Martin Luther University Halle-Wittenberg

Further settings

Login for editors

Comparison of different comminution products of Kupferschiefer-type ore by using “feature analysis” and granulometric data

The study of the present BSc-thesis has been conducted within the Economic Geology and Petrology Research Unit at the Institute for Geosciences and Geography, Martin Luther University Halle Wittenberg. Comminuted ore samples from the German and Polish Kupferschiefer have been studied by particle analyses. Samples have been comminuted in a comparative study of products from traditional ball mills and the new VeRo Liberator®.

“Feature analysis” is an optional tool in the software-suite “Esprit” from Bruker Nano GmbH, known to be specialized in instrumental analytics. This tool automatically detects particles using a scanning electron microscope and delivers chemical compositions for the detected particles by EDX-analyses. Additionally, various particle properties, e.g. particle perimeter or roundness, are also analysed. In order to get reliable analyses, settings must be carried out at the beginning of the measurement process, e.g. the choice of the sector analysed, filter adjustments as well as the binarisation of the particles in the selected area. After the analysis is completed, it is necessary to correct the detected mineral classes manually. If, for example, chalcopyrite is closely intergrown with dolomite, in each case 50% of the total area of the particle is attributed to both mineral classes. Minerals with nearly equal chemical compositions such as chalcocite and covellite are assigned to the same class, i.e. “copper sulphides”. However, the measured particle spectra can and need to be corrected manually. Although this procedure is more exact, it is time-consuming and demands experience with the scanning electron microscope, the analysis tool, and the specific sample material.

The comparison of the mineral distribution from Mansfeld and Rudna black shale ore reveals a strongly variable composition of the gangue minerals. In this case, the black shale ore from Rudna contained significantly higher portions of carbonate minerals in comparison to the Mansfeld black shale ore. Rudna black shale ore is dominated by copper and lead sulphides, whereas black shale ore from Mansfeld is characterised by a zinc-rich mineralisation.

Fig.1 Comparison of the mineral classes (area percent) of samples from Rudna and Mansfeld. Former rocks contain a higher percentage of carbonate but less quartz and aluminium silicates.

Fig.1 Comparison of the mineral classes (area percent) of samples from Rudna and Mansfeld. Former rocks contain a higher percentage of carbonate but less quartz and aluminium silicates.

Fig.1 Comparison of the mineral classes (area percent) of samples from Rudna and Mansfeld. Former rocks contain a higher percentage of carbonate but less quartz and aluminium silicates.

Fig.2 Comparison of the mineral classes (area percent) of samples from Rudna and Mansfeld. The latter ore contains a higher percentage of sphalerite. Other sulphides are less widespread compared with the samples from Rudna.

Fig.2 Comparison of the mineral classes (area percent) of samples from Rudna and Mansfeld. The latter ore contains a higher percentage of sphalerite. Other sulphides are less widespread compared with the samples from Rudna.

Fig.2 Comparison of the mineral classes (area percent) of samples from Rudna and Mansfeld. The latter ore contains a higher percentage of sphalerite. Other sulphides are less widespread compared with the samples from Rudna.

The particle size analyses were filtered electronically to grain sizes < 100 µm and compared with lasergranulometrical data as well as particle size distribution data from Mineral Liberation Analyses. It turned out that the feature analysis is a good application to provide grain size distribution and to estimate the size reduction ratio. However, the feature analysis tool is not able to allow statements about the degree of liberation of minerals. Only subjective assessments, which are observed during the general SEM-investigation, can be provided.

Fig. 3 Grain size distribution from results of the sieve analysis (red) and the feature analysis (green). The shape and geometry of both particle curves show small differences. The sample material is a black shale ore, comminuted by VeRo Liberator®, equipped with round tools.

Fig. 3 Grain size distribution from results of the sieve analysis (red) and the feature analysis (green). The shape and geometry of both particle curves show small differences. The sample material is a black shale ore, comminuted by VeRo Liberator®, equipped with round tools.

Fig. 3 Grain size distribution from results of the sieve analysis (red) and the feature analysis (green). The shape and geometry of both particle curves show small differences. The sample material is a black shale ore, comminuted by VeRo Liberator®, equipped with round tools.

Fig. 4 Grain size distribution of results of the lasergranulometric (blue) and the feature analysis (red). The developing of the grain lines shows divergences from up to 20%, but the course is similar. The feature analysis detect coarser particle than the lasergranulometric data. The sample material is a black shale ore comminuted by VeRo Liberator®, equipped with round tools.

Fig. 4 Grain size distribution of results of the lasergranulometric (blue) and the feature analysis (red). The developing of the grain lines shows divergences from up to 20%, but the course is similar. The feature analysis detect coarser particle than the lasergranulometric data. The sample material is a black shale ore comminuted by VeRo Liberator®, equipped with round tools.

Fig. 4 Grain size distribution of results of the lasergranulometric (blue) and the feature analysis (red). The developing of the grain lines shows divergences from up to 20%, but the course is similar. The feature analysis detect coarser particle than the lasergranulometric data. The sample material is a black shale ore comminuted by VeRo Liberator®, equipped with round tools.

Using the feature analysis, comminution products of the VeRo Liberator® and the ball mill have been compared. It appears that the feed material, comminuted by the VeRo Liberator® is separated preferentially along particle boundaries, which eventually has been proven by the measured particle properties. Generally, the particle size distribution of the ball mill product displays a finer comminution product in the grain size fraction below 100 µm. However, 45 % of the particles in the ball mill product are smaller than 5 µm that represent a particle size range, in which a recovery during following mineral procession steps is typically difficult to impossible. Comminution with the VeRo Liberator® is achieved by a single pass, whereas the ball mill will typically grind until a grain size smaller than a set limit (commonly < 100 µm) is reached. SEM-observations and MLA-data indicate that the degree of liberation of sulphides is generally higher in the VeRo Liberator® products compared to the studied ball mill products.

Fig. 5 Cumulative distribution curves of the degree of particle liberation of the most relevant sulphide ore minerals bornite, chalcopyrite, and chalcocite comminuted by traditional ball mill in comparison to the new VeRo Liberator®.

Fig. 5 Cumulative distribution curves of the degree of particle liberation of the most relevant sulphide ore minerals bornite, chalcopyrite, and chalcocite comminuted by traditional ball mill in comparison to the new VeRo Liberator®.

Fig. 5 Cumulative distribution curves of the degree of particle liberation of the most relevant sulphide ore minerals bornite, chalcopyrite, and chalcocite comminuted by traditional ball mill in comparison to the new VeRo Liberator®.

Up