Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

The imminent challenge in geometallurgical characterization of today’s ore deposits is its inherent heterogeneity and complexity which demands fast, resource-efficient, robust, and cost-effective means of measurement and analysis. To address this concern, mining operations all over the world are incorporating SEM-based automated mineralogy systems. Next generation platforms have been developed to bring out these technologies from the laboratories unto the field for on-site mineralogical characterization. Its widespread use have established its significance in the mining industry, however, cross-validation of results between technologies has not been well studied. 
This study aims to compare and validate results gathered from two SEM platforms: the ZEISS SIGMA 300 Gemini (FEG) with the ZEISS MinSCAN (W-filament) both are coupled with the ZEISS Mineralogic Mining system. Three sets of mixed Cu ore samples (feed, concentrate, and tails) from the Kansanshi deposit were analyzed exploring the effects of SEM operating conditions and sample preparation. The work focused on comparison of the modal mineralogy. The quality of results were assessed by its repeatability and cross-validated with chemical assay (AR-AAS). The limitations of these SEM-based systems were defined and solutions were proposed involving correlative microscopy. The feasibility of employing machine learning algorithms in image classification techniques were proposed to improve data acquisition process and accuracy of results of automated mineralogy systems.
This study establishes the effect of the mineral recipe or SIP to the accuracy of results of a quantitative mineralogical analysis. Expertise in the deposit mineralogy as well as in the capabilities and limitations of the SEM is crucial in achieving reliable results. Cross-validation of back-calculated Mineralogic Cu and Fe grades from modal mineralogy with the measured AR-AAS chemical analysis shows that the error is minimized with the ZEISS SIGMA on samples prepared using carbon black. However, it must be noted that not all numerical values can be directly compared between the results and should be treated merely as indications of accuracy.
Combined features from images obtained from the optical microscope (RGB) and the SEM (BSE) showed the least error in classification using a support vector machine algorithm. The segmented image shows mineral domains where EDX analysis can be performed on a set amount of points veering away from the pixel-by-pixel full mapping acquisition mode. This methodology can be developed and applied to automated mineralogy image processing systems as a domain-based acquisition mode. This has the potential to improve quantitative mineralogical analysis results accuracy while reducing acquisition time and resource costs.

Automated Mineralogy --- Scanning Electron Microscope --- Correlative Microscopy --- Image Analysis --- Mineral Classification --- Process Mineralogy --- Operational Mineralogy --- Ingénierie, informatique & technologie > Géologie, ingénierie du pétrole & des mines

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

This Special Issue, focusing on the value of mineralogical monitoring for the mining and minerals industry, should include detailed investigations and characterizations of minerals and ores of the following fields for ore and process control: Lithium ores—determination of lithium contents by XRD methods; Copper ores and their different mineralogy; Nickel lateritic ores; Iron ores and sinter; Bauxite and bauxite overburden; Heavy mineral sands. The value of quantitative mineralogical analysis, mainly by XRD methods, combined with other techniques for the evaluation of typical metal ores and other important minerals, will be shown and demonstrated for different minerals. The different steps of mineral processing and metal contents bound to different minerals will be included. Additionally, some processing steps, mineral enrichments, and optimization of mineral determinations using XRD will be demonstrated. Statistical methods for the treatment of a large set of XRD patterns of ores and mineral concentrates, as well as their value for the characterization of mineral concentrates and ores, will be demonstrated. Determinations of metal concentrations in minerals by different methods will be included, as well as the direct prediction of process parameters from raw XRD data.

Technology: general issues --- History of engineering & technology --- Mining technology & engineering --- barite --- mineralogy --- industrial application --- beneficiation --- specific gravity --- bauxite overburden --- Belterra Clay --- mineralogical quantification --- Rietveld analysis --- machine learning --- artificial intelligence --- mining --- mineralogical analysis --- bauxite --- available alumina --- reactive silica --- XRD --- PLSR --- lithium --- quantification --- clustering --- Rietveld --- cluster analysis --- spodumene --- petalite --- lepidolite --- triphylite --- zinnwaldite --- amblygonite --- chalcopyrite --- ore blending --- copper flotation --- nickel laterite --- ore sorting --- framboidal pyrite --- sulfide minerals --- flotation --- process mineralogy --- heavy minerals --- ilmenite --- titania slag --- rietveld --- Magneli phases --- n/a

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

This Special Issue, focusing on the value of mineralogical monitoring for the mining and minerals industry, should include detailed investigations and characterizations of minerals and ores of the following fields for ore and process control: Lithium ores—determination of lithium contents by XRD methods; Copper ores and their different mineralogy; Nickel lateritic ores; Iron ores and sinter; Bauxite and bauxite overburden; Heavy mineral sands. The value of quantitative mineralogical analysis, mainly by XRD methods, combined with other techniques for the evaluation of typical metal ores and other important minerals, will be shown and demonstrated for different minerals. The different steps of mineral processing and metal contents bound to different minerals will be included. Additionally, some processing steps, mineral enrichments, and optimization of mineral determinations using XRD will be demonstrated. Statistical methods for the treatment of a large set of XRD patterns of ores and mineral concentrates, as well as their value for the characterization of mineral concentrates and ores, will be demonstrated. Determinations of metal concentrations in minerals by different methods will be included, as well as the direct prediction of process parameters from raw XRD data.

Technology: general issues --- History of engineering & technology --- Mining technology & engineering --- barite --- mineralogy --- industrial application --- beneficiation --- specific gravity --- bauxite overburden --- Belterra Clay --- mineralogical quantification --- Rietveld analysis --- machine learning --- artificial intelligence --- mining --- mineralogical analysis --- bauxite --- available alumina --- reactive silica --- XRD --- PLSR --- lithium --- quantification --- clustering --- Rietveld --- cluster analysis --- spodumene --- petalite --- lepidolite --- triphylite --- zinnwaldite --- amblygonite --- chalcopyrite --- ore blending --- copper flotation --- nickel laterite --- ore sorting --- framboidal pyrite --- sulfide minerals --- flotation --- process mineralogy --- heavy minerals --- ilmenite --- titania slag --- rietveld --- Magneli phases