Food profiling using innovative voltammetric metallic-glassy carbon electrodes evaluated by clustering metrics

Abstract

Abstract This work proved that the quadruple-disk electrodes - iridium-platinum, platinum-glassy carbon, and iridium-glassy carbon can successfully work as a single sensor and act as voltammetric electronic tongue in food profiling. Current samples obtained with square-wave voltammetry (SWV) were additionally interpreted by a novel technique termed double-sampled differential square-wave voltammetry (DSSWV), which allowed for increasing the resolution and extracting the full, hidden information available in the signals. Four clustering validity indices: Calinski – Harabasz index, Davies – Bouldin index, Silhouette index and gap statistics were applied to select the optimal sensor and the current samples interpretation strategy. These four coefficients indicated the best iridium-platinum electrode combined with overall cathodic potential modulation using the technique of DSSWV. Correct clustering of objects, i.e., samples of five different isotonic beverages from one producer, was confirmed using principal component analysis and polar dendrogram. At the same time, it has been shown that in some variants of measurements and signal interpretation, it is not possible to correctly group the samples.