Abstract
Different microoxygenation procedures before and after malolactic fermentation were applied to Pinot noir wines from two consecutive vintages using the same Pinot noir block and a standardized winemaking protocol. The effects of microoxygenation treatments were investigated by spectrophotometry, HPLC-DAD, and descriptive sensory analysis. Microoxygenation before malolactic fermentation was carried out in 250-L red wine fermentors at oxygen dosages of 20 and 100 mg/L/month for 20 days. Oxygen dosages of 1 and 5 mg/L/month were applied for three months after malolactic fermentation using 100-L stainless-steel tanks. CIELab results showed that microoxygenation in 2006 tended to yield light-colored wines with an increase in yellow tonality. Examining monomeric polyphenols after alcoholic fermentation, this oxygen-induced color loss may be related to the high flavan-3-ol to anthocyanin ratio in this vintage. Furthermore, twice the increase in acetaldehyde content upon microoxygenation after malolactic fermentation was observed in 2006, which indicated excess oxygen for these wines. By contrast, a low flavan-3-ol to anthocyanin ratio was determined after alcoholic fermentation in wines of vintage 2007. Predominantly, these wines showed an increase in blueness upon microoxygenation treatments. Low correlations among astringency related attributes proved that the chosen attributes independently characterized the sensory variation in tannic structure among the microoxygenated Pinot noir wines. Descriptive analysis revealed that microoxygenation predominantly diminished green tannins in 2006 wines. Dry tannins, although decreasing in course of microoxygenation with low oxygen doses, rose again at higher oxygen doses. Beneficial aroma changes in 2006 and the increase of color intensity in 2007 were more salient upon microoxygenation prior to malolactic fermentation. An unfavorable impact on Pinot noir aroma occurred primarily due to microoxygenation after malolactic fermentation.
The Pinot noir variety does not possess any acylated anthocyanins (Brouillard et al. 2003), and its pigment content is fairly low compared to other grape varieties (Wenzel et al. 1987). Resulting Pinot noir wines at times show limited aging potential, as they lose their vivid red color, shifting rapidly toward orange or brownish tones. Additionally, winemakers often grapple with Pinot noir’s propensity to exhibit unripe tannins (Hillebrand et al. 1998), which are usually expressed by a combination of sensory attributes such as bitterness, rough astringency, and green flavor. Hence, producing Pinot noir with greater color and balanced tannicity has become a high priority among winemakers.
In numerous studies, the use of microoxygenation (MOX), which relies on the continuous supply of well-controlled doses of subsaturation oxygen during vinification and aging, was examined as a technique to promote intensified and/or stabilized color of red wines (Cano-López et al. 2006, 2007, 2008, De Beer et al. 2008). Reactions involving wine polyphenols are seen as key to processes changing the chromatic characteristics of the wine. Most of these reactions rely on the linking of anthocyanins with tannic structures of low molecular weight to form pigments such as ethyl-linked anthocyanin (epi)catechin dimers (Escribano-Bailon et al. 1996, Es-Safi et al. 1999) or pyranoanthocyanins (Fulcrand et al. 1996, Romero and Bakker 2000, 2001). Previous studies primarily focused on red wines, which are innately rich in anthocyanins, and only a few researchers have examined the effect of MOX on the chromatic characteristics of wines with medium pigment content (Castel et al. 2001, Bernath et al. 2002, Lesica and Kosmerl 2006). However, particularly the latter are of special interest as their color ought to be promoted to meet consumer expectations.
Although not scientifically proven, other benefits claimed to be achieved by MOX include a softening of tannins and a decrease of vegetative aromas. Various authors have reported the influence of MOX on molecular changes in phenolic composition (Castellari et al. 2000, Atanasova et al. 2002, Conte 2008) and aroma of red wines (Ortega Heras et al. 2008, Hernandez-Orte et al. 2009). Only a few studies have investigated sensory changes induced by MOX (Del Carmen Llaudy et al. 2006, Gonzalez-Sanjose et al. 2008) and even fewer have provided sufficient information on training and calibrating judges toward the subtle differences of some sensory descriptors. Changes in tactile impressions, such as mouthfeel and astringency, play a particularly crucial role in the context of MOX. Sensory researchers noted that tasters may be trained to reproducibly discriminate and rate the intensities of astringent subqualities elicited by young dry red wines (Gawel et al. 2001). Furthermore it has been proven that relevant tactile attributes such as green tannins and dry tannins are distinguishable by trained judges (Gawel et al. 2000).
Since its commercial release in 1996, MOX has become a common practice and is now used worldwide, although there is limited scientific information on when and how to apply MOX. Distinction has been made between the influences of MOX before and after malolactic fermentation (MLF) (Parish et al. 2000), related particularly to differences in the concentration of SO2 and its reaction with peroxide and acetaldehyde, thus limiting the condensation of anthocyanins with tannic molecules (Tao et al. 2007). Sulfur dioxide and other aspects are discussed in relation to MOX, such as the varietal and interannual variability among red wines (Perez-Magarino et al. 2007, Durner et al. 2008).
Although some studies on MOX differentiate between pre-MLF and post-MLF treatments, there are no published scientific data comparing the two procedures. In this article, we report the results of pilot-scale experiments, where Pinot noir wines of two consecutive vintages received MOX treatments either pre-MLF or post-MLF. For each practice, two doses of oxygen were applied, one receiving 5-fold the dose of oxygen. The aim of this study was to reveal the sensory changes induced by pre-MLF and post-MLF MOX treatments of Pinot noir by means of an extensive descriptive analysis method that is capable to detect even subtle variances in aroma, color, taste, and tactile attributes. In addition, the spectra of the bottled wines were recorded to allow a direct assessment of the influence of MOX on the chromatic characteristics of Pinot noir wines from different vintages. Several well-established chemical measures were discussed with regard to their ability to potentially act as indicators for MOX-induced color changes.
Materials and Methods
Wine.
Vitis vinifera L. cv. Pinot noir, clone Mariafeld, grapes were harvested from a vineyard of the Staatsweingut Neustadt, Pfalz on 6 Oct 2006 and 9 Oct 2007. In both years the same Pinot noir block had been used and the vinification process had been standardized. Prior to alcoholic fermentation, destemmed and crushed grapes were inoculated with 200 mg/L of the hydrated Lalvin BM4x4 yeast strain according to manufacturer instructions (Lallemand, Rexdale, Canada) and sulfited with 87 mg/L potassium bisulfite (50 mg/L SO2). At the beginning of alcoholic fermentation, two different treatments with oxygen dosages of 20 and 100 mg O2/L/month were started in quadruplicate using identical 310-L stainless-steel fermentors (pre-MLF MOX experiments). After 20 days oxygen delivery was stopped and replicates were pooled and pressed off the skins. MLF was conducted by inoculation with lactic acid bacteria (LAB) (Uvaferm Alpha; Lallemand). After MLF was finished, SO2 addition was omitted, since the depletion of the sugar and malic acid substrates and the complete elimination of any headspace in the storage tanks facilitated a sufficient protection against microbial spoilage. Accordingly, any interference of SO2 with the oxygen-induced condensation of anthocyanins with tannic molecules could be disregarded. Subsequent to MLF the untreated wine from the pre-MLF MOX experiment received two different oxygen dosages at rates of 1 and 5 mg O2/L/month for 92 days in duplicate (post-MLF MOX experiments). Treated and untreated wines from the pre-MLF MOX experiment were stored without post-MLF oxygen addition in duplicate. Regardless of the treatment applied, the batches were kept in identical 100-L stainless-steel tanks. The detailed MOX procedure, specifications on the equipment used, and the standardized vinification protocol are described in a separate article (Durner et al. 2010). Upon completion of post-MLF oxygen addition, the wines were sulfited by consecutive additions of SO2 summing up to 80 mg/L SO2 to achieve a stable level of 30 mg/L free SO2. Thereafter, the wines were bottled and closed with Classic+ stoppers (Nomacorc, Thimister-Clermont, Belgium). After bottling, wines were stored horizontally at room temperature for four months until sensory sessions started.
Chemicals.
All reagents used were of analytical grade unless otherwise stated. Acetaldehyde, potassium metabisulfite, (+)-catechin, and (−)-epicatechin were purchased from Sigma Chemical (St. Louis, MO). Acetonitrile (gradient grade), ethanol, and formic acid were purchased from Merck (Darmstadt, Germany). Malvidin 3-glucoside was purchased from Extrasynthese (Genay, France). Deionized water was purified to HPLC grade using a Millipore Milli-Q water purification system (Bedford, MA). Primary standards were prepared in ethanol:water (12:88 v/v). At least three working standards per compound were then prepared by diluting primary standards using the same solvent.
Standard chemical analyses.
Anthocyanin extractability (AE) and phenolic maturity (PM) of crushed grapes were measured using a described method (Glories 2001). Soluble solids, titratable acidity (TA), malic acid, pH, volatile acidity (VA), reducing sugar, ethanol, and fermentable nitrogen were determined using Fourier transform mid-infrared spectroscopy, including the appropriate calibration method for either must or wine (Foss WineScan FT120 Basic, Hillerød, Denmark). Acetaldehyde was analyzed using the R-Biopharm test combination for the determination of acetaldehyde in foodstuffs (product code 10 668 613 035; Darmstadt, Germany). Flavan-3-ol monomer analysis and anthocyanin monomer analysis were carried out using high-performance liquid chromatography (HPLC) and diode array detection as previously described (Rentzsch et al. 2009). The detection wavelengths were 520 nm for anthocyanins and 280 nm for flavan-3-ols. The concentrations of (+)-catechin and (−)-epicatechin were expressed together as flavan-3-ol monomers. The anthocyanins delphinidin 3-glucoside, cyanidin 3-glucoside, petunidin 3-glucoside, peonidin 3-glucoside, and malvidin 3-glucoside were quantified as malvidin 3-glucoside equivalents and expressed as anthocyanin monomers.
Spectrophotometric analysis.
Prior to analyses, wines were clarified by centrifugation (13,000 rpm for 10 min) and filtered using a 0.45-μm polyester membrane (Macherey-Nagel, Düren, Germany). Absorbance spectra were recorded in undiluted samples between 370 and 830 nm by a Cary 100 double-beam spectrophotometer (Varian, Palo Alto, CA), using a 0.2-cm path-length cell. A resolution of 1 nm and a slit width of 2.0 nm were used for all recordings. After values were corrected to 1 cm path length, absorbance spectra were converted to transmission spectra. Thereafter, rectangular coordinates L*, a*, and b* and cylindrical coordinates C* and h* were calculated using the CIE method (Commission Internationale de l’Eclairage 2004) with the 10° standard observer and the illuminant D65, according to the OIV recommendations (Waterhouse and Laurie 2006).
Sensory descriptive analysis.
Each vintage was evaluated four months after bottling from May to June of the respective year following harvest. The wines were kept under storage conditions as described above during the data collection periods. Prior to descriptive analysis, the individual replicates from each treatment were assessed by an experienced tasting panel (two enology consultants, two winemakers, and four postgraduates from the Viticulture & Oenology Department, Kompetenzzentrum Weinforschung) and considered not to differ by sensory means. The replicate wines within each treatment were then pooled and poured into clear DIN 10960 wine-testing glasses (Schott, Mainz, Germany). All samples were evaluated in isolated booths under standard conditions (Meilgaard et al. 1999).
Sensory panelists.
The sensory panel was composed of 22 and 23 volunteer participants in 2007 and 2008, respectively. All panelists were experienced wine tasters and all but five had previously participated in descriptive analyses. In 2007 the panel consisted of 11 male and 11 female participants; 11 male and 12 female panelists attended the sensory sessions in 2008. The panelists were recruited from the DLR Rheinpfalz and reported to consume red wine at least twice a month. Ages ranged from 22 to 63 with an average of 41 years. All participants signed an informed consent form and were awarded a small monetary incentive for participation.
Sensory training sessions.
An orientation session was held in both years to introduce panelists to red wine sensory evaluation. Three preevaluation sessions were conducted to create a list of attributes that were able to describe all sensory properties that differed among experimental wines. The list was revised as long as all panelists agreed on a final list of attributes and the respective definitions (Table 1⇓); in both years, four training sessions were run using the standards shown. To familiarize panelists with the quantitative evaluation of attributes, base-wine/standard-wine mixes were presented in varying ratios using black glasses (DIN 10960 wine-testing glasses; Schott). Quantitative training was run twice with single attribute standards and three times with blended attribute standards (Table 1⇓). FIZZ for Windows (ver. 2.00 D; Biosystémes, Couternon, France) was used for administration of sensory training and data collection. Judges were requested to rate low intensity of descriptors in the first third on a 15-cm unstructured line scale shown on the screen, medium intensity in the middle third, and high intensity in the last third. For training sessions and evaluation sessions, the scales were anchored with “low intensity” at 1.5 cm and “high intensity” at 13.5 cm.
Sensory evaluation sessions.
Thirty mL of each wine was poured about 30 min before evaluation. The glasses were covered with plastic lids and presented to judges at room temperature. In order to improve sensory calibration, during each session panelists had access to a set of standards (Table 1⇑) and a glass of base wine in each individual booth. The sessions were composed of five samples which were evaluated in triplicate by all judges with at least 24 hr between two sessions. In each of six sessions, the three-digit coded samples were presented in a completely randomized order. Color and aroma attributes were rated comparatively across the five samples, followed by taste attributes and tactile descriptors, which were evaluated monadically with a mandatory 1 min rest between each sample as part of the computerized evaluation session. Panelists were instructed to rinse thoroughly with water between samples because of the in-mouth persistence of the wines.
Statistical analyses.
Statistical analysis of chemical data was performed using one-way analysis of variance (ANOVA). Sensory data were analyzed using a three-way mixed model ANOVA, treating the panelists as a random effect and wines and replications as fixed effects. The least significant difference (LSD) test was used to determine statistically different values at a significance level of α ≤ 0.05. Sensory data points of significant attributes were submitted to a Shapiro–Wilk test to analyze for normality. Depending on whether the scores of two attributes were normally distributed or not, a correlation test was carried out using Pearson’s correlation coefficient if both attributes exhibited normality and Spearman’s correlation coefficient if at least one attribute showed nonnormal distribution. For sensory data, principal component analysis (PCA) was performed on means for judges and repetitions using Spearman’s correlation matrix with no rotation. Prior to further data processing, spectrophotometric data was normalized by calculating the absolute difference between treatments and the control. The PCA on spectrophotometric data was conducted using means for experimental replicates and Pearson’s correlation matrix with no rotation. Statistical analyses were performed using XLSTAT (ver. 2008.7.03; Addinsoft, Paris, France).
Results and Discussion
Chemical must analysis.
A rather high anthocyanin extractability (AE) measured in crushed grapes predicted poor anthocyanin extraction for the upcoming macerations in both vintages (Glories 2001). Anthocyanin extractability exhibited no vintage differences (Table 2⇓), which may be predominantly explained by a rather poor repeatability of the analytical method itself. By contrast, the phenolic maturity (PM) showed significantly higher values in 2006 compared with 2007, indicating a higher tannin to anthocyanin ratio and lower degree of ripeness. Low levels of volatile acidity (VA) in musts of both vintages proved a negligible impact of acetic acid-producing bacteria. Lower pH was measured in the 2006 must (Table 2⇓), probably because of higher content of tartaric acid in 2006. In contrast to the similar titratable acidity (TA) in both vintages, 2007 wines showed significantly higher malic acid (MA) content. As a preponderant acid of unripe grapes (Peynaud 1946), the high MA content in 2007 most probably indicated less light exposure of the grapes from veraison to harvest. As stated earlier, a lower phenolic ripeness was obtained in 2006, accordingly MA and PM showed somehow contradictory results. According to Glories’ argumentation, the risk that the sensory appearance of the wine may be negatively affected by unripe tannins is much higher in 2006 (Glories 2001). Hypothetically, MOX ought to support the softening of unripe tannins (McCord 2003); hence a MOX treatment may be considered for 2006 Pinot noir in particular.
Chemical analysis of bottled wines.
In contrast to the similar must TA in both vintages (Table 2⇑), bottled wines showed lower TA in 2007 (Table 3⇓), which can be explained by the strong contribution of malic acid in 2006 and its conversion into lactic acid by LAB. However, neither TA nor pH was affected by any of the MOX treatments. Although still below sensory recognition threshold, VA slightly increased by MOX treatment of 2007 wines but not 2006 wines (Table 3⇓), possibly explained by the higher pH in 2007 encouraging the viability of acetic acid bacteria as well as Oenococcus oeni and Lactobacillus brevis, which may also produce acetic acid at elevated pH values as well. A previous study revealed that under limited oxygen supply, acetic acid bacteria may reach a viable but nonculturable state in wine (Du Toit et al. 2005). Pre-MLF MOX-treated wines exhibited a more obvious increase in VA than post-MLF MOX-treated wines (Table 3⇓), presumably because of the elevated temperature of 28°C during alcoholic fermentation versus 15°C during wine storage.
Acetaldehyde, produced in the course of oxygen-triggered reactions of polyphenols (Wildenradt and Singleton 1974), significantly increased in bottled wines that received 5 mg O2/L/month post-MLF (Table 3⇑). One study reported an even higher accumulation of acetaldehyde during an 89-day MOX treatment of Merlot with 9.3 mg O2/L/month (Carlton et al. 2007). In contrast to the previous work, which did not differentiate between pre- and post-MLF MOX treatments, our results revealed that acetaldehyde increased in bottled wines that received post-MLF MOX. Any acetaldehyde that had been accumulated due to the pre-MLF MOX treatment was most likely degraded by LAB (Osborne et al. 2006) and/or diminished during the three months of storage after MLF, yielding insignificant differences compared to the control (Table 2⇑). Compared to the control wines of both vintages, the increase of acetaldehyde due to post-MLF MOX treatment was two times higher in 2006 wines as in 2007 wines. This observation may be attributed to the high content of flavan-3-ol monomers in 2006, which belong to the most readily oxidizable constituents of wine. Thus, the pronounced increase of acetaldehyde in 2006 wines may be related to the concurrent decrease of flavan-3-ol monomers (Table 3⇑), as they are the starting point of a cascade of oxidative processes, which are initiated by their oxidation (Danilewicz 2003). Although the concentration of monomeric anthocyanins was higher in 2007, MOX treatments caused a similar decrease of monomeric anthocyanins in both years, presumably because of their participation in polymerization reactions. This strong decrease of monomeric anthocyanins was independent from the amount of applied oxygen.
Chromatic characteristics.
The most obvious changes in wine color were observed in post-MLF MOX treatments of 2006 Pinot noir. Irrespective of oxygen dosage, these wines shifted upward in the a*-b*-color plane, with appreciable differences from the normalized coordinates of the controls represented by the intersection of both axes (Figure 1A⇓). In contrast to the increase in yellowness and decrease in redness due to post-MLF MOX, pre-MLF MOX-treated wine from 2006 showed no significant changes in a* and b* coordinates, indicated by error bars overlapping both axes. Strikingly different results were observed in MOX treatment of 2007 wines. Both post-MLF MOX treatments caused a downward shift in the a*-b*-color plane, indicating an increased blue color hue, which became more pronounced as higher oxygen dosages were applied. Depending on the oxygen dosage, pre-MLF MOX-treated wines from 2007 showed either a decrease in redness or an increase in blue tonality. The latter phenomenon was only observed upon the low oxygen dosage of 20 mg/L/month. Compared to the changes due to post-MLF MOX, the increase in blue tonality was significantly greater after the pre-MLF MOX regime. In addition to changes observed in the a*-b*-color plane, MOX-treated Pinot noir from 2006 were generally characterized by an increased lightness L* (Figure 1B⇓). In contrast to post-MLF MOX, which caused a rather high increase in L* independently from the applied oxygen dosage, a more moderate brightening of wines was observed upon pre-MLF MOX. However, wine from the consecutive 2007 vintage did not show any changes in lightness because of MOX treatments. Summing up, MOX-treated Pinot noir was generally characterized by either decreasing or stable red color hues, which is consistent with results from previous work (Del Carmen Llaudy et al. 2006). However, the influence of MOX on blue and yellow color hues was inconsistent between the 2006 and 2007 vintages, reflecting the inconsistent findings from separate studies (Cano-López et al. 2006, Perez-Magarino et al. 2007). Most likely, the vintage differences in MOX-induced color changes could be traced back to the phenolic composition of the wines (Table 3⇑). Considering the ratio of monomeric flavan-3-ols to anthocyanins (F/A ratio), 2006 control wines showed significantly higher values than 2007 controls (Figure 2⇓). While MOX procedures in 2006 caused dimished F/A ratios, wines from 2007 tended to yield increased F/A ratios upon MOX treatments. Regardless of which MOX procedure was applied, an increase of the F/A ratio was generally associated with a significant color shift toward a blue tonality (Figure 1A⇓) and no change in lightness (Figure 1B⇓). On the other hand, a shift toward a yellow tonality and an increase in lightness were observed only in conjunction with decreasing F/A ratios. In conclusion F/A ratios may have a distinct impact on the MOX-induced formation of polymeric pigments and their chemical structure (D. Durner, author’s unpublished data, 2008) and could account for the observed diverse chromatic shifts.
Panel performance and eligibility of tactile descriptors.
The common understanding of sensory descriptors and consistent use of sensory scales among the panelists were considered crucial to obtain reliable information on MOX-treated Pinot noir wines. Despite comprehensive training sessions, judges contributed significantly to the total variance in all attributes (Table 4⇓, Table 5⇓), possibly because of the variation in individual sensitivity among the panelists. With few exceptions, replications were an insignificant source of variation, demonstrating the consistent scoring of the panel throughout the three replications. In general, no significant judge*replication interaction was detected for tactile descriptors, revealing a common cognitive concept among the panelists.
Correlations among the various tactile attributes of 2006 Pinot noir were generally low, indicating that tactile descriptors were used to describe different astringent subqualities (Figure 3⇓). A weak correlation was detected between the astringency and hard mouthfeel attributes (R2 = 0.202). The descriptors green tannins and dry tannins were correlated neither to each other nor to the attributes astringency and hard mouthfeel, suggesting that panelists used these attributes independently from each other. Regarding the criticism that the proposed lexicon of astringent subqualities (Gawel et al. 2000) is too complex for descriptive panels (DeMiglio et al. 2002, King et al. 2003), the use of four tactile attributes was appropriate for our panel, as judges used them independently to describe the variability among MOX-treated Pinot noir wines. Furthermore, the physical standard development for astringency, bitterness, green tannins, and dry tannins provided a substantial support for the judges to differentiate among those tactile attributes (Table 1⇑).
Sensory analysis.
According to the ANOVA, 11 out of 17 sensory attributes changed significantly during the pre-MLF MOX treatment of 2006 Pinot noir (Table 4⇑). In 2007, only six descriptors accounted for a significant portion of the total variance, and all of them, except for dry tannins, also contributed to a significant variability in 2006. On the contrary, color hue, black currant, dried plum, green pepper, asparagus, and green tannins were only affected in 2006 wines.
The post-MLF MOX treatment of 2006 Pinot noir triggered significant changes in 10 sensory descriptors (Table 5⇑), with most also contributing to the variability induced by pre-MLF MOX. Only raspberry and black currant changed solely during pre-MLF MOX and the tactile descriptor dry tannins only during the post-MLF MOX treatment of 2006 Pinot noir. In 2007, post-MLF MOX treatment yielded only significant changes in color intensity and raspberry. Obviously, the interannual variation in chemical composition of the grape material and subsequent wines affected the sensory impact of MOX between vintages.
The following PCA were restricted to those sensory attributes, which changed significantly at least due to one MOX treatment. According to the PCA for MOX-treated 2006 wines, the first two PCs explained 83% of the variance among all significant sensory attributes (Figure 4⇓). From the right to the left, PC1 distinguishes the MOX-treated wines from the control, showing a decrease in color and tactile attributes. PC2 may be considered as the axis that differentiates pre-MLF from post-MLF MOX treatments. Whereas all MOX procedures yielded wines with a substantial color loss, an increased fruit odor was solely observed due to an early oxygen addition. Wines that received 20 mg O2/L/month pre-MLF exhibited somehow less reduction in color intensity but benefited equally from increased fruit aroma and the decrease in green tannins. On the contrary, post-MLF MOX-treated wines had been characterized by enhanced dried plum aroma, which could be a marker for oxidized fruit character. Summarizing these findings, post-MLF treatments of 2006 Pinot noir mainly caused disadvantageous sensory changes such as color loss or dried plum, regardless of oxygen dose. On the other hand, wines benefited from low-dosage MOX treatment prior to MLF, as fruit intensity improved and tannin perception decreased, validating the empirical records from winemakers and the recommendations given by enology consultants (Parish et al. 2000).
The PCA for 2007 wines indicated that MOX treatments accounted for 81% of the overall variance in sensory attributes, which consisted of a lower number of significant attributes compared to the 2006 vintage (Figure 5⇓). Although defined by a different set of significant sensory descriptors, changes induced by pre-MLF MOX treatments were more noticeable than those resulting from post-MLF oxygen additions. Regardless of the applied procedure, 2007 wines generally inclined to lose raspberry aroma in course of the MOX process, consistent with the diminished berry/plum intensity in MOX-treated Pinotage wines (De Beer et al. 2008). A pre-MLF oxygen dosage of 20 mg/L/month and both post-MLF treatments triggered an increase in color intensity. Wines that had been treated with 100 mg O2/L/month pre-MLF were predominantly characterized by a decline in mouthfeel, astringency, and dry tannins, but color attributes were not affected. Given that pre and post-MLF MOX treatments could not be differentiated along PC1 and PC2, it is impossible to make sweeping statements about potential assets and drawbacks of the different MOX procedures on 2007 Pinot noir.
Regardless of the oxygen dose, pre-MLF MOX treatments in 2006 yielded wines with an increased fruit odor (Figure 6A, B⇓). This finding is in good accordance with previously published work (Gonzalez-Sanjose et al. 2008). However, the simultaneous increase in dried plum also suggests that oxidative changes occurred during pre-MLF MOX, which may eventually result in an early deterioration of wine bouquet. The consecutive vintage was mainly unaffected in fruit aroma; however, a pre-MLF MOX treatment with 100 mg O2/L/month caused an unpleasant decrease in raspberry intensity. Further aroma descriptors such as soy sauce, green pepper, and asparagus decreased in course of the pre-MLF MOX treatments in 2006. This decline had been amplified with an increasing oxygen dose in 2006 but was not observed in 2007 Pinot noir. Losses in color intensity and browning were detected in the pre-MLF MOX treatment of 2006 Pinot noir (Figure 7A⇓). This deterioration of color continued with an increase of the oxygen dosage (Figure 7B⇓). The influence of pre-MLF MOX on color changes in 2007 contradicts these observations. These wines showed enhanced color intensity upon an oxygen dosage of 20 mg/L/month; however, no changes were observed due to an addition of 100 mg O2/L/month. These data reveal a distinct vintage effect and therefore differ from that reported by others (Perez-Magarino et al. 2007). In 2006, the tactile attributes astringency and green tannins decreased due to pre-MLF MOX treatments and attenuations grew stronger as oxygen was increased. In contrast, the 2007 Pinot noir showed lower intensities in mouthfeel, astringency, and dry tannins only if treated with 100 mg O2/L/month pre-MLF.
Post-MLF MOX treatments in 2006 did not affect wine raspberry aroma and black currant odor; however, the raspberry intensity in the 2007 vintage decreased regardless of the applied oxygen dosage (Figure 8A, B⇓). By contrast, dried plum aroma increased only in the treatment of 2006 wines. Comparable to the changes induced by pre-MLF MOX, nonfruit aroma attributes decreased in the course of post-MLF MOX treatments in 2006 and remained unchanged due to oxygen additions in 2007. Observing a substantial deterioration of color intensity and color hue in 2006, post-MLF MOX treatments in 2007 caused an enhancement of color, which grew more intense with an increasing oxygen dosage (Figure 9A, B⇓). Considering the effect on tactile attributes, decreasing intensities observed due to post-MLF MOX treatments in 2006 could not be reiterated in 2007 wines, showing the previously mentioned substantial vintage effect. Fading of the mouthfeel, astringency, and green tannin descriptors was even more pronounced at a higher oxygen dosage; however, dry tannins decreased only significantly in wines that received the low dose of 1 mg O2/L/month. This inconsistency was already addressed by Lemaire (1995), who stated that dry tannin perception decreases upon appropriate oxygen dosage, but increases with excessive oxygen addition.
Through examination of color and tactile descriptors, which are related to phenolic compounds in wine, it became obvious that both desired effects of color enhancement and attenuation of astringency never occurred simultaneously in any of the investigated MOX procedures. This is in accordance with the observation that beneficial changes in the tactile perception of MOX-treated wines solely occurred in combination with diminished color (Del Carmen Llaudy et al. 2006). Various publications on the oxygen-induced changes in wine phenolic composition suggest that MOX triggers the formation of a wide array of phenolic derivatives (Atanasova et al. 2002, Cano-López et al. 2006, 2007, Del Carmen Llaudy et al. 2006). Upon closer examination of those compounds, it is not convincing that they exhibit a similar impact on the two distinct sensory properties of color and astringency. Furthermore, sensory studies suggest that MOX-derived phenolic polymers either enhance color or soften astringency. To elaborate on this point, we presume that derivatives, which account for the softening of tannins, preferably originate from wines with high flavan-3-ol concentrations (Table 3⇑). As the oxygen-induced polymerization of flavan-3-ols is also known to be responsible for the formation of brown pigments, high flavan-3-ol concentrations may also account for any MOX-blamed color reductions (Es-Safi et al. 2002, Li et al. 2008).
Conclusions
Descriptive analysis proved to be a powerful tool to detect and to differentiate sensory changes induced by various microoxygenation treatments of Pinot noir wines. Results here clearly showed that MOX appeared to yield less astringent wines with diminished green tannins. Dry tannins, however, showed contradictory results, as intensities decreased upon low dosages of oxygen but increased again due to the supply of high doses. These results suggest that an adequate dose of oxygen, accustomed to the vintage-dependent composition of phenols, is indispensable to obtain positive effects on sensory properties of Pinot noir. A darker color and a shift toward blue tonality were detected in 2007 wines; however, 2006 wines showed detrimental color loss and browning upon MOX treatments, indicating a distinctive vintage effect. Considering the flavan-3-ol to anthocyanin ratio in both vintages, it seems reasonable to assume that both anthocyanin and flavan-3-ol content contribute to the quality of MOX-induced color changes. Thus, our results suggest that a high flavan-3-ol to anthocyanin ratio is detrimental for red wine color. This hypothesis is based on the fact that browning in wines relies on molecules that form brown pigments due to the oxygen ingress. Presuming that light-colored Pinot noir wines often exhibit a high flavan-3-ol content, the reluctance among winemakers to apply MOX to lighter Pinot noir is reasonable. The results of the present study propose further scientific consideration of the flavan-3-ol to anthocyanin ratio as a meaningful predictor of MOX-induced color changes in Pinot noir wines. To verify the importance of the flavan-3-ol to anthocyanin ratio, further experiments should apply MOX to a broader spectrum of Pinot noir wines that vary even more in the F/A ratio.
In addition to the effects on color, the tactile impressions of Pinot noir wines of both vintages were softened upon pre-MLF MOX. Moreover, beneficial aroma changes in 2006 and the increase of color intensity in 2007 had been more pronounced due to pre-MLF MOX. On the other hand, the most disadvantageous impact on aroma descriptors occurred predominantly during post-MLF MOX treatments, which may question the general suitability of a late MOX application for Pinot noir. Further chemical analyses of both MOX treatments will be conducted to determine the detailed evolution of compositional changes during vinification and aging of MOX-treated Pinot noir wines.
Footnotes
-
Acknowledgments: This research project was supported by the German Ministry of Economics and Technology (via AiF) and the FEI (Forschungskreis der Ernährungsindustrie e.V., Bonn), project AiF 14788 N.
-
The authors express gratitude to Annette Schormann for her support in developing and conducting descriptive analysis. Manfred Jutzi is acknowledged for many fruitful discussions on statistics.
- Received November 2009.
- Revision received May 2010.
- Accepted August 2010.
- Published online December 1969
- Copyright © 2010 by the American Society for Enology and Viticulture