A study of an autogenous slag for steel production with consideration of possible vanadium extraction

Sammanfattning: The thesis discusses the use of the autogenous slag that forms from the residual oxides present in hydrogen reduced iron (H-DRI) pellets during melting. The studies are motivated by a better understanding of how H-DRI affect the steelmaking operations. A possible optimization of the steelmaking process is to recover the vanadium that is contained in the iron ore raw material. Therefore, understanding the role of vanadium during melting of H-DRI is given an extra focus.  Taking advantage of the autogenous slag by utilizing its dephosphorization power to a maximum, or using it to extract vanadium, could make an important contribution to the process economics. To assist the developments in these directions, the phosphorus and vanadium partitions between slag and metal (LP and LV) as well as the phase relationship of the autogenous slag were investigated. The partitions were studied by melting H-DRI with reduction degrees between 91 and 99% in closed systems at 1873 K. The obtained LP and LV were in the ranges 8-26 and 501-1994, respectively. The LV values increased with decreased reduction degrees. The values for LP increased with decreasing reduction degrees until a 97% degree of reduction. Further lowered reduction degrees correlated with decreased LP values. The lowest phosphorus levels encountered in the iron (130 ppm) were obtained after melting of H-DRI with degrees of reduction between 94 and 98%. This indicates that the autogenous slag has a potential to make a significant contribution to the phosphorus refining.  To find out about the phase relationship in the autogenous slag at 1873 K, small (5 g) samples of synthetic slag were equilibrated with 1 g iron in closed systems. Composition-wise, these slags corresponded to the autogenous slags from H-DRI with 98.4-99.7% reduction degrees. Preservation of the high temperature phase relationship required fast cooling; therefore, the samples were quenched in oil. This was also the reason for using small samples. Spinel, magnesiowüstite and liquid phase were identified as the stable phases at 1873 K. The spinel and magnesiowüstite phases were high in V, while the liquid contained almost no V. Increased FeO-contents (decreased degrees of reduction) correlated with a decreased amount of spinel, an increased amount of magnesiowüsite, as well as a decreased content of V in both phases. To increase the understanding about the phases in the autogenous slag, a sub-system containing MgO and V2O3 was investigated under conditions relevant to H-DRI melting, namely temperatures between 1661-1873 K and pO2 values between 1.75×10-11 and 1.75×10-10 atm. The phase boundaries for the three stable phases MgO-halite, spinel and V2O3-corundum were established. The oxygen potential and the temperature had limited impacts on the phase boundaries for the spinel and V2O3-corundum phases, while the maximum solubility of V2O3 in MgO-halite was affected to a somewhat larger extent. As earlier research has shown that an acid slag could be suitable for V extraction, the pseudo-ternary phase diagram between Al2O3, SiO2 and V2O3 at 1873 K and pO2=3.4×10-11-3.4×10-9 atm was also investigated. 5 different phases were identified, namely mullite, Al2O3-corundum, V2O3-corundum, cristobalite, and a liquid phase. The most significant effect of the oxygen potential was on the invariant point representing double Al2O3 and V2O3 saturation of the liquid. The multivalent nature of vanadium is suggested as the reason for the slight impact of the oxygen potential on the phase diagrams. To understand how the autogenous slag forms from the residual oxides, individual pellets with 90 and 99% reduction degrees were studied during heating to either 1773 or 1873 K. It was observed that the autogenous slag forms before iron melts. The slag likely forms as FeO melts and dissolves the other remaining oxides. Thereby, vanadium is transferred to the autogenous slag. Before iron melts, the movement of the autogenous slag is restricted to the pellet’s pore network. Thereafter, when iron melts, the slag starts to coalesce as well as to floatate. As the autogenous slag may contain solid phases, the effect of the fraction of solid phase on the slags foamability was finally investigated. This was done by measuring the maximum foaming heights of slags containing Al2O3, CaO, FeO and SiO2, reminiscent in their compositions to the autogenous slag. The slag compositions were chosen so that the fraction of precipitated magnesiowüstite phase was the main variable. It was found that some amount of solid phase (1.6 vol%) increased the foaming height by approximately 7%, while ≥8.7 vol% more than halved the foaming height.