A mineralogical perspective on the recovery of uranium from brannerite-rich ore at Cooke Section, West Rand Goldfield, South Africa

Abstract

Abstract The recovery of uranium from quartz-pebble conglomerates of the Witwatersrand Basin is accomplished through sulphuric acid dissolution under oxidising conditions. At Cooke Section on the West Rand Goldfield, the extraction process has been plagued by low to moderate yields on the order of 40 to 75%, as opposed to a target recovery of 80%. This has been ascribed to the high abundance of brannerite in the ore, which has traditionally been more problematic to leach. In addition to brannerite, poor metallurgical recoveries may also be associated with processing inefficiencies related to comminution, residence time, acid dosage and leach temperature. In view of this, a range of ore samples (channel samples) were collected from four uranium-bearing conglomerate horizons at Cooke Section (the A1, A5, E9EC and UE1A reefs) for detailed mineralogical and metallurgical characterisation, involving automated mineralogical analysis, and laboratory-scale leach testwork. The mineralogical results show that the major uranium-bearing minerals of uraninite, coffinite and brannerite are fine-grained (~80% passing 32 micron) and exhibit high degrees of mineral exposure to the lixiviant (~99%). Despite these favourable attributes, the elemental deportment data indicate that brannerite accounts for approximately 43% of the combined uranium budget. Further inspection shows that brannerite can be subdivided into three compositional subtypes: uraniferous brannerite (~13% U deportment), brannerite (~25% U deportment) and titaniferous brannerite (~5% U deportment). Baseline laboratory leach tests, which replicated plant leach conditions of 30 kg/ton acid, 4 kg/t oxidant, 24 hour residence time and 60°C leach temperature, yielded elevated dissolutions between ~77% and ~96%, with a combined overall uranium recovery of ~94%. These results are not consistent with the low yields obtained at the processing plant, and suggest that the high level of uranium recovery can be attributed to the effective leaching of brannerite (most likely uraniferous brannerite and brannerite). In view of prevailing market conditions, variability tests were carried out on a representative bulk composite sample to investigate the potential to achieve similar yields under more cost-effective leaching conditions. In these tests, a single parameter was varied (e.g. acid dosage), while the remaining parameters remained at baseline conditions. The results demonstrate that uranium recoveries of ~80% can be achieved on Cooke Section ores at low acid dosages and high temperatures (18 kg/t, 60°C) or at moderate acid dosages and low temperatures (23 kg/t, 30°C). The associated reduction of input costs would represent a significant cost-saving for the Ezulwini gold and uranium recovery plant. It is concluded that the poor uranium yields encountered during commercial processing of the ore are most likely related to undiagnosed inefficiencies in the treatment plant, such as excessive acid consumption related to elevated temperatures/oxidant addition and/or insufficient leach residence times, especially when recirculating, continuous flow-through leaching systems are in use. The broader implication of this study is that uranium processing operations beyond Cooke Section may be able to optimise their process designs and reduce input costs by quantifying the different types of brannerite within their ores through automated mineralogical analyses. The present study thus demonstrates the value of a geometallurgical approach in enhancing the understanding of uranium recovery through acid leaching.