High-throughput Raman platform for microplastics detection on filtration membranes

Abstract

Plastic materials are ubiquitous across modern industries, leading to widespread microplastic contamination in aquatic systems. The diversity of plastic usage yields microplastics of diverse sizes, shapes, and compositions, posing significant analytical challenges. Current techniques cannot provide high-throughput analysis for the chemical and morphological distribution of numerous microplastics. In this study, we propose an integrated line-scan Raman imaging platform for rapid, high-throughput analysis of microplastics on 47-mm-diameter filters. By combining mosaic scanning Raman spectroscopy and optical microscopy, the proposed system enables the efficient acquisition of high-resolution spectral and spatial data. Although Raman measurements on uneven filter surfaces present significant challenges, the proposed platform detects microplastics >= 10 mu m in size on structurally complex filter surfaces. In addition, advanced deep learning segmentation algorithms were integrated to automatically classify particles, quantify microplastic types, and analyze size distributions with high statistical reliability. The proposed platform achieves complete filter analysis, including measurements and processing, within 1 h, indicating a substantial advancement over conventional approaches. The proposed scalable, artificial intelligence-assisted system provides a robust foundation for industrial and environmental applications requiring real-time, quantitative microplastic monitoring.