Identification of virgin and recycled polypropylene based on FTIR spectroscopy combined with machine learning

The circular economy has renewed attention to recycling of polypropylene (PP), including recycled PP used in food packaging, with a focus on reducing production costs and achieving social development goals. However, recycled PP matrices undergo degradation during the recycling process. When utilised in food packaging applications, these materials may leach higher concentrations of potentially hazardous chemicals into the packaged food, posing potential risks to human health. This study proposes a novel identification method that combines Fourier Transform Infra-red Spectroscopy (FTIR) with machine learning algorithms to distinguish between virgin and recycled PP. The results demonstrate that processing the original FTIR data with first-order derivative, followed by Principal Component Analysis (PCA) and Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA) visualisation, achieves optimal discrimination. Machine learning models built on this data also yield highly satisfactory identification performance. Specifically, the Random Forest (RF) model achieved an accuracy of 98.6% on the training set and 86.1% on the test set, and an Area Under the Curve (AUC) value of 0.963. It successfully provided reliable discrimination between virgin and recycled PP.