Transformation Behaviour of Sulfur from Gypsum Waste (CaSO4·2H2O) while Roasting with Tin-bearing Iron Concentrate for Tin Removal and Iron Recovery

Yong Yu; Lei Li

doi:10.2355/isijinternational.ISIJINT-2020-095

Abstract

Gypsum waste, with CaSO₄·2H₂O as the main ingredient, decomposed into SO₂(g) at a proper temperature, according to previous research. It could sulfurize SnO₂ to SnS(g) in the presence of C or CO(g), based on which a method of roasting with gypsum waste (CaSO₄·2H₂O) for removing Sn from Sn-bearing iron concentrates was put forward in this paper. In addition, a traditional curing agent, FeS₂, was introduced and comparatively studied. With FeS₂ used as the curing agent for tin sulfurization and removal, the FeS generated from the FeS₂ decomposition played a major role through its transformation to SO₂(g) and COS(g) in mixed gases of 35 vol.%CO and 65 vol.%CO₂. Besides, the COS(g) had a better sulfurization effect on tin than SO₂(g). The ‘S’ from gypsum waste (CaSO₄·2H₂O) was mainly transformed into FeS and Ca₃Fe₄S₃O₆ upon roasting with tin-bearing iron concentrate in mixed gases of 35 vol.%CO and 65 vol.%CO₂, both of which could sulfurize Sn(l) to SnS(g) also through their transformation to SO₂(g) and COS(g). The amount of sulfur added by CaSO₄·2H₂O could be decreased compared with that added by FeS₂ due to the higher utilization efficiency of ‘S’ in CaSO₄·2H₂O. This process realized the cyclic utilization of gypsum waste, Sn effective removal and the iron pre-reduction from tin-bearing iron concentrate.

1. Introduction

Tin-bearing iron ore is a complex iron ore resource and its reserve in China exceeds 0.5 billion tons.^1,2) Tin-bearing iron ore is generally processed into a tin-bearing iron concentrate, which contains more than 60 wt.% of iron and 0.3 wt.%–1.2 wt.% of tin, using the traditional magnetic separation method for iron utilization;^3,4) however, this concentrate cannot be used directly in the iron making process due to the overly high tin content (>0.08 wt.%).²⁾ Many studies have been carried out for tin effective removal from tin-bearing iron ore mainly using sulfurization or reduction roasting methods.^2,5,6,7,8,9) For the reduction roasting method, the tin removal in the form of SnO(g) is only approximately 80 wt.%, which is ascribed to the considerable formation of Fe–Sn spinel or Fe–Sn alloy.⁸⁾ Tin separation in the form of SnS(g) exceeds 90% while using pyrite, waste tire rubber, ferrous sulfate or high sulfur coal as additives respectively through the sulfurization process.^2,5,6,7) SO₂(g) and COS(g) can be formed in the sulfurization process above and play significant roles in the transformation of the tin phase to SnS(g) through the gas-solid sulfurization reaction with SnO and SnO₂. Proceeding from this point, materials that can decompose into SO₂(g) and/or COS(g) at a proper temperature can be used as curing agents for tin removal from tin-bearing iron concentrate.

Gypsum waste with CaSO₄·2H₂O as the main ingredient (>90 wt.%) is mainly generated from flue gas desulfurization processing, industrial production, construction, renovation or deconstruction sites, etc.^10,11,12,13) In the world, more than 15 million tons of gypsum waste from plasterboard is sent to landfills each year.^14,15) Its accumulation causes serious environmental problems, such as pollution of water, occupation of land and destruction of ecological systems.^10,11,12) Currently, many studies have been performed to treat gypsum waste to reduce its harm through preparing high-strength building gypsum materials,^16,17) fireproof panels,^18,19) and slag-making material,²⁰⁾ etc. However, the utilization rate of gypsum waste through these methods is low, typically less than 50%.²⁰⁾

The decomposition of CaSO₄ is a function of temperature, O₂ and SO₂ partial pressures.²¹⁾ The thermal degradation of CaSO₄ (Eqs. (1) and (2)) in inert atmosphere occurs at a temperature of 1273 K, and it decreases to 623 K with SO₂(g) being the only gaseous sulfur when CaSO₄ is surrounded by carbon. Under a reducing atmosphere, the decomposition of CaSO₄ becomes a coexistent competing reaction among Eqs. (1), (2), (3) and (4), which simultaneously produces CaO, CaS, SO₂(g) and CO₂(g).^{21,22,23,24,25)}

CaS O 4 →CaO+S O 3 ( g )

(1)

S O 3 ( g ) →S O 2 ( g ) +1/2 O 2 ( g )

(2)

2C+CaS O 4 →CaS+2C O 2 ( g )

(3)

4CO( g ) +CaS O 4 →CaS+4C O 2 ( g )

(4)

The gypsum waste decomposes into SO₂(g) at a proper temperature, which can sulfurize SnO₂ to SnS in the presence of C or CO(g).^{2,5,20,22,24)} A method of Sn removal from Sn-bearing iron concentrates using gypsum waste (CaSO₄·2H₂O) in mixed gases of CO and CO₂ is proposed in this paper. The transformation behaviour of ‘S’ in CaSO₄ during the roasting process was mainly researched via thermodynamic, chemical, ICP, XRD, SEM-EDS and EPMA analyses. Moreover, a traditional curing agent, FeS₂, was introduced and comparatively studied.

2. Materials and Methods

2.1. Materials

The raw material for the tin-bearing iron concentrate, containing approximately 65.75 wt.% of iron and 0.94 wt.% of tin (Table 1), was obtained from an iron-making plant located in the Yunnan Province of China. In this concentrate, iron and tin mainly exist in form of Fe₃O₄ and SnO₂, respectively, and most of SnO₂ is embedded in the Fe₃O₄ phase.^2,5) The purity of the FeS₂ and CaSO₄·2H₂O used is 99.0 wt.% and 99.5 wt.%, respectively, both of which were obtained from Tianjin Shentai Chemical Reagent Technology Co. Ltd., China. The gases used in this research, including N₂, CO and CO₂, all have a purity of 99.99 vol.%.

Table 1. Chemical composition of the tin-bearing iron concentrate (wt.%).

Components	Fe	SiO₂	MgO	CaO	Al₂O₃	Sn	S
Contents	65.75±0.25	2.08±0.05	2.01±0.05	1.30±0.03	1.05±0.02	0.94±0.03	0.12±0.01

2.2. Roasting Methods

The roasting tests were carried out in a horizontal tube furnace, which has been illustrated previously.²⁾ For the experimental procedure, first the materials, including the tin-bearing iron concentrate, FeS₂ and CaSO₄·2H₂O, were ground and sieved to 74 μm, respectively. After that, the tin-bearing iron concentrate with CaSO₄·2H₂O or FeS₂ was mixed thoroughly according to various sulfur addition amounts (‘S’ in CaSO₄·2H₂O or FeS₂/tin-bearing iron concentrate mass ratio) and loaded into a ceramic boat. Second, high-purity N₂ was introduced into the tube furnace to purge the air before the furnace reached 1073 K. Then, the ceramic boat was put into the furnace and heated to a proper temperature at a heating rate of 2.5 K/min under a gas mixture of CO and CO₂ with a flow rate of 300 ml/min. Finally, the roasted samples were allowed to cool to room temperature under the high-purity N₂ atmosphere after roasting.

2.3. Characterization

The phase composition of the samples was analysed by XRD (Rigaku, TTR- III), and the XRD patterns were gained with Cu–Ka radiation in 2θ ranging from 10° to 90° with a scan step of 2°/min. The microstructure of the samples was characterized by EPMA techniques (JXA82, JEOL) and SEM (HITACHI-S3400N) coupled with EDS. The different elemental compositions of the samples were characterized by ICP-OES spectrometry (Analytik Jena AG) and chemical analysis. All measurements were conducted three times, and the average value was taken as the final result. Thermodynamic analysis was performed using FToxid, FTmisc and FactPS databases in the FactSage 7.2 software. The S removal ratio during the roasting process was defined according to:

S removal ratio=[ 1-( m 1 w t 1,S / m 0 w t 0,S ) ]×100%.

Here, m₀ and m₁ are the masses of the raw sample and roasted residue, respectively; wt_0,S and wt_1,S are the S contents in the raw sample and roasted residue, respectively.

3. Results and Discussion

3.1. Sn Removal from the Tin-bearing Iron Concentrate Roasted with FeS₂ and CaSO₄·2H₂O, Respectively

The Fe and Sn elements in the tin-bearing iron concentrate exist mainly in the form of Fe₃O₄ and SnO₂, respectively, and the tin sulfurization and volatilization is clearly influenced by the iron phases, as presented in Fig. 1.^2,5) In Fig. 1, a Fe–Sn spinel can be formed in region I, seriously restricting Sn volatilization. Upon increasing the CO content, this Fe–Sn spinel is transformed into the Fe–Sn alloy 1 in region II and then Fe–Sn alloy 2 in region III.⁵⁾ The Fe content increases while the tin content decreases in the formed Fe–Sn alloy in comparing region III to region II.⁵⁾ This leads to a decrease in the tin activity in the reaction (5) and higher residual Sn content from the point of view of thermodynamics.²⁾ Thus, the roasting experiments were evaluated in a gas mixture of 35 vol.%CO+65 vol.%CO₂ (region II), and the temperature range was set at 1073 K–1273 K.

[ Sn ] Fe-Sn alloy +2CO( g ) +S O 2 (g)→SnS( g ) +2C O 2 ( g )

(5)

Fig. 1.

Gas-equilibrium diagram of Fe–Sn–C–O system. (Online version in color.)

Without the addition of a curing agent, Fig. 2(a) shows that the tin residual content decreases little due to the considerable formation of Fe–Sn alloy, as shown in Fig. 3. In comparison, with the same ‘S’ amount added, 1.43 wt.% CaSO₄·2H₂O and 0.5 wt.% FeS₂, respectively, the Sn removal is promoted significantly (Fig. 2(a)). In this process, the SnO₂ is reduced firstly and then sulfurized and removed by gaseous SnS. Figure 3 shows that some Sn could not be reduced effectively and exists mainly in the form of SnO₂ at 1073–1173 K, and only then may be reduced to Sn (l) at the temperatures higher than 1223 K. The reduction process of SnO₂ can be described by a nucleation and growth model and its rate is controlled by interfacial chemical reaction as reported in previous researches,^26,27) which implies that an increase in temperature promotes SnO₂ reduction significantly. To increase the tin removal rate, therefore, the S should be kept in the solid phases at the lower temperatures. At 1273 K, the tin residual content is reduced to 0.022 wt.% and 0.026 wt.% with the addition of 1.43 wt.% CaSO₄·2H₂O and 0.5 wt.% FeS₂, respectively, both of which meet the standard for BF iron-making (Sn<0.08 wt.%). In addition, it is noteworthy that the sulfur residual content with CaSO₄·2H₂O added is obviously higher than that with the addition of FeS₂ at 1273 K, as shown in Fig. 2(b), which implies that the utilization efficiency of ‘S’ for Sn removal in CaSO₄·2H₂O is higher than that in FeS₂. The detailed transformation behaviours of the ‘S’ in FeS₂ and CaSO₄·2H₂O will be discussed in the following sections.

Fig. 2.

(a) Effect of curing agents on the tin residual content in the temperature range of 1073 K–1273 K at a heating rate of 2.5 K/min; (b) Effect of curing agents on the sulfur residual content in the temperature range of 1073 K–1273 K at a heating rate of 2.5 K/min. (Online version in color.)

Fig. 3.

(a) XRD and (b) EPMA analyses of the roasted residues in different temperature ranges without curing agent. (Online version in color.)

3.2. Transformation Behaviour of the ‘S’ in FeS₂ Roasted with the Tin-bearing Iron Concentrate

The equilibrium composition for the reaction systems of FeS₂+CO+CO₂, FeS₂+Fe₃O₄+CO+CO₂ and FeS₂+Fe₃O₄+CO was calculated respectively with the Equilib Module of FactSage 7.2 based on the minimization of the total Gibbs free energy. FeS₂(0.003 mol), Fe₃O₄ (0.0227 mol) and varying amounts of mixed gases of 35 vol.%CO+65 vol.%CO₂ (Table 2) were selected as the reactant precursors. In the temperature range of 1123 K to 1273 K, Fig. 4(a) shows that FeS₂ is mainly transformed to FeS, COS(g) and S₂(g) through Eqs. (6) and (7). With the addition of Fe₃O₄, the amount of S₂(g) and COS(g) decreases while that of FeS increases, as shown in Fig. 4(b), which might be due to the occurrence of reactions (8) and (9). To verify this from the point of view of thermodynamics, the equilibrium composition of FeS₂+Fe₃O₄+CO at 1160.5 K was calculated and given as shown in Fig. 4(c). In Fig. 4(c), the FeS/SO₂ mole ratio of 4 without CO addition is consistent with the synergistic reaction of Eqs. (6) and (8), and the variations of SO₂(g), Fe₃O₄, FeS and CO₂(g) with increasing CO amount are in agreement with Eq. (9). The phase transformation of FeS₂ roasted with tin-bearing iron concentrate is mainly carried out through Eqs. (6), (7), (8) and (9).

Table 2. Amounts of CO and CO₂ used in the calculation for reaction systems of FeS₂+CO+CO₂ and FeS₂+Fe₃O₄+CO+CO₂.

Temperature/K	FeS₂+CO+CO₂		FeS₂+Fe₃O₄+CO+CO₂
Temperature/K	CO/mol	CO₂/mol	CO/mol	CO₂/mol
1123	0.09375	0.17411	0.09375	0.17411
1148	0.14063	0.26116	0.14063	0.26116
1173	0.1875	0.34821	0.1875	0.34821
1198	0.23438	0.43527	0.23438	0.43527
1223	0.28125	0.52232	0.28125	0.52232
1248	0.32813	0.60938	0.32813	0.60938
1273	0.375	0.69643	0.375	0.69643

Fig. 4.

The equilibrium compositions of (a) FeS₂+CO+CO₂ system, (b) FeS₂+Fe₃O₄+CO+CO₂ system, and (c) FeS₂+Fe₃O₄+CO system at 1160.5 K. (Online version in color.)

To confirm the thermodynamic calculation described above, 0.003 mol of FeS₂ and 5.80 g of the tin-bearing iron concentrate (containing 0.0227 mol of Fe₃O₄), which was consistent with the amounts used in the thermodynamic calculation in Fig. 4(b), were mixed and roasted in 35 vol.%CO+65 vol.%CO₂ in the temperature range from 1073 K to 1273 K at a heating rate of 2.5 K/min. The XRD and SEM-EDS analyses for the roasted residues are presented in Figs. 5 and 6.

2Fe S 2 →2FeS+ S 2 ( g )

(6)

S 2 ( g ) +2CO( g ) →2COS( g )

(7)

2F e 3 O 4 +5 S 2 ( g ) →6FeS+4S O 2 ( g )

(8)

3S O 2 ( g ) +10CO( g ) +F e 3 O 4 →3FeS+10C O 2 ( g )

(9)

Fig. 5.

XRD patterns of the roasted residues of tin-bearing iron concentrate roasted with FeS₂ in different temperature ranges (roasting temperature range: 1073 K–1273 K; heating rate: 2.5 K/min; mixed gases flow rate: 300 ml/min). (Online version in color.)

Fig. 6.

The SEM-EDS analysis of S-containing phase in the roasted residues of tin-bearing iron concentrate roasted with FeS₂ in temperature range of (a) 1073 K–1123 K, (b) 1073 K–1173 K, (c) 1073 K–1223 K, and (d) 1073 K–1273 K (heating rate: 2.5 K/min; mixed gases flow rate: 300 ml/min). (Online version in color.)

Figure 5 shows that the roasted residue is mainly composed of phases of Fe₃O₄, FeO and FeS at 1073 K–1123 K and that the intensity of Fe₃O₄ diffraction peak decreases while that of the FeO increases with increasing temperature to 1173 K. The Fe₃O₄ exists in these residues in this temperature range that seems to be different from the results in Fig. 4(b), the reason for which might be that the reduction of Fe₃O₄ occurs slowly at a relatively low temperature. This reduction of Fe₃O₄ by CO (g) is controlled by interfacial chemical reaction kinetically as reported in previous researches and temperature rising increases the reduction rate significantly.^27,28,29,30) As a result, only the FeO and FeS diffraction peaks can be detected when the roasting temperature is higher than 1223 K. Meanwhile, a small amount of FeS–FeO can also be detected in the roasted residues at temperatures higher than 1223 K, as shown in Figs. 6(c)–6(d). The transformation of FeS₂ shown in Figs. 5 and 6 agrees well with the thermodynamic analysis results in Fig. 4.

The FeS₂ decomposes by a large amount through Eq. (6) over the temperature range of 1073 K–1123 K, and then most of the generated S₂ (g) could be fixed in the residue by Fe₃O₄ through forming FeS (Eq. (8)) thermodynamically according to Fig. 4(b). However, Fig. 5 shows that the S removal ratio reaches 40.55%. This high ratio might be attributed to the fact that it is difficult for S₂ (g) to effectively diffuse into Fe₃O₄. Consequently, the reaction of S₂ gas with Fe₃O₄ to produce FeS as in Eq. (8) does not fully take place.^27,31) The S-containing phase in the roasted residue is transformed to FeS at this temperature range seen from Fig. 5. Then combined with Fig. 2(a) in which the Sn residual content obviously decreases for increasing roasting temperature from 1123 K–1273 K at a heating rate of 2.5 K/min, it can be concluded that the FeS plays a major role for tin removal using FeS₂. To study the action mechanism of FeS on the tin sulfurization and removal, thermodynamic analysis of a FeS+CO+CO₂ system and corresponding experiments were carried out. The sample A in Fig. 2(a) was used as the raw material in these series experiments, in which the Sn exists in the form of metallic Sn (Fe–Sn alloy) (Fig. 3(b)).

The Equilib Module of FactSage 7.2 was also used to calculate the equilibrium composition of the FeS+CO+CO₂ system, in which 0.000057 mol FeS and 0.268 mol mixed gases of CO and CO₂ were selected as reactant precursors. At 1173 K, Fig. 7(a) shows that the FeS is mainly oxidized to Fe₃O₄ and SO₂(g) by CO₂(g) through Eq. (10) at a low content of CO in the mixed gases, and the formed SO₂(g) is reduced to COS(g) for increasing CO content through Eq. (11). When the CO content exceeds 70 vol.% and increases, the phases of FeO and SO₂(g) disappear, the COS(g) and Fe amounts increase, and the FeS amount decreases. This indicates that the formation of COS(g) through Eq. (12) is changing into the main reaction.

Fig. 7.

(a) Effect of CO content on the equilibrium compositions of FeS+CO+CO₂ system at 1173 K; (b) Effect of roasting time on the tin and sulfur residual contents in different CO contents using FeS as additive. (Online version in color.)

Figure 7(b) shows that the sulfur residual content in 60 vol.%CO+40 vol.%CO₂ is higher than that in 30 vol.%CO+70 vol.%CO₂, which implies that the total formation amount of gaseous sulfur (COS(g) and SO₂(g)) in 60 vol.%CO+40 vol.%CO₂ is fewer than that in 30 vol.%CO+70 vol.% CO₂. In addition, the increase in CO content promotes the formation of COS(g), as deduced from Eqs. (11) and (12), as a result of which the COS(g) occupies a larger proportion in the formed gaseous sulfur in 60 vol.%CO+40 vol.% CO₂ than that in 30 vol.%CO+70 vol.%CO₂. Combined with the fact that the tin residual content is decreased when the CO content is increased from 30 vol.%CO to 60 vol.%CO, it can be concluded that the COS(g) has a better sulfurization effect on the tin (Eq. (13)) than SO₂(g) (Eq. (14)).

As mentioned above, when FeS₂ is used as a curing agent for tin removal from tin-bearing iron concentrate, the FeS generated from the FeS₂ decomposition plays a major role through its transformation to SO₂(g) and COS(g). Meanwhile, the COS(g) has better sulfurization effects on tin than SO₂(g).

FeS+C O 2 ( g ) →S O 2 ( g ) +CO( g ) +F e x O

(10)

S O 2 ( g ) +3CO( g ) →COS( g ) +2C O 2 ( g )

(11)

FeS+CO( g ) →COS( g ) +Fe

(12)

Sn( l ) +COS( g ) →SnS( g ) +CO( g )

(13)

2Sn( l ) +S O 2 ( g ) →SnS( g ) +Sn O 2

(14)

3.3. Transformation Behaviour of the ‘S’ in CaSO₄·2H₂O Roasted with Tin-bearing Iron Concentrate

The equilibrium composition for the reaction systems of CaSO₄+CO+CO₂, CaSO₄+Fe₃O₄+CO+CO₂, and CaSO₄+Fe₃O₄+CO calculated by the FactSage 7.2 software are shown in Fig. 8. CaSO₄ (0.006 mol), Fe₃O₄ (0.0227 mol) and varying amounts of mixed gases of 35 vol.% CO+65 vol.%CO₂ (Table 3) were selected as reactant precursors for these three reaction systems. The decomposition of CaSO₄ in mixed gases of CO and CO₂ is a coexistent competing reaction between Eqs. (4) and (15). Figure 8(a) shows that most of the CaSO₄ is reduced to CaS at relatively low temperatures and that the other CaSO₄ is transformed into CaO and COS(g) through Eqs. (15) and (11). Figure 8(b) shows that with the addition of Fe₃O₄, CaSO₄ is converted into FeS, (Fe,Ca)_xO and Ca₂Fe₂O₅, which may be due to Eqs. (16), (17), (18), (19). The equilibrium composition of CaSO₄+ Fe₃O₄+CO at 1148 K (Fig. 8(c)) was calculated to show whether Eqs. (16), (17), (18), (19) occur in the CaSO₄+Fe₃O₄+CO+CO₂ system. Figure 8(c) shows that CaSO₄ is mainly transformed to CaS for a CO amount of less than 0.02 mol. Then, the CaS and Fe₃O₄ amounts decrease and disappear, and the FeS and Ca₂Fe₂O₅ amounts increase with CO amount from 0.02 to 0.06 mol. This result indicates that the formation of FeS occurs mainly through Eqs. (16) and (17) and that CaO, which is formed by Eq. (17), will further combine with Fe_xO and Fe₃O₄ to form (Fe,Ca)_xO (Eq. (18) and Ca₂Fe₂O₅ (Eq. (19)), respectively.

Fig. 8.

The equilibrium compositions of (a) CaSO₄+CO+CO₂ system, (b) CaSO₄+Fe₃O₄+CO+CO₂ system, and (c) CaSO₄+Fe₃O₄+CO system at 1148 K. (Online version in color.)

Table 3. Amounts of CO and CO₂ used in the calculation for reaction systems of CaSO₄+CO+CO₂ and CaSO₄+Fe₃O₄+CO+CO₂.

Temperature/K	CaSO₄+CO+CO₂		CaSO₄+Fe₃O₄+CO+CO₂
Temperature/K	CO/mol	CO₂/mol	CO/mol	CO₂/mol
1123	0.09375	0.17411	0.09375	0.17411
1148	0.14063	0.26116	0.14063	0.26116
1173	0.1875	0.34821	0.1875	0.34821
1198	0.23438	0.43527	0.23438	0.43527
1223	0.28125	0.52232	0.28125	0.52232
1248	0.32813	0.60938	0.32813	0.60938
1273	0.375	0.69643	0.375	0.69643

To confirm the thermodynamic calculation in Fig. 8, 0.006 mol of CaSO₄·2H₂O and 5.80 g of the tin-bearing iron concentrate (containing 0.0227 mol of Fe₃O₄) were mixed and roasted in 35 vol.%CO+65 vol.%CO₂ from 1073 K to 1273 K at a heating rate of 2.5 K/min. The XRD and EPMA analysis results for the roasted residues are presented in Figs. 9 and 10.

2CaS O 4 →2CaO+2S O 2 ( g ) + O 2 ( g )

(15)

F e 3 O 4 +CO( g ) →F e x O+C O 2 ( g )

(16)

F e x O+CaS→FeS+CaO

(17)

CaO+F e x O→ (Fe,Ca) x O

(18)

CaO+F e 3 O 4 →C a 2 F e 2 O 5 +F e x O

(19)

Fig. 9.

XRD patterns of roasted residues of the tin-bearing iron concentrate and CaSO₄·2H₂O in different temperature ranges (roasting temperature range: 1073 K–1273 K; heating rate: 2.5 K/min; mixed gases flow rate: 300 ml/min). (Online version in color.)

Fig. 10.

(a) The EPMA analysis of S-containing phase in the roasted residues in different temperature ranges; (b) The corresponding mapping analysis of the roasted residue at the temperature of 1073 K–1173 K (heating rate: 2.5 K/min; mixed gases flow rate: 300 ml/min). (Online version in color.)

Figure 9 demonstrates that the main mineral constituents in the roasted residue are Fe₃O₄ and CaS at 1073 K–1123 K and the corresponding sulfur removal ratio is only 6.34%, which indicates that CaSO₄ was mainly reduced to CaS through Eq. (4) and the little surplus of CaSO₄ was decomposed by Eq. (15). This result is in agreement with Fig. 8(a). When the temperature reached 1173 K, diffraction peaks of FeS and Fe_xO are detected. Meanwhile, the corresponding EPMA analysis in Fig. 10(a) shows that (Fe,Ca)_xO can be found in the roasted residue, which is consistent with the result in Fig. 8(c), except for the phase of Ca₂Fe₂O₅. This result is due to the formed CaO, which reacts with Fe₃O₄ to from CaFe₃O₅ by Eq. (20) instead of Ca₂Fe₂O₅. In addition, as shown in Fig. 9, a diffraction peak of Ca₃Fe₄S₃O₆ is also found at 1073 K–1173 K, and this diffraction peak persists with a further increase in temperature. The Ca₃Fe₄S₃O₆ might be formed between Fe₃O₄ and CaS through Eq. (21). This formation is deduced from Fig. 10(b), which shows that Ca₃Fe₄S₃O₆ coexists with Fe_xO. To confirm this, a mixture of Fe₃O₄ and CaS was roasted over the temperature range of 1073 K–1173 K in 15 vol.% CO+85 vol.%CO₂ (Fe₃O₄ stability region, as shown in Fig. 1) before carrying out XRD analysis for the roasted residue.

CaO+F e 3 O 4 →CaF e 3 O 5

(20)

CaS+F e 3 O 4 →C a 3 F e 4 S 3 O 6 +F e x O

(21)

For Fe₃O₄ roasting with CaS, Fig. 11(a) shows that the phase composition of the roasted residue contains Ca₃Fe₄S₃O₆, which implies that this Ca₃Fe₄S₃O₆ is indeed generated from CaS and Fe₃O₄. To obtain the phase transformation of Ca₃Fe₄S₃O₆ in CO and CO₂ mixed gases, the sample B in Fig. 11(a) was further roasted from 1173 K to 1273 K at a heating rate of 2.5 K/min in 5 vol.%CO+95 vol.%CO₂ and 60 vol.%CO+40 vol.%CO₂, respectively. Figure 11(b) shows that the formed Ca₃Fe₄S₃O₆ can be oxidized to CaFe₃O₅ and Ca₂Fe₂O₅ by CO₂(g) according to Eqs. (22) and (23) in 5 vol.%CO+95 vol.%CO₂. Figure 11(c) shows that, when the CO content increases to 60 vol.%, the ‘S’ in Ca₃Fe₄S₃O₆ is mainly transformed to FeS by Eq. (24), which could be further confirmed by the low sulfur removal ratio of 2.35 wt.%. In addition, Fig. 12 shows that the tin residual content also decreases to a value lower than 0.08 wt.% at roasting temperatures ranging from 1073 K to 1273 K at a heating rate of 2.5 K/min when the tin-bearing iron concentrate was roasted with the sample B under the same experimental conditions used in Fig. 2. The mechanism for tin removal using Ca₃Fe₄S₃O₆ might be due to its transformation to SO₂(g) and FeS first, then to COS(g) by CO(g) and finally Sn sulfurization through Eqs. (13) and (14).

3C a 3 F e 4 S 3 O 6 +25C O 2 ( g ) → 4CaF e 3 O 5 +5CaO+9S O 2 ( g ) +25CO( g )

(22)

CaO+CaF e 3 O 5 →C a 2 F e 2 O 5 +F e x O

(23)

C a 3 F e 4 S 3 O 6 +CO( g ) →FeS+F e x O+CaO+C O 2 ( g )

(24)

Fig. 11.

The XRD patterns of (a) roasted residue of Fe₃O₄ and CaS in 15 vol.%CO+85 vol.%CO₂ from 1073 K to 1173 K, (b) roasted residue of Sample B in 5 vol.%CO+95 vol.%CO₂ from 1173 K to 1273 K, and (c) roasted residue of Sample B in 60 vol.%CO+40 vol.%CO₂ from 1173 K to 1273 K. (Online version in color.)

Fig. 12.

Effect of temperature on the tin and sulfur residual contents using sample B as additive. (Online version in color.)

The transformation behaviour of ‘S’ in FeS₂ and CaSO₄ respectively while being roasted with tin-bearing iron concentrate is summarized in Fig. 13. The S₂(g), which is generated from FeS₂ decomposition, is volatile and plays little effect on the tin sulfurization. According to Fig. 4(a), the formation amount of COS (g) is obviously higher than that of S₂ (g), which seems that the loss of sulfur might be mainly due to the volatilization of COS (g). However, the COS (g) forms mainly through the two consecutive reactions; the first is the self-decomposition of FeS₂ (Eq. (6)) and the second is the combination reaction between S₂ (g) and CO (g) (Eq. (7)). In Fig. 14(a), the Gibbs free energy of Eq. (6) is negative and smaller than that of Eq. (25) under 35vol.% CO+ 65 vol.% CO₂ atmosphere at 1073–1273 K, which indicates that the self-decomposition of FeS₂ occupies a dominant place compared to its reaction with CO. It implies that massive sulfur could be lost in the form of S₂ (g) before the formation of COS (g). Though the generated S₂ (g) could be fixed in the residue by Fe₃O₄ through Eq. (8) according to Fig. 4(b), the slow diffusion of S₂ (g) to Fe₃O₄ restricts it. As a result, massive S is lost.

Fe S 2 +CO( g ) →FeS+COS( g )

(25)

Fig. 13.

The transformation behavior of the ‘S’ from FeS₂ and CaSO₄ respectively while roasted with tin-bearing iron concentrate in region II of Fig. 1. (Online version in color.)

Fig. 14.

(a) Gibbs free energy changes of Eqs. (6) and (25) with temperature at 35 vol.% CO+65 vol.% CO₂ (b) Gibbs free energy changes of Eqs. (4) and (15) under different P_SO2 and P_O2 with temperature at 35 vol.% CO+65 vol.% CO₂. (Online version in color.)

That for gypsum is different. The reaction of CO (g) and CaSO₄ ((Eq. (4)) occupies a dominant place compared to its self-decomposition (Eq. (15)) due to a more negative Gibbs free energy of Eq. (4) as shown in Fig. 14(b). Consequently, the formation of COS (g) through Eqs. (15) and (11) is restricted and the S loss is relatively smaller compared to that in the case of FeS₂. Based on this, the amount of sulfur added by CaSO₄·2H₂O can be decreased compared with that added using FeS₂. Under the temperature range of 1073 K–1273 K with a heating rate of 2.5 K/min and gas mixture (35 vol.%CO+65 vol.%CO₂) flow rate of 300 ml/min, the tin residual content decreases to 0.028 wt.% with the amount of sulfur added by CaSO₄·2H₂O being 0.213 wt.%, which is less by approximately 20 wt.% than the amount of sulfur added by FeS₂ (0.267 wt.%), as presented in Fig. 15. Besides, Table 4 shows that the iron content in this roasted residue is increased to 72.45 wt.% and the S content is low to 0.083 wt.%, indicating this iron concentrate can be utilized effectively through this process.

Fig. 15.

(a) Effect of sulfur amount on the tin residual contents in 1073 K–1273 K at a heating rate of 2.5 K/min; (b) Effect of sulfur amount on the sulfur residual contents in 1073 K–1273 K at a heating rate of 2.5 K/min. (Online version in color.)

Table 4. Chemical composition of the Sample C in Fig. 14 (wt.%).

Components	Fe	Sn	S
Contents	72.45±0.28	0.028±0.008	0.083±0.005

4. Conclusions

(1) The FeS₂ first decomposed into FeS and S₂(g) in mixed gases of CO and CO₂, and most S₂(g) volatilized and played little effect on the tin sulfurization. The FeS played a major role on the tin removal through its transformation to SO₂(g) and COS(g) in mixed gases of 35 vol.% CO and 65 vol.% CO₂, and the COS(g) had a better sulfurization effect on the tin than SO₂(g).

(2) The ‘S’ from gypsum waste (CaSO₄·2H₂O) was mainly transformed into FeS and Ca₃Fe₄S₃O₆ while roasting with tin-bearing iron concentrate in mixed gases of 35 vol.%CO and 65 vol.%CO₂, and the increased CO content promoted the transformation of Ca₃Fe₄S₃O₆ to FeS. Both the FeS and Ca₃Fe₄S₃O₆ could sulfurize Sn to SnS(g) through their transformation to SO₂(g) and COS(g) in mixed gases of CO and CO₂.

(3) The utilization efficiency of ‘S’ for tin sulfurization in CaSO₄·2H₂O was higher than that in FeS₂. Under a temperature range of 1073 K–1273 K with a heating rate of 2.5 K/min and gas mixture (35 vol.%CO+65 vol.%CO₂) flow rate of 300 ml/min, the tin residual content decreased to 0.028 wt.% with the amount of sulfur added by CaSO₄·2H₂O being 0.213 wt.%, which was less by approximately 20 wt.% than the amount of sulfur added by FeS₂ (0.267 wt.%).

Acknowledgments

The authors wish to express their thanks to the National Science Fund for General Projects (51874153) for the financial support of this research.

References

Corresponding author

Register with J-STAGE for free!