Multimodal Detection of Microplastics in Water Based on the Integration of Triboelectric Nanogenerators and Machine Learning

Microplastics (MPs) can enter the human body via the food chain and induce irreversible health risks. Based on triboelectric nanogenerator (TENG) technology, this study developed a droplet microfluidic system for the real‐time detection of MPs in aqueous environments. For the classification of MP types, crystalline polymers exhibit a much greater reduction in open‐circuit voltage ( V OC ) than amorphous polymers, attributed to their stronger electron‐trapping capability. In terms of abundance detection, when the volume fractions of MPs were set at 0.25%, 0.50%, and 0.75%, respectively, the V OC amplitude of droplets containing 5 µm polyethylene (PE) decreased by 24.73%, 43.45%, and 56.25% compared with the deionized water control group. Regarding morphology detection, spherical MPs corresponded to the lowest average V OC amplitude of droplets, whereas flake MPs yielded the highest amplitude. The random forest was introduced to identify MPs with different concentrations and morphologies, achieving a classification accuracy of 100%. In practical applications, the system achieved an identification accuracy of up to 99.9% for flake PE MPs in rainwater. This study verifies that the droplet microfluidic system integrated with TENG and machine learning has substantial feasibility for the multi‐parameter real‐time detection of MPs in water. It provides a novel technical tool for environmental pollution monitoring.