Résumé
Aroma of dark chocolate depends on process and cocoa origin and variety. Repeated sensory analyses of standardized chocolates produced from various cocoa beans batches using a standard fabrication process allowed classifying them into four sensory categories. These categories were confirmed in a PTR-ToF-MS analysis of the volatilome of 206 chocolate samples through supervised multivariate data analyses (PLS-DA). Variable selection using dedicated methods pinpointed some volatile compounds important for the discrimination of the chocolates. Moreover, the nosespace of a subset of chocolates was measured by twelve assessors using PTR-ToF-MS with simultaneous determination of their sensory temporal profile (Temporal Dominance of Sensations methodology, TDS) to better explain the sensory categorization at a perception level. The two sets of data were analysed conjointly by defining an index of abundance of each detected aroma compound while a given attribute was dominant: the Abundance While Dominance (AWD) index. Through correspondence analyses (CA) of these AWD indices some relationships between certain aroma compounds and the sensory attributes expected to be related to them were found.