Semi-scale stirred tank enzymatic bioleaching system for metal recovery from PCBs of end-of-life mobile phones: Process parameter optimization, predictive modelling, and economic assessment

Abstract

Biocatalysts like enzymes have proven to be faster and efficient in metal bioleaching from printed circuit boards (PCBs) than microbe-mediated bioleaching. However, studies on enzymatic metal bioleaching from PCBs are mainly confined to the shake-flask level. Therefore, it is essential to scale-up the process in a semi-scale stirred tank reactor (STR) for commercial applicability. In this study, enzymatic bioleaching of metals from mobile phone PCBs was performed in a semi-scale STR (working volume: 5 L) with optimization and predictive modelling employing response surface methodology (RSM) and machine learning (ML) tools, i.e., support vector machine (SVM) and artificial neural network (ANN), respectively. Process variables, i.e., mixing speed (MS) (200–500 rpm) and pulp density (PD) (1–10 g/L) were optimized and content of glucose oxidase enzyme (300 U/L) and Fe²⁺ ions (20 mM) was kept constant. Selective chemical precipitation was also performed for targeted metals recovery from bioleachate. Further, cost-benefit analysis (CBA) was conducted to assess the economic viability of the integrated technique. Although the 5 L reactor limits commercial-scale analysis, it lays the foundation for future scale-up and cost optimization. Maximum of 90% Cu, 95% Ni, 96% Pb, and 99% Zn were bioleached at optimal conditions, viz., MS: 395 rpm and PD: 5 g/L. ANN-based ML model (R² > 0.99) more accurately predicted enzymatic metal bioleaching than the SVM. Chemical precipitation recovered > 98% of targeted metals. CBA showing a revenue of 0.0423 USD/kg PCB recycling with a payback period of about four years highlights the economic viability of the integrated technique at the semi-scale level.