Discrimination of Apple Liqueurs (Nalewka) Using a Voltammetric Electronic Tongue, UV-Vis and Raman Spectroscopy

Polish liqueurs prepared from five apple varieties have been used as a model of strong liqueurs. Principal Component Analysis has demonstrated that the best discrimination between liqueurs prepared from different apple varieties is achieved using the e-tongue and UV-Vis spectroscopy. Raman spectra coupled to chemometrics have not been efficient in discriminating liqueurs. The calculated Euclidean distances and the k-Nearest Neighbors algorithm (kNN) confirmed these results. The main advantage of the e-tongue is that, using PLS-1, good correlations have been found simultaneously with the phenolic content measured by the Folin-Ciocalteu method (R² of 0.97 in calibration and R² of 0.93 in validation) and also with the density, a marker of the alcoholic content method (R² of 0.93 in calibration and R² of 0.88 in validation). UV-Vis coupled with chemometrics has shown good correlations only with the phenolic content (R² of 0.99 in calibration and R² of 0.99 in validation) but correlations with the alcoholic content were low. Raman coupled with chemometrics has shown good correlations only with density (R² of 0.96 in calibration and R² of 0.85 in validation). In summary, from the three holistic methods evaluated to analyze strong alcoholic liqueurs, the voltammetric electronic tongue using phthalocyanines as sensing elements is superior to Raman or UV-Vis techniques because it shows an excellent discrimination capability and remarkable correlations with both antioxidant capacity and alcoholic content-the most important parameters to be measured in this type of liqueurs.