Mineral liberation by 3D X-ray microtomography and SEM-based image analysis in low-grade iron ores with different mineralogy and texture

Abstract

X-ray microtomography image analysis has demonstrated a great potential in in three-dimensional and non-destructive mineral samples characterization. Therefore, there is a drive from the international scientific community to develop new methodologies for three-dimensional image analysis, with the commitment to seek increasingly accurate results and better predictability in mineral processing. The main argument for this research approach is the mitigation of stereological bias, which is inherent in the analysis of two-dimensional (2D) images from optical and scanning electron microscopy techniques, despite being very robust, reliable and widely used techniques in mineral research for some decades. However, in addition to stereological bias, other sources of error should be considered, from sample preparation to the inherent limitations of the different analytical techniques. Thus, this work aims to contribute to the critical analysis of 2D and 3D mineralogical results and their impact on mineral liberation spectra of low-grade Brazilian iron ores with different mineralogical assemblages and textural relationships. Considering the four iron ore samples studied, no significant differences were observed in the overall 2D and 3D iron oxides liberation for sample IO-1 (not liberated) and sample IO-3 (with a high liberation degree). For sample IO-2 (with significantly higher liberation below 0.053 mm), the 2D results overestimate the overall liberation by 9 % when compared to 3D data. In sample IO-4, although the 2D liberation results range from 84 to 97 % in the range −0.15 + 0.020 mm, the 3D liberation values reach a maximum of 75–79 % in fractions below 0.053 mm, indicating a high stereological bias in all fractions, with an overestimation of 22 % of the overall liberation of the 2D results.