Chemometric tools associated to chromatographic data to assess the quality and authenticity whiskeys

Abstract

Whisky is a distilled beverage consumed all over the world and is often adulterated to obtain commercial advantages. Trace analysis by chromatographic methods is essential for the quality control in the production of whisky. These methods associated with chemometrics provide powerful tools for assessing authenticity and can be used to meet the criteria for differentiating whiskeys. This study proposes a new methodology that combines chromatographic data with chemometric tools to differentiate original and counterfeit whiskeys. Thus, counterfeit (14) and original whiskeys (16) were analyzed by two chromatographic methods. The first method involved the analysis of alcohols and esters using solid phase microextraction (SPME) - headspace (HS) - gas chromatography coupled to barrier discharge ionization detector (BID), HS-SPME-GC-BID. The second method involved the analysis of carbonyl compounds, derivatized with 2,4-dinitrophenylhydrazine (2,4-DNPH), using high performance liquid chromatography coupled to diode array detector (HPLC-DAD). Data analysis was carried out by the chemometric tools, such as principal component analysis (PCA), hierarchical cluster analysis (HCA) and partial least-squares discriminant analysis (PLS-DA). The merit figures for the 2,4-DNPH-HPLC-DAD method were studied, such as selectivity, linearity, recovery and precision. The homoscedasticity evaluation of the variances revealed heteroscedastic data and the quantitative analysis was performed using the weighted linear squares method (WLS). Correlation coefficients varied from 0.990 to 0.999. The significance assessment showed that the linear and angular coefficients were significant. Recovery and precision values of the method, respectively, ranged from 80.84 to 107.21 %, and 0.99 to 3.94 %, for the aldehydes, except for acrolein at one of the concentration levels. The HCA distinguished most of the whisky samples, PCA and PLS-DA distinguished the two classes completely. However, PLS-DA showed a greater predictive capacity than the PCA, as well as high sensitivity, specificity and precision. The validation of the 2,4-DNPH-HPLC-DAD was satisfactory, and chemometric tools allowed the differentiation between the original and counterfeit whiskeys.