A Technical Alternative to Aging Sherry Wine Extraction of Volatile Compounds from Oak Chips in Model Solutions L. Moyano *1, M. Chaves 2, L. Zea 3 Department of Agricultural Chemistry, University of Cordoba Edificio Marie Curie, Campus de Rabanales, 1414 Córdoba (Spain) *1 qe1mocal@uco.es; 2 chmum79@hotmail.com; 3 qe1zecal@uco.es Abstract- The type of oak chips, its concentration and the contact time are variables that could influence the extraction of the oak compounds during the accelerated aging of the wines. The aim of the present work was to study the extraction of aroma compounds by synthetic wine from two different wood pieces, chips and cubes, used in two concentrations (4 g/l and 8 g/l). The results showed that furfural, 5-methylfurfural, 5- hydroxymethyfurfural (HMF), E- y Z-isomers of oak lactone, guaiacol, 4-ethylphenol, siringol and vanillin were the compounds mainly extracted from wood. These compounds contributed with, toasted wood,, and aromas. By adding the odorant activity values (OAVs) of the compounds showing similar aroma descriptor, the descriptor was the most intense one. In addition, significant differences were found between the two different types of wood pieces and concentrations. The synthetic wines in contact with oak chips (8 g/l) showed the highest odorant potency during all the contact time. Therefore, the selection of an oak chip type, of its concentration of use and of a contact time wine/wood should be based on the desired aging level of wine. Keywords- American Oak Chips; Volatile Compounds; Hydroalcoholic Extracts; Wine Aging I. INTRODUCTION The use of barrels for wine storage during aging is a centuries-old practice. However, these barrels are more than storage receptacles; many types of compounds are transferred from the wood to the product: ellagitannins, lactones, coumarins, polysaccharides, hydrocarbons and fatty acids, terpenes, norisoprenoids, steroids, carotenoids and furan compounds. Their positive impact on wine organoleptic properties has been recognized for some years, and it has been shown that contact with wood inside the barrels enhances wine flavors and aromas [1, 2]. Moreover, during barrel-aging, wine undergoes major changes, leading to increased stability [3]. Oak is the main wood used to make barrels, the oak species most commonly used are Quercus alba, also known as American oak, and Quercus petraea and Quercus robur. Oak wood is mainly composed of three large insoluble polymers, cellulose, hemicellulose, and lignin; but it also contains different lower molecular weight compounds, such as volatile and nonvolatile acids, sugars, steroids, terpenes, volatile phenols, lactones, etc. which can be extracted by wine or hydroalcoholic solvents [4]. Nonetheless, due to the heavy investment required in barrels, in equipment for stacking and shifting, in labor, security and maintenance, and in fixed assets, the traditional system is associated with high costs. In addition, as the barrel becomes older, it might become populated with undesirable microorganisms such us Brettanomyces, wich can produce sensorially significant concentrations of ethylphenols with their unpleasant medicinal and horsy aromas [5]. As a result, the winemaking industry is currently seeking alternatives to cask aging, which, without impairing wine quality, will allow some of these costs to be reduced. The addition of wooden pieces to wine to be aged in tanks or in used barrels appears to be a promising additional or alternative aging technique, in that it ensures rigorous technical control of the aging process and also has a positive effect on wine color and aromatic profile [3, 4, 6-12]. This practice was recently approved and legislated by the European Community (CE 2165/25 and CE 157/26), although this practice has been used in some countries such as Australia and South Africa for several years. A number of factors strongly influence the extraction of volatile compounds from the wood, including: the size and shape of the chips, the amount added, the degree of toasting and geographical origin of the wood, duration of contact with wine, the timing of chip addition, and possible oxygenation [13, 8, 14, 15]. For this reason, there are a great number of these kinds of products available on the market: chips, cubes, podwer, shavings, granulates, blocks or segments, up to staves [3, 7]. Different methods have been used to analyze oak wood, most of them based on extraction with organic solvents and subsequent chromatographic analysis of the extracts. Macerating of wood in hydroalcoholic solutions affords the advantage of simulating the extraction conditions occurring during aging of wine while avoiding interference by wine components. On the other hand, while the technical potential of oakchip aging techniques is considerable, the details of the technique used need to be carefully defined and optimized. A further advantage is that the oak-chip technique may enable large European wine producers such as Spain to gain access to new markets, by introducing practices common on other continents and thus competing on an equal footing. The objective of this study is: (i) to know the volatile composition of American oak chips and (ii) to evaluate the -116-
effect of different types and different doses of oak chips on the extraction of volatile compounds from the wood in model solutions. II. MATERIAL AND METHODS A. Wood Samples and Experimental Design Two different types of white oak (Quercus alba) wood fragments were tested, both supplied by Anatride Iberica, S.L (Zaragoza, Spain): chips and cubes, both medium toasted (the details of the toasting process are protected as a trade secret). Chips were irregular fragments measuring roughly 1 cm 1.5 cm.2 cm, and weighing.158 g; cubes measured around 1 cm 2 and weighed 1 g (Fig. 1). Fig. 1 Type of American oak chips: shavings and cubes Each type was tested at two dose-rates: 4 g/l and 8 g/l. Maceration studies were performed in triplicate in 1-L flasks, at a constant temperature of 25ºC, stirring daily. Flasks were then filled to 5 ml with a synthetic wine containing 18% v/v ethanol and 12 mg/l sulfur dioxide, at ph 3.3 (Fig. 2). Samples were collected 2 days, 15 days, 3 days and 9 days after the start of the experiments. Samples were taken in triplicate at each sampling time. Fig. 2 Wood chips in model solutions III. CHEMICALS Reference compounds of the identified volatile compounds and the internal standards were obtained from Sigma-Aldrich (Munich, Germany) and from Le nez du vin (Jean Lenoir, Provence, France). IV. COLOR MEASUREMENT The absorbance at 42 nm was measured on a 1 mm path length by using a Perking-Elmer Lambda 25 spectrophotometer (Fig. 3). This absorbance was considered the browning index. V. DETERMINATION OF THE PERCEPTION THRESHOLDS AND AROMA DESCRIPTORS A panel of thirty-three volunteer panelists of both sexes (thirteen women and twenty men) between 2 and 55 years old from the University of Cordoba participated in the study. Fig. 3 Perking-Elmer Lambda 25 spectro-photometer Thirteen judges of the above mentioned panel had previous experience in wine sensory evaluation. However, all judges were trained in preliminary sessions as described in Ferreira et al. [16]. Reference standards were presented (three per session). During the training, judges discussed about odour terms and modified it by eliminating terms they considered irrelevant or redundant and by adding terms their considered pertinent. The perception threshold is defined as the minor concentration of a substance capable of producing a detectable sensation at least for 5% of the members of a tasting panel. For the determination of this samples were prepared 3 min before the test, to allow time for the vapour pressure to reach equilibrium at ambient temperature. The odour substances (1 ml) are poured directly into the glass flasks containing a piece of cotton and were closed immediately. Judges evaluated three aroma compounds per session by direct method of smelling. The concentration levels of the odorant solutions used were prepared according to annex A (ISO 5496). Starting from the lowest concentration solution, the judges indicated the first solution with an odorant sensation different to the perceived in the control (18% v v ethanol water), according to the annex B (ISO 5496) standard. This sensation must be detected by at least 5% of the judges in a taste panel, as stated above. In addition, the judges were asked for the aroma descriptors, the responses were compiled for all nine-aroma compounds, and those odour descriptors cited by less than 15% of the panel were eliminated. Perception thresholds and odour descriptors are listed in Table I. VI. IDENTIFICATION AND QUANTIFICATION OF AROMA COMPOUNDS Each one of the nine aroma compounds analyzed was identified by means of its retention time, coeluted with a standard solution of commercial product and confirmed by mass spectrometry (Hewlett-Packard 5972 MSD, Palo Alto, CA, USA). Positive ion electron impact mass spectra were acquired in scan mode, with a range of m z 39 3, and scan rate of 1.6 scan s -1. The chromatographic column, injector and oven temperatures, carrier gas and its flow were the same that those used for the quantification, as described below. The volatile compounds were quantified by capillarycolumn gas chromatography after continuous extraction of 1 ml of sample with 1 ml of freon-11 for 24 h (Fig. 4). -117-
TABLE I ODOUR DESCRIPTORS AND THRESHOLDS OF THE AROMA COMPOUNDS IDENTIFIED IN AMERICAN OAK WOOD FRAGMENT Compound Odour Descriptors Threshold (mg/l) furfural burn, 15 5- methylfurfural burn, 16 HMF.9 E-oak lactone Z-oak lactone,,,,.122.35 guaiacol.2 4-ethylphenol spicy 15 vanillin.65 syringol 1.7 information about the yellow-brown color of a solution [17-19]. Absorbance values (at 42 nm) for synthetic wine in contact with each of the two fragment types (chips vs. cubes) at two dose rates (4 g/l vs. 8g/L) are shown in Figure 5. Fig. 4 Continuous extractor Previously, wine was adjusted to ph 3.5 and 5 ml of internal standard (3 mg L -1 of 2-octanol) was added. The freon extract containing the volatile compounds was concentrated to.2 ml in a Kuderna-Danish microconcentrator and 3 µl was injected into a Hewlett- Packard-589 series II gas chromatograph equipped with an HP-INNOWax fused silica capillary column (6 m x.32 mm ID,.25 µm film thickness), with a FID and a sniffing port connected by a flow splitter to the column exit. The oven temperature programmed was as follows: 5 min at 45 ºC, 1ºC min -1 ramp to 185 ºC and 3 min at 185 ºC. Injector and detector temperatures were 275 and 3 ºC, respectively. The carrier gas was helium at 7 kp and split 1:3. The quantification was made by using chromatographic response factors, calculated for each compound in relation to the internal standard, in standard solutions of commercial products supplied by Sigma- Aldrich. VII. RESULTS AND DISCUSSION Although color is not a direct contributor to wine flavour, it is a part of the overall perceived organoleptic impression. Color is the first attribute found in the tasting providing information of potential defects and virtues of a wine, its age and its evolution in time, all contributing to decide on acceptance or rejection. Compounds contributing to color are all phenolic compounds, and many complex reaction products are formed during wine-making and maturation. The commonly accepted analytical measurement of white wine browning is absorbance at 42 nm, which provides Fig. 5 Absorbance values at 42 nm for synthetic wine during maceration of oak fragments Over the first 3 days of contact, absorbance values varied very little, remaining around.1. At 9 days, however, a significant increase in values was recorded. The type of chip used influences the absorbance values thus obtained. Higher values were found for oak chips than for oak cubes, suggesting improved diffusion of colored compounds, mainly in the spectral range corresponding to yellow-brown using this type of chip. Furthermore, as was to be expected, the dose of wood fragment added influences absorbance values at 42 nm remain significantly above those of the synthetic wines with 8 g/l (chips and cubes), for each contact time. The organoleptically-significant volatile compounds released by oak chips into synthetic wine in the course of the study are shown in Table I. The compounds identified belonged to the following chemical families: furanes (furfural, 5-methylfurfural and 5-hydroxymethylfurfural), γ- octalactones (E- and Z-oak lactones), volatile phenols (guaiacol, 4-ethylphenol and syringol) and phenolic aldehydes (vanillin). These were associated with the descriptors burn,,,, spicy and. Each of these substances has a different perception threshold and, depending on their concentration in the wine and the sensitivity of each assessor to each compound, will contribute differently to aroma profile of the wine. The furanes and some volatile phenols have high perception thresholds, and in principle should therefore have a relatively low sensory impact; however, some of them are known to modify considerably the perception of a number of descriptors. Furfural, for example, is reported to display a marked degree of synergy with oak lactones, and at certain concentrations enhances the odor [2, 21]. Variations in the levels of extracted volatile compounds in synthetic wines are shown in Table II. The furane family were the most common volatile compounds. Furfural displayed the highest levels throughout the study, in agreement with the results reported by other authors [11], and followed by vanillin, the only quantified phenolic aldehyde. Furfural and 5-methylfurfural are generated by the Maillard -118-
reaction from cellulose and hemicellulose from the wood during the toasting stage, being responsible for the characteristic aroma of s and toasted s [2, 22]. Vanillin is formed by the thermal degradation of lignin during oak toasting, althougt it is also found naturallyt in green wood [5]. The concentration of this compound was higher in solutions macerated with oak chips than with oak cubes, though levels never exceeded 1 mg/l. Some authors note that the chip size can influence the formation of vanillin during toasting of wood, so that the parts of smaller dimensions, like the chips, are more combustible and hence can be formed more vanillin [7]. For both types of oak fragments (chips and cubes) shows a general increase for vanillin extracted up to 9 days of contact, obtaining values above 7 µg/l and 375 µg/l, respectively. The γ-octalactones were presented in all the samples in relative high concentrations. However, the behavior of γ- octalactones varied: whilst the E- isomer was present in higher concentrations in test solutions containing oak cubes, the Z-isomer displayed higher levels in those containing oak chips. The extraction lactones have been described as being a very fast process and to follow a potencial curve [23]. Usually, it is considered that the isomers of oak lactone are largely responsible for maderizado character of wines, with notes reminiscent mainly of, and toasted oak. Finally, syringol was the only volatile phenol to attain concentrations approaching.1 mg/l after 9 days, with no significant differences between oak-cube and oak-chip solutions. The contribution of a volatile compound to the aromatic profile can be evaluated qualitatively by means of its aroma descriptor and by its odor activity value (OAV). The OAV is obtained by dividing the concentration of each volatile compound (Table II) by its perception threshold (Table I). An aromatic notes or aromatic series are defined as a group of volatile compound with similar aroma descriptor. The value of an aromatic note is obtained as the sum of the OAVs of the compounds that integrate it. According to its aroma descriptors, a volatile compound can be included in one or several aromatic notes. The most representative aroma notes were,,, and. Polygonal graphs constructed using the OAVs for each aroma, chip type and chip dose-rate over the contact period are shown in Fig. 6. TABLE II CONCENTRATION (µg/l) OF AROMA COMPOUNDS IN THE MODEL SOLUTIONS DURING MACERATION OF OAK FRAGMENTS Contact time Compounds 2 days 15 days 3 days 9 days furfural dosage chips cubes chips cubes chips cubes chips cubes 4 g/l 424±79 446±44 621±145 57±26 84±215 1239±35 2377±254 957±63 8 g/l 83±51 658±116 934±58 213±164 1384±263 232±29 9 2519±66 1756±173 4 g/l 65.±1 n.d. 88.±14 22.6±4. 91.6±13 37.±6. 289±6.1 48.6±3. 5- methylfurfural 8 g/l 148±17 42.6±4.5 153±21 85.6±26 157±31 311±33 386±27 343±34 HMF E- oak lactone Z-oak lactone guaiacol 4-ethylphenol syringol vanillin 4 g/l 69.±9. 3.±6. 9.±9.5 44.6±12 95.6±1 41.±4.3 292±4.5 7.6±1 8 g/l 34±37 5.3±11 318±28 81.6±3. 331±53 91.6±3.7 48±4.1 142±47 4 g/l 1.1±3.6 9.1±2. 15.6±2. 82.3±11 28.2±6. 75.3±6. 32.6±4.5 83.±2. 8 g/l 16.±5.2 132±45 26.3±5.7 133±31 54.3±1.5 155±2. 52.3±4.1 16±2.1 4 g/l 38.6±4.1 12.±2. 74.6±11 22.6±5. 82.6±11 5.6±4.5 124±13 75.3±8. 8 g/l 77.1±1 22.±2. 12±31 46.3±8.5 15±18 8.±7. 361±38 1±7.5 4 g/l n.d. n.d. n.d. 6.33±.57 n.d. 7.33±.5 7 11.3±3. 14.±1. 8 g/l 47.3±5.1 n.d. 56.±3.6 8.±1. 66.6±4.6 15.3±3. 65.6±4.9 32.6±3. 4 g/l n.d. n.d. n.d. n.d. n.d. 11.3±.5 8 n.d. 17.±4. 8 g/l n.d. n.d. n.d. 7.33±.57 n.d. 12.6±2.5 n.d. 39.3±14 4 g/l n.d. 17.3±4.9 38.±2.6 36.±1.7 49.3±3.7 3.±7.8 8.±8.7 66.6±9.5 8 g/l n.d. 23.3±3.2 87.3±3.5 43.6±7.5 95.±6. 39.3±1.5 11±7.3 12±7. 4 g/l 141±16 62.6±1.5 241±3.5 11±7.7 366±1 137±6.5 73±149 375±39 8 g/l 159±27 118±2.6 119±5. 135±8.3 731±19 174±.57 888±185 569±82 n.d. = not detected; HMF=5-hydroxymethylfurfural -119-
2 days 8 6 4 2 15 days 8 6 4 2 3 days 2 15 1 5 9 days 3 2 1 Fig. 6 Polygonal graphs constructed using the OAVs for each aroma, chip type and chip dose-rate over the contact period Vanilla notes were particularly marked throughout the study, due primarily to the contribution of vanillin and to a lesser extent of oak lactones. This aroma, characteristic of aged wines and wines stored in oak casks, arises largely from the thermal degradation of lignin during toasting [5, 24]. OAVs for synthetic wines macerated with oak chips were higher than those recorded for wines macerated with oak cubes. Coconut and toasted-wood aromas were mainly attributable to oak lactones, reportedly attributable to the thermal degradation of lipids in oak wood during toasting [1]. The note was due largely to 5- hydroxymethylfurfural, apparently released by the wood polysaccharides cellulose and hemicellulose [12, 25]. All these notes had similar OAV values, though again they were higher in solutions macerated with oak chips than in those macerated with cubes; differences were more marked at the higher dose-rate (8 g/l). The aroma only displayed odor activity (OAV > 1) in synthetic wines macerated with 8g/L of oak chips. It was due mainly to the contribution of guaiacol, also arising from lignin degradation during the wood-toasting process [23, 26, 27]. Like the other aromas, the descriptor displayed higher values in wines macerated with chips than with cubes. In summary, GC/MS analysis of hydroalcoholic extracts of American oak wood allows the characterization of different wood samples (type and dose) using OAV data. The results discussed confirm the existence of quantitative and sensory differences in the aroma extraction process depending on the surface area of the oak chip used and on the dose-rate employed. The greatest differences were observed for synthetic wines macerated with chips rather than cubes, and at 8 g/l rather than 4 g/l. However, neither the contact time nor the other conditions tested here should be ruled out: depending on the type of wine to be aged, other combinations may be of considerable interest in terms of intensifying a given aroma. Finally, it is interesting to note that some of the extracted components present important sensorial notes, and could contribute to the final flavour of wines aged in wood. For this reason, their analysis in model solutions can be used as a first step towards their determination in wines. REFERENCES [1] P. Chatonnet, Influence des procédés de tonnellerie et des conditions d'élevage sur la composition et la qualitédes vins élevés en fûts de chêne, Thèsede Doctoract, Université de Bordeaux II, 1995. [2] P. Dubois, Apports du fut de cheˆne neuf a` l arome des vins, Rev. Franç. Oenol., Vol. 12, pp. 19-24, 1989. [3] M. Del Álamo, and I. Nevares, Wine aging in bottle from artificial systems (staves and chips) and oak woods: Antocyanin composition, Anal. Chim. Acta, Vol. 563, pp. 255-263, 26. [4] M. S. Pérez-Coello, M. A. González-Viñas, E. García- Romero, M. D. Cabezudo, and J. Sanz, Chemical and sensory changes in white wines fermented in the presence of oak chips, Int. J. Food Sci. Technol., Vol. 35, pp. 23-32, 2. -12-
[5] S. Weeks, and M. A. Sefton, Analysis of oak-derived wine flavours, Wine Ind. J., Vol. 14, pp. 42-43, 1999. [6] J. Béteau, and G. Roig, Los chips de roble como herramienta de vinificación y crianza, ACE Rev. Enol., Vol. 73, pp.1-9, 26. [7] M. Del Álamo, I. Nevares, L. Gallego, C. Martín, and S. Merino, Aging markers from bottled red wine aged with chips, staves and barrels, Anal. Chim. Acta, Vol. 62, pp. 86-99, 28. [8] E. Guchu, M. C. Díaz-Maroto, M. S. Pérez-Coello, M. A. González-Viñas, and M. D. Cabezudo, Volatile composition and sensory characteristics of Chardonnay wines treated with American and Hungarian oak chips, Food Chem., Vol. 99, pp. 35-359, 26. [9] L. Monedero, M. Olalla, M. Villalón, H. López-García, and M. C. López, Standardization of the chromatic characteristics of sobretablas wine macerates obtained by an accelerated ageing technique using heating and oak shavings, Food Chem., Vol. 69, pp. 47-54, 2. [1] N. Natali, F. Chinnici, and C. Riponi, Characterization of volatiles in extracts from oak chips obtained by Accelerated Solvent Extraction (ASE), J. Agric. Food Chem., Vol. 54, pp. 819-8198, 26. [11] M. S. Pérez-Coello, J. Sanz, and M. D. Cabezudo, Determination of volatile compounds in hydroalcoholic extracts of French and American oak wood, Am. J. Enol. Vitic.., Vol. 5, pp. 162-165, 1999. [12] J. J. Rodríguez-Bencomo, M. Ortegas-Heras, S. Pérez- Magariño, C. González-Huerta, and M. L. González-San José, Importance of chip selection and elaboration process on the aromatic composition of finished wines, J. Agric. Food Chem., Vol. 56, pp. 512-5111, 28. [13] R. Bozalongo, J. D. Carrillo, M. A. F. Torroba, and M. T. Analysis of French and American oak chips with different toasting degrees by headspace solid-phase microextractiongas chromatography-mass spectrometry, J. Chrom., vol. 1173, pp. 1-17, 27. [14] B. Fernández de Simón, and E. Cadahía, Tratamiento de roble para tonelería, Rev. Enol., Vol. 5, 27. [15] M. T. Frangipane, D. D. Santis, and A. Ceccarelli, Influence of oak woods of different geographical origins on quality of wines aged in barriques and using oak chips, Food Chem., Vol. 13, pp. 46-54, 27. [16] V. Ferreira, M. Pet ka, J. Aznar, and J. Cacho, "Quantitative gas chromatography-olfactometry. Analytical characteristics of a panel of judges usin a simple quantitaive scale as gas chromatography detector", J. Chromatogr. A, Vol. 12, pp. 169-178, 23. [17] V. L. Singleton and T. Kramling, Browning of white wine and accelerated test for browning capacity, Am. J. Enol. Vitic., Vol. 27, pp. 157-16, 1976. [18] T. C. Somers and G. Ziemelis, Spectral evaluation of total phenolic components in Vitis vinifera: grapes and wines, J. Sci. Food Agric., Vol. 36, pp. 1275-1284, 1985. [19] B. W. Zoeccklein, K. C. Fugelsang, B. H. Gump, and F. S. Nury, Wine analysis and production, Chapman and Hall: New York, PP. 146-151, 1995. [2] J. Cacho, Evolución del perfil volátil del vino tinto durante su crianza en barricas de roble, ACE Rev. Enol., pp. 1-12, 26. [21] G. Reazin, Chemical mechanisms of whiskey maturation, Am. J. Enol. Vitic., Vol. 32, pp. 283-289, 1981. [22] F. Zamora, Elaboración y crianza del vino tinto: Aspectos científicos y prácticos, Ed. Mundi-Prensa, Madrid, 23. [23] E. Gómez-Plaza, L. J. Pérez-Prieto, J. I. Fernandez- Fernandez, and J. M. López-Roza, The effect of successive uses of oak barrels on the extraction of oak-related volatilecompounds from wine", Int. J. Food Sci. Technol., Vol. 39, pp. 169-178, 24. [24] P. Arapitsas, A, Antonopoulos, E. Stefanou, and V. G. Dourtoglou, Artificial aging of wines using oak chips, Food Chem., Vol. 86, pp. 563-57, 24. [25] P. Chatonnet, I. Cutzach, M. Pons, and D. Dubourdieu, Monitoring toasting intensity of barrels by chromatographic analysis of volatile compounds from toasted oak wood, J. Agric. Food Chem, Vol. 47, pp. 431-4318, 1999. [26] V. L. Singleton, Maturation of wines and spirits: comparisons, facts, and hipotheses, Am. J. Enol. Vitic., Vol. 46, pp. 98-115, 1995. [27] R. Suarez, J. A. Suarez-Lepe, A. Morata, and F. Calderon, The production of ethylphenols in wine by yeasts of the genera Brettanomyces and Dekkera: A review, Food Chem., Vol. 12, pp. 1-21, 27. -121-