Preliminary Results on the Reproducibility of Sample Preparation and QEMSCAN« Measurements for Heavy Mineral Sands Samples

QEMSCAN« technology is a widely accepted automated e-beam technique providing quantitative and statistically robust mineralogical data to the mining industry. It uses X-ray mapping combined with backscattered electron (BSE) brightness to identify minerals in almost any type of geological material (Gottlieb et al, 2000). Besides the modal mineralogy also the liberation, associations, average grain size and other textural information can be obtained. The aim of this study is to test reproducibility of this technique by setting up and running standard tests on different QEMSCAN« systems. Multiple sample blocks were prepared by our standard operating procedure as epoxy-mounted, carbon-coated, polished blocks. The replicates were measured under the same conditions and measurement parameter settings on different QEMSCAN« instruments. The relative standard deviation (RSD) for the set of measurements was calculated. Generally the reproducibility of the systems themselves was found to be excellent, ranging between 0.4 and 1.5 per cent RSD. The more significant change in the reproducibility was found to be caused potentially by the sample preparation and variation in the mineralogy of the replicates. The measurement parameter settings together with these variations can account for almost 15 per cent RSD for minerals of low abundance. For analysed samples good results for the main phases present in quantities higher than 2 - 5 wt per cent of the total mass of the sample were confirmed with RSD generally below six per cent. Following the experiments, the parameters having the biggest influence on the performance of the QEMSCAN« and its precision are discussed in detail. The importance of sample preparation procedures and measurement parameter settings is emphasised. It is shown that experience and excellence is necessary for getting the most from the automatic mineralogical methods like QEMSCAN«.