Application of short-wave infrared (SWIR) spectroscopy in quantitative estimation of clay mineral contents

Abstract

Clay minerals are significant constituents of soil which are necessary for life. This paper studied three types of clay minerals, kaolinite, illite, and montmorillonite, for they are not only the most common soil forming materials, but also important indicators of soil expansion and shrinkage potential. These clay minerals showed diagnostic absorption bands resulting from vibrations of hydroxyl groups and structural water molecules in the SWIR wavelength region. The short-wave infrared reflectance spectra of the soil was obtained from a Portable Near Infrared Spectrometer (PNIS, spectrum range: 1300~2500 nm, interval: 2 nm). Due to the simplicity, quickness, and the non-destructiveness analysis, SWIR spectroscopy has been widely used in geological prospecting, chemical engineering and many other fields. The aim of this study was to use multiple linear regression (MLR) and partial least squares (PLS) regression to establish the optimizing quantitative estimation models of the kaolinite, illite and montmorillonite contents from soil reflectance spectra. Here, the soil reflectance spectra mainly refers to the spectral reflectivity of soil (SRS) corresponding to the absorption-band position (AP) of kaolinite, illite, and montmorillonite representative spectra from USGS spectral library, the SRS corresponding to the AP of soil spectral and soil overall spectrum reflectance values. The optimal estimation models of three kinds of clay mineral contents showed that the retrieval accuracy was satisfactory (Kaolinite content: a Root Mean Square Error of Calibration (RMSEC) of 1.671 with a coefficient of determination (R2) of 0.791; Illite content: a RMSEC of 1.126 with a R2 of 0.616; Montmorillonite content: a RMSEC of 1.814 with a R2 of 0.707). Thus, the reflectance spectra of soil obtained form PNIS could be used for quantitative estimation of kaolinite, illite and montmorillonite contents in soil.