SEM-based quantitative mineralogical analysis of peridotite, kimberlite, and concentrate

Abstract

Quantitative automated mineralogy is the acquisition of mineralogical and textural data by scanning electron microscopy-based energy dispersive analytical methods. The technique is used in the metals and energy resource extraction industries to provide accurate mineral characterization information over large data sets. We have employed this method to three projects related to diamonds: mantle peridotite, kimberlite, and garnet concentrate. The first example assesses the metasomatic clinopyroxene-phlogopite modal mineralogy in peridotite xenoliths from Premier/Cullinan (South Africa). Understanding mantle mineral variability can be coupled with measureable mineral properties to develop mantle geophysical and geochemical models. The second example compares kimberlites from Letšeng Satellite Pipe (Lesotho) and Ngamiland (Botswana), to assess the variability of kimberlite mineralogy. Kimberlite domains can be identified on a micro scale with potential to understand parameters such as hardness and grindability, or to identify kimberlite clusters with discrete mineral assemblages. The third example applies mineral compositional variability in garnet concentrate samples as a possible tool for evaluating exploration projects. Application of garnet definitions to concentrates through digitally grouping grains into discrete populations results in improved understanding of large sample populations and hence diamond prospectivity. Quantitative mineralogy applied to diamond-related projects utilizes the principles of geometallurgy in evaluating large data sets for greater understanding of the variability of mantle materials.