Foods and beverage aroma results from multicomponent mixtures of volatile compounds present in the food that interact with olfactory receptors and produce a perceptual response in the brain. However, the perceptual interactions that occur when complex odor mixtures are combined are not well understood. Here we used Gas chromatography-Recomposition-Olfactometry (GC-R) to better understand the role that individual compounds have on the perceived sensory aroma of bitters. Bitters are the concentrated alcoholic extract of flavorful plant materials with a wide range of complex sensory and chemical aroma profiles that have not been extensively studied. Previously, we demonstrated that Angostura bitters are characterized by complex aroma attributes described as cola, ginger, orange peel, and black pepper and that the volatile composition of Angostura bitters is predominantly composed of terpenoids. Using GC-R to create in-instrument mixtures of the Angostura headspace extracts, the sensory attributes of Angostura extracts with linalool, α-terpinyl-acetate and caryophyllene omitted were evaluated. The omission experiments demonstrated direct and indirect effects of the individual compounds on the aroma attributes of Angostura bitters, through masking, additive, and synergistic interactions. Caryophyllene in particular, which was present in the headspace extracts at concentration only slightly above sensory threshold levels, had a large and unexpected impact on the sensory properties of the mixtures and may be most responsible for the aromas associated with the whole sample. The GC-R and statistical approaches used here provided valuable tools to reveal relationships among individual compounds and aroma attributes of foods that have not been currently theorized using existing analytical approaches.

This paper describes the design of a new instrumental technique, Gas Chromatography Recomposition-Olfactometry (GC-R), that adapts the reconstitution technique used in flavor chemistry studies by extracting volatiles from a sample by headspace solid-phase microextraction (SPME), separating the extract on a capillary GC column, and recombining individual compounds selectively as they elute off of the column into a mixture for sensory analysis (Figure 1). Using the chromatogram of a mixture as a map, the GC-R instrument allows the operator to "cut apart" and recombine the components of the mixture at will, selecting compounds, peaks, or sections based on retention time to include or exclude in a reconstitution for sensory analysis. Selective recombination is accomplished with the installation of a Deans Switch directly in-line with the column, which directs compounds either to waste or to a cryotrap at the operator's discretion. This enables the creation of, for example, aroma reconstitutions incorporating all of the volatiles in a sample, including instrumentally undetectable compounds as well those present at concentrations below sensory thresholds, thus correcting for the "reconstitution discrepancy" sometimes noted in flavor chemistry studies. Using only flowering lavender (Lavandula angustifola 'Hidcote Blue') as a source for volatiles, we used the instrument to build mixtures of subsets of lavender volatiles in-instrument and characterized their aroma qualities with a sensory panel. We showed evidence of additive, masking, and synergistic effects in these mixtures and of "lavender' aroma character as an emergent property of specific mixtures. This was accomplished without the need for chemical standards, reductive aroma models, or calculation of Odor Activity Values, and is broadly applicable to any aroma or flavor.