This work describes the immobilization of 4-carboxybenzo-18-crown-6 (CB-18-crown-6) and 4-carboxybenzo-15-crown-5 (CB-15-crown-5) assisted by lysine on aryl diazonium salt monolayers anchored to the surface of graphite-epoxy composite electrodes (GEC), and their use for the simultaneous determination of Cd(II), Pb(II) and Cu(II) by differential pulse anodic stripping voltammetry (DPASV). These modified electrodes display a good repeatability and reproducibility with detection and quantification limits at levels of µg L(-1) (ppb), confirming their suitability for the determination of Cd(II), Pb(II) and Cu(II) ions in environmental samples. The overlapped nature of the multimetal stripping measurements was resolved by employing the two-sensor array CB-15-crown-5-GEC and CB-18-crown-6-GEC, since the metal complex selectivity exhibited by the considered ligands could add some discrimination power. For the processing of the voltammograms, Discrete Wavelet Transform and Causal Index were selected as preprocessing tools for data compression coupled with an artificial neural network for the modeling of the obtained responses, allowing the resolution of mixtures of these metals with good prediction of their concentrations (correlation with expected values for an external test subset better than 0.942).

The approach presented herein reports the application of a voltammetric electronic tongue (ET), in contrast with a wine tasting sensory panel, as a tool for standardized wine tasting; concretely, to achieve the discrimination of different wine DOs (Denominación de Origen, a mark related to its geographical region and ensuring high-quality levels) and the prediction of the global score assigned by the trained sensory panel. To this aim, a voltammetric array of sensors based on metallic and bulk-modified graphite electrodes was used as the sensing part, while chemometric tools such as linear discriminant analysis (LDA) and artificial neural networks (ANNs) were used as the qualitative and quantitative modelling tools. Departure information was the set of voltammograms, which were first preprocessed employing fast Fourier transform(FFT), followed by removal of less-significant coefficients employing a stepwise inclusion method and pruning of the inputs. The trend, in global scores, was modelled successfully with a 92.9% of correct identification for the qualitative application, and a correlation coefficient of 0.830 for the quantitative one (with 14 and 20 samples for the external test subsets, respectively)

Electronic tongue technology based on arrays of cross-sensitive chemical sensors and chemometric data processing has attracted a lot of researchers' attention through the last years. Several so far reported applications dealing with pharmaceutical related tasks employed different e-tongue systems to address different objectives. In this situation, it is hard to judge on the benefits and drawbacks of particular e-tongue implementations for R&D in pharmaceutics. The objective of this study was to compare the performance of six different e-tongues applied to the same set of pharmaceutical samples. For this purpose, two commercially available systems (from Insent and AlphaMOS) and four laboratory prototype systems (two potentiometric systems from Warsaw operating in flow and static modes, one potentiometric system from St. Petersburg, one voltammetric system from Barcelona) were employed. The sample set addressed in the study comprised nine different formulations based on caffeine citrate, lactose monohydrate, maltodextrine, saccharin sodium and citric acid in various combinations. To provide for the fair and unbiased comparison, samples were evaluated under blind conditions and data processing from all the systems was performed in a uniform way. Different mathematical methods were applied to judge on similarity of the e-tongues response from the samples. These were principal component analysis (PCA), RV' matrix correlation coefficients and Tuckeŕs congruency coefficients.

Second generation ethanol is produced from the carbohydrates released from the cell wall of bagasse and straw of sugarcane. The objective of this work is the characterization and application of a voltammetric electronic tongue using an array of glassy carbon electrodes modified with multi-walled carbon nanotubes containing metal (Paladium, Gold, Copper, Nickel and Cobalt) oxy-hydroxide nanoparticles (GCE/MWCNT/MetalsOOH) towards a simpler analysis of carbohydrates (glucose, xylose, galactose and mannose). The final architecture of the back-propagation Artificial Neural Network (ANN) model had 36 input neurons and a hidden layer with 5 neurons. The ANN based prediction model has provided satisfactory concentrations for all carbohydrates; the obtained response had a maximum NRMSE of 12.4% with a maximum deviation of slopes in the obtained vs. expected comparison graph of 15%. For all species, the comparison correlation coefficient was of r ≥ 0.99 for the training subset and of r ≥ 0.96 for the test subset