This thesis investigates the mode of action of cationic reagents in the recovery of platinum group metals (PGMs) by anion exchange solvent extraction. Industry uses a range of extractants to achieve efficient concentration and separation of PGMs in hydrometallurgical processes, but an understanding of the processes at a molecular level is limited and restricts the options to improve efficiency. The research, sponsored in part by Johnson Matthey and Anglo American, explores two different aspects of the mode of action of reagents used to recover chloride and platinum. Chapter 1 reviews the extraction of chloridometalates in hydrometallurgy and other methods used to recover PGMs. The chapter also covers current ideas on whether formation of outer-sphere anion-cation molecular assemblies or whether larger supramolecular aggregates are responsible for the extraction of PGMs. These two separate routes of study can be investigated by various computational and experimental methods, which are discussed in Chapter 2. In Chapter 3, the model systems chosen to develop the methodologies utilised throughout the thesis are presented. The development involves the extraction of chloride ions by tributyl-phosphate (TBP) and then is extended to extraction of PGM chloridometalates in Chapter 4. Computational methods are used to probe the atomistic and supramolecular theories in predicting the most likely assemblies which will be formed in the transfer of anions between the aqueous and organic phase. Slope analysis and the determination of the contents of the organic phase is used to validate computational models, along with spectroscopic techniques to determine shape and size of assemblies formed during extraction. The application of these methodologies to an amide extractant is discussed in Chapter 5. Computational methods predict the probability of formation of specific complex assemblies during the extraction of PtCl62- by protonated forms of the amide. Determination of the stoichiometry involved in formation of the complex assemblies by slope analysis is reported along with analysis of water, metal and chloride content to confirm the computational model. Final conclusions on all systems explored within the thesis and suggestions for future work are presented in the final chapter. The combination of experimental and computational methods are shown to be very efficient in defining mechanisms of extraction, involving determination of structures formed during the process and how and why they form.

M. Tech. / The separation of minerals such as coal, diamonds and others have involved the use of a heavy medium. The heavy medium separation technique, used in the diamond industry as a primary concentrator, is a physical process that separates minerals of different specific densities based on their relative movement and resistance to motion in a viscous fluid. It is made up of a suspension of finely ground solid particles constituting a stable suspension. There is a wide range of materials that are used to prepare suspension media in Heavy Medium Separation (HMS), but Ferrosilicon has found a wider usage. Major demand for dense medium Ferrosilicon powder is from the diamond industry which accounts for 77 percent of milled sales. In HMS, either dynamic or static separators can be used, even though; dynamic separators are widely used because they give higher separation efficiencies. A number of accounts have been reported on the effects of stability and viscosity to the medium. It is, however, observed that the effect of ore mineralogy, which is a major contributing factor during ore and medium interaction, during separation, has not been properly taken into account, apart from the fact that ore mineralogy form a basis for HMS efficiency. The literature, however, does not cover a thorough investigation of the relationship between the recovery of diamond bearing ores and stability of Ferrosilicon grades media, as a function of physicochemical properties of ferrosilicon. Eight samples of Ferrosilicon, four milled and four atomized were used to prepare media to recover value from kimberlitic ore, alluvial ore and synthetic ore in a laboratory set-up HMS. Both Ferrosilicon and ore samples were characterized with the following techniques before use: XRD, XRF, SEM-EDAX and Screening, to investigate physico-chemical properties of ferrosilicon and ore mineralogy. The used Ferrosilicon was also characterized to investigate any v changes as Ferrosilicon medium interacted with the ores. Settling tests were performed on each Ferrosilicon grade to investigate media stability before and after recovery exercise. The stability of the media was then related to the recovery of each ore. In-circuit sample of ferrosilicon was collected from Letšeng diamonds to compare any changes with Ferrosilicon used in the laboratory, and also used to study the effects of contamination on the degradated medium. All recovery results were done in a static HMS, but in practice dense medium cyclones are widely used. The screening characterization technique revealed that there was a loss of finer particles size fraction, predominant in sieves 75 microns and 45 microns and that loss was highest with atomized grades than milled ferrosilicon grades. Minimal changes in density and chemical compositions were observed for each ferrosilicon grade. The loss of the finer fraction was found to effect changes in the settling rate of each ferrosilicon grade differently, hence changes in the stability of their media. The efficiency (Ep) of separation was found to vary with each grade of Ferrosilicon used, accompanied by a shift in cut point density indicating the influence of grade on the separation efficiency. The effects of medium stability on recovery for both ores showed that although both ores percentage recoveries differ, the trend of medium stability to recovery, with each Ferrosilicon grade, is the same. However, the recovery was found to be more dependent on the density of the medium, as the effects of loss of finer particle size fraction did not show any significant contribution to recovery, but rather on the medium loss. The mineralogy of the ores also had been observed to give a minimal or no contribution to the separation efficiency, if related to recovery. Milled Ferrosilicon is further recommended over atomized Ferrosilicon, based on cost and stability. The highest loss of atomized Ferrosilicon could escalate the operating costs and affect the stability of the medium. It should be remembered that the purchasing costs of atomized grades is higher than that of milled grades.