Optimisation and validation of an X-Ray fluorescence method for the analysis of mineral sands oxides

Abstract

Namakwa Sands is a heavy mineral mining and beneficiation business within Tronox, and produces two major products, zircon (ZrSiO4 99.9 %) and rutile (TiO2 99.9 %). Heavy mineral sand deposits occur naturally and are mined for minerals such as Zircon (ZrSiO4), Rutile (TiO2) and Ilmenite (FeTiO3). The heavy mineral concentrates are exported to international markets to make specialist coatings for the paints and ceramics industries and both industries are very strict on the purity of the minerals used. Namakwa Sands prides itself in being able to produce zircon and rutile at the customer requirements however, strict requirements, especially in terms of Fe impurities (Fe2O3 content in zircon concentrate must be < 600ppm), limit the productivity and come at a cost to recovery. Production comes from an open pit where dry mining is employed. The ore is transported by loaders and conveyors to a primary and secondary mineral separation plant where wet spirals (seawater) and magnetic and electrostatic separators are used to produce the marketable concentrates. The ability to control the mineral characterisation of feedstock materials or to monitor intermediate, products and waste streams allows better process control and increased efficiency. Heavy mineral sands concentrate samples are analysed by an X-ray fluorescence spectrometer (XRF) in the form of a fused bead. A worldwide shortage of certified reference materials for the calibration of an XRF instrument for the analysis of heavy mineral sands has resulted in the need to synthesise calibration standards. This was done using synthetic standards made from high purity compounds and mineral sands reference materials. This study addresses the optimisation of X-ray fluorescence calibration through the introduction of synthetic standards for the determination of mineral sands oxides. It examines the application of synthetic standards made from high purity compounds and mineral sands reference materials for the calibration of an XRF spectrometer and enabling it to analyse for major, minor and trace elements (Mg, Al, Si, P, Hf, Ca, Ti, Zr, Fe, Th and U) in heavy mineral sands processing. The analytical conditions suitable for the samples were optimized considering sensitivity, precision, and the lower limit of detection. The fusion method was investigated as a universal method that can yield quality fusion beads and retain all the elements of interest on both the calibration standards and samples. Calibration standards were synthesised by mixing pure compounds and mineral sands reference materials, to mimic matrices similar to that of the routine samples and cover the required analytical range. The aliquots were mixed in % fractions and fused with a 1:9 sample/ borate dilution to make glass beads. The matrix-matched standards and dilution have shown that the matrix effect can be reduced. During XRF analysis interference effects such as spectral overlaps and background correction were compensated to ensure accurate analysis. The main parameters studied were the influence of different flux compositions and sample/flux ratio. The optimum sample preparation conditions were evaluated and confirmed by visual inspection of beads to check dissolution, clearness, shattering and infinite thickness. The optimised calibration was validated as per SANAS TR 26-02 criteria for linearity, the limit of detection and quantification, precision, specificity, and accuracy. It passed all the validated parameters. The calibration lines were developed from synthetic standards and the results were accurate, substantiating that the fusion method can eliminate the mineralogical, matrix and particle size effects. The working range for all the lines was satisfactory. The limit of quantification (LOQ) of analytes was reliably low which shows that the predefined goals for bias and imprecision are met. The correlation coefficients (r2) of the resulting calibration curves were > 0.999 showing excellent linearity. The precision of the calibration was sufficiently high, and the accuracy was of adequate quality with z(MAD) < 2. These observations support the successful synthesis and use of a well-selected set of synthetic standards. The low dilution fused glass technique effectively eliminates particle size effect, and allows accurate determination of both major, minor and trace elements from single glass beads. The study confirmed that the newly developed fusion method can be successfully used to analyse a variety of heavy mineral sand samples to assist in the daily routine analysis of ilmenite, rutile, zircon production. The proposed XRF method can replace the Original method since its calibration can enable quick and effective analysis of elements across a wide range of concentrations in different types of heavy mineral sands streams, giving accurate and trustworthy results