A Quantitative and Comparative Analysis of X-ray Computed Tomography, X-ray diffraction, and Mineral Liberation Analysis


A Quantitative and Comparative Analysis of X-ray Computed Tomography, X-ray diffraction, and Mineral Liberation Analysis

Hassanzadehmahaleh, A.; Da Assuncao Godinho, J. R.; Heinig, T.; Möckel, R.; Ebert, D.; Rudolph, M.

A precise characterization of ores from mine to concentration plants is a long-standing aim of geometallurgical surveys. However, only limited number of studies have focused on prediction of materials properties from rock-size to downstream processes (i.e. ground-size) using combined characterization techniques. Additionally, individual uncertainty of most common characterization methods i.e. 3D X-ray Computed Tomography (CT), X-ray diffraction (XRD) and Mineral Liberation Analysis (MLA) and their quantification cannot be directly measured. Thus, validation of results requires a constructive comparison amongst these methods.
This work aims to validate CT as a reliable technique to characterize ore grains from 3D images for a parisite-bearing sample and its processed one in two comminution stages (i.e. crushing and milling). With this purpose, the amount and grain size distribution of parisite (Ca(Ce,La)2(CO3)3F2) in a carbonate sample was measured in three forms: uncrushed rock, after crushing, and after milling, using CT. After milling, the sample was sieved into five size fractions and each fraction was analysed by XRD, MLA and CT. The amount of each fraction was used to back calculate the initial mass of parisite in the initial uncrushed sample. It is found that the mass of parisite estimated from grain mounts and the mass directly measured in the entire sample as measured using CT are in good agreement. CT yields more consistent results for coarser grain size fractions, e.g. >56 µm, but significantly underestimates the mass % and PSD of finer size fractions. Above 56 µm, MLA shows inconsistencies, possibly due to sampling representability of grain mounts. For particles bellow 56 µm, MLA values are more representative although seemingly overestimate the content of parisite. Finally, CT is validated as complementary to traditional techniques commonly used for ore characterization when the grain size is sufficiently large relative to the voxel size.

Keywords: X-ray Computed Tomography; X-ray diffraction; Mineral Liberation Analysis; geometallurgical analysis; communition

  • Contribution to proceedings
    Procemin-Geomet 2019, 20.-22.11.2019, Santiago, Chile

Permalink: https://www.hzdr.de/publications/Publ-29237
Publ.-Id: 29237