Hot Metal Desulfurization Behavior with Dolomite Flux

Abstract

Hot metal desulfurization behavior with dolomite flux was examined in heat treatment tests of pellets and in small-scale and commercial-scale hot metal desulfurization tests by mechanical stirring. In the heat treatment tests of pellets, the Mg gas generation ratio at 1673 K increased as the CaO/MgO ratio increased. The Mg gas generation ratio reached 90% with dolomite, which was 1.4 times higher than that with CaO + MgO at the same CaO/MgO. With calcined dolomite and Al ash, CaO efficiency for hot metal desulfurization at 1673 K was 20–30%, which was around double the results with CaO + MgO and CaO. In the desulfurization slag, sulfur was concentrated not with Mg, but Ca and Al were found in an EPMA analysis and detected as CaS by X-ray diffraction. Desulfurization with calcined dolomite and Al ash is considered to proceed by ① desulfurization by Mg(g) and fixing of MgS as CaS, and ② desulfurization by CaO with Mg de-oxidation. 200 t commercial-scale hot metal desulfurization tests with mechanical stirring were carried out. The obtained Mg efficiency for desulfurization was in the range of 15–25% with Mg consumption of 1–1.3 kg/t as MgO. These results were within the same trend as in other studies using metallic Mg and MgO with Al.

1. Introduction

Recently, the hot metal pretreatment process has become increasingly important, not only to meet demand for high quality steel, but also to reduce steelmaking slag. Various studies have been carried out with the aim of improving the efficiency of hot metal desulfurization in CaO-based flux.^1,2,3,4,5) Fluorite (CaF₂) is well known as an effective additive to CaO-based desulfurization flux.⁶⁾ However, addition of CaF₂ must be avoided in order to recycle slag in materials for roads and construction. To increase desulfurization efficiency, it is important to form a liquidus phase in CaO-based flux without CaF₂. Many studies have been carried out to evaluate the effects of various additives for desulfurization.^7,8,9)

Hot metal desulfurization processes using Mg-containing wire addition^10,11,12) or metallic Mg injection^13,14,15,16) were developed in actual plants. Hot metal desulfurization with Mg-based flux proceeds with gaseous Mg and causes lower temperature drop in the hot metal. CaO is used with Mg to avoid resulfurization due to the instability of MgS in hot metal. A disadvantage of Mg-based flux is its cost. Metallic Mg is relatively expensive compared with CaO-based flux, and its price is changeable.

A pellet form of mixed MgO and Al was proposed^17,18,19) as a cheaper desulfurization flux by using gaseous Mg generated by the reduction of MgO by Al. In this technique, the pellet is immersed in the hot metal using a special lance with carrier gas. However, an appropriate lance design should be developed for practical use in actual plants. In this study, the desulfurization ability of the new dolomite-based flux was examined and discussed based on heat treatment tests and small-scale and commercial-scale hot metal desulfurization tests.

2. Experimental Procedures

2.1. Heat Treatment of Al–CaO–MgO Pellet

In order to confirm the reaction at high temperature with calcined dolomite and Al ash flux, heat treatment experiments with pellets were conducted in a small-scale furnace. Al–CaO–MgO pellets were heated at 1473, 1573 and 1673 K, and their weight changes were measured. Aluminum ash, calcined dolomite, magnesium oxide, and lime were used as sources of Al, CaO, and MgO. The chemical composition and particle size of the flux used in this study are shown in Tables 1 and 2, respectively. The particle sizes of metallic Al and aluminum ash were under 75 μm and 200 μm (average), respectively.

Table 1.Chemical compositions of materials used as desulfurization fluxes.

	CaO	Al2O3	MgO	SiO2	S	M.Al	Ig.loss
Calcined dolomite	63.1	0.34	32.17	0.77	0.01	–	3.2
Lime	98
Magnesium oxide			99
Aluminum ash	1.3	9.2	4.5	17.4	0.11	60.5	–
Metallic Al						99.9

(mass%)

Table 2.Particle size of materials used as desulfurization fluxes.

	Ratio (mass%)
Particle size (μm)	500–355	355–250	250–125	≦125
Flux
Calcined dolomite	11.1	19.0	59.6	10.3
Lime	6.1	12.1	65.5	16.3
Magnesium oxide	9.0	17.6	61.2	12.3

These materials were compounded into the specified compositions and pressed into pellets in a mold (ϕ40 mm×10 mm high). A sample pellet was then set in the furnace and heated at 1473, 1573 or 1673 K for 1 hour in an argon gas atmosphere. The pellets after heat treatment were analyzed by X-ray diffraction in order to identify the crystallographic phases in the pellets. The chemical compositions of pellets before and after heat treatment were analyzed. The experimental conditions are shown in Table 3. In the experiments, the weight ratio of Al/MgO was fixed at 0.45, which is the equivalent ratio of MgO reduction by Al as expressed by Eq. (1). The heat treatment temperature was changed from 1473 K to 1673 K in order to evaluate the effect of temperature on the Mg gas generation ratio. The weight ratio of CaO/MgO was also changed in the range between 0.25 and 1.9 in order to evaluate the effect of the CaO/MgO ratio on the reduction of MgO.   

$$3MgO(s)+2Al(l)=3Mg(g)+A{l}_2{O}_3(s)$$(1)

Table 3.Experimental conditions of heat treatment tests.

Sample		Pellet ϕ40 mm×10 mm
Reaction	Time	1 hour
	Temperature	1473–1673 K
	Atmosphere	Ar
Flux	Material	Calcined dolomite
		Magnesium oxide
		Lime
		Aluminum ash
	Al/MgO	0.45
	CaO/MgO	0.25–1.9

2.2. 70 kg-scale Hot Metal Desulfurization Tests

Hot metal desulfurization tests with mechanical stirring were carried out using a 70 kg-scale high frequency induction furnace. The experimental apparatus and conditions are shown in Fig. 1 and Table 4, respectively. The hot metal was heated and melted at 1673 K, and the carbon and sulfur compositions of the hot metal were adjusted to 4.5–4.7 mass%C and 0.04 mass%S, respectively. A graphite impeller was positioned at the center of the MgO crucible and immersed in the hot metal. The impeller immersion depth, which is defined as the distance from the hot metal surface to the bottom of the impeller, was fixed at 150 mm. In this condition, flux dispersion is the complete dispersion stage, in which the flux is fully dispersed into the hot metal.⁵⁾ The materials in each flux were adequately mixed, and were added when the rotation speed of the impeller reached 700 rpm. Hot metal samples were taken at 1–2 min intervals during the experiment. In addition, slag samples were taken after desulfurization. The slag was analyzed by electron probe X-ray microanalysis (EPMA) and X-ray diffraction in order to observe the sulfur distribution and identify the crystallographic phases in the slag, respectively.

Fig. 1.

Experimental apparatus for 70 kg-scale hot metal experiments.

Table 4.Experimental conditions of 70 kg-scale hot metal desulfurization tests.

Furnace		150 kg IF
Metal		Fe– 4.5–4.7 mass%[C] –0.04 mass%[S] 70 kg
Crucible		Magnesia
Rotation speed		700 rpm
Impeller size	Height	50 mm
	Diameter	100 mm
	Width	25 mm
Impeller immersion depth		150 mm
Temperature		1673 K

The chemical compositions and particle sizes of the fluxes are the same as those used in the heat treatment experiments. The conditions of the fluxes in the hot metal desulfurization tests are shown in Table 5.

Table 5.Conditions of fluxes in 70 kg-scale hot metal desulfurization tests.

No.	Flux	Material content in flux					Content in flux
		Calcined dolomite	CaO (Lime)	Magnesium oxide	Al ash	Metallic Al	CaO	MgO	M.Al
		(kg/t)					(kg/t)
A	Calcined dolomite	4.0					2.5	1.3
B	Calcined dolomite+M.Al	4.0				0.24	2.5	1.3	0.24
C	Calcined dolomite+Al ash	4.0			0.40		2.5	1.3	0.24
D		4.0			0.80		2.5	1.3	0.48
E	CaO+MgO+Al ash		2.6	1.3	0.40		2.5	1.3	0.24
F	CaO+Al ash		4.0		0.40		3.9		0.24

It should be noted that, in this paper, “percent” means mass percent unless otherwise stated. The effect of the CaO and MgO sources can be evaluated by comparing the results of tests C and E. Test F is a conventional CaO-based flux. The effects of the Al sources were evaluated by comparing the result of tests A to D.

3. Results and Discussions

3.1. Heat Treatment of Al–CaO–MgO Pellet

After the heat treatment tests, the pellet weights decreased and the MgO contents in the pellets also decreased. It is considered that these weight changes were caused by the generation of Mg gas. The Mg gas generation ratio was defined by Eq. (2). Here, ΔMgO is the weight loss of MgO in pellets. In Eq. (2), it is assumed that the material involved in weight loss is perfectly consumed by Mg gas generation.   

$$Mg\mathrm{\hspace{0.17em} }gas\mathrm{\hspace{0.17em} }generation\mathrm{\hspace{0.17em} }ratio(\%)=\frac{\mathrm{\Delta }MgO\mathrm{\hspace{0.17em} }for\mathrm{\hspace{0.17em} }Mg\mathrm{\hspace{0.17em} }gas\mathrm{\hspace{0.17em} }(g)}{MgO\mathrm{\hspace{0.17em} }in\mathrm{\hspace{0.17em} }flux\mathrm{\hspace{0.17em} }(g)}\times 100$$(2)

The effect of the heat treatment temperature on the Mg gas generation ratio and the decrease of the MgO contents in the pellets are shown in Fig. 2. ΔMgO was obtained in each case from the MgO content in the pellets before and after heat treatment. Both the Mg gas generation ratio and the decrease of the MgO contents in the pellets increased as the heat treatment temperature increased. It is considered that MgO reduction (Eq. (1)), which is an endothermic reaction, was enhanced by increasing the heat treatment temperature.

Fig. 2.

Effect of heat treatment temperature on decrease of MgO content in pellet and Mg gas generation ratio (heat treatment tests).

The relationship between the CaO/MgO ratio and Mg gas generation ratio is shown in Fig. 3. The Mg gas generation ratio increased as the CaO/MgO ratio was increased up to 90% and became saturated when the CaO/MgO ratio was higher than 0.9.

Fig. 3.

Relationship between CaO/MgO and Mg gas generation ratio (heat treatment tests).

The pellets after heat treatment were examined by X-ray diffraction. 12CaO·7Al₂O₃ and MgO·Al₂O₃ were detected when the CaO/MgO ratio was smaller than 0.9, and 12CaO·7Al₂O₃ and 3CaO·Al₂O₃ were detected when the CaO/MgO ratio was higher than 0.9. From the results of the products after heat treatment, it is considered that the amount of MgO consumed in generating Mg gas decreases by MgO·Al₂O₃ generation when the CaO/MgO ratio is lower than 0.9. At the same CaO/MgO ratio of 1.9, the Mg gas generation ratio with calcined dolomite was higher than those with CaO and MgO. No Al remained after heat treatment when the CaO/MgO ratio was higher than 0.9.

3.2. 70 kg-scale Hot Metal Experiments

3.2.1. Desulfurization Behavior

The effects of the respective Al sources were evaluated by comparing the results of tests A to C. The changes of [S], a_o and [Al] as functions of time in tests A to C are shown in Fig. 4. As shown in Fig. 4, the desulfurization rate in tests B (Al) and C (Al ash) are higher than those in test A (no addition). In tests B and C, the values of [Al] in the early stage of reaction were higher and those of a_o were lower than in test A. From these results, the effect of Al ash is similar to that of Al. In test B with metallic Al, the values of [Mg] were 30 and 10 ppm at 1 and 2.5 minutes after flux addition, respectively. In other tests, [Mg] values were undetectable (≦10 ppm).

Fig. 4.

Changes of [S], a_o and [sol.Al] as function of time (Conditions A, B and C, 70 kg-scale hot metal desulfurization tests).

The effects of the CaO and MgO sources can be evaluated by comparing the results of tests C, E and F. The changes of [S], a_o and [Al] as functions of time in tests C, E and F are shown in Fig. 5. In tests C and E, CaO/MgO and Al/MgO are the same, at 1.9 and 0.18, respectively. Al ash was used as a cheaper Al source, which has the same effect as metallic Al, as shown in Fig. 4. The desulfurization rate in test C (calcined dolomite) is much higher than that in test E (CaO + MgO), even with the same conditions of CaO/MgO and Al/MgO. The desulfurization rate in test C is higher than that in test F, even though the CaO consumption is around half that in test F. In the case of calcined dolomite, [S] decreased to 0.001% in 10 minutes. The levels of [Al] and a_o are similar under these conditions. [Al] increased to 0.006% immediately after addition of the Al source, and after 3 minutes, [Al] decreased to 0%.

Fig. 5.

Changes of [S], a_o and [sol.Al] as function of time (Condition C, E and F, 70 kg-scale hot metal desulfurization tests).

3.2.2. Effects of Flux Composition and Al Consumption on CaO Efficiency

The relationship between CaO consumption and ΔS in tests C, E and F is shown in Fig. 6. The CaO efficiency for desulfurization is defined by Eq. (3), as shown by the dotted lines in Fig. 6.   

$$CaO\mathrm{\hspace{0.17em} }efficiency\mathrm{\hspace{0.17em} }(\%)=\frac{CaO\mathrm{\hspace{0.17em} }that\mathrm{\hspace{0.17em} }became\mathrm{\hspace{0.17em} }CaS\mathrm{\hspace{0.17em} }(kg/t)}{CaO\mathrm{\hspace{0.17em} }in\mathrm{\hspace{0.17em} }flux\mathrm{\hspace{0.17em} }(kg/t)}\times 100$$(3)

Fig. 6.

Relationship between CaO consumption and ΔS (Conditions C, E and F, 70 kg-scale hot metal desulfurization tests).

The amount of desulfurization, ΔS, with CaO and MgO (test E) is less than half of that with calcined dolomite (test C) flux with the same CaO/MgO and Al/MgO. In Fig. 6, ΔS with CaO flux (test F) is the same as that with calcined dolomite flux. However, the CaO consumption of calcined dolomite flux was around half that of CaO. The CaO efficiency of calcined dolomite flux (test C) was around double those of CaO-based fluxes (tests E and F).

CaO efficiency was reported to be about 8% with the conventional CaO-based flux.⁴⁾ CaO efficiency with calcined dolomite flux was 27%, which was significantly higher than that with the conventional CaO-based flux.

The effect of Al consumption on desulfurization behavior with calcined dolomite and CaO fluxes was evaluated by comparing the results of experiments A to D and F. The relationships between Al consumption and CaO efficiency for desulfurization and a_o after flux addition are shown in Fig. 7. The stoichiometric composition of MgO reduction by Al is Al=0.59 kg/t, whereas that of calcined dolomite is Al=4.0 kg/t, respectively. CaO efficiency for desulfurization, defined by Eq. (3), increased with increasing Al consumption. CaO efficiency for desulfurization with CaO + Al ash flux is lower than those with calcined dolomite + Al or Al ash. CaO efficiency for desulfurization with calcined dolomite + Al and calcined dolomite + Al ash were nearly the same under the condition of Al consumption of 0.24 kg/t. In this condition, the a_o levels were the same with both Al and Al ash. There is no significant difference between Al and Al ash as a desulfurization agent.

Fig. 7.

Relationship between Al consumption and CaO efficiency for desulfurization (70 kg-scale hot metal desulfurization tests).

From these results, desulfurization efficiency was enhanced by using calcined dolomite in comparison with CaO and MgO, even at the same CaO/MgO ratio, and desulfurization efficiency with calcined dolomite increased with increasing Al consumption.

3.2.3. Sulfur Distribution in Desulfurization Slag

Figures 8(a) and 8(b) show the observation results of the desulfurization slag in tests C and E, respectively. In Fig. 8(a), sulfur was mainly concentrated at the outer layer of the slag. The thickness of the layer was 50 to 100 μm. In this layer, Mg and Ca were well dispersed. Dispersion of CaO and MgO were also observed inside the layer, which is considered to be unreacted dolomite. In Fig. 8(b), a similar S-enriched layer was observed at the outer layer of the slag. In this case, sulfur was concentrated not with Mg but with Ca.

Fig. 8.

(a) EPMA mapping of desulfurization slag (Conditions C, 70 kg-scale hot metal desulfurization test). (b) EPMA mapping of desulfurization slag (Experiment E, 70 kg-scale hot metal desulfurization test).

Figure 9 shows X-ray diffraction patterns of the desulfurization slag obtained by tests C and E. In test C, 3CaO·Al₂O₃, 12CaO·7Al₂O₃ and MgO·Al₂O₃ were detected as crystallographic phases containing CaO, MgO or Al₂O₃. In both tests C and E, sulfur was fixed not as MgS but as CaS.

Fig. 9.

X-ray diffraction analysis of desulfurization slag (Conditions C and E, 70 kg-scale hot metal desulfurization tests).

3.3. Desulfurization Mechanism with Dolomite Flux

3.3.1. Effect of CaO on Mg Gas Generation

From the results of the heat treatment tests, the reactions of MgO reduction by Al can be expressed by Eqs. (4) and (5) when the CaO/MgO ratio is lower than 0.9. The following Eqs. (4) and (5) show the CaO contribution to Mg gas generation. 3CaO·Al₂O₃ and 12CaO·7Al₂O₃ were detected by X-ray analysis.   

$$\begin{array}{l}4MgO(s)+2Al(l)=3Mg(g)+MgO\cdot A{l}_2{O}_3(s)\\ \hspace{1em}\hspace{1em}\Delta G^o=466-0.287T\mathrm{\hspace{0.17em} }(kJ/mol)\end{array}$$²⁰⁾ (4)

  

$$\begin{array}{l}21MgO(s)+14Al(l)+12CaO(s)\hfill \\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{0.17em} \hspace{1em}=21Mg(g)+12CaO\cdot 7A{l}_2{O}_3(s)\\ \mathrm{\Delta }{G}^o=3\mathrm{\hspace{0.17em} }786\times {10}^3+153.2TlnT-3\mathrm{\hspace{0.17em} }546.4T\mathrm{\hspace{0.17em} }(J/mol)\end{array}$$²¹⁾ (5)

  

$$\begin{array}{l}9MgO(s)+2Al(l)+3CaO(s)=9Mg(g)+3CaO\cdot A{l}_2{O}_3(s)\\ \hspace{1em}\mathrm{\hspace{0.17em} }\mathrm{\hspace{0.17em} }\hspace{1em}\hspace{1em}\mathrm{\Delta }{G}^o=489-0.310T\mathrm{\hspace{0.17em} }(kJ/mol)\end{array}$$²⁰⁾ (6)

According to their Gibbs free energy, the reactions represented by Eqs. (4), (5), (6) could take place at 1673 K. Equation (5) is the most stable among Eqs. (4), (5), (6). Based on these results and thermodynamic considerations, Mg gas generation could be enhanced by a higher CaO/MgO ratio. In the case of higher CaO/MgO, Al₂O₃ can react with CaO. On the other hand, as Al₂O₃ reacts with MgO as shown in Eq. (4), Mg gas generation is restrained by MgO consumption for generation of MgO·Al₂O₃ in the case of lower CaO/MgO.

3.3.2. Effect of Calcined Dolomite

From the results of the heat treatment tests and hot metal desulfurization experiments, the Mg gas generation and desulfurization rates with calcined dolomite are higher than those with CaO+MgO, even though the amounts of CaO and MgO are the same.

The X-ray diffraction pattern of calcined dolomite is shown in Fig. 10. Only CaO and MgO were detected. Figure 11 shows the EPMA mapping of calcined dolomite. MgO and CaO are well dispersed, and their sizes are around 10 μm. It can be observed that CaO and MgO are in good contact in the calcined dolomite. As shown in Fig. 8(a), both CaO and MgO were well dispersed in the S-enriched layer in the case of desulfurization with the calcined dolomite. On the other hand, as shown in Fig. 8(b), MgO was not located with sulfur, but with only CaO in the S-enriched layer in the case of the MgO + CaO flux. The sizes of MgO and CaO in the calcined dolomite are 1/10 as large those in the MgO + CaO fluxes. Mg gas generation (e.g., Eqs. (5), (6)) can be enhanced by the presence of CaO. As the calcined dolomite has a large interfacial area between CaO and MgO, the condition of contact between CaO and MgO has a great effect on the reactions indicated by Eqs. (1), (5) and (6). Dolomite originally has a double salt structure (CaCO₃·MgCO₃),²²⁾ and changes into CaO and MgO independently by baking at over 1073 K.²³⁾ This causes good contact between CaO and MgO in a small scale. From these results, the calcined dolomite has a great effect on both Mg gas and CaS generation.

Fig. 10.

X-ray diffraction analysis of calcined dolomite.

Fig. 11.

EPMA mapping of calcined dolomite.

3.3.3. Desulfurization Mechanism with Dolomite Flux

From the obtained results and discussion, the desulfurization mechanism with calcined dolomite + Al ash flux is estimated to be as follows:

1. Reactions in Mg gas generation by MgO reduction by Al with CaO proceed as shown in Eqs. (7) and (8).   

$$\begin{array}{l}21MgO(s)+14Al(l)+12CaO(s)\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em} \hspace{0.17em}=21Mg(g)+12CaO\cdot 7A{l}_2{O}_3(s)\end{array}$$(7)

  

$$9MgO(s)+2Al(l)+3CaO(s)=9Mg(g)+3CaO\cdot A{l}_2{O}_3(s)$$(8)

Dolomite flux has an advantage for Mg gas generation due to its good contact between CaO and MgO.

2. The desulfurization reaction with Mg gas and CaO is considered to proceed by the following two mechanisms, ① desulfurization by Mg(g), as expressed by Eq. (9), and ② desulfurization by CaO with deoxidization by [Mg], as expressed by Eqs. (10), (11), (12).   

$$Mg(g)+[S]=MgS(s)$$(9)

  

$$CaO(s)+[S]=CaS(s)+[O]$$(10)

  

$$Mg(g)=[Mg]$$(11)

  

$$[Mg]+[O]=MgO(s)$$(12)

① Desulfurization Occurs by Mg (g), and MgS is Fixed as CaS.

In desulfurization slag, only CaS was detected as a desulfurization product. Ender et al.²⁴⁾ investigated the crystal structures of hot metal desulfurization slag by Mg injection, and detected only CaS, and no MgS was observed in the slag. They compared the thermodynamic stabilities of MgS and CaS, and reported that CaS is more stable than MgS in the temperature range of hot metal desulfurization.   

$$MgS(s)+CaO(s)=CaS(s)+MgO(s)$$(13)

The desulfurization product MgS is fixed by CaO as CaS, as shown in Eq. (13).

Yang et al.²⁵⁾ investigated hot metal desulfurization by Mg gas in experiments with a 350 kg-scale furnace. In those experiments, a MgO + Al pellet set in a graphite tube was immersed in hot metal at 1673 K while introducing argon gas. In the early stage of the experiment, 40–80 ppm [Mg] and desulfurization were confirmed. In the later stage of the experiment, the [Mg] concentration and desulfurization rate gradually decreased, and in the last stage, a resulfurization reaction took place. In the experiments, no resulfurization was confirmed in the case of CaO addition from the top. From those results, they estimated that CaO served mainly as a reagent to transform the desulfurization product of MgS into a stable compound of CaS, and as a result, resulfurization was prevented. They also discussed the reaction shown by Eq. (13), and argued that free energy change (–87.28 kJ/mol at 1673 K), which means the reaction in Eq. (13), tends to occur under the conditions of hot metal desulfurization. From those studies, it is considered that the desulfurization product is detected as CaS under hot metal desulfurization conditions due to the instability of MgS.

② Desulfurization by CaO with [Mg] Deoxidation

H. J. Visseri and R. Boom²⁶⁾ reported a hot metal desulfurization mechanism by lime-magnesium injection. From analysis of desulfurization slag, they indicated that desulfurization mechanisms can be understood by both ① and ②. G. A. Irons and C. Celik²⁷⁾ investigated the behavior of hot metal desulfurization by Mg–CaO flux and proposed that desulfurization proceeds through CaO (Eq. (10), and Mg then enhances desulfurization by decreasing the oxygen activity (Eqs. (11) and (12)) around CaO. In this work, [Mg] was detected in test B. In Fig. 7, the oxygen activities in tests B, C and F under the condition of the same Al consumption are nearly the same. In that case, the CaO efficiencies for desulfurization with calcined dolomite (Al ash and Al) were higher than that with CaO and Al ash. Therefore, desulfurization by CaO can be enhanced by Mg deoxidation if the oxygen activity around CaO is locally deoxidized. In the calcined dolomite, MgO and CaO are around 10 μm and are in good contact. Therefore, the oxygen activity around CaO can be decreased when Mg (g) is generated and dissolves in hot metal, which is the same mechanism as that proposed by Irons and Celik.²⁷⁾

In the present work, CaO efficiency for desulfurization under Condition C (calcined dolomite + Al-ash) was 25.7%, which is much higher than that under Condition F (CaO + Al-ash, 10.4%). The dolomite flux has an advantage for both Mg gas generation and desulfurization due to its good contact between CaO and MgO. That difference is considered to be caused by desulfurization through both ① and ②. However, further investigation is needed to clarify the details.

3.4. 200 t Commercial-scale Hot Metal Experiments

200 t commercial-scale hot metal desulfurization tests with the mechanical stirring method were performed using calcined dolomite and Al ash at JFE Steel Corporation, East Japan Works (Keihin District). The experimental conditions are shown in Table 6. The Al/MgO ratio was decided to be 0.19 considering the cost of desulfurization flux. The relationship between CaO consumption and ΔS is shown in Fig. 12. CaO efficiency for desulfurization is defined by Eq. (2). In the experiments, CaO efficiency is around 15%, which is larger than that with CaO–5%CaF₂ flux in commercial (270-KR) plant operation (3.6–9.7%).⁴⁾ Hence, higher efficiency of desulfurization was achieved by using the calcined dolomite and Al ash, which does not contain CaF₂. The CaO efficiency for desulfurization obtained in 70 kg-KR experiments, as shown in Fig. 12, is around 25%, which is higher than that obtained by 200 t-KR experiments, even though the flux compositions are nearly the same in both cases. The hot metal temperature in the 200 t-KR (1583–1643 K) is lower than that in the 70 kg-KR (1673 K) experiments, which has an unfavorable effect on MgO reduction by Al.

Table 6.Experimental conditions of actual plant tests.

Heat size	200 ton /heat
Metal	[S]i = 0.025 – 0.042 mass%
Flux	Mixed Flux
	Al / MgO = 0.19
	(Calcined dolomite + Aluminum ash)
	5.2–6.9 kg/t
Initial temperature	1583–1648 K

Fig. 12.

Relationship between CaO consumption and ΔS (200 t-scale desulfurization tests).

The relationship between Mg consumption and Mg efficiency is shown in Fig. 13. Here, the Mg in the flux is assumed to be the equivalent amount of MgO (W_Mg= 0.6W_MgO). In this experiment, Mg consumption in the flux is equal to 1–1.3 kg/t, and the resulting Mg efficiency for desulfurization was 15–25%.   

$$Mg\mathrm{\hspace{0.17em} }efficiency\mathrm{\hspace{0.17em} }(\%)=\frac{Mg\mathrm{\hspace{0.17em} }that\mathrm{\hspace{0.17em} }became\mathrm{\hspace{0.17em} }MgS\mathrm{\hspace{0.17em} }(kg/t)}{Mg\mathrm{\hspace{0.17em} }in\mathrm{\hspace{0.17em} }flux\mathrm{\hspace{0.17em} }(kg/t)}\times 100$$(14)

Fig. 13.

Comparison of Mg efficiency for desulfurization in various types of hot metal desulfurization.

Here, desulfurization efficiency was evaluated by Eq. (14) for comparison with the results of other investigations of desulfurization with Mg. The obtained Mg efficiency for desulfurization in the 70 kg-KR experiments defined by Eq. (14) was 21–39%. Higher Mg efficiency for desulfurization is supposed to be obtained by due to the higher hot metal temperature compared with those in the 200 t-KR experiments. The results of other research^{11,12,13,14,16,17,18,28,29,30)} are also shown in Fig. 13. As shown in Fig. 13, Mg efficiency for desulfurization decreases with increasing Mg consumption. The data in those studies showed the same trend, even though they were obtained with various methods, such as wire injection, flux injection and MgO + Al. The obtained Mg efficiencies for desulfurization in this work were also within the same trend. Therefore, it can be considered that Mg gas is utilized in desulfurization by the calcined dolomite and Al ash.

4. Conclusions

The results are summarized as follows:

(1) In heat treatment tests both the Mg gas generation ratio and the decrease of the MgO contents in the pellets increased as the heat treatment temperature increased. It is considered that increasing the heat treatment temperature enhanced MgO reduction. The Mg gas generation ratio at 1673 K increased as the CaO/MgO ratio increased and became saturated at CaO/MgO of 1.5. The Mg gas generation ratio reached 90% with dolomite, which was 1.4 times higher than that with CaO + MgO at the same CaO/MgO of 1.9.

(2) In hot metal desulfurization tests in small-scale experiments, with calcined dolomite and Al ash, CaO efficiency for hot metal desulfurization at 1673 K was 20–30%, which was around double the results with CaO + MgO and CaO. In the desulfurization slag, sulfur was not observed with Mg, but Ca and Al were found in an EPMA analysis and detected as CaS by X-ray diffraction.

(3) The Mg gas generation rate caused by MgO reduction by Al was enhanced by the presence of CaO, which can be considered based on the difference of the products, i.e., MgO·Al₂O₃ and 3CaO·Al₂O₃. The effect of calcined dolomite on both Mg gas generation and desulfurization can be explained by considering the microstructures of MgO and CaO.

(4) The desulfurization mechanism with calcined dolomite + Al ash flux is estimated to be as follows:

1. Reactions in Mg gas generation by MgO reduction by Al with CaO

2. Desulfurization reactions with Mg gas and CaO proceeded by ① desulfurization by Mg(g)→MgS fixing as CaS, and ② desulfurization by CaO with [Mg] deoxidation.

The obtained CaO efficiency for desulfurization is 25.7% with calcined dolomite + Al ash, which is much higher than the 10.4% value with CaO + Al ash at the same CaO and Al consumptions.

(5) 200 t commercial-scale hot metal desulfurization tests with mechanical stirring were carried out. The obtained Mg efficiency for desulfurization was in the range of 15–25% with Mg consumption of 1–1.3 kg/t as MgO. Those results were within the same trend as in other studies using metallic Mg and MgO with Al.

References

• 1)   K.  Nakanishi,  N.  Bessyo,  S.  Takada,  A.  Ejima,  M.  Kuga,  J.  Katsuki and  M.  Kawana: Tetsu-to-Hagané, 64 (1978), 1528.
• 2)   N.  Kurokawa,  S.  Matsuo,  H.  Jouguchi,  K.  Yamada and  Y.  Watanabe: Sumitomo Met., 45 (1993), 52.
• 3)   T.  Takaoka,  Y.  Kikuchi and  K.  Yamada: CAMP-ISIJ, 11 (1998), 765.
• 4)   N.  Kikuchi,  S.  Nabeshima,  S.  Takeuchi,  T.  Yamauchi,  Y.  Kitano and  S.  Ogura: Tetsu-to-Hagané, 90 (2004), 322.
• 5)   Y.  Nakai,  I.  Sumi,  H.  Matsuno,  N.  Kikuchi and  Y.  Kishimoto: ISIJ Int., 50 (2010), 403.
• 6)   S.  Maekawa and  K.  Suzuki: Tetsu-to-Hagané, 57 (1971), S60.
• 7)   I.  Shimoda,  T.  Sato,  T.  Nakasuga,  K.  Nakashima and  K.  Mori: Tetsu-to-Hagané, 90 (2004), 401.
• 8)   Y.  Nakamura,  K.  Harashima and  Y.  Fukuda: Tetsu-to-Hagané, 67 (1981), 2138.
• 9)   M.  Ueda,  T.  Iida,  Z.  Morita,  K.  Wakabayashi and  T.  Itoh: Gypsum Lime, 236 (1992), 20.
• 10)   M.  Tateyama,  Y.  Hiraga,  S.  Yaniguchi,  T.  Okimura and  K.  Hirata: SEAISI Q., 29 (2000), 43.
• 11)   M.  Tateyama,  K.  Gennai,  K.  Hirata,  J.  Hirama,  Y.  Hirama and  T.  Okimura: Nisshin Steel Tech. Rep., 76 (1997), 24.
• 12)   Y.  Hiraga,  K.  Gennai,  Y.  Nakasima and  J.  Hirama: CAMP-ISIJ, 9 (1996), 225.
• 13)   T. A.  Engh,  B.  Kristensen and  K.  Venas: Light Metals, TMS-AIME, Warrendale, PA, (1980), 1011.
• 14)   H. J.  Visser and  R.  Boom: ISIJ Int., 46 (2006), 1771.
• 15)   S.  Haimi: Proc. of Scaninject 3, Part.1, MEFOS, Sweden, (1983), 26:1
• 16)   J. P.  Chaussy,  R.  Lecigne,  G.  Denier and  J. P.  Reboul: Dev. Het. Prep. Oxyg. Steelmak., McMaster Univ. Dep. of Metallurgy and Materials Sience, Hamilton, (1983), 148.
• 17)   J.  Yang,  K.  Okumura,  M.  Kuwabara and  M.  Sano: ISIJ Int., 41 (2001), 965.
• 18)   J.  Yang,  K.  Okumura,  M.  Kuwabara and  M.  Sano: ISIJ Int., 42 (2002), 685.
• 19)   S.  Mukawa,  Y.  Ueshima,  M.  Sano,  J.  Yang and  M.  Kuwabara: ISIJ Int., 46 (2006), 1778.
• 20)   E. T.  Turkdogan: Physical Chemistry of High Temperature Technology, Academic Press, New York, (1980), 5.
• 21)   S.  Ban-ya and  M.  Hino: Chemical Properties of Molten Slags, ISIJ, Tokyo, (1991), 17.
• 22)  Gypsum & Lime Handbook, Soc. of Gypsum & Lime, Gihodobobooks, Tokyo, (1972), 146.
• 23)  Gypsum & Lime Handbook, Soc. of Gypsum & Lime, Gihodobobooks, Tokyo, (1972), 152.
• 24)   A.  Ender,  H.  van den Boom,  H.  Kwast and  H.-U.  Lindenberg: Steel Res., 76 (2005), 562.
• 25)   J.  Yang,  M.  Kuwabara,  K.  Okumura and  M.  Sano: ISIJ Int., 45 (2005), 1795.
• 26)   H. J.  Visseri and  R.  Boom: ISIJ Int., 46 (2006), 1771
• 27)   G. A.  Irons and  C.  Celik: Ironmaking Steelmaking, 19 (1992), 136
• 28)   K.  Kimura,  I.  Kikuchi,  S.  Kohira,  Y.  Komatsu,  J.  Fukumi and  T.  Tomita: CAMP-ISIJ, 8 (1995), 105.
• 29)   L.  Brazzoduro,  P.  Buglione,  C.  Borgianni,  A.  Ferretti and  C.  Pietrosanti: Dev. Het. Prep. Oxyg. Steelmak., McMaster Univ. Dep. of Metallurgy and Materials Sience, Hamilton, (1983), 35.
• 30)   K.  Yonezawa,  S.  Sasakawa,  A.  Shinkai,  S.  Kitamura and  S.  Sto: CAMP-ISIJ, 6 (1993), 1070.