Multi-mechanism-driven dual-mode array based on a single PFC-1/QD probe enables AI-assisted on-site identification of biogenic amines and real-time food freshness monitoring.

Biogenic amines (BAs) serve as critical indicators of food spoilage, necessitating portable and real-time monitoring approaches to ensure food safety. To this end, we developed a P-QDot-based fluorescence/colorimetric (FL/CL) dual-mode sensor array integrated with deep learning and smartphone imaging for the classification and quantification of five BAs and the assessment of food freshness. The P-QDot probe was prepared by physically mixing the hydrogen-bonded organic framework PFC-1 with glutathione-capped quantum dots (QD) in aqueous solution. The fluorescence response was governed by a synergistic effect involving aggregation-induced emission (AIE), photoinduced electron transfer (PET), and the inner filter effect (IFE), which were activated by hydrogen bonding between the probe and BAs. Concurrently, a new UV absorption band was generated by P-QDot under the alkaline conditions induced by the BAs, leading to the colorimetric response. Leveraging this dual-mode response mechanism, the P-QDot-based FL/CL sensing array facilitated rapid (<30 s), sensitive, and simultaneous qualitative discrimination and quantitative detection of BAs. Furthermore, a smartphone platform integrated with the YOLOv12 algorithm was established, enabling on-site BAs classification. The practical application of this system was demonstrated by real-time monitoring of shrimp spoilage during storage at 4 °C and 25 °C, where visually recognizable fluorescence color changes and colorimetric signals accurately reflected the degree of spoilage. Additionally, a logic gate device was engineered to directly translate the complex sensor signals into three intuitive freshness levels (fresh, acceptable freshness, and spoiled). This work presented a powerful strategy for food freshness assessment, successfully merging intricate material design with portable intelligence.