Prediction of Acid Rock Drainage (ARD) from calculated mineralogy

The acid-forming potential of ore and waste can be calculated based on a detailed knowledge of mineralogy, especially sulphide and carbonate contents. However, most mineralogical techniques (e.g., semi-quantitative X-ray diffraction (qXRD), scanning electron microscopy/energy dispersive spectroscopy (SEM/EDS) point counting) are too expensive for routine application. Mineralogy can be calculated from assay data using linear programming (simplex method) which is a mature method with application to real world quantities that cannot be negative. In order to apply this method, a table of mineral compositions is required for all the significant minerals in the study area. Unlike least squares methods, the mineral list can exceed the number of elements included in the assay data. Several carbonate compositions with a range of neutralising potential can be included. To use the linear programming method, a calibration must be established based on known compositions. This calibration can be based on qXRD or SEM/EDS point counting methods.

Not all types of assay data are sufficient for calculating mineralogy reliably. The best assay data comes from X-ray fluorescence analysis of major elements, including “loss on ignition” (LOI). Adding measured C content to this analysis provides a robust data set for calculating sulphide and carbonate contents of rocks. The mineralogy can be calculated without measured C, if LOI and SiO₂ are included in the analysis. However, typical mine databases contain multi-element assays based on a four-acid digestion method. In this case SiO₂ is not analysed and there is no “LOI” or total C. With typical four acid digestion data it is not possible to estimate the original carbonate content even when the mineralogy is simple. In rocks with complex mineralogy, mixed carbonates and/or multiple sulphides, qXRD and full chemical analyses are required to calculated acid rock drainage potential from mineralogy.