New Optical Analysis and Machine Learning Technique Enhances Microplastic Detection

New cost-effective method for detecting microplastics in water using optical analysis and machine learning, promising significant advancements in environmental monitoring and waste analysis.

Researchers from Nagoya University, in collaboration with the National Institute for Materials Science in Japan and other institutions, have developed a groundbreaking method to detect microplastics in marine and freshwater environments. Published in Nature Communications, their innovative technique leverages optical analysis and machine learning to identify microplastic pollutants using cost-effective porous metal substrates.

Detecting and identifying microplastics in water is crucial for environmental monitoring. However, it has been a challenging task due to the structural similarities between microplastics and natural organic compounds such as biofilms, algae, and decaying organic matter. Traditional detection methods are not only time-consuming, but also require expensive and complex separation techniques.

The new method developed by Dr. Olga Guselnikova and her team at the National Institute for Materials Science can simultaneously separate and measure the abundance of six key types of microplastics, including polystyrene, polyethylene, polymethylmethacrylate, polytetrafluoroethylene, nylon, and polyethylene terephthalate. It utilizes a porous metal foam that captures microplastics from samples and employs surface-enhanced Raman spectroscopy (SERS) for detection.

An unknown liquid sample containing various microplastics (left) is passed over the porous metal surface. Raman spectroscopy is then performed on the metal foam surface (right), and the scattered light is analyzed with a machine learning algorithm trained to accurately identify microplastics in complex mixtures. Credit: Olga Guselnikova

Dr. Joel Henzie, also from NIMS, explains that the SERS data is intricate but contains patterns that can be deciphered using modern machine learning techniques. To tackle this complexity, the team developed a neural network algorithm named SpecATNet, which interprets the optical measurements to identify target microplastics efficiently and accurately.

“Our procedure holds immense potential for monitoring microplastics in samples obtained directly from the environment, with no pretreatment required, while being unaffected by possible contaminants that could interfere with other methods” stated Professor Yusuke Yamauchi of Nagoya University. The simplicity and cost-effectiveness of this method make it particularly promising for rapid on-site environmental assessments.

The approach not only promises significant cost reductions—by 90 to 95% compared to existing commercial technologies—but also enhances the detection capabilities without the need for advanced laboratory facilities. This could be particularly beneficial for resource-limited labs.

The research team aims to further reduce costs and simplify the replication process of these sensors. They also plan to expand the capabilities of the SpecATNet neural network to include a broader range of microplastics and adapt it to process various types of spectroscopic data beyond SERS.

By facilitating easier and more effective detection of microplastics, the researchers hope to contribute significantly to understanding and mitigating the impacts of microplastic pollution on public health and aquatic ecosystems.

You can read their full paper “” now.

Staff Writer

Our in-house science writing team has prepared this content specifically for Lab Horizons

Leave a Reply

Your email address will not be published. Required fields are marked *