Advanced mineral characterization and petrographic analysis by μ-EDXRF, LIBS, HSI and hyperspectral data merging

Abstract

Petrography and mineralogy are fundamental for understanding processes in various geoscientific fields. Plutonic rock nomenclature is based on mineralogical composition. Therefore, identifying and quantifying minerals is a key for plutonic rock classification. To accomplish this, novel advancements in instrumentations, such as energy dispersive x-ray fluorescence mapping (μ-EDXRF) or hyperspectral imaging (HSI) provide fast and non-destructively spatially resolved and large-scale chemical information. This work describes a comprehensive approach to combine chemical, mineralogical and textural information from μ-EDXRF, Laser Induced Breakdown Spectroscopy (LIBS) and HSI for petrographic analysis of plutonic rocks. Using supervised classification of spectral information, mineral distribution maps are obtained for image analysis including geometrical data of each grain, such as grain size, grain orientation and grain location for subsequent targeted analysis and the modal mineralogy for plutonic rock classification in a QAPF-diagram for 20 rock slabs. The combination of the three mapping techniques can provide valuable information within the limit of each technique such as of spatial resolution or element sensitivity, but with little time needed for sample preparation and measurement. In general, it is an objective, repeatable and quantifiable way for modal mineralogy and petrographic image analysis.