Investigating the effect of grinding method on ore beneficiation behaviour

Commodity projection studies anticipate an exponential increase in copper demand and decreasein mill grade of deposits currently exploited globally. Most of this demand is driven by theindispensable use of copper in electrification and currently, the growing demand in the contextof the green transition. To meet this challenge while supplying society with copper, producerssuch as Boliden AB have plans for exploration and expansion projects to increase their reservesand in turn increase production through efficient measures. As grinding is a bottleneck to mineralprocessing and the most energy intensive, more efficient comminution systems are beingconsidered for environmental sustainability. These are required to have lower energyconsumption while achieving the required liberation or better for subsequent processing byflotation. Consequently, the literature review shows there is a research query on whether more can beexpected from comminution to optimise for downstream processing. With most publishedresearch being on particles with a P80 less than 100 μm, there exists a gap in investigating theeffect of grinding on flotation for coarser size fractions which this degree project focuses on.Importantly, it includes evaluating chalcopyrite liberation and flotation performance when usinga Novel Comminution Device. Three grinding mills were considered, a vertical roller mill (VRM),a Novel Comminution Device (NCD), with a rod mill as the reference mill. The scope of this studywas an analysis of the mill products’ particle size distribution, chalcopyrite liberation andinvestigation of the flotation performance of the products from the three mills. This was done ona Cu-Au-Ag ore from the Boliden Aitik mine in northern Sweden. Particle size distribution (PSD) analysis revealed that at comparable P80 values, the VRMproduced particles had broader PSD curves than the rod mill, while the NCD provided a steeperproduct size distribution than its reference rod mill products. The chalcopyrite liberation analysisusing QEMSCAN automated mineralogy showed that for samples split into three size fractions, -45 μm, +45/-90 μm and +90 μm, the VRM and rod mill products had their highest proportion offree to liberated particles within the intermediate size fraction while for the NCD sample this waswithin the fine fraction. Overall, the NCD product had the highest free to liberated particles acrossall fractions. From shape analysis based on scanning electron microscopy (SEM) images, it wasobserved that the NCD produced particles with the higher elongation than the VRM and rod mill. The flotation performance was assessed with respect to the mass pull, kinetics, grade andrecovery in the concentrate and separation efficiency. The rougher batch flotation tests indicatedthat flotation of the VRM ground samples had higher mass pull, Cu recovery, faster kinetics andlower Cu grade than the samples from the reference rod mill. This was also a similar outcome forthe NCD with the exception of higher mass pull at an increased collector dosage in the rod millcase. For all the mills, the general trend showed that an increase in P80 was associated withreduced mass pull. Overall, at similar flotation conditions, despite having one of the coarsest P80s,the NCD product showed the fastest Cu recovery kinetics, highest total recovery and selectivityindex. The rod mill reference test samples gave the highest grade but the lowest recovery. The findings illustrate that there is a significant difference in PSD broadness between productsfrom different mills at comparable P80. The differences in flotation performance between thethree mill products were mainly attributed to differences in liberation, potential differences insurface activation between wet and dry grinding methods and to some extent, steepness of thePSD curve. Based on the high recovery yet low grade of the NCD ground product, this is best suitedfor the rougher flotation stage. Future studies should therefore include cleaner stage flotation andoptimisation tests for the NCD with respect to reagent dosages and hydrodynamic conditions