Dephosphorization of high carbon ferromanganese using BaCO3 based fluxes

Abstract

Manganese is added to steel in the form of ferromanganese to improve the rolling and forging qualities, strength, toughness, stiffness, wear resistance and hardenability. It is also used extensively as a deoxidizer. In view of its wide applications, the consumption of ferromanganese is more than that of all other ferroalloys combined. However, ferromanganese is one of the main sources of phosphorus (P) contamination in steels because it usually contains more than 0.4% phosphorus. On one hand there is a greater demand on steelmakers to produce steel with stringent limit on phosphorus content (because its presence makes the steel brittle and susceptible to stress corrosion cracking), on the other hand the quality of manganese ores with respect to its phosphorus content goes on deteriorating day by day. In these circumstances, it is virtually impossible for the ferromanganese manufacturers to produce high carbon ferromanganese with low phosphorus content (<0.2%) by carbothermic reduction of inferior grade (>0.2%P) manganese ores. Since ferromanganese is one of the late additives in the steel making process, virtually all the phosphorus present in ferromanganese goes straightway into the product. It is, therefore necessary to restrict the phosphorus content in ferromanganese itself, so that its entry into liquid steel is minimized. Since phosphorus is intimately mixed in the ore, its removal by physical separation technique is not feasible. During last decade, several laboratory scale investigations have been carried out to remove phosphorus by various routes but no established economically viable method exists today for lowering the phosphorus content in ferromanganese to the desired level. The phosphorus removal under oxidizing conditions is usually accompanied by loss of manganese whereas under reducing conditions the product is a phosphide which is not considered environmentally friendly and also it is uneconomical because of the extra cost involved in the post treatment of slag. Gaseous dephosphorization is also not suitable because the partial pressure of manganese is higher than that of phosphorus at smelting temperatures. Therefore, the present investigation focused on the development of a suitable flux for selective removal of phosphorus under oxidizing conditions. Predominance area diagrams for various flux systems ( Ba-P-O and Ca-P-O superimposed by Mn-MnO line) were drawn at standard as well as non standard states and at various temperatures to provide guidelines to select the suitable flux system and temperature for removal of phosphorus without significant loss of manganese. BaO based fluxes rich in MnO at comparatively lower temperature were found to be effective in removing phosphorus with minimal loss of manganese. Feasibility tests were carried out in a graphite crucible fitted in an induction furnace using BaCO3-based fluxes rich in MnO. BaCO3 was selected as a substitute for BaO because of its ready availability at a low cost. Initial trials were unsuccessful because of the difficulties experienced in melting BaCO3, MnO2 mixture at moderate temperatures (<15000C) possibly because of the high melting points of pure BaO and MnO. However, after repeated trials, the flux (BaCO3 ) was found to melt at 1300-13500C in the presence of carbon when it was added from the top without external addition of MnO2. Addition of BaF2/ BaCl2 helped in improving the fluidity of slag. Once it was possible to melt the flux at 13000C, the effects of various parameters such as (i) type of flux, (ii) quantity of flux, (iii) silicon content of the alloy, (iv) temperature were studied on the degree of dephosphorization. The phosphorus level was reduced to 0.18 % from an initial level of 0.56 % when the reaction was carried out at 13000C using 16 Wt % BaCO3-BaF2 fluxes. The manganese loss was restricted in the range of 2-5%. However, it was found that the degree of dephosphorization decreased significantly with increase in the initial silicon content of the alloy. An initial silicon content less than 0.2 % is required to achieve effective dephosphorization. Amongst the various additives used, BaF2 was found most effective. The results clearly indicated that the addition of BaF2 to BaCO3 not only increased the extent of phosphorus removal but also contained the manganese loss to a tolerable level. Dephosphorization tests were also carried out using calcined BaCO3-based flux pellets (major phase BaO) in order to develop a reagent for commercial application and to improve the degree of dephosphorization over BaCO3based powder. It was also possible to study the effect of variation of MnO content in the flux / slag. A comparison of dephosphorization efficiency using calcined BaCO3-MnO2-BaF2 pellets with that using BaCO3-BaF2 powders showed an improvement over powders (80% for calcined pellets against 68% only for powders). The % content of the slag was found to increase with increase in MnO2 content in the reagent up to 30% beyond which it starts decreasing. Therefore, 30 % MnO2 content in the reagent which generates about 30% MnO in slag was considered to be optimum amount, which was attributed to possible increase in softening temperature of BaO-MnO flux beyond this value. The results of the feasibility tests as discussed above using BaCO3-BaF2 fluxes were found encouraging which indicated that more than 60 % phosphorus could be removed but the time taken was about 30 minutes. This large duration for dephosphorization was not considered suitable for plant scale trials. Therefore, a kinetic study was also taken up to reduce the reaction time by enhancing the reaction rate through powder injection. For this purpose an injection system was designed for injecting the flux at laboratory scale (7 Kg melt). The results showed that it is possible to remove 80% phosphorus in 5 minutes by injecting 10 wt% BaCO3 - based flux in liquid ferromanganese using submerged graphite lance from top. A mathematical model developed to understand the injection process showed that transitory reactions contribute approximately 62 % to overall reaction.