Characterization of alteration and mineralization of Chimu Formation quartzite with near-infrared spectroscopy

The use of sensor based technology to distinguish in real-time between economic and sub-economic material could contribute to the optimization of the technology and economics behind mining operations. The goal of this study is to contribute to the development of a near-infrared sensor-based pre-sorting ore classification model. This study is focused on the 11.6 Moz Lagunas Norte high sulphidation epithermal gold deposit in the La Libertad Region, Peru and is part of a research collaboration between Barrick Gold and Delft University of Technology. In the deposit 80% of the gold is hosted in weakly metamorphosed quartzite of the Upper Jurassic to Lower Cretaceous Chimu Fm. A correlation between alteration mineralogy and mineralization is present according to previous authors. In this study a sample set originated from 6 drill holes throughout the deposit was investigated to determine the nature of the correlation between alteration mineralogy, Au and sulphide content by using near-infrared reflectance (NIR) spectroscopy. This method was chosen since alteration minerals typical for the deposit have diagnostic features in this range. Alteration mineralogy was determined with near-infrared reflectance spectroscopy for a whole rock sample suite. This resulted in the identification of mineralogy, typically observed in the advanced argillic zone of a high sulphidation epithermal deposit. The mineralogy was verified by XRD analyses. Furthermore iron bearing mineralogy, typically found in the oxidised breccia zone, was identified. Principal component analyses on the spectra resulted in the observation that the mineral compositions are mixtures between mineral end members, which can be coupled to progressively neutralising of acidic fluids in the hydrothermal system. Subsamples were extracted and were tested both spectrally, petrographic, and geochemical. This resulted in the conclusion that Au was focussed in fault zones. Partial least squares discriminant analyses was used to classify samples based on Au mineralization revealing that pyrophyllite is an indication for Au$0.25%. The Au model had an accuracy of 76% adjusting the threshold would result in a small class which has a high certainty to not contain gold. Hence PLS-DA models showed it is possible to classify ore, based on the near-infrared reflectance spectrum. In which alteration mineralogy can be used as a calibrated proxy for gold and sulphide content at Lagunas Norte. Implementing a pre-sorting step based on NIR would result in a reduction of processed waste, which would result in a reduction of operational expenditure.Applied Earth Sciences / Resource EngineeringGeoscience & EngineeringCivil Engineering and Geoscience