8E Identification of Impact Odorants of Wines 409
evaluate the effect of the omission of the odorant in the model. Obviously, the most
important odorants are those whose omission causes the highest impact. There are
few examples of this technique applied to wine because it requires a complete quan-
tification of all the aromas present in the sample (Guth 1997b; Ferreira et al. 2002a;
Escudero et al. 2004). Furthermore, this technique fails when the sample does not
contain really impact compounds, i.e. when the omission of the compounds (one
at a time) does not have relevant sensory consequences, as was demonstrated in
Escudero et al. (2004). Unfortunately this is something that is going to happen in
many wines. The experience learned from that work makes it possible to extract
some conclusions:
- Compounds showing similar aromas should be considered together as members
of a family, particularly in cases in which the compounds are formed along sim-
ilar biochemical routes. For instance, there is no sense in trying to determine
the impact of ethyl hexanoate, since this compound together with ethyl butyrate,
ethyl octanoate and ethyl decanote forms a family of compounds produced along
the same metabolic pathway and showing aromas which, together, our noses are
not able to differentiate.
- The impact of a compound cannot be predicted from its OAV value, but is the
result of the interaction of the odorant with the other odorants present in the
mixture. In general, experience has shown that impact compounds, whenever
they exist, are odorants present at concentrations higher than the threshold and
showing, usually, an odor quite different from the rest of odors present in the
wine and demonstrating a personal character. For instance, out of all fruity
esters of wine, only isoamyl acetate and phenylethyl acetate, both of which
have clear specific and differentiable aromas, have been found to be able to
act as impact compounds (Van Wyk et al. 1979; Tat et al. 2007). Similarly,
the three mercaptans derived from cysteine, 4-methyl-4-mercaptopentan-2-one,
3-mercaptohexanol and 3-mercaptohexylacetate, all of which have specific aro-
mas, have been found to be key compounds in some wines (Darriet et al. 1995;
Murat et al. 2001; Ferreira et al. 2002a; Campo et al. 2005).
- Instead of making a big reconstitution experiment on a single wine, very good
results can be obtained by studying simultaneously a group of 5–6 wines showing
a wide range of variability in the most important sensory descriptors and trying
to establish correlations between the scores of the main sensory descriptors of
the wines and both the olfactometric scores and the chemical quantitative data of
the most relevant odorants. Different types of easy sensory experiments, such as
ranking tests, can be further carried out to confirm or reject the observed corre-
lations. Examples of this kinds of work showing the effect of impact compound
or of whole families of compounds can be found in Campo et al. (2005) and
Escudero et al. (2007).
AcknowledgmentsThis work has been supported by the Spanish MEC (projects AGL2004-
06060/ALI and AGL2007-65139/ALI)