J Sci Food Agric 1993, 62, 169-174
Changes in Wood Extractives from Oak Cask
Staves through Maturation of Scotch Malt
Whisky
John M Conner, Alistair Paterson* and John R Piggott
Food Science Laboratories, Department of Bioscience and Biotechnology, University of Strathclyde,
131 Albion Street, Glasgow GI lSD, UK
(Received 25 August 1992; revised version received 1 March 1993; accepted 23 March 1993)
Abstract: Cask staves, prepared from American white oak, Quercus alba, were
obtained from unused, new charred Bourbon, used Bourbon, first-fill Scotch and
exhausted Scotch casks. The aim of the experiment was to determine the effect of
repeated maturations of whisky on concentration and location of phenolic
extractives in the cask wood. Contents of vanillic and syringic acids and the
respective aldehydes, arising from hydrolysis of guaiacyl-syringyl lignin, and
whisky lactones (cis- and trans-8-methyl-y-octalactone),arising from charring,
were determined. Staves were sampled at 5 mm intervals from the inner char
surface to 25 mm depth and wood extracted with chloroform, heptane and
aqueous ethanol. Contents of material absorbing at 520 nm, total phenols and
vanillic and syringic acids, vanillin and syringaldehyde, were determined by highperformance liquid chromatography ; both isomers of whisky lactone were
quantified by high-resolution gas chromatography. New cask wood had a
maximum for coloured material in the char layer, but maxima for other
compounds at 5mm below the char. As casks were reused in maturations of
whiskies, the contents of aromatic aldehydes and acids were reduced in the first
20 mm of the wood and maxima for both acids and aldehydes shifted towards the
centre of the staves. This suggested that repeated exposure to aqueous ethanol
resulted in compositional changes in the lignin in the wood. The newly charred
wood had a maximum for cis-8-methyl-y-octalactoneat 5 mm below the char and
the trans isomer at 15 mm. In the first Bourbon maturation both isomers were
largely depleted and with successive extractions the maxima moved into the cask
until in the exhausted cask wood no lactone was detected.
Key words: whisky caskwood, whisky maturation, lignin breakdown, aromatic
acids, aromatic aldehydes, whisky lactones, American white oak, Scotch.
producing colour, flavour and more complex sensory
characters collectively described by consumers as spirit
maturity (Guy et all989; Paterson and Piggott 1989).To
enhance the extent and rate of wood polymer breakdown
casks are toasted or charred to produce both aerobic and
anaerobic pyrolysis reactions that will generate solubles
extracted by the aqueous ethanol (Singleton 1974;
Dubois 1989) and cis and trans isomers of P-methyl-yoctalactones (Maga 1989). Casks are reused for maturation until the ability to mature the spirits is perceived
by the distiller as having been lost. At this point, in recent
years, the internal surface is scraped and recharred to
produce a treated wood that appears to be able to effect
INTRODUCTION
In production of Scotch whiskies, a 3-year maturation in
oak casks is required before new distillates can be sold as
the beverage (Philp 1989).During this period the aqueous
ethanol is exposed to oxygen passing through the cask
staves from the exterior and ethanol is lost with the
establishment of a gradient of the alcohol through the
wood. In the maturation, oxidation and ethanolysis
result in a modification and breakdown of wood
polymers into products that diffuse into the spirit
*
To whom correspondence should be addressed.
169
J Sci Food Agric 0022-5142/93/$06.00 0 1993 SCI. Printed in Great Britain
13-2
J M Conner, A Paterson, J R Piggott
170
further maturations. Typically, much-used casks lose the
ability to mature whiskies to an acceptable quality and
are then discarded by the distiller.
Currently, the bulk of Scotch whisky is matured in
casks manufactured from American white oak, frequently Quercus alba although Q muehlanbegii, Q
macrocauba and Q bicolor are also used, and charred for
Bourbon production (Philp 1989). After a single Bourbon maturation the casks are shipped to Scotland and
put to use by the Scotch industry. Many casks are used
initially in malt whisky maturations and subsequentlyfor
grain whiskies until no longer of value to the distiller.
Other distillers, however, use new Bourbon casks initially
for grain maturations and then in malt whisky production.
Nishimura et a1 (1983) have described four pathways
possibly involved in breakdown of lignins in oak during
oxidation and ethanolysis. However, whisky casks vary
in that individual casks have differing histories and
woods from various species of oak, such as Spanish oaks,
are also used. The present study was undertaken to
determine the effect of repeated maturations of whisky
on the concentration and location of lignin-related
phenolic compounds in staves of casks with various
histories and including the effects on whisky lactone
content and distribution.
MATERIALS AND METHODS
Cask wood staves
The staves used in this study came from Bourbon casks
produced from American white oak charred prior to a
single maturation of Bourbon. Wood was taken from
three staves of cask wood before and after a first
maturation of Bourbon, after a second, single maturation
of Scotch whisky, and from a cask, after multiple
maturations of grain whisky, designated by the distiller
as exhausted.
Sampling of wood
Wood was sampled, in triplicate, by removal of aliquots
(5 g) from staves at the char layer and 5, 10, 15, 20 and
25 mm below the interior surface. Shavings were Soxhletextracted with chloroform, dried and re-extracted with
60 % (v/v) ethanol. The solution was filtered through a
Bond Elut C18 sorbent extraction column (Varian
Associates Inc, Harbor City, CA, USA) prior to analysis.
Quantification of coloured material and total phenols
Colour was quantified as the absorbance of wood
extracts at 520 nm (MacDougall 1989). Total content of
phenolics in extracts was determined using the FolinDennis reagent with gallic acid as a standard (Amerine
and Ough 1974).
Quantification of lignin breakdown products
Contents of syringaldehyde, vanillin, and syringic and
vanillic acids were determined by HPLC using a
Spherisorb ODS 5 pm column (25 cm x 4 mm id; Phase
Separations Ltd, Deeside, UK) and a gradient from
50 ml litre-' ('5 % ') aqueous formic acid to methanol
(Casteele et a1 1983; Conner et a1 1989).
Quantification of whisky lactones
Wood samples (1 g) were shaken with heptane (50 ml) in
sealed tubes overnight at room temperature. Extracts
were evaporated in U ~ C U O to 2ml for analysis by
high-resolution gas chromatography performed using a
Carlo Erba Mega 5300 HRGC equipped with a splitsplitless injector and flame ionisation detector (Fisons
Instruments Ltd, Crawley, UK) with a 25 m x 0.32 mm
Carbowax BP20 column, 0.5 p phase thickness (SGE
UK Ltd, Milton Keynes, UK). Carrier gas was helium at
1.8 ml min-'. Samples were analysed in duplicate with an
initial temperature of 60"C, increasing after 3 min to
230°C at 4°C min-'. Analyte peak areas were calculated
by a TRIO computing integrator (Trivector Systems
International Ltd, Sandy, UK) and standardised on 2,3dimethylphenol (10 pg ml-') in the extract. Variation
between replicates was f5 % standardised peak area or
better. Peak assignments were established by highresolution gas chromatography-mass spectrometry
(GC-MS) using a Finnegan MAT ITS40 integrated
benchtop GC-MS and data analyser (Finnegan MAT
Ltd, Hemel Hempstead, UK) with the same column and
conditions.
RESULTS
Wood samples
It was not possible to obtain staves manufactured from
the same batch of wood or of similar histories and was
thus impossible to compare individual values at a single
depth between staves, but comparisons of profiles of
staves with respect to depth, sampled in triplicate, were
regarded as valid.
Gradients of extract colour
Extract colour (Fig 1) varied in relation to the depth and
cask history. In new wood and Bourbon casks, colour
was maximal in the char layer and decreased markedly
by 5 mm, and continued to decrease in new wood but
was relatively constant and lower after the Bourbon
maturation. After a single Scotch maturation, the second
maturation overall, colour was further reduced and very
low concentrations were obtained in exhausted cask
samples with a small maximum at 20 mm.
Aromatic compounds in whisky caskwood
171
3.:
Used Bourbon
-a-
70
+First-fill Scotch
Used Bourbon
tFirst-fill Scotch
-A-
3
I
Char
5
10
15
25
20
Depth into wood (mm)
04
Char
I
10
5
15
20
25
Depth into wood (mm)
Fig 1. Distribution of material absorbing at 520 nm in inner
25 mm of unused, used Bourbon, single-fill Scotch and
exhausted Scotch cask staves.
Fig 4. Distribution of syringaldehyde in inner 25 mm of the
unused, used Bourbon, single-till Scotch and exhausted Scotch
cask staves.
Total phenols in extracts
1
0.7
Gradients in total phenols content in woods were
observed (Fig 2) in used but not in new casks. After a
single maturation, of bourbon, and a second maturation,
of Scotch, marked maxima were observed at 10mm
depth. In the exhausted cask the maxima were shifted to
20 mm depth and had markedly increased in breadth.
0.3
Aromatic aldehydes: vanillin and syringaldehyde
0.9
0.8
0.2
--t
0.1
First-fill Scotch
0
Char
5
10
15
20
25
Depth into wood (mm)
Fig 2. Distribution of total extractable phenolic material in
inner 25mm of the used Bourbon, single-fill Scotch and
exhausted Scotch cask staves.
Extracts showed maxima for both vanillin and syringaldehyde in used casks in the char and in new wood at
5 mm. Further maxima were observed at 20 mm after
first and second overall spirit maturations, with minima
at 5 and 10 mm, respectively (Figs 3 and 4). There was
two- to three-fold more syringaldehyde than vanillin in
extracts and exhausted caskwood yielded very low
concentrations of aromatic aldehydes. New and used
caskwoods had distinctly different profiles.
Aromatic acids: vanillic and syringic acid
Profiles of vanillic and syringic acids (Figs 5 and 6) were
similar, but approximately two- and four-fold lower than
the respective aldehydes. New wood had maxima at
5 mm, Bourbon and first-fill Scotch caskwoods had
maxima in the char and at 20 mm depth and minima at
5 and I0 mm depth, respectively. Extracts of exhausted
wood had very low acid contents increasing with depth.
/3-methyl-y-octalactone contents
Char
5
10
15
20
25
Depth into wood (rnm)
Fig 3. Distribution of vanillin in inner 25 mm of the unused,
used Bourbon, single-fill Scotch and exhausted Scotch cask
staves.
Neither cis-nor trans-P-methyl-y-octalactonewas present
in exhausted wood, and in newly charred wood maxima
were observed at 5 and 15 mm depth for the isomers
respectively, with a cis to trans ratio of 5 (Figs 7 and
8). After a single Bourbon maturation, a marked
-
J M Conner, A Paterson, J R Piggott
172
300
250
+Used Bourbon
I t First-fill Scotch
h
40
200
50
/
/
lo!
5
0
0
Char
5
10
20
IS
5
Char
25
10
Fig 5. Distribution of vanillic acid in inner 25 mm of the
unused, used Bourbon, single-fill Scotch and exhausted Scotch
cask staves.
t New
20
25
charred wood
Fig 8. Distribution of trans-p-methyl-y-octalactone in inner
25mm of the unused, used Bourbon, single-fill Scotch and
exhausted Scotch cask staves.
T
A
0
Char
5
l6
New charred woodl
-+Used Bourbon
\
t First-fill
Scotch
--c Exhausted Scotch
80
15
Depth into wood (mm)
Depth into wood (mm)
70
30
20
1
10
04
Char
-I
5
10
15
20
25
Depth into wood (mm)
Fig 6. Distribution of syringic acid in inner 25 mm of the
unused, used Bourbon, single-fill Scotch and exhausted Scotch
cask staves.
300
250
T
+New
+Used
t n
5
15
20
25
Fig 9. Ratio of cis to trans-P-methyl-y-octalactone
in inner
25 mm of the unused, used Bourbon, single-fill Scotch and
exhausted Scotch cask staves.
progressive movement of the maxima into the wood with
successive maturations.
charred wood
Bourbon
DISCUSSION
#
500Char
10
Depth into wood (mm)
10
15
20
25
Depth into wood (mm)
Fig 7. Distribution of cis-p-methyl-y-octalactone in inner
25mm of the unused, used Bourbon, single-fill Scotch and
exhausted Scotch cask staves.
depletion of both isomers was observed and this was
continued in the second (ie Scotch) maturation. Plotting
the ratio of cis to trans isomers (Fig 9) showed a
The charring of the inner cask surface results in pyrolysis
of the oak, initially in the presence of oxygen, but
subsequently in the absence as the partial pressure of
carbon dioxide increases, generating aromatic aldehydes
below the surface that can be extracted into the spirit
(Nishimura et al 1983). Charring also generates the two
isomers of /I-methyl-y-octalactone with maxima at
differing depths, according to temperature, from nonlignocellulose wood components (Maga 1989), possibly
2-methyl-3-(3,4-dihydroxy-5-methoxybenzo)-octanoic
acid (Otsuka et a1 1980). The maturation of the 65%
(v/v) ethanol whisky distillates effects ethanolysis, or
more correctly hydrolysis, of the lignin-carbohydrate
matrix as the solution penetrates the wood (Reazin
1983). A third contribution to wood behaviour is the
Aromatic compounds in whisky caskwood
migration of oxygen through the wood from the exterior
of the cask (Nishimura et aZ 1983).
In a detailed study of the first 6 mm of wood in the
internal surface of a Limousin oak (Quercus sessilis)
cask, Chatonnet (1991) showed that after a single wine
maturation there was a distinct gradient of eugenol
maintained during five successive wine maturations and
an inverse relationship with respect to syringol and ethyl4-phenol contents, the last two compounds being at trace
levels in cask wood after one and three maturations but
very marked after five. Concentrations of the two isomers
of P-methyl-y-octalactonewere observed to increase with
depth and were markedly higher after five maturations.
Chatonnet observed that the cis to trans ratio increased
with depth of sampling and cis was more abundant than
the trans isomer. In contrast, Maga (1989) had reported
that in Q alba wood the trans was more abundant than
the cis isomer and both were effectively extracted by 40
and 60% (v/v) ethanol over a 32 month period.
Chatonnet (1991) also studied the inner surface of oak
caskwood after charring for periods between 5 and
20 min, using electron microscopy. It was clear that the
charring had resulted in partial breakdown of the
structure of the wood with erosions of the surface of the
lignocellulose. Scanning electron microscopy revealed
cracking of the surface lamellae and damage to wood
pores. The results in our present and an initial study
(Conner et a1 1989) indicate that charring has a maximal
effect in the first 10 mm of wood from the surface and
was not restricted to the 6 mm analysed by Chatonnet
(1991). In both newly charred oak and after one
maturation, maximal colour, uncorrelated with
concentrations of phenols, was observed in the char. This
may reflect the presence of other wood breakdown
products.
Maximal concentrations of the sensorily-important
phenols were observed at 5 mm below the char in the new
wood with markedly higher concentrations of syringaldehyde and syringic acid than vanillin and vanillic acid
as reported previously (Reazin 1983). Successive maturations shifted minima to 5 and 10mm and, after the
first maturation, maxima were observed in the char
suggesting migration of the aromatic compounds from
the interior of the caskwood towards the aqueous
ethanol. In the exhausted wood, depletion of aromatic
acids and aldehydes, and P-methyl-y-octalactones, was
observed.
These results confirm that charring and pyrolysis
results in formation of syringic and vanillic acids, and
syringaldehyde and vanillin, in the wood immediately
below the char layer. Thermal depolymerisation is
accompanied by reactions leading to formation of
P-methyl-y-octalactones from wood extractives at
abundancies related to their chemical structure. The
successive whisky maturations resulted in extractions
of low-molecular-weight phenolic material into the
maturing spirit followed by progressive ethanolysis of
173
susceptiblelignin bonds from successively deeper regions
of the wood. In parallel there will be continued
breakdown of lignin remnants in the char and lignin
oligomers extracted into the spirit. The progress of
ethanolysis can be observed as the appearance
of acetovanillone (4-hydroxy-3'-methoxyacetophenone)
that progresses from the inner surface of first-fill
caskwood towards the outer surface of the exhausted
cask as reported previously (Conner et aZl989). Oxygen
diffusion through the wood does not appear to be a
limiting factor in the process. With repeated maturations,
bonds susceptible to ethanolysis are depleted. Scraping
and recharring of the inner surface of well-used casks
may result in an initially enhanced ability of the wood to
mature distillates, as the compounds generated beneath
the char are extracted, followed by a rapid decline as
bonds susceptible to ethanolysis had been depleted
by the maturations effected prior to the recharring.
Although lignin is not normally considered as a polymer
with labile bonds it is clear from the work of Freudenberg
et aZ(1964), and the conclusions of Nimz (1974) on beech
lignin structure, that a number of classes of labile bonds
exist. Firstly, acyclic a-0-4 bonds account for 8-10 % of
bonds in spruce lignin and are considered to be
susceptible to hydrolytic cleavage under mild conditions
giving rise to low-molecular-weight cleavage products.
Furthermore, ester linkages in lignin have been reported
by Lapierre et aZ (1982) in poplar and Smith (1955) in
aspen and may be present in oak. Nimz (1 974) has shown
that after several weeks of percolation with water at
100°C beechwood (Fagus silvatica) loses 40% of its
lignin component whereas spruce (Picea excelsa) loses
20% of its lignin into solution.
Neither cis- nor trans-/l-methyl-y-octalactone were
present in significant concentrations in woods used for
maturation of Scotch but were present in the caskwood
used for Bourbon maturation and extracted in this
process. Maga (1989) has recorded the odour thresholds
of cis and trans isomers as 790 and 5MOpgkg-',
respectively: concentrations of these compounds in
newly charred wood were 0-047 and 0.254 pg kg-' at
maxima, and 10-fold less in Scotch casks. These results
suggest that reuse of American white oak casks will not
provide the concentrations of the lactone required for
Bourbon character. Thus, these compounds are unlikely
to play a major role in flavour in Scotch whisky, as was
proposed by Sharp (1982). Other compounds may form
a pool that is converted to the lactones. No quantification
of lactone precursors was performed in this study.
ACKNOWLEDGEMENTS
The authors wish to express their gratitude for support
from the Agricultural and Food Research Council and
Chivas Brothers (Keith) Ltd, who also supplied cask
staves for this study.
J M Conner, A Paterson, J R Piggott
174
REFERENCES
Amerine M A, Ough C S 1974 Wine and Must Analysis. John
Wiley and Sons, New York, USA.
Casteele K V, Geiger H, van Sumere C F 1983 Separation of
phenolics and coumarins by reversed-phase high perfomance liquid chromatography. J Chromatogr 258 111-124.
Chatonnet P 1991 Incidences du bois de chine sur la
composition chimique et les qualitis organoleptiques des
vins. Thesis, University of Bordeaux 11, France.
Conner J M, Paterson A, Piggott J R 1989 The distribution of
lignin breakdown products through new and used staves. In:
Distilled Beverage Flavour: Recent Developments, ed Piggott
J R & Paterson A. Ellis Horwood, London, UK, pp 177-184.
Dubois P 1989 Apport du fi3 de chine neuf a l'ar6me des vins.
Rev Fr Oenoll20 19-24.
Freudenberg K, Harkin J M , Werner H-K 1964 Das
Vorkommen von Benzylarylathern im Lignin. Chem Ber 97
909-920.
Guy C, Piggott J R, Mane S 1989 Consumer profiling of
Scotch whisky. Food Qua1 Pref 1 69-73.
Lapierre C, Lallemand J Y, Monties B 1982 Evidence ofpoplar
lignin heterogeneity by combination of 13C and lH NMR
spectroscopy. Holzforschung 36 275-282.
MacDougall D B 1989 Measurement of food and beverage
colour appearance. In : Distilled Beverage Flavour: Recent
Developments, ed Piggott J R & Paterson A. Ellis Horwood,
London, UK, pp 85-96.
Maga J A 1989 Formation and extraction of cis- and trans$methyl-y-octalactone from Quercus alba. In : Distilled Bev-
erage Flavour: Recent Developments, ed Piggott J R &
Paterson A. Ellis Horwood, London, UK, pp 171-176.
Nimz H 1974 Beech lignin: proposal of a constitutional
scheme. Angew Chem Int Edn 13 313-321.
Nishimura K, Ohnishi M, Masuda M, Koga K, Matsuyama R
1983 Reactions of wood components during maturation. In:
Flavour of Distilled Beverages-Origin and Development, ed
Piggott J R. Ellis Horwood, Chichester, UK, pp 241-255.
Otsuka K, Sat0 K, Yamashita T 1980 Structure of a precursor
of P-methyl-y-octalactone, an aging flavor compound of
distilled liquors. J Ferment Techno1 58 395-398.
Paterson A, Piggott J R 1989 The contributions of the process
to flavour in Scotch malt whisky. In: Distilled Beverage
Flavour: Recent Developments, ed Piggott J R & Paterson A.
Ellis Honvood, London, UK, pp 151-169.
Philp J M 1989 Cask quality and warehouse conditions. In:
The Science and Technology of Whiskies, ed Piggott J R ,
Sharp R & Duncan R E B . Longman, London, UK,
pp 264294.
Reazin G H 1983 Chemical analysis of whisky maturation. In:
Flavour of Distilled Beveragesorigin and Development, ed
Piggott J R. Ellis Horwood, Chichester, UK, pp 225-240.
Sharp R 1982 Analytical techniques used in the study of whisky
maturation. In : Current Developments in Malting, Brewing
and Distilling, ed Priest F G & Campbell I. Institute of
Brewing, London, UK, pp 143-156.
Singleton V L 1974 Some aspects of wooden container as a
factor in wine maturation. In: Chemistry of Winemaking, ed.
Webb A D. A C S Pub1 Co., New York, USA, p 310.
Smith D C C 1955 p-Hydroxybenzoate groups in the lignin of
aspen (Populus tremula). J Chem SOC2347-2351.