This paper examines the longer-term predictive capacity of the GAMSPATH computer program (developed by the Alberta Research Council) for a waste rock (sample MS) obtained from the Kaltim Prima Coal (KPC) mine, Kalimantan that contained significant framboidal pyrite._x000D_
The model was refined by comparison with the net acid generation (NAG) and column leach tests to provide a basis for calibration of the model for both short term and longer-term reaction trends. The oxidation of pyrite during the kinetic NAG test using H2O2 provides oxygen directly at the mineral surface, generating a measured oxidation rate for pyrite of 1.16_x000D_
10-6 mol m-2 s-1, whereas the oxygen diffusion limited rate in the leach column air equilibrated reaction was of the order 10-8 mol m-2 s-1 as calculated from the GAMSPATH thermodynamic database. Results indicated that the oxygen derived from hydrogen peroxide had to be included as a reactant (eg a reacting mineral) in the simulation to account for the direct availability of oxygen at the sulfide mineral surfaces._x000D_
Optical investigations identified the proportion, particle size, and surface area, of framboidal and euhedral pyrite in sample MS. The reactive surface areas of the two morphologies were very different (eg ratio: 5:1). To manage these two different pyrite morphologies the GAMSPATH thermodynamic database was updated to include framboidal pyrite, with similar thermodynamic properties and oxidation rate to euhedral pyrite but with a particle size based on the size of the individual internal framboid microcrysts. This modification, with framboidal pyrite reaction based on the size of its microcrysts produced results that correlated well with the kinetic NAG and column leach test results when the measured (kinetic NAG test rate) and calculated (column test rate) pyrite oxidation rates were employed. This provided some confidence that the model could be extrapolated to indicate longer-term geochemical reaction trends.