Determination of the predominant minerals in sedimentary rocks by chemometric analysis of infrared spectra

Abstract

The objective of the present study was to determine the predominant minerals in sedimentary rocks using Fourier-transform infrared (FTIR) spectroscopy and chemometric analysis. The chemometric analysis was performed on three types of sedimentary rock samples (claystones, clay slates, and sandstones), each with different predominant mineral components. Chemometric models were created to determine the major minerals of the rock samples studied chlorite, muscovite, albite, and quartz. The FTIR spectra were obtained in transmission mode from pressed pellets of KBr-sample mixtures or by diffuse reflectance from hand-packed mixtures of samples with KBr. Spectral regions measured were 4000-3000 and 1300-400 cm-1, which contained important spectral information for the creation of the chemometric models. Principal component analysis was used in the chemometric method, with calibration models being created by a partial least-squares regression method. The mean relative error, standard error of prediction, and relative standard deviation were calculated for the assessment of accuracy, precision, and reproducibility. The value of the mean relative error was 15-20% for most of the calibration models; the value of the standard error of prediction was up to 6 w/w % for most of the calibration models. The values of the standard relative deviation ranged from ∼2 to 8% for calibration models based on diffuse reflectance spectra whereas calibration models based on transmission spectra had values of relative standard deviation of ∼15-20%.