Raman spectroscopy offers a nondestructive means to identify minerals in rocks, but the ability to use the technology for quantitative mineralogical analysis is limited by fluorescence that can mask the spectral features of minerals. In this paper we apply continuous wavelet transformation (CWT) to remove fluoresence from Raman data acquired from 26 carbonate rock samples. We then record the intensity values of individual spectral features, proxies for mineral abundances, using the original Raman data and the thus inferred CWT data. The intensity values are then compared against the known mineral abundances determined using the scanning electron microscope (SEM) technology. This comparison is conducted using a linear regression model to determine whether fluorescence removal enhances the mineral abundance predictions. Our results suggest that CWT enhances the accuracy of mineral abundance estimates, thus highlighting the importance of fluorescence removal when using Raman for quantitative mineralogical analysis.
01/09/2015 → 31/12/2019
Projekti: Academy of Finland: Other research funding