Abstract
High-value recycling of WEEE plastics requires particle-level information on polymer identity, filler loading and restricted additives. This study developed an automated FTIR-XRF-AI workflow for characterising mixed WEEE plastic flakes. More than 14,000 flakes were analysed and 2,800 manually verified spectra were used to train an AI classification model for seven WEEE-relevant polymer classes, achieving a test F1 score of 0.993. A separate FTIR-AI model estimated talc and CaCO₃ contents in PP with MAEs of 2.19 wt.%, 1.48 wt.% and 3.08 wt.% for talc, CaCO₃ and total filler content. Integrating FTIR-AI outputs with XRF-based bromine screening and material-flow analysis revealed particle-level links between polymer identity, filler loading, Br concentration and recycling-route distribution. The automated system is commercially available, and the AI and MFA tools are released as open-source resources to support independent application. These resources establish a particle-resolved workflow for improving WEEE plastic sorting, recyclate quality control and compliance-oriented recycling decisions.

This publication has 16 references indexed in Scilit: