Assessment of Automated Mineralogy in Characterising Zn-Pb nonsulphide ore

Automated Mineralogy using SEM-EDS analysis has been routinely applied to ore characterisation, notably using the QEMSCAN. The main application of this technique has been to quantitatively characterise the mineralogy as applied to mineral processing and other applications of geometallurgy. This preliminary study explores the capabilities of the next generation of Automated Mineralogical analyses using Mineralogic Mining (ZEISS) technology. Mineralogic utilizes modern quantitative EDS technology to allow minerals to be classified based on the wt% element abundance (stoichiometry) which is sensitive to subtle elemental variations. In this study, samples from the Jabali Zn-Pb-Ag Mine and 3 from the Reef Ridge prospect were analysed using Mineralogic Mining. Samples were taken from different oxidation zones representative of the deposits comprising complex and varied Zn-Pb nonsulphides that require thorough characterisation before economic assessment can be made. The main nonsulphide Zn mineral at Jabali is smithsonite, which replaces both dolomite and sphalerite. Previous studies revealed zoned smithsonite, variably enriched in CaO, MgO and FeO and the occurrence of up to 20% Zn-dolomite, considered the main reason for Zn loss during processing. The Reef Ridge nonsulphide ore concentrations also consists of smithsonite commonly associated with Fe-(hydr)oxides (mainly goethite), with remnant sphalerite, and minor pyrite/marcasite. Mineralogic Mining, as well as the previously used QEMSCAN, was able to build highresolution maps and to clearly identify and quantify the major phases such as smithsonite and goethite (up to ~70wt.% and ~30 wt.% respectively in the highly oxidized samples), dolomite and sphalerite (up to ~45 wt.% and ~50wt.% respectively in the un-weathered samples). It was also possible to identify the “impure” metal bearing minerals, i.e. Zn-dolomite (from 14 to 26 wt.% Zn on average), Fe-dolomite (from ~4 to ~6wt% Fe on average), Mg-smithsonite (10 wt.% Mg on average), Fe-smithsonite (~28 wt% Fe on average). The technique was clearly able to distinguish between two very similar mineral phases (i.e. ankerite and Fedolomite). The software was also able to automatically calculate the distribution % of Zn in the different mineral phases: from Jabali, results show that the Zn occurs mostly in smithsonite and Zndolomite, while at Reef Ridge Zn is largely in smithsonite and goethite. This information is critical for geometallurgical modelling, feasibility studies and process planning, allowing for quantification of the metal deportment between minerals, predicting metal losses during the treatment. This study outlines the effectiveness of the new generation of automated mineralogy (Mineralogic) in ore characterisation at the exploration stage as an early aid to the evaluation of possible processing problems.