Analysis of Volatile Components of Varietal English Wines Using Stir Bar Sorptive Extraction / Gas Chromatography-Mass Spectrometry

Aroma is an important property of wine and it can be influenced significantly by enological practices. The aim of this work was, by use of stir bar sorptive extraction/gas chromatography-mass spectrometry (SBSE/GC-MS), to compare semi-quantitative concentrations of the volatile constituents of stainless steel tank-fermented/matured Huxelrebe, Ortega, Schönburger and Siegerrebe varietal wines from a commercial English vineyard, with corresponding wines produced by oak cask (‘barrel’) fermentation/maturation. Aroma profiles of tank and barrel wines were different, with more volatiles detected and net concentrations being higher in barrel wines. Long chain ethyl carboxylate esters were generally more abundant in barrel wines, whereas acetate esters were generally more prominent in tank wines. By conducting a short (~7 month) maturation period in secondhand (third or fourth fill) casks, it was possible to make wines with more complex aromas, but without obvious oak aroma.


Introduction
Aroma is a very important property of wine, contributing much to its overall quality and commercial value.The aroma profile of a wine is the result of the combined contributions of hundreds of organic compounds, some of which are present in concentrations lower than µg/L (parts per billion) [1].Some of these low concentration components have exceptionally low odour threshold values (OTVs; OTV is the lowest concentration of a specific odour compound that is perceivable by the human sense of smell), sometimes in the ng/L (parts per trillion) range [2], and hence make a significant individual contribution to the overall wine aroma.Many others, present in concentrations lower than their OTV, may make a valuable background contribution.
The low concentration of most wine volatile compounds makes the use of an extractive enrichment or focusing technique essential.Such a technique is stir bar sorptive extraction (SBSE) [3].SBSE has been used to extract volatile compounds from grape juice [4][5][6][7][8][9][10] and wine [9][10][11][12][13][14][15][16][17], including wines that have had prolonged contact with oak wood, such as maturation in oak casks [12,13], and also including liquid (solvent) desorption in one case [15], as opposed to the standard thermal desorption.Similarly, SBSE has been used to extract volatile compounds from grape juice and wines of grapes to which French oak extract was applied at the ripening stage (veraison) [9,10].In the case of wine, all SBSE studies have been carried out on wines made from warm climate varieties, such as Malvasia [12], Merlot [14], Pinotage [16], Verdejo [9] and Petit Verdot [10], all growing in warm regions.Of two SBSE investigations on the influence of oak maturation on wine aroma, one was on Madeira wines [12] and the other on red wine [13], and both studies involved wines that mostly had prolonged maturation periods (>1 year) in oak casks.
The aim of this work was to determine for the first time the aroma profiles of English varietal wines made from Huxelrebe, Ortega, Schönburger and Siegerrebe grapes from a commercial vineyard.Samples of these wines were derived from the same musts as those used in previous studies [4][5][6].
In particular, we report here a comparison of aroma profiles of varietal wines (Huxelrebe, Ortega, Schönburger) fermented and aged in stainless steel tanks, and wines from the same must that were fermented and matured in secondhand oak casks.By using a short (~7 month) residence in secondhand (third or fourth fill) Burgundian Nevers oak casks, the winemaker's aim was to make wines with greater aroma complexity and earlier maturity than the standard tank wines, but without obvious oak aroma character, as it was felt that this would clash with the 'aromatic' aromas (Muscat or Gewürztraminer type) of the base wines.All the varieties here have either Muscat or Gewürztraminer as part of their ancestry [5].

Wine Samples
Huxelrebe, Ortega and Schönburger and Siegerrebe wines of the 2004 vintage were obtained as gifts from the commercial 22-acre Chilford Hundred vineyard (Cambridgeshire, UK).

Winemaking and Ageing
Processing of the Chilford grapes was conducted at the vineyard and the adjacent winery.Each variety was harvested, destalked, crushed and pressed separately, using a horizontal pneumatic press.The varietal musts were pumped into separate stainless steel tanks, sulphited (30-50 mg/L) and allowed to settle for 24 h.The clarified juices were then pumped to separate clean stainless steel tanks, whence samples (1 L) were taken of each and analysed by SBSE/gas chromatography-mass spectrometry (GC-MS) as described by Caven-Quantrill et al. [5] (see Figure 1).To ensure the start of fermentation, yeast (Lallemand EC1118, Montréal, QC, Canada) was added to each tank and fermentation was allowed to proceed for 24 h, after which the fermenting musts were transferred to either fresh tanks or barrels to finish fermentation at ca. 12-15 • C and to begin the ageing process (a total of 7 months for each wine, with no further racking).The bulk of the wines spent the time in stainless steel tanks, but small amounts were fermented and aged in secondhand Nevers oak casks (228 L) from Burgundy.Fermentation proceeded to dryness, all the wines having 1.5-3.8g/L of reducing sugars, and subsequent maturation took place under airlock at 10-15 • C. The free SO 2 level was maintained at 30-50 mg/L (ppm) and temperature was kept low in all cases in order to discourage malolactic fermentation (MLF).After 7 months, our sample wines (1 L of each) were taken from tank and cask and bottled.The bottles, stoppered with sterile conventional corks, were stored horizontally in a cool (ca.10-15 • C), dark place until opened for analysis nearly 2 years after the vintage.

Determination of Routine Viticultural and Oenological Parameters
Specific gravities of musts were determined using a wine hydrometer, with adjustment for temperature.Malic and tartaric acids in musts and wine were determined by reversed-phase highperformance liquid chromatography (HPLC), as described previously [18].D-and L-lactic acids in wine were determined by a previously reported chiral ligand-exchange HPLC method [18].Free SO2 levels were determined by the Ripper method [19].Determination of ethanol content (% v/v) was carried out using a standard gas chromatography-flame ionisation detection GC-FID procedure [20].Residual reducing sugar in the wines was assayed by the Lane and Eynon method [19].

Sampling Conditions for SBSE
Subsequent SBSE and thermal desorption gas chromatography/mass spectrometry TD-GC/MS analyses (triplicate analyses for each sample) were conducted as previously optimized [4] using a 20 mL sample volume and a 5:1 injection split ratio [17].
Prior to use, the stir bars were conditioned at 300 °C in a helium stream (100 mL/min) for 1 h using a TC-1 tube conditioner (Gerstel).A preconditioned Twister TM stir bar was added to each of the sample vials before being capped and placed onto a Gerstel Twister™ stirrer plate (TS-1).Samples were stirred under the previously optimised ambient temperature juice conditions (1000 rpm for 2 h [4].On completion, stir bars were removed from the vials, washed with 'super pure' water (5 mL) and blotted dry on a lint-free tissue.The stir bars were finally spiked directly with internal standard solution (n-tetradecane 0.02% w/v in acetone, 1 μL) then transferred to a clean preconditioned thermal desorption tube and placed onto a TDS-A autosampler for analysis.Addition of n-tetradecane directly to the stir-bar allowed it to be used for semi-quantifying the extracted volatile compounds, that is, it acts as both an extraction internal standard and a GC internal standard, since its recovery into the stir bar polydimethylsiloxane (PDMS) phase is 100% [4].

Instrumentation and Conditions
For an account of the method development and optimisation of the TD-GC/MS procedure, see references 4 and 17.The SBSE analyses were performed using an automated TDS-2/TDS-A thermal desorption unit (Gerstel) mounted on an Agilent 6890 gas chromatograph system coupled to a quadrupole Agilent 5973 electron ionisation (70 eV) mass spectrometric detector (Agilent Technologies, Palo Alto, CA, USA), equipped with an Agilent Innowax (crossed linked polyethylene glycol) capillary column (30 m × 0.25 mm i.d., 0.25 μm film thickness).The carrier gas was helium with a constant column flow rate of 1 mL/min (mean velocity 36 cm/s).
The analytes were cryofocused in a programmed temperature vaporising injector (PTV) (CIS-4, Gerstel) held at −50 °C with liquid nitrogen prior to injection.A packed liner containing 20 mg of Tenax TA was used in the PTV.Stir bars were thermally desorbed in a stream of helium carrier gas at a flow rate of 70 mL/min and by programming the TDS 2 from 20 °C to 300 °C (5 min) at a rate of

Determination of Routine Viticultural and Oenological Parameters
Specific gravities of musts were determined using a wine hydrometer, with adjustment for temperature.Malic and tartaric acids in musts and wine were determined by reversed-phase highperformance liquid chromatography (HPLC), as described previously [18].D-and L-lactic acids in wine were determined by a previously reported chiral ligand-exchange HPLC method [18].Free SO 2 levels were determined by the Ripper method [19].Determination of ethanol content (% v/v) was carried out using a standard gas chromatography-flame ionisation detection GC-FID procedure [20].Residual reducing sugar in the wines was assayed by the Lane and Eynon method [19].

Sampling Conditions for SBSE
Subsequent SBSE and thermal desorption gas chromatography/mass spectrometry TD-GC/MS analyses (triplicate analyses for each sample) were conducted as previously optimized [4] using a 20 mL sample volume and a 5:1 injection split ratio [17].
Prior to use, the stir bars were conditioned at 300 • C in a helium stream (100 mL/min) for 1 h using a TC-1 tube conditioner (Gerstel).A preconditioned Twister TM stir bar was added to each of the sample vials before being capped and placed onto a Gerstel Twister™ stirrer plate (TS-1).Samples were stirred under the previously optimised ambient temperature juice conditions (1000 rpm for 2 h [4].On completion, stir bars were removed from the vials, washed with 'super pure' water (5 mL) and blotted dry on a lint-free tissue.The stir bars were finally spiked directly with internal standard solution (n-tetradecane 0.02% w/v in acetone, 1 µL) then transferred to a clean preconditioned thermal desorption tube and placed onto a TDS-A autosampler for analysis.Addition of n-tetradecane directly to the stir-bar allowed it to be used for semi-quantifying the extracted volatile compounds, that is, it acts as both an extraction internal standard and a GC internal standard, since its recovery into the stir bar polydimethylsiloxane (PDMS) phase is 100% [4].

Instrumentation and Conditions
For an account of the method development and optimisation of the TD-GC/MS procedure, see references 4 and 17.The SBSE analyses were performed using an automated TDS-2/TDS-A thermal desorption unit (Gerstel) mounted on an Agilent 6890 gas chromatograph system coupled to a quadrupole Agilent 5973 electron ionisation (70 eV) mass spectrometric detector (Agilent Technologies, Palo Alto, CA, USA), equipped with an Agilent Innowax (crossed linked polyethylene glycol) capillary column (30 m × 0.25 mm i.d., 0.25 µm film thickness).The carrier gas was helium with a constant column flow rate of 1 mL/min (mean velocity 36 cm/s).
The analytes were cryofocused in a programmed temperature vaporising injector (PTV) (CIS-4, Gerstel) held at −50 • C with liquid nitrogen prior to injection.A packed liner containing 20 mg of Tenax TA was used in the PTV.Stir bars were thermally desorbed in a stream of helium carrier gas at a flow rate of 70 mL/min and by programming the TDS 2 from 20 • C to 300 • C (5 min) at a rate of 60 • C/min.After desorption and cryofocusing, the CIS-4 was programmed from −50 • C to 260 • C (20 min) at 12 • C/s to transfer the trapped aroma volatiles onto the analytical column.The TDS-2 was operated in the split mode to provide a 5:1 split ratio, as previously optimized [17].The GC oven temperature was programmed from 40 • C (5 min) to 240 • C (20 min) by increasing the temperature at 3 • C/min, and the MS was operated in scan mode (35-300 amu).The temperature of the mass selective detector transfer line was retained at 250 • C throughout.

Identification of Volatile Components and Determination of Semi-Quantitative Concentrations
Agilent MS Chemstation software, equipped with the Wiley 275L, Nist 98 and Frutarom's proprietary mass spectral libraries, was employed for identification of the volatile compounds, along with ethyl ester retention indices (as secondary confirmation), which were determined, whenever possible, by co-injection of authentic aroma standards under identical GC-MS conditions and calculated by linear interpolation relative to the retention times of C 2 -C 20 ethyl esters [21].Confirmation of the presence of oak lactones in barrel wines (and of their absence in tank wines) was achieved by selected ion chromatography (SIC) using the m/z = 99 ion.
Semi-quantitative concentrations were calculated using the expression (peak area/IS peak area) × IS concentration (IS = internal standard), and hence are only approximate values.However, our major interest in this work is with regard to differences in wine volatile component levels of varietal tank and corresponding barrel wines of a single vintage.

General
The SBSE/GC-MS method used in this work was optimised, as described previously, using a mixture of 46 pure volatile compounds of known concentrations in a synthetic grape must medium [4], and in a corresponding synthetic wine medium containing 12.0% ethanol by volume [17].By changing the GC injector split ratio from 20:1 for the synthetic must volatiles to 5:1 for the synthetic wine volatiles, it was possible to compensate for the decreased SBSE sensitivity caused by the presence of ethanol in the wine medium.In this way, the GC peak areas of the extracted volatiles were generally much more similar for the two media.Hence, it was possible to monitor changes in volatile component composition caused by the alcoholic fermentation of real grape must to wine by direct comparison of data for wines in this work with data for the corresponding musts, as reported in our previous reports [5,6], and as outlined in Figure 1.Briefly, compared with the original musts, the wines were characterized by much higher overall levels of volatiles, especially with regard to ethyl alkanoate and acetate esters, fatty acids, 2-phenylethanol and β-damascenone, whereas certain alcohols, carbonyl compounds and some terpenoids were more abundant in the musts.
The casks used for barrel fermentation/aging at Chilford Vineyard were secondhand Nevers Oak 228 litre 'pièces', that previously held up to three vintages worth of Burgundy Chardonnay wine.By using secondhand casks and a short maturation period (ca.seven months), the winemaker at Chilford Vineyard intended to produce more complex and more intense wines, but without pronounced oak aromas and flavours, because it was thought that these would clash with the natural strongly aromatic character of the wines.At the same time, the barrel wines would still benefit from typical accelerated maturation compared with the tank wines.
The level of free SO 2 (30-50 mg/L) in all the wines after primary fermentation, combined with the relatively low fermentation temperatures (12-18 • C) and maturation temperatures (10-15 • C), completely discouraged malolactic fermentation (MLF) [22] (Table 1), so the difference in aroma profiles of corresponding tank and barrel wines is due entirely to ageing for seven months in different containers.MLF is a secondary fermentation caused by resident or added bacteria (e.g., Lactobacillus, Oenococcus and/or Pediococcus spp.); it converts L-malic acid to the 'softer' L-lactic acid, but also produces changes to the aroma profile.The occurrence of MLF in some of the wines would have invalidated the comparisons made here.Typically, very low levels of the major MLF product L-lactic acid (25-44 mg/L) [23] (Table 1) confirm that none of the wines underwent even partial MLF.This is supported by the observation of (at best) only trace levels of diacetyl in the wines.The small depletions of malic acid observed (Table 1) are typical of those caused by ordinary yeast metabolism during alcoholic fermentation [23].MLF causes a significant reduction in the total acidity of wine and, hence, would be undesirable here, since the juice acidities (except Huxelrebe) were low-moderate, and in any case they were lowered naturally as a result of alcoholic fermentation (Table 1).
Analysis of all the wines was performed nearly two years after the vintage (consumers typically begin to drink English wines of about this age) over a short period of time.a All juices were 'fermented to dryness', having residual reducing sugar content of 1.5-3.8(±0.2) g/L.For all the wines made from these juices, malic acid levels dropped by 8-12% (±1%), which is normal for alcoholic fermentations using the S. cerevisiae EC1118 strain at 15-20 • C [23].Tartaric acid levels dropped by ca.10% (presumably as a result of precipitation of potassium hydrogen tartrate during and after alcoholic fermentation).L-Lactic acid levels were only 25-44 (±2) mg/L and D-lactic acid levels were 195-326 (±6) mg/L, thus indicating the absence of malolactic fermentation (MLF).The % ethanol by volume range was 11.7-12.5 (±0.2); b • Brix = sugar content in g/100 g juice.

Wine Volatiles
Since the semi-quantitative concentrations of volatiles quoted in Table 2 are only approximate, a comparison of these with their corresponding literature odour threshold values (OTV) will not give a definitive indication of extent of contribution to global wine aroma, especially as literature OTVs vary according to the medium.However, it is possible to make a tentative suggestion on the contribution of a particular wine volatile, especially if its semi-quantitative concentration is much higher than its OTV.Certainly, the suggestions here are in broad agreement with the informal organoleptic assessments made by a number of independent qualified or experienced tasters, as well as by the authors.The major wine volatiles are summarized next.

Alcohols
The 'fusel alcohols' 2-methyl-1-propanol (isobutyl alcohol) and 2-and 3-methyl-1-butanols, together with the normal straight chain volatiles such as 1-butanol and the benzenoid 2-phenylethanol, are present in all the wines and their combined concentrations are similar for tank and barrel wines.Individually, their concentrations are well below their literature OTVs, but their presence is likely to make a significant beneficial contribution to the background aroma [9,24,25].

Esters
All the wines, especially the barrel wines, had low levels of ethyl acetate and, since the OTV of this compound from wine is 1.60 × 10 8 ng/L, it is unlikely it played an individual role in organoleptic quality [26].
The C 5 and C 6 acetate esters are the most abundant acetate esters.The OTV of hexyl acetate is 2000-480,000 ng/L [27], so it is possible that this component contributed to the perceived pronounced fruity, green, apple/pear notes of the tank wines in particular.
From Table 2, certain ethyl esters were the most abundant of all the wine volatile compounds, the highest of which was ethyl octanoate, present in the concentration range 206,000 ng/L (Ortega barrel wine) to 370,000 ng/L (Huxelrebe wines).Since this ester's literature OTV is 5000-92,000 ng/L [27], it is likely to make a significant contribution to the overall organoleptic profile of all of the wines studied, providing a sweet odour reminiscent of apricot, banana and pineapple [28].
Another abundant ester is ethyl hexanoate, present in the concentration range 43,000 ng/L (Schönburger barrel wine) to 91,000 ng/L (Huxelrebe barrel wine).Again, because the literature OTV of this compound is reported to be 1000-36,000 ng/L [24] and 300-5000 ng/L [27], it is feasible that this ester contributed to the bouquet of all of the wines studied.The same could be said of ethyl decanoate (OTV 8000-12,000 ng/L [27].The higher homologues (ethyl palmitate, etc.) are all present at low levels, and since the only available literature OTV for these oily/waxy esters was for ethyl palmitate (2 × 10 6 ng/L), their contribution is unknown.

Fatty Acids
Other comparatively abundant common components of the wines were the long chain fatty acids, especially octanoic, decanoic, dodecanoic (lauric), hexadecanoic (palmitic), oleic and stearic acids.However, the levels of these compounds were all well below their reported odour threshold values, and so they will contribute only to background odour (and possibly to 'mouthfeel').

Terpenoids and Norisoprenoids
Forty eight volatile terpenoids and norisoprenoids were identified in the wines of this study and many are also common to grape juices [5,26].The most abundant terpenoid in grape musts was linalool [5], but fermentation and other winemaking processes appear to have reduced its concentrations to well below its literature OTVs (1500-50,000 ng/L) in all the wines studied [24,26].Likewise, all of the remaining terpenoids/norisoprenoids identified (apart from (Z)-rose oxide and (E)-β-damascenone) were also below their OTVs in the wines of this study, but it is likely that they provide an important synergistic background odour contribution [26].Recent examples include vineyard elevation [29], climatic or vintage variations [6,14,30], vine variety (cultivar) [5,15,[30][31][32] and the use of oak or oak products [9,10,13,[33][34][35].To our knowledge, this is the first report of an SBSE/GC-MS based comparison of aroma profiles of 'unoaked' white wines with corresponding wines that have had a limited oak contact.
Overall, Table 2 indicates that tank wines and their barrel equivalents have a similar range of volatile components, but with different concentration profiles and with net concentrations of volatiles being rather greater in barrel wines.In fact, ca.60% of the total concentration differences of volatile compounds observed in Table 2 are associated with higher levels in the barrel wines.Additionally, extra volatiles were detected in barrel wines (Table 2), including oak lactones.
Other than very low levels of oak lactones, typical oak volatile phenols (such as eugenol, guaiacol, 4-methylguaiacol and vanillin), as well as other oak volatiles (such as cyclotene, ethoxylactone and maltol) [22], were not detected in the barrel wines of this study and the low p-vinylphenol levels were similar in tank and barrel wines.This indicates that a short period of storage in used oak casks in the presence of lees fulfilled the winemaker's aim to produce more complex wines, but without obvious oak aroma and flavour.The typical oak aroma compounds mentioned above may largely have been removed from the wood in previous wines (in Burgundy) and/or may have been adsorbed as a result of prolonged contact with yeast: it is known that volatile phenols are selectively adsorbed by yeast lees [33].The greater flavour complexity of the barrel wines, as perceived by experienced tasters at informal tastings, may be partly due to the presence of non-volatile oak compounds at low levels: such compounds are known to alter human perception of wine aroma [26].
Some major volatile compositional differences between corresponding tank and barrel wines seen in Table 2 are discussed in the next paragraphs.
Alcohols.The 'fusel alcohols' are present at similar levels in tank and barrel wines (see Section 3.2), but 1-hexanol, responsible for grassy, green fruit notes, is more abundant in tank wines, thus possibly making a contribution to the perceived more 'green' aroma of these wines.
Esters.Ethyl acetate and other acetate esters (such as hexyl acetate, 2-phenethyl acetate and the C 5 acetates) are generally more abundant in tank wines and could well contribute to the overall simple fruity aroma ('light acetate ester notes') of the tank wines in particular, as described informally by several independent experienced and qualified tasters.Conversely, Table 2 shows that ethyl carboxylate esters (ethyl octanoate, hexanoate and others) are collectively more abundant in barrel wines, possibly contributing to the organoleptically perceived softer, more exotic fruity notes of these wines, compared with the corresponding tank wines.
Acetals.Table 2 shows that, although acetals are not unique to the barrel wines, they are more abundant in these wines.Concentrations of acetals in oak-aged wines have been linked to the type of wood and the aging of the wine in oxidative conditions [34].
Lactones.Oak lactones (β-methyl-γ-octalactone diastereoisomers) occur in many wines that have matured in oak, but are not found in wine that has had no contact with oak.Already present in oak wood, and also formed during toasting of the cask during cooperage, these compounds are extracted slowly into the wine.Consequently, it may take from six months to one year in new oak casks for the coconut-like fragrance of oak lactones to become apparent.In most situations, the total content of these lactones rises during oak maturation, and there is also selective extraction of the more aromatic cis-isomer [32].Continued use of casks (vintage after vintage) causes extensive leeching of volatile compounds from the woody material, so that later wines will have much lower levels of these compounds [35].Indeed, in the barrel wines of this study, these two compounds are both well below (≤108 ng/L) their OTVs-74,000 ng/L and 320,000 ng/L for the cis and trans isomers (respectively) as determined in red wine [13].

Conclusions
The SBSE/GC-MS method described here allowed the identification and semi-quantification of the volatile components (numbers in parentheses) of single vintage (2004), single varietal wines produced from Vitis vinifera cultivars Huxelrebe (tank wine: 158; barrel wine: 160), Schönburger (tank wine: 148; barrel wine: 158), Ortega (tank wine: 148; barrel wine: 155), and Siegerrebe wine (tank wine: 146).The results show that both greater numbers of volatiles, and also higher net levels of volatiles, were associated with the barrel wines.The major volatiles of wines were esters, with long chain fatty acid ethyl esters being generally more abundant in barrel wines and acetate esters being generally more abundant in tank wines.Long chain carboxylic acids, norisoprenoids and acetals were generally more abundant in barrel wines, and low levels of oak lactones-much below their odour threshold values (OTVs)-were found only in barrel wines.The use of a short (seven month) maturation time in used oak casks (third or fourth fill) gave more complex wines, but with no typical oak aroma, as perceived in informal tastings by experts.

Figure 1 .
Figure 1.Sample origin and relationships.The full scheme applies to the musts and wines of the varietals Huxelrebe, Ortega and Schönburger.Tank wine only was produced from the must of Siegerrebe.

Figure 1 .
Figure 1.Sample origin and relationships.The full scheme applies to the musts and wines of the varietals Huxelrebe, Ortega and Schönburger.Tank wine only was produced from the must of Siegerrebe.

a
Semi-quantitative concentrations (ng/L) calculated from (peak area/IS peak area) × IS concentration; b Retention indices (RI) on a 30 m × 0.25 mm × 0.25 µm Innowax column, relative to C 2 -C 20 ethyl esters [21]; c Identification based on retention index, comparison of mass spectra and co-injection with authentic compounds, except for components marked *, which were based on comparison of mass spectra only; d Coefficient of variation (% CV) estimated from the SBSE-GC-MS analyses of a single wine sample (in triplicate; n = 3); -Indicates compound not detected; tr.Indicates concentration of <10 ng/L.

3. 2 . 5 .
Comparison of Volatile Compounds Identified in Tank-and Barrel-Fermented/Aged Wines Over the years, it has been demonstrated by use of a variety of extraction/chromatographic techniques, that several viticultural/enological parameters can influence the aroma profile of wine.

Table 1 .
Summary of routine enological parameters a .