Mineral Identification and Trace Analysis in Complex Mineral Specimens

A suite of heavy mineral sand specimens from the Tista River in Bangladesh were analysed by quantitative x-ray diffraction (QXRD), bulk x-ray fluorescence (XRF), and hyperspectral energy- and wavelength-dispersive spectrometry (EDS and WDS) in an electron probe micro-analyser (EPMA). Specimens were analysed after collection, and again after a series of mineral separation processes aimed at producing concentrates of valuable heavy minerals (VHMs) of zircon, garnet, and ilmenite, which are respectively used for the production of ceramic glazes, abrasives, and titania for pigment production.

The efficacy of the mineral separation processes was assessed by measuring the abundance of all minerals (valuable and gangue) in all samples by both XRD and EPMA mapping. Cluster analysis of the EPMA map data proved particularly successful for identifying minerals and quantifying the abundance and composition of rare and/or unexpected phases, with a total of over 40 phases identified in the suite of samples, many with abundances below 0.1%. The abundances measured by XRD and EPMA were found to be in good agreement.

Phase patched image

Clustered phase map of minerals in the ‘D4 shaker’ heavy mineral sand sample from the Tista River, Bangladesh.

A key achievement of this study was to track changes in the deportment of valuable rare earth elements (REEs) through the mineral separation processes, as well as tracking the deportment of the penalty radioactive elements thorium and uranium. REEs were found to be enriched in the garnet concentrate (6.5% REE phases), compared to 1.7% in the unprocessed material, and were negligible in both the zircon (0.12%) and ilmenite (0.097%) concentrates. Uranium and thorium were likewise enhanced in the garnet concentrate, mostly due to the greater abundance of monazite (5.1 wt% Th, 0.66 wt% U), allanite (1.1 wt% Th), xenotime (0.80 wt% Th, 0.33 wt% U), and florencite-Ce (0.36 wt% Th). The radioactive content of the zircons within the bulk sample and zircon concentrates were both below the minimum detection limits (<60-80ppm), indicating that the zircons of the Tista river are relatively free of these penalty radioactive elements that are present at higher concentrations in other commercial deposits of zircon. The low thorium and uranium content of the zircon concentrate was independently confirmed by trace analysis by bulk XRF.

Energy dispersive x-ray (EDS) spectrum (left) and cathodoluminescence (CL) spectrum (right) of selected phases, extracted from the hyperspectral EDS and CL map datasets using automated data clustering. Phases include xenotime ([Y,HREE,Th]PO4), apatite (Ca5[PO4]3.[OH,F]), zircon (ZrSiO4), titanite (CaTiSiO5), sillimanite (Al2SiO5), quartz (SiO2), plagioclase (NaAlSi3O8 – CaAl2Si2O8), albite (NaAlSi3O8), K feldspar (KAlSi3O8) and grossular (Ca₃Al₂[SiO₄]₃).

The composition of each phase in each sample was determined using standards-based quantitative EDS analysis of spectra extracted post hoc from the hyperspectral map datasets using in-house software (Chimage and EQuiP). These analyses demonstrated minimum detection limits of <100ppm for Th and U, ~100-200ppm for rare earths, and less than 10ppm for elements in abundant phases (e.g. 7ppm Mn and 9ppm Ti in quartz). The analysis method successfully quantified complex REE minerals such as monazite [(LREE,Th)PO4], xenotime [(Y,HREE,Th)PO4] and allanite [(Ce,Ca,Y,Th)2 (Al,Fe)3(SiO4)2(OH)], including the quantification of a challenging suite of 24 elements in xenotime.

These analyses demonstrated the great versatility of hyperspectral mapping in an EPMA for characterising mineral compositions and abundances, as well as finding and quantifying trace elements, all without a priori knowledge of the suite of minerals or elements present. ​​​​​​​

Publications

Aaron Torpy, Nicholas C. Wilson, Colin M. MacRae, Mark I. Pownceby, Pradip K. Biswas, Md Aminur Rahman and Mohammad N. Zaman, Deciphering the Complex Mineralogy of River Sand Deposits through Clustering and Quantification of Hyperspectral X-Ray Maps, Microscopy and Microanalysis (2020), 26, 768–792. DOI: https://doi.org/10.1017/S143192762000135X

P. K. Biswas, S. S. Ahmed, M. I. Pownceby, N. Haque, S. Alam, M. N. Zaman and M. A. Rahman, Heavy mineral resource potential of Tista river sands, Northern Bangladesh. Appl Earth Sci (2018), 127(3), 94–105 DOI: https://doi.org/10.1080/25726838.2018.1488357