Handbook of Vanilla Science and Technology

An evaluation of 93 samples of vanilla beans of different grades from the 5 producing countries was reported by Givaudan experts (Gassenmeier et al. 2008). The samples were extracted with ethanol in a Soxhlet unit, and analyzed by HPLC. Sensory analyses of the extracts found little correlation between “good vanilla” and vanillin content. Their HPLC analyses revealed a significant lack of compliance of the “ratios” (Table 10.2) with the accepted and regulated values notably for the case of Madagascar, even when the approved extraction conditions were used. The research discusses each bean source critically. They suggest that the strict enforcement of these ratios has lead to products produced to meet these parameters but that may be less flavorful. This confirms earlier research findings that vanillin explains less than 12% of the organoleptic variability (Hoffman et al. 2005). “A parameter that focuses on the purity of an extract or the amount of vanilla beans per solvent seems to be more appropriate,” according to these researchers. This would be very similar to the US Standard of Identity for vanilla (21CFR169.175). However, these researchers acknowledge that the vanillin/p-hydroxybenzaldhyde ratio is still useful to detect the addition of vanillin to vanilla extracts, perhaps as an initial investigation. 

Table 10.2 Ratios for vanillin: p-hydroxybenzaldehyde from various sources compared to the IOFI expected values. (I.O.F.I. International Organization of the Flavor Industry. Information letter No. 775, France: Vanilla Flavor 25.01.1998.)

Source | Sample Preparation | Number of samples/type | Vanillin/p-hydroxybenzaldehydeHPLC-RATIO

Ranadive, 1992 | Ethanol/Water | 7 various | 7.9-25.7

Lampprecht et al., 1994 | Ethanol/Water NF | 6 Bourbon | 11.4-24.4

Charvet and Derbesy, 2001 | AFNORNF ISO 5565 | 93 Mad.agascar, 29 Comores | 15.1-20.0

Charvet and Derbesy, 2001 | AFNORNF ISO 5565 | 265 various | 16.2-16.4

Scharrer and Mosandl, 2001 | Ethanol/Water (AOAC) | 15 various | 10-28

Gassenmeier et al., 2008 | Ethanol (French) | 85 various | 10-27.5

Expected IOFI | ― | ― | 10-20

In addition to HPLC methods, thin layer chromatography (TLC) is often investigated and useful. A thin layer variation of the liquid chromatographic method was developed by Belay and Poole (1993), which uses automated multiple development (AMD-TLC). With this technique, multiple samples can be analyzed simultaneously providing greater throughput while requiring minimal sample preparation by the analyst. This technique was capable of separating and quantifying the same vanilla flavor constituents as the common HPLC methods, and could be used to verify the authentication of products based on the constituent ratios.

A Russian group (Gerasimov et al. 2003) also developed a TLC procedure, which could be used to authenticate vanilla flavor in products. The methodical work optimized the chromatographic conditions specifically to separate vanillin and ethyl vanillin, as well as enabling the analysis of other chemicals commonly added to vanilla flavorings.

10.4.2 Isotopic techniques

The isotopic methods used for vanilla authentication involve a variety of analytes, ¹⁴C, ¹³C, ¹²C, H, D, ¹⁶O, ¹⁸O, ³H, as well as a number of techniques, liquid scintillation (LS), stable isotope ratio analysis (SIRA), gas proportional counting (GPC), nuclear magnetic resonance (NMR) including site-specific nuclear isotopic fractionation (SNIF-NMR) and accelerator mass spectrometry (AMS).

10.4.3 Radiometric and stable isotope ratio analysis

Researchers (Culp and Noakes 1992) at the University of Georgia’s Center for Applied Isotope Studies (CAIS) investigated the application of ¹⁴C,¹²C/¹³C and D/H isotopic ratios to establish the naturalness of a variety of 11 flavors, including vanilla. The targeted flavor compound was isolated and then the stable isotope ratio determined by an isotope ratio mass spectrometer (IRMS). They reported that D/H ratios were better at differentiating the source of these flavor compounds. For example, vanillin from guaiacol had a D/H value at —20%c, while pulp vanillin was at — 186%c, and vanillin from vanilla beans had a D/H value of —77%c. They advocated the use of a combination of all isotopic measurements in order to enhance the interpretation and define the flavor materials’ origin.