When thinking of vanilla’s aroma, vanillin may be the first substance that comes to mind. Yet vanilla beans smell not only sweet: they also suggest flowers, fruit, and dry plant material. If we select and combine the compounds supporting those characteristics, how closely can we approach the original aroma?
In a study published in 2013, Makoto Takahashi and colleagues analyzed aroma-related compounds in Tahitian vanilla beans, mixed 13 compounds based on their measured amounts, and compared the mixture with the beans. Rather than stopping at identifying compounds with analytical instruments, they returned to the overall aroma perceived by people to test their findings.
What did the Tahitian beans smell like?
The subjects were beans of Vanilla tahitensis grown on Tahiti and subjected to the postharvest processing known as curing. Cured Vanilla planifolia beans from Madagascar were used for comparison.
The initial comparison involved 13 people trained to recognize and evaluate aromas. Ground beans were placed in containers labeled with numbers rather than sample names. Participants rated seven attributes on a seven-point scale: floral, dried fruit-like, resinous, hay-like, metallic, phenolic, and sweet. Here, “sweet” describes the smell, not a taste experienced while eating.
The Tahitian sample was rated more floral than the Madagascar sample, and less resinous and dried fruit-like. No statistically significant differences were found for the remaining four attributes.
This was, however, a comparison of two commercial samples. They differed in plant species, origin, and postharvest treatment, and their moisture contents were 45% for the Tahitian beans and 25% for the Madagascar beans. The differences cannot all be attributed solely to origin or solely to species.
Separating compounds and smelling successive dilutions
The researchers extracted aroma compounds from the Tahitian beans using a solvent and distilled the extract under reduced pressure to prepare an analytical sample. Gas chromatography separated the components sequentially, while mass spectrometry helped identify candidate molecules and people smelled the compounds at the instrument’s outlet. This technique is called GC-O.
A compound that produces a large instrumental signal is not necessarily prominent to the nose. The team therefore diluted the extract in successive fivefold steps to determine how far each compound could be diluted before its odor disappeared. This procedure is called aroma extract dilution analysis, or AEDA. Three experienced assessors each performed the GC-O analysis twice.
The results are expressed as FD factors, the greatest dilution factors at which odors remained detectable. A high-FD compound is one whose presence remains easy to detect after dilution of that extract. FD is not the same as the original concentration or perceived odor intensity, however, nor is it a value whose contribution to a mixture can simply be added up.
Twenty odor-active regions were found above the specified dilution threshold, and compounds were assigned to 19 of them. Two of those assignments were tentative. Not every component was identified with the same degree of certainty.
Vanillin and three anisyl compounds
Vanillin and anisaldehyde had the highest FD factors. Vanillin was described as sweet and vanilla-like, and anisaldehyde as anise- or raspberry-like. Anisyl alcohol, reminiscent of flowers and anise, and anisyl acetate, reminiscent of flowers and raisins, followed with high FD factors.
Quantification found approximately 3.17 g of vanillin and 3.15 g of anisyl alcohol per kilogram of beans, making them abundant among the compounds measured. Anisaldehyde, by comparison, was present at about 132 mg/kg. Despite being less abundant than vanillin, it reached the same high FD factor in the dilution-smelling test.
Thus, “present in large amounts” and “detectable after dilution” do not necessarily produce the same ranking. Alongside vanillin, the researchers considered these three anisyl compounds major candidates supporting the Tahitian sample’s character.
Rebuilding the aroma with 13 compounds at measured levels
Next, the team quantified 13 compounds above the FD threshold and mixed them at concentrations based on those measurements. Besides vanillin and anisyl compounds, the mixture included acetic acid, guaiacol, eugenol, and other substances. Water, an oil called medium-chain triglyceride, and an emulsifier were used to prepare a sample for comparison with the aroma released by the beans.
Eleven people compared the reconstituted aroma with the original Tahitian vanilla beans. This was a separate comparison from the 13-person test of beans from the two origins. The same seven aroma attributes were evaluated.
No significant differences were found for any of the seven attributes. The mean scores were not identical, but the panel size and assessment method did not detect clear differences. The authors concluded that the 13-compound mixture closely approximated the aroma characteristics of the Tahitian beans.
This demonstrates that the main aroma characteristics can be approached without assembling every natural constituent. It does not prove that every subtle odor is identical or that nobody could distinguish the samples. Nor was this an equivalence test designed to show that differences lay within a specified range.
The study also did not directly compare the mixture with vanillin alone or test mixtures with one component removed at a time. It therefore cannot establish whether all 13 compounds are indispensable or what changes when particular compounds are reduced.
Changing the extraction method can change the apparent key players
The study also raised an important methodological question. Earlier research had identified anisaldehyde and guaiacol as key odorants of Tahitian vanilla. Here, vanillin was also prominent, while guaiacol’s FD factor was lower than those of the main anisyl compounds.
The authors noted that earlier work had used distillation and extraction in water at 100°C for 2 hours, as well as a different way of classifying odors. Their study used room-temperature solvent extraction and reduced-pressure distillation, captured a broad range of odors through dilution analysis, and then evaluated a reconstituted mixture.
This method still involved heating during reduced-pressure distillation, so it was not entirely heat-free. Nor did the study directly test whether extraction alone caused the differences from earlier findings. Nevertheless, the point is important: how compounds are extracted, smelled, and selected can affect conclusions about the key contributors to an aroma.
Both the authors and the sensory panel were affiliated with T. Hasegawa. The study drew on specialized evaluation experience, but it was not a broad survey of consumer preferences.
Summary
In the Tahitian vanilla sample, anisaldehyde, anisyl alcohol, and anisyl acetate emerged alongside vanillin as important aroma candidates. Combining 13 quantified compounds produced an aroma with no significant differences from the original beans across the seven attributes evaluated by 11 panelists.
The study’s value lies in going beyond a list of ingredients: the team separated and smelled compounds, measured their amounts, and mixed them again to test the result. Understanding the aroma of a natural material requires both its composition and its effects on the nose. Moving between analysis and reconstruction makes it possible to examine the combinations underlying a complex aroma in concrete terms.
References
Takahashi, M., Inai, Y., Miyazawa, N., Kurobayashi, Y., & Fujita, A.(2013). Identification of the Key Odorants in Tahitian Cured Vanilla Beans (Vanilla tahitensis) by GC-MS and an Aroma Extract Dilution Analysis. Bioscience, Biotechnology, and Biochemistry, 77(3), 601–605. DOI: 10.1271/bbb.120840.Full original article.
Original article: © Japan Society for Bioscience, Biotechnology, and Agrochemistry 2013.
