Profiling of volatile substances by direct thermal desorption gas chromatography high-resolution mass spectrometry for flagging a characterising flavour in cigarette tobacco

Abstract

This paper describes an analytical method that supports the implementation of articles 9 and 10 of the WHO Framework Convention on Tobacco Control (FCTC) regarding the provisions on the reduction of the palatability and attractiveness of tobacco products regarding flavour ingredients. This study aimed to develop a screening method to identify cigarettes that may have a characterising flavour to support the implementation of the ban of characterising flavours of tobacco products, as laid down in the US and EU law. An analytical method combining direct thermal desorption and GC-QTOF MS was developed for acquiring the profile of volatile and semi-volatile substances in tobacco. A database of flavour additives was created comprising 133 compounds. A group of cigarettes without a declared characterising flavour was used to establish a reference profile of flavouring chemicals commonly present in tobacco products. A reference profile was modelled both by the means of principal component analysis (PCA) and based on the calculation of threshold values specified as 95th percentile of measured compounds' relative responses. Cigarettes and roll-your-own tobacco labelled as flavoured were analysed to evaluate the discrimination power of the method. A constructed model of the reference cigarettes allowed the differentiation of the flavoured tobacco products from the reference group. The method allows drawing conclusions on the chemical profiles of flavour constituents of tobacco products at even sensorial subliminal concentration levels and is suitable for both the initial screening of products on the market for characterising flavours and for confirmatory purposes after sensory analysis.

Keywords: Cigarettes; Flavour; Mass spectrometry; Tobacco filler; Volatiles.

Fig. 1

a PCA score plot of PC1 and PC2 of the developed model for the group of reference WDCF cigarettes. PC1 and PC2 captured 29.3% of total variance. b Control (Shewhart) chart of the quality control sample of the PC1 score. Green line represents variable average of the model. The warning (yellow lines) and control (red lines) limits are set as 2 and 3 times standard deviations from the average, respectively

Fig. 2

PCA score plot of PC1 and PC2 of the reference model with the flavoured tobacco products projected in the principal components space (a) and the zoomed plot with the six flavoured samples positioned closest to the reference group (b). Red dots represent the flavoured products; blue dots represent the reference group of WDCF cigarettes

Fig. 3

DModX PS+ plot of a distance to the model observed for the flavoured tobacco products. Plot is in normalised unit (absolute distance divided by the pooled residual standard deviation of the model). The red dashed line represents the critical value (DCrit), equal to 1.18. The left part of the plot is magnified to show the samples with the lowest distance values

Fig. 4

PCA score plot of PC1 and PC2 of the reference model with projected diluted flavoured tobacco products in the principal components space. Blue dots represent the reference group of WDCF cigarettes, green dots the original flavoured samples, and red dots the diluted flavoured samples. First plot (a) visualises only the menthol cigarette and its dilution; second rescaled plot (b) displays the dilution of other four flavoured samples

Fig. 5

Upper: Overlay of the profile of threshold values (red line) with the profile of the relative responses measured in the a strawberry-flavoured cigarette and b clove cigarette (blue lines). Lower: Overlaid total ion chromatograms (TIC) of tobacco fillers of six reference WDCF cigarettes (red lines) and tobacco fillers of a strawberry-flavoured cigarette and b clove cigarette (blue lines)