UDK 669.18:669.187.28 Original scientific article/lzvirni znanstveni članek ISSN 1580-2949 MTAEC9, 48(3)327(2014) MODELLING OF THE SULPHIDE CAPACITY OF STEELMAKlNG SLAGS MODELIRANJE SULFlDNE KAPACITETE JEKLARSKIH ŽLlNDER Zarko Radovic, Nebojsa Tadic, Milisav Lalovic, Dragoljub Blecic University of Montenegro, Faculty of Metallurgy and Technology, Cetinjski put bb, 81000 Podgorica, Montenegro zarkor@ac.me Prejem rokopisa - received: 2012-12-06; sprejem za objavo - accepted for publication: 2013-08-27 In this paper the influence of the CaF2 amount in steelmaking slags on the CaO and Al2O3 activities is presented. The new analytical model (RMJ) for determining the sulphide capacity was derived. The aim of this model is to adjust and correct some models, according to Tsao, Daffy and Young, that had limitations regarding the presence of CaF2 in the slag. Using the RMJ model, the effect of CaF2 on the sulphide capacity and desulphurisation degree is more precisely defined. Calcium fluoride decreases the negative effect of Al2O3 on the sulphide capacity. For an investigation under production conditions, several types of carbon and low-alloy steels were chosen. The steels used for the analysis of specified process parameters were produced using a mixture of CaF2-CaO and CaO-CaF2-white bauxite (WB) as the flux. For specified cases, the results obtained with the RMJ model as well as with the Young model, were presented. The RMJ model defines correction factors k1(CaF2) and k2(CaF2). It can be seen that the certainty factor of linear dependence is a bit larger in the case of the RMJ model than the one obtained with the Young model. Keywords: calcium fluoride, sulphide capacity, RMJ model V tem članku je predstavljen vpliv vsebnosti CaF2 v jeklarskih žlindrah na aktivnost CaO in Al2O3. Izpeljan je bil nov analitični model (RMJ) za določanje sulfidne kapacitete. Namen modela je prilagoditev in poprava modelov po Tsaoju, Daffyju in Youngu, ki imajo omejitve glede prisotnosti CaF2 v žlindri. Z uporabo modela RMJ je mogoče bolj natančno opredeliti vpliv CaF2 na sulfidno kapaciteto in stopnjo razžvepljanja. Kalcijev fluorid zmanjša negativni vpliv Al2O3 na sulfidno kapaciteto. Več vrst ogljikovih in malo legiranih jekel je bilo izbranih za industrijski preizkus. Jekla, izbrana za analizo specifičnih procesnih parametrov, so bila izdelana z uporabo mešanice CaF2-CaO in CaO-CaF2-beli boksit (WB) kot talilo. Za določene primere so predstavljeni rezultati RMJ-modela, kot tudi Youngovega modela. RMJ-model določa korekcijske faktorje k1(CaF2) in k2(CaF2). Opazi se, da je faktor zanesljivosti linearne odvisnosti nekoliko večji pri RMJ-modelu v primerjavi z dobljenim pri Youngovem modelu. Ključne besede: kalcijev fluorid, sulfidna kapaciteta, RMJ-model 1 INTRODUCTION Synthetic-slag practice is employed to obtain clean steels and to desulphurize molten steel (50-60 % of the original sulphur). These slags contain CaO, CaF2, Al2O3 and a small amount of SiO2. Besides, slag contains insoluble particles (CaO, MgO) and a chemical analysis of the slag gives unrealistic results for the slag basicity. ln this case, the calculated basicity is significantly higher than the basicity of fully liquid slags. An important parameter to characterize synthetic slag with respect to its suitability to desulphurize molten steel is the sulphide capacity. On the basis of the ionic theory, the sulphide capacity and sulphur activity must be correlated with the optical basicity because the standard calculation methods neglect the influence of many oxides, except for CaO, MgO and SiO2.1 The model by Tsao2 has been used for defining the chemical-composition influence on the sulphide capacity: ) - (1) lg C S = 3.44(Z CaO + 0.1Z MgO - 0.8Zai2O3 - ^ SiO 9894 +2.05 where: Xi is the molar fraction of oxide i, T is the slag temperature in K. However, the above correlation is not applicable to the metallurgical slags containing CaF2 because the effect of CaF2 must be substituted with an oxide. On the basis of the optical-basicity index, the following correlation was suggested by Gaye3: B 12364 lg C S = -+1.445 (2) where: B = 5.62 (% CaO) + 4.15 (% MgO) - 1.15 (% SiO2) + 1.46 (% Al2O3), D = (% CaO) + 1.39 (% MgO) + 1.87 (% SiO2) + 1.65 (% Al2O3). ln this case the amount of CaF2 is assumed to be part of the CaO amount. Since aluminum was used for steel deoxidation, the activity of Al2O3 and the Al amount are considered while calculating the sulphur-distribution coefficient4: lg Ls = lg C S - 3lg(fl • Al^O,) + 3(%Al) +- 20397 —5.482 (3) ln the secondary-metallurgy processes Si was used as a deoxidizer. Therefore, the formation of a stable SiO2, its interaction with the other oxides and its influence on the sulphide capacity should not be neglected. This correlation can be empirically defined with the Young expression5: Ig CS =-13.913+48.84A-23.82A2 - r11710 (4) -0.02223(%SiO2) - 0.02275(%Al2O3) where A is the theoretical optical basicity of the slag calculated with the Nakamura method6. An experimental analysis of EAF steelmaking slags shows that this model can be used as the basis for further consideration. Young et al.5 showed that equation (4) can be applied only to the range of A < 0.8, i.e., to the slags examined in the experimental part of this research. Larger values of the CaF2 amount cause high values of Cs and Ls. In this way, it is possible to obtain a value of Ls that is higher than 1000, and the literature reports show that such large values are indeed obtained in industrial practices7. The main aim of this paper is to derive a new model (RMJ) that will more precisely define the impact of CaF2 on the sulphide capacity of steelmaking slags. 2 EXPERIMENTAL WORK The first stage of the experimental investigations is a chemical analysis of the slag samples. The chemical compositions of the investigated slags for the two slag mixtures are given in Table 1. The steels are produced in a EAF 60 t, while Si and Al are used as deoxidizers. The slag and metal were manually sampled by the EAF operator at the temperature of ~1570 °C, after a vacuum treatment of the steel, 10-12 min before casting. The mixture for the synthetic slag contains CaO, CaFa and white bauxite in the mass ratio of 3 : 1 : 2, with the ratio of 3 : 1 for the CaO-CaFa mixture. For degassing the steel, a vacuum treatment and stirring with argon purging from the ladle bottom are used. The aim of using a triple slag mixture is to investigate the impact of CaFa in the case of an increased presence of Al2O3 in white bauxite. 3 RESULTS AND DISCUSSION A presence of CaFa increases the CaO saturation in the slags. For example, at comparable CaO concentrations, activity ^(CaO) in the CaO-CaFa slag is much higher than in the CaO-Al2O3 and CaO-SiOa systems. The slags in secondary steelmaking are basically unstable at high temperatures since two or more components in the slag react8 and the composition changes continuously while the fluorides and oxides react as follows: 3CaF2 + A12O3 = 2A1F3 (g) + 3CaO (5) The slag composition shows a steady increase in the CaO concentration and a decrease in AI2O3, so that the part of expression (4) that is correlated with Al2O3 must be corrected with k1(CaF2). Besides, the effect of moisture when dealing with the slags containing CaF2 is very important because of the following reaction: CaFa + H2O = CaO + 2HF(g) (6) In this case, CaF2 increases together with the CaO activity. Therefore, the Young model (equation 4) must be adjusted with the correction factors, k1(CaF2) and k2(CaF2), defining the secondary influence of CaF2 on Cs as follows: ^11710 — _ I -i 11 I -i _1_ /IV V/1 A _ J-i V'J A'- _ Ig CS = -13.913 + 48.84A- 23.82A2 - (7) -0.02223(% SiO2) - 0.02275(% A12O3) ■ k 1(CaF2) ■k2, (CaF2) Figure 1 shows sulphide capacities in the CaO-CaF2-Al2O3 ternary system at 1500 °C, as determined by Kor and Richardson4. Using the correlations shown in Figure 1 for the chemical compositions of the investigated slags (Table 1), obtained with the regression-analysis method, the following expressions are obtained: k 1(CaF. ) =-0.0001(%CaF2 )2 +0.024(%CaF2)+0.031 (8) k 2(ca^) =-0.0002(%CaF2)2 +0.015(%CaF2)+0.015 (9) Table 1: Chemical compositions of the investigated slags after steel vacuuming (mass fractions, w/%) Tabela 1: Kemijska sestava preiskovanih žlinder po obdelavi jekla v vakuumu (masni deleži, w/%) Slag mixture Steel (EN) Amounts of components (w/%) CaO MnO MgO SiO2 AI2O3 FeO Fe2O3 CaF2 S CaO-CaF2 1.6582 40.00 1.90 13.70 18.20 16.90 1.77 1.30 5.00 0.27 1.2714 43.52 0.87 11.31 13.77 21.20 1.14 1.90 4.65 0.50 1.1181 40.5 0.1 18.1 16.5 15.6 0.6 1.4 6.00 2.50 1.7147 56.5 0.3 7.8 15.1 6.2 0.9 0.9 8.90 2.30 1.0503 51.0 0.6 10.6 16.6 7.1 1.1 1.3 9.20 1.20 1.7005 39.2 1.9 24.3 17.2 9.5 1.5 0.2 5.10 0.30 1.1191 41.8 3.8 12.3 21.1 20.1 1.9 0.2 5.80 0.30 1.7225 50.7 0.8 10.1 18.2 13.1 1.2 0.9 4.90 0.50 CaO-CaF2-WB 1.6582 41.2 0.3 14.7 27.2 12.30 2.30 1.0 2.30 0.40 1.7225 44.0 0.5 9.2 23.1 9.1 1.8 1.1 2.5 0.32 1.1141 45.3 0.8 19.1 18.5 18.8 1.33 0.92 3.1 0.41 1.7147 46.5 0.6 17.8 15.1 9.2 1.4 1.4 3.9 0.39 1.0503 42.2 0.52 16.8 22.2 13.1 1.21 1.32 2.9 0.35 1.2714 39.52 0.87 19.31 21.8 20.1 1.94 1.8 2.95 0.33 Figure 1: Sulphide capacity for the CaO-CaF2-Al2O3 system at 1500 °C Slika 1: Sulfidna kapaciteta sistema CaO-CaF2-Al2O3 pri 1500 °C Equation 8 shows that CaF2 decreases the negative effect of Al2O3 on CaS and it is defined by k1(CaF2). With a calculation using equations 8 and 9, the effect of CaF2 on Cs, for the constant amounts of Al2O3 and CaO, was examined. Therefore, equations 7, 8 and 9 provide the basis for deriving the RMJ model. The procedure was repeated for different amounts of these components under the conditions of secondary metallurgy that are analyzed. The values of Cs were analyzed cumulatively so that the influence of Al2O3 and CaO was presented indirectly through CaF2. Equations 8 and 9 show that the k1(CaF2) and k2(CaF2) values are always positive. This is in accordance with the fact that an increase in the CaF2 amount increases the sulphide capacity. The primary effect of CaF2 was defined on the basis of the optical-basicity values. However, the secondary effect of CaF2 can be considered through a decrease in the slag viscosity and an increase in the CaO activity. Based on the above-mentioned procedure, the Cs' values, using the RMJ model (eq.7), and Cs, using the Young model (eq.4), were determined. On the basis of Figure 3: Influence of basicity on sulphide capacity Slika 3: Vpliv bazi~nosti na sulfidno kapaciteto the chemical composition of the slag, optical basicity was calculated with the Nakamura method4. In all the considered cases, the compared values of Cs' are higher than the Cs values. The relationship between the sulphide capacity and the optical basicity for the two above-mentioned models is shown in Figure 2. It is likely that the Cs values increase with an increase in the optical basicity. This is the basis of the CaF2 effect on the sulphide capacity and desulphurisation. All the values of the sulphide capacity calculated with the new RMJ model are higher than the ones calculated with the Young model. It was noticed that Cs increases faster with the increasing CaF2 amount when the RMJ model is used because of its effect on k1(CaF2) and k2(CaF2). Figure 3 shows the variation of Cs as the function of the slag basicity, indicating that the sulphide capacity increases with the increased B values. For the slag with B <2, the values of Cs obtained with the RMJ and Young models are approximately equal. This is a direct result of the small values of correction factors k1(CaF2) and k2(CaF2) in the RMJ model. In the later stages, for B > 2, its values significantly increase and the differences between the Cs values of the two models are higher. The Cs values are higher for the CaO-CaF2 mixture than for the triple CaO-CaF2-WB mixture because of the Al2O3 presence in white bauxite. Figure 2: Sulphide capacity as the function of optical basicity Slika 2: Sulfidna kapaciteta v odvisnosti od opti~ne bazi~nosti Figure 4: Influence of CaF2 on the values of k1(CaF2) factor Slika 4: Vpliv CaF2 na vrednosti faktorja k1(CaF2) Table 2: Coefficient of sulphur distribution for the investigated slags Tabela 2: Koeficient razporeditve žvepla pri preiskovanih žlindrah Slag i 2 3 4 5 6 7 8 Ls 53.7 73.3 69.i 76.6 80.8 i29.3 iii.4 99.8 Ls'(CaO-CaF2) 58.8 79.2 75.6 83.8 88.5 i4i.6 i22.0 i09.8 Ls'(CaO-CaF2-WB) 54.i 75.3 7i.i 8i.i 85.0 i3i.2 - - 0,16 0.14 0.12 0.10 0.08 0.06 0.04 0.02 0,00 CaO-CaF2 jt*^,o-tiü CaO-CaF2-WB 10 CaF; (mass.%) Figure 5: Influence of CaF2 on the values of k2(CaF2) factor Slika 5: Vpliv CaF2 na vrednosti faktorja k2(CaF2) Figure 6: Effect of Cs on sulphur distribution Slika 6: Vpliv C s na razporeditev žvepla The typical values of ki(CaF2) and k2(CaF2), as functions of the CaF2 amount, are shown in Figures 4 and 5. In general, it may be noticed that ki(CaF2) and k2(CaF2) are proportional to the CaF2 amount. Besides, for w(CaF2) < 8 %, the relationship is approximately linear, but for w(CaF2) >8 %, the intensity of the kCaF2 changes is lower. The values of ki(CaF2) and k2(CaF2) are lower for CaO-CaF2-WB than for the CaO-CaF2 slag mixture. It is noticeable that a higher amount of CaF2 and a lower Al2O3 amount cause an increase in the correction-factor values. Besides, the results in the case of the CaO-CaF2-WB mixture are more scattered than for CaO-CaF2. The coefficient of sulphur distribution is calculated with equation 3. The values of Ls (the RMJ model) and Ls' (the Young model) are presented in Table 2. The effect of Ls on the sulphide capacity for the two considered models is shown in Figure 6. It can be seen that, at the lower Cs values, the relation between Ls and Cs is almost linear. However, at Cs > 0.0015, the Ls values strongly increase, and the results are more scattered in all the cases. It is likely that the degree of desulphurisation is higher for the RMJ model than for the Young model. This is because positive values of ki(CaF2) and k2(CaF2) allow a larger CaO activity that is generally observed in practice. The probability factors of polinomial dependence are higher for the RMJ model (R2 = 0.97 and R2 = 0.972) than for the Young model (R2 = 0.956). 4 CONCLUSION The results of the new analytical RMJ model, presented in this paper, indicate that a usage of correction factors ki(CaF2) and k2(CaF2) is required. The secondary effect of CaF2 on the degree of desulphurisation is fully defined with these factors. Using white bauxite as a slag component, the values of k2(CaF2) decrease and the degree of desulphurisation is lower. The relative effect of CaF2 on the values of kCaF2 and Ls at the basicity of B >2 was indicated. The differences in the values of Cs for the RMJ and Young models increase with the increasing kCaF2. This difference was caused by the secondary effect of CaF2 on the CaO and Al2O3 activities as well as the properties of the slag. The presence of Al2O3 in white bauxite causes a lower sulphide capacity, and an optimum component amount in the steelmaking slag mixture is a condition for good desulphurisation. 5 REFERENCES i A. Mc Lean, Y. Yang, I. Sommervile, The Science and Technology of Slags for Iron and Steelmaking, Proceedings of 6th Int. Conference, Stockholm, 2000, i27 2T. Tsao, H. G. Katayama, Trans. ISIJ, 26 (1986), 7i7 3 H. Gaye, P. Riboud, Proceed. of 5th Int. Iron and Steel Con., 6 (1986), 63i 4 A. Gosh, Secondary Metallurgy, CRC Press, New York 2000, i95 5 W. Young, A. Duffy, Use of optical basicity concept for determining sulphur slag-metal partitions, Ironmaking and Steelmaking, i9 (1992), 20i 6 M. M. Nzotta, M. Andersson, P. Jonsson, S. Seetaraman, A study on the sulphide capacities of steelmaking slags, Scand. Journal of Metallurgy, 29 (2000), i77 7 N. Bannenberg, H. Lachmud, Proceedings of Steelmaking Conf. Proc., Chicago, 77 (1994), i35 8 K. Mills, The estimation of slag properties, Journal of The Southern African Institute of Mining and Metallurgy, iii (2011), 649