Effect of Si on Desulfurization in Fe–Si–S, Fe–Si–Cr–S and Fe–Si–Ni–S Melts

Abstract

The effect of silicon content varied from 0 to 45 mass% on the desulfurization in Fe–Si–S, Fe–Si–Cr–S, and Fe–Si–Ni–S melts was investigated at 1873 K (1600°C). It was found that sulfur can be remarkably removed when the silicon content is more than 30 mass%. From the thermodynamic analysis, it was concluded that sulfur may be primarily evaporated as SiS(g), and the activities of silicon and activity coefficient of sulfur in alloys will increase as increasing the silicon content, which makes the evaporation of SiS(g) more feasible and more sulfur removed.

1. Introduction

During the productions of stainless steel, lots of pickling sludge are generated, and the valuable metals such as Fe, Cr, and Ni contained in the pickling sludge are worthwhile to recover. However, the high sulfur content of the produced alloy by the method of directly carbothemic reduction is a big problem.^1,2) Furthermore, in the industrial practices, it is found that the sulfur contents of high silicon bearing ferroalloys such as ferrosilicon and ferrosilicochromium, as well as the low carbon ferroalloy prepared by using silicon bearing ferroalloys as the raw material, such as low carbon ferroalloy, are always very low, even though the basicity of slag during the melting process is very low, and no extra desulphurization process is conducted. The high content of silicon may be the cause of a low sulfur content of alloy melts. Some investigations^3,4,5) have also indicated that silicon is a positive element for desulfurization, and the reaction [Si]+[S]=SiS(g) can take place in Si-containing liquid alloys.^6,7,8,9) However, most of the researches^{3,4,5,6,7,8,9)} were carried out on Fe–Si–S alloys in a low content of silicon. Meanwhile, there are few studies about the alloy systems containing Cr and Ni. Whether it is feasible to remove sulfur from Fe–Si–S, Fe–Si–Cr–S, and Fe–Si–Ni–S melts only by increasing the silicon content is unclear. Therefore, in the present work, the effects of silicon content varied from 0 to 45 mass% on desulfurization in Fe–Si–S, Fe–Si–Cr–S, Fe–Si–Ni–S liquid alloys will be investigated.

2. Experimental

The compositions of alloys are shown in Table 1, and the alloys are classified into three different groups: groups A, B and C are Fe–Si–S, Fe–Si–Cr–S and Fe–Ni–Si–S system, respectively. In groups A, B, and C, the initial content of sulfur is about 0.71 mass%; in groups B and C, mass ratio of Cr to Fe, Ni to Fe are set as 2/3. Fe powder (purity > 98.35 mass%), FeS powder (purity > 99 mass%), Ni powder (purity > 99.8 mass%), Cr powder (purity > 99.9 mass%), and Si powder (purity > 99.99 mass%) were used as the raw materials. According to the composition of alloys, the powders were accurately weighed, and homogeneously mixed in an agate mortar. Then about 5 g mixtures were put into an alumina crucible with the diameter of about 15 mm. The schematic diagram of the experimental apparatus was shown in Fig. 1. When the furnace was heated to 1873 K (1600°C) under a flowing argon atmosphere, alumina crucible together with the sample was put into the constant temperature zone and reacted for 1 h. Furthermore, the influences of holding time for three systems were investigated. Samples after reaction were cooled down to the room temperature in the argon atmosphere after being taken out from the furnace quickly. The final S content of sample was analyzed by using an infrared carbon-sulfur analyzer (EMIA-920V2, HORIBA, Japan).

Table 1. The compositions of alloys and S content after reaction.

Samples	Initial composition					After reaction [%S]
	[%Fe]0	[%Cr]0	[%Ni]0	[%Si]0	[%S]0
A0	99.29	–	–	0	0.71	0.71000
A1	94.29	–	–	5	0.71	0.65000
A2	84.29	–	–	15	0.71	0.38075
A3	69.29	–	–	30	0.71	0.00480
A4	54.29	–	–	45	0.71	0.00660
B0	59.57	39.72	–	0	0.71	0.66411
B1	56.57	37.72	–	5	0.71	0.67548
B2	50.57	33.72	–	15	0.71	0.66031
B3	41.57	27.72	–	30	0.71	0.00357
B4	32.57	21.72	–	45	0.71	0.00476
C0	59.57	–	39.72	0	0.71	0.71581
C1	56.57	–	37.72	5	0.71	0.69561
C2	50.57	–	33.72	15	0.71	0.46912
C3	41.57	–	27.72	30	0.71	0.00438
C4	32.57	–	21.72	45	0.71	0.00268

Fig. 1.

The schematic diagram of the experimental apparatus. (Online version in color.)

3. Results

Since the holding time required for the equilibration of desulfurization should be dependent on the composition of alloy, three typical compositions (A2, B2, C2) in Fe–Si–S, Fe–Cr–Si–S and Fe–Ni–Si–S systems with the silicon content of 15 mass% were selected to investigate the influences of holding time and composition. According to Fig. 2, it can be seen that when Si content is 15 mass%, the sulfur content in Fe–Si–S, Fe–Si–Cr–S, and Fe–Si–Ni–S alloys almost keep constant as the reaction time is 40 min. For the alloys with a silicon content higher than 15 mass%, it can be deduced that a shorter equilibration time will be required (less than 40 min) due to the higher reaction kinetics resulted from the higher silicon content. Considering that the equilibration time for alloys with a silicon content of 5 mass% may be longer, a holding time of 60 min was selected in the current study.

Fig. 2.

Changes of S content in alloys (A2, B2, C2) with the holding time at 1873 K (1600°C). (Online version in color.)

The S contents of different samples after holding for 60 min are shown in Table 1. The changes of S content with the initial Si content of alloys are given in Fig. 3. From which it can be seen that in groups A and C, the final content of S decreases as increasing the initial Si content from 0 to 15 mass%; but in group B, the increase of Si content in this range has little effect on the final S content. When the Si content increases to 30 mass%, the final S contents of A3, B3, and C3 are very low compared with the silicon-free samples A0, B0, and C0, respectively. As further increasing the initial Si content to 45 mass%, the S contents of A4, B4, and C4 are very low and almost keep constant. Accordingly, it is feasible to remove sulfur from Fe–Si–S, Fe–Si–Cr–S, and Fe–Si–Ni–S liquid melts by increasing the Si content.

Fig. 3.

Changes of S content in alloys with the initial Si content at 1873 K (1600°C). (Online version in color.)

4. Discussion

From Fig. 2, According to the previous studies,^10,11,12,13) the sulfur can evaporate from liquid alloys as SiS(g), S(g), and S₂(g), and the reactions can be described with Eqs. (1), (2), (3).   

$$[\text{Si}]+[\text{S}]=\text{SiS(g)}$$(1)

  

$$[\mathrm{S}]=\text{S}(\mathrm{g})$$(2)

  

$$\text{2[S]}={\text{S}}_2(\mathrm{g})$$(3)

  

$$K_1=\frac{P_{\text{SiS}}}{a_{\text{Si}}{a}_{\text{S}}}=\frac{P_{\text{SiS}}}{a_{\text{Si}}{f}_{\text{S}}[\%\mathrm{S}]},\hspace{1em}[\%\mathrm{S}]=\frac{P_{\text{SiS}}}{K_1{a}_{\text{Si}}{f}_{\text{S}}}$$(4)

According to the work of Fruehan,¹⁰⁾ the equilibrium vapor pressure of SiS(g) is much greater than that of S(g) and S₂(g), and the evaporation of SiS(g) is the primary way of the desulfurization. In Eq. (4), where K₁ is the equilibrium constant of reactions (1) at 1873 K (1600°C); P_SiS is the equilibrium vapor pressure of SiS(g); a_Si and a_S are the activity of Si and S, respectively; f_S is the activity coefficient of sulfur; [%S] is the sulfur content.

It should be point out that, because of the use of argon as protecting gas, the vapor pressure of SiS(g) is difficult to be determined. Thus, in the following part of thermodynamic calculation, it is not the aim of the current study to compare the equilibrium value with the experimentally measured data. However, from Eq. (1), it can be seen that the final sulfur content will be decreased by increasing the activity of silicon and activity coefficient of sulfur, which may qualitatively give some indication for the desulphurization process. Next, the variations of them will be discussed in detail.

When the Si content is high, the high order interaction parameters are necessary for calculating the activity of Si. Unfortunately, there are few thermodynamic investigations on the Fe–Si–S, Fe–Si–Ni–S, Fe–Si–Cr–S systems, and the corresponding parameters to calculate the influence of sulfur on the activity of silicon of these systems are not available. However, in the current study, the contents of Si, Cr and Ni are much higher than of the content of the S, thus the main factor affecting the silicon activity may be Si, Cr and Ni, relative to S. Therefore, the effect of sulfur on the activity of silicon was ignored, and the Fe–Si–S, Fe–Si–Cr–S, and Fe–Si–Ni–S alloys are simplified to Fe–Si, Fe–Si–Cr, and Fe–Si–Ni alloys, respectively, as calculating the activity of silicon.

Fe–Si liquid melts are widely considered as the regular solution, and the activity coefficient of Si can be calculated by Eq. (6).¹⁴⁾   

$$a_{\text{Si}}={\gamma }_{\text{Si}}\times x_{\text{Si}}$$(5)

  

$$lg{\gamma }_{\text{Si}}=-2.12{\left(1-x_{\text{Si}}\right)}^2-0.22$$(6)

where γ_Si and x_Si are the activity coefficient of silicon and the mole fraction of silicon, respectively. The activity of silicon in Fe–Si–Cr alloy was investigated at 1873 K (1600°C) by Slye et al.,¹⁵⁾ it was found that the effect of chrome on activity of silicon is not significant when silicon content is about 0.5 mass%. However, in the present work, the silicon content is high, and Eq. (7) by the theories of Shi¹⁶⁾ is used to calculate the activity coefficient of silicon in Fe–Si–Cr alloy at 1873 K (1600°C).   

$$ln{\gamma }_{\text{Si}}=\frac{-13.2008y-8.3997}{1+y}{(1-x_{\text{Si}})}^2$$(7)

where y and x_Si are the mole ratio of Fe/Cr and mole fraction of silicon, respectively. The activity coefficient of silicon in Fe–Si–Ni liquid alloy can be calculated with the non-regular ternary Margules expansion described by Eq. (8).¹⁷⁾   

$$\begin{array}{l}RTln{\gamma }_i=2W_{ji}{x}_i{x}_j(1-x_i)+W_{ij}(x_j^2-2x_i{x}_j^2)\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}+W_{kj}{x}_i{x}_k(1-x_i)+W_{ik}(x_k^2-2x_i{x}_k^2)-2W_{kj}{x}_j^2{x}_k\\ \hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}-2W_{jk}{x}_j{x}_k^2+W_0(x_j{x}_k-2x_i{x}_j{x}_k)\end{array}$$(8)

where W is Margules parameter; i, j and k represent elements Si, Fe and Ni, respectively; x is the mole fraction of component. The values of Margules parameters are shown in Table 2.

Table 2. Values of Margules parameters (kJ/mol).¹⁷⁾

WSiFe	WNiFe	WSiNi	WFeNi	WFeSi	WNiSi	W0
−98.425	−8.267	−159.634	−15.375	−93.799	−220.862	−195.352

According to reaction (1), there may have some silicon evaporates as SiS(g). It is assumed that the silicon only consumed by the evaporation of SiS(g), and the mole ratio of evaporated Si and S is 1/1. Based on this, the Si content after reaction can be calculated and used on the calculation of γ_Si in Eqs. (6), (7), (8), and it is assumed that elements Cr and Ni have no losses after reaction. The activities of silicon in Fe–Si, Fe–Si–Cr, and Fe–Si–Ni alloys are eventually calculated by Eq. (5) and shown in Fig. 4(a).

Fig. 4.

Values of a_Si and lgf_S in melts after reaction. (Online version in color.)

In Fe–Si–S system, the activity coefficient f_s is approximately calculated by Wagner equation Eq. (9),   

$$\begin{array}{l}\text{lg}{f}_{\text{S}}=e_{\text{S}}^{\text{Si}}\text{[\%Si]}+e_{\text{S}}^{\text{S}}\text{[\%S]}+r_{\text{S}}^{\text{Si}}{\text{[\%S]}}^2+r_{\text{S}}^{\text{S}}{\text{[\%Si]}}^2\\ \hspace{1em}\hspace{1em}+r_{\text{S}}^{\text{S,Si}}\text{[\%Si][\%S]}\end{array}$$(9)

where $e_S^{Si}$, $e_S^S$, $r_S^{Si}$, $r_S^S$, $r_S^{Si,\mathrm{\hspace{0.17em} }S}$ are the interaction parameters. The quadratic terms of [%S] is omitted because the content of S is low, then Eq. (9) becomes,   

$$\text{lg}{f}_{\text{S}}=e_{\text{S}}^{\text{Si}}\text{[\%Si]}+e_{\text{S}}^{\text{S}}\text{[\%S]}+r_{\text{S}}^{\text{S}}{\text{[\%Si]}}^2+r_{\text{S}}^{\text{S,Si}}\text{[\%Si][\%S]}$$(10)

Then, the values of lgf_S in Fe–Ni–Si–S and Fe–Si–Cr–S alloy systems are approximately calculated by Eqs. (11) and (12), respectively,   

$$\begin{array}{l}\text{lg}{f}_{\text{S}}=e_{\text{S}}^{\text{Si}}\text{[\%Si]}+e_{\text{S}}^{\text{S}}\text{[\%S]}+e_{\text{S}}^{\text{Ni}}\text{[\%Ni]}+{\gamma }_{\text{S}}^{\text{Ni}}{\text{[\%Ni]}}^{\text{2}}\\ \hspace{1em}\hspace{1em}+r_{\text{S}}^{\text{S}}{\text{[\%Si]}}^2+r_{\text{S}}^{\text{S,Si}}\text{[\%Si][\%S]}\end{array}$$(11)

  

$$\begin{array}{l}\text{lg}{f}_{\text{S}}=e_{\text{S}}^{\text{Si}}\text{[\%Si]}+e_{\text{S}}^{\text{S}}\text{[\%S]}+e_{\text{S}}^{\text{Cr}}\text{[\%Cr]}+r_{\text{S}}^{\text{Cr}}{\text{[\%Cr]}}^2\\ \hspace{1em}\hspace{1em}+r_{\text{S}}^{\text{S}}{\text{[\%Si]}}^2+r_{\text{S}}^{\text{S,Si}}\text{[\%Si][\%S]}\end{array}$$(12)

The values of S and Si contents after reaction are used to calculate f_S. The values of interaction parameters used in Eqs. (10), (11), (12) are shown in Table 3 and the values of lgf_S are shown in Fig. 4(b).

Table 3. Values of interaction parameters at 1873 K (1600°C).

$e_S^{Si}$	$e_S^S$	$r_S^{Si}$	$r_S^{Si,\mathrm{\hspace{0.17em} }S}$	$e_S^{Cr}$	$r_S^{Cr}$	$e_S^{Ni}$	$r_S^{Ni}$
0.06707)	−0.033718)	0.002919)	−0.00127)	−0.010620)	020)	0.006121)	020)

From Figs. 4(a) and 4(b), we can see that in Fe–Si–Cr–S alloy, both the activity of silicon and activity coefficient of sulfur in Fe–Si–Cr–S alloy are the lowest compared with those in Fe–Si–S and Fe–Si–Ni–S alloys, and the values of a_Si × f_S in Fe–Si–Cr–S alloy system are very low and almost keep constant as increasing the silicon content from 0 to 15 mass% as shown in Table 4. Consequently, the increase of silicon content in that range has little effect on the sulfur content in Fe–Si–Cr–S alloy system as shown in Fig. 3. In addition, from Table 4, at the same initial Si content, the values of a_Si × f_S in Fe–Si–Ni–S alloy system are slightly lower than those in Fe–Si–S alloy system, and by Eq. (4), which can explain that the sulfur content in Fe–Si–Ni–S alloy is slightly higher than that in Fe–Si–S alloy. Furthermore, in all three alloy systems, both the activity of silicon and activity coefficient of sulfur increase as increasing the silicon content, and the increase rate becomes rapid when the silicon content is high, which may be the reason for that the sulfur is remarkably removed when silicon content is up to about 30 mass%. From the Eq. (4) and Table 4, according to the characteristic of inverse function, the decrease rate becomes slow when the value of a_Si × f_S is large enough, which makes the sulfur almost keep constant when silicon content varies from 30 to 45 mass% as shown in Fig. 3.

Table 4. Values of a_Si × f_S in melts after reaction with the initial Si contents.

aSi × fS	[%Si]0=0	[%Si]0=5	[%Si]0=15	[%Si]0=30
Fe–Si–S	0	0.0025	0.4134	1941
Fe–Si–Cr–S	0	≈ 0	0.0096	236.0
Fe–Si–Ni–S	0	0.0013	0.2343	1566

5. Conclusions

Based on the above analyses, it can be concluded that using silicon to remove sulfur from Fe–Si–S, Fe–Si–Cr–S, and Fe–Si–Ni–S system is feasible and efficient, and sulfur can be remarkably removed when the silicon content is more than 30 mass%. The decrease of sulfur content in alloys is well explained by the increases of Si activity and activity coefficient of sulfur as increasing the silicon content.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (51734002).

References

• 1)   J. X.  Zhao,  Z. Y.  Zhao,  R. M.  Shi,  X. M.  Li and  Y. R.  Cui: JOM, 70 (2018), 2825.
• 2)   M.  Ito,  M.  Yoshioka,  Y.  Seino,  M.  Suzuki,  M.  Sakuta,  Y.  Maki and  Y.  Kawabata: ISIJ Int., 37 (1997), 391.
• 3)   J. P.  Morris and  A. J.  Williams: Trans. Am. Soc. Met., 41 (1949), 1425.
• 4)   C. W.  Sherman,  H. I.  Elvander and  J.  Chipman: JOM, 2 (1950), 334.
• 5)   C. W.  Sherman and  J.  Chipman: JOM, 4 (1952), 597.
• 6)   R. J.  Fruehan and  E. T.  Turkdogan: Metall. Trans., 2 (1971), 895.
• 7)   H.  Shoji and  U.  Tatsuji: Bull. Nagoya Inst. Technol., 27 (1975), 249 (in Japanese).
• 8)   R. J.  Fruehan: Metall. Trans. B, 9 (1978), 287.
• 9)   G. R.  Belton,  R. J.  Fruehan and  E. T.  Turkdogan: Metall. Trans., 3 (1972), 596.
• 10)   Y.  Kobayashi,  H.  Todoroki and  K.  Mizuno: ISIJ Int., 59 (2019), 277.
• 11)   J.  Lee and  K.  Morita: ISIJ Int., 42 (2002), 588.
• 12)   R.  Nakao,  S.  Fukumoto,  M.  Fuji and  H.  Takeuchi: ISIJ Int., 32 (1992), 685.
• 13)   Y.  Miyazaki,  F.  Kirihara and  H.  Todoroki: ISIJ Int., 59 (2019), 1205.
• 14)   R. J.  Fruehan: Metall. Trans., 1 (1970), 865.
• 15)   W.  Slye and  R.  Fruehan: 57th Electric Furnace Conf. Proc., Iron and Steel Society, Pittsburgh, PA, (1999), 401.
• 16)   Z. C.  Shi: J. Iron Steel Res., 3 (1991), 17 (in Chinese).
• 17)   J.  Matas,  Y.  Ricard,  L.  Lemelle and  F.  Guyot: Contrib. Mineral. Petrol., 140 (2000), 73.
• 18)   S.  Ikada,  S.  Hayashi and  T.  Uno: Tetsu-to-Hagané, 61 (1975), 2321 (in Japanese).
• 19)   C. H. P.  Lupis and  J. F.  Elliott: Acta Metall., 14 (1966), 1019.
• 20)   G. K.  Sigworth and  J. F.  Elliott: Met. Sci., 8 (1974), 298.
• 21)   T. H.  Ju,  X. Y.  Ding,  Y. Y.  Zhang,  W. L.  Chen,  X. K.  Cheng,  B.  Wang,  J. X.  Dai and  X. L.  Yan: Entropy, 20 (2018), 808.