Proximal sensor analysis of mine tailings in South Africa: an exploratory study

Abstract

Proximal sensors have rapidly developed in-situ analysis of soil and sediment. Two systems in particular, portable X-ray fluorescence (PXRF) spectrometry and visible near infrared diffuse reflectance spectroscopy (VisNIR DRS), have gained widespread acceptance for agronomic, pedological, and environmental quality assessment, but have scarcely been utilized to investigate mine tailings in South Africa. This study investigated 419 samples across four mines in South Africa using both PXRF and VisNIR DRS. Specifically, the PXRF was used to provide elemental data for comparison to X-ray diffraction (XRD) and energy dispersive X-ray spectroscopy (EDAX) coupled with scanning electron microscopy (SEM) for confirmation of tailings mineralogy. Next, PXRF data was used to model first derivative (1D) VisNIR reflectance spectra. Results revealed many satisfactory calibrations (R2 > 0.70) relative to PXRF analysis with VisNIR DRS predictive models showing fair (RPD 1.4-2.0) to stable, accurate (RPD > 2.0) prediction of multiple elements. Use of better data partitioning methods and considering the target variability is likely to improve model accuracy further. Summarily, this study has established the effectiveness of using VisNIR DRS for the evaluation of mine tailings, justifying further application development with mine-specific models