Phase equilibria studies and beneficiation of titaniferous slags

Doctoral Thesis


Permanent link to this Item
Journal Title
Link to Journal
Journal ISSN
Volume Title
Titaniferous magnetite (titanomagnetite) offers a unique opportunity for the production of three valuable products from one resource. It generally contains economically appreciable reserves of vanadium and iron as well as significant contents of titanium. Titanomagnetite is typically smelted in blast or electric furnaces in the presence of reductant and fluxes (dolomite and quartz) to produce a valuable vanadium bearing pig iron and a virtually valueless titaniferous slag. The titaniferous slag by-products are generally defined by the Ca-Mg-Al-Si-Ti-O system. These fluxed slags can contain as high as 20-40wt% TiO2 (titania). The lack of interest in processing titaniferous slags to produce saleable titania materials is attributed to the presence of chemically inert phases, like the spinel solid solution [Mg(Al,Ti,V)2O4] that cannot be handled by the available titania slag upgrading technologies. The understanding of phase relations in titaniferous slags is thus important in order to be able to implement a suitable fluxing strategy for the production of a treatable titaniferous slag with no inert spinel phase. The available phase equilibria data established in air for titaniferous slags is inconclusive about the possible crystallisation of the detrimental spinel. However, literature on phase compositions of plant titaniferous slags are conclusive about the crystallisation of Mg(Al,Ti,V)2O4 in high MgO and Al2O3 bearing slags. It is thus clear that the understanding of phase relations in titaniferous slags requires further development. The objective of the current project was to investigate the phase equilibria and beneficiation of titaniferous slags to produce a saleable titania product. As a development to previous work, the composition of the slag for review was based on the available work, namely; TiO2 = 37.19wt%, SiO2 = 19.69wt%, and Al2O3 = 13.12wt%, at varying proportions of CaO (30- 0wt%) and MgO (0-30wt%). The phase equilibria studies followed a systematic approach involving the review and validation of the available equilibrium phase diagram produced in air, followed by the determination of updated phase equilibria at low oxygen partial pressures (pO2) of 10-16 atm applicable to titanomagnetite smelting. The generic approach of studying phase equilibria in multicomponent systems was followed, namely; (1) literature survey of the available thermodynamic and phase equilibria data applicable to the reviewed system, (2) calculation and re-drawing of the equilibrium phase diagrams using FactSage thermochemical software, and (3) equilibration-quench-(electron probe micro analysis) (EPMA) experiments to the validate calculated equilibrium phase relations. A titaniferous slag with little crystallisation of the inert spinel phase, based from the best fluxing strategy with an MgO-free limestone, was produced by smelting in conventional (using alumina crucible) and cold copper crucible induction furnaces for subsequent beneficiation using the established Upgraded Slag (UGS) process. A conceptual flowsheet for the production of vanadium, steel and titanium products was therefore designed and subsequently subjected to economic evaluation using the discounted cash flow (DCF) modelling approach. Thermodynamic and phase equilibria literature for the Ca-Mg-Al-Si-Ti-O system demonstrated that this system and some subsystems are well researched in air, and not as much in low pO2 atmospheres applicable to smelting operations. The FactSage software used in the current study for the calculation of phase equilibria in the Ca-Mg-Al-Si-Ti-O system applicable to titaniferous slags does not have tialite (Al2TiO5) modelled as a component in the customary pseudobrookite solution database - Al2TiO5 is a component of the pseudobrookite solution reported in literature for the current system. Hence, a private pseudobrookite solution database applicable to the reviewed system, i.e. MgTi2O5-Al2TiO5-Ti3O5, was developed and incorporated into FactSage before any calculation could be conducted. Thermodynamic modelling of the MgTi2O5-Al2TiO5-Ti3O5 system was conducted through the CALculation of PHase Diagram (CALPHAD) method. The sublattice model coupled with compound energy formalism (CEF) and Redlich-Kister polynomial were adopted. The model information was incorporated into FactSage to create a private database for subsequent calculations of phase equilibria of titaniferous slags. The equilibrium phase diagram for the Ca-Mg-Al-Si-Ti-O system in the same compositional range as in the available literature was then calculated in air. The liquidus surfaces and phase relations in the equilibrium phase diagrams of available literature and FactSage calculation are fairly comparable. However, at high MgO concentrations: FactSage calculation predicts that Mg(Al,Ti)2O4 is the primary phase, followed by successive crystallisation of pseudobrookite solid solution (MgTi2O5-Al2TiO5) and forsterite (Mg2SiO4); and the available literature reports MgTi2O5-Al2TiO5 as the primary phase, followed by Mg2SiO4. The crystallisation of spinel phase in the available phase diagram produced in air is not predicted. The crystallisation of the spinel solid solution phase in titaniferous slags is extensively reported in the open literature. Equilibration-quench-EPMA experimental results produced in air generally compared well to the FactSage calculations. The inability of the available phase diagram to predict spinel phase crystallisation was attributed to the lack of sensitive analytical techniques in the late 1960s, when the available phase diagram was developed. The phase equilibria of titaniferous slags were further calculated at low pO2 atmospheres of 10-16 atm. In the reviewed compositional range of titaniferous slags, the liquidus surface and Ti3+/Ti4+ mass fraction ratio increased with decreasing the pO2. There was no significant difference in terms of the crystallisation of phases between the calculated results in air and at pO2 of 10-16 atm, except that the size of the primary phase field at higher MgO concentrations than the composition for the minimum liquidus temperature increased and the pseudobrookite solution included Ti3+ bearing phase, i.e. MgTi2O5-Al2TiO5-Ti3O5. Equilibration-quench-EPMA experimental results produced at pO2 of 10-16 atm generally compared well to the FactSage calculations. The new phase equilibria at low pO2 of 10-16 atm shows that the crystallisation of the chemically inert Mg(Al,Ti)2O4 in titaniferous slags would occur if the slag contains high Al2O3 concentration and MgO concentration of 2wt% and above. However, the crystallisation of Mg(Al,Ti)2O4 in titaniferous slag is not significantly sensitive to variation in the TiO2 concentration in, and basicity of the slag. To produce a titaniferous slag with minimum possible inert spinel content for subsequent beneficiation, the South African Main Magnetite Layer (MML) titanomagnetite concentrate was smelted in the presence of an MgO-free lime and low ash Sasol carbon (SASCARB) reductant. This smelting approach would produce a titaniferous slag with about 4wt% MgO, which would come from the titanomagnetite - based on the phase equilibria, this slag should crystallise a small amount of the inert spinel. When the smelting was conducted in an alumina crucible placed inside a conventional induction furnace, the slag was inevitably contaminated by Al2O3 from the refractory container. This slag crystallised a significant Mg(Al,Ti)2O4 with the content approximated by the MgO concentration - the significant Mg(Al,Ti)2O4 crystallisation was attributed to the Al2O3 saturation in the titaniferous slag. A titaniferous slag containing a treatable ulvospinel phase was produced in a cold copper crucible induction furnace - the crystallisation of the ulvospinel, instead of the chemically inert spinel solid solution was attributed to the saturation of the slag by iron from the incomplete reduction process due to the inevitable stoppage of the heat supply to the induction furnace as soon as the susceptor iron metal and produced pig iron settled at the bottom of the copper crucible. For the purpose of demonstrating the feasibility of producing titania products from titaniferous slags, this slag was also subjected to beneficiation using UGS process. The current study successfully demonstrated that the titaniferous slags can be beneficiated to saleable titania products using the UGS process: the TiO2 in the Mg(Al,Ti)2O4 bearing waste titaniferous slag produced by the defunct Evraz Highveld Steel and Vanadium Corporation (EHSV) was upgraded from about 33wt% to 75wt%, while the TiO2 in the titaniferous slags produced in conventional and cold crucible induction furnaces were upgrade from about 30wt% to 67wt% and 22.00wt% to 90.45wt%, respectively. The remaining impurities in the 75wt% and 67wt% TiO2 UGS products were mainly MgO and Al2O3 contained in the refractory spinel solid solution. In the case of the 67wt% TiO2 product, there was excess Al2O3 in the spinel structure - the excess Al2O3 remained in the glass phase. Though the 90.45wt% TiO2 product is attractive for use as feedstock for the production of the preferred chloride pigment, this product however contained finer PSD and higher concentrations of impurities such as SiO2, Al2O3, and CaO, than the specification for the chloride process feedstock. This product was thus not a suitable feedstock for the chloride pigment production. Further optimisation of the UGS conditions has a potential to reduce the concentrations of impurities to levels suitable for feedstock for the preferred chloride pigment production process. Further investigations are also required to study the feasibility of the chlorination of micro-pellets of the UGS product. Since the UGS product is mainly composed of rutile structure, it would not be a suitable feed for the sulfate pigment production as the sulfuric acid lixiviant is unable to dissolve the rutile structure. Only if soluble in sulfuric acid, this high TiO2 bearing UGS product produced from titaniferous slags could be used as advantageous feedstock for the sulfate pigment production in terms of the minimization of the reagent consumption and the amount of the toxic sulfate waste. Based on the work of the current study, literature data and Pyrosim simulations, a conceptual process flowsheet for the production of vanadium slag, steelmaking pig iron and titania product was proposed. The economic modelling of the conceptual flowsheet for a 20 year operational projection showed that the process is economically viable. The process economic viability is sensitive to variation in the Opex and Revenue. In addition, additives, such as the amount and type of reductant, fluxes, and reagents account for about 75% of the Opex. It is possible that the additives are overestimated in the process as the recycle streams were not included in the proposed process and economic model. At the same time, the economic model does not consider the environmental and waste management costs. Hence, the economic analysis is considered to be preliminary in nature, or indicative at best. The current study has demonstrated that (1) a titaniferous slag containing little or no inert spinel phase that is suitable for upgrading can be produced when the MgO content in the slag is below 2wt% - the best approach to producing a slag with minimum possible MgO content would be to smelt the titanomagnetite in the presence of an MgO-free limestone flux and low ash reductant, and (2) it is technically and economically feasible to produce three products, i.e. V slag, steelmaking pig iron, and titania product, from titanomagnetite.