Plastics come in many different varieties based on their chemical composition and structure, and these varieties can be difficult to distinguish and separate to the degree necessary for effective recycling. Because plastic typically must be at least 96% pure by polymer type to be recycled by current industrial standards, available methods are not always accurate or efficient enough to keep up with the immense amount of plastic waste that must be sorted and processed. Researchers at Aarhus University, in collaboration with plastic recycling companies Vestforbrænding, Dansk Affaldsminimering Aps and PLASTIX, have now developed a new system using hyperspectral imaging and machine learning to more quickly and accurately identify and separate a dozen different varieties of plastic as they come down the conveyor belt.
Current separation methods involve near-infrared (NIR) technology or density tests such as flotation separation, which can separate some plastic fractions such as PE, PP and PET but are not as accurate or versatile as the hyperspectral system. The new system includes a hyperspectral camera and spectrograph with a spectral resolution of 8.3 nm in the range of 955 to 1700 nm, which are installed inline on a conveyor belt system through which the plastic samples are transported. The researchers established an unsupervised machine learning model to distinguish between different plastic types based on the spectral information that is recorded and transferred to a computer as the plastics are scanned on the conveyor belt. The system can accurately identify a total of 12 different plastic varieties: PE, PP, PET, PS, PVC, PVDF, POM, PEEK, ABS, PMMA, PC and PA12. This work was recently published in Vibrational Spectroscopy.
“With this technology, we can now see the difference between all types of consumer plastics and several high-performance plastics. We can even see the difference between plastics that consist of the same chemical building blocks, but which are structured slightly differently,” said Mogens Hinge, study co-author and head of Aarhus University’s “Re-Plast” project. “We use a hyperspectral camera in the infrared area, and machine learning to analyse and categorise the type of plastic directly on the conveyor belt The plastic can then be separated into different types. It’s a breakthrough that will have a huge impact on plastics separation.”
The system has already been tested in a pilot program and is planned to be implemented at PLASTIX and Dansk Affaldsminimering Aps facilities in the spring of 2022. The collaborators also hope to further improve the system to distinguish between even more polymer types and additives in the future. This research was part of “Re-Plast,” a project that aims to improve the quality and purity of recycled plastic.