Machine learning-enhanced detection of chlorpyrifos using molecularly imprinted polymer-coated optical fibers.

This paper explores the use of large core declad optical fibers coated with molecularly imprinted polymers for chlorpyrifos detection, a key marker of organophosphate pesticides. The performance of sensor is evaluated using artificial neural networks and principal component analysis. By varying the declad length, the performance of molecularly imprinted polymer-coated fibers is compared to uncoated fibers, and both are used to identify commercial and pure samples of chlorpyrifos pesticides. Molecularly imprinted polymer-coated declad fiber sensors particularly those with longer declad lengths, exhibit significantly lower detection limits and higher sensitivity. The obtained maximum sensitivity, and minimum detection limit at 4 cm declad fiber length are 0.0027 mV/nM and 60.70 nM respectively. The results obtained also demonstrate that the artificial neural network can make an accurate prediction and the principal component analysis validates the efficacy of our molecularly imprinted polymer-coated fibers in chlorpyrifos detection.