Aggregation-induced fluorescence switching of melamine-tailored carbon dots: Dual-recognition strategy coupled with deep learning for sensitive salbutamol assay

Sensors based on multiple fluorescence materials have been widely applied in detection because of reliability and reproducibility. High-performance on-site sensing strategies are vital to public safety and human health. Here, by modifying the carbon dots (CDs) with melamine (MA) molecules, we developed a novel dual-recognition fluorescence sensor for efficient and sensitive detection of salbutamol. Upon increasing salbutamol concentration, the fluorescence intensity of blue emission under 365 nm excitation decreased, while the yellow emission under 460 nm kept enhancing. The quench of blue emission leads to the yellow fluorescence enhancement by Fluorescence resonance energy transfer (FRET) from monodisperse CDs to clustered CDs. Both the two emission-intensity showed linear relationship with salbutamol concentration within a certain concentration range. Dual detection of fluorescence signals under 365 nm and 460 nm, with detection limits (3δ/K) as low as 7.12 and 16.67 nM, is possible. Machine learning is then employed to analyze its detection performance. Therefore, the quantitative detection of salbutamol can be achieved. Moreover, the dual-recognition assay endows salbutamol detection built-in self-calibration to ensure high assay sensitivity and precision. We demonstrated the application in the fluorescence tracking of salbutamol in real samples with satisfactory results. The study would anticipate this dual-recognition fluorescence sensing system would inspire fast high-precision sensor design for environmental, safety and health monitoring.