A Raman spectral database of naturally aged plastics: A proof-of-concept study for waste plastic sorting

A Raman spectral database of naturally aged plastics for years is developed aiming to support efficient and proper plastic waste management based on Raman sorting technology. The database is designed as a website and contains the Raman spectral data for 146 plastic waste samples degraded for years in natural environmental conditions. Plastic samples aged for years in indoor conditions are also included along with a few undegraded, common plastics. The natural environmental degradation occurring for years of exposure to sun radiation, terrestrial and aquatic factors, or temperature variations is poorly understood in terms of spectral changes of representative plastics. The samples of choice are represented by the widespread types of plastics such as polyethylene terephthalate, high-density polyethylene, low-density polyethylene, polyvinyl chloride, polypropylene, and polystyrene. Each plastic sample is described individually on the database's website, and its Raman spectrum is available as a txt file for both educational and analytical purposes. Moreover, the database was validated by external validation of the machine-learning classifiers: k-nearest neighbor, support vector machine, logistic regression, random forest, and Naive Bayes resulting in a classification accuracy between 0.902 and 0.973.