CORESTA Congress, Online, 2022, Smoke Science/Product Technology Groups, ST 74

Quantitative determination of 28 flavour compounds (lactones and pyrazines) related to potential additives in tobacco products

Labstat international Inc., Kitchener, Ontario, Canada

Many regulatory bodies require or will require the quantification of various flavour components in tobacco products addressing potential attractiveness or enhanced addictiveness issues related to these products. The objective of this study was to develop ‘fit for purpose’ methodology for the determination of a series of lactone and pyrazine flavour components. This study describes the quantitative determination of 28 compounds using gas chromatography - mass spectrometry (GC-MS).

The methodology used two grams of ground tobacco extracted with methyl tert-butyl ether (MTBE), a saturated potassium chloride (KCl) aqueous solution and internal standard solution containing four deuterated compounds (plus 4,4'-Dibromobiphenyl). After extraction and phase separation, an aliquot of the organic phase (MTBE) was used for GC-MS analysis using a 30 m DB-17MS column and single ion monitoring (SIM) data acquisition mode.

The methodology had many challenges related to commercially available reference materials limited to a mixture of isomers (e.g., 2-ethyl-3,5(or 6)-dimethyl pyrazine and 6,10-Dimethyl-5,9-undecadien-2-one). A change in the instrument response was also observed for some compounds (4-Vinyl guaiacol, trans-β-ionone, isoamyl phenylacetate and sclareolide), requiring a ‘bracketing standards’ approach for quantitation. Calibrations were prepared using standards ranging from 0.05 µg/mL to 10 µg/mL. However, a low S/N ratio for (-)-β-Citronellol, and poor level recoveries caused by matrix effects for γ-Valerolactone, required the lowest standard be 0.10 µg/mL. Recoveries of all target compounds from laboratory fortified matrix (LFM) at three different levels of fortification were within 79.5 - 123 %, except γ-Valerolactone (55 - 62 %) related to the matrix effects. The LFM samples were also used to evaluate method precision with the coefficients of variation (CV) for all analytes ≤ 15.

A suitable method for the quantification of 28 compounds with adequate sensitivity has been developed. However, the method does not differentiate between compounds naturally present in tobacco, those chemically bound to the matrix, and those that might have been added.