SEARCH

SEARCH BY CITATION

Keywords:

  • volatile benzene compounds;
  • wine and spirits aging;
  • barrel wood;
  • GC/MS;
  • PICI;
  • MS/MS

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

Extracts from acacia, chestnut, cherry, mulberry, and oak wood, used in making barrels for aging wine and spirits were studied by GC/MS positive ion chemical ionization (PICI). Wood chips were extracted by a 50% water/ethanol solution and a tartrate buffer pH 3.2–12% ethanol (model wine) solution. The principal compounds identified in extracts were guaiacol-containing aldehydes and alcohols, such as benzaldehyde and derivatives, vanillin and syringaldehyde, cinnamaldehyde and coniferaldehyde, eugenol and methoxyeugenol, guaiacol and methoxyguaiacol derivatives. PICI using methane as reagent gas produced a high yield of the protonated molecular ion of volatile phenols, compound identification was confirmed by collision-induced-dissociation (CID) experiments on [M + H]+ species. MS/MS fragmentation patterns were studied with standard compounds: guaiacol-containing molecules were characterized by neutral methyl and methanol losses, benzaldehyde derivatives by CO loss. Acacia wood extracts contained significant syringaldehyde and anisaldehyde, but no eugenol and methoxyeugenol. Significant syringaldehyde, eugenol and methoxyeugenol, and high vanillin were found in chestnut and oak wood extracts; low presence of volatile benzene compounds was found in mulberry wood extracts. Cherry wood extracts were characterized by the presence of several benzaldehyde derivatives and high trimethoxyphenol. Copyright © 2007 John Wiley & Sons, Ltd.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

Aging wine and spirits in wooden barrels is a process widely used in wine making. This process is finalized to stabilize the color, to improve limpidity, and to enrich the sensorial characteristics of the product. Oxygen permeation through the wood promotes redox processes, formation of new anthocyanin derivatives and tannins with consequent stabilization of color in red wines and loss of astringency.1 Aging in barrels is also used to produce spirits such as armagnac, whisky, brandy, and grappa.2–6 Several wood compounds transfer from wood to the product: ellagitannins, lactones, coumarins, polysaccharides, hydrocarbons and fatty acids, volatile phenols and benzene, aldehydes, terpenes, norisoprenoids, steroids, carotenoids, furan compounds.7–17 Furan and pyran derivatives, and other compounds characterized by toasty caramel aroma, are formed as a consequence of heat treatment of wood made in barrel making.18, 19 The principal wood compounds characterized by sensorial proprieties are vanillin (vanilla note; sensory threshold 0.02–0.2 ppm) and eugenol (clove, spicy; threshold 10 ppm).20

Oak is the main material used in making barrels destined to enology, but also chestnut and cherry, more rarely acacia and mulberry, are used.21

In a previous study, by performing positive ion chemical ionization (PICI) MS analysis of volatile benzene aldehydes and phenols using methane as reagent gas, a high yield of protonated molecular ions was observed.22 In the present study, volatile benzene compounds released from wood were studied by GC/PICI-MS in acacia, chestnut, cherry, mulberry and oak extracts. Wood chips were extracted by both a model spirit (50% ethanol) and a model wine (tartrate buffer pH 3.2–12% ethanol) solution. Identification of compounds was confirmed on the fragmentation patterns produced by ion trap collision-induced-dissociation (CID) and MS/MS experiments on [M + H]+ species.

EXPERIMENTAL

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

Samples and standards

Acacia, chestnut, cherry, and mulberry wood samples were from trees grown in northern Italy, and oak wood sample from France. Samples were provided as staves for making barrels by Veneta Botti s.r.l. (Conegliano, Veneto, Italy). Woods were naturally seasoned for 24–36 months and not subjected to any toasting treatments. Staves were reduced to chips of 0.5–2 × 0.5 cm in size and 1 mm thick. Syringaldehyde, trans-3,5-dimethoxy-4-hydroxycinnamaldehyde (sinapinaldehyde), eugenol, 4-allyl-2,6-dimethoxyphenol (methoxyeugenol), were purchased from Sigma-Aldrich (Milan, Italy); benzaldehyde, coniferyl alcohol, vanillin, 4-hydroxybenzaldehyde were purchased from Fluka (Milan, Italy); cinnamaldehyde, anisaldehyde, guaiacol and 1-heptanol were purchased from Carlo Erba Reagenti (Milan, Italy). Coniferaldehyde was synthesized by oxidation of coniferyl alcohol: 0.4 mg were suspended in 200 µl of H2SO4 6 M solution and added 100 µl of K2Cr2O7 0.01 M aqueous solution. The reaction was carried out for 5 min at room temperature, then the mixture was extracted with 2 ml of dichloromethane. The organic phase was dried with sodium sulfate and filtered, GC/MS PICI analysis of extract was performed in SCAN mode. The coniferaldehyde/coniferyl alcohol peak area ratio was 0.22.

Extraction

Thirty grams of chips were extracted with 500 ml of 50% water/ethanol (v/v) solution for 6 days in the dark at room temperature and kept under stirring on the last day. Extraction of others 30 g of chips was repeated with 500 ml of tartrate buffer pH 3.2–12% ethanol (v/v) model wine solution for 30 days in the dark at room temperature (static conditions). After extraction, solutions were filtered and placed in refrigerated storage. All extractions were repeated twice.

Samples preparation for analysis

Fifty millilitres of 50% hydroalcoholic extracts were reduced to 30 ml by rotavapor at room temperature to reduce the ethanol content, a 200 µl volume of a 1-heptanol 185 mg/l solution was added as internal standard (IS). The solution was transferred to a 100-ml volumetric flask, the volume adjusted by water and the sample centrifuged at 4000 rpm for 10 min. Fifty millilitres of solution were passed through a 1-g C18 Sep-Pak cartridge previously washed with dichloromethane and activated by the passage of methanol and water. After the sample was passed through, the cartridge was washed with 10 ml of water (to remove salts and more polar compounds) and analytes were recovered with 6 ml of dichloromethane. The organic phase was dried with sodium sulphate and filtered. Before analysis the sample volume was reduced to 500 µl under nitrogen flow. For analysis of model wine extracts, a volume of 30 ml sample was adjusted to 90 ml by water, a volume of 200 µl 1-heptanol 185 mg/l solution was added, and 45 ml of solution were passed through a 1-g C18 Sep-Pak cartridge as described above. Before analysis, the sample volume was reduced to 300 µl under nitrogen flow. Two samples of each extract were prepared for analysis.

Quantitative analysis

Calibration curves were calculated by analysis of four 50% ethanol aqueous standard solutions at concentration between 0.005–5 mg/l (0.1–100 µg/g wood). A volume of 50 ml solution was added of 200 µl of 1-heptanol 185 mg/l (IS) and sample preparation was performed as described for 50% of the ethanol extracts. Analysis of each standard solution was repeated twice, mean data were used to calculate calibration curves as (analyte concentration)/(IS concentration) versus (area of [M + H]+ signal)/(area of IS signal at m/z 55). Calibration curves: anisaldehyde Am/z137/Am/z55 = 14.754 · CST/CIS(R2 = 0.9992); hydroxybenzaldehyde Am/z123/Am/z55 = 0.622 · CST/CIS(R2 = 0.9999); cinnamaldehyde Am/z133/Am/z55 = 9.097 · CST/CIS(R2 = 0.9995); eugenol Am/z165/Am/z55 = 12.559 · CST/CIS(R2 = 0.9999); methoxyeugenol Am/z195/Am/z55 = 13.785 · CST/CIS(R2 = 0.9998); vanillin Am/z153/Am/z55 = 2.350 · CST/CIS(R2 = 0.9998); syringaldehyde Am/z183/Am/z55 = 3.678 · CST/CIS(R2 = 0.9993); guaiacol Am/z125/Am/z55 = 2.746 · CST/CIS(R2 = 1). Quantitative analysis was performed on the signal of [M + H]+ ion of compound from the total ion current (TIC), data are expressed as µg/g wood.

GC/MS analysis

Analyses were performed with a ThermoFinnigan TraceGC gas chromatograph (Rodano, Italy) equipped with an HP Innowax fused silica capillary column (30 m × 0.25 mm i.d.; df = 0.25 µm) (Agilent Technologies Italia S.p.A., Cernusco sul Naviglio, Italy) and coupled with a ThermoFinnigan PolarisQ Ion Trap mass spectrometer (Austin, TX, USA) operating in PICI mode using methane as reagent gas (reagent gas flow: 0.8 ml/min). Ion source temperature: 200 °C; GC injection port: 240 °C; volume injected: 1 µl splitless injection. Programmed oven temperature: 3 min at 70 °C, 2 °C/min to 160 °C, 3 °C/min to 230 °C, 25 min at 230 °C. Transfer line temperature: 280 °C; carrier gas: He; flow mode: constant flow 1.3 ml/min. The ion trap was operating in normal scan mode with m/z 40–550 scan range. To confirm identification of compounds, CID experiments on the [M + H]+ ions were performed using helium as collisional gas (dumping gas flow: 0.3 ml/min) at an excitation voltage of 225 mV. The principal fragments produced by CID experiments are reported in Table 1.

Table 1. Principal fragments produced by CID experiments on the [M + H]+ ions using helium as collisional gas (dumping gas flow: 0.3 ml/min; excitation voltage: 225 mV). b.p.: base peak of fragmentation spectrum
CompoundPrecursor ion m/z [M + H]+MS/MS m/z fragment ions (abundance > 5%)
Benzaldehyde10779 (b.p.)
Methylbenzaldehyde12193;43 (b.p.)
Hydroxybenzaldehyde12395 (b.p.);91;81
Guaiacol125110 (b.p.);96;93;91;65
Cinnamaldehyde133115 (b.p.);105;91;79;55
Anisaldehyde137122;109 (b.p.);94
Vinylguaiacol151136;123;119 (b.p.); 115;95;91;81
Vanillin153138;125 (b.p.);93
2,6-dimethoxyphenol155140 (b.p.);123;95;91;65
Eugenol165150 (b.p.);137;133;105
Trimethoxybenzene169154;138 (b.p.);126
Coniferaldehyde179164;161;147 (b.p.); 133;119;105;55
Coniferyl alcohol181166 (b.p.);153;138
Syringaldehyde183168;155 (b.p.);140;123;95
Trimethoxyphenol185170;153 (b.p.);125
Methoxyeugenol195180 (b.p.);167;163;135;107
trans sinapinaldehyde209194;191;177 (b.p.); 149;121;107;93

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

GC/PICI-MS and MS/MS characterization of volatile phenols and benzene aldehydes

PICI using methane as reagent gas produced a high yield of protonated molecular ions for all the volatile phenols studied. The reconstructed ion chromatogram of [M + H]+ species (oak extract analysis) is showed in Fig. 1. Fragmentation patters from ion trap CID experiments on [M + H]+ species of standards were studied and compared with those of compounds in extracts. Scheme 1 shows MS/MS fragmentation patterns of cinnamaldehyde (5), coniferaldehyde (6) and sinapinaldehyde (17). [M + H]+ species showed methyl and water losses, from coniferaldehyde and sinapinaldehyde ion at m/z 147 and 177, respectively, formed by methanol loss as principal fragmentations (Table 1). For coniferaldehyde and sinapaldehyde, ions at m/z 105 and 107, respectively, were observed. Cinnamaldehyde also showed formation of an ion at m/z 91, probably due to chetene molecule loss. MS3 and MS4 experiments on the m/z 147 and 177 ions, and then on the m/z 119 and 149 ions produced from the first experiment, revealed two consecutive losses of 28-Da fragments, probably corresponding to CO and C2H4 residues. To characterize benzaldehyde derivatives no methylbenzaldehyde standard was used. Fragmentation of [M + H]+ species showed CO loss, and anisaldehyde also showed a methyl loss. Fragmentation patterns of guaiacol and of monomethoxy- and dimethoxy-guaiacol derivatives were characterized without standards. As in cinnamaldehyde and its derivatives, methyl and methanol losses were observed; in MS3 experiments on ions at m/z 93, 123 and 153, all three compounds lost CO. CID experiments on [M + H]+ ion of vanillin and syringaldehyde showed methyl and methanol losses, and the CO loss as principal fragmentation. Lastly, Scheme 2 shows MS/MS fragmentation patterns of eugenol (8) and methoxyeugenol (9). Losses of methanol and a vinyl group, and the 15-Da loss as principal fragmentation, were observed. MS3 experiments on ions at m/z 137 and m/z 167 produced the species at m/z 105 and 107, respectively.

thumbnail image

Scheme 1. Fragmentation patterns and principal ions produced by CID and MS/MS experiments on the [M + H]+ ion of cinnamaldehyde (5), coniferaldehyde (6) and sinapinaldehyde (17).

Download figure to PowerPoint

thumbnail image

Scheme 2. Fragmentation patterns and principal ions produced by CID and MS/MS experiments on the [M + H]+ ion of eugenol (8) and methoxyeugenol (9).

Download figure to PowerPoint

thumbnail image

Figure 1. PICI-reconstructed ion chromatogram of [M + H]+ signals from the 50%-ethanol oak wood extract analysis. Principal benzene compounds identified: 3 anisaldehyde (m/z 137); 8 eugenol (m/z 165); 10 vinylguaiacol (m/z 151); 11 vanillin (m/z 153); 9 methoxyeugenol (m/z 195); 12 syringaldehyde (m/z 183); 16 trimethoxyphenol (m/z 185); 6 coniferaldehyde (m/z 179). i.s.: internal standard 1-heptanol (m/z 55).

Download figure to PowerPoint

Study of wood extracts

As the efficiency of 50%-ethanol aqueous solutions in wood extraction has been reported to be good, this was used as a model spirit solution.23 Sixteen benzoic and cinnamic volatile compounds were identified in extracts. Quantitative data for the 50%-ethanol extracts analyses are listed in Table 2 and expressed as µg/g wood. Sinapinaldehyde was also identified, but, due to its low volatility, reliable quantification was not possible.

Table 2. Benzene volatile compounds determined in 50%-ethanol solution wood extracts. Compounds are quantified on the signal of [M + H]+ ion
Compoundm/z[M + H]+µg/g wood
acaciachestnutcherrymulberryoak
  • a

    quantified with anisaldehyde calibration curve.

  • b

    quantified with cinnamaldehyde calibration curve.

  • c

    quantified with guaiacol calibration curve.

1 Benzaldehydea1072.25
2 Methylbenzaldehydea1211.05
3 Anisaldehyde1370.810.050.01
4 Hydroxybenzaldehyde1231.131.120.330.59
5 Cinnamaldehyde1330.010.05
6 Coniferaldehydeb1790.170.250.030.250.28
7 Coniferyl alcoholb1810.02
8 Eugenol1650.732.01
9 Methoxyeugenol1950.200.01
10 Vinylguaiacolc1510.07
11 Vanillin1531.655.150.130.051.96
12 Syringaldehyde18310.304.230.370.539.25
13 Trimethoxybenzenec1690.29
14 Guaiacol1250.04
15 2,6-dimethoxyphenolc1550.110.04
16 Trimethoxyphenolc1850.3429.942.070.20

A characteristic profile was observed for each wood extract: acacia contained significant benzoic aldehydes, particularly, vanillin and syringaldehyde, but no eugenol or methoxyeugenol. In chestnut and oak extracts, the latter two compounds were noteworthy, and high vanillin was found. Cherry wood extracts were characterized by the presence of several benzoic aldehydes, although not in high abundance, a high level of trimethoxyphenol was revealed in hydroalcohol solution. Oak wood extracts had the highest qualitative and quantitative richness in volatile benzene compounds, and mulberry the lowest.

The presence of vanillin and eugenol was followed by that of methoxy derivative, with high vanillin and syringaldehyde in acacia, chestnut and oak, and considerable eugenol and methoxyeugenol in chestnut and oak. Data in Table 2 evidence the greater presence in extracts of benzoic aldehydes with respect to cinnamic ones that is characteristic of untoasted woods.

The attitude of wood in releasing volatile phenols into wine was studied by comparing quantitative data of 50%-ethanol aqueous extracts with those of the model wine solution extracts. The data shown in Table 3 are percentual recoveries of model wine solution with respect to model spirit solution, calculated for the four principal compounds. The higher efficiency of the model spirit solution in coniferaldehyde and eugenol extraction is evident; instead the model wine solution (more polar and with lower pH) enhanced the recovery of vanillin. The different percentual recoveries in the different samples observed in particular for vanillin and syringaldehyde by using the same extracting solution are probably due to the different porosity of the wood. Both solutions showed similar efficiency in vanillin and syringaldehyde extraction from acacia, oak and chestnut. The higher percentage of vanillin extraction from cherry and mulberry of the model wine solution was probably affected by the low content of compound in wood. Higher eugenol and coniferaldehyde were extracted from the model spirit solution.

Table 3. Percentual recoveries of model wine solution with respect to 50%-ethanol model spirit solution calculated for the four principal compounds identified in extracts
CompoundRecovery by model wine solution (%)
acaciachestnutcherrymulberryoak
  1. n.f., not found.

Coniferaldehyde3140702733
Eugenoln.f.66n.f.n.f.37
Vanillin126122224226101
Syringaldehyde8110219211967

CONCLUSIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

In general, CI using methane as a reagent gas showed a high yield of [M + H]+ species of volatile phenols studied (signal of [M + H]+ was the base peak of mass spectra). MS/MS experiments on the [M + H]+ ion of molecules with a guaiacol residue showed losses of methyl and methanol. Benzaldehyde and its derivatives were characterized by loss of CO.

Chestnut and oak wood released significant amounts of eugenol and methoxyeugenol into wine; high vanillin was released from chestnut, and high syringaldehyde from acacia and oak. Cherry was characterized by methoxyphenols and high trimethoxyphenol. On the basis of compounds identified in the wood and the different extraction efficiency of two solutions studied, it is inferred that chestnut, oak and acacia barrels (for a high vanillin recovery) and eugenol-containing wood barrels (chestnut, oak) are recommended for aging spirits. Cherry and mulberry are more suitable for aging wine.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES

This research is supported by Veneta Botti s.r.l. and Confraternita del Raboso Piave.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. EXPERIMENTAL
  5. RESULTS AND DISCUSSION
  6. CONCLUSIONS
  7. Acknowledgements
  8. REFERENCES
  • 1
    Ribéreau-Gayon P, Glories Y, Maujean A. Dubourdieu D. Trattato di Enologia II—Chimica del Vino Stabilizzazione Trattamenti. Dunod: Paris 1998; 389.
  • 2
    Puech JL. Extraction and evolution of lignin products in armagnac matured in oak. Am. J. Enol. Vitic. 1981; 32: 111.
  • 3
    MacNamara K, van Wyk CJ, Brunerie P, Augustyn OPH, Rapp A. Flavour components of whiskey. III. Ageing changes in the low-volatility fraction. S. Afr. J. Enol. Vitic. 2001; 22: 82.
  • 4
    Delgado T, Gómez-Cordovés C, Villarroya B. Relationship between phenolic compounds of low molecular weight as indicators of the aging conditions and quality of brandies. Am. J. Enol. Vitic. 1990; 41: 342.
  • 5
    Mattivi F, Versini G, Sarti S. Study on the presence of scopoletin in commercial wood-aged brandies. Riv. Vitic. Enol. 1989a; 3: 23.
  • 6
    Mattivi F, Versini G, Sarti S. The presence of Scopoletin as an indicator of ageing parameters in distillates. Riv. Vitic. Enol. 1989b; 2: 3.
  • 7
    Puech JL, Feuillat F, Mosedale JR. The tannins of oak heartwood : structure, properties, and their influence on wine flavor. Am. J. Enol. Vitic. 1999; 50: 469.
  • 8
    Pérez-Coello MS, Sanz J, Cabezudo MD. Determination of volatile compounds in hydroalcoholic extracts of french and american oak wood. Am. J. Enol. Vitic. 1999; 50: 162.
  • 9
    Guichard E, Fournier N, Masson G, Puech JL. Stereoisomeres of β-Methyl-γ-Octalactone. Quantification in brandies as a function of wood origin and treatment of the barrels. Am. J. Enol. Vitic. 1995; 46: 419.
  • 10
    Feuillat F, Moio L, Guichard E, Marinov M, Fournier N, Puech JL. Variation in the concentration of ellagitannins and cis- and trans-β-Methyl-γ-Octalactone extracted from oak wood (Quercus robur L., Quercus petracea Liebl.) Under model wine cask conditions. Am. J. Enol. Vitic. 1997; 48: 509.
  • 11
    Masson E, Baumes R, Moutounet M, Puech JL. The effect of kiln-drying on the levels of ellagitannins and volatile compounds of european oak (Quercus petracea Liebl.) stave wood. Am. J. Enol. Vitic. 2000; 51: 201.
  • 12
    Matricardi L, Waterhouse AL. Influence of toasting technique on color and ellagitannins of oak wood in barrel making. Am. J. Enol. Vitic. 1999; 50: 519.
  • 13
    Hale MD, McCafferty K, Larmie E, Newton J, Swan JS. The influence of oak seasoning and toasting parameters on the composition and quality of wine. Am. J. Enol. Vitic. 1999; 50: 495.
  • 14
    Masson G, Puech JL, Moutounet M. The chemical composition of barrel oak wood. Bull. O.I.V. 1996; 785–786: 634.
  • 15
    Sauvageot F, Feuillat F. The influence of oak wood (Quercus robur L., Q. Petracea Liebl.) on the flavor of burgundy pinot noir. An examination of variation among individual trees. Am. J. Enol. Vitic. 1999; 50: 447.
  • 16
    Ibern-Gómez M, Andrés-Lacueva C, Lamuela-Raventós RM, Lao-Luque C, Buxaderas S, de la Torre-Boronat MC. Differences in phenolic profile between oak wood and stainless steel fermentation in white wines. Am. J. Enol. Vitic. 2001; 52: 159.
  • 17
    Chatonnet P, Dubourdieu D, Boidron JN. Influence of fermentation and ageing conditions of dry white wines in barrels on their composition in substances yielded by oak wood. Sci. Aliments 1992; 12: 665.
  • 18
    Cutzach I, Chatonnet P, Henry R, Dubourdieu D. Identification of volatile compounds with a “Toasty” aroma in heated oak used in barrelmaking. J. Agric. Food Chem. 1997; 45: 2217.
  • 19
    Chatonnet P. Discrimination and control of toasting intensity and quality of oak wood barrels. Am. J. Enol. Vitic. 1999; 50: 479.
  • 20
    Porretta S. In Analisi Sensoriale & Consumer Science, Chiriotti Editori: Pinerolo, Italy, 2000; 159.
  • 21
    Salagoity-Auguste MH, Tricard Chr, Marsal F, Sudraud P. Preliminary investigation for the differentiation of enological tannins according to botanical origin: determination of gallic acid and its derivatives. Am. J. Enol. Vitic. 1986; 37: 301.
  • 22
    Flamini R, Dalla Vedova A. Fast determination of the total free resveratrol content in wine by direct-exposure-probe, positive-ion chemical ionization and collision-induced-dissociation mass spectrometry. Rapid Commun. Mass Spectrom. 2004; 18: 1925.
  • 23
    Caldeira I, Per eira R, Climaco MC, Belchior AP, de Sousa RB. Improved method for extraction of aroma compounds in aged brandies and aqueous alcoholic wood extracts using ultrasound. Anal. Chim. Acta Neth. 2004; 513: 125.