Porous materials and machine learning promise cheap method to monitor microplastics

Optical analysis and machine learning techniques can now easily detect microplastics in marine and freshwater environments using inexpensive porous metal substrates. Details of the method, developed by researchers at Nagoya University and collaborators including Japan's National Institute for Materials Science, are published in the journal Nature. Nature Communications.

Detecting and identifying microplastics in water samples is essential for environmental monitoring, but is challenging due in part to the similarity of their structure to natural organic compounds derived from biofilms, algae and decaying organic matter. Existing detection methods typically require complex, time-consuming and costly separation techniques.

“Our new method can simultaneously separate and measure the abundance of six major types of microplastics: polystyrene, polyethylene, polymethyl methacrylate, polytetrafluoroethylene, nylon and polyethylene terephthalate,” said Dr. Olga Guselnikova of the National Institute for Materials Science (NIMS).

The system uses porous metal foam to capture microplastics from solution and detects them optically using a process called surface-enhanced Raman spectroscopy (SERS). “The SERS data acquired is highly complex, but contains identifiable patterns that can be interpreted using modern machine learning techniques,” explains Dr. Joel Henzie from NIMS.

To analyze the data, the research team created a neural network computer algorithm, called SpecATNet, that learns how to interpret patterns in the optical measurements and identifies targeted microplastics more quickly and accurately than traditional methods.

“Our method has great potential for monitoring microplastics in samples taken directly from the environment, without the need for pretreatment and without being affected by contaminants that can interfere with other methods,” says Professor Yusuke Yamauchi of Nagoya University.

The researchers hope that this innovation will contribute significantly to society in assessing the impact of microplastic pollution on public health and the health of all living organisms in marine and freshwater environments. By creating an inexpensive microplastic sensor and an open-source algorithm to interpret the data, they hope that rapid detection of microplastics will become possible even in resource-limited laboratories.

Currently, the materials needed for this new system offer a 90% to 95% cost reduction over commercially available alternatives. The research group plans to further reduce the cost of these sensors and make the method easily reproducible without the need for expensive equipment. Additionally, the researchers hope to expand the capabilities of the SpecATNet neural network to detect a wider range of microplastics and to accept different types of spectroscopic data in addition to SERS data.