Taste Classification by Data Fusion at a Single-Channel Hybrid Potentiometric and Impedimetric Electronic Tongue

The application of electronic tongues (e-tongues) in analyses relies on a series of measurements with expensive functionalized sensor arrays. In the current work, we establish a hybrid e-tongue by utilizing data fusion from potentiometry and impedimetry at a single channel composed of functionalization-free metal electrodes supported with pH electrodes. Potentiometric response encompasses pH and open circuit potential, whereas impedimetric response includes equivalent electric circuit parameters: charge transfer resistances and double-layer capacitances. Taste discriminability of the developed inexpensive and single-channel hybrid e-tongue was validated for the combination of two metal electrodes, Pt and Ni, with two different sample preparations of fundamental taste analytes at a concentration around the threshold of the human tongue. Qualitative discrimination was reported by dimensionality reduction and clustering of analytes in principal component analysis (PCA) plots. The quality of clustering was benchmarked with silhouette coefficients (SC) and hierarchical cluster analysis. A correlation among taste analytes and physicochemical parameters was established using heatmap plots and PCA loading plots to obtain insight into hybrid e-tongue. Taste discriminability diminishes because of the buffer activity of the standard reference solution (aqueous KCl and tartaric acid) used for sample preparation. Pt and Ni electrodes show strong clustering of taste analytes with SC of 0.83 and 0.74, respectively, without any specific sample preparation. Hybrid e-tongue, by data fusion at the single channel, has improved accuracy and clustering compared to independent potentiometric and impedimetric measurements. Reported single-channel hybrid e-tongue can potentially pave the path for commercial e-tongues.