SILICON AND MANGANESE PARTITION BETWEEN SLAG AND METAL PHASES AND THEIR ACTIVITIES PERTINENT TO FERROMANGANESE AND SILICOMANGANESE PRODUCTION
Hakan Cengizler1 and R Hurman Eric2
1TMYO, Celal Bayar University, Turgutlu, Turkey
2 Metallurgy and Materials Engineering, University of the Witwatersrand, Johannesburg, South Africa [email protected] , phone: +27117177537
Keywords: Ferromanganese, silicomanganese, slag, metal, equilibrium
Abstract
Equilibrium between MnO-CaO-MgO-SiO2-Al2O3 slags and carbon saturated Mn-Si-Fe-C alloys was investigated under CO at 1500oC. Manganese and silicon activities were obtained by using the present data and the previously determined MnO and SiO2 activities of the slag. Quadratic multi-coefficient regression equations were developed for activity coefficients of manganese and silicon. The conclusions of this work are:(i)increase in the basicity and the CaO/Al2O3 ratios decreases the Mn distribution ratio,(ii)increase in the silica concentration and the MgO/CaO ratio increases the Mn distribution ratio, iii)carbon and manganese as well as carbon and silicon of the metal phase are inversely proportional,(iv)as Mn/Fe and Mn/Si ratio increases in the metal the carbon solubility increases,(v)decrease in the basicity increases the silicon content of the metal and (vi)increase in the silica content of the slag increases the silicon content of the metal and this effect is more pronounced at the higher Mn/Fe and Mn/Si ratios.
Introduction
High-carbon ferromanganese is the most common alloy used as an additive in the steel industry to control deoxidation of steel and to produce manganese-bearing alloys. Two processes are employed in the manufacture of high carbon ferromanganese referred to as the low- (discard) and high- (enriched) slag practice. The low slag practice is the least economical in terms of production costs per ton of ferromanganese and it produces the normal grades of high-carbon ferromanganese with secondary slags ranging from 8 to 15 per cent contained MnO. This practice is employed when the manufacturer does not produce silicomanganese for which a slag of higher manganese content can be used as a raw material or has no sale for the higher MnO containing slag. In high slag practice, the slags produced during manufacture of high-carbon ferromanganese contain between 25 and 40 per cent MnO. These manganese- rich slags are used for the production of silicomanganese by carbothermic reduction due to their extremely high Mn-to Fe ratio, high silica content and low phosphorus content. In manufacturing high- carbon ferromanganese and silicomanganese in the smelting furnace, the factor which determines the quality of the product, the efficiency of the process and the final recovery rate of manganese is the distribution of the valuable manganese between the alloy, slag and gas phases and the phase equilibria among them. In other words, the efficiency is governed, to a large extent, by the thermodynamic activities of MnO in the slag1, 2 and that of Mn in the alloy phase. An extensive experimental data and information on the thermodynamic activity of MnO in MnO-CaO-MgO-Si0-AlO synthetic ferromanganese slags
Advances in Molten Slags, Fluxes, and Salts: Proceedings of The 10th International Conference on Molten Slags, Fluxes and Salts (MOLTEN16) Edited by: Ramana G. Reddy, Pinakin Chaubal, P. Chris Pistorius, and Uday Pal TMS (The Minerals, Metals & Materials Society), 2016
were gathered and was reported1-3 elsewhere. Some noteworthy relationships from literature pertinent to this work are the following. Firstly, the inverse relationship between the silicon and carbon in the metal phase and the increase in carbon solubility of the melt with a decrease in Fe/Mn ratio can be mentioned4, 5-
10. Although, the reported data5-10 are not in good agreement with each other, a general tendency is that the increasing carbon and silicon additions lower the manganese activity leading to large negative deviations from ideality in liquid manganese. Although it appears that a substantial amount of information has been available on manganese containing liquid alloy systems, the majority of these are related to conditions in iron and/or steelmaking where manganese is in low concentrations. Furthermore, there are few studies on thermodynamics and slag-metal equilibria pertinent to high-carbon ferromanganese and silicomanganese smelting at typical process temperature of 1500oC. The present research work is directed to fill in some of these gaps.
Experimental procedure
A gas-tight vertical molybdenum resistance furnace has been assembled and employed for the study of slag-metal equilibrium. The reaction tube was continuously purged with chemically pure CO gas which was introduced from the bottom end of the alumina furnace work tube and was vented to the fume hood through the gas outlet on the top of the work tube. Actual sample temperatures in the furnace were measured with a Pt-6 %Rh/Pt-30%Rh (B-type) thermocouple. In the slag-metal distribution experiments, graphite crucibles were used. The slag was prepared as a homogenous mixture made from its pure components. In total, 55 slag samples of different compositions containing MnO, CaO, MgO, SiO2 and Al2O3 were prepared from analytical-reagent grade oxide powders. However, prior to use, all the oxides except MnO were first calcined in an electrical resistance-heated muffle furnace for about 24 hours at 1200°C. After cooling, the required amount of CaO, MgO, SiO2 and Al2O3 were carefully weighed and thoroughly mixed in the desired proportions to give slags of the designed compositions in an agate mortar under a liquid blanket of acetone until all the acetone had vaporized. The dried mixtures were then pressed into disks. These disks were re-calcined for 12 hours at about 1200°C in order to promote sintering, and thereby homogenization of the samples. On cooling, the homogenized disks were crushed and ground in an agate mortar and mixed with required quantities of reagent-grade MnO in an agate mortar as described above before every individual run. The slag basicity ranged from about 0.4 to 1.2 and MnO contents ranged from 5 to 30 mass per cent. The initial Al2O3 content of slags were kept at 5 mass per cent level. CaO content varied in the range 20 to 35 mass percent, MgO in the 5 to 17.5 mass percent and SiO2 in the 27 to 58 mass percent range. The metal charge which consisted of Mn, Si, C and Fe was prepared from electrolytic manganese, iron, high-purity silicon and spectrographic-grade graphite powders. From these, a number of master alloys were prepared by melting the elements under an argon atmosphere in graphite crucibles. The composition of the alloys for the experimental runs was adjusted by adding the necessary pure elements in powder form to these master alloys. The ranges of initial alloy compositions as determined by chemical analysis are shown in Table 1. It was found that a minimum of 10 hours was necessary for equilibration1. In the slag-metal equilibrium distribution experiments, the charge consisted of about 6 g of metal and 4 g of slag of selected compositions and was contained in graphite crucibles. The reaction tube of the furnace was flushed for about 15 minutes with spectrographically pure argon before the CO atmosphere was employed. The flow rate of CO was kept at 600 cm3/min. After equilibration at 1500oC the CO gas was turned off and the argon was allowed to pass through the reaction tube and the crucibles were lowered to the bottom water-cooled end of the furnace quickly. The apparatus was opened from the bottom, and the graphite crucibles with their contents were immediately quenched in water. The crucible was broken and the slag and metal phases were removed from the crucible and separated. Both phases were crushed, ground, pulverized and were
Table 1. The initial metal-phase compositions.
kept in sealed bottles in desiccators until their contents were analysed. The analysis of alloys was done for Mn, Si, Fe and C. The whole composition of slag was analysed and the total composition amounted to within 98 and 101 mass per cent.
Results and Discussions Manganese distribution
The effect of basicity ratio on the logarithm of the distribution ratio of manganese is seen in Figure 1.
The distribution ratio of a species “i” is defined as per cent of “i” in the slag divided by per cent of “i”
in the metal phase. In all the figures, round and square brackets indicate the slag and metal phases respectively. Figure 1 sows that an increase in the basicity ratio causes an increase in the manganese content of the metal phase relative to that of the slag. It is well known11-14 that the interaction of Ca2+
ions with SiO2 is stronger than those of mixed Ca2+ and Mg2+ ions and Mg2+. Thus, one can conclude that an increase in the amount of CaO can lead to an increase in the formation of Ca2SiO4 orthosilicate, leading to higher aMnO values in the slags. The increase in the aMnO values favours the transfer of manganese into the metal phase.
Figure 2 shows the relation between CaO-to-Al2O3 ratio and the Mn distribution between slag and metal phases. The Si-to-Fe ratio is 0.23 in the metal phase. As it is seen in the figure, when CaO replaces Al2O3 in the slag, the Mn transfer to the metal phase increases. This can also be explained by referring to increased MnO activities by an increase in the concentration of CaO in the slag.
In the metal phase, the carbon and manganese concentrations are directly proportional to each other as seen in Figure 3. An increase in manganese concentration increases the saturation limit of carbon in Mn-Fe-Si-C ferroalloys. Linear regression equation was derived to predict the carbon concentration (in mass per cent) and very high correlation coefficient (r2=0.930) was obtained. At 1500oC, the regression equation is: C% = 30.6952 + 5.4453 Mn%.
Silicon distribution
The relationship between the equilibrium solubility of carbon (at saturation) and silicon in Mn-Fe-Si-C alloys is illustrated in Figure 4 which is applicable for ferromanganese alloys where three different data groups with different Mn-to-Fe ratios (3.7, 6.7, and 7.4) are plotted. It can be seen that, at constant silicon contents, as the Mn-to-Fe ratio increases, the carbon solubility in the metal phase increases. The carbon and silicon concentrations in the metal phase are inversely related as expected and this behaviour can be explained in terms of the stability of carbides. As silicon lowers the solubility of carbon, the conditions becomes less favourable for carbide formation. Thermodynamically, this is an indication that silicon lowers the activity of carbon.
Phase Mn, % Fe, % Si, % C, %
Ferromanganese 82.19 9.34 1.50 7
79.77 11.73 1.50 7
74.56 16.94 1.50 7
Silicomanganese 66.64 19.30 12.34 1.7 62.55 19.30 16.46 1.7 57.04 19.30 21.94 1.7
Figure 1. The effect of basicity ratio on Figure 2. The effect of CaO/Al2O3 ratio on manganese distribution at 1500oC manganese distribution at 1500oC under CO atmosphere. under CO atmosphere. Si/Mn=0.23
Figure 3. The effect of the Mn content on the C Figure 4. The effect of the Si content on content of the metal phase (Si:Fe0.18) at the C content of the metal phase 1500oC under CO atmosphere. with different Mn-to-Fe ratios at at 1500 oC under CO atmosphere.
The silicon content of the metal phase in ferromanganese and silicomanganese compositions are to a great extent affected by the silica content of the slag with which they are in equilibrium. Figure 5 show the effect of the Mn-to-Fe ratios of the metal phase on the relationship between the silicon content of the metal and silica content of the slag in ferromanganese compositions. In Figure 5, three data groups with different Mn-to-Fe ratios (3.7, 5.5, and 6.7) are superimposed. The increase in the silica activity in the slag with increasing silica concentration results in the transfer of silicon from the slag to the metal.
Figure 5. The effect of silica content of the slag on the Figure 6. Iso-activity curves of Si in
Si content of the metal at 1500oC under CO. silicomanganese melts saturated with C at 1500oC.
It is also established that as the CaO-to-Al2O3 ratio increases (at more or less constant Al2O3 content) the silicon content of the metal decreases. As the Mn-to-Fe and Mn-to-Si ratio increases, the effect of increasing CaO-to-Al2O3 ratio on the silicon content of metal is even greater. This indicates that the activity of silica in the slag is decreased by the substitution of CaO in accordance with the network modification power of CaO.
Manganese and silicon activities in Mn-Si-Fe-C melts
The results of the previous investigation by the present authors on MnO activities1 in manganese slags were in excellent agreement with those predicted from the slag model developed by Gaye15. Therefore, SiO2 activities in the slag were predicted through the same slag model15 in order to calculate the Si activities in Mn-Si-Fe-C metal phase and the compositional information of the slag and metal phases were provided from the results of slag-metal equilibrium experiments making use of the following reaction at the experimental temperature of 1500oC.
C Si COSiO) 2 2
( 2 G55968.28J/mole (1)
The equilibrium constant of this reaction can be written as:
2 2
) (
) (
2 C
SiO CO Si
a a
p K a
(2)
In equation (2),
a
C 1
andp
CO 1
atm. Thus, the equation (2) reduces to:SiO2
Si
a
K a
(3)Where K0.02244.
The Mn activities were also calculated similarly using known activities of MnO in the slag1,2 through the equation
C Mn CO MnO) ( G23622.80J/mole (4) The equilibrium constant of this reaction can be written as:
MnO C
CO Mn
a a
p K a
(5)
In equation (5),
a
C 1
and pCO 1atm, therefore, the equation (5) reduces to:MnO Mn
a
K a
(6)Where K4.953.
The activities of Si and Mn can be easily calculated through the equations (3) and (6).
The model equations for activity coefficients of Si and Mn for silicomanganese compositions are:
) 7 ...(
...
5894 . 79 7
. 199 5236
. 28 5 . 123 4747
. 13 2237 . 8
ln
Si X
Si X
Fe X
2Mn X
MnX
Fe X
SiX
Fe ) 9 ...(...
3 . 332 8
. 143 8 . 122 5
. 320 1
. 300 4176 . 74
ln
Mn XMn XMn2 XSi2 XMnXFe XSiXFe and the model equations for activity coefficients for ferromanganese compositions are:) 8 ..(
...
0 . 1106 9
. 352 9
. 247 7
. 108 4553
. 14 6390 . 9
ln
Si XMn XSi XFe XMnXFe XSiXFe) 10 ..(
...
2 . 1263 9
. 303 3 . 1427 2975
. 30 0995 . 8 8048 . 6
lnMn XMn XSi XSi2 XFe2 XSiXFe Activities of manganese and silicon display strong negative deviations from ideality in both silicomanganese and ferromanganese compositions. The iso-activity contours of silicon for
silicomanganese and ferromanganese compositions were derived through the model equations (7) and (9) at 1500oC and are shown in Figures 6 and 7.
As it was mentioned before, the figures represent a carbon-saturated system and if atomic fraction of silicon (XSi) increases, then, atomic fraction of carbon (XC) decreases. Namely, the carbon concentration
decreases as it goes up to the top part of the figure. In this system, XC takes a certain value if the atomic fraction of iron; XFe and XSi are given, and therefore, the compositions of all carbon-saturated alloy melts of the Mn-Si-Fe-C system can be represented on a plane with XFe and XSi on both axis as shown in Figures 6 and 7.
In silicomanganese composition (Figure 6), the predicted iso-activity contour values vary between 0.012 and 0.021. However, in ferromanganese composition (Figure 7), the silicon activities are even smaller (between 5×10-4 and 3×10-3). The difference between the silicon activities in these two compositions can to a certain extent, be explained according to the difference in silicon concentrations (being higher in silicomanganese composition) in these compositions. Concerning the carbon saturated system Mn-Si-Fe- C; the related thermodynamic data is very scarce. Therefore, it is difficult to explain the order of magnitude change in iso-activity curves.
Figure 7. Iso-activity curves of Si in ferromanganese Figure 8. Iso-activity curves of manganese melts saturated with C at 1500oC (x10-4). in silicomanganese melts saturated with C at 1500oC.
Figures 8 and 9 show the iso-activity contours of manganese which are derived through the model equations (8) and (10) for silicomanganese and ferromanganese compositions respectively. In Figure 8, in silicomanganese composition, an increase in XFe decreases aMn values. A decrease in XFe will increase XMn
and XSi in carbon-saturated system and this will lead to higher aMn values. In Figure 9, a decrease in XSi at constant XFe first increases and then decreases the aMn values in ferromanganese composition. The increase in aMn is clear due to increase in XMn at fixed XFe and XC values. However, even a further decrease in XSi causes aMn to start decreasing which may be due to the fact that at very low silicon concentrations, the saturation solubility limit of carbon increases substantially favouring stronger Mn-C interactions and a tendency forMn5C2formation which will tend to decrease aMn values. When the results of the present study, are compared to the results of Tanaka16, 17 who studied the same system, it is seen that the iso-activity contours of silicon and manganese have very similar patterns and there is reasonable agreement between the actual activity values.
Figure 9. Iso-activity curves of manganese in ferromanganese melts saturated with C at 1500oC.
Conclusions
This research work was carried out to study the equilibrium distribution of Mn and Si between carbon- saturated Mn-Si-Fe-C alloys and MnO-CaO-MgO-SiO2-Al2O3 slags that are typical of the production of ferromanganese in submerged arc electric furnaces in South Africa. Furthermore, an attempt was made to determine the activities of Mn and Si in the metal phase. The conclusions are outlined as follows:
- An increase in basicity ratio of the slag decreases the Mn distribution ratio, - An increase in silica concentration of the slag increases the Mn distribution ratio, - An increase in CaO-to-Al2O3 ratio of the slag decreases the Mn distribution ratio, - An increase in MgO-to-CaO ratio of the slag increases the Mn distribution ratio, - The carbon and manganese contents of the metal phase are directly proportional,
- The carbon and silicon concentrations in the metal phase are inversely related and, as the Mn-to- Fe or Mn-to-Si ratio increases, the carbon solubility in the metal phase decreases,
- An increase in the silica content of the slag phase increases the silicon content of the metal phase and this effect is even more pronounced at the higher Mn-to-Fe or Mn-to-Si ratios,
- The activities of Mn were computed by making use of slag-metal equilibrium data and MnO activity data gathered in earlier part of the present work1, 2.
- The activities of Si were computed by making use of slag-metal equilibrium data and the SiO2
activities predicted through the slag model15
- Statistical model equations which represent
ln
Mnandln
Sidata successfully were developed in order to use to predict the activity coefficients of Mn and Si in the compositional range of the present work.- The iso-activity contours of Si and Mn in silicomanganese and ferromanganese compositions were derived.
References
1. Cengizler, H. (1993): Thermodynamic activity of MnO in manganese slags and slag- metal equilibria. Ph.D. Thesis”, University of the Witwatersrand, Faculty of Engineering, Johannesburg, South Africa.
2. Cengizler, H., Eric, R.H. Thermodynamic activity of manganese oxide in ferromanganese slags and the distribution of manganese between metal and slag phases, “Proceedings of the 6th International Ferroalloys Congress”, Cape Town, South Africa, 1992, Vol. 1, pp. 167- 174.
3. Cengizler,H., Eric, R.H. and Reuter, M. Modelling of Activity of MnO in Ferromanganese and Silicomanganese slags by Neural Nets, Proceedings of the 5th International Conference on Molten Slags, fluxes and salts, January, 1997, Sydney, Australia, ISS, Warrendale-PA, pp. 75-90.
4. Tuset, J.Kr., Sandvik, J., and Venås, K. SINTEF Research Report No. 340420, Trondheim, Norway, 1970.
5. Esin, O.A. and Vatolin, N.A. Application of the electro-force method to the study of the liquid manganese alloys, Zhur Priklad Khim, vol. 27, 1954, pp. 1252-1256.
6. Turkdogan, E.T., Hancock, R.A. and Herlitz, S.I. The solubility of graphite in manganese, cobalt and nickel, Journal of Iron and Steel Institute, vol. 182, 1956, pp. 274-277.
7. Petrushevskii, M.S. and Gel’d P.V. Equilibrium of carbon with liquid alloys of Fe, Mn, Si and C, Journal of Applied Chemistry of USSR, vol. 32, 1959, pp. 86-94.
8. Skiredj, O. and Elliot, J.F. carbon content of graphite-saturated Fe-Si-Mn alloys, 1400o to 1650
oC, Transactions of the Metallurgical Society of AIME, vol. 227, 1963, pp. 536- 538.
9. Tanaka, A. Activities of manganese in Mn-Fe-C, Mn-Si-C and Mn-Fe-Si-C melts at 1673 oK, Transactions of the Japan Institute of Metals, vol. 21, Oct. 1980, pp. 27-33.
10. Ni, R., Ma, Z., and Wei, S. Thermodynamic of Mn-Fe-C and Mn-Si-C system, Steel Research, vol. 61, no. 3, 1990, pp. 113-116.
11. Kubachevski, O., Evans, E.U., and Alcock, C.B. Metallurgical Thermochemistry, 5th ed. Oxford:
Pergamon Press, 1979, pp. 114-178.
12. Barin, I. Thermochemical Data of Pure Substances, VCH Verlagsgesellschaft mbH, D-6940 Weinheim (Federal Republic of Germany), 1989, Part 2.
13. Pankratz, L.B., Stuve, J.M. and Gokcen, N.A. Thermodynamic data for Mineral Technology,Bulletin 677, U.S. Bureau of Mines, (1978).
14. Robie, R.A., Hemingway, B.S., and Fisher, J.R. Thermodynamic Properties of Minerals and Related Substances at 298,17K and 1 Bar (105 pascals) Pressure and at Higher temperatures.
Bulletin 1452, Washington, U.S.A., Geological Survey, (1978), p. 456.
15. Gaye, H. A model for the representation of the thermodynamic properties of multicomponent slags, Centenary of Metallurgy Teaching in Glasgow, June, 1984, pp. 1-14.
16. Tanaka, A. Activities of manganese in Mn-Fe-C, Mn-Si-C and Mn-Fe-Si-C melts at 1673 oK, Transactions of the Japan Institute of Metals, vol. 21, Oct. 1980, pp. 27-33.
17. Tanaka, A. Equilibria between Mn alloy melts saturated with C and various kinds of molten slags and calculation of the activities, Tetsu-to-Hagane, vol. 66, 1980, pp. 1474-1483.