Modelling pyrite oxidation in refractory gold-bearing stockpiles to evaluate gold recovery by direct cyanide leaching

Abstract

The recovery of sulphuric refractory gold requires pre-treatment of the material for the liberation of gold particles from sulphide-bearing minerals (mainly pyrite). This pre-treatment is expensive and can increase significantly the total processing cost. However, for low-grade materials stockpiled for a long period of time, this cost can be reduced if the material naturally oxidised. When exposed to air and water, the pyrite in the stockpiles can be oxidised spontaneously. Over a prolonged period of time, this process may result in partial or complete oxidation of the contained pyrites, which may enable gold extraction by direct cyanide leaching and reduce the need for pre-treatment, hence increase the profitability of reclaiming the gold from the stockpiled material. The aim of this research is to investigate the possibility that the natural oxidation of pyrites in stockpiles of refractory gold-bearing materials may facilitate gold recovery without pre-treatment. To solve this problem, pyrite oxidation under stockpile conditions was studied and two models were developed to predict the level of pyrite oxidation in stockpiles. The first model describes the oxidation rate of pyrite grains under unsaturated conditions and/or circum-neutral to alkaline pH environments, in which a diffusion barrier develops on the fresh pyrite surface during the reaction. This reaction rate model was derived using the shrinking core model and it incorporates the effects of oxygen concentration, temperature and degree of water saturation on the reaction. The second model is a coupled multi-component numerical model that can simulate the pyrite oxidation in three-dimensional stockpiles together with related processes such as oxygen transport and heat transfer. This numerical model includes the reaction rate model as one of its components and the simulation incorporates the above-mentioned factors as well as other stockpile properties such as size distributions of rock fragments and pyrite grains. The outputs from the numerical model include oxygen concentration, temperature distribution, air velocity field, pyrite oxidation level and, more importantly, the oxidation profile of pyrite grains, which is an essential input for the estimation of gold recovery without pre-treatment. The application of these models was demonstrated in this research using a case study of the Kapit Flat stockpile on Lihir Island in Papua New Guinea. The simulation results were compared with those measured for samples taken from the stockpile and an acceptable estimation of the level of pyrite oxidation was obtained after calibrating the model. The models developed in this research have been demonstrated to provide a practical solution framework for estimating the level of pyrite oxidation in refractory gold-bearing stockpiles so that the recovery of gold without pre-treatment can be evaluated.