Aroma improvement and modulation are important concerns in the winemaking process for which oenology offers a wide range of applicable methodologies and technologies. Microbial modulation [1
], grape post-harvest degradation [2
], temperature control during fermentation [1
], and stem contact fermentation [4
] can all be considered examples of traditional practices that aim to change wine flavor. More recently, other techniques have been developed, such as the recovery of aroma losses during fermentation [7
] or grape ozone treatment [9
]. These methodologies significantly change the wine aroma by modifying the typologies and/or the concentrations of the grape derived aroma compounds, and the microbially derived secondary metabolites [10
Traditional winemaking techniques that significantly impact wine aroma use different carbon dioxide pressures; examples include carbonic maceration [5
] and second fermentation of sparkling wines [13
]. In the carbonic fermentation, whole berries are maintained in an atmosphere saturated with carbon dioxide for one or two weeks. During this period, small amounts of ethanol are produced [12
] and, consequently, pressure in the maceration tank rises slightly. Sparkling wine production employs a similar second fermentation process that can occur either in the bottle, or in a pressurized stainless-steel tank. This step is required to form their characteristic carbon dioxide bubbles. Another alcoholic beverage that is produced with the fermentation at high pressure is beer [16
]; a profound change occurs in the volatile profile, with ethyl esters, acids, higher alcohols, and their acetates being the most affected compounds. During the production of both sparkling wine and beer, high carbon dioxide pressure (i.e., higher than 1 bar) significantly reduces the concentration of the above compounds, since pressure affects yeast activity [14
Recently, new technologies that produce a pressure slightly higher than the atmospheric pressure during the alcoholic fermentation have been developed. For example, the Nectar Tank (Trecieffe, Italy) uses carbon dioxide produced by the fermentation to enliven the juice and protect wine from oxidation. Such tanks are called isobaric fermenters. Similarly, the ACDF system (Dynamic Analysis of Fermentation Kinetics; Parsec, Sesto Fiorentino (FI), Italy) applied for fermentation monitoring, uses a pressure switch (set to 0.15–0.30 bar) to monitor fermentation kinetics and to enhance the pressure value. Furthermore, the Ganimede system uses a slight overpressure to achieve the pump-over effect of grape must on skins. For this reason, it could be grouped with the other systems conducting the fermentation at pressure higher than the atmospheric. According to the marketing information available in the respective company’s websites, these systems are able to enhance the extraction from berries. However, the effect of the slightly increased pressure on the aroma of the produced wines has not been studied yet since the studies found in the literature tested higher pressure conditions (higher than 1 bar e.g., sparkling wines and beers). Furthermore, the wine literature is mainly focused on the impact of overpressure conditions after the second fermentations, whereas scant (or no) information is reported on the effect of these pressures after the first fermentation. The hypothesis behind the present study is that a slight carbon dioxide overpressure affects both microbial metabolism and metabolites solubility, resulting in different wine aroma profiles. As carbon dioxide also contributes to extraction from skins, the experiment was conducted on juice only with the aim of separating those 2 effects. Hence, we investigated at the laboratory scale, the effect of a slight increase in pressure (between 0.2 and 1 bar) during the first fermentation of wines on the concentration of the microbially derived secondary metabolites, hence on the volatile profile of the produced wines. Finally, we verified the sensory difference with a discriminant test, using 48 judges, and with quantitation of selected volatile compounds, using head space, solid phase micro extraction, gas chromatography coupled with mass spectrometry (HS-SPME-GC-MS).
3. Results and Discussion
The fermentation was monitored every two days. It ended on day six, and a similar trend was observed in all the wine samples. No significant difference was found for ethanol content (mean 12.8 ± 0.2%) for all pressure levels tested and sampling times (days 0, 2, 4, and 6). Hence, pressure did not affect ethanol production. Consistently, at the end of the fermentation, all the samples showed less than 1 g/L of residual sugars. The acidity of the obtained wines was on the average 6.9 ± 0.2 g/L, without showing significant differences among the five pressure levels.
Seven days after fermentation ended, two sensory tests were performed, and the 26 compounds reported in Table 3
were quantitatively measured. The PCA clearly separated the control wines fermented at atmospheric pressure from all of the other samples fermented in overpressure conditions. PC1 explained 52.5% of the total variance, while PC2 explained 13.5% (Figure 1
). The P0 samples were placed in the left side of the score plot, while the P02, P05, P07, and P1 samples in the right side (Figure 1
a). Hence, PCA first loading was of particular interest since it allowed to identify the compounds that produced the aroma differences between atmospheric and overpressure conditions (Figure 1
b). The wine samples placed in the right side of the score plot were characterized by higher concentrations of compounds with positive loadings and lower concentrations of compounds with negative loadings. The opposite result was obtained for the wines placed in the left side of the plot. The highest loadings were found for esters (i.e., ethyl acetate, isoamyl acetate, butanol-3-methyl acetate, hexyl acetate, ethyl hexanoate, ethyl octanoate, ethyl decanoate, etc.), alcohols (i.e., 1-buanol, 1-hexanol), and some acids (i.e., acetic, octanoic and decanoic acid). Thus, the above chemical compounds can be considered as the main contributors to the aroma difference between the control samples and the samples produced in overpressure conditions. On the other hand, the highest negative scores were found for isobutyric and isovaleric acid, which were found to be higher in the P0 samples.
The ANOVA highlighted significant (p
< 0.05) differences for 13 compounds making up the aroma profile (Table 3
). Seven of the above compounds belong to the ester class (i.e., ethyl acetate, ethyl hexanoate, hexyl acetate, isoamyl acetate, ethyl dodecanoate, ethyl tetradecanoate, and phenylethyl acetate), five to the acid class (i.e., isobutyric acid, isovaleric acid, hexanoic acid, octanoic acid, and decanoic acid), and one to the alcohol class (i.e., 1-butanol).
Esters are usually related to the fruity flavor [20
] and, in general, they were found to increase in the wine samples produced under overpressure conditions compared to the control samples [21
]. Ethyl acetate is a well-studied volatile compound. In our trials, ethyl acetate concentration in the P
0 samples remained under the odor threshold (OT), while in the P02, P05, P07, and P1 samples ethyl acetate concentrations were significantly higher. Thus, the overpressure treatment, regardless the pressure level, made the ethyl acetate perceivable in the obtained wine samples. Some authors considered ethyl acetate a positive compound if its concentration remains below 200 mg/L [22
], as in our trials. However, ethyl acetate could have a suppressive effect on the perception of other esters and, consequently, it could negatively affect the wine flavor [20
The same result was obtained for ethyl hexanoate. In the literature, this compound is associated to fruity notes, in particular to a red berry aroma associated with “strawberry jam/red berry fruit/raspberry jam” [23
], and can be considered a positive compound for the aroma of fermented beverages [24
]. Similar to ethyl acetate, concentrations of ethyl hexanoate were significantly lower in the control wine samples compared to the samples fermented at higher pressures.
Isoamyl acetate concentrations were around 5.6 times higher in the P1 samples compared to the P0 samples. This compound is described as “sweet, fruity, banana, solvent” in wines [14
]. Thus, conducting a fermentation in overpressure significantly impacted the concentration of this ester. A regression (removing the P0 samples) to test for a decrease in isoamyl acetate related to overpressure increase was performed. Results showed that, for almost all of the aforementioned compounds, the highest concentrations in volatile compounds were obtained at the lowest pressure level (i.e., 0.2 bar), and the lowest concentrations at the highest-pressure level (i.e., 1 bar). An example is the significant decrease in the isoamyl acetate concentration as pressure increased (p
= 0.02). The highest isoamyl acetate concentration was produced when the lowest level of overpressure was applied; further pressure increases significantly reduced the concentration of this compound (Figure 2
). The literature reports contrasting results regarding the isoamyl acetate produced during fermentation in overpressure conditions. Martínez-García et al. [14
] reported a significant decrease of isoamyl acetate concentration during the second fermentation of a sparkling wine. Similarly, Renger et al. [16
] reported a reduction of the concentration of this compound during beer fermentation in overpressure (in the range 1–3 bar). Conversely, Tesniere and Flanzy [12
] showed a significant increase of isoamyl acetate concentration during the carbonic maceration (at slight overpressure) for the production of a Beaujolais wine. A possible explanation of the obtained results could be that a slight overpressure condition is responsible for the increase of the isoamyl acetate concentration in wines, by increasing the compound solubility, whereas further increases in pressure cause a significant reduction of this compound. Indeed, the wine samples produced at moderate overpressure conditions showed a higher concentration of this ester, while the wine samples obtained at higher pressure conditions were characterized by lower concentrations. Similar to isoamyl acetate, a significant linear decrease (p
= 0.04) was found for ethyl hexanoate concentrations as function of overpressure level. Finally, ethyl acetate reached the highest concentration at the lowest overpressure level (0.2 bar), and, like isoamyl acetate and ethyl hexanoate, its concentration significantly decreased when increasing the pressure level.
An explanation for the decrease in ester concentrations with increasing the overpressure level can be found in the beer and wine literature. A decrease in ethyl acetate when fermentation occurs at a pressure of 1 bar or more is reported in literature [25
]. Ethyl acetate biosynthesis is mediated by acetyl-CoA activity, which, at pressures above 1 bar (i.e., in the latter studies) is inhibited by carbon dioxide [27
]. This phenomenon does not occur at lower pressure levels (i.e., the pressure levels tested in the present study) where other phenomena are responsible for the significant increase of its concentration. Furthermore, Tesniere and Flanzy reported an increase in ethyl acetate concentration during carbonic maceration (at slight pressure) in the production of a Beaujolais wine [12
], while Martínez-García et al. [14
] reported a decrease during the second fermentation of a sparkling wine at pressure higher than 1 bar. Similarly, a reduction during the fermentation of beer produced under pressure conditions (in the range 1–3 bar) is reported in literature [16
]. These findings could reveal that a slight increase in pressure is able to increase isoamyl acetate concentrations in wines, while pressure higher than 1 bar may decrease it.
Hexyl acetate is another pleasant compound, related to the red berry aroma [28
]. Similar to the results obtained for ethyl acetate and ethyl hexanoate, hexyl acetate concentrations were found to be significantly higher in the wine samples produced under overpressure conditions compared to the control samples. In all the wine samples produced with overpressure the concentrations were over the OT (1.50 mg/L). Their odor activity values (OAV) ranged from 1.9 to 2.4, whereas the control samples showed a value lower than the OT threshold (OAV = 0.5). This compound is related to the carbonic maceration of grapes [11
] and has been found to decrease during the second fermentation of sparkling wines [14
Other esters showing higher concentrations in the samples produced under overpressure conditions were ethyl dodecanoate and ethyl tetradecanoate. The former is known to be a contributor to the fruity aroma [29
] and has been found in sparkling wines [14
], carbonic maceration wines [11
] and beers [24
]. The latter is described as fruity, like the aforementioned esters, but also as “butter, fatty” [14
]. Hence, ethyl tetradecanoate can be considered as the least-desirable ester among those detected. However, since both compounds were found to be below their respective OT (0.8 mg/L and 2.0 mg/L) in all the analyzed samples, they had very little impact on the overall taste of the resultant wines. Their average OAV were 0.5 for ethyl dodecanoate and 0.1 for ethyl tetradecanoate.
Phenylethyl acetate differed from the other esters since the lowest concentration was found in the P07 samples (i.e., 0.7 bar). This compound is described as “rose, honey, tobacco” [20
]. In the literature, it has been related to the second fermentation of sparkling wines [14
], and has been found to decrease during storage in contact with lees [29
To summarize, slightly increased pressures applied during fermentation significantly increased the concentration of the compounds related to fruity aromas. This result is consistent with the literature reporting that ethyl acetate, isoamyl acetate, and ethyl hexanoate are the main contributors of the fruity flavor in wine [7
Overpressure conditions applied during fermentation significantly changed the acid profile of the wine samples. Isobutyric, isovaleric, hexanoic, octanoic, and decanoic acids were found to be significantly different. Concentrations of isobutyric and isovaleric acids showed a significant decrease as pressure increased. Both of the above acids are undesirable compounds in wines. They are described as “rancid, butter, cheese” (isobutyric), and as “sweat, acid, rancid” (isovaleric) [20
]. Concentrations of both compounds were found to be the highest in the control wine samples; furthermore, it was found that the concentration of both these acids decreased as pressure increased. Isovaleric acid was perceivable in all the wine samples (i.e., OAV > 1), while isobutyric acid was only perceivable in the P0 samples (OAV = 1.1).
Similar to esters, concentrations of octanoic and decanoic acids significantly grew as pressure increased. However, unlike esters, these compounds can be considered undesirable in wine since both of them are related to a fatty and rancid aroma [20
]. Similar changes have been found in carbonic maceration wines (see [11
] for a discussion of octanoic acid), while a decrease has been found during the second fermentation of sparkling wines [14
]. Nevertheless, in our study, OAVs of these compounds remained lower than 1 in all the wine samples. Finally, concentrations of hexanoic acid significantly increased as pressure was enhanced; in all cases, concentrations were above the OT (OAV > 1), hence perceivable in all the wine samples.
The concentration of one alcohol (1-butanol) significantly increased as increasing the fermentation pressure. The difference was significant after seven days of fermentation. The literature reports that 1-butanol increases during carbonic maceration [11
] despite no particular effect on the wine aroma profile was reported for this compound.
The above discussed differences prove that a slight pressure increase during the alcoholic fermentation can change the volatile composition of a wine from a chemical perspective. However, from volatile compositional data, it is almost impossible to understand if pressure caused a sensory change of the wines at racking. Moreover, not all the active compounds for the wine aroma were quantified, and it is very difficult to consider all the interactions among volatiles (i.e., enhancement and masking effects). For these reasons we performed a discriminant sensory test. With the triangular test, we do not want to decide if a wine made with overpressure is better or worse than one made at atmospheric pressure. We only want to prove that the chemical differences due to the overpressure are perceivable by tasters, and consequently, that overpressure fermentation could be a low-cost strategy to modulate wine aroma.
The results of the triangle test were statistically significant (p
< 0.001): 32 of the 48 judges were able to distinguish the samples produced under pressure conditions from those fermented at atmospheric pressure. This finding was consistent with results of the chemical analyses. The significant differences in concentrations of the volatile compounds identified by GC-MS allowed panelists to discriminate between wine samples obtained in overpressure conditions from the control samples obtained at atmospheric pressure. The results of the preference test are shown in Figure 3
Approx. one third of judges (i.e., 17) preferred the control samples (0 bar), and approx. one third (i.e., 16) preferred the wine samples produced at the lowest pressure level (0.2 bar). The remaining judges were distributed across the other three wine samples (0.5, 0.7, and 1 bar). Sensory test were consistent with the chemical findings, showing that the greatest difference (in terms of volatile composition) was found between 0.2 bar and 0 bar wine samples. Furthermore, data from GC-MS allowed to split the wine samples in 2 groups: the control wine samples and the wine samples produced in overpressure conditions. In the same way, some judges preferred the control wine samples, whereas some others preferred the increased pressure wines (the wines were different but the preference is individual). Hence, applying pressure during fermentation caused significant chemical differences, which further resulted in wine samples significantly different in their sensory profile (i.e., “odor” attribute) as assessed by judges. Within the wine samples produced at pressures higher than the atmospheric pressure, the samples produced at the lowest overpressure level (0.2 bar) showed the highest concentrations of compounds, measured with GC-MS. This is consistent with the claimed differences in wine aroma profile by the commercial solutions using overpressure during fermentation (which work at low overpressures comparable with the 0.2 bar trial), where we measured the maximum effect on wine.
These results highlighted that a simple and low-cost technological solution was able to significantly change the wine aroma profile.