A Smart Electrochemical Biosensor for Arsenic Detection in Water

Biosensors contain biological receptors for the accurate detection of a variety of analytes. However, the efficacy of these bioreceptors, when immobilized on the surface of working electrodes, tends to diminish over time. This necessitates frequent replacement, consequently inflating the costs and adversely affecting the commercial viability of biosensors. In this study, first, a three-electrode electrochemical biosensor incorporating Au nanoparticles was constructed to facilitate the measurement of arsenite, the trivalent form of arsenic commonly found in water sources. Subsequently, machine learning was employed in the structure of the biosensor, considering electrochemical data, sample pH, enzyme lifespan, and storage temperature as input features. To enhance the performance of the models, the optimized values of the parameters belonging to artificial neural networks (ANN) and support vector machines (SVM) were obtained using the Harris hawks optimization (HHO) and whale optimization algorithm (WOA). The hybrid models, HHO-SVM and HHO-ANN, exhibited promising results, with coefficients of determination (R
²
) of 0.89 and 0.85, respectively. These results were obtained from data collected by the biosensor over a 45-day period following the immobilization of arsenite oxidase and Au nanoparticles on the electrode. This study underscores the role of metaheuristic optimization techniques in enhancing the efficiency of intelligent biosensors.