On the Recovery of Associated Minerals in the Flotation Process Based on Size-Liberation Data

Flotation is a separation process in which mineral recovery depends on the particle surface properties. The selective collection of particles by air bubbles requires hydrophobic mineral surfaces. Hydrophobicity can be selectively induced on a valuable mineral surface by means of reagents. For this purpose, it is required that mineral grains in the ore become physically liberated or exposed in the particle surface after grinding. The grinding process generates three types of particles according their exposed surface in the floatable mineral: locked particles (0-10%), associated particles (10-90%) and liberated particles (90-100%). Locked particles have low floatability, and they typically show low recovery (<20%), while liberated particles have a high floatability and can be easily recovered (98-100%) at different particle sizes (10-150 µm). The main challenge in the flotation process is the recovery of associated particles. These particles typically have recoveries of 20-80%, but strongly depend on the particle size, particle surface liberation and collector dosage. Recent studies on particle floatability by size-liberation have significantly contributed to improving the characterization of flotation processes. In this work, a study of flotation recovery of associated particles of 10-90% liberation is presented for particle sizes (10-150 µm) and collector dosages (5-40 g/ton). In this work historical data on galena recovery based on particle cross-section composition of galena was used, assuming they represent an acceptable estimate of particle surface composition, as shown by some previous researchers. The results showed that galena recovery can be well represented by a model consisting of the interaction of a linear function and a Gamma distribution function to account for the liberation and the particle size effects. Consequently, an alternative approach for modelling the recovery of associated mineral (10-90% liberation), with a reduced number of parameters, was developed and validated.