Raman spectroscopy as an alternative approach for prediction of silicate mineral content in sedimentary rocks

This work presents Raman spectroscopy as an alternative approach for rapid prediction of silicate mineral content in sedimentary rocks. Raman spectra of clay shale and claystone samples were measured by dispersive Raman spectrometer using the excitation laser wavelength of 633 nm. Due to the presence of strong fluorescence and overlapping bands in Raman spectra of claystones and clay shales, partial least squares regression (PLSR) was employed as a multivariate calibration method for semi-quantitative estimation of silicate mineral content. The PLSR models were developed for four predominant silicate minerals of claystones and clay shales – chlorite, muscovite, quartz and albite. The correlation coefficients for calibration and validation samples were > 0.91 and > 0.84, respectively. The root mean square error of prediction (RMSEP) ranged from 1.8 to 4.9 wt %. The developed PLSR models showed good predictive capabilities for control samples and the repeatability of measurements expressed as relative standard deviation (RSD) was < 20%. This approach has the potential to be used for rapid and non-destructive on-site estimation of chlorite, muscovite, quartz and albite content in samples of claystones and clay shales using mobile Raman instruments.