Sulphide capacities of CaO–MgO–AlO 1·5,MgO–MnO–AlO 1·5 and Cao–MgO–MnO

Abstract

Modelling of sulphide capacities of CaO–MgO–AlO_1·5, MgO–MnO–AlO_1·5 and CaO–MgO–MnO–AlO_1·5 slag systems was investigated. The Reddy–Blander model was applied to calculate the sulphide capacities a priori with the knowledge of chemical and solution properties of oxide and sulphides for the above slag systems. The sulphide capacities were determined as a function of temperature (1500–1650°C), composition (0< <0·6) and basic oxides (CaO, MnO and MgO). Sulphide capacities of aluminates increased with increase in temperatures and decreased with the increase in alumina content in the slags. Very good agreement was observed between model and experimental data.

Keywords

Sulphide capacities Reddy–Blander model slags

Introduction

The Reddy–Blander (RB) model for a priori prediction sulphide capacities of binary and multicomponent silicate melts using the model1 for polymeric silica chains was developed. They showed that sulphide capacities could be calculated a priori based on knowledge of the chemical and solution properties of oxides and sulphides.1^– 6 In another publication,7 the model was extended to binary aluminate melts. They applied the model to ternary silicate melts containing FeO.8 In all systems, excellent agreement between the model and experimental data was observed. In this work, the RB model is extended to calculate the sulphide capacities of binary, ternary and multicomponent CaO–MgO–MnO–AlO_1·5 aluminate slags.

Theoretical analysis

By using CaO–SiO₂ as an example, the binary is divided into two regions: one basic and another acidic.

The sulphide equilibrium reaction in the melt is written as (1) The equilibrium constant K_M for the reaction (1) is given by (2) The sulphide capacity C_s of slags is expressed as9 (3) where (wt‐%S) is the weight percentage of sulphur in slag, and and are the partial pressures of oxygen and sulphur respectively.

By combining equations (2) and (3), and C_s can be obtained as (4) where a_MS and a_MO are the activities of MS and MO respectively.

Reddy–Blander model for basic binary:

The silicate melt is completely depolymerised with oxygen given by the dissociation of MO. Some oxygen anions will be left unattached in the solution since the molar ratio MO/SiO₂ is >2. These anions (O²⁻) are called free oxygen. For this system, the model assumes that the only species in solution are one cation M²⁺ and three anions O²⁻, SiO₄⁴⁻ and S²⁻. The latter is present in very small amounts. The solution is thus a simple ternary MO–M₂SiO₄–MS. Substituting AlO_1·5 for SiO₂ in the melts and the equation for C_s can be written as5 (5) where M_S is the atomic weight of sulphur, K_M is the equilibrium constant for the sulphur reaction, a_MO is the activity of the basic metal oxide in the binary system and is the average molecular weight of solution given by (6) Given the slag's temperature and composition, K_M and a_MO were obtained from HSC 4·110 and FactSage 5·111 respectively.

Reddy–Blander model for acidic binary:

In this system, the anions are assumed to be randomly mixed. The CaO is completely consumed and allowing for the assumption that the concentration of O²⁻ ions is negligible. The model assumes a lattice consisting of sulphur (monomer) and a distribution of polymers (aluminate polymeric ions). The volume fraction of S²⁻ sites is given by (7) But since the solution is dilute, is small compared to N_Al, and as a result it can be neglected in the denominator. As discussed earlier, is also negligible. Equation (7) thus reduces to (8) Acidic binaries are divided into two regions. For intermediate acidic binaries ( ), the solution is a mixture of polymeric species and monomer ions. The model equation for this region is (9) where m is the polymer chain length given by (10) For the remaining portion of the acidic binary ( ), the solution is mostly polymeric. The model equation for this region is (11) The most complicated multicomponent system investigated in this work has four components. In this system, four oxides are basic: CaO, MgO and MnO, while one acidic oxide AlO_1·5. To calculate the sulphide capacity of this system, it is broken down into three ternaries: CaO–MgO–AlO_1·5, MgO–MnO–AlO_1·5 and CaO–MnO–AlO_1·5. Equations (5) and (9)–(11) can be used to calculate the sulphide capacities of each ternary system with few changes. The activity of the basic component is obtained from FactSage 5·111 using the same mole fractions of the acidic components as in the multicomponent system. The average molecular weight of solution is obtained for the ternary system. Once the sulphide capacity of each ternary has been calculated, the following equation was used to calculate the sulphide capacity of the multicomponent system (12) where is the equivalent cationic fraction of species i and can be calculated from the following equation (13) where the basic mole fractions are the same as those in the multicomponent system. Equations similar to equation (13) were used for other basic components. The logC_s(i) is the sulphide capacity of the ternary system. The sulphide capacity of multicomponent system using equation (12) was calculated at a constant composition of acidic component (i.e. AlO_1·5) for binary melts.

Results and Discussion

The model equations were used to calculate sulphide capacities for binary, ternary and higher order systems given a slag's temperature and composition. The model was used to calculate sulphide capacity of many systems, but only aluminate systems with available experimental data are reported in this work. Experimental results for sulphide capacities of binary CaO–AlO_1·5 were obtained from Refs. 9 and 12–19.

Sulphide capacities of ternary slag systems

The sulphide capacities of ternary slag system experimental and model calculated data are presented in Table 1. The sulphide capacities of the ternary system CaO–MgO–AlO_1·5 at 1550, 1600 and 1650°C as reported by Hino et al.13 are compared to model results in Fig. 1. The ternary system MgO–MnO–AlO_1·5 was studied by Nzotta et al.20 at 1600 and 1650°C. Figure 2 compares reported experimental values to model results for the ternary MgO–MnO–AlO_1·5. The liquidus data for basic sulphides (i.e. MgS) and also experimental sulphide capacities of MgO–MnO–AlO_1·5 are not available. The available solid data for solid sulphides and activity of basic activity in aluminates were used in calculating the sulphide capacities as a function of temperature. The predicted sulphide capacities for CaO–MgO–AlO_1·5 are slightly lower than the corresponding experimental data. The sulphide capacities of CaO–MgO–AlO_1·5 decreases with the increase in composition of AlO_1·5. But the decrease is much smaller for the MgO–MnO–AlO_1·5 system. Effect of temperature on sulphide capacities of these ternary systems is not significant.

Figure 1

Comparison of available experimental data to model results of CaO–MgO–AlO_1·5 at 1550, 1600 and 1650°C: experimental data are from Hino et al.13

Figure 2

Comparison of available experimental data to model results of MgO–MnO–AlO_1·5 at 1600 and 1650°C: experimental data are from Nzotta et al.20

Table 1

Sulphide capacity model and experimental data for ternary slag systems at various temperatures*

T, °C	X _CaO	X _MgO	X _MnO		RB model C_s	Experimental data C_s
1550	0·521	0·053		0·427	2·5827×10⁻³	6·1518×10⁻³
1550	0·427	0·103		0·470	9·6587×10⁻⁴	2·2233×10⁻³
1550	0·471	0·052		0·477	1·3332×10⁻³	2·7102×10⁻³
1550	0·412	0·052		0·536	6·9279×10⁻⁴	1·3032×10⁻³
1550	0·384	0·039		0·577	4·9629×10⁻⁴	4·6238×10⁻⁴
1550	0·336	0·039		0·625	2·7639×10⁻⁴	2·0277×10⁻⁴
1550	0·380	0·090		0·530	5·3700×10⁻⁴	1·1995×10⁻³
1550	0·351	0·090		0·559	3·9897×10⁻⁴	4·0884×10⁻⁴
1600	0·530	0·053		0·417	3·8896×10⁻³	9·8401×10⁻³
1600	0·504	0·117		0·380	3·9067×10⁻³	9·1622×10⁻³
1600	0·426	0·116		0·458	1·3917×10⁻³	3·9355×10⁻³
1600	0·481	0·052		0·467	2·0452×10⁻³	4·5290×10⁻³
1600	0·432	0·052		0·516	1·1518×10⁻³	2·2803×10⁻³
1600	0·374	0·039		0·587	6·2475×10⁻⁴	9·0365×10⁻⁴
1600	0·316	0·039		0·645	3·1022×10⁻⁴	4·5709×10⁻⁴
1600	0·367	0·128		0·505	6·9496×10⁻⁴	1·9679×10⁻³
1600	0·322	0·090		0·588	4·1348×10⁻⁴	9·9083×10⁻⁴
1650	0·504	0·117		0·380	5·1045×10⁻³	1·9364×10⁻²
1650	0·511	0·053		0·437	3·9352×10⁻³	1·6596×10⁻²
1650	0·461	0·052		0·487	2·0964×10⁻³	6·7764×10⁻³
1650	0·402	0·052		0·546	1·1834×10⁻³	3·1842×10⁻³
1650	0·374	0·039		0·587	8·6713×10⁻⁴	1·2078×10⁻³
1650	0·399	0·090		0·510	1·2145×10⁻³	5·4450×10⁻³
1650	0·380	0·090		0·530	1·0205×10⁻³	2·9309×10⁻³
1650	0·333	0·077		0·590	6·2580×10⁻⁴	1·3900×10⁻³
1600		0·091	0·528	0·381	6·1422×10⁻³	8·1670×10⁻³
1600		0·091	0·528	0·381	6·1422×10⁻³	7·9400×10⁻³
1600		0·218	0·505	0·277	1·1842×10⁻²	1·5490×10⁻²
1600		0·218	0·505	0·277	1·1842×10⁻²	1·5750×10⁻²
1650		0·091	0·528	0·381	7·7209×10⁻³	1·2160×10⁻²
1650		0·091	0·528	0·381	7·7209×10⁻³	1·1940×10⁻²
1650		0·123	0·534	0·343	8·9671×10⁻³	1·2790×10⁻²
1650		0·123	0·534	0·343	8·9671×10⁻³	1·3100×10⁻²
1650		0·133	0·543	0·324	1·5074×10⁻²	1·4400×10⁻²
1650		0·133	0·543	0·324	1·5074×10⁻²	1·6500×10⁻²
1650		0·218	0·505	0·277	1·5140×10⁻²	1·9290×10⁻²
1650		0·218	0·505	0·277	1·5140×10⁻²	1·8360×10⁻²

^*CaO–MgO–AlO_1·5 at 1550, 1600 and 1650°C experimental data are from Hino et al.,13 MgO–MnO–AlO_1·5 at 1600 and 1650°C experimental data are from Nzotta et al.20

Sulphide capacities of quaternary slag systems

The sulphide capacities of the system CaO–MgO–MnO–AlO_1·5 as measured by Nzotta et al.20 at 1600 and 1650°C are compared to model results in Fig. 3. The sulphide capacities of quaternary slag system experimental and model calculated data are presented in Table 2. A priori predicted sulphide capacities for this system match reasonably well with the experimental sulphide capacities of quaternary slags.

Figure 3

Comparison of available experimental data to model results of CaO–MgO–MnO–AlO_1·5 at 1600 and 1650°C: experimental data are from Nzotta et al.20

Table 2

Sulphide capacities model and experimental data for quaternary slag systems at various temperatures*

T, °C	X _CaO	X _MgO	X _MnO		RB model C_s	Experimental data C_s
1600	0·431	0·067	0·054	0·448	2·4685×10⁻³	2·9500×10⁻³
1600	0·431	0·067	0·054	0·448	2·4685×10⁻³	2·7560×10⁻³
1600	0·427	0·067	0·039	0·467	1·9150×10⁻³	2·4990×10⁻³
1600	0·427	0·067	0·039	0·467	1·9150×10⁻³	2·4830×10⁻³
1600	0·414	0·093	0·032	0·460	1·6845×10⁻³	2·3060×10⁻³
1600	0·414	0·093	0·032	0·460	1·6845×10⁻³	2·1570×10⁻³
1600	0·423	0·067	0·048	0·463	2·0378×10⁻³	1·9480×10⁻³
1600	0·423	0·067	0·048	0·463	2·0378×10⁻³	1·8750×10⁻³
1650	0·431	0·067	0·054	0·448	3·2366×10⁻³	3·5700×10⁻³
1650	0·431	0·067	0·054	0·448	3·2366×10⁻³	4·1640×10⁻³
1650	0·427	0·067	0·039	0·467	2·5291×10⁻³	2·7390×10⁻³
1650	0·427	0·067	0·039	0·467	2·5291×10⁻³	2·0250×10⁻³
1650	0·414	0·093	0·032	0·460	2·2363×10⁻³	2·9520×10⁻³
1650	0·414	0·093	0·032	0·460	2·2363×10⁻³	3·0130×10⁻³
1650	0·423	0·067	0·048	0·463	2·6857×10⁻³	2·2800×10⁻³
1650	0·423	0·067	0·048	0·463	2·6857×10⁻³	2·1110×10⁻³

*CaO–MgO–MnO–AlO_1·5 at 1600 and 1650°C experimental data are from Nzotta et al.20

Conclusion

The RB model can be used to calculate sulphide capacities of any system a priori given the slag's temperature and composition. The model agrees very well with experimentally measured values. For molten metal desulphurisation, slag temperature and composition can be adjusted to obtain high sulphide capacity. Slags with high sulphide capacities can be used effectively and economically to absorb and retain sulphur from molten metal. The low sulphur steels can be used in more demanding applications.

Footnotes

Acknowledgements

The authors would like to acknowledge the financial support for this research by American Cast Iron Pipe Company (ACIPCO) and the National Science Foundation (NSF).

References

Flory

: ‘Principles of polymer chemistry’; 1953, Ithaca, Cornell University Press.

Reddy

, Blander

: Metall. Trans. B, 1989, 20B, 137–140.

Chen

, Reddy

, Blander

: Proc. 3rd Int. Conf. on ‘Molten slags and fluxes’, 270–272; 1989, London, Institute of Metals.

Reddy

, Hu

, Blander

: Proc. 4th Int. Conf. on ‘Molten slags and fluxes’, 144–148; 1992, Sendai, ISIJ.

Reddy

, Zhao

: Metall. Mater. Trans. B, 1995, 26B, 925–928.

Reddy

, Blander

: Metall. Trans. B, 1987, 18B, 591–596.

Zhao

, Reddy

: in ‘EPD Congress 1995’, 39–47; 1995, Warrendale, PA, TMS.

Derin

, Reddy

: Proc. Int. Metallurgy and Materials Cong., 54–62; 2002, Istanbul, UCET.

Fincham

CJB

, Richardson

: Proc. R. Soc. Lond. A, 1954, 223A, 40–62.

10.

Roine

: ‘Outokumpu HSC Chemistry for Windows program, version 4.1’; 2001, Pori, Outokumpu Research Oy.

11.

Bale

, Pelton

, Thompson

: ‘FactSage 5·1: thermochemical Software for Windows’; 2002, Montreal, Que., Thermfact Ltd.

12.

Drakaliysky

, Nilsson

, Sichen

, Seetharaman

: High Temp. Mater. Process., 1996, 15, 263–272.

13.

Hino

, Kitagawa

, Ban‐Ya

: ISIJ Int., 1993, 33, 36–42.

14.

Abraham

, Richardson

: JISI, 1960, 195, 313–317.

15.

Sharma

, Richardson

: JISI, 1961, 198, 386–390.

16.

Carter

, Macfarlane

: JISI, 1957, 185, 54–66.

17.

Kor

GJW

, Richardson

: Trans. Metall. Soc. AIME, 1969, 245, 319–327.

18.

Ozturk

, Turkdogan

: Met. Sci., 1984, 18, 299–332.

19.

Zhao

: ‘Modeling of sulfide capacities of binary silicate and aluminate melts and ternary silicate melts’, Reno, NV, USA, University of Nevada, Master's dissertation, 1995.

20.

Nzotta

, Sichen

, Seetharaman

: ISIJ Int., 1998, 38, 1170–1179.

Sulphide capacities of CaO–MgO–AlO 1·5,MgO–MnO–AlO 1·5 and Cao–MgO–MnO–AlO 1·5 slags

Abstract

Keywords

Introduction

Theoretical analysis

Reddy–Blander model for basic binary:

Reddy–Blander model for acidic binary:

Results and Discussion

Sulphide capacities of ternary slag systems

Sulphide capacities of quaternary slag systems

Conclusion

Footnotes

Acknowledgements

References