A dual-channel optical sensor for real-time microplastic classification via ratiometric fluorescence analysis

Microplastics (MPs) are widespread contaminants of water and food, yet available methods for polymer identification remain costly, time-consuming, and poorly suited to continuous monitoring. We describe SMILE, a compact, low-cost fluorescence platform developed as a proof of concept for the automated detection and classification of selected MPs in flowing water. MPs are stained with DANS, whose fluorescence emission depends on the surrounding polymer environment, and detected through two spectral channels to obtain a ratiometric descriptor for polymer discrimination. The system was evaluated under controlled laboratory conditions using single-polymer suspensions of PP, PE, PS, PMMA, PVC, and PET in Milli-Q water. Polymer standards were verified by Raman spectroscopy, with fluorescence microscopy providing complementary qualitative observations. Five-fold cross-validation yielded an overall classification accuracy of 77.8%, with high recall for PP, PE, and PS (93.2–98.6%) and lower recall for PMMA, PVC, and PET (24.5–59.6%). Apparent particle-count recovery ranged from 40% to 65%. Preliminary PS size-response calibration yielded diameter-equivalent LOD and LOQ estimates of 1.3 and 4.3 μm, respectively, as size-response indicators rather than concentration-based detection limits. Overall, these results demonstrate the feasibility of dual-channel ratiometric fluorescence for MP detection and polymer discrimination under controlled conditions.