MLA-based partition curves for magnetic separation

• MLA and magnetic susceptibility data are combined.
• A large section transformation is applied to the data.
• Bootstrap resampling with replacement is applied to the data.
• Partition curves for magnetic separation are calculated.
• Parameters are studied as a function of particle size.

Mineral liberation analysis (MLA) provides detailed information on the composition of single particles as well as particle populations exposed on the cross-sectioned surface of a grain mount. This information can be used to evaluate process efficiency related to the liberation distribution of valuables or to the distribution of a feature used for separation of valuable from barren particles. When separation processes are studied, the separation feature has to be added to the particle population data. In case of density separation (particle size and particle density) this is already been done by MLA software. For magnetic separation, however, the MLA data has to be combined with mineral susceptibilities using additional software.This contribution describes the calculation of partition curves for magnetic separation based on the liberation analysis of the feed and the products. Magnetic susceptibility data from measurements of pure minerals and from literature are combined with particle composition data from liberation analysis. Using sieved fractions in magnetic separation experiments the evolution of cut susceptibility and separation efficiency with particle size is studied. The effects of a stereological correction and of the width of susceptibility classes as well as of the mineral susceptibilities used in calculation are discussed.