Machine learning - optimized molecularly imprinted electrochemical sensor based on Bi-rich BiOBr/BC@AuNPs for trace determination of sulfachlorpyrazine sodium in food.

A highly sensitive and selective molecularly imprinted electrochemical sensor was developed for the detection of sulfachloropyrazine sodium (SPZ), assisted by machine learening (ML) optimization. The sensing platform was fabricated using novel Bi-rich BiOBr/BC@AuNPs nanocomposite, where the synergistic combination of Bi-rich BiOBr, conductive biochar, and gold nanoparticles markedly enhanced electron transfer and provided abundant electroactive sites. Critical imprinting parameters, including the functional monomer-to-template ratio, electropolymerization cycles, elution time, and incubation time, were optimized using a central composite design coupled with ML algorithms like support vector regression. Under optimal conditions, the sensor displayed a wide linear range of 0.031-61.34 ng/mL, with a low detection limit of 0.0025 ng/mL and a quantification limit of 0.0077 ng/mL. The sensor exhibited excellent selectivity, effectively distinguishing SPZ from structurally analogous sulfonamide. Successful application to spiked chicken, milk, and honey samples yielded satisfactory recoveries and reproducibility, confirming its strong potential for practical food safety.