91av

Whisky drinkers are true romantics. They believe that making whisky
is a mysterious and wonderful art, its flavour the result of subtle differences
between Highland or Lowland air and water, or the effect of sea salt, heather
or peat. You can almost hear chemists darkly muttering ‘hocus-pocus’. But
sceptics should beware: despite the predominance of alcohol and water, demystifying
the molecules in a bottle of Scotch is trickier than it sounds.

Thirty years ago, chemists intent on discovering the secrets of Scotch
would have concentrated their efforts on long-chain alcohols produced during
fermentation. During the past decade, however, advances in analytical techniques
have revealed that Scotch contains minute quantities of as many as 500 different
molecules. Now it is beginning to look as though the unique flavour of whisky
depends less on the influence of one or two dominant compounds than on an
extremely complex series of chemical reactions, going on at different rates
and various times. Not all the 500 molecules produced in these reactions
may be important, but finding out where they come from and how they fit
into the chemistry is essential before we can understand the flavour of
whisky. Although fermentation mattera, the real key seems to lie in the
final step of the whisky-making process when many of these chemical reactions
take place. What happens at this last stage – maturation – remains a puzzle.

Malt whiskies, renowned for their distinctive flavours, are traditionally
produced in a copper ‘pot still’. While chemists look for answers in the
various stages of this process, according to traditional practice the last
word in the individuality of a whisky belongs to the blender. Whisky manufacturers
rely on skilful blenders with a nose for whisky, often acquired through
long years of apprenticeship in the trade. Their job is to blend together
spirit from different casks – each of a different colour, aroma and flavour
– so that each batch is consistent with the last. Even single malts are
blends of whiskies from different casks. A good blender, helped by a tasting
panel of another half-dozen people, can ‘nose’ 2000 samples a week, and
has no fear of being replaced by chemical technology. Using gas-liquid chromatography,
a sensitive technique familiar to analytical chemists, it would be possible
to analyse the content of only 140 samples a week, working round the clock.
Even then, such an analysis would be a poor guide to flavour. Blenders use
a rich and detailed vocabulary to describe the various aromas of spirits
from different distillations, from rancid to rubbery and peaty to phenolic,
and the ‘formula’ for a particular blend is a closely guarded secret within
each distillery.

So how does a whisky acquire its distinctive flavour? Both blenders
and chemists already have some clues. They know that using different types
of barley has a strong effect, for example. And most people can recognise
a peated malt, produced from barley that has had peat smoke blown through
it, a relic of the days when barley was dried over fires fuelled by peat
or heather. Peat smoke contains organic ring compounds that have double
carbon-carbon bonds. Chemists call all kinds of carbon-based rings ‘aromatic’,
but the ones found in peat smoke have a distinctive aroma and they give
the whisky a pungent flavour. Other flavours derive from the distillation
process itself. At this point the type of yeast is important, and distillers
often add carefully selected varieties of spent brewer’s yeast to give particular
flavours. Even the shape of the still is crucial; complex chemical reactions
take place on its surface, and different shapes will select and condense
different compounds. Especially important during distillation is the breakdown
of sulphur-containing hydrocarbons such as dimethyl sulphide. Trace amounts
of copper from the still probably catalyse this reaction, which may explain
why malt whisky is traditionally distilled in copper, not glass.

When Scots first distilled whisky, they would have drunk the new spirit
without waiting for it to mature. But this meant that whisky making had
to be an activity for winter, just after the barley harvest and at a time
when there were adequate supplies of water. Maturation in oak casks probably
evolved as a form of storage, but today newly distilled spirit is considered
undrinkable. Blenders know it as a liquid with a rather unpleasantly sharp,
pungent aroma. Chemists have a good idea what it contains, in addition to
ethanol and water: phenols, from malt and peak smoke, along with esters,
lactones, aldehydes and some compounds that contain sulphur and nitrogen.
These molecules provide the basic flavour, but by law the spirit is not
Scotch whisky until it has been left to mature in an oak cask for at least
three years.

Chemists began to study the changes that take place in maturing whiskies
in the early years of this century. They still have a long way to go. All
they can say for sure is that a decade or so is long enough for the slowest
moving alcohol to be transformed into something more interesting. It is
a complex and lengthy process, involving many different reactions. Oxygen
diffuses into the cask, mixing with the vapours already present in the ‘headspace’
above the new spirit, and ethanol and water evaporate out. In the liquid
itself, some alcohols are oxidised to aldehydes, and aldehydes to acids;
other alcohols react with acids to give esters, or with aldehydes to form
acetals. As many as 17 years later, the whisky is pronounced mature.

In the early 1980s, George Reazin and his colleagues at the whisky manufacturer
Joseph Seagram of Louisville, Kentucky tried to work out what goes on in
a cask of maturing whisky. They studied American whisky (‘whiskey’), but
their results are just as valid for Scotch because the same basic processes
apply to the maturation of both. They found that there were some physical
effects: the concentration of the whisky when it enters the cask affects
the levels of some flavour molecules, while the temperature at which the
cask is stored affected how quickly the molecules form. Traditionally, malt
whisky begins maturation at about 63 per cent alcohol and Scotland’s climate
is allowed maximum influence over casks stored in unregulated warehouses.

As for the chemistry, Reazin and his colleagues concluded that some
of the final flavours come from chemical reactions that take place between
molecules already in the distillate, while some aromatic compounds, such
as syringaldehyde and vanillin – the origin of sweet, ‘woody’ aromas – come
from the breakdown of the cask. This was a challenge. Although chemists
know the structure of cellulose, they know much less about the other compounds
in wood, such as hemicelluloses and lignin.

Lignin is the stuff that makes trees ‘woody’. It seems to be a random
polymer that binds to hemicellulose in the plant cell walls, forming a resin
that surrounds the cellulose fibres which give the cell wall its structural
strength. The polymer is made up of three monomers, each containing a benzene
ring with side chains containing carbon atoms joined by double bonds. The
monomers can be linked in various ways: side chain to ring, ring to ring
and side chain to side chain. Bonds between side chains are the easiest
to break and so these can be attacked by ethyl alcohol and water in a process
called hydrolysis. It seems likely that during maturation, the young whisky
extracts wood components from the cask in this way.

Russian chemists were the first to realise that the molecules produced
by the breakdown of lignin were important to maturation when, during the
1950s, they identified the aromatic aldehydes syringaldehyde and vanillin
in brandy. Later, other researchers turned their attention to whiskies.
In 1983, Kiichi Nishimura and his colleagues from the Suntory company in
Osaka worked out in more detail how compounds related to lignin become part
of mature whisky.

Despite their efforts, the breakdown of lignin is still not entirely
understood. It is thoght to give rise to molecules such as coniferyl alcohol,
sinapaldehyde and vanillic acid, precursors of related molecules such as
syringaldehyde and vanillin that are known to contribute to flavour. With
funding from the Agricultural and Food Research Council and the Scotch whisky
industry, John Piggott, Alistair Paterson and John Conner from the University
of Strathclyde have just completed a three-year study of how molecules formed
from the breakdown of lignin and tannin from cask wood affect whisky maturation.
They studied one type of breakdown product, aromatic acids and phenols,
to find out how much is extracted from the cask wood, and exactly where
in the cask the lignin breaks down.

During maturation, the alcohol and water penetrate the entire depth
of cask wood. The Strathclyde team discovered that as the whisky matures,
the inner faces of the cask quickly become devoid of phenols and easily
degradable lignin. As the cask ages, these molecules are found in the highest
concentrations deeper inside the wood.

Maturation limits the life of any cask. Less than 4 per cent of the
total lignin in an oak cask can be broken down easily by the spirit it contains.
In practice, most casks contain whisky for 50 years or more, until the aroma
of whisky gives telltale signs that all the easily degradable lignin has
been used up. From this point onwards the casks no longer produce a good
maturation; the oak contributes little to the final taste and aroma. ‘The
spirit that comes out of the barrel smells worse than when it was put it,’
says Conner. ‘You start to get some very peculiar flavours.’

In 1974, many of the largest whisky manufacturers in Scotland formed
a consortium to study ways of improving whisky and its methods of production.
Most of the research done at Pentlands Scotch Whisky Research Institute
near Edinburgh remains firmly under wraps, but the consortium is interested
in developing new strains of yeast and in analysing maturation. Chemists
at Strathclyde and Pentlands have now identified many of the chemical changes
that go on during maturation. For example, flavours associated with sulphur-based
compounds and certain aldehydes decrease, whereas there is an increase in
flavours associated with vanillin and wood components as these are extracted.
Chemists can identify many of the compounds – the problem is tying them
into particular flavours. Some molecules, such as those containing sulphur
and nitrogen, are present only in minute quantities, parts per billion,
but you can smell them nevertheless. Other molecules less crucial to the
flavour may be present in much larger quantities.

Timing is important for many of the reactions. The whisky industry would
like to speed up the maturation process – manufacturers lose a costly 2
per cent per year in evaporation through the casks, which is officially
recognised by the Customs & Excise in Scotland. Conner believes this
is impossible, because no one can predict how speeding up one reaction will
affect the others. ‘You can follow the changes that take place in aroma
over time, but pinning down the change to particular components is difficult
because they all change at the same time,’ he says.

Perhaps the most difficult quality of scotch to define and relate to
chemistry is ‘smoothness’. What makes people describe one malt as ‘smooth’
or ‘mellow’, and another as ‘harsh’ appears to be more of a physical than
a chemical characteristic. A theory favoured by Nishimura and his colleagues
is that the spirit matures, hydrogen bonds form between the alcohol and
the water molecules, so that they form stable clusters. Piggott and Conner
aim to repeat the Japanese work and extend it, using both whiskies and model
solutions, by measuring energy changes caused by hydrogen bonding and identifying
the structures that form between molecules.

They are particularly interested in tracking the fate of compounds containing
sulphur and nitrogen because their contribution to the flavour decreases
steadily as the new spirit is transformed into a mature Scotch. Wood components
appear to be closely linked with this decrease. Conner thinks that as the
concentration of oak extracts increase, so does the solubility of other
components such as hexadecanol and tetradecanol – long chain alcohols with
at least 14 carbons – and esters of organic acids with between 10 and 18
carbons. It seems likely that when the concentration of wood components,
especially phenols, reaches a certain level, the alcohols and esters begin
to form loose groups, perhaps with the sulphur and nitrogen-containing hydrocarbons.
This makes them less free to move in the liquid and therefore less available
for evaporation into the ‘headspace’ above the liquid, where they can contribute
to the aroma.

Researchers in countries not noted for their whisky are equally keen
to discover the secrets of Scotch. Lalli Nykanen heads the analytical department
of the research laboratories of Alko, the Finnish state alcohol company,
and his team of researchers has spent years trying to find out what makes
whisky – Canadian, American and Scotch – the way it is. So far they have
pinpointed some 400 different flavour components in whisky, mainly fatty
acids, esters, alcohols and aldehydes. But even after a decade they have
not managed to track down a basic chemical recipe common to the best brands.
‘It’s very difficult to say what the main components of the flavour are,’
says Nykanen. ‘Whisky contains many fusel (long chain) alcohols but aldehydes
are also very important to the flavour, even though they are present in
much smaller quantities. Almost all fatty acids and fatty acid esters from
acetic to stearic occur in whisky. Esters of fatty acids are very volatile,
so they give the whisky a strong flavour.’

In one experiment, Nykanen and his colleagues pickled oak chips in alcohol
to mimic the maturation process. The alcohol freed 44 different varieties
of acids, lactones and esters from the oak. ‘We have found that there are
600 to 800 compounds in whisky, but how they are related and how they affect
the flavour is not clear,’ says Nykanen. ‘You smell and taste whisky using
your nose and your mouth, but combining sensory tests with chemicals analysis
is very difficult. We have tried using statistical methods and computers,
but it’s not easy. I think it will be 5 or 10 years before we can say we
know what makes the flavour.’

Alko has been producing two brands of whisky for the past 10 years,
based on the results of their research. In the search for excellence, the
company imports brewer’s and distiller’s yeasts from several countries,
including Scotland. There is good evidence that Alko has not yet succeeded
in developing a whisky to rival Scotch, however: Nykanen prefers to drink
The Glenlivet, a single Highland malt that matures for a leisurely 12 years.

In time, chemists will eventually analyse and identify all the flavours
in the distillate. Using the latest techniques, combining gas chromatography
and mass spectrometry, they can detect compounds present at the parts per
billion level. One problem with the most sensitive method is that what the
chemists are measuring is not precisely the same thing that the blenders
are nosing – the headspace of vapours above a standard measure of whisky
in a standard, tulip-shaped glass. The nearest they get is probably to take
a syringe full of vapour from above a sample of whisky. Even then, some
of the heavier compounds may not always be present in the vapour in sufficient
quantities to be detected.

So far, not even chemists working in the heart of Scotland have managed
to find the right mix of molecules in the right proportions. It seems that
Scotch will be baffling chemists for some years yet. ‘The way it’s made
is an art,’ argues Conner. ‘You can’t do without the blenders. What the
analytical chemists should be able to do is provide some predictions of
how a whisky will mature and which barrel to use, so the whisky from a particular
distillery is consistent. But you couldn’t produce a whisky in a test tube
and compare it with one that has been maturing for 8 years. Anyway, it wouldn’t
be Scotch.’ After all, perhaps 15 years isn’t too long to wait for a good
malt.

* * *

Defining the art of distillation

Malt whisky is made in batches by the ‘pot still’ process. This begins
by fermenting malted barley, unadulterated with any other type of grain.
Malting is the process of steeping and drying barley. First, the barley
is allowed to germinate for about 10 days. During this time it produces
a collection of enzymes, called diastase, that convert stored starch into
a soluble form. Just as the barley is beginning to sprout, the distiller
stops the germination by heating it. The barley is dried, ground, then mixed
with hot water. This sugary liquid, the wort, is transferred to a fermentation
vessel where yeast is added to convert the sugar to alcohol. After about
48 hours this impure liquid, the wash, is distilled twice, traditionally
in pear or onion-shaped stills made from copper.

The first distillation converts the wash into ‘low wines’ which are
about 26 per cent alcohol. These low wines are then distilled in a second
copper still. The first portion of this distillate, foreshots, contains
many pungent and unwanted flavours and is recycled to the wash still. Then
comes the new whisky, at about 75 pe cent alcohol, which is collected in
a spirit receiver. The final portion, feints, contains much less alcohol
and is partly recycled and partly thrown away, A skilled operator relies
on nothing more complicated than a hydrometer to indicate when to start
and finish collecting the whisky. Then the spirit is transferred to an oak
cask ready for the long process of maturation. Finally, it is blended and
bottled.

Single malt whisky is usually produced by a single distillery, but is
a blend of whiskies from several distillations. The specialist ‘single-cask’
malt whiskies are produced from a single distillation. Grain whisky is distilled
in a continuous process from a mixture of malted barley and unmalted cereals,
usually maize. Most people – including Scots – drink blended whiskies. These
are mixtures of malt and grain whiskies, deriving from as many as 50 distilleries,
each with its individual flavour. The best blends comprise as much two-thirds
malt whisky, diluted with the more neutral-tasting grain whisky. The process
used to distil grain whisky is designed to be much more efficient than the
pot still process, so that grain whisky contains far fewer flavour compounds.
Most of the flavour in a blended whisky comes from the malt it contains.

* * *

Roll out the barrel for flavour

Scotch is matured in casks made from the wood of two groups of oak species:
the American white oak and the European oak. Almost half of all white oak
casks are of Quercus alba, which grows in Missouri, Kentucky and Arkansas,
while European oak is mainly Quercus robur from parts of France, Spain and
Portugal.

Scotch is never matured in new casks because these quickly release strongly
flavoured compounds that would spoil the final flavour. Bourbon, on the
other hand, has to be matured in new casks by law, and it has a sweet, vanilla,
very woody aroma as a result. Fewer secondary flavour constituents develop
in whisky after 12 years in an aged cask than after two years in a new white
oak cask.

Historically, many scotch whiskies owed at least some of their distinctive
flavour to a taste of Spain. Scotch manufacturers have relied for more than
century on used sherry butts (a butt holds between 480 and 520 litres) in
which to mature their spirit, and many still do. Paul Rickards, director
of the cooperage company which supplies Lang Brothers, a malt whisky distiller
based in Glasgow, still makes frequent visits to Spain to choose the casks
he wants. Spanish sherry producers use Amercian oak for fino and amontillado
sherries and the European or Spanish oak for oloroso. The different varieties
of sherry give a subtly different flavour and colour to the cask and therefore
to the whisky that will inhabit it for at least the next three years. Olorosos
give a darker colour; finos a lighter one.

In the 1960s sherry producers began to use stainless steel casks for
shipping sherry, and many Scotch manufacturers now mature their spirits
in used bourbon casks made of American oak. This has changed the flavour
of the whiskies matured in them to some extent, but by the time the cask
is used for whisky up to 70 per cent of the woody flavour will already have
been lost to the bourbon. So the individuality of Scotland’s distilleries,
which also depends on minor differences in the type of yeast, malt, and
the shape of the stills, is retained.

The heartwoods of the oaks Q alba and Q robur are very similar: around
40 per cent is cellulose, a fifth is hemicellulose, and almost a third is
lignin. The remaining 10 per cent, which can be extracted by hot water or
ether, is a chemist’ delight – volatile oils, acids, sugars, steroids, tannins,
pigments and inorganic compounds, a vast repertoire of complexes, chains
and rings. Many of these compounds react with ethanol as the whisky matures.
The main volatile components in wood are oak lactones – rings with four
carbons and an oxygen that are also found in mature Scotch.

Sometimes the casks are charred on the inside. The reactions that take
place on the surface of the wood during charring are complex and poorly
understood. Chemists think they include the production of a layer of ‘active’
carbon which removes some unwanted compounds. Charring also begins the process
of breaking down lignin in the inner face of the cask, releasing flavour
molecules more quickly into the whisky. Some whisky producers scrape and
rechar casks once they have been used. This seems to prolong the life of
the cask, but no one yet knows how often they will be able to repeat the
process before the structure of the cask wood begins to break down.