The Björkdal gold deposit is situated in the Skellefte mining district, northern Sweden. The ore is mined from a complex system of quartz veins ranging from a few cm to a meter in width. The mineral processing steps at Björkdal comprises a flotation circuit, which performance is dependent on the pH of the ore feed. A total of ten samples from five different ore zones from the underground development was investigated with the purpose of improving the understanding of factors that causes the pH value of rock samples to vary which is deleterious to the metal recovery. The samples were investigated with optical microscopy together with scanning electron microscopy (SEM) and automated mineralogy (QEMSCAN) to carry out the modal mineralogy. The analyses identified a total of 37 different minerals in the samples. Chlorite was identified with significantly higher values in parts of the investigated ore zones and especially in the sample with the highest pH value, suggesting chlorite-hosting shear zones that crosscut the sampled ore zones to be a contributing factor that could be affecting the elevated pH values, resulting in a decreased recovery of gold.

The Björkdal gold deposit is located in the eastern part of the Skellefte district, northern Sweden. Twenty thin sections from four production areas in the open pit and four drifts from the underground mine were analysed for mineral association and grain size distribution of gold. In addition, the texture of gold was investigated in order to find out how that affects the recovery of gold. The overall gold grain size distribution shows an interval from very fine-grained (2 μm) to coarse grained(856 μm) while the overall median size is 7 μm. Gold from the Quartz Mountain production area displays the smallest median size of 4 μm, whereas gold from the sampled drifts at 340m- and 385m- level has the largest median size of 14 μm. Gold at grain boundary is the dominant textural mode of gold from all sampled locations and varies from 62% to 92%. This is followed by intergrown which ranges between 8% and 29%. Of the sulfides, pyrite, chalcopyrite and pyrrhotite are the most common. Galena and was also present in the samples. Gold is significantly and positively correlated with tellurium (Appendix 10.1.1), and weakly positive correlated to silver and mercury. Gold show a close association to bismuth-tellurides in the samples. Apart from native gold, which is the dominant mineral phase of gold, two additional gold-bearing tellurium minerals were detected with SEM-EDS, a Au-Te-mineral and a Ag-Au-Te-mineral. One additional bismuth-telluride mineral aside from the most commonly occurring tsumoite (BiTe) was also detected with SEM, with a elemental composition of Bi-Te-S. Liberated gold in the tailings was optically identified in two thick sections, TB1-02feb-1 and TB1-07feb-1 (Fig. 32A and B), where the flotation circuit failed to float the free gold. One grain of gold was also identified intergrown with bismuth-telluride as an inclusion in silicate (Fig. 33), where the flotation properties of the larger silicate grain likely dominated in the flotation process. This thesis highlights the importance of further quantitative analysis utilizing SEM/QEMSCAN/MLA to retrieve representative mineralogical data to benefit the mineral processing of the ore from the active mine. Keywords: Björkdal gold deposit, gold, gold-telluride, SEM, mineral association, grain size,geometallurgy.

The Aurora zone is an ore zone which was recently discovered in the Björkdal gold mine, northern Sweden, and it has been the main focus of mining and exploration activities for the past few years (Pressacco et al., 2020). The purpose of this project is to determine how gold occurs in the Aurora zone. A three-day long campaign was therefore done at the processing plant at the Björkdal mine where 11 000 tonnes of ore from the drive Aurora 370/1650 E+W were processed. The issue regarding the ore from the Aurora zone is that it has a lower recovery compared to the rest of the mine.  Six chip samples, 12 samples from the ingoing plant feed and two tailing samples were analyzed using optical microscopy, scanning electron microscopy (SEM), and automated mineralogy (QEMSCAN). This was done to determine the mineralogy, grain size distribution, mineral associations, textures and modal mineralogy which all are factors that could influence the metal recovery at the processing plant. The samples were prepared using the cold mounting method and epoxy mounts were created. After cutting, grinding, and polishing, the sections were ready to be examined. Optical microscopy was performed using a Nikon ECLIPSE E600 POL microscope. Ten epoxy mounts were carbon coated and automated mineralogy was performed on nine of them in a ZEISS Sigma 300 VP using a recipe (analysis mode) for “bright phase search”. Manual point-ID analysis was done using a ZEISS MERLIN SEM. Fifty gold grains were identified in this study, 48 of them in the chip samples and two of them in the ingoing-feed samples. 64% of them were associated with silicates, 22% were quartz associated, 12% were associated with bismuth minerals and 2% of them were associated with sulfides. The grain size distribution has a range between 0.7 and 19 μm and the median grain size is 4.8 μm. The gold grains identified from the Aurora zone have a significantly smaller median grain size than gold from other parts of the mine. The majority of the gold grains identified in this study, have a very fine grain size, are mainly associated with silicates and most prominently occur as inclusions. Gold that occurs in this way is typically difficult to recover in the processing plant and it seems like this is the main reason for the lower gold recovery from the Aurora ore zone.  No gold was found in the tailings, suggesting that the mineral process is performing well although no thorough conclusion can be made in regards of the processing. The lack of data for the different sample types in this project is an issue. Gold from the tailings must be identified and examined to draw any clear conclusions regarding the processing. For future work, it is therefore recommended to analyze more tailing samples and to implement hydroseparation at the sample preparation stage, to separate the heavier gold particles from lighter minerals. Then more gold will most likely get detected in the tailing samples.