Simultaneous quantification of analytes in quaternary mixtures using dual pulse staircase voltammetry and genetically optimised neural networks

This paper describes the simultaneous quantification of four aliphatic compounds (ethanol, methanol, fructose and glucose) mixed in varying concentrations. The method used employs dual pulse staircase voltammetry (DPSV) to acquire an electrochemical voltammogram from the mixture. An artificial neural network, optimised using an elitist genetic algorithm, is then used to determine the concentration of each individual analyte from the voltammogram. The best average RMS errors achieved when testing with unseen data were 5.03%, 7.72%, 3.29% and 4.00% for maximum analyte concentrations of ethanol, methanol, fructose and glucose respectively. This work represents an important step forward because DPSV data is notoriously difficult to calibrate due to complex electrode–analyte interactions, and had not previous been shown to be amenable to quaternary mixtures.