Novel techniques in ore characterisation and sorting

Abstract

Ore sorting is becoming increasingly important to the minerals industry but is a difficult technology to apply effectively to many ore types. Microwave excitation and infrared detection (MW/IR) is a potential new sensor technology for sorting copper sulphide ores. This technology utilises the high dielectric permittivity of copper sulphide minerals and relatively low dielectric properties of gangue minerals to selectively heat copper rich particles and thereby gauge particle grade from temperature. To date, the dielectric properties of minerals and the role of mineral structures within particles are not fully understood and can vary significantly from deposit to deposit due to differences in geochemistry and genesis. These variations in mineralogy and dielectric properties can mean that some ores are unsuitable for sorting. The aim of this thesis is to develop a characterisation methodology to assess the suitability of ores for MW/IR based sorting from small sets of particles. The microwave heating of six different ores, chosen to represent a range of minerals and mineral textures, was characterised using randomly split particles, a domestic microwave oven and thermal imaging. The dielectric properties of core and powder subsamples of particles from four ores were then measured using a resonant cavity perturbation experiments at The University of Nottingham. Three effects were identified that may influence the observed relationship between microwave heating and dielectric properties: strong interaction with the magnetic field, mineral structures and sulphide heterogeneity in samples. These effects were investigated using radiofrequency magnetic permeability measurements, rotation of rock cylinders in a rectangular waveguide resonant cavity and automated scanning electron microscopy (MLA). The modal mineralogy of ore samples was measured using MLA while geochemistry was investigated using electron probe microscopy. The dielectric properties of mineral groups were then calculated using systems of linear equations based on the Landau-Lifschitz-Looyenga mixture equation as well as modal mineralogy and dielectric property data. Finally, numerical simulations of microwave heating were performed using data on mineralogy, mineral dielectric properties and variability in particle heating to determine to assess the potential of rocks types for sorting. MW/IR characterisation showed that quartzite, monzonite and skarn ores have distinct particle temperature distributions and that variation in the pixel temperatures of individual particle surfaces might be useful in distinguishing particle petrology. A power law was found to describe the relationship between microwave heating and dielectric loss tangent (R2 = 79%) although a good fit could be achieved with a linear relationship (R2 = 89%) when specific, termed anomalous, particles were excluded. Magnetic permeability measurements indicated that there was no significant difference in permeability between normal and anomalously heating particles. Changes in cavity perturbation were observed with core rotation and were consistent with bulk dielectric anisotropy resulting from mineral structures. MLA imaging of the solid sections of skarn samples demonstrated a high degree of sulphide heterogeneity. Calculation of dielectric constant using least squares regression of overdetermined systems of linear equations produced statistically significant results that compared well with values reported in literature but had large standard errors (> 10%). Dielectric loss factors calculated in this manner were not reliable as regressions only produced statistically significant data for some minerals and standard errors were much higher. Grade recovery curves produced by simulation of microwave heating indicated that quartzite ore is likely to be a suitable candidate for MW/IR sorting, however, some monzonite type ores may be unsuitable. Future work on this topic may include: • development of particle surface temperature variation as a metric for petrology • MW/IR characterisation at 915 MHz • Measurements of permeability at microwave frequencies • Development of the resonant cavity texture measurement technique • Development of the regression method for dielectric mixture calculation and validation of the numerical simulation of MW/IR detection.