Evaluation of mineral compositions is a widely used approach in resource exploration strategies where preparation time and cost may prove to be an important factor. In research institutes it is highly beneficial to determine the major element composition of minerals prior to their destructive analysis for trace elements and radiogenic isotopic ratios thus allowing a comprehensive interpretation of mineral petrogenesis. For the analysis of unique and small (submilligram) samples, avoiding sample loss is a key issue in ultimately producing high quality geochemical data. Consequently here we evaluate the precision and accuracy of electron probe microanalysis of unpolished garnet, olivine, orthopyroxene and clinopyroxene grains by comparison of analyses performed on polished thin sections of the same minerals. By utilizing a protocol that focuses on flat mineral surfaces, rejects analyses with low totals (<. 90%) and major element compositions, magnesium numbers and stoichiometry outside two standard deviation, results had on average a reproducibility of 1.3 times the relative standard deviation of the results of polished thin sections. Major element ratios are indistinguishable from the thin section results. For example, the Mg# for clinopyroxene and olivine is within 0.4% and for garnet within 1-1.5%. Individual analyses of minerals with flat surfaces such as clinopyroxene had a higher rate of success (73%) than minerals with a more variable surface topography such as conchoidally fractured garnet (40%), underlining that a flat topography is the controlling factor in EPMA analyses. These tests establish that accurate and reproducible EPMA analysis can be produced on unpolished minerals that are within error of conventional thin section analyses. The technique is predicted to be of particular use in diamond exploration strategies where knowledge of the geotherm beneath exploration areas is a key parameter. Integrated studies of composition and geochronology of mineral inclusions in diamonds have the potential to significantly improve the understanding of diamond formation processes and the imposed octahedral morphology of the inclusions mean that they have flat crystal faces, ideal for analysis using the proposed methodology.
Timmerman, S., Matveev, S., Gress, M. U., & Davies, G. R. (2015). A methodology for wavelength dispersive electron probe microanalysis of unpolished silicate minerals. Journal of Geochemical Exploration, 159, 243-251. https://doi.org/10.1016/j.gexplo.2015.09.016