An ANN based biosensor for the detection of Pyrocatechol and hydroquinone in aqueous solutions

Abstract

This study presents a novel method using an artificial neural network (ANN) to detect Pyrocatechol and hydroquinone in an aqueous solution. The technique utilizes a nanocomposite Poly 3,4-ethylenedioxythiophene-Graphene Oxide Nanosheets (PEDOT-GONs) Laccase biosensor, which demonstrates excellent electrocatalytic activity and electrochemical performance for the redox reactions of the target compounds. The study successfully develops a radical cation electropolymerization process involving poly EDOT with GONs and Lac as dopants to ensure the effective entrapment of the enzymes. Experimental techniques such as electrochemical characterization and SEM confirm the successful entrapment of the enzymes. Further, the ANN was trained using the backpropagation algorithm by applying the experimental results, with current and Pyrocatechol as inputs and hydroquinone concentrations as output. The application of ANN modeling to biosensor technology is validated, demonstrating its utility. The biosensor's linear response detection limit is 0.041 μM within the 0.040–0.41 μM and 0.41–3.0 μM ranges. Finally, the model's performances were evaluated by testing it within the test samples. The performances were measured using precision, recall, accuracy, F1-score, sensitivity, etc., and validated with existing mechanisms.