In this work different electrode materials were investigated as sensors in a voltammetric electronic tongue. Basically, the electronic tongue is based on the combination of nonspecific sensors (electrodes) and pattern recognition tools, for example principal component analysis (PCA). Copper. glassy carbon, nickel, palladium, silver, tin, titanium and zirconium together with more traditional electrode materials such as gold, iridium, and platinum were studied. Cyclic voitammetry was applied to study typical model reactions in solutions containing different electroactive compounds, like ascorbic acid, glucose, histidine and potassium hexacyanoferrate(II). Different sensitivity and selectivity were obtained with the electrodes. Large responses were for example found for the amino acid and the carbohydrate using the copper, nickel and silver electrode. Some of the electrodes were employed in multicomponent solutions, i.e., liquid washing detergents from different suppliers together with differential pulse voltammetry. Responses from the electrodes in combination with PCA showed that they separated the detergents to different extents. This was further used when information from the sensors was merged together for successful discrimination of the detergents. It was found that two detergents close to each other in the score plot were from the same supplier. Furthermore. scanning electron microscopy (SEM) was used to monitor surface changes at the nonnoble electrodes (copper, nickel, and silver).

A hybrid electronic tongue is described based on a combination of potentiometry, voltammetry and conductivity. It was used for classification of six different types of fermented milk. Using ion-selective electrodes, pH, carbon dioxide and chloride ion concentrations were measured. The voltammetric electronic tongue consisted of six working electrodes of different metals (gold, iridium, palladium, platinum, rhenium and rhodium) and an Ag/AgCl reference electrode. The measurement principle is based on pulse voltammetry in which current transients are measured due to the onset of voltage pulses at decreasing potentials. The data obtained from the measurements were treated by multivariate data processing based on principal components analysis and an artificial neural net. The hybrid tongue could separate all six samples. Also, the nature of the micro-organisms in the different fermentations was reflected in the principal component analysis. Copyright (C) 2000 Elsevier Science B.V.

In this article, drift correction algorithms were used in order to remove linear drift in multivariate spaces of two data sets obtained by an electronic tongue based on voltammetry. The electronic tongue consisted of various metal electrodes (Au, Ir, Pt, Rh) combined with pattern recognition tools, such as principal component analysis. The first data set contained different types of liquid, from well defined to more complex solutions. The second data set contained different black and green teas. Component correction (CC) was compared to a simple additive correction. In CC, the drift direction of measured reference solutions in a multivariate space was subtracted from other types of solution. In additive correction, responses from reference samples were subtracted from other samples. CC showed similar or better performance in reducing drift compared to additive correction for the two data sets. The additive correction method was dependent on the fact that the differences in between samples of a reference solution were similar to the changes in between samples of other liquids, which was not the case with CC.

The use of experimental design as a tool to optimise electrochemically cleaned electrodes applied in a voltammetric electronic tongue is described. A simple and quick activation of electrode surfaces is essential for this type of device, especially for on-line applications in industrial processes. The electronic tongue consisted of four metal electrodes, e.g. Au, Ir, Pt, and Rh in a three-electrode configuration. Current was measured as a function of large potential pulses of decreasing amplitude applied to each electrode. Preliminary results showed that electrochemical cleaning activated the electrode surfaces to similar extent as polishing. Settings of potential and time for each electrode was determined with experimental design in a solution containing 1.0 mM K 4[Fe(CN)6] in 0.1 M phosphate buffer (pH 6.8). Electrode surfaces were deactivated in-between measurements in a complex liquid, like tea. Optimal settings for potential and time in the electrochemical cleaning procedure at each electrode were chosen at recoveries of 100% (compared to polished electrodes). The recoveries were larger than 100% when too large potentials and times were applied. This could be explained by the fact that the electrode areas increased and therefore also the current responses. Principal component analysis (PCA) was used to investigate the stability of the electrode settings at 100% recoveries. No obvious trends of drift in the signals were found. © 2004 Elsevier B.V. All rights reserved.

In this paper, three data compression methods are investigated to determine their ability to reduce large data sets obtained by a voltammetric electronic tongue without loss of information, since compressed data sets will save data storage and computational time. The electronic tongue is based on a combination of non-specific sensors and pattern recognition tools, such as principal component analysis (PCA). A series of potential pulses of decreasing amplitude are applied to one working electrode at a time and resulting current transients are collected at each potential step. Voltammograms containing up to 8000 variables are subsequently obtained. The methods investigated are wavelet transformation (WT) and hierarchical principal component analysis (HPCA). Also, a new chemical/physical model based on voltammetric theory is developed in order to extract interesting features of the current transients, revealing different information about species in solutions. Two model experiments are performed, one containing solutions of different electroactive compounds and the other containing complex samples, such as juices from fruits and tomatoes. It is shown that WT and HPCA compress the data sets without loss of information, and the chemical/physical model improves the separations slightly. HPCA is able to compress the two data sets to the largest extent, from 8000 to 16 variables. When data sets are scaled to unit variance, the separation ability improves even further for HPCA and the chemical/physical model. © 2001 Elsevier Science B.V.