As global plastic waste continues to rise, accurately identifying and quantifying recycled content in plastic products is critical for developing a circular economy. At present, there is no method that can accurately determine the percentage of recycled plastic content in a plastic product. Here, we demonstrate a multi-modal, non-destructive sensing technique to determine the percentage of recycled plastic in plastic products. We have developed a multi-modal, multi-physics approach that integrates triboelectric properties, dielectric/impedance spectroscopy, capacitance measurements, mid-infrared spectroscopy, combined with machine learning and artificial intelligence to quantify the recycled content in plastics. Experimental results reveal that increasing recycled content leads to enhanced charge retention, reduced permittivity, and increased dielectric loss, consistent with polymer chain scission and defect-induced polarization. The machine learning model trained on the multi-modal dataset achieves we achieved over 97% classification accuracy across PET samples ranging from 0% to 50% recycled content, which is the expected regime of required recycled content in plastic products. This method offers a solution for control and regulatory compliance for recycled plastics. Yaoli Zhao and colleagues have developed a non-destructive method to determine the percentage of recycled plastic content in a plastic product. By combining multi-modal sensing with machine learning, this approach enables reliable verification of recycled content, supporting regulatory compliance, quality control, and circular plastics manufacturing.
Zhao et al. (Mon,) studied this question.