Potential prediction and coupling relationship revealing for recovery of platinum group metals from spent auto-exhaust catalysts based on machine learning

As hazardous waste, the massive generation of spent auto-exhaust catalysts (SACs) puts enormous pressure on environmental management, but provides a rare opportunity for platinum group metals (PGMs) recycling. In this study, machine learning (ML) method was firstly applied to accurately predict regional SACs generation in China for 2025–2050 under five shared socio-economic pathways (SSPs) scenarios, based on which economic and carbon emission reduction potential of PGMs recycling were estimated. Population-GDP-GDPII-GDPIII and Random Forest were determined as key variables and the predictive model. Results indicate that SACs will reach 28.15 million sets (1.7 times that of 2020) and PGMs have economic potential of $890 million by 2050 (SSP1). Furthermore, based on environmental impact assessment, the capture enrichment-electrodeposition purification process is proposed as the best low-carbon recycling solution for SACs. And the integrated recovery process based on copper capture can realize 1.51 million tons of carbon emission reduction in China in 2050 (SSP1). This study can provide decision-making guidance for PGMs recovery and environmental management, as well as technical references for SACs recovery program selection.