(177e) Evaluation of Mechanical and Thermal Properties of Recycled Plastics Using ?-Caprolactam Based Coupling Agents and Prediction of Properties through Machine Learning

Plastic is one of the most innovative inventions in human history, but it is increasing the amount of waste generated. In particular, plastic products discarded by end users are one of the major contributors to environmental pollution. Post-consumer recycled (PCR) plastic is consist of the immiscible polymers PP and PE. PCR plastics exhibit low interfacial adhesion and poor thermodynamic miscibility resulting in decreased mechanical properties. Through anionic ring-opening polymerization, it is possible to form reactive blends of ε-caprolactam with immiscible polymers such as PP and PE. This enables the enhancement of interfacial adhesion and improvement in mechanical properties. Moreover, through machine learning, it is possible to predict the properties of PCR plastics by training models on analysis results. In this study, we determined the optimal catalyst and initiator content for different contents of PP and PE. Mechanical properties (tensile strength, flexural strength, and impact strength), MFI, thermal properties (thermal stability and long-term life analysis), and morphological properties were evaluated. Furthermore, predictions of mechanical properties through machine learning were made using various algorithms such as linear regression, decision trees, etc. The predicted results were evaluated using metrics like the coefficient of determination, mean squared error, and so on. PCR plastics using ε-caprolactam coupling agents can expand their applications through improved mechanical properties.