The effect of particle size distribution on the extraction of gold from printed circuit boards using ammonium thiosulphate

Abstract

The rapid advances in the fields of electronics and telecommunications have resulted in the growing generation of electronic waste (e-waste) worldwide. This waste contains hazardous heavy metals that can endanger the ecosystem. As such, environment-friendly recycling techniques are envisaged as mitigating measures. The use of hydrometallurgy to recover base and precious metals from waste computers and mobile phones has seen unprecedented attention in the past decade. Non-cyanide leaching agents such as ammonium thiosulphate show potential in the recovery of gold from waste printed circuit boards (PCBs), the central part of electronic devices. This research project aimed to investigate the effects of particle size distribution (PSD) on the shrinking core model (SCM) for the determination of the control regime of the ammonium thiosulphate leaching of gold from waste mobile phone PCBs. The Gates-Gaudin-Schuhmann (GGS) and Rosin-Rammler (RR) models were used to describe the PSDs used. It was determined that the GGS model was a good fit for large particles (658 μm mean size) whereas the RR model was a good fit for smaller particles (503 to 433 μm mean sizes). These PSD models were used to estimate the central values (median and mean sizes) and covariances of the PSDs. The leaching experiments were carried out in a batch setup. The PSD was varied from a mean size and median size of 658 and 592 μm to 433 and 354 μm, respectively. The pulp density was subjected to three variations: 40, 80 and 120 g/L. The other leaching conditions were fixed at levels prescribed by the existing literature. The PCBs were found to contain 172 g Au/ton-PCB based on the aqua regia characterisation. In thiosulphate leaching, almost complete gold extraction (97% in 2 hours) was obtained at 40g/L pulp density, 0.1 M thiosulphate, 0.04 M copper sulphate, pH 9.5, 30°C and 350 rpm mixing rate. These results were obtained with PCB particles with median and mean sizes of 433 and 354 μm, respectively. Reducing the pulp density from 120 to 40 g/L improved the gold dissolution from averages of 35 to 80%, irrespective of the PSD used. The variation of PSD had little to no significant impact on gold extraction at high pulp densities but did influence gold extraction at low pulp density. iii The PSD covariances obtained in this study (CV < 0.3) indicated that particle size distribution could safely be excluded from the shrinking core model in determining the rate-limiting step of the leaching process. Based on the SCM, the leaching process was found to be controlled by chemical reaction and ash layer diffusion at 40 g/L pulp density. In contract, higher pulp densities hindered the shrinking of the unreacted core and corresponding gold conversion and limited the application of the shrinking-core model to describe, predict and provide insight into the rate-limiting mechanism of the leaching process. Possible reasons for the SCM not fitting the leaching data, for larger particles at low pulp density and all particle sizes at higher pulp densities, were discussed. This research will provide insights into the overlooked particle size distribution aspect of the extraction process; assist in devising optimised mathematical models for the design of new leaching reactors; retrofitting of existing ones to include PCBs in the existing process streams and optimise the size reduction step for improved dissolution efficiency.