Smartphone-integrated multi-color ratiometric fluorescence portable optical device based on deep learning for visual monitoring of Cu2+ and thiram

Dual emission ratiometric fluorescence probes have been widely used for naked visual individual detection of hazardous chemicals, but it is still limited to simultaneously detecting multiple targets in a complex system. In this work, a deep learning-assisted smartphone-integrated tricolor ratiometric fluorescence optical device has been designed for visual monitoring of Cu2+ and thiram. The tricolor sensing probes are consist of blue-emission carbon quantum dots (B-QDs), green-emission cadmium telluride quantum dots (G-QDs), and red-emission cadmium telluride quantum dots (R-QDs). The sensing system presents a three emission response to Cu2+ and thiram based on electron transfer effect, complexing effect, and inner filter effect (IFE), respectively. The reaction mechanisms were verified by density functional theory (DFT). Interestingly, the G-QDs and R-QDs’ fluorescence intensity are simultaneously quenched by Cu2+, whereas the fluorescence intensity of B-QDs remained unchanged and used as the internal reference, resulting in a distinct color shift from orange-red to blue with a detection limit (LOD) of 0.05 μM. Subsequently, the addition of thiram restored the G-QDs and R-QDs’ intensity while the fluorescence intensity of B-QDs was quenched accompanied by a distinguishable color transition from blue to red with LOD of 0.073 μM. Moreover, colorimetric detection of thiram is realized based on the variation of UV absorption intensity with LOD of 0.142 μM. Besides, deep learning-YOLO v3 algorithm-assisted smartphone-integrated tricolor portable optical device for in-situ monitoring of Cu2+ and thiram with ultralow LOD of 0.344 μM and 1.840 μM, respectively. In conclusion, the sensing system shows excellent stability, sensitivity, and specificity, which provides a great capacity for the convenient evaluation of Cu2+ concentration and thiram residue.