Characterization of feldspar texture and liberation by automated image analysis

An aplite ore from Guarda in northern Portugal, containing albite, microcline and quartz and muscovite gangue was studied to identify conditions for extracting the feldspar minerals. The large amounts of quantitative mineralogical data needed to evaluate grinding behaviour were provided by automated image analysis. The images were acquired on polished surfaces by measuring back-scattered electron signals produced with a microprobe. Minerals were identified from X-ray spectra and discriminated by grey level and element-characteristic X-rays. The quantities and size distributions of the minerals were determined by measuring thin sections of unbroken ore to predict the grind required to liberate the feldspars. The ore was then ground to -1.7 mm, polished sections were prepared from sieved fractions and the quantities of mineral species and apparent total liberation were determined by analysis. The predicted and measured liberations were compared and the grinding behaviour of the ore interpreted. In an appended correction, the correct version of Equation 1 is represented and Tables 3 and 6 are revised to consider all size fractions, including those coarser than 300 micrometres, in calculating the total liberation from King's fractional liberation