Search

Computed tomography (CT) can capture volumes large enough to measure a statistically meaningful number of micron-sized particles with a sufficiently good resolution to allow for the analysis of individual particles. However, the development of methods to efficiently investigate such image data and interpretably model the observed particle features is still an active field of research. When image data of particles exhibiting a wide range of shapes and sizes is considered, traditional image segmentation methods, such as the classic watershed algorithm, struggle to recognize particles with satisfying accuracy. Thus, more advanced methods of machine learning must be utilized for image segmentation to improve the validity of subsequent analyzes. Moreover, CT data does not include information about the mineralogical composition of particles and, therefore, additional SEM-EDS image data has to be acquired. In this paper, micro-CT image data of a particle system mostly consisting of zinnwaldite-quartz composites is considered. First, an image segmentation method is applied which uses deep convolutional neural networks, in particular an adaptation of the U-net architecture. This has the advantage of requiring less hand-labeling than other machine learning methods, while also being more flexible with the possibility of transfer learning. Then, fully parameterized models based on vine copulas are designed to determine multivariate probability distributions of descriptor vectors for the size, shape, texture and composition of particles -- allowing for the estimation and interpretable characterization of interdependencies between particle descriptors.

Data publication: The spatial association of accessory minerals with biotite in granitic rocks from the South Mountain Batholith, Nova Scotia, Canada D. Barrie Clarke; Axel D. Renno; David C. Hamilton; Sabine Gilbricht; Kai Bachmann Related to publication Geosphere (2022) 18 (1): 1–18; https://doi.org/10.1130/GES02339.1 We use mineral liberation analysis (MLA) to quantify the spatial association of 15,118 grains of accessory apatite, monazite, xenotime, and zircon with essential biotite, and clustered with themselves, in a peraluminous biotite granodiorite from the South Mountain Batholith in Nova Scotia (Canada). A random distribution of accessory minerals demands that the proportion of accessory minerals in contact with biotite is identical to the proportion of biotite in the rock, and the binary touching factor (percentage of accessory mineral touching biotite divided by modal proportion of biotite) would be ~1.00. Instead, the mean binary touching factors for the four accessory minerals in relation to biotite are: apatite (5.06 for 11,168 grains), monazite (4.68 for 857 grains), xenotime (4.36 for 217 grains), and zircon (5.05 for 2876 grains). Shared perimeter factors give similar values. Accessory mineral grains that straddle biotite grain boundaries are larger than completely locked, or completely liberated, accessory grains. Only apatite-monazite clusters are significantly more abundant than expected for random distribution. The high, and statistically significant, binary touching factors and shared perimeter factors suggest a strong physical or chemical control on their spatial association. We evaluate random collisions in magma (synneusis), heterogeneous nucleation processes, induced nucleation in passively enriched boundary layers, and induced nucleation in actively enriched boundary layers to explain the significant touching factors. All processes operate during the crystalliz

Compilation of all available raw data for the publication "Metal Deportment in Complex Secondary Raw Materials: The Case of Vanadium in Basic Oxygen Furnace Slags". List of Supplementary Information Table A 1: Compilation of elements analyzed by XRF including lower and upper limits of determination (LoD). Table A 2: Compilation of the subsample designations of the three BOS samples analyzed for the XRF and XRD as well as the MLA and EPMA analyses. All subsamples correspond to representative subsets of the bulk sample. Table A 3: Compilation of the EPMA measurement parameters with spectrometer position (Spec), lower background (LB), upper background (UB), dwelltime on peak (DTP) and background (DTB) and the complete list of reference materials. Table App 4: Compilation of all major and minor element contents determined by XRF recalculated as “water-free” and normalized to 100 wt.-%. Table A 5: Comparison of LOI values for one sample each of the delivered material "as delivered" and fully hydrated. Table A 6: Compilation of significant differences between the samples with regard to the chemical composition. Table A 7: Detailed compilation of the results of the MLA measurements of all investigated samples. Table A 8: Compilation of the phases predefined for EPMA analyses and the number of measurements performed and usable for further MLA and deportation analyses. Table A 9: Complete summary of all EPMA data used for the deportment analysis (xls File).

A peculiar type of stratabound tungsten mineralization in metacarbonate rocks was discovered and explored at Mallnock (Austria) during the late 1980s. The tungsten prospect is located in the Austroalpine Drauzug-Gurktal Nappe System recording polyphase low-grade regional metamorphism. It is the only tungsten occurrence in the Eastern Alps where scheelite is associated with wolframite (96 mol% ferberite). Scheelite and ferberite occur as polyphase stockwork-like mineralization in Fe-rich magnesite in the northern ore zone (Mallnock North), while in the western ore zone (Mallnock West), only scheelite is hosted in dolomitic marbles. LA-ICP-MS analyses of scheelite and ferberite yield low concentrations of Mo, Nb, Ta and REEs, while Na and Sr are enriched; particularly U has high concentrations of up to 200 µg/g making scheelite a suitable target for U-Pb dating. In-situ U-Pb dating of scheelite yielded Early Permian (293 ± 18 Ma) and Middle Triassic (241 ± 6 Ma) ages. A monzodioritic dyke close to Mallnock yielded an U-Pb apatite age of 282 ± 5 Ma. Raman spectroscopy of carbonaceous material yielded maximum temperatures of 296 ± 27 °C and 258 ± 27 °C, that are assigned to the Variscan and Eoalpine metamorphism, respectively. The data are used to interpret tungsten mineralization in the geodynamic context. The new age data show that the metallogenetic model of tungsten mineralization in the Eastern Alps must also consider Permian-Triassic events and indicate that ore formation at Mallnock could be seen in the context of magnesite-siderite mineralization in the Eastern Alps.

In the context of mineral processing, the separation of particles of some feed material into a valuable and a waste fraction is commonly achieved by processes like magnetic (1) or flotation separation (2). The outcome of such processes often depends on particle properties like wettability, but it can also depend on the composition, size and shape of particles. Therefore, with regard to the goal of optimizing separation processes, it is useful to analyze different particle descriptors when studying the influence of wettability on the particle separation behavior. A common tool for classifying particle separation processes are Tromp functions (3). Recently, multivariate Tromp functions, computed by means of non-parametric kernel density estimation, have emerged which characterize the separation behavior with respect to multidimensional vectors of particle descriptors (4). However, estimating multivariate probability densities of particle descriptors using methods of kernel density estimation requires a relatively large sample size (5). Therefore, we propose an alternative parametric approach based on copulas (6) in order to compute multivariate Tromp functions and, in this way, to characterize the separation behavior of particle systems, see Figure 1. Moreover, the parametric modeling approach is extended by an optimization routine to handle separation processes when measurements are not available for all separated fractions. A potential application of the optimization routine is to reduce the measurement effort in a series of separation experiments for a given feed material and various separated fractions. A simulation study has been performed in order to quantitatively compare the parametric with the non-parametric modeling approach. This comparison focuses especially on scenarios for which only a relatively small number of particles is observed in measurements. Such scenarios occur, for example, when image data of particle systems are considered, from which descrip

Many separation processes are based on particle properties like wettability, composition, size and shape. Therefore, it is necessary to analyze the influence of different particle properties on the particle separation behavior. A common tool for classifying particle separation processes are Tromp functions. Recently, multivariate Tromp functions, computed by means of non-parametric kernel density estimation, have emerged which describe the separation behavior with respect to multidimensional particle properties. In this paper, an alternative flexible parametric modeling approach is proposed to model the separation behavior of particle systems with multivariate Tromp functions observed by mineral liberation analyzer (MLA) image measurements. However, different particle properties such as particle wettability cannot be observed in MLA data, although many separation processes such as flotation-based particle separation processes rely on differences in wettabilities. In order to analyze the influence of wettability on particle separation behavior, bivariate Tromp functions for area-equivalent diameter and aspect ratio of differently shaped glass particle systems with differently modified wettabilities are computed in a case study. Comparing the computed Tromp functions reveals the influence of particle wettability on the separation behavior. In addition, we extend the parametric approach to model multivariate Tromp functions to handle separation processes when image measurements are not available for all separation streams.

The quality of results obtained by a scanning electron microscope-based automated mineralogy systems strongly depend on the project-specific library of spectra of the minerals considered to form the samples analysed, the mineral table. Mineral tables can be selected from existing universal libraries according to the expected or suspected minerals, and/or can be constructed based on spectra collected in the samples of a specific project. One can prefer extense mineral tables to enhance the chances of capturing variations in chemical composition of specific minerals; or else smaller, compact mineral tables to reduce misclassification. No general, perfect balance within generality-specificity or extension-compaction exists, so that users end up interactively and iteratively building the spectral library for each project in tedious steps of adding and removing mineral spectral candidates. This process is also project- and ore-specific. While some automated mineralogy devices provide operating modes for automatically constructing mineral lists throughout a measurement, these commonly offer only limited settings and are not clear about the data processing steps. The goal of this contribution is to compare the performance of several components of a strategy to automatically construct automated mineralogy mineral lists - making use of several machine learning algorithms, for the specific case of dataset collected with the Mineral Liberation Analyser (MLA). The strategy has five steps: (1) preliminary data transformation, (2) dimension reduction, (3) endmember detection, (4) phase detection, and (5) unmixing. For each of these steps, several options were tested. These included for data transformation peak extraction and Box-Cox transformations, which at the same time embraces logarithm/log-ratio transformations, square root transformations and the identity transformation. Regarding dimension reduction, principal component analysis and minimum-maximum autocorrelation factors

Reevaluating tailings material originating from a high-sulfidation epithermal gold deposit has unveiled the potential for a holistic remining endeavour. The primary objective of this operation would be the mitigation of minerals containing penalty elements, specifically sulphur in pyrrhotite, a major contributor to acid mine water drainage, alongside the concentration of precious elements like gold. Furthermore, exploring the applicability of the silicate fraction for industries such as ceramics, glass, and geopolymer production has been considered. To accomplish these objectives, an initial drill core campaign featuring six drill holes was executed, accompanied by a thorough material characterization using automated mineralogy, geochemical assays, and hyperspectral analysis. The subsequent step involved a geometallurgical domaining process based on 64 geochemical assays, particle size measurements, and mineralogy assessments. A Mahalanobis distance hierarchical cluster analysis was employed to differentiate domains, and predictions for these domains were extended to all hyperspectral imaging samples. The outcome of this comprehensive approach revealed the delineation of four distinct domains, each characterized by variations in modal mineralogy and trace elemental contents. This strategic analysis provides valuable insights into the heterogeneity of the tailings material, laying the groundwork for targeted interventions to address environmental concerns and maximize the extraction of valuable resources.

Practically all particle separation processes depend on more than one particulate property. In the case of the industrially important froth flotation separation, these properties concern wettability, composition, size and shape. Therefore, it is useful to analyze different particle descriptors when studying the influence of particle wettability and morphology on the separation behavior of particle systems. A common tool for classifying particle separation processes are Tromp functions. Recently, multivariate Tromp functions, computed by means of non-parametric kernel density estimation, have emerged which characterize the separation behavior with respect to multidimensional vectors of particle descriptors. In the present paper, an alternative parametric approach based on copulas is proposed in order to compute multivariate Tromp functions and, in this way, to characterize the separation behavior of particle systems. In particular, bivariate Tromp functions for the areaequivalent diameter and aspect ratio of glass particles with different morphologies and surface modification have been computed, based on image characterization by means of mineral liberation analysis (MLA). Comparing the obtained Tromp functions with one another reveals the combined influence of multiple factors, in this case particle wettability, morphology and size, on the separation behavior and introduces an innovative approach for evaluating multidimensional separation. In addition, we extend the parametric copula-based method for the computation of multivariate Tromp functions, in order to characterize separation processes, also in the case when image measurements are not available for all separated fractions.

Discordant iron-rich ultramafic pegmatites (IRUPs) intersect the UG2 chromitite at many places in the Bushveld Complex. The effects of IRUP interactions on the UG2 ore mineralogy and PGE grade are assessed at the Thaba mine, north-west Bushveld, based on a borehole profile through the UG2 layer and detailed analysis of mineral textures and compositional variations across the UG2-IRUP contacts using micro-XRF element mapping. The UG2-IRUP interaction operated at different scales and probably by different mechanisms. At the local scale (< 10 cm), a thin layer of Fe–Ti–Cr spinel and ilmenite formed on the IRUP side of the contact with UG2, whereas the UG2 chromitite developed grain coarsening, loss of interstitial silicates, and chemical gradients in Cr, Al, Fe, and Ti that extend a few centimeters from the contact into the seam. These local effects are attributed to the intrusion of IRUP melt into the solidified UG2 layer, followed by re-equilibration of the oxide minerals across the contact during cooling. On a larger scale, changes in the ore and gangue mineral assemblages in UG2 took place throughout the entire meter-thick main seam. Compared with regional UG2 compositions, chromite has higher TiO2 and lower Mg#, and there is an anomalously low abundance of interstitial plagioclase. The IRUP-affected UG2 shows relatively abundant secondary hydrous silicates, replacement of PGE sulfides by PGE alloys and PGE-As–Sb–Bi–Te–Pb phases, and formation of secondary Ni–Cu–Fe sulfides after pentlandite and chalcopyrite. These large-scale effects are attributed to hydrothermal fluids derived from IRUP melts. The IRUP bodies at the Thaba mine caused redistribution of PGE within the UG2 layer but did not significantly reduce the overall grade. However, significant changes in the ore mineral assemblage and an increased abundance of secondary silicates can reduce the efficiency of PGE recovery.

With the increasing demand for metalliferous and mineral raw materials and the consequent depletion of the global natural resource base, the possible utilization of secondary raw material sources is receiving more and more attention. In the present study, we present results from a detailed vanadium deportment study of three basic oxygen furnace slag (BOS) samples known to containing elevated bulk concentrations of vanadium. Complementary analytical methods that were used to quantify the abundance and composition of V-containing phases include SEM-based automated mineralogy, X-ray fluorescence analysis, and X-ray powder diffraction as well as electron probe microanalysis. The vanadium deportment was quantified using Monte-Carlo simulations of the data obtained from automated mineralogy and electron microprobe analysis. The total V concentrations identified by XRF are between 1.7 and 2.2 wt.% V. The most important hosts of vanadium are larnite-, brownmillerite- and portlandite-solid solutions. In two samples Ca carbonates also significantly contribute to the V deportment, while wuestite, lime, and native iron do not contribute significantly to the vanadium deportment. A thorough consistency check identifies considerable uncertainties in the density of the V-bearing phases as the most likely reason to explain remaining discrepancies between measured and calculated V values

In this study, synthetic pure cassiterite and cassiterite doped with two different Fe contents were successfully recrystallized by means of sintering. Their crystal structure and chemical compositions are characterized by X-ray powder diffraction (XRD) as well as scanning electron microscopy (SEM) combined with energy-dispersive X-ray (EDX) analysis. Their floatability was studied by microflotation with a diphosphonic acid surfactant named Lauraphos301 as a collector. Unlike the addition of ferric ions in solution, which strongly depressed the floatability of all the cassiterite samples, a much higher flotation efficiency of the Fe-doped cassiterite samples was found especially at lower collector concentrations. The cassiterite floatability is proportional to the Fe content in the cassiterite at a broad range of pH, and the recovery has the following order: Cassiterite with 1417 ppm Fe > cassiterite with 1165 ppm Fe > pure cassiterite The electrokinetic behavior of the cassiterite samples with and without the collector was studied by electrophoretic measurements and revealed that the chemical interaction dominated the adsorption. With the help of the particle shape analysis, a more angular shape was found for the Fe-doped cassiterite samples. Moreover, without the influence of particle shape, much abundant adsorption of Lauraphos301 was found on the Fe-doped cassiterite samples by AFM topography imaging. The minor amount of Fe in the cassiterite lattice as well as a more angular shape of the Fe-doped cassiterite samples were believed to enhance floatability collectively. The study reveals that the influence of the chemical composition of the minerals on flotation was almost inextricably bound up with particle morphology and emphasizes the importance of considering both factors and investigating them individually for the flotation study.