ISIJ International, Vol. 49 (009), No., pp. 56 63 Sulphide Capacities of CaO Al O 3 SiO MgO MnO Slags in the Temperature Range 673 773 K Yoshinori TANIGUCHI,,) Nobuo SANO 3) and Seshadri SEETHARAMAN ) ) Division of Materials Process Science, Royal Institute of Technology, S-00 44 Stockholm Sweden. ) Nippon Steel Corporation, -6-3, Otemachi, Chiyoda-ku, Tokyo 00-807 Japan. 3) Professor Emeritus, The University of Tokyo, 7-3- Hongo Bunkyo-ku Tokyo 3-8656 Japan. (Received on August, 007; accepted on November 7, 008) With a view to estimate the sulphide capacities of slags used in hot metal pretreatment processes, the sulphide capacities of CaO Al O 3 SiO MgO, CaO Al O 3 SiO MnO and CaO Al O 3 SiO MgO MnO slags in low SiO concentration region were measured in the temperature range 673 773 K. The gas slag equilibrium technique has been used for these measurements. From the results obtained, it was found that the MgO and MnO increased the sulphide capacity values of slag. A new empirical model based on optical basicity for sulphide capacity prediction of these slags was developed using the measured values of the present work and literature. KEY WORDS: sulphide capacity; hot metal pretreatment; desulfurization; slag; CaO Al O 3 SiO MgO; CaO Al O 3 SiO MnO; CaO Al O 3 SiO MgO MnO; sulphide capacity model.. Introduction In Steelmaking processes including hot metal pretreatment, CaF containing fluxes had an important role in desulfurization for the past several decades to increase the sulfur absorption ability of desulfurization agent. However, it has become increasingly difficult to use CaF because of the environmental problems, caused by fluoride emissions. Serious efforts are being made to find a fluoride-substitute. In the case of hot metal treatment, the practically used flux is CaO-based. In view of the carry-over of a certain quantity of blast furnace slag in the hot metal ladle and materials which are added to improve desulfurization reaction, high basicity CaO Al O 3 SiO -based multi-component slag is likely to be formed during the hot metal pretreatment. While the experimental studies of the sulphide capacity values for CaO Al O 3 SiO -based slag are abundant in literature, 5) to the knowledge of the authors, only a few measurements of the sulphide capacity of slags at hot metal pretreatment temperature have been reported. Considering that the solubility of CaS in CaO is low, 6) it is very important to find a low melting slag which coexists with solid CaO and has high solubility of sulfur in order to improve the desulfurization reaction in hot metal pretreatment process. MgO has been known 7) to decrease the melting point of CaO Al O 3 SiO slags. Thus, it would be useful to investigate the effect of the addition of MgO to the ternary CaO Al O 3 SiO slags in low SiO concentration on the sulphide capacities. Further, it is noted that MnO increases the sulphide capacity of a slag. ) However, the sulphide capacities of CaO Al O 3 SiO MgO MnO slag at low SiO contents have not been reported in literature so far. In the present work, the sulphide capacities of CaO Al O 3 SiO MgO, CaO Al O 3 SiO MnO and CaO Al O 3 SiO MgO MnO were determined in the temperature range 673 773 K.. Experimental The classical gas slag equilibrium technique was employed in the present study for the measurement of sulphide capacities. The schematic diagram of experimental set-up is shown in Fig.. A detailed description of the experimental setup and the procedure adopted are given elsewhere. 4) In a general run, two to four Pt crucibles (ID: 0.06 m, Depth: 0.008 m) or one to two CaO crucibles (ID: 0.05 m, Depth: 0.07 m) with the slag samples of g were pushed inside the furnace kept at 673 K and kept for h in purified Ar gas stream for pre-melting. The Ar CO CO SO gas mixture was then introduced into the reaction tube and temperature was changed to the target value. A constant flow rate of 500 ml/min was maintained during the equilibration of the slag with the gas mixture at the experimental tempera- Fig.. Schematic diagram of experimental setup. 009 ISIJ 56
ISIJ International, Vol. 49 (009), No. ture for at least 6 h. This time interval was found to be sufficient for the slag to reach the equilibrium with gas.,4) After equilibration, the samples were pulled out toward the cold end of the furnace and quenched with the gas mixture flowing. The gas atmosphere was changed to argon and the samples were taken out and subjected to chemical analysis. Sulfur content was analyzed using a LECO combustion-infrared spectrometer (LECO is a trademark of LECO Corporation, St. Joseph, MI). Some of the slag samples were analyzed for Ca, Si and Mn in order to ascertain any possible composition change during the experiment. These analyses were carried out using an ICP emission spectrometry. The results of chemical analysis were in good agreement with the original weighed-in compositions. 3. Results and Discussion The concept of sulphide capacity as expressed by Richardson and Fincham ) is as follows: C S / PO (%S) / P S...() where, (%S) is the mass% of sulfur content in slag. In the present experiments, the partial pressures of sulfur and oxygen were calculated using Thermo-Calc version R with the SSUB3 database. In order to check the reliability and accuracy of the SSUB3 database, Hayakawa et al. s 5) results were recalculated using the database. The calculated oxygen and sulfur partial pressures were in good agreement with their results. The partial pressure of oxygen in the gas mixture just above the sample was maintained between 3.5 0 4 and.4 0 3 Pa and the partial pressure of sulfur was found to vary between 8.4 0 and. 0 3 Pa. 3.. Sulphide Capacity Measurement 3... CaO Al O 3 SiO MgO System The experimental compositions and results are summarized in Tables and. The values of optical basicity were calculated using the following coefficients for each oxide reported by Sosinsky and Sommerville 8) : CaO.0, Al O 3 0.6, SiO 0.48 and MgO 0.78. The error ranges of log C S are mainly due to standard deviations in the chemical analysis of sulfur. Figure shows the results of the present sulphide capacity measurements of the CaO Al O 3 SiO MgO slags at 773 K. In Fig., the results for the same slag system reported by Abraham and Richardson 4) as well as Hayakawa et al., 5) the results for CaO Al O 3 SiO slags by Carter and Macfarlane ) as well Table. Experimental compositions in CaO Al O 3 SiO MgO slag system. Fig.. Sulphide capacities (log 0 C S ) of CaO Al O 3 SiO MgO slags at 773 K. Table. Experimental results in CaO Al O 3 SiO MgO slag system. 57 009 ISIJ
ISIJ International, Vol. 49 (009), No. Fig. 3. Sulphide capacities (log 0 C S ) of CaO Al O 3 SiO MgO slags at 73 K. Fig. 5. Effect of MgO on sulphide capacity at 673 K. Fig. 4. Sulphide capacities (log 0 C S ) of CaO Al O 3 SiO MgO slags at 673 K. as Drakaliysky et al. 0) and the results for CaO Al O 3 slags by Carter and Macfarlane, ) Sharma and Richardson, 9) Kor and Richardson 0) as well as Drakaliysky et al. ) are also plotted for comparison. The dashed line in the figure indicates an iso-sulphide capcity line. All these results agree well with the results obtained in the present work. The results with 4 mass% MgO at 73 and the results with 0 and 4 mass% MgO at 673 K are shown in Figs. 3 and 4 respectively. In Figs. 3 and 4, the results for CaO Al O 3 SiO MgO slags by Hayakawa et al. are plotted for comparison. The dashed lines in these figures indicate iso-sulphide capacity line as with Fig.. It is evident from Figs. 3 and 4 that the relationship between log C S and the slag composition obtained by the present authors and those by Hayakawa et al. at all three temperatures, viz. 673, 73 and 773 K are quite similar indicating thereby that the present results are in good agreement with earlier results. Figure 5 shows the effect of MgO on the sulphide capacity at 673 K. As shown in Fig. 5, the sulphide capacity is found to increase by the addition of MgO. Kalyanram et al. 3) measured the activity of CaO in CaO Al O 3 SiO and CaO Al O 3 SiO MgO slag in high SiO concentration at 773 K. The sulphide capacity was not reported directly in their work, but can be calculated from their results. The calculated results indicate that the addition of MgO increases the sulphide capacity of CaO Al O 3 SiO slag. Hino et al. 7) as well as Ohta et al. 3) reported that the addition of MgO increases the sulphide capacity of CaO Al O 3 slag. However, it is also evident from Fig. 5 that there is no substantial effect of MgO on the sulphide capacity when it increases from to 6 mass%. Table 3 shows the effect of exchanges of CaO and MgO on sulphide capacities in MgO Al O 3 SiO, ) CaO Al O 3 MgO, 7) CaO SiO MgO 3) and CaO Al O 3 SiO MgO 6) systems. As shown in Table 3, CaO and MgO do not have the equivalent effect on sulphide capacities but it is also clear that the difference is very small in case of CaO SiO MgO and CaO Al O 3 SiO MgO systems. This indicates that MgO has a similar effect as CaO on sulphide capacities when CaO and SiO coexist in the slag system but the effect becomes small if the slag system does not have both of them. Thus, it is possible that the effect of MgO addition on sulphide capacities becomes small with the increasing of MgO if the SiO content of the slag is low. In the present work, the SiO content is lower than 8 mass%. This may explain the change of the effect of MgO addition at around mass%. The difference in the effect of MgO addition may perhaps be explained on the basis of the solubility of MgS in the slag; but no information on the MgS solubility in slags is available in literature. Further work in this regard is planned currently in the present laboratory. In Fig. 5, the predicted values of sulphide capacity by Sommerville s model, 8) Young s model 3) and the KTH model,4,5) are also plotted for comparison. It was found that all these models underestimate the effect of MgO on the sulphide capacity at 673 K. This is probably because these models had been optimized using the sulphide capacity values measured in high SiO concentration at high temperatures. It is to be noted that the predictions of the KTH model with respect to the change of log C S with MgO concentration show an opposite trend. Figure 6 shows the relationship between log C S and temperature. In this figure, the results by Nzotta et al., ) Shankar et al. as well as Hayakawa et al. are plotted for comparison. It is evident from Fig. 6 that the slope of log C S vs. 0 4 /T of the present work is in very good agreement with those by Hayakawa et al. 5) in the C S range examined and the plots are linear. It also can be seen in the figure that the slope slightly increased with the increasing of optical basicity. 3... CaO Al O 3 SiO MgO MnO System The experimental compositions and results are summarized in Tables 4 and 5. The values of optical basicity were also calculated using the following coefficients for each oxide reported by Sosinsky and Sommerville 8) : CaO.0, Al O 3 0.6, SiO 0.48, MgO 0.78 and MnO.. The error ranges of log C S are mainly due to standard deviations in the chemical analysis of sulfur. In the present work, two different partial pressures of oxygen were employed for all 009 ISIJ 58
ISIJ International, Vol. 49 (009), No. Table 3. Effect of exchanges CaO and MgO on sulphide capacities in MgO Al O 3 SiO, CaO Al O 3 MgO, CaO SiO MgO and CaO Al O 3 SiO MgO systems. Fig. 6. Relationship between log 0 C S and temperature in CaO Al O 3 SiO MgO slags. Numerical values show optical basicity. the measurement temperatures. Figure 7 shows the sulphide capacity in both cases at 73 K. It is evident from this figure that there is a difference in the sulphide capacity values except for the slag which had no MnO. The difference between the sulphide capacity values with 5 mass% MnO is about 0. in base 0 logarithm. A similar tendency was seen at 673 as well as 773 K. According to Tamura et al., 6) it is considered that the valency of Mn in the slag phase is almost Mn at the partial pressures of oxygen used in the present work. However, the sulphide capacity values of the lower partial oxygen pressure were used for the analysis in the present work. The reasons for this are that the basicity of the present work is higher than Tamura et al. s experiments, the difference was seen at all the measured temperatures and the sulphide capacities of 0 mass% MnO slags showed excellent agreement with each other. Because the durability of a Pt crucible at the lower oxygen partial pressure was not good, the measurement at an even lower partial pressure of oxygen was not conducted in the present work. Figure 8 shows the relationship between log C S and MnO content. As shown in the figure, the sulphide capacities increased with the increasing of MnO content at all the measurement temperatures. Figure 8 also shows the effect of MgO on the sulphide capacity of the slags. It is evident from this figure that, at 673 K, MgO shows a slight tendency to increase the sulphide capacity within the range of 0 to 5 mass% MnO but there is no noticeable effect beyond 0 mass% MnO. The reason for this may be that the increase of sulfur content by the addition of MnO is much larger than the increase by the addition of MgO. The results with 5 mass% MnO and 0 mass% MnO are plotted in phase diagrams in Figs. 9 and 0 respectively. The dashed lines in these figures indicate iso-sulphide capacity lines. It is evident from these figures that the relationship between log C S and the slag composition is similar to the case of CaO Al O 3 SiO MgO system. The measured values were compared with the values calculated using Sommerville s model, 8) Young s model 3) and the KTH model.,4,5) The results of a comparison at 673 and 773 K are shown in Figs. and respectively. As shown in these figures, Sommerville s model agreed well with the measured values at 773 K but the measured values at 673 K were significantly higher than the calculated values using all of the models. As mentioned above, the sulphide capacities of this slag system at low SiO contents have not been reported and these models had been optimized using the experimental values measured at higher than 673 K, the predict values at 673 K may have errors from extrapolation. Figure 3 shows the relationship between log C S and temperature. The slopes of slags containing 5, 0 and 5 mass% MnO in the present work were 59 009 ISIJ
ISIJ International, Vol. 49 (009), No. Table 4. Experimental compositions in CaO Al O 3 SiO MgO MnO slag system. Table 5. Experimental results in CaO Al O 3 SiO MgO MnO slag system. 009 ISIJ 60
ISIJ International, Vol. 49 (009), No. Fig. 7. Effect of MnO on sulphide capacity in CaO Al O 3 SiO MgO MnO slags at 73 K at two different oxygen partial pressures. Fig.. Comparison between calculated and measured sulphide capacity for CaO Al O 3 SiO MgO MnO slags at 673 K. Fig. 8. Effect of MnO on sulphide capacity in CaO Al O 3 SiO MgO MnO slags. Fig.. Comparison between calculated and measured sulphide capacity for CaO Al O 3 SiO MgO MnO slags at 773 K. Fig. 9. Sulphide capacity (log 0 C S ) of CaO Al O 3 SiO MgO 5 mass% MnO slags at 673 K. Fig. 3. Relationship between log 0 C S and temperature in CaO Al O 3 SiO MgO MnO slags. Fig. 0. Sulphide capacity (log 0 C S ) of CaO Al O 3 SiO MgO 0 mass% MnO slags at 673 K. found to be.04, 0.88 and 0.7 respectively. In Fig. 3, the results for CaO Al O 3 SiO MgO MnO slags by Nilsson et al. 9) and Nzotta et al. ) are also plotted for comparison. The slopes of Nilsson et al. containing 6.9 mass% MnO and Nzotta et al. containing 3.83 mass% MnO were found to be 0.68 and.04 respectively. Their values are in very good agreement with the slopes of the slags containing 5 and 5 mass% MnO in the present work. This indicates that MnO addition affected the temperature dependence of the sulphide capacity and increased the slope of log C S vs. 0 4 /T. Shankar et al. 4) mentioned a relationship between the enthalpy value for desulfurization and the 6 009 ISIJ
ISIJ International, Vol. 49 (009), No. slope of log C S vs. 0 4 /T. For the slag system having one basic oxide, desulfurization takes place by the following reaction: MO(s) S (g) MS(s) O (g)...() DG DH TDS (J/mol)...(3) 4 0 log CS 0. 57 0 5 H 0. 057 S T γ log (%S) γ MO MS X X...(4) where, g MO and g MS are the activity coefficients of solid MO and MS respectively. The standard state of the activity of them is pure solid. X MO and X MS are the mole fractions of MO and MS respectively. If it is assumed that DH, DS, g MO and g MS are constant, the slope of log C S vs. 0 4 /T can be expressed by DH. In the present work, the slag system has CaO and MnO. Therefore, sulfur removal takes place by the following reaction: CaO(s) S...(5) (g) CaS(s) O (g) DG 9 000.55T (J/mol) 7)...(6) MnO(l) S (g) MnS(l) O (g)...(7) DG 74 606 7.65T (J/mol) 8,9)...(8) It is clear in the above equations that the enthalpy for desulfurization by MnO is smaller than that by CaO. Thus, there is a possibility that the slope of log C S vs. 0 4 /T increases with the increasing of MnO content. In fact, the activity coefficients of all components must be taken into account but it is expected that the MnO addition is more effective at lower temperatures. As mentioned above, the slope of the slag containing 5 mass% MnO was 0.7. In the case of this slag, the slope of Sommerville s model was.06 and the slope of Young s model was.7. There was a significant difference between these values. This could explain the difference between the calculated and measured values at 673 K. It should be mentioned that in the experiment in which CaO crucibles were used or optical basicity of the slag is more than 0.83, the slags crept out of the crucible. Thus, it was not possible to analyze the samples. The same phenomenon was also seen in CaO Al O 3 SiO MgO slags. This indicates that the slag shows a high wetting behavior with respect to the crucible wall. In real process, with solid CaO present in the slag and the prevalence of relatively low oxygen and sulfur potentials in the hot metal, there is likely to be a strong affinity between the CaO particles and liquid slag phases leading to highly wetting conditions. MO MS Fig. 4. Comparison between predicted and measured sulphide capacities for CaO Al O 3 SiO MgO MnO slags. 3.. Sulphide Capacity Model As mentioned above, Sommerville s model, 8) Young s model 3) and the KTH model,4,5) underestimate the sulphide capacity of CaO Al O 3 SiO MgO MnO slags in low SiO concentration at 673 K. Thus, it would be appropriate to have a new sulphide capacity model which is applicable to the prediction at 673 K. For this reason, a new model which is based on optical basicity has been developed using the measured values of the present work and literature. 4,6 9,,5,8,9) Prediction of sulphide capacity by this model is shown in Fig. 4. It is evident from this figure that this model predicts sulphide capacity very well over the whole range of measured sulphide capacity for the present slag system as compared to other models. log CS 7. 350 94. 89 log Λ 0 05 Λ( 338(% MgO) 87(% MnO)) T 0. 84(% SiO ) 0. 379(% Al O ) 3 0. 0587(% MgO) 0. 084(% MnO)...(9) This equation has been developed for slags having CaO: 0 to 63 mass%, Al O 3 : 0 to 65 mass%, SiO : 0 to 68 mass%, MgO: 0 to 5 mass%, MnO: 0 to 30 mass% and temperature: 673 to 98 K. A total of 306 data points were used for the regression and the multiple correlation coefficient was 0.99. The range of this model is applicable to iron-making, hot metal pretreatment as well as secondary refining slags. 4. Conclusions In the present work, sulphide capacity of CaO Al O 3 SiO MgO, CaO Al O 3 SiO MnO and CaO Al O 3 SiO MgO MnO were experimentally measured in the temperature range 673 773 K using the gas slag equilibrium method. In CaO Al O 3 SiO MgO system, the MgO increased the sulphide capacity of slag but there was no substantial effect between and 6 mass% MgO. In the case of CaO Al O 3 SiO MgO MnO system, the MnO increased the sulphide capacity of slag and its addition was more effective at lower temperatures. The MgO addition increased the sulphide capacity of slag within the range of 0 009 ISIJ 6
ISIJ International, Vol. 49 (009), No. to 5 mass% MnO but there was no substantial effect above 0 mass% MnO. Using the measured value of the present work and data from literature, a new empirical model based on optical basicity for sulphide capacity estimation of CaO Al O 3 SiO MgO MnO slag was developed. This model is applicable to iron-making, hot metal pretreatment as well as secondary refining slags. Acknowledgement The authors are thankful to Nippon Steel Corporation, Japan for sponsoring the present work and the financial support for the same. REFERENCES ) F. D. Richardson and C. J. B. Fincham: J. Iron Steel Inst., 78 (954), 4. ) P. T. Carter and T. G. Macfarlane: J. Iron Steel Inst., 85 (957), 54. 3) M. R. Kalyanram, T. G. Macfarlane and H. B. Bell: J. Iron Steel Inst., 95 (960), 58. 4) K. P. Abraham and F. D. Richardson: J. Iron. Steel Inst., 96 (960), 33. 5) J. Cameron, T. B. Gibbons and J. Taylor: J. Iron Steel Inst., 04 (966), 3. 6) K. Kärsrud: Scand. J. Metall., 3 (984), 44. 7) M. Hino, S. Kitagawa and S. Ban-ya: ISIJ Int., 33 (993), 36. 8) M. Görnerup and O. Wijk: Scand. J. Metall., 5 (996), 03. 9) R. Nilsson: Doctor Thesis, Royal Institute of Technology, Sweden, (996), Supplement 6. 0) E. Drakaliysky, R. Nilsson and S. Seetharaman: Can. Metall. Q., 36 (997), 5. ) M. M. Nzotta, D. Sichen and S. Seetharaman: ISIJ Int., 38 (998), 70. ) M. M. Nzotta: Doctor Thesis, Royal Institute of Technology, Sweden, (999), Supplement VI. 3) M. Ohta, T. Kubo and K. Morita: Tetsu-to-Hagané, 89 (003), 8. 4) A. Shankar, M. Görnerup, A. K. Lahiri and S. Seetharaman: Metall. Mater. Trans. B, 37B (006), 94. 5) H. Hayakawa, M. Hasegawa, K. Oh-nuki, T. Sawai and M. Iwase: Steel Res., 77 (006), 4. 6) Slag Atlas, nd ed., ed. by VDEh, Verlag Stahleisen GmbH, Düsseldorf, (995), 0. 7) Slag Atlas, nd ed., ed. by VDEh, Verlag Stahleisen GmbH, Düsseldorf, (995), 60. 8) D. J. Sosinsky and I. D. Sommerville: Metall. Trans. B, 7B (986), 33. 9) R. A. Sharma and F. D. Richardson: J. Iron Steel Inst., 98 (96), 386. 0) G. J. W. Kor and F. D. Richardson: J. Iron Steel Inst., 06 (968), 700. ) E. Drakaliysky, R. Nilsson, D. Sichen and S. Seetharaman: High Temp. Mater. Proc., 5 (996), 63. ) M. M. Nzotta: Scand. J. Metall., 6 (997), 69. 3) R. W. Young, J. A. Duffy, G. J. Hassall and Z. Xu: Ironmaking Steelmaking, 9 (99), 0. 4) R. Nilsson, D. Sichen and S. Seetharaman: Scand. J. Metall., 3 (994), 8. 5) M. M. Nzotta, D. Sichen and S. Seetharaman: Metall. Mater. Trans. B, 30B (999), 909. 6) Y. Tamura, S. Nakamura and N. Sano: Tetsu-to-Hagané, 73 (987), 4. 7) E. T. Turkdogan, B. B. Rice and J. V. Vinters: Metall. Trans., 5 (974), 57. 8) E. T. Turkdogan: Physical Chemistry of High Temperature Technology, Academic Press, New York, (980), 5. 9) Steelmaking Data Source Book, The Japan Society for the Promoting of Science, The 9th Committee on Steelmaking, Gordon and Breach Science Publishers, New York, (988), 6. 63 009 ISIJ