Name Surname
AI-assisted mineralogical and textural characterization of lithium-cesium-tantalum (LCT) pegmatites: Applications for pegmatite crystallization models and mineral processing
Main Supervisor: Anouk Borst (KUL)
Co-Supervisors: Laurent Arbaret (University of Orleans), Jeannette Meima (BGR)
Objectives
This project aims to develop new protocols for textural and mineralogical characterization of coarse-grained LCT pegmatites using data fusion modelling and machine learning algorithms to integrate multimodal (spectral and geochemical) imaging data at multiple scales and resolutions. Pegmatites are coarse-grained igneous rocks and provide important sources for a range of critical metals such as Li, Ta and Be, among others. Detailed mineralogical and textural data are essential to evaluate pegmatite genesis models and improve metallurgical flowsheets for lithium extraction. However, the characterization of LCT pegmatites is often hampered by 1) their anomalously coarse grain size, 2) dominantly white or transparent mineral assemblages, and 3) the difficulty of detecting lithium using X-ray-based methods such as EDS or EPMA. Multimodal data (2D multispectral imaging, LIBS and RAMAN mapping) will be collected from LCT pegmatite samples (already available and at disposal for the project) with different grain sizes, intergrowths and mineral assemblages, using case studies and available samples from the world-class Manono-Kitotolo pegmatites (Democratic Republic of Congo. Data fusion modelling and machine learning algorithms will be applied to develop an automated workflow to extract information and improve our knowledge of the evolution of Li-bearing pegmatite systems.
Epected Results
- Define new multimodal analytical acquisition protocols for the characterization of LCT pegmatites;
- Significantly improve our knowledge of the evolution of LCT-pegmatite systems.