A Full-Color Carbon Quantum Dots Fluorescence Sensing Array Combined with Machine Learning for Rapid Bacterial Detection and Classification.

Abstract

Rapid bacterial identification represents a critical priority in food safety, medical diagnostics, and environmental monitoring. Conventional methodologies predominantly rely on time-intensive cultivation processes, specialized analytical equipment, and bacterial recognition receptors, rendering them unsuitable for rapid and high-throughput applications. To address these limitations, a novel multichannel fluorescence sensing array based on carbon quantum dots (CQDs) is developed for rapid bacterial detection and classification. Utilizing a scalable acid reagent engineering strategy, water-soluble full-color CQDs exhibiting emission wavelengths ranging from 422 to 679 nm are successfully synthesized. The sensing array exploits differential fluorescence responses of CQDs to bacterial cell wall structures, surface potentials, and quantum yields, thereby enabling multidimensional fluorescence signal acquisition. Through integration with machine learning algorithms, the system demonstrates successful identification of five common pathogenic bacteria (Escherichia coli, Staphylococcus aureus, Salmonella typhimurium, Listeria monocytogenes, and Pseudomonas aeruginosa) with 100% accuracy. Further evaluation in a complex pork matrix demonstrates the array's capability for differentiating five bacterial species, quantitative detection, and identification of mixed bacterial populations, attaining classification accuracies exceeding 93%. This investigation presents a versatile and effective analytical platform for rapid bacterial detection and classification, with significant potential for application in food safety and related fields.