Estimating modal mineralogy using Raman spectroscopy: Multivariate analysis models and Raman cross-section proxies

• DOI: 10.2138/am-2023-9224
• Publication Date: 2024-04-26T19:03:35Z
• Authors: Laura B. Breitenfeld; M. Darby Dyar; Timothy D. Glotch; A. Deanne Rogers; Miriam Eleazer
• Keywords: Section (typography)

Abstract Raman spectroscopy is a powerful technique in the context of planetary exploration because it provides mineral identification, chemistry, and abundance information. For spectrometers with large spot sizes, multiple phases can be investigated via collection single spectrum. There lack methodology for quantifying species mixtures due to independent signal strengths different materials spectra. Two techniques are presented this work quantification common rock-forming minerals: partial least squares multivariate analysis novel approach called cross-section proxies (numerical metrics associated specific features). This paper targets 20 relevant mineralogy planet Mars. Mineral end-member samples 187 binary mineral-mineral (mixture two distinct minerals) used modeling. Eighteen diamond-mineral derive proxies. proportions predicted known mineralogical content evaluate efficacy quantitative methods. Technique performance mineral-dependent. The root-mean-square error unseen predictions (RMSE-P) using ranges from ±3.2-17.0 volume%. models, cross-validated (RMSE-CV) ±8.8 26.2 Although these estimates not directly comparable, they provide most accurate estimate currently available. Different scenarios may favor use one or other fundamental groundwork that applied minerals on terrestrial planets including Quantification abundances aids answering critical geologic questions related ancient primary altered rocks as well processes conditions.