Kyung Hee University Develops AI-Powered Sensor to Detect Microplastics

Researchers at Kyung Hee University have developed a novel sensor technology that can detect micro and nanoplastics in water with high sensitivity. The innovation, led by Professor Yoo Jeong-mok from the Department of Convergence Bio-Auto & Materials Engineering and Professor Jeong Dae-hyun from the Department of Smart Farm Science, utilizes surface-enhanced Raman scattering (SERS) to achieve this breakthrough.

The core of the technology lies in its unique design: gold nanoparticles are embedded in a regenerative cellulose hydrogel. When the hydrogel comes into contact with water, it re-expands, actively rearranging the gold nanoparticles around the nanoplastics. This self-assembly process creates highly dense "hotspots" that significantly amplify the Raman signal, enabling the detection of even low concentrations of nanoplastics.

This advancement addresses a critical challenge in environmental monitoring. Micro and nanoplastics are pervasive in various environments, including water bodies, soil, and food, raising concerns about their potential impact on ecosystems and human health. However, their small size and low concentrations have made accurate detection difficult.

“

This research is significant because it utilized regenerative cellulose hydrogel not just as a support but as a functional material that actively changes the structure of gold nanoparticles.

To demonstrate the sensor's practical applicability, the research team tested its performance using polystyrene nanoplastics of varying sizes (80, 150, and 850 nm). They confirmed that the hydrogel's re-expansion and the subsequent rearrangement of gold nanoparticles effectively enabled SERS detection of these different-sized particles. The sensor also showed potential for use in complex samples, including reservoir water, tap water, milk, and artificial seawater.

Further enhancing its capabilities, the team integrated an artificial intelligence (AI) system. This AI, trained on multi-label classification, can automatically identify individual pollutants, such as micro/nanoplastics, organic dyes, and pesticides, even when they are present together in a sample. The AI's decision-making process was also verified against the unique Raman fingerprints of the substances, ensuring accuracy and reliability. The research was recently featured in Nanowerk, a global nanotechnology publication, highlighting its international recognition.

“

We plan to expand this technology into a sensor system for rapid on-site analysis of various environmental pollutants.