The use of Laser Induced Breakdown Spectroscopy for the mineral chemistry of chromite, orthopyroxene and plagioclase from Merensky Reef and UG-2 chromitite, Bushveld Complex, South Africa

Abstract

A new approach using scanning Laser Induced Breakdown Spectroscopy (LIBS) was developed to unravel cryptic variations in the mineral chemistry of chromite, orthopyroxene, and plagioclase in drill core from Merensky Reef and UG-2. The LIBS method provides continuous spatially resolved chemical data, elucidating variations in mineral chemistry and in element partitioning, taking into account textural features of the rock itself, of individual minerals, and of immediate neighbourhoods. Since the method is relatively fast and easy to apply to numerous core sections, it is of special interest when comparing the lateral variability of diagnostic horizons such as Merensky Reef and UG-2. This approach for measuring Mg/Fe, Cr/Al and Ca/Na ratios is based on the strength of LIBS in detecting chemical variations using intensity ratios within a single matrix, which in this application is one single particular type of mineral phase. Supervised classification followed by k-means clustering was applied to identify those spectra that consist of pure chromite, orthopyroxene, and plagioclase. Measurements and data processing strategies were focused on minimizing the standard deviation in the LIBS-based ratios to be able to detect relatively small variations in mineral chemistry. The LIBS-based Cr/Al and Mg/Fe ratios were successfully validated with respect to electron probe microanalysis on multiple polished sections. Distinct trends could indeed be extracted from a continuous LIBS measurement of 6 m drill core through the Merensky Reef. The data correlate with whole-rock geochemistry and hydrous phases such as phlogopite. The LIBS-based approach opens up unique insights into chemical and mineralogical expressions of large- and local-scale processes. This allows unique and detailed interpretations relevant for the genesis of mineralizations and metal distributions, based on a comprehensive dataset.