Differentiation of wood-derived vanillin from synthetic vanillin in distillates using gas chromatography/combustion/isotope ratio mass spectrometry for δ13 C analysis.

Abstract

Typical storage in oak barrels releases in distillates different degradation products such as vanillin, which play an important role in flavour and aroma. The addition of vanillin, as well as other aroma compounds, of different origin is prohibited by European laws. As vanillin samples from different sources have different δ13 C values, the δ13 C value could be used to determine whether the vanillin is authentic (lignin-derived), or if it has been added from another source (e.g. synthetic). The δ13 C values for vanillin derived from different sources, including natural, synthetic and tannins, were measured by gas chromatography/combustion/isotope ratio mass spectrometry (GC/C/IRMS), after diethyl ether addition and/or ethanol dilution. A method for analysing vanillin in distillates after dichloromethane extraction was developed. Tests were undertaken to prove the reliability, reproducibility and accuracy of the method with standards and samples. Distillate samples were run to measure the δ13 C values of vanillin and to compare them with values for other sources of vanillin. δ13 C values were determined for: natural vanillin extracts (-21.0 to -19.3‰, 16 samples); vanillin ex-lignin (-28.2‰, 1 sample); and synthetic vanillin (-32.6 to -29.3‰, 7 samples). Seventeen tannin samples were found to have δ13 C values of -29.5 to -26.7‰, which were significantly different (p < 0.05) from those of the natural and synthetic vanillins. The vanillin δ13 C values measured in distillates (-28.9 to -25.7‰) were mainly in the tannin range, although one spirit (-32.5‰) was found to contain synthetic vanillin. The results show that synthetic vanillin added to a distillate could be differentiated from vanillin derived from oak barrels by their respective δ13 C values. The GC/C/IRMS method could be a useful tool in the determination of adulteration of distillates.