Applicability of multivariate data analysis to improve the sorting degree of recycled polyethylene

A key target of the Circular Plastic Alliance Declaration is to include 10 million tons of recycled plastic per year into new plastic products in Europe by 2025. To meet this objective mechanical recycling has to overcome two main obstacles. First, new reliable, and cost-efficient technologies with high material separation accuracy are required. Secondly, the quality and performance of recyclates must be improved significantly. Polyethylene (PE) is a globally dominant polymer as a result of its wide range of variations in molecular structure and morphology.
State-of-the-art near-infrared sorting systems easily identify base materials, but they can hardly ac-count for specific characteristics in the molecular structure of PE, hence the applicability of such materials for new plastic products are limited. Applying multivariate data analysis (MVDA) on FTIR and Raman spectroscopy data of 13 PE grades allowed the prediction of density. The results showed a good agreement with the measured values, indicating that processing relevant information can be extracted from FTIR and Raman data. Improved sorting would enhance the usage of post-consumed polymers for new products and assist to achieve the yearly 10 million tons of recycled plastics.