were all CaS. The desulfurizing reaction at the interface of slag–iron is given in Equation (12) [26].

[S] + (Ca²⁺ + O²⁻)_slag = [O] + (Ca²⁺ + S²⁻)_slag

(12)

The slag–iron sulfur distribution ratio (L_S) can be calculated from the molecule–ion coexistence theory of slag structure, as described in Equations (13) and (14):

$$\begin{array}{ll}{K}_{CaS}& ={\displaystyle \frac{a_{(CaS)}}{a_{(CaO)}}}\times {\displaystyle \frac{a_{[O]}}{a_{[S]}}}={\displaystyle \frac{N_{(CaS)}}{N_{(CaO)}}}\times {\displaystyle \frac{a_{[O]}}{a_{[S]}}}\\ & ={\displaystyle \frac{2(\%S)/32/{\displaystyle \sum n_i}}{N_{(CaO)}}}\times {\displaystyle \frac{a_{[O]}}{f_{[S]}\times [\%S]}}\\ & ={\displaystyle \frac{1}{16N_{CaO}\times {\displaystyle \sum n_i}}}\times {\displaystyle \frac{(\%S)}{[\%S]}}\times {\displaystyle \frac{a_{[O]}}{f_{[S]}}}\end{array}$$

(13)

$$L_S={\displaystyle \frac{(\%S)}{[\%S]}}=K_{CaS}{\displaystyle \frac{16N_{(CaO)}\times {\displaystyle \sum n_i}}{a_{[O]}}}\times f_{[S]}$$

(14)

The relationship between the equilibrium constant and temperature is given in Equation (15) [27].

$$\ln K_{CaS}=3.455-{\displaystyle \frac{12723}{T}}$$

(15)

Given a carbon content of about 4.3 wt.% and a silicon content of about 0.08 wt.% in the hot metal, the activity of [O] is controlled by the carbon–oxygen reaction. The calculation formulas are shown in Equations (16) and (17) [28].

$$[C]+[O]=C{O}_{(g)}, \Delta G_{CO}^o= -23720- 33.91T (\mathrm{J}·{\mathrm{mol}}^{-1})$$

(16)

$$a_{[O]}={\displaystyle \frac{1}{e^{-\frac{\Delta G_{CO}^{°}}{RT}}}}\times {\displaystyle \frac{1}{w{[C]}_{\%}}}$$

(17)

where [C] is in the carbon-saturated iron liquid, and [O] is in the 1 wt.% iron solution as the standard state.

The activity coefficient of sulfur was calculated using Equation (18), and the activity interaction coefficient is shown in

Table 9. Activity interaction coefficient of elements in molten iron [5,29].

	C	Si	Mn	P	S	V	Ti
$e_S^j$	0.11	0.063	–0.026	0.029	–0.028	–0.016	–0.072

.

$$\lg f_i={\displaystyle \sum _i^{n}w{[j]}_{\%}\times e_i^j}$$

(18)

where $w{[j]}_{\%}$ is the mass fraction of component j in molten iron, wt.%; $e_i^j$ is the activity interaction coefficient of component j to component i.

Based on the composition of the molten iron, the equilibrium L_S at the slag–iron interface is calculated. The calculated value is compared with the experimental result, as shown in Figure 8.

Ls gradually decreases with increasing contents of V₂O₃ and TiO₂. This is true for both the experimental and calculated value of L_S. However, the calculated value is greater than the experimental value. The reactions of vanadium–titanium oxides with CaO consume the free CaO content. This is in agreement with the findings of Dong et al. [30], where titanium oxides act as acid oxides in the form of TiO₆⁸⁻ in a high basicity slag. It is logical that vanadium oxides are acidic. This is also supported by the XRD results in Figure 3.

The calculated value of the sulfur partition ratio is larger than that of the experimental value, which may be due to the larger than average sulfur content of the slag near the slag–iron interface, and the slower mass transfer of sulfur in a slag containing vanadium and titanium oxides.

Another explanation for Figure 8 maybe that the calculation process also considers the solid-phase CaO in desulfurization. However, the experimental value shows that this portion of CaO does not have desulfurization ability at the slag–iron interface. On the other hand, some desulfurization products such as CaS exist in solid phase and are more retained at the interface. The actual amount of CaS in the upper part of the top slag is small, resulting in lower actual test results. However, the calculation results are consistent with the test results in terms of the influence of top slag content and vanadium–titanium oxide content on desulfurization. This can also provide guidance to real production.

5. Conclusions

By means of experiments and the ion–molecular coexistence theory, this paper investigates the effect of the amount of blast furnace slag and vanadium–titanium oxides content in top slag on the influence of hot metal pre-desulfurization. The main conclusions are as follows:

I. The top slag of CaO-SiO₂-Al₂O₃-MgO-TiO₂-VO_x, composed of blast furnace slag and CaO desulfurization agent, has a certain capacity of desulfurization. The slag can reduce the sulfur content in molten iron from 0.08 wt.% to 0.02 wt.%. This result shows that the top slag has desulfurization capacity at pre-desulfurization process.

II. The desulfurization capacity of the top slag is reduced with increasing VO_x and TiO₂ content. Titanium oxides react with CaO in the slag to form CaTiO₃, thereby reducing the amount of free CaO in the slag content. Vanadium oxides in the pre-desulfurization slag are acidic and react with MgO to form MgV₂O₄, which consumes free CaO indirectly. The vanadium–titanium oxides in the top slag can be reduced by reducing the amount of blast furnace slag, and increasing the injection amount of desulfurization agent.