Nitrogen-doped carbon dots as fluorescent probes for a web system-assisted intelligent detection model for flavonoids
摘要
Flavonoids are a class of polyphenolic compounds which are widely distributed in plants, and its excessive intake may cause adverse reactions. Traditional detection methods, which rely on costly instruments and consume a large amount of human resources, are difficult to use for on-site detection. In this study, a fluorescence sensing platform based on nitrogen-doped carbon dots (N-CDs) was constructed by a simple and easy-to-operate microwave method. Under the optimal detection conditions, the N-CDs fluorescent probe had a wide detection range (20 mg/L–100 mg/L) and a low detection limit (LOD) (myricetin 6.27 mg/L, morin 3.79 mg/L, kaempferol 3.74 mg/L). A paper sensor based on N-CDs was designed to realize the semiquantitative visual analysis of flavonoids. In addition, we have improved the detection performance of kaempferol by smartphone-assisted fluorescence detection, with the LOD of 1.45 mg/L. In order to further overcome the limitations of smartphone-assisted detection method on smartphone models, we integrated the Faster Region-based Convolutional Neural Networks (Faster R-CNN) target detection algorithm and machine learning (ML) regression algorithm. Consequently, a web-based concentration detection system for flavonoids was designed, establishing a portable intelligent detection platform.
Graphical abstract