Machine Learning-Assisted Microwave-Microfluidic Platform for Microplastic Detection

Area of Science:

• Environmental Science
• Analytical Chemistry
• Sensor Technology

Background:

• Microplastics (MPs) pose a significant environmental and potential human health risk.
• Current detection methods for MPs are often complex, limited in particle size detection, and lack continuous monitoring.
• A need exists for advanced, efficient MP detection technologies.

Purpose of the Study:

• To develop and validate a novel microwave-microfluidic platform for the flow-through detection and identification of microplastics.
• To assess the system's capability for single-particle detection and material classification.
• To establish a foundation for real-time MP analysis in environmental samples.

Main Methods:

• Utilized a microwave-microfluidic system with disposable microfluidic chips and split-ring resonator (SRR) sensors.
• Evaluated single-particle detection for particle sizes ranging from 20-300 μm.
• Employed a k-nearest neighbors (k-NN) model for material classification (polyethylene, polystyrene, glass, brine shrimp eggs) in the 165-300 μm range.
• Assessed system robustness with various aqueous carrier liquids using baseline referencing.

Main Results:

• Successfully demonstrated single-particle detection across multiple size ranges (20-300 μm).
• Achieved material classification for specific particle types within the 165-300 μm range using the k-NN model.
• Confirmed the platform's robustness and adaptability to different carrier liquids.

Conclusions:

• The developed microwave-microfluidic platform is feasible for continuous, in-flow, single-particle monitoring and material classification.
• This technology provides a foundation for analyzing microplastics in complex environmental matrices.
• The study highlights the potential of microwave sensing for advanced environmental monitoring applications.