Hyperspectral imaging of coal core: A focus on the visible-near-shortwave infrared (VN-SWIR) region

Abstract

Analytical technology is constantly being developed, refined, and applied to different materials. A key objective is to develop technologies that are non-destructive, rapid, and improve data accuracy. Hyperspectral imaging (HSI) is a non-destructive analytical technique that measures the spectral response of molecular bonds within mineral crystals or organic matter, caused by their excitation by light. The technique has a potential to save time and money for the coal exploration and mining companies. Typically, minerals within borehole cores are characterised based on their unique spectral properties within specific infrared ranges and presented as a function of reflectance versus wavelength. to examine spectra generated on coal core samples using HSI. The HSI spectral data were compared to traditional approaches X-ray fluorescence, X-ray diffraction, proximate data, and Fourier-transform infrared spectroscopy (FTIR). A coal core from Witbank Coalfield, South Africa (Medium Rank C bituminous, inertinite-rich, generally high ash), was examined within the visible-near infrared (VNIR) (350–1000 nm) and shortwave infrared (SWIR) (1000–2500 nm) regions. The HSI coal spectra exhibit positive slopes with low reflectance values within the VNIR region and gradual increase of reflectance values in the SWIR region. The spectra are influenced by very-fine grained clay and Fe-rich minerals (pyrite and siderite) included in the coal; the latter was verified by XRD as pyrite and siderite. The spectra with higher amounts of organic matter are flat and the absorption features are weaker due to the absorbing nature of the carbon. The identified absorption features for coal functional groups within VN-SWIR are 1700 nm (CH), 2200–2206 nm (CH₂, CC, CO) and ∼ 2310 nm (CH₃), which were confirmed by FTIR data. However, the absorption features between 2200 and 2450 nm are affected by overlapping bands of inorganic phases, resulting in uncertainty. The bright banded coal (vitrinite-rich) can be adequately separated from the dull coal (inertinite-rich) through the extraction (D) of D2200 and the deepest feature between D2100 - D2450. The technique can also distinguish the carbonaceous shale from coal, demonstrating the ability to differentiate rock types based on the mineral composition and proportions.