Real time aroma reconstruction using odour primaries

A methodology was developed to derive odour primaries based on the actual composition of a set of target aromas to be reconstructed. These odour primaries are mixtures of pure aroma components. The methodology was applied to reconstruct the aroma of five Citrus species (lime, lemon, orange, grapefruit and mandarin) in real time using an aroma synthesiser with four nozzles dispersing odour primaries blended using 12 pure odorants that was built for this purpose. The composition of the actual and reconstructed Citrus aromas was analysed using gas chromatography-mass spectrometry (GC-MS). Citrus species could be discriminated based on their headspace volatile composition. The aroma synthesiser was shown to be able to blend aromas in real time. The theoretical and actual correct classification accuracies of the reconstructed aromas were 80% and 73%, respectively. The required number of primaries and also pure odourants to construct these primaries was higher than anticipated, suggesting that further research with respect to the dimensionality of aroma spaces is required.