An electrostatic potential modulation strategy for fluorescent probe coupled with deep learning enabling ultrafast and visual detection of meat freshness

Abstract

Biogenic amines (BAs), as the metabolic byproducts of proteolysis, are essential indicators in assessing the freshness of meat. Excessive BAs intake can lead to a range of diseases, highlighting the importance of detection BAs in a rapid, on-site, accurate way. To this aim, a strategy of modulating sensing performance of the fluorescence probes is put forward and a portable detection platform is built up by a smartphone as well as a deep convolutional neural network (DCNN). Here, three fluorescence probes applying diphenylamine (DPA), thiodiphenylamine (PTZ), and carbazole (Cz) as the fluorophore cores is synthesized, which feature a specific recognition site of “N-H” and stepwise increased electrostatic potential (ESP). It is found that, CF₃Cz, with the highest ESP, enables superior detection performance, including a low detection limit (11 ppb), rapid response (<5 s) and noticeable bathochromic shifts (90 nm). Further, a portable smartphone-based fluorescence platform allows for the visual recognition and on-site quantitative detection of BAs. By combining with DCNN, this platform is able to predict meat freshness with an accuracy of 99.7 %. This work not only provides a design methodology by precisely modulating ESP, but also presents an intelligent detection platform for nondestructive assessment of BAs and meat freshness.