The characterization of platinum group minerals (PGMs) from concentrator flotation circuits provides valuable information towards understanding PGM recovery under given milling and flotation conditions. Additional mineralogical characterization also provides information on ore variability, which may affect recovery. Considering the low grades involved, automated scanning electron microscopy (SEM) characterization of PGMs has enabled relatively quick analyses, compared with older, time consuming and less accurate optical microscopy methods. over the years, parameters, such as species, size distribution, liberation, gangue and base metal sulfide (BMS) associations, have been used to characterize the PGMs in a process mineralogical context, to aid metallurgical interpretation. This paper considers whether flotation indices determined for PGMs can be empirically validated, using initial results from analysis of PGMs in a concentrate bank from a South African concentrator treating UG2 ore from the western limb of the Bushveld Complex. The aim is to determine whether floatability is adequately described by these parameters, or whether other mineralogical factors must be considered when assessing PGM floatability, and hence accounting for recovery vs. losses to tails. The data obtained to date, show that flotation parameters (including liberation index, BMS/gangue relationships and size distribution) have merits where large changes in PGM concentration occur, as observed in the first four cells of the concentrate bank. Larger data sets, however, are required to assess floatability, particularly where concentrate cell grades are similar, as observed in the last six cells of the concentrate bank. Furthermore, the floatability indices could be enhanced by the incorporation of selected associated gangue mineral information, chief amongst these being the content and mode of occurrence of naturally floatable talc and associated orthopyroxene.
