8F Interactions Between Wine Matrix Macro-Components and Aroma Compounds 423
perception of some aroma compounds, such as citral (Zamora et al. 2005). Because
it can influence viscosity of the beverages, it could modify aroma release and thus,
aroma perception (Nurgel and Pickering 2005).
Nevertheless, the most studied ethanol effect is related to its capacity to mod-
ify solution polarity, thus altering the gas-liquid partition coefficient. An increase
in ethanol content has been shown to decrease the activity coefficients of many
volatile compounds in wine because of an increase in solubility (Voilley et al. 1991).
Hartmann et al. (2002) showed a decrease in the recovery of 3-alkyl-methoxy-
pyrazynes extracted with a divinylbenzene/carboxen SPME fibre from wine model
systems when the ethanol content increased from 0% to 20%. Similarly, Whiton
and Zoecklein (2000) reported that a small increase in ethanol content (from 11%
to 14%), in general, reduced the recoveryof typical wine volatile compounds. Both
of these studies suggest that increasing the alcohol content will reduce the release
of volatile compounds from wines.
Aznar et al. (2004) used static headspace-APCI-MS to study the release of
volatiles from water and hydroalcoholic systems (12 vol.%). They found a decrease
in the headspace concentration of volatile compounds with an increase in the log
P values (hydrophobicity values) until log P=3. Nevertheless, for very non-polar
compounds (log P>3), they did not find this trend; this could be due to changes in
hydrophobic interactions in the solution.
At higher ethanol concentrations (17–20 mL/100 mL), a decrease in volatility
of ethyl esters and aldehydes has been found and this effect cannot be explained
only by an increase in the solubility of the aroma compounds afforded by the added
ethanol. This effect has been attributed to changes in the structure of the solution
where ethanol molecules can aggregate at molar fractions above 0.05–0.06 (corre-
sponding to 15–17 mL/100 mL of ethanol) creating hydrophobic areas (or ethanol
clusters) able to retain other low water-soluble components (Conner 1994; Escalona
et al. 1999). It has been found that the addition of wood extracts (in the case of model
spirit solutions) increases this effect (reduction in activity coefficients) for some
volatile compounds at ethanol strengthsabove 10 mL/100 mL (Conner et al. 1999).
Nevertheless, Escalona et al. (2001) did notfind the same effect after the addition
of wood extracts to model wines. This author suggested that this discrepancy may
have been due to differences in the wood extracts used in the studies. Those used
in Conner’s work were obtained in whisky aging, extracting fractions with higher
ethanol solubility, while the extracts used in the wine study were obtained from
wine aging and were mainly composed of more water soluble compounds that had
no contribution to the formation of ethanol clusters.
There are few studies reporting on the effect of ethanol on the release of aroma
compounds using dynamic methodologies. In one study (dynamic headspace anal-
ysis and APCI-MS), Tsachaki et al. (2005) observed that in aqueous systems (no
ethanol) there was a rapid decrease in MS signal intensity until the rate of replen-
ishment equalled the rate of loss from headspace purging. However, above ethanolic
solutions, there was a similar initial rapid decrease followed by a “steady state”loss
at much higher levels (at 50–90%) of the initial relative intensity depending on the
volatile compound (Fig. 8F.2). In contrastto the aroma release effects noted under