Different Chemometric Methods for the Discrimination of Commercial Aged Cachaças

Abstract

Cachaca, like other distilled beverages, often passes through an aging process and a peculiar fact is that Brazilian native woods are used for producing the barrels used for aging, in addition to oak. Cachacas can be aged in barrels of amendoim (Pterogyne nitens), balsam (Myroxylon peruiferum), jequitiba (Cariniana estrellensis), umburana (Amburana cearensis), and oak (Quercus sp.). Cachacas aged in wooden barrels have their final price for the consumer substantially increased, making them susceptible to fraud. So, methods to check the authenticity of aged cachaca are increasingly required. In this paper, 109 samples of commercial aged cachacas were used for the construction of supervised classification models—PLS-DA (partial least squares for discriminant analysis) and NPLS-DA (N-way PLS-DA)—by using UV-Vis spectrophotometry and spectrofluorimetry. The best results were obtained for the model based on low-level data fusion, providing good reliability rates for the training and test sets (ranging from 56 to 100 %) for all the analyzed classes.