Multi-Analysis Approach to Characterisation of Mining Ores – A Critical Mineral Focus

Le résumé fourni par la source

In modern society, the characterisation of mineral ores has become increasingly important due to the growing demand for critical minerals. Critical minerals comprise a group of both metallic and non-metallic elements essential for our future technologies such as computing, telecommunications, energy and storage applications (Table 1) [1, 2]. Their significance arises from their unique metallurgical, catalytic, nuclear, electrical, magnetic and luminescent properties that underpin many technologies required for economic development and the global energy transition [1, 2]. Australia has identified more than 30 critical minerals vital to clean energy technologies, aerospace and defence industries (Figure 1) [1, 2]. While many elements are geochemically abundant, economically viable concentrations of critical minerals are often of very low concentrations and difficult to extract. There is increasing demand in identifying and characterising alternative sources of these materials, including historic tailings, stockpiles, extraction from recycled materials and preprocessing of waste stockpiles from legacy mining operations [1, 2]. Accurate characterisation of such complex materials requires detailed knowledge of elemental composition, mineral phases, and textural relationships within the ore. Traditional field and bulk analytical techniques alone are insufficient to resolve this complexity. Ores commonly contain multiple phases, fine-grained intergrowths, and chemically variable minerals, particularly in systems enriched in rare earth elements (REEs). Consequently, no single analytical technique can provide all the information required to support effective processing and extraction strategies. This paper presents a multi-technique characterisation approach that integrates bulk and micro-analytical methods, including X-ray fluorescence (XRF), X-ray diffraction (XRD), scanning electron microscopy with energy-dispersive spectroscopy (SEM-EDS), automated mineralogy (MLA-AMICS), X-ray mapping (XRM), Raman spectroscopy, and micro-computed tomography (Micro-CT) [3-15]. The strengths and limitations of each technique are discussed throughout the investigation of several critical mineral–bearing ore samples that were sent for analyses from different regions across Australia. As is often the case, only limited information is provided about the samples and external clients request either 1) analysis for all elements present, which they believed the sample had over 20 different elements, 2) the phases present and 3) what elements were associated with further different elements and phases. Further to this, some requested a rare earth assessment. In this paper, results are mainly shown for a Li ore sample, with some results also shown from a second Zn ore sample. Ore characterisation techniques range from traditional bulk analyses to advanced micro-analytical methods capable of resolving mineral chemistry and textural relationships at the micron scale. Bulk techniques such as XRF and powder XRD provide essential information on elemental composition and dominant mineral phases but offer limited insight into mineral associations, liberation, and fine-scale heterogeneity. SEM-EDS enables detailed examination of grain morphology, phase chemistry, and mineral associations, while automated mineralogy systems extend this capability by allowing large-area, statistically robust mineral mapping. Automated mineralogy (MLA) has been successfully applied across a wide range of commodities, including iron ore, bauxite, mineral sands, laterites, and diamonds [3–9]. In this study, Automated mineralogy (MLA-AMICS) was employed to provide quantitative mineralogical data. Complementary techniques were required to address the inherent limitations of SEM-based chemical analysis. Raman spectroscopy and XRD were used to confirm mineral structures and crystallography, while Micro-CT provided spatial contextual information on particle morphology and internal phase distribution. The integration of these techniques forms a comprehensive workflow capable of resolving both bulk mineralogy and fine-scale chemical variability [11-15]. With regard to the Li ore provided, the initial work involved Micro-CT of a sample of the Li ore, which was prepared in resin ready for analysis with SEM-EDS, XRM and AMICS analysis. Results from the Micro-CT and XRM’s are shown in Figure 2. From the Micro-CT reconstruction of this mixed mineral sample (Li ore), different minerals and phases of different densities can easily be spatially located in the sample. Areas of higher density were then investigated by XRM, which could be easily correlated (Figure 2a-c). For XRD analysis, a subsample was taken and ground to a fine powder using a McCrone mill then spiked with corundum (20 wt.%) as an internal standard. The resulting mixture was then mounted in a Bruker backloaded XRD plate and analysed on a Bruker D8 Advance XRD using a Cu source with a Lynxeye XE-T detector. Phase identification was undertaken using Bruker Eva (V7.1) in conjunction with the ICDD PDF5+ database. A Rietveld refinement was then carried out using Bruker TOPAS (V6) with structure files obtained from the ICDD database. The results are shown in Figure 3 and the quantification of phases determined. In this work, AMICS was also used to characterise a Zn ore concentrate. Large-area mineral maps were generated by integrating backscattered electron imaging with EDS spectra, enabling automated assessment of mineral abundance, associations, and liberation characteristics. Advanced trigger and filter settings were used to switch between conventional mineral liberation analysis and higher-resolution X-ray mapping modes, improving the identification of chemically variable phases [10-15]. Because AMICS primarily provides chemical information, mineralogical interpretation was supported by XRD and Raman spectroscopy to confirm phase identity and crystallography. This integrated workflow enabled iterative refinement of the analytical method and improved confidence in the resulting mineralogical models. XRD experiments were carried out with an internal standard spike of corundum 20 wt.%, Reitveld refinements carried out are shown in Figure 3. Rietveld refinements of XRD experiments of the Li ore (3a) show the dominance of spodumene, followed by quartz, with remaining phases including feldspar minerals, trace phases and amorphous phase/s. While the Zn ore (Figure 3b) indicated that the concentrate was dominated by dolomite, quartz, mica, sphalerite and pyrite with other phases observed as minor phases. AMICS mineral mapping (Figure 4) on Zn ore produced comparable results, however some discrepancies were observed. Discrepancies between the XRD and AMICS datasets, may be attributed by fine texture of Zn ore grains with some observed features below that of the electron beam interaction volume as well as the intrinsic nature of XRD not sensitive to chemical composition and likewise SEM microanalysis not sensitive to crystal species. Due to the inability to directly detect Li X-rays, AMICS was more difficult to perform on the Li ore. Work is ongoing on this topic to identify possible solutions so that AMICS may be carried out on ores with significant Li containing phases. Similar preparation-related effects have been reported in automated mineralogy studies [13-15]. The sample contained a high degree of fine grain material which gives rise to mixed EDS spectra. To reduce the data, generated mixes was employed by comparing to clean spectra and generating a 5% mix of each to obtain the mixes above e.g. Calcite_1(55)-Quartz(45) = a 55% to 45% spectrum mix of Calcite and Quartz. Depending on the scientific question the user is after the next steps in AMICS would be: 1) conduct a higher magnification and resolution scan, 2) target specific grain or mineral types through the use of triggers, and 3) use the TMQ workflow to extract grains, chemistry and X Y coordinates for other

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.