h1

%0 Thesis
%A Krebbers, Leonard Theodor
%T Computed tomography of critical raw materials: method development and mineralogical characterisation for graphite and tungsten ores
%I Rheinisch-Westfälische Technische Hochschule Aachen
%V Dissertation
%C Aachen
%M RWTH-2025-04496
%P 1 Online-Ressource : Illustrationen
%D 2025
%Z Veröffentlicht auf dem Publikationsserver der RWTH Aachen University
%Z Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2025
%X Critical raw materials (CRMs) are essential to modern society, with graphite and tungsten serving as two examples in this work. Graphite and tungsten are primarily sourced from complex ores requiring rigorous characterisation for evaluating ore deposit quality and maximising recovery. Conventional characterisation methods, while providing essential insights into the mineralogy and geochemistry of ores, involve destructive sample preparation, require careful sectioning and are limited to two-dimensional (2D) information. Moreover, translating this 2D information into three-dimensional (3D) information can introduce stereological bias. X-ray computed tomography (CT) allows for non-destructive 3D analysis of both bulk volume and spatially resolved microstructures of scanned objects, based on the X-ray attenuation information of their constituents, with a resolution down to 1 µm. Over the past two decades, CT has gained increasing attention in the study of ores, enabling, for example, the spatial visualisation of mineral distributions and quantification of grain sizes, shapes and volumes. However, applying CT in ore characterisation presents challenges, particularly due to the varying X-ray attenuation properties of different minerals and the lack of spectral information. To address these challenges, several acquisition and image processing methods have been developed and applied to enhance image quality and the amount of extractable mineralogical information, optimising CT for specific ores. Against this background, CT has been particularly successful in analysing gold and PGE ores. In contrast, the comprehensive mineralogical characterisation of graphite and tungsten ores remains relatively underexplored, highlighting the need for proper CT analysis protocols. This thesis addressed these gaps and aimed to answer the main research question: ‘How can CT be optimised for the characterisation of the critical raw materials graphite and tungsten ores, to obtain reliable qualitative and quantitative mineralogical information?’ The main research aim was to investigate the potential of CT for the characterisation of graphite and tungsten ores. The main research question and the main research aim were addressed through three peer-reviewed scientific studies, forming the experimental work of this thesis. The methods employed included CT with corresponding reconstruction and processing software, X-ray diffraction (XRD), optical microscopy (OM), ultraviolet (UV) light, inductively coupled plasma optical emission spectroscopy (ICP-OES), scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDS) as well as SEM-based automated mineralogy software. Study 1 evaluated the potential of CT for characterising flake graphite ore, focusing on the development of a CT analysis protocol in order to establish critical mineralogical aspects. By considering the ore’s mineralogical composition, accounting for the heterogeneous attenuation characteristics identified using conventional mineralogical methods, high-quality CT data were acquired. The application of deep learning (DL)-based segmentation strategies successfully reduced inherent limitations associated with CT image data, resulting from the similar X-ray attenuation properties of graphite and silicate minerals. The protocol developed enabled accurate quantitative analysis of graphite flake size distribution, shape, impurities, and overall graphite content. Validation against conventional methods confirmed the accuracy of the data, demonstrating the added value of CT for the mineralogical characterisation of flake graphite ore. Study 2 addressed the constraints related to the limited grey value contrast of similar X-ray attenuating materials such as graphite and silicates. Considering that the X-ray attenuation properties of the phases vary with the X-ray energy applied, this study investigated the effectiveness of dual-energy CT for improving image contrast. A sequential fusion approach was employed on a graphite ore sample to combine CT data obtained from different single-energy CT (SECT) scans at high spatial resolution prior to reconstruction, establishing dual-energy CT (DECT) data. In addition, varying weighting factors were applied to determine the optimal contribution of each energy level and spectrum. To evaluate the image quality obtained, a method was developed for quantitatively measuring the image contrast between individual phases. The findings demonstrated that all DECT datasets showed significantly improved image contrast compared to SECT datasets. Furthermore, the image quality measure method developed demonstrated to be an effective tool for comparing image quality between multi-material datasets comprising heterogeneous grey value information. While studies 1 and 2 primarily focused on method development, study 3 formed a case study. Two scheelite ore samples from the Australian Kara Fe-W deposit were examined using CT to establish modal mineralogy, mineral textures, scheelite distribution and tungsten grade. The CT analysis workflow developed in study 1 was applied and optimised to the ores’ specific mineralogical properties. In addition, the fusion approach developed in study 2 was applied to increase the image quality, mitigating acquisition issues. The results showed that scheelite was primarily associated with hydrous phases and occurred predominantly as massive or disseminated vein-fill mineralisation at minor and trace concentrations. The study demonstrated that CT of scheelite ore enabled accurate 3D texture visualisation and yielded valid quantitative data on modal mineralogy and WO3 grade of the samples investigated, ultimately providing relevant information on ore formation and for comminution strategies of scheelite at the Kara Fe-W deposit. The synthesis of the research findings showed that by developing dedicated analysis protocols, reliable mineralogical 3D information of graphite and tungsten ore was generated, showcasing the added value of CT applied to these ores. The ability to reduce greyscale contrast-based analysis constraints, particularly through the use of DL-based segmentation, marks a significant advancement in CT ore analysis. While CT was optimised for graphite and tungsten ores, the strategies developed hold potential for application to other complex ores with similar mineralogical characteristics, contributing to broader applications of CT. However, inherent methodological constraints persist, which, in conjunction with extended analysis time, sample complexity and limited standardisation, present a significant barrier to adoption in the raw materials sector, particularly for operators. Future research focusing on integrating AI along the CT analysis chain and automatisation could significantly lower this barrier.
%F PUB:(DE-HGF)11
%9 Dissertation / PhD Thesis
%R 10.18154/RWTH-2025-04496
%U https://publications.rwth-aachen.de/record/1011028