AI tool enables recycled plastic tracking for regulatory compliance

Researchers at the University at Buffalo (UB), US, have developed a new method to differentiate recycled from virgin plastic in packaging formats such as bottles, shopping bags, and yogurt cups.

The method combines scientific testing and AI to help companies, regulatory agencies, and organizations monitor plastic recycling.

“Our goal is to create a quick and reliable tool that can be used to verify recycled material content, as well as enforce recycling regulations,” says corresponding author Amit Goyal, professor at the UB Department of Chemical and Biological Engineering.

He adds that the tool aims to “improve the quality of plastic products and help reduce plastic waste, which will support a more circular economy where plastic pollution and its associated health and environmental risks are reduced.”

The study was published in Communications Engineering.

Spotting differences

The researchers explain that recycling plastic involves melting, cleaning, and remolding. The chemical makeup of the resulting material is “very similar” to that of virgin plastic.

Nonetheless, there are “subtle” differences, such as microscopic impurities and broken polymer chains in recycled plastic.

Amit Goyal, PhD, professor at the UB Department of Chemical and Biological Engineering.The research team outlines four sensing techniques to determine these differences: triboelectric testing, dielectric/impedance spectroscopy, capacitance analysis, and mid-infrared spectroscopy.

Triboelectric testing measures how plastic gains and holds an electrical charge when surfaces touch — static electricity. Recycled plastics often hold a charge longer due to structural defects from repeated processing, the researchers point out.

Dielectric/impedance spectroscopy is a technique that applies an electric field to measure how a plastic stores and loses energy. Recycled plastics show lower energy storage and higher energy loss.

Meanwhile, capacitance analysis tracks how quickly plastic charges and discharges in a circuit. Differences in timing are said to reveal changes in electrical properties resulting from recycling plastic.

Mid-infrared spectroscopy involves shining a light on the chemical structure of plastics to show the fragmented polymer chains found in recycled plastic.

Machine learning predictions

As part of the study, the researchers tested their method by examining new and recycled PET. They analyzed and combined the test data using machine learning.

The machine learning model studied the test results and learned to recognize patterns in the data that correlate with recycled plastic percentages, with accuracy higher than 97% when determining the percentage of recycled content in PET samples that contained 0–50% recycled material.

Goyal says: “This is an ideal example of combining cutting-edge innovation in science and engineering with AI for social good, and to potentially realize significant societal impact.”

He adds that in the future, the team will combine the method’s different sensing techniques and machine learning model into a portable device.

“By fabricating such a device, we hope to enable widespread, real-time monitoring of recycled plastics in commercial products.”

Goyal argues that his team’s work will grow in significance as more countries adopt regulations requiring the use of recycled materials.

The European Commission recently introduced new measures to boost demand for recycled plastics in the EU. Meanwhile, rising oil and virgin plastic prices are expected to push packagers toward greater use of recycled materials.

A consumer survey by Amcor revealed that the majority of EU shoppers support the use of recycled material in packaging.