Abstract

Hyperspectral analysis has had an increasing acceptance and application in the mining industry for mineral identification and characterization due to its rapid data acquisition, relatively low cost and non-destructive nature. Recognition of mineral species with this technique is achieved by identifying their distinctive spectral features in specific wavelengths of the electromagnetic spectrum. Until now, hyperspectral analysis in mining industry has been focused on gangue minerals as oxides, silicates, sulfates, carbonates, among others. Even though scientific literature has reported that sulphide minerals have characteristic spectral features in the far infrared range of the electromagnetic spectrum (FIR: 15-100 μm), the hyperspectral techniques have not been successfully applied in detecting those minerals, in fact, there is currently no industrial hyperspectral analysis equipment that allows identifying sulphides. The most known equipment of this type, as Corescan or HyLogging Systems, work within visiblenear infrared to thermal infrared ranges (from 0.4 to 14.5 μm approximately). In the present research, by doing mathematical analysis of the measured spectra, it was possible to detect indirectly the presence of non-disseminated sulfides in the thermal infrared range (TIR: 6.5-14.5 μm), since reflectance around 7.6 μm was found to be significantly higher in samples that contain these minerals (e.g. pyrite, chalcopyrite or bornite). Based on this particular behavior, spectral filters were created, in order to separate sulfides from other mineral species that also raise reflectance values around 7.6 μm wavelength, such as iron oxides (e.g. specular hematite). Sulfide detection through hyperspectral analysis in TIR range will help to improve the interpretation of ore and gangue minerals associations, providing new tools to optimize and improve decision-making within mining processes.