Formulation and Plant Trial of a Mixed Collector Suite For Eland Platinum

"The role of mixed collectors in enhancing the metallurgical performance of sulphide flotation processes was reviewed in 2008, and later published as an integrated theory by Lotter and Bradshaw, 2009. The theory proposed that xanthates, dithiophosphates and dithiocarbamates could be synergistically combined in an optimal formula to deliver improved grade and recovery. The approach uses a five-step structure, viz.1. A mineralogical study of the ore to be treated, identifying and quantifying the economic minerals (with hosting and association) and gangue species present.2. Selection of candidate collectors from this mineralogical information and from the known pulp conditions of that operation (e.g. Eh, pH).3. Design and execution of a factorial experiment to estimate and test the main effects and interactions, and to model the response surface for an optimum formulation [Box et al., 1978].4. Use of a High-Confidence Flotation Test platform in the factorial to reduce overall errors and allow for clearer observations [Lotter, 1995; Lotter and Fragomeni, 2010].5. Statistical design of a plant scale trial to measure and validate the gains [Napier-Munn, 1995; Lotter et al., 2009]. Eland Platinum, which was commissioned in November 2007, appointed a team to apply this theory to formulate an optimized mixed collector suite for use in the operation. The West Pit ore (UG2 - Normal ore) was sampled at site and brought to Xstrata Process Support, Sudbury, Canada, for mineralogical and metallurgical testing in 2008. Microprobe and QEMSCAN measurement characterized the type, grain size and textural association of the discrete Platinum Group Minerals (PGM) present. From the mineralogical data, it was clear that a mixture of dithiocarbamate and two xanthates would be a likely reagent suite selection. A factorially designed set of High-Confidence Flotation Tests studying these was performed, concluding that an optimal suite using certain amounts of these collectors was a viable formulation that would advance the grade and recovery performance at Eland above the standard PIBX by approximately 2.13% 4E PGE. The plant trial, which used a form of the replicated block design after Napier-Munn, 1995, and Lotter et al., 2009, exceeded this estimate with an overall PGE recovery gain of 2.48% (absolute basis) and a concentrate grade gain of 26.4 g/t PGE (absolute basis), as well as reducing chrome in concentrate from 2.04 to 1.62% (absolute basis). These gains proved to be statistically significant using the analysis of variance method, and have validated the theory."