Rapid and on-site detection of carbaryl in complex matrices using cellulose-based organogel sensor with assist of machine learning.

Rapid and on-site monitoring of carbaryl residues in complex matrix is important for ensuring environmental and food safety, but it also presents significant challenges. We have developed a cellulose organogel sensor, which, with the aid of smartphones and portable imaging devices, enables accurate and quantitative detection of carbaryl. Specifically, p-aminobenzoic acid was covalently immobilized onto the cellulose. The cellulose organogel sensor was fabricated using a solvent-induced strategy, which exhibits good stability. The naphthol produced by the hydrolysis of carbaryl could react with p-aminobenzoic acid through an azo coupling reaction, generating a red product, thereby enabling the visual detection of carbaryl. The cellulose organogel features a dense and porous internal network structure, which effectively excludes interfering substances from the sample matrix and absorbs carbaryl into the organogel. The cellulose organogel sensor demonstrates excellent resistance to matrix interference from complex samples. The organogel sensor can directly detect carbaryl in extracts from river water, rice, cabbage, and soybean oil samples without additional separation or enrichment steps. We used the You Only Look Once (YOLO) algorithm to analyze the organogel images captured by the smartphone, effectively reducing interference and improving the sensor's detection accuracy. The organogel sensor achieves a detection limit of 0.19 mg kg-1 for carbaryl, with recoveries higher than 85.2 % in the samples. The entire detection process including sample pretreatment takes less than 40 min. This study presents a new and reliable solution for the on-site, accurate detection of carbaryl in complex samples by combining organogel sensors with machine learning.