You are on page 1of 13

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/257372983

Relation between volatile composition, ellagitannin content and sensory


perception of oak wood chips representing different toasting processes

Article  in  European Food Research and Technology · April 2013


DOI: 10.1007/s00217-013-1930-0

CITATIONS READS

28 472

2 authors:

Kleopatra Chira Pierre-Louis Teissedre


Université de Bordeaux Institut des Sciences de la Vigne et du Vin University of Bordeaux - Institut des Sciences de la Vigne et du Vin
36 PUBLICATIONS   882 CITATIONS    280 PUBLICATIONS   9,738 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Innovations in the study of astringency of wine related to analytical methods, composition and structure of tannins, and polysaccharides View project

stilbenoids and health properties View project

All content following this page was uploaded by Pierre-Louis Teissedre on 31 May 2014.

The user has requested enhancement of the downloaded file.


Eur Food Res Technol (2013) 236:735–746
DOI 10.1007/s00217-013-1930-0

ORIGINAL PAPER

Relation between volatile composition, ellagitannin content


and sensory perception of oak wood chips representing
different toasting processes
Kleopatra Chira • Pierre-Louis Teissedre

Received: 19 November 2012 / Revised: 20 January 2013 / Accepted: 23 January 2013 / Published online: 28 February 2013
Ó Springer-Verlag Berlin Heidelberg 2013

Abstract The evolution of both volatile and nonvolatile respectively, whereas furfural levels justified 5.66 % of the
(ellagitannins) compounds extracted from oak wood chips total astringency variance.
while being macerated for 3 months in wine model solution
with eight different commercial chips [Toasted LT (Light Keywords Oak wood  Volatile compounds 
Toast), LT? (Light plus Toast), MT (Medium Toast), MT? Ellagitannins  Astringency  Predictive model
(Medium plus Toast), HT (Heavy Toast), Noisette, Special)
and un-toasted (UN)] was studied. Sensory analysis con-
cerning vanilla, spicy, overall woody, astringency, bitterness Introduction
and sweetness descriptors was conducted in parallel. A
2-month time is needed for the maximum extraction of el- The wood from which a barrel is made contributes to the
lagitannins and of furfural compound. Relating to other characteristics of the alcoholic beverages stored in it. Oak
volatile compounds, their concentration increased progres- wood has commonly been used in wine aging, during oak
sively over time. Spicy, woody and vanilla aromas as well as wood contact, red wine undergoes important modifications,
sweetness perception increased during contact time with oak as spontaneous clarification, slow and continuous diffusion
wood. Additionally to pyrolytic toasting stage, watering of oxygen through the wood pores of the oak barrel, and the
process during toasting influences significantly oak wood extraction of many substances from the oak wood (e.g.,
composition (lactones, vanillin guaiacol compounds and aromatic compounds and ellagitannins) which modulate its
ellagitannins). Correlation analysis between chemical and organoleptic quality and complexity such as aroma, struc-
sensory data demonstrated significant correlations between ture, astringency, bitterness, persistence color and astrin-
chemical and sensory parameters and provides some insight gency. The ellagitannins (hydrolysable tannins) are among
into the implication of volatile and nonvolatile composition these substances. In oak heartwood, they may represent
on the final quality perception of astringency. A predictive 10 % of the dry weight and are responsible for the high
model between astringency intensity and ellagitannins, fur- durability of this wood [1]. From a quantitative point of
fural, guaiacol and methyl guaiacol contents was estab- view, castalagin and vescalagin are the most abundant el-
lished. Ellagitannins levels explained 45 %, guaiacol and lagitannins in oak wood [2] but lyxose/xylose derivatives
methyl guaiacol contents interpreted 13.20 and 8.86 %, (grandinin and roburin E) and dimeric forms (roburins A,
B, C and D) are also present [3, 4]. The hydroalcoholic
nature of wine allows the extraction of these compounds
from wood to wine, but due to their high reactivity, their
K. Chira  P.-L. Teissedre (&)
ISVV, EA 4577, Œnologie, Univ. Bordeaux, levels in wine are much lower than could be expected. The
210 Chemin de Leysotte, 33140 Villenave d’Ornon, France total ellagitannins levels in wine, being in contact four
e-mail: p.teissedre@u-bordeaux2.fr months with oak staves, may range from 0.8 to 3.49 mg/L
of released ellagic acid [5].
K. Chira  P.-L. Teissedre
INRA, ISVV, USC 1366 Œnologie, 210 Chemin de Leysotte, Human sensory experiments revealed that ellagitannins
33140 Villenave d’Ornon, France have an impact on astringency and bitterness sensation [6, 7].

123
736 Eur Food Res Technol (2013) 236:735–746

The oral sensation imparted by ellagitannins was described many factors that affect oak qualities used: origin; species
as astringent at relatively low threshold concentrations [18]; the method used to obtain the staves; the stave-drying
spanning from 0.2 to 6.3 lmol/L, whereas bitterness was technique (natural or artificial) [19].
strongly depending on ellagitannin structure, a bitter taste However, in order to know the compounds that oak
was perceived at threshold concentrations between 410 and wood yields to the wine, the use of synthetic wine has the
1650 lmol/L. Recently, Stark et al. [8] observed that all advantage of ensuring the absence of any microbial activ-
ellagitannin derivatives induce a smooth and velvety ity. At the same time, by using synthetic wine, this ensures
astringent oral sensation at rather low threshold concen- that no fermenting compounds, and which could also come
trations ranging from 0.9 to 2.8 lmol/L. from the barrels, will exist. The bibliography consulted
Other important wood chemicals, which might be reveals few studies on the characterization of oak wood in a
impact molecules, include the naturally present cis- and synthetic medium [20].
trans-methyl-octalactone (whiskey lactones, with a coco- Therefore, to clarify the impact of oak wood on wine
nut flavor in their pure state), trans-2-nonenal (sawdust composition, a model wine was used. This approach mini-
smell), volatile phenols released via toasting such as mized any effect which may have arisen due to compounds
guaiacol, 4-methylguaiacol and eugenol (spicy and smoky naturally present in wines, and microbial activity was kept to
smells) and vanillin (vanilla smell). All of these com- a minimum. The accumulation of oak volatiles and non-
pounds can be present in barrel-aged wines in concentra- volatiles compounds was monitored over a 3-month time. In
tions above the sensory threshold; their concentrations parallel, sensory evaluation of wine model solution over a
depend on the chemical composition of the barrel’s wood. 3-month time was realized. The toasting level impact on
Some of them undergo modification in a wine medium. both volatile, nonvolatile compounds and sensory perception
The increasing demand for wood that is suitable for was studied in parallel. The relationship between the
barrel making and the rapid development of alternative chemical composition and the sensory assessment of oak
aging systems have led to the possibility of using wood wood extract was also investigated.
pieces. This practice recently was approved and legislated
by the European Community (CE 2165/2005 and CE
1507/2006), but in some countries such as Australia, South Materials and methods
Africa and Chile, this practice has been used for several
years. Different shapes of oak wood pieces can be found on Wood origin and drying conditions
the market: chips, cubes or beans, powder, shavings or
granulates, dominoes, and blocks or segments. Factors such The wood samples were constituted from two oak species
as piece size, amount of added wood and contact time (Quercus robur and Quercus petraea) from the same forest
between wood and wine affect the sensory and chemical located in center in France. The oak wood was stored for
characteristics of wines [9–11], especially their wood- natural seasoning during 24 months in the Tonnellerie
related volatile composition [12]. Nadalié (Ludon-Medoc, France) seasoning park. Then,
There are several factors that determine the final char- they were crushed to oak chips dimensions (L = 1–2 cm,
acteristics of wines macerated with chips, and some of h = 0.2–0.8 cm) and submitted to different toasting pro-
these factors have been studied in recent years. Some cedures according to the desired final product using oak
papers have been written after studying the effect of chips fire. Each time 150 kg of chips is put into a stainless drum
on the phenolic composition of different wines [9, 10, 13– (1 m3) and this drum turns above an oak fire. The intensity
15], whereas others have focused on volatile compounds, levels used are as follows: light (LT), 1 h and 30 min at
mainly on those extracted from wood. Thus, Guchu et al. 165°C; light plus (LT ?), 2 h at 170°C; medium (MT), 2 h
[16] evaluated the influence of the geographical origin of at 180°C; medium plus (MT ?), 2 h and 25 min at 190°C;
the oak (American or Hungarian), the toasting degree Noisette, 3 h at 175°C; Special 2 h at 180°C and heavy
(toasted and nontoasted) and the time of contact on volatile toast (HT), 2 h and 30 min at 200°C. In the case of Special,
compounds in a white wine. Arapitsas et al. [12] also a watering process takes place. After 2 h of toasting, 30 L
studied the effect of the contact time, but only for 14 days, of water is put to the drum, and afterward the drum is
and the size of the chips on red wines. Frangipane et al. placed above the fire in order to be heated for 30 min more.
[11] studied four types of chips from four different French
forests, added to a red wine stored in 2-year-old barrels. Sample preparation
Campbell et al. [17] studied the origin of the wood and the
type of heating in model wines. An analysis of the results Each experiment was carried out in triplicate in hermeti-
discussed in these papers indicates a great variability in the cally closed 5-L glass jars. Different model wines were
results, which is probably due to the fact that there are prepared for this experiment, all in triplicate. Model wines

123
Eur Food Res Technol (2013) 236:735–746 737

were prepared with deionized water, HPLC grade ethanol furfuryl alcohol, m/z 98 and m/z 83 for the internal stan-
(Prolabo-VWR, Fontenay-sous-Bois, France) and tartaric dard (dodecan-1-ol). The concentrations of each substance
acid (Prolabo-VWR, Fontenay-sous-Bois, France). Titrat- were measured by comparison with calibrations made with
able acidity was adjusted to 4.5 g tartaric acid/L, and pure reference compounds analyzed under the same con-
ethanol level was 12 % (v/v). ditions. The corresponding calibration was made for each
Small pieces of French oak wood (L = 1–2 cm, compound, and linear regression coefficients between 0.98
h = 0.2–0.8 cm) from the same forest located in the center and 0.999 were obtained.
of France were supplied by Nadalié cooperage (Tonnellerie
Nadalié, Ludon-Medoc, France). Three grams of medium Sample preparation prior to total ellagitannin level
size, UN (Untoast), LT (Light Toast), LT? (Light plus determination
Toast), MT (Medium Toast), MT? (Medium plus Toast),
HT (Heavy Toast), Noisette, Special chips were added to Model wine solution (50 mL) was evaporated under
the model wine solution. The jars were shaken once a day reduced pressure, and the resulting residue was dissolved in
to facilitate the extraction of oak compounds. Sampling methanol (20 mL); then, 4 mL of this mixture was loaded
was carried out at 15 days, 1, 2 and 3 months from the in the hydrolysis tubes for the determination of the total
initiation of contact between the oak wood and the syn- ellagitannin level.
thetic wine. Afterward, elagitannins and aromatic com-
pounds quantification were performed by HPLC–UV and Total ellagitannin level determination HPLC–UV
GC–MS analysis, respectively. Sensory analysis was also analysis
conducted.
The total ellagitannin concentration was determined by the
Extraction of volatile compounds quantification of ellagic acid released during acidic
hydrolysis (2 h at 100 °C, 2NHCl in MeOH) as previously
Model wine solutions were extracted with dichlorometh- described [22]. Each sample was analyzed in triplicate, and
ane; 200 lL of a solution of dodecan-1-ol as internal each reaction mixture was subjected to HPLC–UV using a
standard was added to 50 mL of samples. Three extractions LiChrospher 100 RP-18 column, 250 9 4.6 mm, 5 lm.
were then carried out using 4, 2 and 2 mL of dichloro- The equipment used for the HPLC analysis consisted of a
methane. The organic fractions were combined and dried Finnigan Surveyor UV–Vis detector (UV–Vis 200), a
on sodium sulfate anhydrous and then concentrated to Finnigan autosampler and a Finnigan ternary pump. The
500 lL under a nitrogen stream. In all cases, the samples mobile phases used were solvent A [H2O/H3PO4 (99.9/1)]
were analyzed in duplicate. and solvent B [methanol/H3PO4 (99.9/1)], and the gradient
elution was 0–35 % of B in 5 min, 35–45 % of B in 25 min
Gas-chromatography analysis and 45–100 % of B in 5 min. The flow rate was set at
1 mL/min with detection set at 370 nm.
A simple and reliable GC method for quantitative deter-
mination of the volatile compounds arising from oak wood Sensory analysis
was used according to an adaptation of a previous method
[21]. An Agilent HP 5890 GC (Hewlett–Packard, Wil- The sensory assessment was done by a committee of 20
mington, DE, USA) was coupled with a mass spectrometer expert judges from the Oenology faculty of Bordeaux. The
(HP 5972, electronic impact 70 eV, eMV = 2 kV). One judges were specially trained in the employment of scales
lL samples of organic extract were injected in splitless and aroma descriptors according to ISO 8586-1 [23]. The
mode. The column was BP21 (SGE) (50 m 9 0.32 mm, attributes selected were grouped in two families: olfactive
0.25 lm); carrier gas was helium (pressure: 70 kPa); descriptors related to wood–wine interaction (vanilla, spicy,
temperatures were injector, 250 °C; detector, 280 °C; overall woody) and gustative descriptors (sweetness astrin-
oven, 60 °C for 1 min programmed at a rate of 3 °C/min to gency and bitterness). The ‘‘overall woody’’ descriptor was
240 °C, the final step lasting 40 min; the splitless time was chosen by tasters to describe all olfactory sensations brought
30 s with a split flow of 30 mL/min. about by the wood.
The compounds were identified by comparing their Panellists attended 16 training sessions over a period of
retention times and mass spectra with those of the pure 2 months. The training period included a first general phase
reference standards. Working in the SIM mode, the fol- a second and a third, product-specific training phase. The
lowing ions were used: syringaldehyde, m/z 182; vanillin, general phase was dedicated to the recognition of sensa-
m/z 151; eugenol, m/z 164; guaiacol, m/z 124; b-methyl-c- tions and aromas perceived. Aqueous solutions of vanillin
octalactone, m/z 99; 5-methyl-2-furfuraldehyde, m/z 110; (20 lg/L), eugenol (60 lg/L), furfurylthiol (0.8 ng/L) and

123
738 Eur Food Res Technol (2013) 236:735–746

oak wood chip (5 g/L medium toast) were proposed for Data analysis
vanilla, spicy and overall woody character, respectively.
Aqueous solutions of quinine sulfate (0.25 g/L), aluminum Statistical data analysis was performed using the analysis
sulfate (3 g/L) and sucrose (4 g/L) were proposed to set of variance (ANOVA) of Statistica V.7 software (Statsoft
bitterness, astringency and sweetness. During this session, Inc., Tulsa, OK). Tukey’s HSD and Duncan’s tests were
the discriminative ability of participants was assessed. used as comparison tests when samples were significantly
Samples were presented, and participants were instructed different after ANOVA (p \ 0.05) for chemical and sen-
to identify the solutions as sweet, bitter, astringent, spicy, sory analysis, respectively. Principal component analysis
overall woody or vanilla. All participants correctly identi- (PCA) was performed on the correlation matrix using the
fied all solutions. attributes that differed significantly by ANOVA. Pearson’s
During the second phase (8 sessions), the judges have correlation analysis was used to investigate relationships
been trained to evaluate the descriptors: sweet, bitter, between chemical composition and sensory, perception.
astringent, spicy, woody or vanilla in various concentra- Once correlations established, multiple regression analysis
tions [23]. In order to improve panel performance, scaling was used to determine the extent to which the variance in
training (ranking of solution according to concentration of the sensory measures (astringency) could be predicted by
descriptor) was used. Fourth sessions were used for the the remaining variables.
olfactory attributes and four for the sensory gustative
attributes. Wine model solutions containing different con-
centrations of table sugar (0–24 g /L) for sweetness [24], Results and discussion
quinine sulfate (0–15 mg/ L) for bitterness, aluminum
sulfate (0–4 g /L) for astringency, vanillin (0–320 lg/L) Oak wood volatile composition
for vanilla, eugenol (0–500 lg/L) for spicy and medium
toast oak wood chip (0–10 g/L) for woody were presented The following wood volatile compounds were studied:
to the panel to aid them to discriminate between the dif- furanic aldehydes, furfural and 5-methylfurfural, the two
ferent concentrations. After this second phase, the dis- isomers of methyl-c-octalactone, cis and trans (commonly
criminative ability of participants was assessed. known as oak lactones or whiskey lactones); the volatile
The third phase (7 sessions) was allocated to familiarize phenols guaiacol, eugenol and cis- and trans-isoeugenol;
the judges with the intensity scale used (0–7). The first two and the aldehyde phenols, vanillin and syringaldehyde.
sessions have been dedicated to the overall assessment of
the descriptors of interest. Oak wood chips of different Furfuryl compounds
concentrations (0–8 g/L) and of different toast (light,
medium and heavy) were added in model wine solutions in The evolution of furfural and 5-methylfurfural concentra-
order to determine the repeatability of judges from one tions during aging was different and dependent on the
session to another. For the last five sessions, oak wood toasting (Fig. 1). Globally, furfural concentration was much
chips of different concentrations (0–8 g/L) and of different higher than that of 5-methylfurfural and was extracted
toast (light, medium, high) were added in red wine. After rapidly, reaching a maximum level at 60 days (2 months).
these last sessions, the judges became familiar with Using two pieces of oak wood (12 9 2 9 2 cm) in two
intensity rating of spicy, woody, vanilla, sweetness, bit- liters of model wine solution, an 105-day assay period was
terness and astringency using a 0–7 point scale. needed to extract the maximum of furfural in model wine
In the formal sessions, the panelists were provided with solutions [27]. Using 10 grams of medium size in one liter
30 mL of wine in coded standard clear wine glasses [25], model wine solution [20], an 120-day assay period was
covered with a watch glass to minimize the escape of vol- required for extracting the maximum of furfural. In the
atile components and coded with random three-digit num- current experiment, less time was demanded for the maxi-
bers. Assessment took place in a standard sensory-analysis mum extraction, probably due to the lower concentration of
chamber [26], equipped with separate booths, and with a oak chips used or even due to the toasting method used, oak
uniform source of lighting, absence of noise and distracting chips are toasted in oak fire.
stimuli, and ambient temperature between 19 and 22 °C.
Wines were sniffed and tasted. In every session, the cis- and trans-b-methyl-c-octalactones
expert judges had to start with evaluation of the orthonasal
odor (first without moving the glass, then moving it cis-b-Methyl-c-octalactone was detected in bigger con-
gently), and then, after a short break, they evaluated the centrations than trans-b-methyl-c-octalactone (Fig. 1). The
taste (sweetness, bitterness) and tactile sensation (astrin- cis isomer is regarded as among the most important of oak
gency). The experiment was carried out in duplicate. wood volatile components and is generally found at above

123
Eur Food Res Technol (2013) 236:735–746 739

UN (not toasted), LT (Light Toast), LT+ (Light plus Toast), MT (Medium Toast), MT+ (Medium Plus Toast), HT (Heavy Toast), Noisette, Special (toast with watering)

6000 450

Furfural 400 Methyl-Furfural


5000
Concentration in µg/L

350

Concentration in µg/L
4000 300

250
3000
200

150
2000
100

1000 50

0
0 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4
T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4
UN LT LT+ MT MT+ HT Noisette Special
UN LT LT+ MT MT+ HT Noisette Special

200 200

180 180 cis b methyl -γ -octalactone


Trans b methyl -γ -octalactone
160 160

Concentration in µg/L
Concentration in µg/L

140 140

120 120

100 100

80 80

60 60

40 40

20 20

0 0
T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4
UN LT LT+ MT MT+ HT Noisette Special UN LT LT+ MT MT+ HT Noisette Special

60
60
Guaiacol Methyl -Guaiacol
50
50
Concentration in µg/L

Concentration in µg/L

40
40

30
30

20
20

10
10

0
0
T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4
T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4
UN LT LT+ MT MT+ HT Noisette Special
UN LT LT+ MT MT+ HT Noisette Special

20
20

18 Syringol
Eugenol+Isoeugenol 18
16
16
Concentration in µg/L

Concentration in µg/L

14
14
12
12
10
10
8
8
6
6
4
4
2
2
0
T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4
0
UN LT LT+ MT MT+ HT Noisette Special T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4
UN LT LT+ MT MT+ HT Noisette Special

4500
1600
Synrigaldehyde
4000
1400 Vanillin
Concentration in µg/L

3500
Concentration in µg/L

1200
3000
1000
2500
800
2000

600
1500

1000 400

500 200

0 0
T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4 T1 T2 T3 T4
UN LT LT+ MT MT+ HT Noisette Special UN LT LT+ MT MT+ HT Noisette Special

Fig. 1 Evolution of volatile oak compounds during different contact times (T1 = 15 days, T2 = 1 month, T3 = 2 months, T4 = 3 months)

123
740 Eur Food Res Technol (2013) 236:735–746

its perception threshold [28]. The profile obtained for lac- 970.88 lg/L, respectively) than samples with Special oak
tone extraction from the wood is depending on their (940.45 lg/L) at the end of 12-week period. This finding
toasting level. For almost all the samples, concentration suggests that watering during toasting process may
increases during the maceration time. Samples with UN decrease significantly vanillin. Thus, except of alcohol
oak chips presented more important concentrations than content and acidity [20], watering process during toasting
these with MT. This is due to the fact that b-methyl-c- may modulate the extraction of lactones and vanillin.
octalactone can be found naturally in oak wood [29, 30]. In all samples, syringaldehyde content, like vanillin,
Moreover, light-toasted chips released more oak lactones to increased during maceration time, independently of the oak
wines than heavy toasted, probably due to the thermo- type used. The samples with medium toast (Noisette, MT,
degradation of these heat-sensitive compounds or their loss Special, MT?) presented the most important concentra-
by volatilization when the oak wood is subjected to very tions, following by light (LT, LT?) and heavy toast (HT).
high temperatures or even charring [31]. Noisette toast Indeed, toasting increases the quantities of aromatic
showed the highest concentrations of lactones, and this compounds and also leads to the formation of other new
observation could be attributed to the fact that for Noisette, compounds. Furanic aldehydes caused by carbohydrate
the temperature used was lower and lasted more time than degradation and responsible for the ‘‘toasty’’ aromas, and
this of MT. Comparing Special to MT oak chips, it is lignin degradation compounds (vanillin, syringaldehyde
observed that samples with Special extract more lactones, and guaiacol) tend to be formed during the toasting pro-
implying that watering process during toasting enhance cess. However, HT chips presented a low potential of
lactones extraction. extracting volatile compounds, showing that toasting
increased concentrations of volatile compounds up to cer-
Volatile phenols tain level of toasting and then decreases afterward [33]. In
general, it is observed that Noisette oak wood extracts
Guaiacol and 4-methylguaiacol have smoky aromas and more volatile compounds following by MT, MT? and
indicate the wood toasting level [32] since they are formed Special oak wood extracts. These observations indicate that
almost exclusively by the degradation of lignin during the 1-hour prolongation in toasting process along with a 5 °C
toasting process. The rate of extraction in model solutions decrease in toasting temperature enhance oak volatile
has been shown as consistent with their generation by extraction.
toasting, with the largest concentration presented at the end
of maceration time in samples with heavy toast oak chips Oak wood total ellagitannin concentration
(Fig. 1). It is interesting to note that samples treated with
Noisette and Special extract less guaiacol compounds than The ellagitannin level and composition in each oak stave
samples with MT, even if the toasting temperature used is were determined by HPLC–UV. In a first approach, the
the same. Suggesting that the prolongation in toasting time total ellagitannin level was estimated by the determination
along with the watering during toasting is factors that of the amount of ellagic acid released after acidic hydro-
decrease guaiacol levels. lysis. During this reaction, each ellagitannin monomer or
Eugenol increased its concentration from the beginning, dimer released one molecule of ellagic acid. The total el-
its concentration ranged from 2.72 lg/L (15 days for LT) lagitannin level, expressed as mg/L of released ellagic acid
to 17.16 lg/L (after 3 months for Noisette) (Fig. 1). In in wine model solution, revealed a large diversity of con-
samples treated with HT oak, it was detected only after centrations ranging from 0.95 to 13.72 mg of released
1 month in a slight concentration of 0.74 lg/L, whereas in ellagic acid/L of wine model solution (Fig. 2). The UN and
samples with UN oak, it was presented from the beginning the HT presented the highest and the less important
to the end of maceration time (3.5–4.61 lg/L). ellagitannins concentrations. Fig. 2 reveals the drastic
changes in wood chemical composition that occur during
Phenolic aldehydes toasting. Such differences were expected since ellagitan-
nins undergo thermolytic degradation during the toasting
Vanillin, which is very important to wine flavor, is present process [34, 35]. UN oak chips presented the highest values
in green wood but is also associated with the heating in mg of released ellagic acid/L of wine model solution
process, so higher concentrations are to be expected in the (from 8.96 (T1) to 13.72 (T4)), following by LT (from 7.87
first few millimeters of the toasted wood. Its concentration (T1) to 9.52 (T4)), LT? (from 6.75 (T1) to 7.92 (T4)), MT
increased progressively over time, with the highest values (from 3.10 (T1) to 6.23 (T4)), MT? (from 6.75 (T1) to 7.92
reached in the model solution with Noisette oak chip (T4)), Noisette (from 2.16 (T1) to 5.71 (T4)), Special (from
(1431.08 lg/L after 3 months). Samples with MT and 2.06 (T1) to 4.01 (T4)) and finally by HT (from 0.95 (T1)
MT ? oak chips extract more vanillin (1048.45 and to 1.21 (T4)). Moreover, the Fig. 2 demonstrates that

123
Eur Food Res Technol (2013) 236:735–746 741

16 16
UN LT
mg eq.released ellagic acid/L of model
14 14

mg eq.released ellagic acid/L of model


wine solution 12 12
10 10

wine solution
8 8
6 6
4 4
2 2
0 0
T1 T2 T3 T4 T1 T2 T3 T4
mg eq.released ellagic acid/L of model

16
16

LT+ 14 MT

mg eq.released ellagic acid/L of model


14

12 12
wine solution

10

wine solution
10

8 8

6 6

4 4

2 2

0 0
T1 T2 T3 T4 T1 T2 T3 T4
mg eq.released ellagic acid/L of model

16 16
MT+ HT
mgeq.releasedellagicacid/L of

14 14
model wine solution

12 12
wine solution

10 10
8 8
6 6
4 4
2 2
0 0
T1 T2 T3 T4
T1 T2 T3 T4
mg eq.released ellagic acid/L of model

mg eq.released ellagic acid/L of model

16
16
14
Noisette 14 Special
12
12
wine solution

wine solution

10
10
8
8
6
6
4
4
2 2
0 0
T1 T2 T3 T4 T1 T2 T3 T4

Fig. 2 Evolution of ellagitannins during different contact times (T1 = 15 days, T2 = 1 month, T3 = 2 months, T4 = 3 months)

independent toasting time and method, a maximum the first millimeter of the wood, the solution needs to go
extraction of ellagitannins is obtained after a 2-month time, deeper in the wood to extract more ellagitannins, conse-
afterward a stabilization or even a decrease is observed. quently, at a slower rate. Samples with Special oak chips
During the first 2 months, wine model solution extracts not only presented less important ellagitannins concentra-
ellagitannins at a faster rate than between 2 and 3 months. tions than samples with MT oak chips but they also showed
When most of the ellagitannins have been extracted from as well the greatest decrease in these compounds during

123
742 Eur Food Res Technol (2013) 236:735–746

Fig. 3 Sensory evaluation of UN Vanilla


model wine solution treated 5 LT Vanilla
5
with different oak chips during 4
4
different contact times Bitterness
3
Woody 3
Bitterness Woody
(T1 = 15 days, T2 = 1 month, 2
2
T3 = 2 months, 1
1
T4 = 3 months) 0
0

Astringency Spicy
UN T1 Astringency Spicy
LT T1
UN T2
LT T2
Sweetness UN T3 LT T3
Sweetness
UN T4 LT T4

LT+ Vanilla MT Vanilla


5 5
4
4
3
3 Bitterness Woody
Bitterness Woody 2
2
1
1
0
0

Astringency Spicy
LT+ T1
Astringenc MT T1
Spicy
y LT+ T2
MT T2
LT+ T3
Sweetness
MT T3
LT+ T4
Sweetness MT T4

MT+
Vanilla HT
5
Vanilla
4 5

3 4
Bitterness Woody
2 3
Bitterness Woody
1 2

0 1

0
MT+ T1
Astringenc
Spicy MT+ T2 HT T1
y Astringenc
Spicy
MT+ T3 y HT T2

MT+ T4 HT T3
Sweetness HT T4
Sweetness

Noisette Special
Vanilla Vanilla
5 5

4 4

3 3
Bitterness Woody Bitterness Woody
2 2

1
1
0
0

N T1 Astringenc S T1
Astringenc Spicy
Spicy y
y N T2 S T2

N T3 S T3

N T4 Sweetness S T4
Sweetness

123
Eur Food Res Technol (2013) 236:735–746 743

Fig. 4 Principal component


analysis (PCA) representation N
of chemical and sensory data for
model wine solution treated
with different oak chips during
different contact times
(T1 = 15 days, T2 = 1 month, HT
T3 = 2 months, LT
T4 = 3 months)

LT+

UN

after 2 months. Thus, not only the pyrolytic toasting stages more astringent and with stronger bitterness. This may be
diminish the quantity of these compounds but also the attributed to their higher levels of ellagitannins.
watering procedure.
Correlations between sensory and chemical results
Sensory evaluation of oak chips
The analysis of variance revealed that the toasting method
Figure 3 shows ‘‘spider web’’ diagrams for the average has a significant impact (p \ 0.05) on chemical composi-
scores of the olfactory attributes for all the trials. For all the tion and sensory perception of oak wood extracts. After-
trials, it is observed that vanilla, smoky, grilled characters ward correlations between sensory descriptors were
along with sweetness build up, whereas astringency assessed. The PCA factor loading plot for the sensory
intensity decreases slightly during the contact time. In variables and the means for oak volatile compounds and
general, MT and MT? are perceived more woody, whereas ellagitannins are shown in Fig. 4. This figure is comple-
UN less woody; vanilla intensity varies from 1.9 (for UN mented by the Pearson’s correlation coefficient values
oak chips) to 4.4 (for Noisette oak chips), woody intensity presented in Table 1. The first two principal components
ranges from 0.4 (for UN oak chips) to 4.9 (for Noisette oak captured 67.53 % of the sample variance. The first principal
chips); spicy character fluctuates from 1.1 (for HT oak component (PC) is strongly and negatively correlated with
chips) to 3.3 (for Noisette oak chips); in parallel, samples both sensory descriptors (vanilla, smokey/grilled, spicy,
with Noisette after 3 months, presented the most important sweetness) and volatile compounds (furanic compounds
levels of eugenol and vanillin contents, observation that guaiacol, eugenol and vanillin), whereas it is strongly pos-
may justify their intense spicy and vanilla aroma. Samples itively correlated with ellagitannins, astringency and bit-
with Special were perceived less woody than samples with terness. Concerning the second principal component, it
MT oak chips which is accordance with the above results, represents strongly and negatively the lactones. Apparently
samples treated Special extract less guaiacol compounds oak chips can easily be discriminated according their
than samples with MT. chemical composition and sensory parameters. There seem
Sweetness perception goes from 2.0 (for UN oak chips) to be six groups, namely HT, LT, LT?, UN, N (Noisette)
to 3.2 (for MT oak chips); astringency perception varies and S (Special). MT and MT? oak chips seem to have
between 1.4 (for HT oak chips) and 4.3 (UN) and bitterness similar behavior pattern, they show a good correlation with
varies between 2.0 (MT ?) and 4.2 (UN). Model wine guaiacol compound and smoky aroma. UN along with LT
solutions treated with untoasted oak chips were generally and LT? oak chips has a good affinity with ellagitannins

123
744 Eur Food Res Technol (2013) 236:735–746

and both astringency and bitterness. Special oak chips Astringency ¼ 2:17 þ 1:57  ðellagitanninsÞ  2:06
demonstrate a compatibility with lactones, whereas Noi-  ðfurfuralÞ  5:37  ðguaiacolÞ þ 6:11
sette oak chips are perceived to have more vanilla and  ðmethyl guaiacolÞ
grilled flavor, and at the same time, they perceived sweeter.
Each sensory descriptor was correlated with the chem-
ical concentration of the oak wood compounds of interest These results confirm ellagitannins importance in
(Table 1). This procedure allowed us to measure the extent astringency perception; it intensifies significantly with
to which sensory and chemical variables are correlated. ellagitannins concentration. It is also noteworthy that the
Based on the correlation analysis, it can be observed that above model suggests that volatile compounds like guaia-
perceived astringency is positively and significantly cor- col and furfural contents play a relevant role in sensory
related with elagitannins levels (R = 0.93) and negatively astringency, and the highest astringency was perceived for
with guaiacol (R = -0.65), methyl guaiacol (R = -0.46) UN oak wood extracts (4.3) containing 13.72 mg/L of
and furfural (R = -0.47). These are the variables con- released ellagic acid, 47.14 lg/L of furfural and 2.46 lg/L
tributing the most to the predictive model of astringency. guaiacol.
When these three variables are used for a multiple linear The overall woody intensity is positively correlated with
regression analysis (Fig. 5), they explained 72.6 % of the guaiacol, methyl guaiacol and furfural contents which is
perceptual astringency variance; ellagitannins levels reasonable since oak wood sensation is complex and
explained 45 %, guaiacol and methyl guaiacol contents influenced by the presence of various odor-active wood
account 13.20 and 8.86 %, respectively, whereas furfural extractives [18, 36, 37]. The negative correlation with
levels justified 5.66 % of the total astringency variance ellagitannins content may be attributed to the fact that
(Table 2). The importance of these variables can be also toasting up to a certain level decreases ellagitannins con-
seen in the predicted model: centration. The perceived intensity of vanilla aroma is

Table 1 Pearson’s correlations between chemical and sensory data


Elagitannins Furfural Methyl Guaiacol Methyl trans-b methyl-c- cis-WL b methyl-c- Eugenol Vanillin
furfural guaiacol octalactone octalactone

Vanilla -0.27 NS 0.39* 0.25 NS 0.76*** 0.88*** -0.04 -0.10 0.40* 0.56**
Woody -0.46** 0.50** 0.29 NS 0.88*** 0.89*** -0.17 -0.26 0.35 0.45**
Spicy -0.01 NS 0.18 0.04 NS 0.26 NS 0.45** 0.02 0.04 0.11 0.28
Sweetness 0.03 NS 0.23 0.10 NS 0.47** 0.63*** -0.35 -0.34 0.06 0.30
Astringency 0.93*** -0.47** -0.25 NS -0.65*** -0.46** -0.06 0.15 -0.25 -0.22
Bitterness 0.35 NS -0.34 -0.33 NS -0.47** -0.54** 0.19 0.19 -0.08 -0.18
NS not significant ([5 %)
*, **, *** Significant at 5, 1 and 0.1 %, respectively

Fig. 5 Prediction of sensory


astringency from chemical
composition

123
Eur Food Res Technol (2013) 236:735–746 745

Table 2 Predictive model of astringency Acknowledgments The authors gratefully thank Dr. Tempère
Sophie for her essential advices in the training of the judges as well as
Variable Astringency the judges who participated in the sensory analyses. We also thank
Dr. Warren Albertin for her suggestions in statistical analysis.
Sum of squared errors p value
of prediction (SSE)
Conflict of interest None.
Ellagitannins 45.00 6.681 3.59e-07***
Compliance with Ethics Requirements This article does not
Furfural 5.66 0.02535*
contain any studies with human or animal subjects.
Guaiacol 13.20 0.00121**
Methyl guaiacol 8.86 0.00510**
Residuals 27.27 8.832 1.90e-09***
*, **, *** Significant at 5, 1 and 0.1 %, respectively References

1. Scalbert A, Monties B, Favre JM (1988) Polyphenols of Quercus


correlated with vanillin, eugenol and furfural concentra- obur: adult tree and in vitro grown calli and shoots. Phyto-
tions. Thus, vanilla flavor is correlated with vanillin, a chemistry 27:3483–3488
2. Nonier MF, Vivas N, Vivas De Gaulejac N, Absalon C, Vitry C,
marker of vanilla smell but it is also enhanced by the Fouquet E (2005) Global fractionation of oak heartwood
presence of other odorous chemicals like eugenol, furfural extractable polymers (lignins, polysaccharides and ellagitannins)
and methyl guaiacol. Overall woody aroma is positively by selective precipitations. J Sci Food Agric 85:343–353
correlated with methyl guaiacol and furfural compounds, 3. Herve du Penhoat CLM, Michon VMF, Ohassan A, Peng S,
Scalbert A, Gage D, Roburin A (1991) A dimeric ellagitannin
which is logical since these molecules enhance oak woody from heartwood of Quercus robur. Phytochemistry 30:329–332
aromas. Surprisingly, the spicy flavor is not closely related 4. Hervé Du, Penhoat CLM, Michon VMF, Peng S, Viriot C,
to eugenol content (pure eugenol is described as clove Scalbert A, Gage D (1991) Structural elucidation of new dimeric
like), but it is correlated with methyl guaiacol, it is likely ellagitannins from Quercus robur L. roburins A-E. J Chem Soc
Perkin Trans 1:1653–1660
that this sensation is more complex and influenced by the 5. Michel J, Jourdes M, Silva MA, Giordanengo T, Mourey N,
presence of other odor-active wood extractives. Sweetness Teissedre PL (2011) Impact of concentration of ellagitannins in
and bitterness perception are related to guaiacol com- oak wood on their levels and organoleptic influence in red wine.
pounds, judges characterized samples with important J Agric Food Chem 59:5677–5683
6. Glabasnia A, Hofmann T (2006) Sensory-directed identification
guaiacol levels sweeter and less bitter; implying that a of taste-active ellagitannins in American (Quercus alba L.) and
strong reduction in bitterness takes place, when samples European oak wood (Quercus robur L.) and quantitative analysis
have important guaiacol levels. This may also indicate that in bourbon whiskey and oak-matured red wines. J Agric Food
judges interpret less bitter the samples that they found Chem 54:3380–3390
7. Sáenz-Navajas MP, Fernández-Zurbano P, Ferreira V (2012)
sweeter. Contribution of nonvolatile composition to wine flavor. Food Rev
In summary, this work reveals that cooperage oak woods Int 28:389–411
from different toasting treatments differed in terms of their 8. Stark T, Wollmann N, Wenker K, Lösch S, Glabasnia A, Hof-
principal chemical composition, and these differences were mann T (2010) Matrix-calibrated LC-MS/MS quantitation and
sensory evaluation of oak ellagitannins and their transformation
reflected by perceived sensory differences. The oak chips, products in red wines. J Agric Food Chem 58:6360–6369
representing different toasting methods (UN, LT, LT?, 9. Del Alamo Sanza M, Escudero JAF, De Castro Torı́o R (2004)
MT, MT?, HT, Noisette and Special), were very well Changes in phenolic compounds and colour parameters of red
discriminated according to their chemical composition and wine aged with oak chips and in oak barrels. Food Sci Technol Int
10:233–241
sensory assessment. Independent of toasting, a 2-month 10. Del Alamo Sanza M, Nevares Domı́nguez I (2006) Wine aging in
time is demanding for extracting the maximum of furfural bottle from artificial systems (staves and chips) and oak woods:
and ellagitannins compounds. Additionally to pyrolytic Anthocyanin composition. Anal Chim Acta 563:255–263
toasting stage, watering process during toasting influences 11. Frangipane MT, Santis DD, Ceccarelli A (2007) Influence of oak
woods of different geographical origins on quality of wines aged
significantly oak wood composition. Special extracts more in barriques and using oak chips. Food Chem 103:46–54
lactones, less vanillin and guaiacol compounds and at the 12. Arapitsas P, Antonopoulos A, Stefanou E, Dourtoglou VG (2004)
same time less ellagitannins. Artificial aging of wines using oak chips. Food Chem 86:563–570
The correlation analysis evidenced significant correla- 13. Del Alamo Sanza M, Nevares Domı́nguez I, Cárcel Cárcel LM,
Navas Gracia L (2004) Analysis for low molecular weight phe-
tions between chemical and sensory parameters and pro- nolic compounds in a red wine aged in oak chips. Anal Chim
vides some insight into the implication of volatile and Acta 513:229–237
nonvolatile composition on the final quality perception of 14. De Coninck G, Jordão AM, Ricardo-Da-Silva JM, Laureano O
astringency. A predictive model between astringency (2006) Evolution of phenolic composition and sensory properties
in red wine aged in contact with Portuguese and French oak wood
intensity and ellagitannins, furfural, guaiacol and methyl chips. Journal International des Sciences de la Vigne et du Vin
guaiacol contents was established. 40:25–34

123
746 Eur Food Res Technol (2013) 236:735–746

15. Gallego L, Del Alamo M, Nevares I, Fernández JA, De Simón 27. Ancı́n C, Garde T, Torrea D, Jimenez N (2004) Extraction of
BF, Cadahı́a E (2012) Phenolic compounds and sensorial char- volatile compounds in model wine from different oak woods:
acterization of wines aged with alternative to barrel products effect of SO2. Food Res Int 37:375–383
made of Spanish oak wood (Quercus pyrenaica Willd.). Food Sci 28. Cerdán TG, Rodrı́guez Mozaz S, Ancı́n Azpilic C (2002) Volatile
Technol Int 18:151–165 composition of aged wine in used barrels of French oak and of
16. Guchu E, D-MM C, Pérez-Coello MS, González-Viñas MA, American oak. Food Res Int 35:603–610
Cabezudo Ibáñez MD (2006) Volatile composition and sensory 29. Prida A, Heymann H, Balanuta A, Puech JL (2009) Relation
characteristics of Chardonnay wines treated with American and between chemical composition of oak wood used in cooperage
Hungarian oak chips. Food Chem 99:350–359 and sensory perception of model extracts. J Sci Food Agric
17. Campbell JI, Sykes M, Sefton MA, Pollnitz AP (2005) The 89:765–773
effects of size, temperature and air contact on the outcome of 30. Cutzach I, Chatonnet P, Henry R, Dubourdieu D (1997) Identi-
heating oak fragments. Aust J Grape Wine Res 11:348–354 fication of volatile compounds with a ‘‘Toasty’’ aroma in heated
18. Sauvageot F, Feuillat F (1999) The influence of oak wood oak used in barrel making. J Agric Food Chem 45:2217–2224
(Quercus robur L., Q. petraea Liebl.) on the flavor of Burgundy 31. Singleton VL (1995) Maturation of wines and spirits: compari-
Pinot noir. An examination of variation among individual trees. sons, facts, and hypotheses. Am J Enol Vitic 46:98–115
Am J Enol Vitic 50:447–455 32. Spillman PJ, Pocock KF, Gawel R, Sefton MA. (1996). The
19. Masson E, Baumes R, Moutounet M, Puech JL (2000) The effect influences of oak, coopering heat and microbial activity on oak-
of kiln-drying on the levels of ellagitannins and volatile com- derived wine aroma. In: Paper presented at the ninth Australian wine
pounds of European oak (Quercus petraea Liebl.) stave wood. industry technical conference, Winetitles, Adelaide, South Australia
Am J Enol Vitic 51:201–214 33. Chatonnet P, Boidron JN, Pons M (1989) Incidence du traitement
20. Rodrı́guez-Rodrı́guez P, Gómez-Plaza E (2011) Differences in thermique du bois de chêne sur sa composition chimique. 2e
the extraction of volatile compounds from oak chips in wine and Partie: Evolution de certains composés en fonction de l’intensité
model solutions. Am J Enol Vitic 62:127–132 de brûlage. Connaiss Vigne Vin 23:223–250
21. Barbe JC, Bertrand A (1996) Quantitative analysis of volatile 34. Doussot F, De JÃÓso B, Quideau S, Pardon P (2002) Extractives
compounds stemming from oak wood. Application to the aging of content in cooperage oak wood during natural seasoning and
wines in barrels. J Cooperage Sci Technol 2:83–88 toasting; influence of tree species, geographic location, and sin-
22. Peng S, Scalbert A, Monties B (1991) Insoluble ellagitannins in gle-tree effects. J Agric Food Chem 50:5955–5961
Castanea sativa and Quercus petraea woods. Phytochemistry 35. Mosedale JR, Puech J-L, Feuillat F (1999) The influence on wine
30:775–778 flavor of the oak species and natural variation of heartwood
23. ISO (2008) 8586-22 Sensory analysis—General guidance for the components. Am J Enol Vitic 50:503–512
selection, training and monitoring of assessors—Part 2: expert 36. Garcı́a-Carpintero EG, Gómez Gallego MA, Sánchez-Palomo E,
sensory assessors González Viñas MA (2012) Sensory descriptive analysis of Bobal
24. Jackson RS (2009) Wine tasting: a professional handbook. Aca- red wines treated with oak chips at different stages of wine-
demic Press, Burlington making. Aust J Grape Wine Res 17:368–377
25. ISO (1997) 3591 Sensory analysis—Apparatus—Wine-tasting 37. Boidron J-N, Chatonnet PMP (1988) Influence du bois sur
glass, p 3 certaines substances odorantes des vins (Effects of wood on
26. ISO (1988) 8589 Sensory analysis. Guide for the installation of a aroma compounds of wine). Connaiss Vigne Vin 22:275–294
chamber for sensory analysis, p 9

123

View publication stats

You might also like