Potential of model cakes to study reaction kinetics through the dynamic on-line extraction of volatile markers and...

This study presents a novel strategy for the dynamic analysis of volatile compounds extracted from baking vapors using a fit-for-purpose model cake. This model imitates a real sponge cake in terms of structure and processing but it is not reactive towards Maillard and caramelization reactions. When implemented with precursors (glucose (G) or glucose + leucine (G + L)), the reactions are activated and volatile markers can be monitored dynamically during baking. A method for the on-line sampling of vapors during baking using sorbent tubes coupled to thermal desorption (TD-GC-MS) has been developed and proven to be an appropriate and rapid technique to analyze a large number of volatile compounds within a broad range of physical and chemical characteristics. Volatile markers such as acetic acid, furfural, furfuryl alcohol and 5-hydroxymethylfurfual were identified using both models: glucose (G) and glucose + leucine (G + L) because they arise from both caramelization and the Maillard reaction. On the other hand, 3-methylbutanal and 2,5-dimethylpyrazine were only identified in the (G + L) model cake as they arise from the Strecker degradation pathway induced by the presence of leucine. Moreover, the relative abundance of all markers of reactions covers a broad range. On-line sampling coupled to TD-GC-MS enabled the collection of kinetic data on these markers throughout the baking operation and discrimination of the two formulas (G vs G + L) and two baking temperatures (170 °C and 200 °C) used. These results offer promise for the further use of this approach to study reaction kinetics in model cakes.