Development and Calibration of a Quantitative, Automated Mineralogical Assessment Method Based on SEM-EDS and Image Analysis: Application for Fine Tailings

Abstract

Quantitative mineralogy has seen significant developments from the combination of scanning electron microscopy (SEM) with automatic image analysis and energy dispersive X-ray spectrometry (EDS). The mining industry is one of the fields that has benefited from this progress. In this paper, the authors present a newly developed quantitative method based on SEM-EDS and image analysis (IA), which is used to determine the mineralogical and environmental characteristics of mine tailings. The main objectives of the method are to be able to characterize sulphides and carbonates as monomineral particles, which control the acid generation from the tailings. Pure sulphides, calcite and quartz were blended to make mineralogical standards that represent typical mine tailings environmental behavior. The SEM-EDS-IA method achieved good mineralogical precision for medium (1-20 Wt%) and abundant (> 20 Wt%) minerals, with a relative error below 10 %. However, some corrections had to be applied to account for typical stereological effects (apparent particle diameter from polished surface) and preparation modes (particle segregation during resin hardening). Particle size analysis was used to calibrate the method and identify the corrections to be applied. Since mineralogical quantifications are based on the area of the observed particles, the most reliable particle size analyses (also obtained from particle area) typically lead to the best mineralogical characterization. However, the SEM based techniques may show some limitations for fine-grained particle quantification (< 10 µm), which required additional corrections. In this article, the technique is described, and it is applied to characterize fine-grained mine tailings with a size-by-size mineralogy (with

Optimization and Quality Control of Automated Quantitative Mineralogy Analysis for Acid Rock Drainage Prediction

“…Mermillod-Blondin et al [46] developed mathematical corrections for automated mineralogical results generated from hardened epoxy sections with the objective of reducing the impact of the segregation effect. However, the validity of using Stoke's law, in which particles are assumed to be smooth and spherical and the material homogeneous, to account for the phenomenon is questionable.…”