Advances in Flavours and Fragrances

From the Sensation to the Synthesis

Advances in Flavours and

Fragrances
From the Sensation to the Synthesis

Edited by
Karl A.D. Swift
Quest International, Ashford, Kent, UK

RSeC
ROYAL SOCIETY OF CHEMISTRY

The proceedings of Flavours and Fragrances 2001 : From the Sensation to the Synthesis held
on 16-18 May 2001 at the University of Wanvick, Coventry, UK.

Special Publication No. 277

ISBN 0-85404-82 1-9

A catalogue record for this book is available from the British Library
0 The Royal Society of Chemistry 2002

All rights reserved.

Apart from any fair dealing for the purpose of research or private study, or criticism or
review as permitted under the terms of the UK Copyright, Designs and Patents Act, 1988,
this publication may not be reproduced, stored or transmitted, in any form or by any means,
without the prior permission in writing of The Royal Society of Chemistry, or in the case of
reprographic reproduction only in accordance with the terms of the licences issued by the
Copyright Licensing Agency in the UK, or in accordance with the terms of the licences
issued by the appropriate Reproduction Rights Organization outside the UK. Enquiries
concerning reproduction outside the terms stated here should be sent to The Royal Society of
Chemistry at the address printed on this page.
Published by The Royal Society of Chemistry,
Thomas Graham House, Science Park, Milton Road,
Cambridge CB4 OW,UK
Registered Charity No. 207890
For further information see our web site at www.rsc.org
Printed in Great Britain by TJ International Ltd, Padstow, Cornwall

Preface
This book is a compilation of sixteen of the twenty papers presented at the 2001
RSC/SCI flavours and fragrances conference at Scarman House, University of Warwick.
The meeting was spaced over two and a half days and saw speakers and delegates
from all corners of the world exchanging ideas and information.
The aim of the meeting was to bring together scientists from both industry and the
academic world, who have a cornmon interest in the chemistry of flavours and fragrances. The subject matter was intentionally broad, covering areas such as biochemistry of receptors/structure activity relationships, analytical techniques, natural products/essential oils, organic and bioorganic chemistry, and flavours/foods. The book is
divided into the same sections as the original meeting.
The chapters contained in this book have been rapidly edited and proof read by
the editor only. Every effort has been made to ensure that no mistakes are present but
inevitably it is likely that some still exist! The editor also asks that the reader is
understanding of the fact that most chapters have been written by people who are not
native English speakers.
Finally, I would like to thank everybody who contributed to the 2001 conference
and made it such a success.

Contents
Structure Activity Relationships
Structure Activity Relationships and the Subjectivity of Odour Sensation
Thomas Markert
Relationship of Odour and Chemical Structure in 1- and 2-Alkyl Alcohols
and Thiols
Y. Sakoda and S. Hayashi

15

Analytical
New Developments in Sorptive Extraction for the Analysis of Flavours and
Fragrances
P. Sandra, F. David and J. Vercammen
Application of Chromatographic and Spectroscopic Methods for Solving
Quality Problems in Several Flavour Aroma Chemicals
Michael Zviely, Reuven Giger, Elias Abushkara, Alexander Kern,
Horst Sommer, Heinz-Juergen Bertram, Gerhard E. Krammer,
Claus Oliver Schmidt, Wolfgang Stumpe and Peter Werkhoff

27

39

Natural Products and Essential Oils

Commercial Essential Oils: Truths and Consequences
Brian Lawrence

57

Stable Isotopes for Determining the Origin of Flavour and Fragrance

Components: Recent Findings
Daniel Joulain

84

Fragrant Adventures in Madagascar: The Analysis of Fragrant Resin from

Canarium madagascariense
Robin Clevy

92

The Effect of Microgravity on the Fragrance of a Miniature Rose, Overnight

Scentsation on Space Shuttle (STS-95)
Braja D. Mookherjee, Subha Pate1 and Weijia Zhou

99

vii

...

Advances in Flavours and Fragrances

Vlll

Organic and Bioorganic Chemistry

Ambergris Fragrance Compounds from Labdanolic Acid and Larixol
Aede de Groot

113

The Synthesis of Fragrant Cyclopentanone Systems

Helen C. Hades

127

Designing Damascone- and Ionone-like Odorants

Philip Kraft

138

Creation of Flavours and the Synthesis of Raw Materials Inspired by Nature

Mark L. Dewis and L. Kendrick

147

Flavours/Foods
New Results on the Formation of Important Maillard Aroma Compounds
Peter Schieberle and Thomas Hofmann
Out of Africa: The Chemistry and Flavour Properties of the Protein
Thaumatin
Steve Pearce and Hayley Roth

163

178

Stability of Thiols in an Aqueous Process Flavour

Chris Winkel,Paul B. van Seeventer, Hugo Weenen and Josef Kerler

194

High Impact Aroma Chemicals

David J. Rowe

202

Subject Index

227

Structure Activity Relationships

STRUCTURE ACTIVITY RELATIONSHIPS AND THE SUBJECTIVITY OF ODOUR

SENSATION

Dr. Thomas Markert

Cognis Deutschland GmbH, Henkelstr 67, D 4055 1 Duesseldorf, Germany

1 INTRODUCTION

Structure activity relationships (SAR) are one of the most useful sets of tools in both
pharmaceutical and fragrance research. Ever since Amoore carried out his studies and
formulated his theory of odour recognition, chemists have been looking at the shape of
molecules and their associative possibilities to find clues that would explain perceived
odour sensations. How difficult it is to go down this research path in finding new chemical
entities with interesting odour qualities is clear from the broad variety of odours the human
nose is able to detect and identify. I will now attempt to explain how complex the activity
side of SAR can be and what the consequences of this complexity are.
In this context I will again follow up the question which Pieter Aarts recently put at
the top of an article [l], although he was dealing with a totally different subject: The
Optimal Fragrance - Lucky Shot or Organised Hunt?
The sense of smell is even able to discriminate between the antipodes of chemical
structures like R- and S-carvone or R- and S-p-menthene-8-thiol [2]. When a perfume
layman, like a chemist, tries to verify the reported odour descriptions, he becomes aware
that the difference between the odours of chemically similar substances is dependent on
purity, concentration in your nose, your sniffing technique, the way the air streams through
your nose [3], and much more.
As Charles Sell tells us in a remarkable report about structure/odour correlations
entitled The Mechanism of Olfaction and the Design of Novel Fragrance Ingredients [4],
it is sometimes a trace impurity which fundamentally changes the scent of a substance or a
mixture of substances.
2 AMOORES CONCEPT OF PRIMARY ODOURS
Let us start with John E. Amoores [5] theory of odour reception (figure l), which is based
on specific anosmia and the concept of primary odours. What I understand about his idea is
that he tried to find chemical structures by using the holes in the olfactory epithelium and a
negative selection of substances that were reported as resulting in specific anosmia.

Advances in Flavours and Fragrances

L<

HIGH

ODOR THRESHOLD CONCENTRATION (log2)

Figure 1 John E. Amoore 's theory of odour reception

In terms of SAR, this would mean he was

searching for chemicals with no activity. And
from the shape of the molecules he found in this
way he tried to reconstruct a receptor site which
could in size and shape accept this chemical
structure (figure 2). The goal of his studies was a
classification of odours by collecting groups of
similar molecules, which could fit, specifically
into the same receptor. Amoore was limited in his
approach to the choice of known substances and
he was also dependent on the odour descriptions
he was given by the experts. My opinion in this
context is that Amoore could never definitely
know whether a substance, which would bind to
the same specific receptor, would cause the same
odour sensations and associations. In other words,
he grouped various chemicals together, guided by
the similar odour descriptions for those materials.
Figure 2
3

SPECIFIC ANOSMIA AND THE CONCEPT OF PRIMARY ODOURS

I have to admit at this point that I have a problem. My problem is with specific anosmia,
which is the basis of Amoore's theory of olfaction. The way Amoore measured specific
anosmia demonstrated the usefulness of his approach and proved the reality of this
phenomenon. However, the results are not useful to classify scents; they only caused
chemists to focus on molecules for which there would probably be a specific receptor in
the nasal mucous membrane. When a chemist looks at the structures found by Amoore
they are surprised to find four small molecules like trimethylamine and isobutyric
aldehyde, alongside two very large molecules like androstenone and pentadecanolide
(figure 3).

Structure Activity Relationships

Figure 3

Structure activity relationships and the subjectivity of odour sensation

Those who are able to smell androstenone with 19 carbon atoms describe it as reminding
them of stale sweat. Isovaleric acid, a molecule with 5 carbon atoms, is almost officially
said to smell sweaty. So, am I to believe that a molecule with 19 carbon atoms is bound to
the same specific receptor as a similarly smelling compound containing 5 carbon atoms?
The Axnoore approach is most interesting because, when you think about it, in the end it
doesnt tell you much about the structural side of SAR, nor does it tell you much about the
activity on the side of the receptor, but it raises the question of what specific anosmia
means. What is the sense of lacking receptors?
When we at Cognis were searching for new sandalwood substances, I noticed that I
became anosmic to Sandelice@;first only on Fridays, then later all the time (figure 4).

Figure 4

Sandelice, lost in a forest of numerous sandalwood trees

Then I noticed that my anosmia was a hyperosmia. I was so sensitive to Sandelice@

that I had the odour impression for a fraction of a second and then my nose had adapted.
So adaptation can also look like anosmia. By contrast, I am truly anosmic to androstenone.
True specific anosmics smell the impurities in the compounds. So, although Im training

Advances in Flavours and Fragrances

on our androstenone sample, to me it smells a little bit cedar-woody but not at all like urine
or stale sweat. Others nearly had their noses blasted off when they opened the bottle. So I
consider that the purity of our androstenone sample is very good.

3.1 What anosmics smell?

When the results of The Smell Survey were published by National Geographic in 1987 [6],
I thought how unhappy the 1.2% of people who were suffering from total anosmia must
feel. I thought those people would neither smell nor taste anything so delicious as truffles
or foie gras. This is by far not the case. I learned from one of my neighbours who lives a
few houses away from ours that her bulbus olfactorius had been severed in a car crash. But
she is still able to taste and to smell. - Though she might need more cigarettes or beer to
have the same activity effect as osmic people. - And I wondered how this could happen
without the ability to smell. Then I read [7] about people who, though lacking a sense of
smell, were able to cook, detect dry or humid air, and more. At least taste is working well
in anosmics.
By thinking about the odour impressions of people lacking olfaction, I found the
explanation for some unusual odour descriptions. What do you think a powdery or dusty
scent should mean? Would it be a powder or dust that would enter your nose? We were
once purifying the essential oil of pinus longifolia and when I smelled the fractions, I
immediately imagined smelling powdered bellpepper from a pepper pot. The visual picture
of a liquid in a distillation bulb did not fit the odour impression of a powder. Then
suddenly I had an idea about what could be the explanation for this curious phenomenon.
Like every mucous membrane, the olfactory epithelium is sensitive to touch as well. In
other words, your nose does not just smell things, it also feels them (figure 5).

Figure 5
Our sense of taste is based on touch. The only reason we boil our soup or coffee is that
we like it hot. Umami (monosodium glutamate) is discussed as a fifth taste quality but it
could also work as a transporter of tastemakers (my personal name for taste enhancers) comparable to the odour binding proteins - by distributing aroma components in your
mouth. The result is called mouthfeel. Touch, pain, or trigeminal reception is what
anosmics smell, and probably osmics as well.

3.2 The Kallmann syndrome

Kallmann's syndrome is a neuronal migration defect, which also affects olfactory system
development. To test the functioning of olfaction with patients suffering from Kallmann's
syndrome, doctors use common fragrance materials. In this way it was found that many
fragrances have a strong trigeminal component. Anosmic patients were able to assign
odour descriptions to fragrances without using olfactorial nerves. So the information must

Structure Activity Relationships

have been transported by a nerve other than the bulbus olfactorius. In his report about
Trigeminal Perception of Odorant Quality in Congenitally Anosmic Subjects [8],
Matthias Laska presents a list of compounds eliciting strong trigeminal responses, which
sounds like Amoores list of primary odours (figure 6).

OI,,
A

acetic acid

\OH
ethanol

L-menthol

s-

1,&cineole

acetone

-OH

propanol

Figure 6
Again there are four smaller molecules like acetic acid, acetone or ethanol and two larger
entities (-)-menthol and 1,8-cineole. The existence of trigeminal sensation, which I would
like to call feeling chemicals, is well known from von Skramliks [9] experiments that were
published in 1926. He found a trigeminal component in odorants by monorhinal
application. When the test person could detect whether the odorant entered the left or right
nostril this was by trigeminal irritation, because olfactorial reception is not able to identify
the direction from which the smell has approached the nasal cavity. To test the extent of
the trigeminal nerve stimulus within a sniffing process, monorhinal application is
recommended.
So what do we learn from those results?

THE ACTIVITY SIDE OF SAR REASSESSED

In a report called Clinical Testing of Olfaction Reassessed A.J.Pinching [ 101 speaks of

the poor smell vocabulary of most humans, which was regarded as a barrier to
interpretation of olfactory tests. However it has become clear that the great majority of
odours have a trigeminal component to their detection. Let us now take a closer look at
odour descriptions. In SAR they represent the activity part of the relationship, and the
accuracy of this part should be as scientific as the knowledge about the chemical structure.
But this is by no means the case. I do not dispute the trigeminal component of odour
sensation. What I think is rather that you have a pain sensation in your odour description
that does not come from an olfactory reception site. That means that many impressions
may stem from stimulating trigeminal nerve endings and you consider them to be your
odour reception, not knowing or even wanting to know that you as an individual suffer
from specific anosmia regarding this particular scent.
As a perfumer you cannot tell everybody that you are not able to smell floral or musk
substances, because you would have all the marketing people crowding round trying to sell
you those substances. That was how I learned that a little cedarwood effect could be my
reception of androstenone. This is enough to live with, but not enough to detect truffles,
which contain markers similar to androstenone.

Advances in Flavours and Fragrances

Methyl dihydrojasmonate (figure 7) is said

[ l l ] to smell less intensive as its purity
increases. When you have perceived this
substance once, you have the impression of
blossoming flowers everywhere in nature,
especially in springtime. The sensation is not
a smell for me but a kind of radiation, which
conjures up the picture of a sunbeam sizzling
your nose into a springtime feeling.
Substances with the same effect, later
evaluated by innocent perfumers, were always
attributed the quality smells like paint. So Figure 7
cis-Methyldihydrojasmonate
which activity would you propose to search (Hedione, FIRM)
for a receptor for methyl dihydrojasmonate,
the paint or the flowery activity?
Being aware of how difficult it is to describe and identify odours, many companies
have invented descriptor systems, which they put in a graph showing the intensity for each
of 160 descriptors of an aroma (figure 8).

Figure 8
This is a method used to characterise scents more accurately. This is also the way the
so-called electronic noses work. With their different sensors they adsorb or oxidise
vapours of organic material and identify the vapour composition by pattern analysis with
neural networks. However, the electronic noses do not give any odour description that
could be used in SAR analyses. They are only able to discriminate between headspaces
that they have stored.

Structure Activity Relationships

4.1 Activity in a chemical sense

This is highlighted in an excellent review about trends in fragrance chemistry [ 121 in the
German magazine Angewandte Chemie (Applied Chemistry) (figure 9).

Figures 9 and 10
What Givaudan researchers make visible and sniffable in this review is a olfactophor
model; i.e. known molecules with known scents were put together in one olfactophor,
similar molecules with other scents were used to define certain exclusion zones around the
olfactophor. This olfactophor model was used to explain scents of new materials that
belonged, olfactorily, to the same family. So nobody except the concept users can know
what came first, the concept or the molecule ... or the chicken or the egg?
Why dont chemists use the knowledge of, for example, the research results of Hanns
Hatt [ 131 (figure 10) who reports that the first cloned human olfactory receptor OR 17-40
exhibits a remarkable ability to discriminate structurally closely related molecules like
helional and piperonal. Interestingly, to humans, both chemicals smell differently as well.
He wonders about the thresholds of some odorants in mammals, especially in humans,
which can be as low as a few parts per million. Such high sensitivity is not observed with
cloned receptors. Multiple factors may explain the higher sensitivity observed in vivo,
including the presence of odorant binding proteins in the nasal mucus. In one figure (figure
11, Diagram B in [13] p.122), the protein encoded by the human OR 17-40 is presented as
traversing the plasma membrane seven times).

Figure 11

10

Advances in Flavours and Fragrances

The finding that the odorant receptors react more sensitively in vivo to odorants than
in vitro is analogous to what Amoore found with his dilution method. Some people were
extremely sensitive to some molecules, which he recognised as primary odours. The
success in isolating and identifying human receptors may mean that special sensitivity to
special odorants has nothing to do with receptors but with those multiple factors that may
explain higher sensitivity in vivo like a vomeronasal organ.

4.2 The quality of fragrances reassessed

Dietrich Kastner tells us [14] that taste and smell apparently may not be qualities of
molecules. This is true in a philosophical sense because it is our mind that makes
perception happen by offering the basic conditions for our ability to perceive in time and
space. Immanuel Kant therefore created the term a priori. At the point when Hatt was
wondering why in his mind structurally related molecules like helional and piperonal
smelled different, Kastner knew that from the nearly 20,000 substances which he smelled
and characterised, he hadnt found any 2 odorants with a totally identical scent. The
conclusion from this observation was, that odour is what we make in our mind out of the
reception of odorants. On the other hand, this would mean that the same molecule would
also smell different to the same nose at different locations and occasions. This is what I
myself am wondering about, since I found in many cases that new synthesised molecules
smelled different in Krefeld, where our perfumers work, compared to Holthausen, where I
work. This I can explain through what I think is trigeminal nerve stimulation dependency.
From small molecules to bigger ones is like switching trigeminal reception to olfactorial
perception.
When Giinther Ohloff once held a lecture in Diisseldorf about his triaxial rule of
odour sensation in the ambergris odorants family he was asked during the discussion
about the odours of hydrogen cyanide or hydrogen sulphide. His remarkable answer was:
What you perceive from those molecules is not an odour reception (figure 12, The nose
as spectroscopist [19]) This is also my theory in explaining the subjectivity of odour
reception: Smaller molecules are felt through irritation of trigeminal nerve endings.
Examples of these touch sensations are the cooling effect of (-) menthol or the burning
sensation of chilli capsaicin or the stinging of acetic acid, the mucous layer membrane
wrinkling of acetone, or the pain sensation of carbonic acid. As the studies of impulses
with congenitally anosmic subjects have shown, they are well able to receive odour
sensation from small molecules and can even make statements about the qualities of the
trigger (figure 13).

A
(-)menthol

xo

acetic acid

Figure 12

Figure 13

acetone

0
capsaicin (chilli)

O K0o H
0
piperin (bellpepper)

carbonic acid

Structure Activity Relationships

11

This is a very common phenomenon, though we normally are not aware of it. For example,
the difference between the taste of alcohol-free and alcohol-containing beer is made by
ethanol. Or the effect of champagne compared to the effect of wine on your hangover is
caused by carbon dioxide via the trigeminal nerve, which means a measurable activity in
the pain centre of the brain. The individual has the impression of becoming drunk sooner
than by consuming the same amount of wine.
This is also the explanation for different odour sensations of the same molecule in
different places. Humidity and temperature are influencing factors for odour reception.
Humidity means that odorants and receptors are covered by different amounts of water and
therefore may have different scents. Perfumers know the importance of humidity from GCsniffing and normal people know that it is very difficult to taste wine properly during
flights. This is because the air conditioning dries the mucous membranes and food and
drink is not tasty any more, except champagne and spirits. These have carbonic acid and
ethanol, which is enough to compensate the climate effect. Thus, the Lufihansa sommelier
recommends drinking lots of water before tasting wine whilst on board.

Mugetanol(H8R)

Y O - C H O

Troenan S521

Lilial,E 034 (GIV)

Lilestralis (BBA)
Lily Aldehyde (SODA)
Lysmeral(BASF)
Rhonadelhyde(RHO)
Floriiys extra (GLID)
Lilion (TOYO)
Novafleur(FIRM)

Muguetaldehyd.E 136
(Citronellyloxyacetaldehyd,
BED. DRAG, IFF, PCAS, SODA)

Figure 14
The reason why aldehydes in general have more intensive scents than the
corresponding alcohols is in my opinion the trigeminal sensation. This clearly does not
influence the low threshold of aldehydes like vanillin, as trigeminal reception is not as
sensitive as odour reception. Many people do not perceive anything when sniffing acetals
like Troenan@[ 15,161 (figure 14) but have strong odour responses from aldehydes of the
same muguet family, like with Bourgeonal@.It is well known that aldehydes might be
oxidised, forming the corresponding acid on their way to the receptors, and it might be that
the acids are responsible for triggering stinging impulses.
4.3 Ways of explaining subjectivity in the perception of odorants

From tests with anosmics we learned that there are trigeminal sensations in
chemoreception which chemists do not normally take into account when searching for new
structures using S A R methods. In my opinion, it is mostly larger molecules where a
stringent SAR approach is useful, for example in the musk, amber or sandalwood family. It
seems as if those molecules with a more complex structure lack trigeminal distribution in
odour perception. It is therefore well known that people who are not able to smell
macrocycles are able to perceive PCMs or nitromusk molecules (figure 15) and vice versa.

12

Advances in Flavours and Fragrances

civetone (mcm)

Figure 15

Galaxolide (pcm, IFF)

musk ketone (nm)

A macrocyclic musk (mcm),polycyclic musk (pcm), and nitro musk (nm)

It seems also as if musks are normally ketones and sandalwoods are normally
alcohols, but there are still some exceptions, just as there are exceptions to the triaxial rule
in the family of ambergris molecules. There will always be a degree of uncertainty in
assigning a chemical structure to a particular perception. Expert perfumers are trained to
find odour descriptors that make it possible for them to identify the same molecule again
and again. The odour descriptors themselves are normally not enough to tell chemists what
kind of molecule was received. This turns out to be worse with mixtures.
Consumers and chemists too, to a certain extent, also associate scents
with individual memories like the smell of granny's bathroom or some
other situation in their youth. The most narcotic smell of my youth is
Opium from Yves Saint Laurent with which I was taught disco dancing
at school. Therefore, odour description in itself is, between the extremes,
both too general and too individual. Climate differences and ethnic odour
preferences are well known.
The most intriguing question in my opinion is that of the sense of specific anosmia. As
an expert perfumer or chemist I cannot work with fragrances which have an odour only for
me but for many other people have little or no scent. Lack of receptors for specific
substances is normal in every individual and Amoore could measure this phenomenon by
his dilution method. He used his results in forming his concept of primary odours to find
out about receptors. When I consider this, I have to say that this would be the right way in
SAR. You have to search for substances to which there is a high proportion of anosmics.
Because this is the only way you can be sure that you have a compound at hand which
binds to a specific odour receptor. This concept, but with a different explanation, led to the
patents for Timberol@and Norlimbanol@(figure16).

Tirnberol (Dragoco)

Norlirnbanol (Firrnenich)

Figure 16
The high amount of anosmics to those substances was claimed to be an indicator of
high substantivity or fixating power [17, 181. So specific anosmia should be the only
argument to test the attractiveness of new compounds. It is the only way to be sure of
having compounds that fit perfectly into one receptor site. Perfumistic evaluation and
selling of those fragrances is especially difficult, for obvious reasons. With this knowledge
in mind some perfumers use mixtures of, for example, PCMs and macrocyclic musks so

Structure Activity Relationships

13

that anosmics in one of those areas have an impression of the other similar-smelling
odorant. Does this also mean that the different musks do not bind to the same receptor?

CONCLUSION

Subjectivity in odour sensation could be described as mixture of trigeminal and odour

sensation, where smaller molecules could influence especially the emotional part of odour
perception, the likes and dislikes. People lacking specific receptors partly take trigeminal
odour responses for odour sensation and might thus have different interpretations of those
scents than other people have. So the field where the chemical senses are essential for life
is the most difficult to speak about generally. The landscape of transcribed and expressed
receptor proteins in humans seems to be as complex as the immune system, and this should
be considered in SAR studies as well. Therefore, my not very optimistic - and maybe not
very scientific - view of SAR is that it is not able to predict new odour molecules and not
even the odour of existing ones. And what is a revolution in medical science, that the first
human odour receptors are already known, will not lead to the construction of new
chemical entities, because one still needs a known fragrance to find an unknown receptor.
It could be more difficult to find new aroma chemicals for known receptors. As Charles
Sell put it [4] there are multiple factors which influence recognition of fragrance
ingredients. In searching for the differences in odour perception and identification, the
most difficult problem will be answering the question why we obviously do have so many
different receptors and also lack so many common receptors. Is our nose of such little
importance in modern times that we have lost the ability to perceive special scents? Or is it
a matter of individuality that we became aware of the incomparability of odour reception?
The multiplicity of facets in chemical reception may have led D.Kastner to the idea that
smell is not a quality of fragrance molecules. His argument is that no 2 molecules out of
20,000 smelt the same [13], but I think the more likely argument for his hypothesis would
be that a highly purified substance may well give different odour impressions on different
occasions. If this were the case, the conclusion would have been that odour is not a
chemical quality.
That molecules of the same kind sometimes give a sensation as if you could feel the
shape of the molecule is unbelievable. But it is in my mind the one and only argument
speaking for the fact that odours are qualities of molecules. The general public, unaware of
the chemistry involved, knows that tastes differ and this is also true for colours, or was it
pain.
Finally we are the crew:

14

Advances in Flavours and Fragrances

The latter would like to thank his interpreters Alice Milne and Dave Brandt for their
enthusiasm and professionalism as ExperTeam.

References
[I]
[2]
[3]
[4]
[5]

P.Aarts, Perfumer & Flavorist July/Aug. 2000, p. 1

A. Mosandl, Kontakte (Darmstadt) 1992 (3), p.38
M.Schrope, Sniffing danger, New Scientist, A ~ g . 2 62000,
~ p. 16
C.Sel1, Perfumer & Flavorist Jan./Febr. 2000, p.67
Amoore, John E. Specific anosmia and the concept of primary odors.
Chem. Senses Flavor (1977), 2(3), 267-8 1
[6] Avery N. Gilbert, Charles J. Wysocki, The Smell Survey , Results; National
Geographic 1987, pp. 5 14-525
[7] I.Ebberfeld, Dragoco-Report 6/1998, pp. 264-70
[8] M.Laska, H.Diste1 and R.Hudson, Chem.Senses 22:447-456, 1997
[9] E.von Skramlik, Handbuch der Physiologie der niederen Sinnne, Vol.l. Die
Physiologie des Geruchs- und Geschmackssinns. Thieme, Leipzig, 1926
[ 101 A.J.Pinching, Brain (1977) 100,377-388
[ 113 DKastner, Parfumerie und Kosmetik, 66 (1) 1985, 5-16
[ 121 P.Kraft, J.A.Bajgrowicz, C.Denis, G.FrBter, Angew. Chem. 2000,112, 3106 - 3138
[ 131 H.Hatt, Zoology 102, ( 1999/2000), 120-126
[ 141 D.Kastner, Parfumerie und Kosmetik, 74 (4)1993, p.208
[ 151 K.J.Rossiter, Chem.Rev. 1996,96, 3201-3240
[ 161 R. Pelzer, U.Harder, A.Krempe1, H.Sommer, H.Surburg and P.Hoever, Synthesis of
New Floral Fragrance Substances Supported by Molecular Modelling, in Recent
Developments in Flavor and Fragrance Chemistry, VCH, Weinheim 1993,29-67.
DE
2807584 (Dragoco Gerberding & Co GmbH, 22.2.1978)
[ 171
[ 181 K.H. Schulte-Elte, W.Giersch, B.Winter, H.Pamingle, G.Ohloff, Helv.Chim.Acta
68( 1989, 1961 - 1985
[ 191 L.Turin, Chemistry & Industry (London) 3.Nov. 1997,866-870

RELATIONSHIP OF ODOUR AND CHEMICAL STRUCTURE IN 1- AND 2-ALKYL

ALCOHOLS AND THIOLS

Y. Sakoda and S. Hayashi

Nagaoka Perfumery Co., Ltd., Research & Development Centre, 1-3-30, Itsukaichi,
Ibaraki, Osaka 567-0005, Japan

1 INTRODUCTION
In recent years, with the explosion of new tastes and combinations of tastes, food has taken
on more than simply the functional role of the maintenance of life. Flavour is one of most
significant factors in taste. Among the constituents of flavours, some compounds influence
the characteristics of flavour greatly. They are called key compounds, and therefore it is
very important for flavour companies to research and develop them. In order to analyse
odour-structure correlation, the methods based on concepts of quantitative structureactivity relationships (QSAR) - I 3 and comparative molecular field analysis (CoMFA) have
been mainly used. The relationship of odour and chemical structure in 1- and 2-alkyl
alcohols and thiols having carbon number from 5 to 11 as synthetic flavour materials was
investigated using sensory evaluation. The respective odour characteristics were analysed
by plotting radar charts. The obtained data was also treated with a principal component
analysis in order to investigate the relationship between the odour and chemical structure.

2 EXPERIMENTAL

2.1

1- and 2-alkyl alcohols and 1-alkyl thiols

The following commercial reagents were used: 1-pentanol, 1-hexanol, 1-heptanol, 1octanol, 1-nonanol, 1-decanol, 1-undecanol, 2-pentanol, 2-hexanol, 2-heptanol, 2-octanol,
2-nonanol, 2-decanol, 2-undecanol, 1-pentanethiol, 1-hexanethiol, 1-heptanethiol, 1octanethiol, 1-nonanethiol, 1-decanethiol and 1-undecanethiol.

2.2 Synthesis of 2-alky thiols

Bromination of 2-hexanol, 2-heptanol, 2-nonanol and 2-undecanol with PBr3 yielded the
respective bromide^.'^^ l5 The bromides were reacted with NaSH to give 2-hexanethiol, 2heptanethiol, 2-nonanethiol and 2-undecanethiol respectively.16 The thiols, 2-pentanethiol,
2-octanethiol and 2-decanethiol were prepared from the corresponding bromides by

16

Advances in Flavours and Fragrances

reaction with NaSH. The 2-alk 1 thiols were purified by distillation and analysed by GC,
GC-MS, FT-IR,'H-NMR and C-NMR. The purities of these thiols were over 97%.

'Y

2.3

N M R and GC-MS

'H-NMR spectra were obtained with a JNM-EX 270 spectrometer (JEOL, Tokyo, Japan) at
270 MHz. 13C-NMR spectra were obtained with a JNM-EX 270 spectrometer (JEOL,
Tokyo, Japan) at 67.5 MHz.
GC-MS analysis was performed on a Hewlett-Packard HP6890 series. The
chromatograph was equipped with a TC-WAX column (60 m x 0.25 mm with 0.25 pm
film) and was programmed from 70C (5 min) to 240C at 3"Clmin; injector temperature,
240C; detector temperature, 240C. The detector ionisation potential was 70eV.
2-Pentanethiol: 'H-NMR: 6 0.91 (t, 3H, J = 7.3), 1.33 (d, 3H, J = 6.6), 1.36-1.59 (m, 4H),
1.48 (d, -SH, J = 6.3), 2.90-3.00 (m, 1H) ppm. I3C-NMR: 6 13.7, 20.6, 25.6, 35.3, 43.0
ppm. MS: mlz (%) = 104 (M', 44), 71 (40), 70 (M' - H2S, 32), 61 (66), 60 (15), 59 (12), 55
(58), 47 (16), 45 (1 l), 43 (loo), 42 (30), 41 (57), 39 (41), 29 (28).
2-Hexanethiol: 'H-NMR: 60.90 (t, 3H, J = 7.1), 1.27-1.60 (m, 6H), 1.33 (d, 3H, J = 6.6),
1.48 (d, -SH, J = 5.9), 2.88-2.98 (m, 1H) ppm. 13C-NMR: 6 14.0, 22.4, 25.6, 29.6, 35.6,
40.6 ppm. MS: mlz (%) = 118 (M', 38), 85 (26), 84 (M' - H2S, 22), 69 (29), 61 (62), 60
(171, 59 (lo), 57 (17), 56 (38), 55 (39), 47 (lo), 43 (loo), 42 (35),41 (66), 39 (31), 29 (24).
2-Heptanethiol: 'H-NMR: 60.89 (t, 3H, J = 6.9), 1.25-1.62 (m, 8H), 1.33 (d, 3H, J = 6.6),
1.48 (d, -SH, J = 5.9), 2.88-2.98 (m, 1H) ppm. 13C-NMR: 6 14.0, 22.5, 25.6, 27.1, 31.5,
35.6,40.8 ppm. MS: mlz (%) = 132 (M', 28), 99 (2), 98 (M+ - H2S, 22), 70 (33), 69 (29),
61 (58),60 (18), 59 (lo), 57 (loo), 56 (74), 55 (42), 43 (42), 42 (20), 41 (77), 39 (28), 29
(31).
2-Octanethiol: 'H-NMR: 6 0.89 (t, 3H, J = 6.8), 1.28 (br s, 6H), 1.33 (d, 3H, J = 6.6),
1.35-1.59 (m, 4H), 1.48 (d, -SH, J = 5.9), 2.88-2.95 (m, 1H) ppm. 13C-NMR: 6 14.0, 22.6,
25.6, 27.4, 29.0, 31.7, 35.6, 40.9 ppm. MS: rnlz (%) = 146 (M', 30), 113 (2), 112 (M' H2S, 2% 84 (25), 83 (40), 82 (lo), 71 (55), 70 (74), 69 (30), 61 (71), 60 (23), 59 ( l l ) , 57
(79), 56 (56), 55 (78), 47 (1 l), 43 (76), 42 (41), 41 (loo), 39 (33), 29 (41).
2-Nonanethiol: 'H-NMR: 6 0.88 (t, 3H, J = 6.8), 1.27 (br s, 8H), 1.33 (d, 3H, J = 6.6),
1.36-1.59 (m, 4H), 1.48 (d, -SH, J = 6.3), 2.88-2.98 (m, 1H) ppm. 13C-NMR: 6 14.1, 22.6,
25.6, 27.4,29.2, 29.3, 31.8, 35.6,40.9 ppm. MS: mlz (%) = 160 (M', 26), 127 (3), 126 (M'
- H2S, 25), 98 (15), 97 (34), 85 (24), 84 (29), 83 (25), 82 (lo), 71 (43), 70 (44), 69 (44), 61
(64), 60 (21),59 (1 l), 57 (52), 56 (69), 55 (78), 47 (lo), 43 (91), 42 (37), 41 (loo), 39 (31),
29 (44).
2-Decanethiol: 'H-NMR: 60.88 (t, 3H, J = 6.6), 1.27 (br s, lOH), 1.33 (d, 3H, J = 6.6),
1.35-1.59 (m, 4H), 1.48 (d, -SH, J = 6.3), 2.88-2.98 (m, 1H) ppm. 13C-NMR:6 14.1, 22.6,
25.6, 27.4, 29.2, 29.3, 29.5, 31.8, 35.6, 40.9 ppm. MS: rnlz (%) = 174 (M+, 23), 141 (3),
140 (M+ - H2S, 26), 112 (12), 111 (20), 98 (15), 97 (32), 85 (26), 84 (23), 83 (26), 82 (lo),
71 (33), 70 (55), 69 (53), 61 (61), 60 (19), 59 (lo), 57 (72), 56 (72), 55 (81), 43 (88), 42
(30), 41 (loo), 39 (27), 29 (43).
2-Undecanethiol: 'H-NMR: 60.88 (t, 3H, J = 6.8), 1.27 (br s, 12H), 1.33 (d, 3H, J = 6.9),
1.43-1.57 (m, 4H), 1.48 (d, -SH, J = 5.9), 2.88-2.98 (m, 1H) ppm. 13C-NMR: 6 14.1, 22.7,
25.6, 27.4, 29.3 (2C), 29.5 (2C), 31.9, 35.6, 35.6, 40.9 ppm. MS: mlz (%) = 188 (M', 18),
155 (3), 154 (M' - H2S, 23), 111 (18), 98 (12), 97 (28), 85 (20), 84 (25), 83 (33), 82 (11),
71 (31), 70 (49), 69 (51), 67 (lo), 61 (52), 60 (16), 57 (70), 56 (53), 55 (76), 43 (SO), 42
(27), 41 (loo), 39 (26), 29 (44).

Structure Activity Relationships

17

3 RESULTS AND DISCUSSION

3.1

Odour characteristics

Odour characteristics of 1- and 2-alkyl alcohols and thiols were described respectively by
using ten sensory descriptive terms: sweet, fruity, tropical, floral, fresh, refreshing, spicy,
roasty, fishy and oily. The odour strength in each descriptive term was scored as 0 none; 1
weak; 2 a little weak; 3 normal; 4 a little strong; 5 strong. The average score from
assessment of 1- and 2-alkyl alcohols and thiols by three trained flavourists was plotted on
a radar chart l7 (Figure 1).
For the 1-alkyl alcohols, 1-hexanol had the strongest fresh and refreshing factors. 1heptanol had a strong fruity factor. 1-octanol had a stronger fruity factor than 1-heptanol,
and also had a strong fresh factor. 1-decanol and 1-undecanol had a stronger oily factor
than the other molecules. Fruity and floral factors decreased from a peak at 1-octanol. The
fresh factor decreased from a peak at 1-hexanol. The oily factor increased according to
increasing carbon number.
For the 2-alkyl alcohols, the sweet factor of 2-decanol and 2-undecanol was markedly
weaker than the other molecules. The fruity factor decreased from a peak at 2-heptanol.
Tropical and floral factors decreased from a peak at 2-octanol. The fresh factor gradually
decreased in relation to increase of the number of carbons. A strong refreshing factor was
observed with 2-pentanol and 2-hexanol, but it decreased in relation to an increase in the
number of carbon atoms. Fishy and oily factors increased according to increasing carbon
number.
For the 1-alkyl thiols, sweet and tropical factors decreased from a peak at 1heptanethiol. The floral factor was scarcely noticed in any compounds, although 1decanethiol had little of it. The refreshing factor was also scarcely noticed in all
compounds. The spicy factor increased according to a decrease in the number of carbons.
For the 2-alkyl thiols, sweet, fruity and tropical factors of 2-pentanethiol and 2hexanethiol were weak, but strong in other compounds such as 2-heptanethiol which was
very strong. The fruity and tropical factors of 2-octanethiol were also very strong. The
floral factor was very weak in all compounds, but increased according to an increasing
number of carbons. The fresh factor was slightly felt in all compounds except 2pentanethiol and 2-hexanethiol. The refreshing factor was extremely weak in almost all
compounds, but was slightly felt in 2-heptanethiol, 2-octanethiol and 2-nonanethiol. The
spicy factor was slightly felt in 2-pentanethiol, 2-hexanethiol and 2-undecanethiol, but was
almost negligible for the other compounds. The fishy factor was very strong in 2pentanethiol and 2-hexanethiol and relatively weak in the other compounds.

18

Advances in Flavours and Fragrances

Figure 1 Odour profiles of 1- and 2-alkyl alcohols and thiols

19

Structure Activity Relationships

3.2

Principal component analysis

The relationships between odour and chemical structure in 1- and 2-alkyl alcohols and
thiols was investigated by analysis of the data using principal component analysis. 18, l 9
The result of this analysis with 1- and 2-alkyl alcohols is shown figure 2. The
contribution of principal component 1 (PC 1) and 2 (PC 2) summed up to nearly 70%
(Table 1). Fruity, tropical, fresh, sweet and refreshing factors contributed to the plus side of
PC 1, while fishy and oily factors contributed to the minus side. Oily, roasty and floral
factors contributed to the plus side of PC 2, while the refreshing factor contributed to the
minus side. Pentanols, hexanols, heptanols and octanols were positioned on the plus side of
PC 1, while nonanols, decanols and undecanols were on the minus side. The change of
odour between 1-alkyl alcohols and 2-alkyl alcohols was more contributed to PC 2 than PC
1. In addition, 1-alkyl alcohols were positioned on the plus side of PC 2, while 2-alkyl
alcohols were on the minus side. The result indicates that the oily and floral factors in 1alkyl alcohols are stronger than those in 2-alkyl alcohols.
PC 2

PC 1

-3
-5

1-alkyl alcohols

2-alkyl alcohols

Figure 2 The score-plots on the plane of PC 1 vs. PC 2 in I - and 2-alkylalcohols

20

Advances in Flavours and Fragrances

~~

Sweet
Fruity
Tropical
Floral
Fresh
Refreshing
Spicy
Roasty
Fishy
Oily

PC1
0.896
0.795
0.872
0.503
0.796
0.859
-0.326
-0.351
-0.922
-0.499

PC2
-0.070
0.41 1
0.284
0.610
0.156
-0.249
0.166
0.758
0.046
0.734

PC3
-0.336
0.192
-0.266
-0.345
0.548
0.361
0.743
-0.196
-0.141
0.299

PC4
0.081
-0.284
0.134
0.43 1
-0.146
-0.162
0.535
-0.21 1
-0.101
-0.163

PC5
-0.040
0.169
-0.038
0.193
-0.025
-0.029
-0.132
-0.462
0.219
0.244

PC6
0.166
-0.139
-0.157
0.102
0.065
0.176
-0.009
0.062
0.124
0.046

Eigenvalue
Proportion
Cumulative proportion

2.269
51.495
5 1.495

1.362
18.558
70.054

1.213
14.707
84.761

0.840
7.063
9 1.824

0.640
4.091
95.9 15

0.372
1.383
97.297

Table 1 Eigenvalues and eigenvectors of the correlation matrix in 1 -and 2-alk3d alcohols
The result of 1- and 2-alkyl thiols is shown figure 3. The contribution of PC 1 and 2
summed up to nearly 83% (Table 2). Floral, refreshing, fruity, fresh and sweet factors
contributed to the plus side of PC 1, while fishy, roasty and spicy factors contributed to the
minus side. Oily, fishy and floral factors contributed to the plus side of PC 2, whilst the
spicy and sweet factors contributed to the minus side. Pentanethiols and hexanethiols were
positioned on the minus side of PC 1. The change of odour between 1-pentanethiol and 2pentanethiol was relatively little. The change of odour between 1-hexanethiol and 2hexanethiol was also relatively little. 2-Heptanethiol, 2-octanethiol and 2-nonanethiol were
positioned on the plus side of PC 1. The reason for this is that 2-alkyl thiols have stronger
sweet, fruity and tropical factors than the 1-alkyl thiols, and 1-alkyl thiols have a stronger
oily factor than the 2-alkyl thiols. The changes of odour from 1-nonanethiol, 1-decanethiol
and 1-undecanethiol to the corresponding 2-alkyl thiols were very large, and they had a
high contribution for both PC 1 and PC 2. The reason for this is that 1-nonanethiol, 1decanethiol and 1-undecanethiol have stronger oily and fishy factors than the
corresponding 2-alkyl thiols, while 2-nonanethiol, 2-decanethiol and 2-undecanethiol have
stronger sweet, floral, fresh and spicy factors than the corresponding 1-alkyl thiols. The
changes of odour from 1-heptanethiol to 2-heptanethiol and from 1-octanethiol to 2octanethiol contributed more to PC 1 than PC 2. The reason for this is that 2-heptanethiol
and 2-octanethiol have much stronger sweet, fruity, tropical, floral, fresh and refreshing
factors than the corresponding 1-alkyl thiols, and have weaker spicy roasty and oily
factors. These results show that 2-alkyl thiols with a carbon number in the range of 7 to 11
have stronger fruity, fresh, tropical and sweet factors than the corresponding 1-alkyl thiols.
It also indicated that 2-alkyl thiols have a brighter odour than 1-alkyl thiols.

Structure Activity Relationships

21

PC 2
5 -

3 -

PC 1

o 1-alkyl thiols
a 2-alkyl thiols
Figure 3 The score-plots on the plane of PC 1 vs. PC 2 in 1 - and 2-alkyl thiols

Sweet
Fruity
Tropical
Floral
Fresh
Refreshing
Spicy
Roasty
Fishy
Oily

PC1
0.933
0.944
0.924
0.748
0.944
0.934
-0.692
-0.849
-0.820
-0.392

PC2
-0.292
0.037
-0.033
0.360
-0.146
0.048
-0.639
-0.239
0.346
0.690

PC3
0.030
-0.216
-0.197
0.390
0.059
-0.186
-0.147
-0.229
0.4 16
-0.587

PC4
0.058
0.145
0.244
0.364
-0.027
-0.225
0.119
0.355
0.060
0.052

PC5
-0.079
0.048
-0.076
0.075
0.23 1
-0.1 11
0.243
-0.186
0.012
0.133

PC6
0.144
-0.065
-0.143
0.008
0.091
-0.130
-0.110
0.065
-0.111
0.058

Eigenvalue
Proportion
Cumulative proportion

2.639
69.664
69.664

1.141
13.018
82.682

0.932
8.682
91.364

0.643
4.140
95.504

0.444
1.970
97.474

0.320
1.027
98.500

Table 2
thiols

Eigenvalues and eigenvectors of the correlation matrix in 1 -and 2-alkyl

22

Advances in Flavours and Fragrances

The result for the analysis of all compounds is shown in figure 4. The contribution of
PC 1 and 2 summed up to nearly 74% (Table 3). Looking at the odour correlation between
1-alkyl alcohols and 1-alkyl thiols, the changes of odour among the compounds with a
carbon number range from 5 to 9 contributed more to PC 1 than PC 2. The changes of
odour for pentyl and hexyl compounds had high contribution for PC 1.This is due to the
fact that 1-pentanol and 1-hexanol have stronger fresh and refreshing factors, whilst the
corresponding thiols have stronger roasty and spicy factors. 1-Heptanol and 1-octanol have
very strong floral and fresh factors, whilst the corresponding thiols have a stronger .roasty
factor. 1-Nonanol has strong floral and fresh factors, whilst 1-nonanethiol has stronger oily
and fishy factors. The changes of odour for decyl and undecyl compounds contributed
more to PC 2 than PC 1. This is due to I -decanol and 1-undecanol having stronger floral
and fresh factors, whilst the corresponding thiols have stronger tropical, roasty and fishy
factors. Looking at the odour correlation between 2-alkyl alcohols and 2-alkyl thiols, the
changes of odour from 2-pentanethiol and 2-hexanethiol to the corresponding alcohols
contributed to PC 1, as already seen in the case of the 1-alkyl compounds. The rational
behind this is that 2-pentanol has stronger fresh and refreshing factors than 2-pentanethiol,
whilst 2-pentanethiol has stronger roasty and fishy factors. Comparing 2-hexanol and 2hexanethiol, 2-hexanol has a stronger refreshing factor, whilst 2-hexanethiol has stronger
spicy and fishy factors. The changes of odour for 2-heptanethiol, 2-octanethiol, 2nonanethiol, 2-decanethiol and 2-undecanethiol contributed to PC 2 because 2-heptanol, 2octanol and 2-nonanol have a very stronger floral factor, whilst the corresponding thiols
have stronger fruity and tropical factors. Also, 2-decanethiol and 2-undecanethiol have
stronger sweet, fruity, tropical and spicy factors than the corresponding alcohols. These
results indicated that alkyl thiols have stronger fruity, sweet and tropical factors than the
corresponding alcohols, with a carbon number above 7.

Sweet
Fruity
Tropical
Floral
Fresh
Refreshing
Spicy
Roasty
Fishy
Oily

PCl
0.657
0.657
0.378
0.744
0.872
0.875
-0.7 17
-0.796
-0.899
-0.424

PC2
0.684
0.613
0.877
-0.396
-0.066
-0.270
0.302
0.495
0.077
-0.198

PC3
-0.175
0.369
0.182
0.158
0.060
-0.098
-0.164
-0.019
0.051
0.865

PC4
-0.I35
0.064
-0.129
-0.177
0.419
0.279
0.529
0.159
-0.195
0.122

PC5
0.062
-0.092
0.053
0.470
-0.026
-0.148
0.257
0.075
-0.047
-0.007

PC6
0.011
0.006
-0.010
0.102
0.192
0.114
-0.076
0.134
0.356
-0.050

Eigenvalue
Proportion
Cumulative proportion

2.285
52.200
52.200

1.492
22.274
74.474

1.007
10.149
84.624

0.826
6.816
9 1.439

0.577
3.329
94.768

0.462
2.131
96.899

Table 3

Eigenvalues and eigenvectors of the correlation matrix in I -and 2-alkyl

alcohols and thiols

Structure Activity Relationships

23

PC 2

4.5

1
PC 1

-3.5

1-alkyl alcohols
2-alkyl alcohols

-5.5

Figure 4

The score-plots on the plane of PC I vs. PC 2 in I - and 2-alkyl alcohols and

thiols

CONCLUSION

The relationship between odour and chemical structure in alkyl alcohols and thiols, which
have a carbon number range from 5 to 11, was investigated. The results indicated that
odour characteristics classified alcohols and thiols to carbon number 5 and 6, or 7 and 8, or
10 and 11. Compounds with 9 carbons, in short, nonyl compounds had both odour
characters, which are found in the groups of heptyl and octyl compounds, and decyl and
undecyl compounds. In alkyl alcohols, 1-alkyl alcohols with a carbon number from 5 to 8
had a stronger oily factor than the corresponding 2-alkyl alcohols, whilst 2-alkyl alcohols
had a stronger sweet factor. 1-Alkyl alcohols with a carbon number from 9 to 11 had
stronger fresh and oily factors, but 2-alkyl alcohols had little odour character. For the alkyl
thiols, 1-alkyl thiols with a carbon number of 5 or 6 had stronger fruity, tropical and oily
factors than the corresponding 2-alkyl thiols. The 2-alkyl thiols had a stronger fishy factor.
This was reversed in thiols with a carbon number of 7 or 8. 1-Heptanethiol and 1octanethiol had stronger fishy and oily factors than the corresponding 2-alkyl thiols. 2Heptanethiol and 2-octanethiol had stronger sweet, fruity, tropical, floral and fresh factors,

Advances in Flavours and Fragrances

24

and had a very bright odour. In thiols with a carbon number from 9 to 11, 1-alkyl thiols had
stronger spicy, roasty and oily factors than the corresponding 2-alkyl thiols. The 2-alkyl
thiols had stronger sweet and fresh factors.
From the result of the principal component analyses, it was shown that the change
tendencies of odour from 1-pentanol, 1-hexanol, 1-decanol and 1-undecanol to the
corresponding 1-alkyl thiols are similar to those from 2-alkyl alcohols to 2-alkyl thiols.
The change tendencies of odour from 1-heptanol and 1-octanol to the corresponding 1alkyl thiols differ from those of 2-heptanol and 2-octanol to the corresponding 2-alkyl
thiols. It is seen that the change in tendencies of odour from 1-alkyl compounds to 2-alkyl
compounds and from alkyl alcohols to alkyl thiols have a regular trend. In addition, it
indicated that 2-heptanethiol and 2-octanethiol have a brighter and unique odour when
compared to other alkyl thiols.

ACKNOWLEDGEMENTS

W e are grateful to Honorary Professor Hiromu Kameoka, Kinki Unversity, and Associate
Professor Araki Masuyama, Department of Materials Chemistry, Faculty of Engineering,
Osaka Unversity, for their advice and support.

References

10
11
12
13
14
15
16
17
18
19

M. H. Abraham, J. Andonian-Haftvan, J. E. Cometto-Muniz and W. S . Cain, Fundamental and

Applied Toxicology, 1996,31,7 1.
F. Yoshii and S . Hirono, Chem. Senses, 1996,21,201.
M. Akamatsu, K. Nishimura, H. Osabe, T. Ueno and T. Fujita, Pestic Biochem. Physiol., 1994,
48, 15.
G. Buchbaur, A. Hillisch, K. Mraz and P. Wolshann, Helv. Chim. Actu., 1994,77, 2286.
A. J. Calder, A. J. Wyatt, A. D. Frenkel and E. J. Casida, J. Computer-Aided Mol. Design,
1993,7,45.
D. R. Cramer 111, D. J. Bunce, E. D. Patterson and E. I. Franck, Quunt. Struct.-Act. Relat.,
1988,7, 18.
H. K. Engel, A. R. Flath, G . R. Buttery, R. T. Mon, W. D. Ramming and R. Teranishi, J. Agric.
Food Chem., 1988,36,549.
R. S . Krystek Jr., T. J. Hunt, D. P. Stein and R. T. Stouch, J. Med. Chem., 1995,38,659.
I. Moriguchi, S. Hirono, I. Nakagome and H. Hirono, Chem. Phurm. Bull., 1994,42,976.
P. J. J. Stewart, Quantum Chemistry Program Exchange Catalogue (QCPE), 1992, IV, 19.
F. Yoshii, S. Hirono and I. Moriguchi, Helv. Chim. Acta., 1993,76,2279.
F. Yoshii, S . Hirono and 1. Moriguchi, Quant. Struct.-Act. Relut., 1994,13, 144.
C. Y. Martin, Quantitative drug design; Marcel Dekker; New York, 1978.
J. D. Bartleson, R. E. Burk and H. P. Lankelma, J . Am. Chem. Soc., 1946,68,25 13.
H. Hauptmann, Tetrahedron Lett., 1974,3593.
L. M. Ellis and E. E. Reid, J. Am. Chem. Soc., 1932,54, 1674.
A. Hatanaka, T. Kajiwara, H. Horino and K. Inokuchi, Z. Naturforsh., 1992,47c, 98.
Y. Sakoda, K. Matsui, A. Hatanaka and T. Kajiwara, 2. Naturjiwsh., 1995, SOc, 757.
Y . Sakoda, K. Matsui, Y. Akakabe, J. Suzuki, A. Hatanaka and T. Kajiwara, 2. Naturforsh.,
1996, ~ O C 84
, 1.

Analytical

NEW DEVELOPMENTS IN SORPTIVE EXTRACTION FOR THE ANALYSIS OF

FLAVOURS AND FRAGRANCES

P . SandralT2,F. David2 and J. Vercammen'

1

Department of Organic Chemistry, University of Gent, Krijgslaan 28 1-S4, B-9000 Gent,

Belgium
2Research Institute for Chromatography, Kennedypark 20, B-8500Kortrijk, Belgium

1 INTRODUCTION
Developments in separation sciences gained momentum in recent years and there are
hardly any analytical challenges left that can not be tackled with state-of-the-art pressureor electrodriven separation methods. For a large number of applications sample preparation
is still the bottleneck for high accuracy, precision and sample throughput, and new
developments in this respect are more than welcome.
In recent years, several new sample preparation methods have been introduced for gas
chromatographic analysis like gum phase extraction (GPE) for gaseous samples, solid
phase extraction (SPE) and solid phase microextraction (SPME) for liquid samples,
ultrasonic (UE) and accelerated solvent extraction (ASE) for solid samples, etc. For an
overview on recent developments in sam le preparation for chromatographic analysis, we
refer the reader to a recent review article.P
Sorptive extraction (SE) in its different forms is slowly but surely taking a unique
position in sample preparation. Sorptive extraction refers to extraction of organic
compounds from a sample matrix being it a gas, a liquid or a solid, into the bulk of a
polymeric retaining phase. The mechanism is based on partitioning or dissolution in a
polymer that is above it's glass transition temperature and behaves as a liquid. This differs
from adsorptive extraction in which compounds are retained on the surface by temporary
bonds to active sites of the adsorbent like the inorganic materials charcoal or silica gel or
the organic materials Tenax, polyurethane foam and divinylbenzenestyrene copolymer. In
sorptive extraction the silicone rubber polydimethylsiloxane (PDMS) is mostly used as
enrichment phase (sorbent). PDMS is also the best known GC stationary phase. Features
of PDMS for enrichment include: PDMS/gas distribution coefficients are known from GC
retention times or can be calculated, PDMS/water distribution coefficients are very close to
octanol/water distribution coefficients that are tabulated as log P or can be calculated by
e.g. software programs like LOGWIN2, high inertness and thermal stability (from -100C
to 350"C), good enrichment performance for polar and reactive analytes, and last but not
least degradation products of PDMS are low in intensity, very specific i.e. cyclic
dimethylsilicones, and can easily be identified and differentiated from target solutes with a
mass spectrometer (MS).

28

Advances in Flavours and Fragrances

Sorptive extraction has already been studied intensively and forms the basis of the
enrichment techniques open tubular trapping (OTT), solid phase microextraction (SPME),
and gum phase extraction (GPE) in the breakthrough and equilibrium mode. In OTT a
capillary column with a thick PDMS layer (up to 100 pm) is used for enrichment of
gaseous and liquid samples.3347s96
SPME uses a special syringe with a fused silica fibre
In GPE PDMS particles 9 , 1 0 ~ 1 1are
coated with a PDMS or polyacrylate (PA)
employed. After enrichment by OTT, SPME or GPE, the solutes are thermally desorbed
on-line with a capillary GC system. The sensitivity that can be obtained with the abovementioned techniques solely depends on the amount of sorbent employed. In the case of
OTT and SPME the amount of PDMS available for sorption is very low (typically 0.5 1.11
for SPME and 1.75 pl/lO cm for OTT), thus causing limited sensitivity. In GPE with
particles (typically 300 p1) this problem is circumvented but intensive drylng is needed for
the enrichment from water samples and volatile analytes are thus lost.
The most recently introduced sorptive extraction techni ues are stir bar sorptive
extraction (SBSE)12and head space sorptive extraction (HSSE).' SBSE and HSEE present
solutions to most of the above mentioned problems and are simple, robust, sensitive and
solvent less sample enrichment methods. In both SBSE and HSSE, a stir bar consisting of
a magnetic core encapsulated in glass, is coated outside with a thick layer of PDMS. In
SBSE, the stir bar is placed in an aqueous sample while in HSSE the stir bar is placed in
the headspace above a solid material. The extraction mechanism is the same as for SPME
but because of the large increase in sorbent phase (typically 25 pl to 125 pl vs 0.5 pl for
SPME) sensitivity is up to 100 times higher. Commonly, after extraction of the analytes of
interest from the sample matrix, the stir bar is thermally desorbed and the analytes are
transferred to a cold (-100C) programmed temperature vaporisation (PTV) injector for
refocusing followed by capillary GUMS or GC/AED analysis. SBSE and HSSE have been
applied for the analysis of micropollutants in water Sam les12, for the determination of the
corkiness flavour in wine14, for pesticide analysis", for the determination of the
preservative benzoic acid in foodstuffs16 and for the determination of PCB's in human
sperm.I7 As an alternative, the stir bars can be desorbed by liquid extraction followed by
conventional or large volume injection in capillary GC. Liquid desorption can also be
applied for solutes that are not amenable to GC analysis. The extracts are then analysed by
liquid chromatographic or electrophoretic techniques.
In this contribution, the principle of SBSE and HSSE is presented and their
applicability in flavour and fragrance research illustrated with some typical analyses
covering the broad application range of SBSE and HSSE like the analysis of flavour
carriers in tea, beer, yoghurt and bananas, with the analysis of off-flavours in beer, and
with the analysis of the bitter compounds in beer.

2 METHOD AND RESULTS

Stir bars coated with 25 to 125 p1 PDMS are commercially available from Gerstel GmbH
(Mulheim a/d Ruhr, Germany). Thin PDMS coatings are applied for the enrichment of
semi-volatiles and thermolable solutes while thick coatings are used for SBSE enrichment
of volatiles and for HSSE. SBSE extraction of a liquid sample is performed by placing a
suitable amount of sample in a headspace vial or another piece of glassware e.g. an
Erlenmeyer flask, depending on the chosen volume and the sensitivity to be reached. The

29

Analytical

stir bar is added and the sample is stirred for 30 to 120 min. Full equilibration is not
necessary for quantitative analysis. For HSSE, the stir bar is hung in a closed vial over the
solid material for a given time. After enrichment, the stir bar is removed and introduced in
an empty glass thermal desorption tube (187 mm L x 4 mm i.d.) and transferred to a
thermal desorption unit. Desorption temperatures depend primarily on the volatility of the
analytes of interest, and are between 150 to 300C at which the stir bar is desorbed for 5 to
15 min under a flow of helium. Cryofocussing in a PTV inlet is required to obtain narrow
inlet bands in the capillary GC analysis. Typical manipulations in SBSE are illustrated in
Figure 1.
A. Add stir bar to vial and stir for a given
time
B. Remove stir bar with forceps and rinse
shortly in pure water
C. Dry with lint-free tissue

D. Introduce stir bar in thermal desorption

tube
E. Thermally extract

F. Thermal desorption unit with rack for

twenty stir bars

G. Headspace sorptive extraction (HSSE)

Figure 1

DifSerent manipulations and devices for SBSE and HSEE

For thermal desorption a TDS-2 or a TDS-A system was used (Gerstel GmbH,
Mulheim a/d Ruhr, Germany). The thermal desorption unit was mounted on a HP 6890
G C N S D (Agilent Technologies, Wilmington, DE, USA) equipped with a CIS 4 PTV inlet
(Gerstel). As an alternative to thermal desorption, liquid desorption may be used as will be
described for the analysis of the bitter acids in beer by micellar electrokinetic
chromatography (MEKC).
The first application of SBSE concerns the analysis of two tea samples with different
flavours namely a herb-mix flavour (Figure 2A) and a raspberry flavour (Figure 2B).

30

Advances in Flavours and Fragrances

The tea's samples were prepared by simply adding boiling water. At 60"C, 10 ml samples
were placed in headspace vials and SBSE extraction was performed on 25 yl PDMS stir
bars during 90 minutes. The stir bars were then thermally desorbed by programming the
thermodesorption unit from its initial temperature, 20"C, at 60"C/min to a final
temperature of 240C at which the stir bar was desorbed for 10 minutes. During thermal
desorption the TDS-2 was operated in the splitless mode so that the entire amount of
desorbed analytes flows towards the cryotrap. The desorbed solutes were cryofocussed at
-150C using liquid nitrogen in an empty glass tube. After desorption, the PTV was
programmed to 280C for re-injection of the trapped compounds. Since the tea samples
were found to be rather concentrated, PTV re-injection was performed in the split mode at
a split ratio of 1:20. The analytical column applied was a HP-5 MS (30 m L x 0.25 mm i.d.
x 0.25 ym df), which was operated in the constant flow mode at a flow of 1 ml/min helium.
The GC oven temperature was programmed from 4OoC, which was kept for 1 min, at a rate
of 5"C/min to 300C. The detector used was a mass spectrometer. The herb-mix tea
contains several terpenoids with anethole (peak 2) as main compound. The raspberry tea is
characterised by high concentrations of amyl propionate (peak 7), cis-3-hexenyl propionate
(peak 8) and a-ionone (peak 11). Surprisingly, iso- and n-butylphthalate (peaks 13 and 14)
are present in relatively high concentrations. The occurrence of the phthalates is not so
strange because they are known as aroma keepers i.e. they are responsible for a slower
release of the aroma solutes.
Fatty matrices (milk, fresh cheese, yoghurt, etc.) have also been analysed with SBSE.
A typical example is shown in Figure 3 representing the profile for yoghurt flavoured with
strawberries. For this application, the yoghurt sample was diluted 1: 1 with distilled water
and extracted during 60 minutes at 1400 rpm with a 55 yl PDMS stir bar. The column was
a Stabilwax (30 m L x 0.25 mm i.d. x 0.25 pm df). The GC oven temperature was
programmed from 40C (1 min) at a rate of 5"C/min to 240C. The compounds responsible
for the strawberry flavour namely ethyl-2-methylbutyrate (peak 3) and y-decalactone (peak
10) are clearly present. It is surprising that the lipid matrix did not disturb the SBSE
enrichment for this quality control application.
HSSE is similar to SPME for gaseous samples but uses a larger amount of sorbent.
The performance of HSSE is illustrated with the analysis of the aroma carriers in bananas
(Figure 4). A slice of 1 g of an unripe and a ripe banana was put in a 250 ml Erlenmeyer
and HSSE sampling was performed for 60 minutes using a 55 pl PDMS coated stir bar.
Thermal desorption of the HSSE stir bar was performed and for cryofocusing of the
analytes prior to injection, a PTV injector with liquid nitrogen cooling was applied. In the
liner of the PTV a small plug of Tenax (ca. 20 mg) was placed. Splitless thermal
desorption was done by ramping the TDS from 40C to 250C at a rate of 60"C/min and
holding the upper temperature for 5 min. During thermal desorption, the PTV was cooled
at -150C and then ramped to 250C at a rate of 600"C/min. The injector was operated in
the split (1/50) mode. Capillary GC analyses were performed on a 30 m L x 250 pm I.D. x
1 pm HP-1 column (Agilent Technologies), with helium as carrier gas. The oven was
programmed from 30C (1 min) to 300C at a rate of 10"C/min. The mass spectrometric
detector was operated in the scan mode ( d z 20-400). The green unripe banana profile (A)
is going over into the typical banana aroma profile (B) in which 2-pentanol acetate and 2methyl- 1 -methylbutyl and 2-methyl-3-methylbutyl propionate play an important role.

Analytical

31

Abundance

l /

Abundance

30 min

11

14

A
Figure 2

'i

10

12

d
30
min

SBSE-TD-CGC/MS
of a herb-mix flavoured (A) and a raspberry flavoured
tea (B). Compounds: 1, menthone; 2, anethole; 3, bisabolol oxide B; 4,
bisabolol oxide A; 5, butyl acetate; 6, amyl acetate; 7, amyl propionate; 8,
cis-3-hexenyl propionate; 9, iso-menthone; 10, damascone; I I, a-ionone;
12, p-ionone; 13, di-isobutylphthalate; 14, dibutylphthalate.

Advances in Flavours and Fragrances

32

4bundance
5
1
8

10

'

"

"

"

"

'

'

'

'

'

'

25 min
Figure 3

SBSE-TD-CGC/MS
of yoghurt flavoured with strawberries.
Compounds: 1, methyl-2-methyl butyrate; 2, ethyl butyrate; 3, ethyl-2methyl butyrate; 4, cis-3-hexenol A; 5, ethyl caproate; 6, cis-3-hexenyl
acetate; 7. isoamyl butyrate; 8, methyl cinnamate; 9, vanilline; 10, ydecalactone.

The elucidation of the "hoppy aroma note" in beer has always been a challenging
task. The main reason without any doubt is the lack of sensitive and selective analytical
methods. In the brewery industry, organoleptic panels very often identify a hoppy note in
beers, but analytical evidence cannot support their observations. SBSE was evaluated for
the enrichment of hop derived solutes in beers. Beer extraction was performed by placing a
10 ml beer sample in a 20 ml headspace vial, adding a stir bar of 24 pl PDMS and stirring
at 700 rpm during 45 min. The following conditions were used: thermal desorption 20C 60"C/min - 240C (10 min); splitless mode ; 50 ml/min desorption flow; PTV : -150C
(cryofocussing temp) - 600"C/min - 280C (2 min) ; split mode (1/20); column 30 m x 0.25
mm i.d. x 0.25 pm Stabilwax, carrier gas 1 ml/min helium; oven temp. 40C (1 min) S"C/min - 240C (30 min). Detection was by 1:l effluent splitting to MSD operated in the
full scan mode and to an PFPD detector.
The chromatograms obtained by parallel MS and PFPD (S-mode) detection for a beer
after stir bar sorptive extraction are very complex. The enriched compounds range from

Analytical

33

4bundance

6
9

1
~

Abundance

Figure 4

HSSE-TDS-CGC/MS of a ( A ) unripe and (B) ripe banana.

Compounds: 1, butyl acetate; 2, 2-pentanol acetate; 3, 1 -hexanol; 4, 1butanol-3-methyl acetate; 5, butyl butyrate; 6, hexyl acetate; 7. 2-methyl-1methylbutyl propionate; 8, 2-methyl-3-methylbutylpropionate; 9, hexyl
butyrate; 10, 2-ethylcyclohexyl butyrate.

Advances in Flavours and Fragrances

34

acetaldehyde to palmitic acid and thus cover a wide volatility range. Different sulphur
compounds
were
detected
such
as
dimethylsulphide,
methylthioacetate,
dimethyldisulphide, methional and 2-(2-furfanyl)thiazole. Hop derived solutes in the
aromagram could be identified in different beers by ion extraction.
Abundance

22 min

Figure 5

SBSE-TD-CGC/MS
ion extraction chromatogram at 204 m/e

32 mir
of a hoppy

beer

Figure 5 shows part of the extracted ion chromatogram at m/e 204 and important
sesquiterpenes and sesquiterpenoids like humulene and P-damascenone could be detected.
Moreover off-flavour compounds, such as the sunstruck flavour that originates from the
bitter acids, could be detected and quantified at the ppt level. This is shown in Figure 6. A
fresh beer sample was analysed directly after opening the bottle and after irradiation with
UV-light. The peak at 9.34 min increases as function of irradiation time and this compound
could be identified by full scan mass spectrometry as 3-methyl-but-2-ene-thiol (insert
Figure 6). It is known that this compound is responsible for the sunstruck flavour of beer
at the ppt level. Stir bar sorptive extraction opens new perspectives for the quality control
of beer (ppm level), for the elucidation of hop derived aroma compounds in beer (ppb and
sub-ppb level) and for ultra-trace analysis of important off-flavour compounds such as 3methyl-but-2-ene-thiol.

Analytical

35

Abundance

- ,

Figure 6

<
45 min

SBSE-TD-CGC/PFPD-MS.
Identification of the sunstruck flavour

The bitter taste of beer is derived from hop (Hurnulus lupus L.) or hop extracts added
to the wort during brewing. In the boiling process the hop a-acids or humulones are
converted into the iso-a-acids or isohumulones. These products are not only responsible
for the bitter taste of beer, but also exhibit bacteriostatic properties and furthermore play an
essential role in the foam stability of beer as well as in the formation of off-flavours like
the sunstruck flavour. Liquid chromatography (LC) is intensively used for the analysis of
a- and P-acids in
and of iso-a-acids in beer.20,21Problems related with interaction
of the solutes with trace metals in the chromatographic system resulting in poor resolution
and quantitation have been reported. 20,21 The electrodriven separation technique micellar
electrokinetic chromatography (MEKC)22,23has been evaluated for the analysis of the hop
acids and the iso-a-acids. For the analysis of the iso-a-acids in beer, a pre-concentration
step is normally performed although direct beer injection in MEKC is possible. Enrichment
is commonly carried out by liquid-liquid extraction or by solid phase extraction on
reversed phase materials. SBSE-LD has been applied for the MEKC analysis of the bitter
compounds in beer. The SBSE-LD-MEKC procedure is as follows. Beer (25 ml) is poured
in a headspace vial of 40 ml and 1 ml of 6 M HCl is added. A stir bar containing 25 p1
PDMS is introduced and the sample is stirred at 1500 rpm at 25C for 60 minutes. The stir
bar is dipped in bidistilled water and dried on a tissue paper. The stir bar is then placed in 2
ml of acetonitrile containing the internal standard ethylparaben at 8 mg/l and is stirred at
1500 rpm at 25C for 60 minutes. 0.5 ml is taken and evaporated to dryness under a
nitrogen stream. The residue is dissolved in 0.5 ml bidistilled water and the sample is
injected. MEKC analysis was performed on a HP3DCE capillary electrophoresis system
equipped with diode array detection (Agilent Technologies, Waldbronn, Germany). Data
analysis was done with the ChemStation (Rev. A.06.01) software (Agilent). Bare fused
silica capillaries (Composite Metal Services Ltd., Worcester, UK) with an internal

36

Advances in Flavours and Fragrances

diameter of 75 pm and a total length of 77.7 cm was used (Leff= 55.7 cm). The buffer
consisted of 50 mM phosphate (pH 7.6) containing 40 mM sodium dodecylsulphate (SDS).
Injection was performed hydrodynamically for 2 s. The MEKC method provides more than
adequate resolution for the separation of the 6 iso-a-acids within a sufficiently short time.
Interference from other beer components is eliminated when using SBSE for sample
preparation. This is evident from figure 7 which shows a comparison between a sample
obtained by direct injection of beer (A) and by SBSE-liquid desorption (B). The
repeatability of the MEKC analysis was found to be adequate as can be seen from the data
in Table 1. The RSD (%) for the areas (measured as normalised areas, i.e. peak area
divided by migration time) of the hop acids were calculated relative to the normalised area
for ethylparaben.

Acid
cis-cohumulone
t rans-co humulone
cis-adhumunlone
cis-humulone
trans- adhumulone
trans-humulone
ethylparaben

Relative areas
RSD( %)*
RSD(%)* Relative
migration normalised
time
area

09141

0,15
0,23
0,26

4,50
5,52
4,21
5,33
4,12
3,46

*n=5

Table 1
3

Repeatability of SBSE-LD-MEKCf o r the bitter compounds in beer.

CONCLUSION

Sorptive extraction is a powerful technique for flavour and fragrance analysis. SBSE and
HSEE are two novel techniques broadening the applicability of sorptive extraction. More
developments like passive sampling, enfleurage on PDMS, etc. are expected in the near
future.

Analytical

37

Abundance

AA
1

Figure 7

The bitter compounds in beer b y direct injection ( A )and SBSE-LD-MEKC

(B).
Compounds: 1, cis-cohumulone; 2, trans-cohumulone; 3, cis-adhumulone,
4, cis-humulone; 5, trans-adhumulone, 6, trans-humulone; 7, ethylparaben
(internal standard).

References
1. P. Sandra and F. David, in Gas Chromatographic Techniques and Applications,
A.J. Handley and E.R. Adlard (Eds), Sheffield Acad. Press, England, 2001.
2. W.M. Meylan, Software KowWIN Version 1.66, SRC-LOGKOW, SRC-ESC,
Syracuse, USA, 2000
3. C. Bicchi, A. DAmato, F. David and P. Sandra, Flavor and Fragr. J., 1987,2,49.
4. J. Roeraade and S. Blomberg, J. High Resolut. Chromatogr., 1989,12, 138.
5. B.V. Burger, M. LeRoux and W.J.G. Burger, J. High Resolut. Chromatogr., 1990,13,
777.
6. C. Bicchi, A. DAmato, F. David and P. Sandra, J. High Resolut. Chromatogr. and
Chromatogr. Comm. 1989,12,317.
7. J. Pawliszyn, Solid Phase Microextraction, Theory and Principles, Wiley-VCH, New
York, 1997.

38

Advances in Flavours and Fragrances

8. J. Pawliszyn, Applications of Solid Phase Microextraction, Royal Society of

Chemistry, Hertfordshire, England, 1999.
9. E. Baltussen, H.-G. Janssen, P. Sandra and C.A. Cramers, J. High Resolut.
Chromatogr., 1997,20, 385
10. E. Baltussen, A. den Boer, P. Sandra, H.-G. Janssen and C.A. Cramers,
Chromatographia 1999,49,520.
1 1 . E . Baltussen, F. David, P. Sandra, H.-G. Janssen and C.A. Cramers, Anal. Chem., 1999
71, 5193.
12. E. Baltussen, P. Sandra, F. David and C.A. Cramers, J. Microcolumn Sep., 1999, 11/10
737.
13. B. Tienpont, F. David, C. Bicchi and P . Sandra, J. Microcolumn Sep., 2000, 12(11),
577
14. A. Hoffmann, W.R. Sponholz, F. David and P. Sandra, Proc. 23rd International
Symposium on Capillary Chromatography, Rive del Garda, Italy, Poster D35,2000.
15. P. Sandra, B. Tienpont, J. Vercammen, A. Tredoux, T. Sandra and F. David,
J. Chromatography A , 2001,928, 1 17.
16. A. Tredoux, H. Lauer, T. Heideman and P. Sandra, J. High Resolut. Chromatogr.,
2000, 23,644.
17. T. Benijts, J. Vercammen, R. Dams, H. Pham Tuan, W. Lambert and P. Sandra,
J. Chromatography B , 2001,755, 137.
18. A.C.J. Hermans-Lokkerbol and R. Verpoorte, J. Chromatogr. A, 1994,669,65.
19. A.J.P. Hofte, R.A.M. van der Hoeven, S.Y. Fung, R. Verpoorte, U.R. Tjaden and J. van
der Greef, J. Am. SOC. Brew. Chem., 1998,56(3), 118.
20. M. Verzele, G. Steenbeke, L.C. Verhagen and J. Strating, J. High Resolut.
Chromatogr., 1990, 13, 826.
21. D. De Keukeleire, J. Vindevogel, R. Szucs and P. Sandra, Trends Anal. Chem., 1992,
11,275.
22. J. Vindevogel and P. Sandra, J. High Resolut. Chromatogr., 1991,14,795.
23. J. Vindevogel, R. Szucs, P. Sandra and L. Verhagen, J. High Resolut. Chromatogr.,
1991, 14,584.

APPLICATION OF CHROMATOGRAPHIC AND SPECTROSCOPIC METHODS

FOR SOLVING QUALITY PROBLEMS IN SEVERAL FLAVOUR AROMA
CHEMICALS

Michael Zvielv", Reuven Giger' , Elias Abushkara', Alexander Kern', Horst Sommer2#,
Heinz-Juergen Bertram2, Gerhard E. Gammer2, Claus Oliver Schmidt2, Wolfgang
Stumpe2 and Peter Werkhoff2

1) Research and Development, Frutarom Ltd., Haifa, Israel

2) Flavour Division, Research and Development Flavours, Haarmann & Reimer GmbH,
Holzminden, Germany
1 INTRODUCTION
The confusion in the interpretation of the molecular structure of some aroma chemicals
causes many misunderstandings with the users of these products. Several important cases,
i.e. keto-enol tautomerisation of a-diketones, isomerisation of a-angelica lactone and the
exact substitution pattern of pyrazines, will be discussed.
The anal ical methods used to unravel the correct molecular structures were GC, GCMS and 'G-NMR spectroscopy especially polarisation-transfer for sensitivity
enhancement (INEPT) and 2D-correlation techniques.

2 a-DIKETONES
a-Diketones are an important group of aroma chemicals that are characterised by two
adjacent keto functionalities. Their structure consists of 4 to 7 carbon atoms as shown in
table 1.
The flavour characteristics of a-diketones range from creamy, sweet, buttery,
caramely and milky, to fruity, roasted and burnt. This group of natural products occur in
many food flavours, butter, milk, fruits, wine, roasted products, vinegar, yoghurt, coffee,
raspberry and cheddar cheese. These materials are applied to dairy flavourings (e.g.
margarine), ice cream flavours, butter, strawberry, caramel, butterscotch, brown sugar,
custard and cream. The a-diketones are also used as key raw-materials for pyrazines, e.g.
diacetyl for 2,3-dimethylpyrazine, 3,4-hexanedione for 2,3-diethylpyrazine etc.

Advances in Flavours and Fragrances

40

r I
4 Carbon atoms

2,3-Butanedione diacetyl

5 Carbon atoms

2,3-Pentanedione acetyl propionyl

6 Carbon atoms

2,3-Hexanedione acetyl butyryl

'ir

3,4-Hexanedionedipropionyl
7 Carbon atoms

2,3-Heptanedione acetyl valeryl

0

5-Methyl-2,3-hexanedioneacetyl
isovaleryl

Table 1 a-Diketones
2.1 On the keto-en01 tautomerism of a-diketones
The GC analysis of certain diketones has caused some confusion amongst several of our
clients. The following explanation should help remove this confusion.
Aldehydes and ketones exist in solution as equilibrium mixtures of two isomeric
forms, the keto form and the enol form. We have noticed that this equilibrium in 3,4Hexanedione (figures 1 & 3) especially, but also in the other a-diketones, is dependent on
the GC conditions, the insert, the insert filling and age, the column, etc.
It has been shown that when injecting the material under different GC conditions, the
equilibrium changes, especially when using an apolar column, e.g. HP-5. When a polar
column, i.e. HP-FFAP was used, only the keto form was detected, thus summing up the
area of both peak areas that are seen when using an apolar column. Even when using an
apolar column with different inserts, the same material would give either 2 peaks or 1 peak
(which is the sum of both areas). The enol form can be readily trapped by reaction with
BSTFA (N,O-Bis(trimethylsilyl)trifluoroacetamide)(figures 2 & 4).

keto form

enol form

Figure 1 The tautorners of 3,$-hexanedione which are seen in GC analysis

+
0

Figure 2 The enol form can react with BSTFA

41

Analytical
Abundance

TIC: 5B5.D

1.&+07.

1.4e+07.

GC-MS of 3.4-Hexanedione
1.2eN7.
1e+07.

Enol

8000000-

GMx)o-

3.91

4000000-

>1

2000000-

C 1 1 M Q :3

. 4 - H a

tf e ( p o o & - n b - t

--lo-

MS of 3.4-Hexcnedione - Keto Form

so0
0

. q . ?.5
. . 20 . .

37,:z
.., . ..
.

! ' ! ' I . .

50

. I ! ' !

. .

7 1

. .

, a. 5.

. . ea
.
,

. ,

.',

.-

osoo
0000

-500

aooo

MS of 3.4-Hexanedione - Enol Form

7500
7000

esoo
-000

s=oo
SO00
4600
4 000

9-00

1000
2600

2000
1 ~ 0 0

-3 0 0 0

SO0
D
50

Figure 3

. . . . . . . . . . . . . . . . . . . . . . . .
35

*o

46

50

6 C

Advances in Flavours and Fragrances

42

Silylated
Enol 1
728
i
150

73

'7
!

Figure 4
As a consequence of these results, one has to take into account the sum of the 2 peaks
(keto and enol), when calculating the purity of a-diketones, and not only the major keto
form peak.

3 a-ANGELICA LACTONE
a-Angelica lactone occurs in grape (dried), bread, soybean and liquorice. The material has
a sweet, creamy, coconut, and vanilla flavour with hay and coumarin-like nuances. It is
applied in aromatic and dairy formulations; coconut, vanilla, molasses, coumarin, cream,
milk, nut and tobacco flavours.

43

Analytical

Angelica lactone (5-Methyl-2,3H-furanone) can be described by the following

structure:
_____-

AQ

Angelica Lactone

Practically, this molecule exists in an equilibrium of 3 isomers (figure 5 ) of which the

major isomer is a.

a-Angelica Lactone

P-Angelica Lactone

y-Angelica Lactone

Figure 5 The 3-isomers of angelica lactone

As can be seen, there is a possibility for the double bond electrons to migrate. This
phenomenon is affected by different parameters, e.g. temperature, pH, etc. When a sample
of a-angelica lactone is injected onto a gas chromatograph (figure 6), it is influenced by the
injector temperature, the form and acidity of the insert, the column conditions, its history
and other parameters.

Figure 6

S u n 1 1 2 0 (8.146 min): 2 . 0

QOOO

8000
7000

43

MS of An-

SOW

I ado@

5000
4000

3000
2000

39

1000
0

25
TI1z-w

Figure 7

30

35

40

45

60 55

60

85

70

75

80

8s

SO

QC

6LIoos '

Advances in Flavours and Fragrances

44

Due to its sensitivity, a-Angelica lactone isomerises during a GC run to give a certain
isomeric partition, as shown in figure 8.
F 101 A. (ANGtlAC; I W N G l t i ( 5 1 .U)

-l----T--7

7.5

2.5

10

12.

Figure 8
Another result of this phenomenon that can be observed (figure 8) is the cohesion of
the peaks, which arises from the isomerisation taking place on column, during the GC run.
Figures 9 and 10 show the same material after being injected after an injection of 5 pL
diethylamine, to change the column pH.
FlDl A, (ANGEIACTWNGl675G.D)

PA 7
180 -

10-

140-

120100-

8060-

4020-

L A

04--r-----T-

2.5

7.5

1
10

Figure 9 GC of a-angelica lactone after diethylamine injection

12

Analytical

45

Figure 10 shows the ratio of isomers before and after diethylamine injection
As in the former example, one has to take into account that different partition patterns
of the three isomers peaks of angelica lactone are obtained in relation to different
parameters in the GC, e.g. temperature, pH, etc.

4 ON THE SUBSTITUTION PATTERN OF PYRAZINES

Pyrazines are materials obtained during the Maillard reaction as by-products of the
browning reaction of sugars and proteins or amino acids. These reactions occur during
roasting, cooking, baking, etc. of different food products. The importance of these
materials motivated organic chemists to synthesise and use them in the flavour industry as
ingredients in flavour formulations for roasted nuts, meat flavours, etc.
By 1970, the first pyrazines obtained GRAS status in the US, for use as flavouring
substances. However, as a result of the lack of modern analytical tools, the exact structures
were not defined. For example, a mixture of isomers of methyl-methoxy- pyrazines was
described in an non-exact manner, namely 2-methoxy-3(5 or 6)-methyl- pyrazine.
The problem of the exact structure of several commercially significant pyrazines
arises when examining the product lists of some important pyrazine manufacturers. The
ambiguous pyrazines are mainly those which contain a substituent at position number 5 or
6, in addition to a substituent at position number 2, for example 2-methyl-5 or 6-methoxypyrazine (figure 11).

Figure 11 2-Methyl-5-methoxypyrazine and 2-methyl-6-methoxypyrazine

In this work, the correct substitution pattern of disubtituted pyrazines was evaluated.
The method which was used to elucidate the structures of these rather simple, but puzzling
molecules was nitrogen or "N-NMR spectroscopy].

Advances in Flavours and Fragrances

46

The substitution pattern of di- or trisubstituted pyrazines can be elucidated by a

combination of NMR methods (figures 12 & 13), especially in mixtures by gradient
selected 'H,15N HMBC experiments (figure 13) at natural abundance level. In the case of
disubstituted pyrazines it allows one to distinguish between 2,3-/2,5- and 2,6-disubstituted
isomers. Another advantage is the determination of and assessment of I5N chemical shifts
even in relatively low concentrations. The reliability of this method was verified by
experimental data resulting from 'H NMR and long-range heteronuclear correlation
experiments such as FLOCK or 'H,13C HMBC experiments. Carbon-carbon coupling
constants derived from INADEQUATE experiments were taken into account as additional
tests. Thirty-one commercially available di- and trisubstituted pyrazines were measured in
order to identify the substitution pattern and to obtain the chemical shift information.
I

N1

N4

* <
.

I'

-!

14

LNycH3
1

o J

CH,CH,CH,O

yycH3

'

CH3CH2CHZO

6
7:
;

0-1

IT

?-.,-,--T,

10

r -

20

,.
30

eL- N4
r .

p - - -

40

cln

r,-

,-,---:-r-i7-'T1

7c

En

r'

l3rm)

-00

-00

-;w

-'lo

i;O

:30

+l;o

--

Figure 13 Heteronuclear correlated 2 0 NMR spectrum (HMBC) of 2-methyl-5-propoxyor 2 -methy 1-6-propoxypyrazine

Analytical

47

Figures 14 to 19 show the range of 15Nchemical shifts that can be expected for various
functional groups. Figure 20 details proton, carbon, and nitrogen shifts for several wellknown pyrazines. The effect of substitution on the NMR of pyrazines is detailed in figures
21 and 22.

Alkylpyrazine

N-1

N-4

N-1

N-4

-46.6

-46.6

-48.1

-45.6

-45.8

-45.8

-44.8

-46.8

-44.5

-47.2

-45.1

-46.0

-48.1

-46.4

Alkylpyrazine

Alkyl groups (only) place the 15Nchemical shifts in the range of -44 to -48 ppm.
Figure 14 I5N NMR data of di- and trisubstituted alkyl pyrazines

Advances in Flavours and Fragrances

48

I=-=Alkox

razine

~~

N-1

N-4

-100.9

-41.7

Alkoxypyrazine

N-1

N-4

-94.4

-37.9

-100.9

-43.6

-100.1

-42.8

-99.8

-45.7

-100.8

-40.3

-94.6

-38.6

-100.3

-44.5

-100.8

-42.6

-100.4

-42.0

-93.5

-38.3

- 100.7

-42.1

-93.6

-38.3

-99.8

-42.4

-93 .O

-4 1.7

Analytical

49

Alkoxy groups place the 15N chemical shifts in the range of -93 to -101 ppm. Alkyl
groups in presence of alkoxy substituents place the 15Nchemical shifts in a range of -38 to
-46 ppm.
~

Thioalkoxypyrazine

N-1
-60.7

N-4
-50.0

Thioalkoxypyrazine

N-1
-59.0

N-4
-54.0

-58.4

Figure 16 "N-NMR data of substituted thioalkoxy pyrazines

Thioalkoxy groups place the "N chemical shifts in the range of -58_to
- -61 ppm.
13
The alkyl groups in presence of thioalkoxy substituents place the N chemical shifts
within the range of -46 to -54 ppm.

N-1
-42.3

Acetylpyrazine

N-4
-40.6

Acetylpyrazine

N-1
-44.1

N-4
-38.6

Figure 17 "-N-NMR data of disubstituted acetyl pyrazines

Acetyl groups place the 15Nchemical shifts in the range of -42 to -44 ppm.
The alkyl groups in presence of acetyl substituents place the 15N chemical shifts within the
range of -38 to -44 ppm.

-52.0

N-4
-37.0

Chloropyrazine

Figure 18 "N-NMR data of di- am ' substituted chloro pyrazines

N-1
-52.9

N-4
-35.1

50

Advances in Flavours and Fragrances

15

Chloro groups place the N chemical shifts in the range of -52 to -53 ppm.The
15
alkyl groups in presence of chloro substituents place the N chemical shifts within the
range of -37 to -53 ppm.

-100
-90
-80
-70
-60
-50
-40
-30
Figure 19 Summary of I'N-NMR chemical shifts for substituted pyrazines

tippm

145.2 145.2 145.2 145.2

O=C-CHs

-45.6 -41.01 9.236 8.753 8.646 2.728 147.7 143.5 147.8 143.6 25.8

.c":>:

N 2 R
1

Figure 20 ''N, ' H and I3C NMR shift data of pyrazine and
p y raz ines

four monosubstituted

To summarize:
Alkyl substituents cause the nitrogen atoms to resonate in the range of -35 to -55 ppm.
Alcoxy groups cause the nitrogen atoms to resonate in the range of -93 to -101 ppm.
Thioalcoxy substituents cause the nitrogen atoms to resonate in the range of -58 to -61
PPm.
Acetyl groups cause the nitrogen atoms to resonate in the range of -42 to -44 ppm.
Chlorine atoms cause the nitrogen atoms to resonate in the range of -52 to -53 ppm.

Analy tica 1

51

Substituent
H
CH3
OCH3

N-1 (ppm)
-43.9
-45.2

SCH3

-97.6
-59.4

COCH3

-45.6

C-3 (ppm)
145.2
144.9
136.0
143.5
143.5

H-3 (ppm)
8.6
8.47
8.24
8.47
9.24

Figure 21 Effect of electronic properties of substituents on pyrazines chemical shifts

The largest low-frequency shift of N-1 was found in 2-methoxypyrazine which is
attributed to the strong .n-donatingnature of the methoxy group.

52

Advances in Flavours and Fragrances

Substituent

H-3
'H-NMR

CH3
OCH3

Upfield

SCH3

Downfield

COCH3

Downfield

Upfield

c-3
13C-NMR
Upfield
Upfield
Down field
Downfield

N-1
l

Upfield
Upfield
Downfield
Down field

Figure 22 The efSect on the chemical shifts of substituted pyrazines

Similar trends are observed in the chemical shifts movements of H-3 (in the 'H-NMR), C-3
(in the 13C-NMR)and N-1 (in the "N-NMR) resonances:
Methyl and methoxy groups shift upfield, thiomethoxy and acetyl groups shift downfield.

4.1 Effect on Chemical Shifts of SubstitutedPyrazines vs. Benzenes

Substituehnt
CH3
OCH3
COCH3

H (ortho)
-0.16
-0.46
+0.62

Effect
Downfield
Downfield
Upfield

H-3
+O. 13
+0.36
-0.64

Figure 23 EfSect on chemical shifts of substituted pyrazines vs. benzenes

Effect
Upfield
Upfield
Downfield

53

Analytical

SUMMARY

While calculating the purity of a-diketones, one has to take into account the sum of the
2 peaks (keto and enol), and not onlv the major keto form peak.
Angelica lactone (5-Methyl-2,3H-furanone) exist as an equilibrium of three isomers of
the double bond, which show different GC patterns under different GC conditions.
15

N NMR data of 31 different pyrazines was measured and evaluated. Chemical shifts
ranges were determined for alkyl, alcoxy, thioalcoxy , acetyl and chloro substituents.
The effect of electronic properties of substituents on chemical shifts of pyrazines in
15
N, 13Cand 'H NMR spectra was studied.
The combined data enabled us to better understand the "N-NMR of pyrazines and to
evaluate and predict their structure more accurately.

References
1.

Studies Towards Structure Determination of Substituted Pyrazines, Zviely, A. Kern, I.

Gozlan and R. Frim, Perfumer & Flavorist, 23, 27, 1998.

Natural Products and Essential Oils

COMMERCIAL ESSENTIAL OILS: TRUTHS AND CONSEQUENCES

Brian M. Lawrence
Journal of Essential Oil Research, R. J. Reynolds Tobacco Company, 950 Reynolds
Boulevard, Winston-Salem, North Carolina 27 105 USA

1 INTRODUCTION
For an essential oil to be a genuine commercial essential oil it must be isolated from a
whole plant or plant part of known taxonomic origin by physical means only.
The physical methods used are distillation (steam, steadwater and water) or
expression (also known as cold pressing, a unique feature for citrus peel oils). There is one
other method of oil isolation known as maceratiorddistillation that is specific to a very
limited number of essential oil plants. In the process, the plant material is macerated in
warm water to release the enzyme-bound essential oil. Examples of oils produced by
maceration are onion, garlic, wintergreen, bitter almond, etc. A sub- or supercritical CO2
extract does not constitute an essential oil, even if it has been fractionated to remove the
non-volatile and waxy components found in the extract. This product is best described as a
volatile concentrate. Obviously, if any extracting solvent is used to isolate the volatiles
such as a halohydrocarbon, the end product is always an extract. There is one further
product that has been classified as an oil (although in this authors mind this is a mixture of
an oil and the wood thermal degradation products) and that is destructive, dry or
empyreumatic distillation.
If we consider a partial inventory of the vegetable kingdom, it is estimated that,
although there are approximately 17,500 aromatic plants in the world, only about 300 are
used as raw material sources in the flavour and fragrance industries world-wide. These
plants are used to produce essential extracts, oleoresins, concretes, absolutes, resinoids,
pomades, tinctures, etc. Of these materials, ca. 50% are produced from cultivated plants
while the other 50% are obtained either as by-products of a primary industry or are
harvested in the wild (wild-crafted). In addition, many of the raw materials are not
distilled or extracted within close proximity of where they were grown or harvested.
Wild populations of aromatic plants can either be homogeneous or heterogeneous.
Individual members of a population can differ both in essential oil yield and their
quantitative composition. However, oil compositional differences in commercial oils is
more complex than described above. There are several reasons why essential oils can have
differing chemical compositions, although the taxonomic origin of the plant from which
the oil was isolated is known, e.g.

Advances in Flavours and Frugrunces

58

1. Genetic: clone, hybrid, cultivar or population specific taxon variation

2. Ontogenetic: variation in the growth stage of the plant harvested for oil isolation.
3. Extrinsic: effect of climate, soil, season, time, disease or insect damage, geographical
origin, etc.
4. Infraspecific: variation in chemotype of plants harvested (wild plants but also some
seed sown plants)
5. Endogenous: within plant variation which can be affected by how much of plant was
harvested
6. Processing: variation in the method of oil isolation
7. Exogenous: addition of components either prior to distillation or directly
into the oil

2 REASONS FOR DIFFERING CHEMICAL COMPOSITION

2.1

Genetic

As expected the genotype of the plant from which the oil is obtained has a marked effect
on the oil composition. Some cultivated aromatic plants are reproduced clonally; however,
different clones of the same botanical species can and are often used for oil production.
An example of this can be seen in the oil composition of three cultivars of lavandin
produced in France, which are reproduced clonally (see Table 1).

Compound

Abrialis

1,8-cineole
(Z)- p-ocimene
(E)-P-ocimene
camphor
linalool
linalyl acetate
lavandulyl acetate
terpinen-4-01
lavandulol
a-terpineol
borneol
Lawrence

6.7-10.4
2.1-2.6
4.0-5.5
8.2-12.2
19.6-39.6
18.6-28.0
1.1-2.7
1.0-1.2
0.5-1.o
0.5- 1 .O
2.4-3.7

Grosso

Super

5.2-10.2

2.0-4.9
1.3-1.4
1.9-2.4
3.8-6.0
29.3-31 .O
30.4-45.0
1.6-2.5
1.4-1.6
0.4-0.9
0.4-0.5
1.6-2.3

0.7- 1.2

0.5-0.6
5.9-8.2
25.7-34.1
26.2-36.7
2.0-2.4
2.8-3.6
0.8-2.4
1 .o-1.2
2.0-3.6

Table 1 Comparative Composition of Commercial Lavandin Oils

An example of marked captionin oil composition due to clonal difference can be seen
with Tagetes minuta (Table 2).
The effect of cultivar on oil composition can be seen with Salvia officinalis (Table 3).
A more complex situation of variation in oil composition is found with geranium (Table 4).
Geranium is a hybrid that has been produced from the cross between Pelargonium
graveolens LHerit ex Arton or P. radens H. E. Moore with P. capitatum (L.) LHerit.
Until someone performs a DNA fingerprint on the plant material used to produce the
various oils their true hybrid origin will remain unknown.

Natural Products and Essential Oils

59

Pre-Flowering
Early Flowering
Clone
Clone
Clone
Clone
1' ~
1 ~
"
~
~
2 (Z)-S-ocimene
16.9-17.8
2.0
21.9-33.3
20.4
dihydrotagetone
46..4-5 1.3
16.5 30.0-36.0
22.9
(E)-tagetone
1.4-3.2
16.9
17.5-24.1
11.4
(Z)-tagetone
18.5-22.4
21.5
1.3-1.4
23.4
(Z)-tagetenone
1.5-4.4
0.1
6.9-8.6
8.3
(E)-tagetenone
2.0-3.0
0.4
1.5-1.9
2.2
1 Lawrence 2 Chalchat et al. (1995)

z2

Full Flowering
Clone
Clone
1
2
35.3
19.0
30.5
3 1.3
2.2
7.3
16.9
29.7
6.9
0.2
1.8
0.2

l
_
l
l
^

Post Flowering
Clone
Clone
1
241.3-45.9
11.8
14.8-24.1
33.4
7.5-8.2
3.7
18.2-20.4
25.8
0.3-0.4
0.6
2.0-3.6
0.4

Table 2 Ontogenetic EfSect on Composition of Oils of Two Clones of Tagetes minuta

Czech'
Hungarian'
--3.4
6.2
a-pinene
camphene
4.3
2.6
1.4
3.0
P-pinene
1,8-cineole
9.5
4.8
22.8
a-thujone
22.4
2.2
7.1
P-thujone
camphor
25.9
25.6
a-humulene
9.6
5.5
1 Chalchat et al. (1998) 2 Asllani (2000)

.
I
"
x
,
.
l
"
"
I
"
l
.
"
l
l
"
l
"
l
.
.

French'
1.9
2.6
1.9
11.0
26.4
5.7
19.9
4.0
l"l.",

""

Yu_goslavian*'
_-.
.-I"-I.2.0-6.6
1.7-8.5
0.7-3.4
5.8-22.5
7.2-36.3
3.9-40.1
7.0-23.1
0.8-2.9
-,"I

"
I

Albanian*'
I."_.
1.1-5.2
4.6-7.2
0.6- 1.5
7.4-19.6
16.6-39.8
4.2-7.8
17.2-30.6
2.1-7.4
~

__

Table 3 Percentage Composition of Commercial Oils of Salvia ogicinalis

cis-rose oxide
trans-rose oxide
menthone
isomenthone
citronellol
geraniol
citronellyl formate
geranyl formate
guaia-6,9-diene
10-eDi-Y-eudesmol

0.7- 1.1
0.3-0.4
0.9-1.3
5.3-5.4
22.9-27.9
17.1-25.0
7.6-9.4
5.9-6.5
t-0.3
4.2-5.4
India
7.0-9.0
0.6- 1.2
0.3-0.5
0.4-0.6
5.3-7.7
24.4-29.0
17.1-21.9
6.0-8.6
3.7-6.3
0.1-0.2
3.8-6.4

"""--

linalool
cis-rose oxide
trans-rose oxide
menthone
isomenthone
citronellol
geraniol
citronellyl formate
geranyl formate
guaia-6,9-diene
1O-epi-y-eudesmol
Lawrence

0.4
0.2
0.2
7.6
31.7
9.8
12.8
3.4
4.6

1.0-1.4
0.5-0.7
1.4-2.4
5.4-5.7
36.5-39.1
8.7-8.9
9.2- 12.5
1.9-2.1
6.5-6.8

M_cl~OCc!?"
5.6-9.9
0.8-1.3
0.3-0.6
0.8-2.1
4.2-5.6
18.6-28.0
18.6-20.6
6.0-7.6
4.1-6.0
0.2-0.5
2.5-5.2
I

Reunion
9.1- 10.8
0.4-0.6
0.3-0.4
0.4- 1.1
8.1-9.5
20.7-23.2
18.1-20.7
7.4-10.4
5.5-5.9
5.8-6.9

l
l
_

0.9- 1.O
0.3-0.4
0.5- 1.3
5.4-7.0
24.8-33.0
14.0-18.0
6.5-7.4
3.1-4.8
0.3-0.5
4.0-5.7
_
_
^
_
I
_

_-___

Table 4 Comparative Percentage Composition of Selected Compounds in Geranium Oil

60

Advances in Flavours and Fragrances

--

Ck~~Om2~~E%-~~~---.Ploidv

2n=18
diploid
2n=36
tetraploid
2n=54
hexaploid
2n=72
octaploid
Oswiecimska (1973, 1974)

Pollen Grain Size

- - - - ~Prochamazu!!ene-Biotll?s<23 1.42 pm
+
21.65-23.46 pm
+
23.89-25.96 pm
_l_l_

--

Table 5 Polyploidy in Achillea millefolium

-*wew-mp-

----*

-*----

~~

Hexaploid Oil
12.4
11.5
1.1
0.1
1.o
5.1
7.5
6.8
6.0
6.5

Diploid Oil
15.3
7.5
1.o

Co?P?tfd.
sabinene
P-pinene
1,s-cineole
linalool
camphor
terpinen-4-01
P-caryophyllene
germacrene D
caryophylleneoxide
(E,E)-farnesol

<o.1

0.1
1.8
22.7
11.4
0.8

Octaploid Oil
1.2
1 .o
14.3
26.2
6.6
1.7
1.9
1.3
1.3

Table 6 Percentage composition of major components of diploid, hexaploid and octaploid

forms of Achillea millefolium
Compound
- -Initiation

a-pinene
1i monene
y-terpinene
octanal
nonanal
decanal
camphor
linalool
(E)-2-decenal
dodecanal
(E)-2-undecenal
geranyl acetate
tridecanal
(E)-2-dodecanal
geraniol
tetradecanal
(E)-2-tridecenal
Lawrence

0.1
t

0.1
1.2
0.5
30.0
0.1
0.3
20.6
3.3
2.6
4.2
3.1
7.6
0.2
0.7
0.5
4.5

-F!o-wering
0.3
t

0.4
0.9
0.1
11.9
0.5
17.5
46.5
1.o
1.4
0.8
2.0
6.0
0.4
0.1
0.1
-. 1.7

_Fruit
1.6
0.2
1.6
0.4
0.1
6.2
2.2
40.9
30.2
0.5
0.8
1.o
1.1
4.8
0.9
0.1
0.1
1.6

Fruit5.1
I .3
6.3
0.4
0.1
1.6
2.4
60.4
3.9
0.4
0.2
1.6
0.5
2.5
1.4
0.2
0. I

Oil3.1
1.3
4.2
0.5
0.3
10.4
1.2
33.6
21.7
1.o
0.9
1.2
1.2
3.6
0.8
0.1
0.1

1.7J
a

Table 7 Ontogenetic Influence on Oil Composition of Whole Plants of Coriandrum

sativum
A further example worth mentioning is yarrow (Achillea millefolium L. ssp. millefolium)
because it exists in various polyploidal forms. Polidy or the existence of polyploids means
that the cells of yarrow can contain more than two sets of chromosomes. As the quality of
yarrow oil is judged by its chamazulene content, ploidy is of great importance. In the

Natural Products and Essential Oils

61

1970s, it was determined that diploids and tetraploids contained achillicin a

prochamazulene precursor, whereas higher polyploids did not. It was also shown that the
existence of the prochamazulene precursor and pollen grain size were correlated (Table 5).
This correlation was confirmed by Hofman et al. (1992) as shown in Table 6.

2.2

Ontogenetic

Many plants, particularly herbaceous plants grown in temperate zones of the world, exhibit
growth effects on their oil composition. Although this phenomenon is generally well
known, the magnitude of variation can be quite dramatic. Also, it must be realised that an
ontogenetic effect is influenced by extrinsic conditions. Again, this is more pronounced in
the temperate and sub-tropical zones. Examples of ontogenetical effects on the oil
composition of whole coriander plants (Table 7) and two clones of Tugetes minutu (Table
2 ) can be seen.

2.3

Extrinsic

The extrinsic conditions such as climate, water, sunlight, day length, pressure, nutrients,
diurnal fluctuation, soil type, disease and insect damage, etc. can have an effect on the
composition of an oil. An example of the effect of extrinsic conditions on the composition
of Midwest (USA) peppermint oil, both within a season (1988, 1993 or 1997) and over the
three seasons is shown in Table 8. Similarly, the effect of seasonal variation on the
composition of clary sage oil produced in North Carolina over 4 years can be seen in Table
9.
The effect of disease on the composition of an oil produced from an Indian cultivar of
geranium can be seen in Table 10. These differences, although not major except for
geraniol, could be the result of necrotic leaves, thereby changing the leafhtem ratio in the
distillation charge.

_CO?Pou"--

limonene
1,8-cineole
menthone
menthofuran
isomenthone
menthyl acetate
menthol
germacrene D
piperitone
Lawrence

__!?S---

1.4-1.7
5.6-6.7
18.1-29.1
0.8-5.3
3.0-4.3
2.0-5.9
33.1-41.1
2.1-3.5
0.5-0.6

Season
L99L
I

1.4-1.7
5.6-6.2
24.5-29.9
1.6-3.0
3.8-4.3
2.9-3.6
33.2-31.5
3.2-2.6
0.5-0.6

3 Season Range
I

1997c
1.O-2.6

3.6-6.1
13.2-34.6
0.2-2.9
2.0-4.6
I .9-8.8
34.1-54.4
0.9-2.1
0.3-0.7

___1 .O-2.6
3.6-6.7
13.2-34.6
0.2-5.3
2.0-4.6
1.9-8.8
33.1-54.4
0.9-3.5
0.3-0.7
-

Table 8 Comparative Percentage Composition of Selected Components in Midwest

Peppermint Oils Over Three DifSerent Seasons

Advances in Flavours and Fragrances

62
Compound
myrcene
limonene
(Z)-p-ocimene
(E)-p-ocimene
linalool
linalyl acetate
a-terpineol
germacrene D
neryl acetate
nerol
geraniol
Lawrence

1979
0.9
0.4
0.3
0.4
27.6
44.4
3.6
2.3
2.9
0.6
2.7

1981
0.1
0.1
0.4
0.2
22.3
53.4
4.5
2.8
3.5
1.o
2.5

1984 -

1982
1.5
0.8
0.5
1 .o
24.1
47.6
4.7
2.6
3.2
1 .o
2.7

1.8
0.6
0.7
1.4

25.3
44.9

5.4
3.6
3.O
1.1

3.3

Table 9 Effect of Extrinsic Conditions on Clary Sage Oil Composition

Healthy Plants
.l_-lll___ll-"
Oil
_""!!!mEEClinalool
10.5
cis-rose oxide
0.5
trans-rose oxide
0.4
0.8
menthone
isomenthone
7.7
citronellol
28.5
geraniol
24.3
citronellyl formate
5.O
geranyl formate
2.1
1O-epi-y-eudesmol
5.7
citronellyl tiglate
1.3
geranyl tiglate
1.2
Rajeswara Rao et al. (2000)

Diseased Plants
Oil

9.5
0.8
0.5

0.9
8.9
31.4
12.4

6.2
1.9
8.2
3.9
3.0

Table 10 Effect of Disease on Geranium Oil Composition

2.4

Infraspecific

The existence of infraspecific (within a species) chemical differences in the essential oils
of some aromatic plants is well known. Such plants are generally referred to as
chemotypes or chemo-varieties. In the context of this manuscript, a chemotype is a
morphologically identical plant that possesses different essential oil from the essential oil
normally encountered.
The reason why the existence of infraspecific differences is of importance to the
essential oil industry is that many plants that are used to produce an oil are collected from
the wild. If all of the plants collected are of the same chemotype a uniform oil composition
can be obtained; however, this is not always the situation.
Compound
a-pinene
myrcene
a-terpi nene
y-terpinene

Percentage
0.6-1.O
1.O-2.8

0.9-2.6
5.0- 10.3

Compound
trans-sabinene hydrate
linalool
terpinen-4-01
thymol

Table 11 Thymus zygis - ISO/AFNOR Specifications

Percentage t-0.5

4.0-7.0
0.7-2.7
36.0-55.0

.o

Natural Products and Essential Oils

63

Spanish thyme oil illustrates this infraspecific point well. Examination of the
ISO/AFWOR Standard (Table 11) reveals that the oil of commerce is rich in thymol.
However, Thymus zygis (Lofl.) L. is collected from the wild, and it exists in three
subspecies forms that are morphologically very similar. In addition, these morphologically
similar subspecies have been found (Velasco Negueruela and Alonso 1984, Salgueiro and
ProenGa da Cunha 1989, ProenGa da Cunha and Salgueiro 1991, Salgueiro et al. 1993, Saez
1995, and Sanchez Gomez et al. 1995) to exist in a number of chemotypic forms (Table
12). It is no wonder that some discernible variations have been found in commercial
Spanish thyme oils.
.I-

2
3
4

5
6
7
8
9
10
11
12
13
14
15

carvacrol
th ymollcarvacrol
linalool
linalool/trans-sabinene hydrate
linalool/a-terpineol
linalool/terpinen-4-ol
linalool/thymol
geraniol/geraniol acetate
camphene/geran y 1 acetate
1,8-~ineole/linalool
1,8-~ineole/thymol
1,8-~ineole/linalool
1,8-cineole/linalooI/thymol
a-terpineolla-terpinyl acetate

Table 12 Chemotypes of Thymus zygis subspecies

The existence of polychemism can also be found with Hyssopus oflicinalis L. ssp.
officinalis. Examination of the oils containing bicyclic components (Table 13) reveals that
none of the oils produced from thirty-five strains of H. officinalis grown in North Carolina
met the ISO/AFNOR standard. This data set represents an excellent example of the effects
of infraspecific variation on hyssop oil.
-COTP?!?!!!!_

P-pinene
1,%cineole
pinocamphone
pinocarvone

- --1(15)*
~~~

7.4- 15.4
<o.1
26.4-62.8
0.1-3.4

2(15)*
8.5-20.3

<o.1
0.2-24.9
0 1-0.4

3(2)

-"

16.6-17.2
<o.1
0.2
19.8-20.0

4(3)
7.7-8.1
18.8-29.3
0.1-18.8
0.1-1.0
"

ISO/AFNOR
"_13.5-23.0
5.5-17 5

* no. of strains
1. pinocamphone > isopinocamphone > P-pinene > pinocarvone
2. isopinocamphone > pinocamphone >P-pinene > pinocarvone
3. isopinocamphone > pinocarvone > P-pinene > pinocamphone
4 (i) isopinocamphone > 1&cineole > pinocarvone and
(ii) 1,8-cineole > isopinocamphone > pinocarvone

Table 13 Comparative Percentage Composition of Major Components of Oils of Thirtyfive Strains of Hyssopus officinalis and ISOIAFNOR Standard for Oil

Advances in Flavours and Fragrances

64

2.5 Endogenous
All essential oil bearing plants possess a variation in the oil composition across the plant.
This within-plant variation is often ignored by essential oil scientists and, as a result,
erroneous conclusions can be drawn from the analysis of an oil isolated from small
quantities of plant material. This is not critically important for commercial essential oils
but there is a ramification that can effect an oil composition. For example, examination of
the data presented in Table 14 shows the variation in oil composition of oils produced from
leaves collected from different stalk positions of Mentha pulegium L. compared with a
typical commercial oil. As can be seen, compositional differences across the plant are very
evident.

Compound
~~

~~

a-pinene
P-pinene
limonene
3-octanol
menthone
isomenthone
neomenthol
neoisomenthol
pulegone
Strengele ( 1994)

Lower
Leaves
1.3
1.o
0.3
1.3
15.3
10.3
35.4
6.2
7.0

Middle
Leaves
0.7
0.4
0.2
2. I
18.3
22.6
8.6
1.2
37.7

Upper
Leaves
0.4
0.3
0.2
2.2
3.6
2.5
0.4
0.2
85.3

Typical
Pennyroyal
Oil
0.4
0.3
0.4
0.7
2.3
0.8
0.4
0.1
86.7

II

Table 14 Comparative Composition of Mentha pulegium oil by Stalk Position

-- Compound

.--_I

menthyl heptenone
(E)-P-ocimene
ci tronellal
citronellol
neral
geraniol
methyl citronellate
geranial
geranyl acetate
P-caryophy Ilene
germacrene D
germacrene D-4-01
caryophyllene oxide
nd = not detected.

Percentage
nd-4.5
nd-5.8
0.7-39.0
nd-6.3
0.4-3 1 .O
nd-22.6
nd-4.9
1.1-41.1
nd-3.3
nd-19.9
nd-26.2
nd-2.9
nd-23.5

Lawrence

Table 15 Percentage Composition of the Major Components of Melissa officinalis oils

The percentage composition of the major components of melissa oil can vary widely
(Table 15). In pursuance with the concept that the oil composition of different leaf
positions might explain the reasons why the neral + geranial content and citronella1
contents vary widely in commercial melissa oil. Hose et al. conducted a study on the
ontogentic effect of oil composition variation. They examined the oil composition of
leaves of M. oflicinalis obtained from different stem positions on a single plant and
compared these to the oil composition of the whole plant. As can be seen in Table 16,

Natural Products and Essential Oils

65

there were dramatic differences-in the oil composition between the leaf oils produced from
the different stem positions. It was concluded that the composition of the leaf oil was
dependent on the position of the leaves on the stem and not, as had been previously
assumed, the age of the plant. To complicate this further, Hose et al. found that there were
also age dependent changes in oil composition that took place within the secretory
structures on the leaf. They found that although the sum of citronellal, neral and geranial
remained approximately constant, neral and geranial were converted to citronellal as the
leaf aged. The varied oil composition of melissa is of distinct importance for the
pharmaceutical and aromatherapy use of the plant (antispasmodic and sedative properties)
because the biological activity of neral and geranial has been found to be greater than
citronellal.

--

Whole Plant
c O _ m J ? ~ u n d - ~ ~Oil
(E)-P-ocimene
4.6
ci tronellal
22.8
neral
10.1
geraniol
3.4
methyl citronellate
2.2
geranial
16.3
P-caryophy llene
10.4
germacrene D
10.5
~

Basal Leaf
Oil
-~
I .3
~

52.4

<o. 1
17.9
0.5
7.3
6.3

_
-

Middle Leaf
Oil

"

2.6
25.6
3.2
<o. 1
1.9
7.1
17.5
21.6

Top Leaf
"

.--- Oil

I_ _ _ x _ - _

2.2
1.1

11.1
1.8
26.1
13.0
20.0

Hose et al. (1 997)

Table 16 Comparative Percentage Composition of Major Components of the Leaf Oil of

Melissa officinalis Produced From Leaves of DifSerent Stem Positions

2.6 Processing
Commercial oils are distilled either by water distillation (hydrodistillation), steam and
water distillation or steam distillation. Hydrodistillation and steam and water distillation
are still used in many developing countries for oil isolation because the stills are less
expensive to construct and the process does not require the high cost of a boiler (e.g.,
satellite steam generator). In developed countries most oils are isolated by steam
distillation. As with all commercial processes the parameters of use can have an effect on
the end product. A comparison between the three distillation processes can be seen in
Table 17. Steam distillation offers several advantages over hydrodistillation or steam and
water distillation.
The effects of different distillation parameters on oil composition can be seen in three
examples. First, a comparison of an oil produced from dried Origanum majoruna under
vacuum, a hydrodistilled oil and a number of commercially available majoram oils reveals
some significant differences in composition as can be seen in Table 18. The effect of
steam versus hydrodistillation on the oil composition of coriander is less dramatic as can
be seen in Table 19.

Advances in Flavours and Fragrances

66

---I-^
xI.Ix_
I-II-_
IParameter
-II-.

"__I.-^--__IxxI_"I^-.-"---~-

Steam & Water

Distillation
Medium
Medium
Medium-high

Hydrodistillation

Oil yield
Speed of distillation
Formation of still
notes
Loss of oxygenated
constituents
Susceptibility to
hydrolysis
Susceptibility to

Lowest
Lowest
Highest
Highest (phenols dissolve
in water)
Highest (pH increase in
water)
Highest (prolonged heat)

Wetness of plant material can affect hydrolysis

----

Medium-low (refluxing a
problem)
Medium-higha (refluxing a
problem)
Medium to high (thermal
conductivity of still wall)
Lawrence

"

Steam
Distillation
Highest
Highest
Lowest
"

"

Lowest
Lowest
Lowest
----=
--

Table 17 Processing: Comparison of Hydrodistillation, Steam & Water Distillation and

Steam Distillation
waw

.-

.
P
p
l
l
l

C!EE?!!d.""

_..~."

sabinene
a-terpi nene
p-cymene
y-terpinene
cis-sabinene hydrate
trans-sabinene hydrate
linalool
terpinen-4-01
a-terpineol
cis-sabinene hydrate acetate
linalyl acetate
terpinen-4-yl acetate
aPlants dried prior to distillation

Commercial
Oil---.l_,__l_-_^-__._.-.--I
2.0-5.3
0.9-7.0
0.9-9.5
3.2-10.2
2.4-6.7
7.1-14.0
2.3-5.6
16.4-36.3
3.8-8.3
"~

1.9-5.4
2.3-5.4

Hydrodistilled
Oil
2.9
3.0
3.5
7.6
0.7
6.9
2.0
48.2
7.8
~"

I-.

2.4
2.0
Lawrence

Table 18 EfSect of Processing on Herb Oil of Origanum majorana

_sE!.?e,-"-."-"a-pinene
camphene
sabinene
P-pinene
myrcene
limonene
y-terpinene
p-cymene
terpinolene
linalool
camphor
geranyl acetate

Hydrodistilled
Oil
1.2
0.2
0.1
0.1
0.5
1.1
2.2
0.3
0.3
81.8
4.8
2.0
--""

_
_
I
_

Steam Distilled
Oil
3.7
0.6
0.1
0.4
0.8
2.0
4.1
0.7
0.5
75.7
4.5
3.2

---.---

Bandoni et al. (1998)

Table 19 Effect of Distillation Process on Oil Composition of Coriander

Vacuum
Microwave
Distilled Oil"
4.6
0.9
8.2
11.1
4.6
8.2
2.9
29.8
4.6
4.6
2.2
0.9

67

Natural Products and Essential Oils

LOPLO?@_

myrcene
limonene
(Z)-p-ocimene
(E)-P-ocimene
linalool
linalyl acetate
or-terpineol
germacrene D
neryl acetate
geranyl acetate
nerol

Fresh Plant Oil

1.7
0.8
0.7
1.4
28.6
44.9
3.1
2.6
1.o
1.9
1.2

-.-

Lawrence

Dried Plant Oil

0.1
0.2
t
0.1
19.1
60.2
2.5
2.0
1.7
3.O
0.5
1.2

Table 20 Comparative Composition of Clary Sage Oil Produced from Fresh and Dried
Plant Material by Steam Distillation
*
-

co!!!Pound___
or-pinene
myrcene
y-terpinene
p-cymene
cis-linalool oxidet
trans-linalool oxidet
camphor
linalool
Bandoni et al. (1998)
t furanoid form

----

TYfica!Oil.
1 .O-6.5
0.4-1.O
2.2-5.1
0 3-3.0
0.1-0.4
0.1-0.3
2.1-4.4
68.9-83.7

- -

_-

- __ .AgeFfOil
2.3
0.1
01
3.8
4.9
4.3
nd
64.0

W g A g e d Oil
0.5

t
4.9
14.1
12.3
6.3

nd = not determined

Table 21 Comparative Percentage Composition of Aged Coriander Oil

The effect of dryness of plant material on the on the oil composition of c l a y sage oil
produced by steam distillation can be seen in Table 20. An example of age as it relates to
oil composition can also be seen for coriander (Table 21).

2.7 Exogenous
A number of years ago, the lavandin producing areas of Provence was visited during
distillation season. On the visit to a distillation site, it was observed that the distiller was
wetting down his lavandin flowers once they were compressed in the still. When
questioned why this process was performed the distiller said that it helped him increase his
yield. Examination of the liquid used to wet down the flowers revealed a mixture of
synthetic linalool and linalyl acetate. As a result, it is not surprising that the yield was
increased. It is quite likely that this was an isolated incident; nevertheless, the addition of a
wetting mixture was also found to have been applied in dillweed oil production in the
United States. The wetting mixture in this case was (+)-limonene and racemic carvone.
Before discussing the addition of materials to the oil or adulteration, the term oil
standardisation requires attention.

68

Advances in Flavours and Fragrances

3 OIL STANDARDISATION
The term standardisation is often used in the essential oil industry to indicate that a
coupage (a blend of components from other sources) has been added to an oil so that the
seasonal or geographical variation of the oil blend in minimised so that oil composition
consistency is maintained from shipment to shipment. In some situations the essential oil
seller compounds a coupage so that the end product oil is specific to a customer.
In their discussion on mint oil standardisation, Moyler and Moss (1988) noted that
commercial gain was not a valid reason for blending mint oils. They stated that the sole
purpose was to achieve consistency in the commercial availability and supply of mint oils
irrespective of crop year and origin. As a result, oil blends should not have a geographical
origin on the label or they should be labelled as with other natural flavours added (WONF).
Oil standardisation for consistency can, therefore, be either:
A blend of oils produced from the same botanical source irrespective of the
geographical origin of the oils
m
An oil with other natural ingredients added
= A blend of oils with other natural ingredients added
The only difference between these three is that the first one is a genuine oil while the other
two are WONF oils.

3.1 Adulteration
The addition of both natural and synthetic components to an oil either to standardise it or
for financial gain when the oil is labelled as a genuine oil of known geographical origin or
just a genuine oil is unfortunately not an uncommon practice in the essential oil industry.
Adulteration can take many forms such as:
The addition of synthetic (foreign) compounds unrelated to the oil composition
The addition of synthetic compounds related to the oil composition
= The addition of oils of or fractions of oils of similar composition to all or part of the
oil.
The addition of natural compounds produced enzymatically or from other oil sources.
The addition of more than one of the above

3.2 Analytical procedures for determining whether an oil is naturavgenuine

There are a number of analytical procedures used to determine whether an oil is
natural/genuine :
Unstable isotope measurement
Stable isotope measurement (I3C/I2C,D/H)
Site-specific natural isotope fractionation (SNIF-NMR)
Foreign component identification
Hydrocarbon or oxygenated component quantitation
Oxygen heterocyclic quantitation (Citrus oils)
Selected ion monitoring of MS of known impurity
Component ratio measurement
Chiral GC analysis
Any combination of the above methods

Natural Products and Essential Oils

69

Although it is beyond the scope of this manuscript to discuss the analytical procedures
in detail, the theory behind some of the procedures does require some explanation.

3.2.1 Unstable Isotope Measurement. According to Culp and Noakes (1990),

differentiation between synthetic compounds produced from fossil fuels and natural
compounds can be achieved by measuring the unstable 14Cisotope content.
In 1962, Friedlander and Kennedy reported that the formation of 14C in the upper
atmosphere occurred through the interaction of cosmic radiation with nitrogen. As C02 is
an atmospheric gas, a certain proportion of "CO2 naturally exists in the atmosphere. As
plants fix C02 through photosynthesis, a portion of 14C02 is fixed. In a report by Bricout
and Koziet (1978), they showed that all recent plant materials have a well-established I4C
content as their source of carbon is atmospheric C02. They noted that prior to 1950 the 14C
activity of CO2 was ca. 13.6dpdgC; however, this activity increased to ca. 26 dpm/gC in
1964 as a direct result of nuclear weapons testing. By the late 1980s, the activity had
decreased slightly to ca. 15 dpm/gC following the Limited Test Ban Treaty. Due to their
extreme age, petroleum (fossil fuel) based synthetic compounds are devoid of 14Cbecause
the isotope has a half-life of 5730 years. Such a difference can, therefore, be used to
unequivocally differentiatekharacterize between a natural compound and a synthetic
compound produced from fossil fuels.
However, according to Culp and Noakes (1990) a natural essential oil component
cannot be distinguished from its synthetic counterpart by this test alone for two major
reasons. One is that many so-called synthetic compounds have natural materials as their
starting product. A second is that the simulation of a natural 14C activity could be
produced by the addition of a 14C-labeled compound. Theoretically, a measured I4C
activity of a known oil component could be compared with the expected activity for a
certain year as a determinant of year of harvest; however, this has not really been used to
any great extent. 14C-activity has been used sparingly to determine whether petroleum
based materials have been added to natural benzaldehyde (Krueger 1987), cinnamaldehyde
(Hoffman and Salb 1989) and ethyl butyrate (Byme et al. 1986). This analytical procedure
requires the isolation of the natural component in question followed by the liquid
scintillation measurement of the unstable radioactive 14Cisotope.
3.2.2 Stable Isotope Measurement - Carbon. As a consequence of photosynthesis,
plants are selectively enriched with 12C and depleted in 13C relative to atmospheric C02,
which contains approximately 13C02 (1.1%) and l2C02 (98.9%), as a reflection of their
environment and metabolism (O'Leary 1981). This means that the depletion of 13C is
different depending upon the photosynthetic pathway utilised by the plant.
There are three possible photosynthetic pathways known in plants. The first, which is
known as the Calvin pathway (C-3 pathway) is used by most plants (Calvin and Bassham
1962). For C-3 plants atmospheric C02 is fixed by carboxylation of ribulose 1,5disphosphate to yield 3-phosphoglycerate. A second pathway, which is known as the
Hatch Slack pathway (Hatch and Slack 1966, C-4 pathway) is used only by a few aromatic
plants such as Cymbupogon species. For C-4 plants atmospheric C02 is fixed by
carboxylation of phosphoenolpyruvate, which occurs with a lower carbon isotope
fractionation than the CO2 pathway. The third pathway, (Bender 1971), which is known as
the Crassulacean Acid Metabolism pathway (CAM-pathway), is used by Vanilla planifolia
and other succulent plants. For CAM plants atmospheric CO2 is fixed at night by
carboxylation of phosphoenolpyruvate with malic acid accumulation. During the day

70

Advances in Fluvours and Fragrances

malic acid is decarboxylated and the liberated CO2 is fixed by ribulose 1,5-diphosphate.
The two steps in the C02 fixation are influenced by environmental factors.
Stable carbon isotopic measurement is performed with specially designed and
modified mass spectrometric equipment so that the isotopic ratio of the C02 obtained from
an isolated essential oil component can be compared with the C02 obtained from a
standard or a standard C02. The result is a unique ratio of 13C02:12C02.For essential oil
or parts
constituents this ratio is more commonly encountered as a 6I3C value per mil ('loo)
per thousand deviation of the sample isotopic ratio, which is expressed as follows:

613C (Oleo) =

13c0.,
/ I2co2(sample)
"CO2 / l2CO2 (standard)

-1

x 1,000

(Bender 1971)

Examples of the use of this technique to determine genuineness can be seen in Table 22.
"

Component--"--.-""-..----,"""P-pinene, limonene, y-terpinene, nerol, geraniol,

neral, geranial, neryl acetate, geranyl acetate
octanal, nonanal, decanal, dodecanal
myrcene, limonene, y-terpinene, p-cymene,
linalool, geraniol
neral, geranial, citronellal, nerol, geraniol,
geranyl acetate, P-caryphyllene
linalool, linalyl acetate, a-terpineol
a-thujene, a-pinene, sabinene, myrcene, octanal,
limonene, y-terpinene, terpinolene, linalool, a sinensal, methyl N-methyl anthranilate
a-pinene, P-pinene, myrcene, limonene, octanal,

Lemon

__----.
Reference
Braunsdorf et al. ( 1 993)

Orange
Coriander

Braunsdorf et al. (1 993)

Frank et al. (1995)

Melissa, citronella,
lemongrass, lemon
Neroli, petitgrain,
bergamot
Mandarin

Hener et al. (1995)

Mosandl & Juchelka (1997)

Bitter orange

Juchelka ( 1998)

-----

Oil
~ _ I

. I

Faulhaber et al. (1997)

Table 22 Use of Stable Isotope Ratio Measurements ('3C/'2C) to Examine Oil Genuineness
3.2.3 Stable Isotope Measurement - Hydrogen. It is known that as water evaporates
from the oceans and lakes it is progressively depleted of deuterium (D, the heavier H
isotope) as the evaporation/condensation cycle proceeds. According to Craig ( 196 l), this
phenomenon is most evident in the mountains and polar regions where water vapour is
condensed at higher altitudes and latitudes with the associated temperature decrease
thereby depleting deuterium in both cloud and rain water. Isotopic fractionation of D
occurs in plants as water evaporates through leaf stomata through the normal
evapotranspiration cycle. However, according to Yapp and Epstein (1980) the D/H ratio is
only affected to only a minor extent by this phenomenon in comparison with the relative
D/H ratio in the local water and the local environment in which the plants are grown.
Although there is not a clear distinction between the isotope ranges of various hydrogen
pools, 6D variations occur as a direct result of photosynthetic pathways, biochemical
subgroups and botanical species (Culp and Noakes 1990).
6D o/oo values can be used to differentiate between natural and synthetic sources of
major essential oil constituents. In the case of 6D o/oo ratios they can be measured against
one or two standards such as Standard Mean Oceanic Water (SMOW), which has a 6D o/oo
value set at zero and/or Standard Light Antarctic Precipitation (SLAP), which has a
6D o/oo value of -428 (Gonfiantini 1978).
6D values are obtained using a similar calculation as used for 6I3C values e.g.,

Natural Products and Essential Oils

71

sample - D/H standard)

D/H (standard)

x 1,000

In 1978, Bricout and Koziet determined the 6D values for (E)-anethole from a variety
of sources showing that synthetic anethole (Table 23) was easily differentiated from
natural sources. They also showed how synthetic L-menthol, linalool and neral/geranial
could be differentiated from their natural counterparts (Table 24).
*
-"
--

Anethole Source
"

sweet fennel bil bitter fennel oil

star anise oil
synthetic
Bricout & Koziet (1978)

6D Value
SM9-W
-86
-9 1
-84 to -96
-45

Table 23 SD O/oO Valuesfor fE)-anethole

Constituent
~------

111111^_-

L-menthol

ex Peppermint
ex Cornmint
Synthetic
linalool
ex Coriander
ex Thyme
ex Petitgrain
Synthetic
neral/geranial ex Lemongrass
ex Bitter orange
ex Litsea cubeba
Synthetic
Bricout & Koziet (1978)

6D SMOW -Value
.I-Ix
-394
-358
- 196 to -242
-269
-257
-244
-170
-276
-258
-25 1
-174
-

1
1
1

Table 24 60 Valuesfor Specific Oil Constituents

3.2.4 Site-specific Natural Isotope Fractionation. As a way to exploit the joint
structural and quantitative dimensions of proton NMR spectroscopy, Martin and Martin
(1981, 1985) determined that deuterium was far less randomly distributed in organic
molecules than was once thought. They found that large variations characterised in the
deuterium contents of different molecular sites were related to specific enzymatic or
biochemical (kinetic) or thermodynamic isotope effects. To examine this phenomenon
Martin et al. (1982) pioneered the technique used to determine site-specific natural isotope
fractionation using NMR and gave it the acronym SNIF-NMR. With this technique direct
measurement of isotopic ratios of protons at several positions in a single molecule is
performed, thereby enhancing the performance of an isotopic method to draw an
irrefutable conclusion as to the naturalness of a specific component.

72

Advances in Flavours and Fragrances

I s

OCH3
Figure 1

(E)-anetholecontains six non-equivalent sites associated with six

monodeuterated isotopomers. Ratios of the protons at each site with those
at other sites throughout the molecule leads to site specific natural isotope
fractionation (SNIF NMR) which results in the ability to differentiate
between the natural and synthetic anethole (Martin et al. 1982).

Since the values of the specific D/H isotopic ratios measured by D-NMR are sensitive
to the method of purification and to instrumental conditions, internal ratios are also taken
into account. For example with (E)-anethole (see Figure I), the isotopic distributions of
the aromatic, ethylenic and methyl sites can be characterised by the ratios 2/1, 4/3 and 6/5,
respectively (Martin et al. 1982).

3.2.5 Foreign Components. A comparative capillary GC profile on both a polar and

non-polar column of a sample oil and a genuine oil will readily reveal the existence of a
stranger peak that can be quickly identified by a combination of GUMS and retention
indices.
If the unrelated constituent is a non-volatile additive, add a known weight of a
selected internal standard and re-run the GC profile and determine whether the internal
peak is found in the amount close to expected. If not, then the oil was adulterated with a
non-volatile components. It should be noted that the internal standard should be carefully
selected so that its retention index will not coincide with any of the oil components.
Examples of some unrelated synthetic components that have been identified in some
commercial oils can be seen in Table 25.
3.2.6 Hydrocarbon or Oxygenated Component Quantitation. Because oils can be
quite complex, it is useful to separate the oxygenated components from the hydrocarbons
by column or flash chromatography to examine their genuineness. It is easier to examine
the quantitative data for each group and determine their reproducibility. Also, if chiral GC
is planned this procedure simplifies the mixture somewhat. An example of the
hydrocarbons and oxygenated constituents of lemon oil can be seen in Tables 26-27.

Natural Products and Essential Oils

73

- Oil
- - - - - - - - ~ ~Adulterant
~ - ~ - " - - _ - - - - - ~ - _ - . - Reference
~

cassia
rose
geranium
cinnamon
citronella
juniperberry
nutmeg
palmarosa
patchouli

dibenzyl ether
dibutyl phthalate
dibenzyl phthalate
benzyl benzoate
diethylene glycol monoethyl ether
hercolyn D
ethyl palmitate
triacetin
propylene glycol
hexylene glycol
hercolyn D, etc.

Lawrence
Lawrence
Lawrence
Lawrence
Lawrence
Lawrence
Lawrence
Lawrence (1976)
Bruns (1978)

Table 25 Examples of Synthetic foreign) Compounds Unrelated to the Oil Compositions

that have been found as Oil Adulterants
Compound
a-thujene
a-pinene
camphene
sabinene
P-pinene
myrcene
a-phellandrene
a-terpinene
p-cymene
limonene
(Z)-p-ocimene
(E)-P-ocimene
y-terpinene
teminolene
Chamblee et al. (1991)

Sicilian Oil
0.4 1
1.74
0.05
1.70
10.25
1.60
0.04
0.22
0.13
65.16
0.05
0.1 1
9.28
0.4 1

--.,..._....-..-_I Californian Oil _ _ _

I

0.39
1.75
0.06
1.83
11.14
1.58
0.05
0.17
0.25
65.93
0.05
0.09
8.31
0.35

Table 26 Selected Monoterpene Hydrocarbons in Lemon Oil

_--CO!!!!P!L

- Sicilian

I__

octanal
1,8-cineole
nonanal
linalool
citronella1
decanal
neral
geranial
perillaldehyde
undecanal
neryl acetate

0.10
0.05
0.13
0.18
0.11
0.05
1.25
2.02
0.02
0.02
0.5 1

Oi!

_
I
_

Californian Oil

I
-

0.08
0.04
0.13
0.14
0.08
0.05
0.73
1.14
0.03
0.03
0.57

Chamblee et al. (1991)

Table 27 Selected Oxygenated Constituents in Lemon Oil

Another way to determine the authenticity of an oil is to examine its unique
sesquiterpene hydrocarbon content as can be seen in Table 28 for the citrus oils.

74

Advances in Flavours und Fragrances

Sesquiterpene HC
-a-copaene
P-cubebene
p-caryophyllene
trans-a-bergamotene
(Z)-P-farnesene
germacrene D
valencene
a-selinene
(E,E)-a-farnesene
P-bisabolene
6-cadinene
G. Dug0 et al. (1999)

Bergamot
Bitter
----.J!!axE-!*!

Grape-

24 .O
2 1.9

27.8

13.4
13.3
38.2

4.1

3.7
51.8

1.4
9.8

4.9

Lemon

Lime

Mandarin

13.8
24.5
1.6

11.8
16.5
1.7
4.0

0.9
32.7

37.1
v

3.7
2.5
32.2

8.5
7.0
6.8

2.8

9.0
6.3

26.5
10.6
40.9

5.4

3.9

10.9

38.5

13.0
-

__

_
I
-__--I

1 .o

Orange
-

.-x

-&
-*.--*--

Table 28 Comparative Sesquiterpene Hydrocarbon Content of Citrus Oils

Recently, Feger et al. (2000) also examined the sesquiterpenes in citrus oils; however,
they determined the germacrene content of the oils (Table 29), which can also be used to
determine oil genuineness.
Citrus Oil
bergamot
grapefruit
lemon
lime distilled
lime key ( N B )
lime Persian
mandarin
orange bitter
orange sweet

A
0.01
0.02-0.03

B
0.01

C
0.02

D
0.05-0.07
0.07-0.1 1

_I1

t
-

0.04-0.08
0.78-0.90
0.16-0.19

0.36-0.46
0.12-0.1 8
t-0.02
t-0.01
t-0.02

Feger et al. (2000)

0.01-0.02

0.50-0.59
0.09-0.12
<0.01-0.01
0.01-0.02

0.08-0.10
0.05-0.07
*co-elutes with valencene

0.30-0.4 1
0.07-0.16
0.02-0.04*
0.07-0.1 1*
0.07-0.33*
0.06-0.0

--__BiYC!L.
0.0 1
0.02-0.04
0.04-0.12

t-0.0 1
t-0.0 1
t-0.01
0.01-0.02

Table 29 Percentage of Germacrenes in Citrus Oils

3.2.7 Oxygen Heterocyclic Quantitation in Citrus Oils.Cold pressed citrus oils are
known to contain a few percent of non-volatile oxygen heterocyclic compounds such as
coumarins, psoralens and flavones. The major types of components found can be seen in
Figures 2-4.

Herniarin:
Osthol:
Citropten:
Aurapten:
Meranzin:
Isomeranzin:
Meranzin hydrate:
Epoxyaurapten :

Figure 2

Coumarins in Citrus Oils

7-methoxycoumarin

7-methoxy-8-isopentylcoumarin
5,7-dimethoxycoumarin
7-geran ylox ycoumarin

7-methoxy-8(2,31-epoxy)isopentenylcoumarin
7-methoxy-8(21-oxo)isopentenylcournarin
7-methoxy-8(21,31-dihydroxy)isopentenylcoumarin
7-(61,71-epoxy)geranyloxycoumarin

75

Natural Products and Essential Oils

Bergapten:
Bergamottin:
Imperatorin:
Phellopterin:
Isopimpinellin:
Onidillin:
Expox ybergamottin:
Epoxybergamottin hydrate:
Ox ypeucedanin:
Oxypeucedanin hydrate:
B yakangelicin:
B yakangelicol:

Figure 3

5-(6',7'-epoxy)geranyloxypsoralen
5-(6',7'-di hydrox y)-geran yloxypsoralen
~-(2',3'-epoxy)-isopentyloxypsora1en
5-(2',3'-dihydroxy)-isopentyloxypsoralen
8 4 2',3'-dihydroxy)-isopentyloxy-5-methoxypsoralen
8-(2'3'-epoxy)-isopentyloxy-5-methoxypsoralen

Psoralens in Citrus Oils

tangeretin:
sinensetin:
nobiletin:
tetra-0-methylscutellarein:

Figure 4

5-methoxypsoralen
5-gerany loxypsoralen
8-isopentenyloxypsoralen
5-methoxy-8-isopenteny loxypsoralen
5,8-dimethoxypsoralen
5-isopentenyloxy-8-methoxypsoralen

5,6,7,8,4'-pentamethoxyflavone
5,6,7,3',41-pentamethoxyflavone
5,6,7,8,31,41-hexamethoxyflavone
5,6,7,41-tetramethoxyflavone

Methoxyjlavones in Citrus Oil

Examples of these can be seen in Tables 30-31. In Table 32, it can be deduced that
the commercial bitter orange oils have been adulterated with sweet orange oil as all of the
coumarins and psoralens are reduced except meranzin hydrate and the levels of
heptamethoxyflavone and tetra-0-methylscutellarein are increased.
..!!29oEn!~---"
----tangeretin
3,3',4',5,6,7,8-heptamethoxyflavone
nobiletin
tetra-0-methylscutellarein
3' ,3' ,4',5,6,7-hexamethoxyflavone
sinensetin
Mondello et al. (1993)

s_l

2E%e-~AndErkoi!38-73
148-339
55- 123
16-69
34-80
36-151
2 1-42
3-1 1
6-23
5-16
1-4
mg/100g oil

_l___

Table 30 Methoxyflavone Content of Orange and Mandarin Oils

Finally in Table 33, it can be deduced that the bergamot oil has been adulterated with
lime oil because of the increase in 5-geranyoxy-7-methoxy-coumarinand the addition of
herniarin.

76

Advances in Flavours and Fragrances

"m-.&-K!LE3154

- Persian
~ ~ . -

com~~d"_____._.~_.._____-_.___-__

--_^"

bergamottin
5-geranyloxy-7-methocoumarin
onidillin
citropten
herniarin
bergapten
isopimpinellin

3 157-3283*
3065-4045
25-35
49 1-632
86-96
100-124
350-365

2221-391 8
1943-3780
5-8
326-569
339-594
158-250
169-293
210-328

4093
24

484
74
89
33 1
IAA
'M

P.Dugo et al. (1998)

*mg/100g oil

Table 31 Main Oxygen Heterocyclic Components in Lime Oil

- x

Commercial
"Oils
105-197
27-45
0- 127
0-40 1
69- 126
34-78
4-30
16-46
0-24
24-43
31-98

Spanish
_--"-"
?tali?!!
366-37 1*
154-184
71
52-73
304-328
188-322
307-332
788- 1 172
208-2 13
154-2 1 1
95-99
59-156
23-25
5-14
76-85
34-87
10-18
5-8
24-42
13-45
18-39
10-70
P.Dugo et al. (1996)

_cP_~POu"t!
- --- - .-"l"
osthol
bergapten
epoxybergamottin
meranzin
isomeranzen
tangeret i n
heptamethoxy flavone
nobiletin
tetra-O-methy lscutellarein
epoxybergamottin hydrate
meranzin hydrate
*mg/100g oil
"

11"

"

_ i

"

"

P
I
-

Table 32 Main Oxygen Heterocyclic Components in Bitter Orange Oil

.we

ComEound
bergamottin
5-geranyloxy-7methoxycoumarin
citropten
herniarin
*mg/100g oil

C.P. oil
1000-2750*
80-270

Recovered oil
930-1660
110-180

Bergapten-free oil
1 170-1620
150-200

120-350

90-220

0-50

I
_
-

C o m - m ~ c ~ a 1__- o ~
680- 1 160
180-370
100- 1 30

Mondello et al. (1993)

Table 33 Main Oxygen Heterocyclic Components in Bergamot Oil

3.2.8 Selected Ion Monitoring of MS of Known Impurity. Agnel and Teisseire (1984)
determined adulteration in lavender oil with synthetic linalool and linalyl acetate by
determining the existence of dihydrolinalool ( 1.5-2.0%) and dehydrolinalool (0.05-0.10%)
in linalool and dihydrolinalyl acetate (1 5 2 . 0 % ) and dehydrolinalyl acetate (0.05-0.10%);
however, this was difficult if the amounts of synthetic linalool and linalyl acetate added to
the oil were quite low. Frey (1988) determined that through the use of selected ion
monitoring (SIM) using fragments m/e 138, 123 and 73 the addition of as little as 2%
synthetic linalool and linalyl acetate could be determined in commercial samples of neroli
and bergamot oils. He also determined that as phenylpentadienal was found as an impurity
of synthetic cinnamaldehyde using fragments m/e 158, 129 and 128, the addition of as little
as 2% synthetic cinnamaldehyde could be determined in a commercial sample of Chinese
cassia oil.

Natural Products and Essential Oils

77

3.2.9 Component Ratio Measurement. A number of years ago Lawrence et al. (1989)
used component ratios rather than direct quantitative data to differentiate between
peppermint oils produced from the same clone in different regions of the Midwest and far
western United States. To simplify the component ratio data obtained from the analyses of
ca. more than 50 samples from each growing area selected, component ratios were
presented as polygons in a pattern recognition pictorial representation. More recently, this
data was expanded to include two seasons data for peppermint (Mentha piperita L.) so that
regional authenticity could be monitored. However, the data for peppermint was also used
to determine adulteration of U.S. peppermint with Indian peppermint oil (Table 34). Also,
with the use of other component ratios the adulteration of US peppermint oil with
dementholised oil obtained from M. amensis L. f. piperascens Malinv. ex Holmes (Table
35). Component ratios have also been used to determine adulteration in a number of citrus
oils as shown in Tables 36-38.

__.

C ~ ~ ! ! F - ~_ --"~ ~ ---_
~ MidvvGt--~ ! ! ! Willgmette _"_"Madras- _Ya-ima
cineole/menthol x 50
d
d
dcineole/germacreneD x 2/3
d
d
d
menthone/piperitonex 1/10
d
d
d
d
isomenthone/tmeritone x 2/5
/
d
Lawrence ( 1999)
" "

Idaho_"_
d

Table 34 Component Ratio that can be successfully used to detect adulteration of U S .

Peppermint Oil (1993) with Indian Peppermint Oil

- C E ! ? ? ! Ratio

,"",--,--EECE!

I
_

menthone/piperitonex 1/10
cineolehmonene
isomenthone/piperitonex 2/5
menthyl acetate/germacrene D x2
cineole/menthofuran
menthofurardisomenthonex2
Lawrence ( 1999)

--Wi!kmettp- ".--Ma*a?d

YaEma

L d , h o-

d
d
d
/

Table 35 Component Ratio than can be successfully used to detect adulteration of U.S.
Peppermint Oil (1993) with Chinese or Indian Dementholised Cornmint Oil
Ratios
I

..-.-.I

--

Genuine oil
.-I_."

....._._I

Ixx-"

__-_.I.""

..I._..l^l__.._..,.l

citronellal/terpinen-4-01
octyl acetate/a-terpineol
y-terpinenelsabinene+ P-pinene
trans-sabinene hydrate acetatel or-terpineol
Verzera et al. (1996)

"".","ll.

0.167- 1.875
0.842-4.742
0.661-1.279
0.704-3.323

"_

2 Adulterated
commercial oils 0.1 19-0.136
0.500-0.561
0.670-0.733
0.303-0.354

Table 36 Component Ratio of Genuine and Two Adulterated Bergamot Oils

Compound

_Rat_ip_--

--

1
"
1
1
1
1

6-3-carene
6-3-carendor-terpinene
6-3-carenelcamphene
Verzera et al. (1978)

"

k!!!!?!?Oil_--~

0-0.008
0-0.044
0.005-0.129

Orange Terpenes added to Lemon Ojl--_--n5%

10-_--%
20 %
_-".
0.008-0.014
0.012-0.022
0.020-0.032
0.044-0.086
0.069-0.146
0.104-0.254
0.145-0.274
0.226-0.468
0.328-0.744
."ll_lll"l.l

Table 37 Detection of Lemon Oil Adulterated with Orange Terpenes

xl__-

Advances in Flavours and Fragrances

78

-C

O
!
~
6-3-carene
&-3-~arene/camphene

Bitter
OrangeOil
Max 0.006
Max 1.092

Orange
- Oil Added
3%
5%
0.005-0.009 0.008-0.012
0.714-1286 1.000-1.714
I

Orange Terpenes Added

3%~
-5 % _
_
0.006-0.007
0.009-0.011
0.857-1.000
1.111-1.375

P. Dug0 et al. (1996)

Table 38 Bitter Orange Oil Adulterated with Orange Oil or Orange Terpenes
3.2.10 Chiral GC Analysis. Many components of essential oils possess one or more
asymmetric carbon atoms that exhibit optical activity. These chiral compounds of natural
origin are generally found in a characteristic enantiomeric distribution because they evolve
via enzymatically controlled biosynthetic synthesis.
The advent of chiral GC columns and multidimensional GC techniques in which the
components to be examined for enantiomeric distribution are "heart cut" from a standard
capillary GC column and the analyses are performed on the chiral column has become
almost a routine procedure for determining adulteration in commercial essential oils.
Examples of the use of enantiomeric distributions to determine adulteration in coriander,
mandarin and neroli oils can be seen in Tables 39,40 and 41, respectively,
to name just a few.
Genuine

Commercial Oils

50
10
90

(4R)-(+)-limonene
(3s)-(+)-linalool
(3R)-(-)-linalool
Frank et al. (1 995)

62
33
67

88
27
73

93
34
66

85
20
80-

Table 39 Percentage Composition of Commercial Coriander Oil Adulterated with (+)Limonene and Racemic Linalool
Compound

Lemon Oil
_ I - ^ x _

(lR5R)-(+)-P-pinene
(1S,SS)-(-)-P-pinene
(1R,5R)-(+)-sabinene
( 1S,5S)-(-)-sabinene
(4R)-(+)-limonene
(4s)-(-)-limonene
(3s)-(+)-linalool
(3R)-(-)-linalool
(4S)-(+)-terpinen-4-01
(4R)-(-)-terpinen-4-01
(4R)-(+)-a-terpineol
(4S)-(-)-a-terpineol
Mondello et al. (1998)

5.1
94.9
15.1
84.9
98.1
1.9
28.5
71.5
19.7
80.3
22.6
77.4

Mandarin Oil
^

"

97.6
2.4
78.8
21.2
98.0
2.0
86.9
13.1
11.4
88.6
28.9
71.3

l_ll_

Mandarin
__
Oinemon
_- _ _ Oil 9 9 : L
- 90:l
90.3
50.2
49.8
9.7
75.0
45.6
54.4
25.0
97.8
98.0
2.2
2.0
73.1
80.7
26.9
19.3
15.7
14.4
84.3
85.6
25.6
26.5
74.4
73.5

_ _ I _ _ x -

I_

Table 40 Chiral Analysis Used to Determine Addition of Mandarin Oil in Lemon Oil

Natural Products and Essential Oils

79

Nero'i Oil
Comg__ound ,l__.l_- ".
(lR,5R)-(+)-a-pinene
(1S,5S)-(-)-a-pinene
(1RSR)-(+)-p-pinene
(lS,SS)-(-)-p-pinene
(4R)-(+)-limonene
(4S)-(-)-limonene
(3s)-(+)-linalool
(3R)-(-)-halo01
(3s)-(+)-linalyl acetate
(3R)-(-)-linalyl acetate
(4S)-(+)-terpinen-4-01
(4R)-(-)-terpinen-4-01
(4R)-(+)-a-terpineol
(4S)-(-)-a-terpineoI
(R)-(-)-nerolidol
(S)-(+)-nerolidol
Mondello et al. (1994)

--I_"_..".-

....-^I_I_.l."

4.6- 13.6
86.4-95.4
0.1-0.8
99.2-99.9
93.1-98.5
1.5-6.9
9.4-29.2
70.8-90.6
1.8-4.6
95.4-98.2
35.3-49.0
5 1 .O-64.7
69.1-72.3
27.7-30.9
0.4-1.8
98.2-99.6

- _ _ _ _ I - - ~Commercial
. - ~ - - . - I _ "Oils

. _-I.__- 1
l_.."."."

"I

11.9
88.1
2.9
97.1
97.1
2.9
27.4
72.6
37.2
62.8
35.1
64.9
60.2
39.8
1.6
98.4

__ ..

"~

2 ..... 13.1
86.9
4.0
96.0
96.7
3.3
36.9
63.1
11.4
88.6
58.8
41.2
71.9
28.1
33.4
66.6
,

___-_I_I

""

_""

Table 41 Enantiomeric Distribution of Selected Compounds in Neroli Oils

Co-ound (Threshold pym)
octanaTr@EG67j--"^
3590 (47%)
1870 (25%)
5460 (6%)
1 inaloo12(0.0008)
1320 (2%)
nonana13 (0.000 1)
320 (3%)
500 (5%)
citronella14(0.046)
490
320
decanal' (0.0001 )
3360 (3 1%)
970 (9%)
nera14(0.074)
780
90
geranial' (0.0032)
480
1200
thymo14(0.0155)
590
dodecana14(0.055)
500
300
methyl N-methyl anthranilate4(0.00115)
4580 (4%)
340 (6%)
p-sinsensa13(0.00005)
a-sinsensa13(0.00005)
250 (5%)
3050 (57%)
Percentages in parentheses refer to importance to overall aroma of orange and mandarin oils.
'Buttery et a1 (1978) 'Padrayuttawat et al. (1997) 30hloff (1994) 4Devos et al. (1990)

Table 42 Comparative Selected Oxygenated Constituents of Orange and Mandarin Oils

4

CONSEQUENCES

The consequences of using commercial essential oils that are not genuine are well known.
However, there is a tendency to ignore the fact that they may be adulterated either with
synthetic (foreign) components unrelated to the oil composition (the worst case scenario),
synthetic compounds related to the oil composition or the addition of oils of a similar
composition, fractions of other oils or natural compounds from other sources. The
resultant blended oil will, therefore, have a reduced olfactory character, will be less potent
than the genuine oil and will have less impact on its end product which defeats its purpose
for use. Examination of selected highly odoured constituents of orange and mandarin oils
(Table 42) shows the concentration and threshold of these constituents. If the
concentration in ppm is divided by its threshold (also in ppm) the percentage importance of

Advances in Fluvours and Fragrances

80

these constituents reveals that in orange oil the major components responsible for the
odour character of this oil are octanal, decanal, linalool, p-sinensal, a-sinensal and
nonanal. Similarly, a-sinensal, octanal, decanal, methyl N-methyl anthranilate, nonanal
and linalool are the major components responsible for the odour of mandarin oil.

___-"
Compound (Thresholdppm) -_-.--octanal' (0.00007)
nonana12(O.OOO~)
decanal (0.0001)
a-ter ineol' (0.035)
nerol (0.029)
nera13 (0.074)
geranial' (0.0032)
indole3(0.000032)
methyl anthranilate3 (0.00115)
neryl acetate4 (2.0)
geranyl acetate4 (0.1 5)
methyl N-methyl anthranilate3 (0.00115)
(E)-nerolidol' (0.012)
linalool' (0.0008)

'

Concentrationin m m
Bitter Orange Oil - Petitgrain
1910 (49%)
360 (6%)
500 (58%)
1940 (35%)
5 90
56300 (19%)
70
9900 (4%)
4300
980
3800 (14%)
240
1140

17300
31600

830
3300 (7%)

800
32,550,000*

Neroli -_

17900 (1%)
6900
4100
6500 (4%)
600 (36%)
1100 (2%)
9200
13800
3 1800 (53%)
17600 (3%)
15,590,000*

*Major odour characters of petitgrain and neroli oils. Percentages in parentheses refer to importance to
overall aroma of bitter orange and the odour characters of petitgrain and neroli oils.
'Buttery et a1 (1978) 'Ohloff (1994) 3Devos et al. ( 1990)4Yanget a1 (1992) 5Padrayuttawat et al. (1997)

Table 43 Comparative Selected Oxygenated Constituents of Bitter Orange, Petitgrain and

Neroli Oils
Using this same procedure of dividing the concentration (pppm) by the threshold
(ppm) for bitter orange (Table 43), it can be seen that the most important odour characters
in the oil are octanal, decanal, linalool, nonanal and linlyl acetate. Also in this same table,
the odour characters for petitgrain and neroli are obviously linalool and linalyl acetate;
however, the other odour characters of lesser importance in these oils are nonanal, aterpineol, geranial and nerol for petitgrain oils and methyl N-methyl anthranilate, indole,
geranial, (E)-nerolidol, methyl anthranilate and a-terpineol in neroli oil. Because of the
importance of linalool and linalyl acetate in petitgrain and neroli oils, the addition of
racemic linalool as an adulterant can greatly affect the potency of these oils as (+)-halo01
has a detection threshold of an order of magnitude higher than the naturally occurring (-)linalool.
In Tables 44-45 a similar treatment of lemon oil and three oils of grapefruit can be
seen. It is important to note in the grapefruit oil example that the naturally occurring (+)nootkatone has a threshold of 1 ppb, whereas its (-)-enantiomer has a threshold of 1 ppm.
Hence, any addition of racemic nootkatone can drastically reduce the odour character of a
grapefruit oil.
It is hoped from this discussion of oil composition variation and the methods used to
determine genuineness or adulteration that there is a realisation that diluents or adulterants
have a deleterious effect on the odour intensity of the oil. Finally, it is hoped that the users
of commercial essential oils will become more demanding about their purchase and use of
genuine essential oils so that genuine oils become the norm rather than the exception.

Natural Products and Essential Oils

81

Concentrationin ppm
Compound
(Thresholdmm)
---- ___
octanaF3k0007)
octano12(0.01 1)
300
100
linaloo13 (0.0008)
2000 (6%)
1400 (5%)
nonanal' (O.OOOI)
1300 (31%)
1300 (37%)
citronella14(0.046)
1100
800
decanal' (0.OOOl)
500 (12%)
500 (1 4%)
nero14 (0.029)
150
300
nera14 (0.074)
12500
7300
geranial' (0.0032)
11400 (10%)
20200 (15%)
neryl acetate' (2.0)
5700
5 100
geranyl acetate' (0.15)
3800
4900
nootkatone' (0.0001)
Percentages in parentheses refer to importance to overall lemon oil aroma.
'Buttery et al. (1978) 20hloff (1994) 3Padrayuttawat et al. (1997) 4Devos et al. (1990)
'Yang et al. (1992)

.
I

Table 44 Comparative Selected Oxygenated Constituents of Lemon Oil

Compound

Concentrationin DDm
Grapefruit Oil
Marsh
Red Blush
5800 (63%)
3400 (54%)
1400
1400
2500 (2%)
2000 (6%)
600 ( 5 % )
500 (6%)

Duncan
octanal' (0.00007)
2300 (27%)
octanol' (0.01 1)
700
linaloo13 (0.0008)
4500 (5%)
nonana12 (O.OOO~>
700 (6%)
citronella14(0.046)
2200 (1 8%)
2600 (20%)
2600 (29%)
decanal' (0.000 1)
nero14 (0.029)
300
200
300
nera14 (0.074)
600
600
500
geranial' (0.0032)
800
1000
700
neryi acetate' (2.0)
50
50
100
geranyl acetate' (0.15)
800
500
600
nootkatone2 (0.OOOl)
5400 (44%)
1300 (10%)
700 (8%)
Percentages in parentheses refer to importance to overall grapefruit oil aroma.
'Buttery et al. (1978) 'Ohloff (1994) 3Padrayuttawat et al. (1997) 4Devos et al. (1990)
'Yang et al. (1992)
_
l
_
l
l
_
l
l
l
l
l
l
l
_
l
i

Table 45 Comparative Selected Oxygenated Constituents of Grapefruit Oil

References

= G. Friedlander and J. W. Kennedy, Nuclear and Radiochemistry. p. 388, J . Wiley &

Sons, New York, (1955)
M. Calvin and J. A. Bassham, The Photosynthesis of Carbon Compounds. W. A.
Benjamin, Inc., New York (1962).
M. D. Hatch and C. R. Slack, Photosynthesis by Sugar-cane Leaves. Biochem. J., 101,

.
.
.

103-111 (1966).

M. M. Bender, Variations in the 12C/'3C ratios of plants in relation to the pathway of

photosynthetic carbon dioxide fixation. Phytochemistry, 10, 1239-1244 (197 1).

82

Advances in Flavours and Fragrances

M. Oswiecimska, Chemotaxonomic investigations of an azulene in Achillea. Herba

Pol., 19, 207-215.
M. Oswiecimska, Correlation between the number of chromosomes and
prochamazulene in eastern European Achilla. Planta Med., 15,389-395 (1974).
P. Teisseire and A. GalfrC, Contribution to the knowledge of Ylang Ylang 3rd oils
adulteration of exotic basil oils, Recherches, 19, 269-274 (1974).
L. Peyron, J. Acchiardi and P. Pellerin, Sur la presence anormale dhexylene glycol
dansles huiles essentials commerciales de vktiver et de cananga. Ann. Fals. Exp.
Chim., 68,485-499 (1975).
B. M. Lawrence, Progress in Essential Oils, Perfum. Flavor. 1(3),45-46 (1976).
K . Bruns, Contribution to the study and quality evaluation of Patchouli oil, Parfum.
Kosmet., 59(4), 109-115 (1978).
R. Gonfiantini, Standards for stable isotope measurements in natural compounds.
Nature, 271, 534-536 (1978).
J. Bricout and J. Koziet, Characterization of synthetic substances in food jlavors b y
isotopic analysis. In: Flavors of Foods and Beverages, Chemistry and Technology.
Edits., G. Charalambous and G. E. Inglett, p. 199-208, Academic Press, New York
(1978).
R. G. Buttery, D. G. Guadagni and L. C. Ling, Volatile aroma components of cooked
artichoke. J. Agric. Food Chem., 26,791-793 (1978).
P. G. Hoffman and M. Salb, Radiocarbon (I4C) method for authenticating natural
cinnamic aldehyde. J. Assoc. Offic. Anal. Chem., 63, 1181-1183 (1980).
M. H. OLeary, Carbon isotope fractionation in plants. Phytochemistry, 20, 553-567
(1981).
L. Mondello, P. Dugo, K. D. Bartle, B. Frere and G. Dugo, On-line high performance
liquid chromatography coupled with high resolution gas chromatography and mass
spectrometry (HPLC - HRGC - MS ) for the analysis of complex mixtures containing
highly volatile compounds. Chomatrographia, 39, 529-538 ( 1994).
F. Saez., Essential oil variability of Thymus zygis growing wild in southeastern Spain.
Phytochemistry, 40, 819-825 (1995).
U. Hener, S. Faulhaber, P. Kries and A. Mosandl, On the authenticity evaluation of
balm oil (Melissa oflicinalis L.), Pharmazie, 50,60-62 (1995).
C. Frank. A. Dietrich, U. Kremer and A. Mosandl, GC-IRMS in the authenticity control
of the essential oil of Coriandrum sativum L. J. Agric. Food Chem., 43, 1634-1637
(1995).
J-C. Chalchat, R-P. Garry and A. Muhayimana, Essential oil of Tagetes minuta from
Rwanda and France: Chemical composition according to harvesting location, growth
stage and part of plant extracted. J. Essent. Oil Res., 7, 375-386 (1995).
P. Sanchez Gomez, J. A. Sotomayor Sanchez, M. C. Soriano Cano, E. Correal
Castellanos and M. C. Garcia Vallejo, Chemical composition of the essential oil of
Thymus zygis ssp. gracilis cv. linalool type und its per$ormance under cultivation. J.
Essent. Oil Res., 7,399-402 (1995).
AFNOR, Huiles Essentielles. Tome 2. Huile essentielles de thym a thymol (Thymus
zygis (Loejl.) L.), type Espagne. NFT 75-349, pp. 473-48 1: Huile essentielles dhysope
(Hyssopus oflicinalis Linnaeus). NFT 75-355, pp. 509-5 13, AFNOR, Paris (1996).
Padrayuttawat, T. Yoshizawa, H. Tamara and T. Tokunaga, Optical isomers and odor
thresholds of volatile constituents in Citrus sudachi. Food Sci. Techol. Int. (Tokyo),3,
402-408 ( 1997).

Natural Products and Essential Oils

83

S. Hose, A. Zhglein, Th. van den Berg, W. Schultze, K-H. Kubeczka and F-C.
Czygan, Ontogenetic variation of the essential leaf oil of Melissa oflicinalis L.,
Phannazie, 51,247-253 (1997).
D. Juchelka, T. Beck, U. Hener, F. Dettmar and A. Mosandl, Multidimensional gas
chromatography coupled on-line with isotope ratio mass spectrometry (MDGC-IRMS):
Progress in the analytical authentication of genuine flavor components.1 J. High
Resol. Chromatogr., 21, 145-151 (1998).
D. Moyler and N. Moss, Mint Oils: Potential for standardizing profiles with natural
flavoring substances. Perfum. Flavor., 23(2), 37-42 ( 1998).
L. Mondello, M. Catalfamo, A. R. Proteggente, I. Bonaccorsi and G. Dugo,
Multidimensional capillary GC-GC for the analysis of real complex samples. 3.
Enantiomeric distribution of monoterpene hydrocarbons and monoterpene alcohols of
mandarin oils. J. Agric. Food Chem., 46,54-61 (1998).
J. S. Chalchat, A. Michet and B. Pasquier, Study of clones of Salvia oflicinalis L. yields
and chemical composition of essential oil. Flav. Fragr. J., 13, 68-70 (1998).
L. Bandoni, I Mizrahi and M. A. Juhez, Composition and quality of the essential oil of
coriander (Coriandrum sativum) from Argentina. J. Essent. Oil Res., 10, 581-584
( 1998).
B. M. Lawrence, Mint oils of commerce. In: 4' Colloque Produits Naturels D'origine
Vkge'tale. Edits., G. Collin and F. X. Garneau, pp. 102-131, Univ. Quebec at
Chicoutimi, Quebec, Canada (1999).
G. Dugo, K. D. Bartle, I. Bonaccorsi, M. Catalfamo, A. Cotroneo, P. Dugo, G.
Lamonica, H. McNair, L. Mondello, P. Previti, I. Stagno d'Alcontres, A. Trozzi and A.
Verzera, Advanced analytical techniquesfor the analysis of citrus essential oils. Part 2.
Volatile fraction: LC-HR C and MD GC. Essenz. Deriv. Agrum., 69, 159-217 (1999).
G. Dugo, K. D. Bartle, I. Bonaccorsi, M. Catalfamo, A. Cotroneo, P. Dugo, G.
Lamonica, H. McNair, L. Mondello, P. Previti, I. Stagno d'Alcontres, A. Trozzi and A.
Verzera, Advanced analytical techniques for the analysis of Citrus essential oils. Part
3. Oxygen heterocyclic compounds: HPLC, HPLUMS, OPLC, SFC, Fast HPLC
analysis. Essenz. Deriv. Agrum., 69,251-283 (1999).
U. Asllani, Chemical composition of Albanian sage oil (Salvia oflicinalis L.). J. Essent.
Oil Res., 13, 79-84 (2000).
B. R. Rajeswara Rao, P. N. Kaul, G. R. Mallavarapu and S. Ramesh, First observation
of little leaf disease and its impact on the yield and composition of the essential oil of
rose-scented geranium (Pelargonium sp.). Flav. Fragr. J., 15, 137-140 (2000).
W. Feger, H. Brandauer, H. Ziegler and E. Ziegler, Analytical investigation of the
increasingly valued essential oils of Murcott (honey) tangerine and Sweetie
(oroblanco). Paper presented at 2ndInternational Citrus Symposium, Orlando, Feb. 1519 (2000).
B. M. Lawrence, unpublished data.

STABLE ISOTOPES FOR DETERMINING THE ORIGIN OF FLAVOUR AND

FRAGRANCE COMPONENTS: RECENT FINDINGS.

Dr. Daniel Joulain

Robertet, France

1 INTRODUCTION
During the last couple of decades, the use of stable isotopes has developed in parallel with
that of very sensitive and precise methods for their detection. Such methods rely on
physical characteristics of these species: mass differences are measured by isotopic mass
spectrometry, resonance differences exhibited by nuclei are measured by NMR, and
photon emission from excited atoms is measured by atomic emission spectrometry.
Nowadays, isotopic ratios of hydrogen, carbon, nitrogen, oxygen and sulphur are
mostly determined using two methodologies: the global stable isotope ratio analysis
(SIRA) by mass spectrometry, and the quantitative positional isotopic ratio analysis by
NMR. These ratios have been widely used for assessment of the origin, mainly natural
versus synthetic, of defined isolated compounds in flavour mixtures
Both methods have
their own advantages and disadvantages, and should preferably be used in combination. In
any case, definitive information is obtained only if comprehensive databases have been
compiled beforehand, using samples of known origin. The possible effects of isotope
fractionation during isolation procedures should not be overlooked either.

2 NATURAL vs SYNTHETIC AND STABLE ISOTOPES

Stable isotope methods have gained increasing importance in the authenticity assessment
of flavour compounds. Measurement of carbon and hydrogen stable isotope ratios, which
required multi-milligram quantities of pure compounds when using off-line methods,
became more straightforward and convenient with the introduction of commercial online GC-IRMS instruments in 19902.The latest GC-P-IRMS instrumentation involves a
pyrolysis interface (P),
which allows one to measure oxygen isotope ratios. Over the 10
past years, a significant amount of data have been published on the differentiation of
flavour and essential oil components, from (or versus) their synthetic (or so-called natureidentical) counterparts3. Therefore, these data are considered as an authentication tool
only. In this respect, a further example of this possibility is provided by the case of ( E ) - p damascenone, a ubiquitous key flavour contributor in natural flavours. When isolated from

85

Natural Products and Essential Oils

rum (made from cane sugar), damascenone shows a 13Cenrichment that is typical of a C4
plant, whereas the same material obtained by a biotechnological process from a
hydroxylated carotenoid precursor present in a C3 plant would display a 13C/'*C ratio
rather similar to that of synthetic damascenone (Table 1).

natural (rum)
natural (biotech)
I synthetic (fossil)

Table 1

- 16.2 k 0.2
- 31.9 k 0.6

I - 29.9 k 0.2

I3Cf2Cratio for dlfSrent sources of damascenone

Some additional information is given by the measurement of *H/'H and 180/'60

ratios,
but definitive information is provided by Site-Specific Natural Isotope Fractionation NMR
(SNIF-NMR) (table 2).

c
origin
natural (rum)
ssynthetic
nthetic

= %,

Table 2

16
6l80a

-7.7
20.9

I 6 *H
~~

-162
-126

1 Site 1 *H/'H
48
137

Site 2 2 ~ / 1 Site
~ 32
34
84
145
217

SNIF-NMR values for different sources of damascenone

Not only may stable isotopes be used for determining the origin of natural products, as
in the case of biotechnology-derived isolates (see below), but they can be also used to
extract more intimate clues which may help to characterize some parameters involved in
fermentation processes. Hence, we reported for the first time in 1988 that 6-pentyl-apyrone (6-PP), a metabolite of Trichoderma viride, can be characterized by carbon SIRA4.
As expected, when fed with either grape juice sugars (a C3 plant) or corn steep li uor (a
C4 plant) as a sole carbon source, the fungus biosynthesises 6-PP with specific 'C/I2C
ratios typical of source (Table 3).
carbon source
grape juice sugars
corn steep liquor
synthetic

Table 3

6 13C%o k 0.2 ( PDB)

-27.1
-12.4
-28.8

Carbon SIRA of 6-PPof different origins

Therefore, one can see that the isotopic signature of de novo fermentation-derived
products can be tailor made. Indeed, as of today, very little has been published on the use
of stable isotopes as tracers to support manufacturing claims in flavours and fragrances.
Interesting1 a recent patent' claims the use of perfume or flavour ingredients labelled
with *H, I%,
I8O and 13C for tracing formula duplication, a method that had been

'

86

Advances in Flavours and Fragrances

suggested long ago by Brazier6. However, there is little doubt that isotope labelling is used
today in drug tagging for counterfeit control (at least to a certain extent), and in forensic
applications (explosives, illicit drugs etc.).

3 STABLE ISOTOPES AND BIOGENETIC CONSIDERATIONS

Fruit flavour volatiles may be generated through different biosynthetic pathways.
Pineapple, banana and apple flavour volatiles have been investigated by carbon SIRA7.
Pineapple is an example of a lant following the unusual Crassulacean Acid Metabolism
(CAM). Therefore, typical C enrichment is observed for all the volatile aroma
components. Measurement of the relative concentration and carbon SIRA of flavour
components even allows one to characterize the fruits according to their geographical
origin and variety (Table 4).
Queen Victoria
rel.

613c(o/o0)

Compound

Smooth Cay nne

rel.
6I3c(%o)

Methyl 2-methylbutyrate
-19.1 f 0.35
Diethyl malonate
n.m.
Ethyl 2-methylbutyrate
-19.3 k 0.35
Methyl hexanoate
-20.9 5 0.6
Methyl 3-methylthiopropionate
n.m.
Ethyl hexanoate
-20.5 +_ 0.23
Methyl octanoate
-23.0 f 1.0
Methyl 3-acetoxyhexanoate
n.m.
ethyl octanoate
-24.0 f 1.4
n.m.: not measured; n.d.: not detected

Table 4

6.60 -20.5 f 0.22

0.13 - 16.4 k 0.2
n.d.
9.40
13.8 -22.6 f 0.2
4.2 -25.2 k 0.2
n.m.
21.7
5.1 -19.7 f 0.5
0.2 -22.1 k 0.6
10.4
n.m.

%
-

8.25
3.10
n.m.
38.9
11.7
2.0
3.0
8.7
0.5

Carbon isotopic deviations of some pineapple volatile compounds

In the case of purple passionfruit (PassiJlora edulis Sims), a C3 plant, the usual carbon
isotope deviations are observed for all the flavour components, but, interestingly, the
putative mevalonic route to Edulan I and p-ionone is associated with a significant I3C
depletion* (Table 5).
compound
ethyl butyrate
ethyl hexanoate
n-octyl butyrate
hexyl hexanoate

Table 5

613C (%o)
-35.6 k 0.3
- 3 1.9 k 0.7
-32.2 f 0.6
-30.4 +_ 0.8

compound
Edulan I
p-ionone

813C (%o)
-37.9 f 0.6
-35.0 k 0.4

Carbon isotopic deviations of certain purple passion fruit components

Natural Products and Essential Oils

87

STABLE ISOTOPES AS CLUES OF MANUFACTURING

TECHNIQUES EMPLOYED FOR NATURAL FLAVOUR ISOLATES

4.1 Raspberry ketone

Deuterium SIRA finds some usefulness in the case of raspberry ketone since the natural
modification is significantly enriched in deuterium, relative to the synthetic counterparts.
Using SNIF-NMR, Martin et aL9 had even shown more precisely that in the case of
anethole and other p-coumaric-related phenylpropanoids, the natural products exhibit a
high aromatic deuterium enrichment. This is understandable if we bear in mind that the
aromatic ring originates from photosynthesized carbohydrates having a high overall D/H
content. Consequently, raspberry ketone would be expected to display the same feature,
and this is indeed the case, as we show in Table 6.

sum of aromatic

Synthetic raspberry ketone

Table 6

0.355 (mean)

D/H content in raspberry ketone

It had also been shown previously that compounds which derive from p-coumaric acid
(anethole, estragole, vanillin ...) are usually deuterium-enriched in the a-position with
respect to the hydroxyl or methylated hydroxyl group (meta from the C3 side chain) as a
consequence of a specific isotope effect operating during the aromatization step of
prephenic acid into tyrosine, the precursor of p-coumaric acid (figure 1).
Dvruvic acid
I

<

CH,-CO-COOH
+
>
HOCO-CHOH-CHOH-CHO

erythrose

HO

?H

QCOOH
I

OH

shikimic acid

Figure 1

-H~@~~-cooH

prephenic acid

Overview of route to p-coumaric acid from pyruvic acid and erythrose

Again, this feature is also present in the raspberry ketone: when obtained from natural
precursors, it exhibits a higher D/H content on site 3, ortho relative to the hydroxyl group,
than on site 2. In aromatic rings of synthetic origin (anethole, raspberry ketone etc.), the
D/H ratios are practically identical on both sites 2 and 3 (Table 7).

Advances in Flavours and Fragrances

88

deuteriuh
enrichment

HO

Natural anethole (fennel)

Synthetic anethole
Natural raspberry ketone
Synthetic raspberry ketone

Table 7

raspberry ketone

D/H molar fraction fi D/H molar fraction f3

(mean values)
(mean values)
0.164
0.228
0.176
0.170
0.190
0.247
0.175
0.175

D/Hvalues in anethole and raspberry ketone

Concerning the specific aspect of the double bond reduction of the raspberry ketone
precursor, one should be able to distinguish between bioconversion and catalytic
hydrogenation processes, even though the final result seems to be the same. SNIF-NMR
would be expected to provide definitive information on the mode of reduction that is used
to convert the unsaturated precursor into raspberry ketone. We were able to perform
tedious deuterium natural abundance measurements on carbon atoms a (C3) and p (C4)
relative to the carbonyl group, and we could establish that site p is significantly enriched
relative to site a,in all products generated through biological means. Thus, in these
cases, the ratio of deuterium molar fractions on site p versus site a is 1.16 f 0.02, which
appears to be quite sufficient for determining the mode of reduction that has been used.

4.2 E-2-hexenal
The biochemical oxidative degradation of unsaturated C- 18 fatty acids into C-6 green
compounds is very well documented. Among these compounds, E-2-hexenal is of great
interest since it occurs as such in many fruit species, and it is also generated upon tissue
disruption during fresh plant processing. Today, this feature is used for the
biotechnological production of both E-2-hexenal and 2-3-hexenol or E-2-hexenol.

Natural Products and Essential Oils

89
2

6W

C
4

(Q-2-hexenal

Table 8 summarizes the deuterium abundance values recorded by SNIF-NMR on two

selected sites 1 and 3 of the E-2-hexenal molecule from various origins.

Origin

Site 1

Site 3

(DM, ppm) ( D m PPm)

Synthetic
Tomato fruit juice
Natural a
Natural b
Natural c
Natural d
Natural e
Alfalfa leaf
Microbial oxidation of natural E-2-hexenol
Chemical oxidation of natural E-2-hexenol

Table 8

123
77
74

65
63
67
67
63
78
232

95
142
132
136
164
153
157
116
171
172

D/H values for E-2-hexenal of diflerent origins

Not only can the SNIF-NMR method provide a clear-cut image of the natural origin of
E-2-hexena1, but, as in the case of raspberry ketone, it can also allow one to distinguish
products obtained from natural E-2-hexenol by oxidation using either a microbial process,
for example using Candida boidinii'', or a chemical reaction. Indeed, in the case of the
chemical oxidation, the deuterium abundance at site 1 bearing the aldehyde function of E2-hexenal is dramatically increased, whereas in the case of the bio-converted product, it is
similar to that of natural counterparts directly isolated from plants. Interestingly, it has
been shown recently that when E-2-hexenol and E-2-hexenal co-exist in a given plantderived mixture, the global deuterium content is lower in the aldehyde than in the
alcohol'2.

5 THE CASE OF A NATURAL RAW MATERIAL FOR FRAGRANCES: LICHEN

RESINOIDS
As of today, about 3000 tons per year of lichen are processed into resinoids, mostly by
French manufacturing companies located in the Grasse area. The raw material belongs to
two main species: Evemia prunastri (oak moss) and Pseudevemia furjiuracea ("tree
moss"). Mono-aromatic compounds derived from depsides are among the main odourdonating compounds in both species. Some of them (e.-g. 2) have been available to
perfumers as synthetic chemicals for more than 25 years. Lichens are symbiotic organisms
of fungi and algae, and appear to utilize C3 photo~ynthesis'~.
We were able to confirm that
this is indeed the case, in taking p-orcinol 1, methyl P-orcinolcarboxylate 2,atranol 3 and
methyl haematommate 4 as examples. Not surprisingly, a significant deuterium depletion
is again observed in the natural products, relative to their synthetic modifications (Table 9).

Advances in Flavours and Fragrances

90

compound origin
1
natural
2
natural
natural
3
natural
4
synthetic
1
2
synthetic
synthetic
3
5
synthetic

613C %O f 0.2 (PDB)

- 31.3
- 30.1
- 28.4
- 29.1
- 32.0
- 29.7
- 33.6
- 31.3

DEIk0.2 (ppm)
128.5
123.1
139.0
132.3
142.6
148.0
140.4
136.3

Table 9
9

y43

y 4 3

V H

HOQ

5\

COOCH3

'5

CHO

CHO

For the first time, we could measure the site specific deuterium abundances in
depsidic compounds, though not without some technical difficulties due to the relatively
large amounts of pure compounds required, and problems related to their concentration in
the solvent, and to the internal standard. Partial results are depicted in Table 10:

Compound
Natural 1
Natural 2
Natural 3
Synthetic 1
Synthetic 2
Synthetic 3

-CHO
-COOCH3
8-CH3
9-CH3
D/H (ppm) D/H (ppm) D/H (ppm) D/H (ppm)
11 1.2 k 1.7 96.4 k 1.0
98.0 f 1.6 1 12.8 f 0.9
92.8f 1.2
112.5f 7.4
106.6 k 1.0
125.6 0.2 145.3 f 1.1
143.7 f 4.3 126.5 f 0.6 142.4f 2.2
127.9 & 0.5
102.0f 6.6

Table 10

*CH360-S-CoA
14

+3

14

14

HOOC-CH2-CO-S-CoA
3 coz

*CH3
orsellinic acid

Figure 2

Natural Products and Essential Oils

91

It is noteworthy that both the 9-CH3 in 2 and the methanol involved in the esterasemediated esterification of natural P-orcinolcarboxylic acid are similarly deuteriumdepleted. The C1 fragment incorporation to form the 9-CH3 in 2 and then the CHO group in
3 or 4, is thought to be achieved by formate prior to the completion of the aromatic ringI4.
Therefore, although a possible isotopic effect during esterification cannot be ruled out,
these preliminary results are consistent with the involvement of these two one-carbon units
(formic acid and methanol) at a later stage during the biosynthesis of 2. As the biosynthetic
route to orsellinic acid (figure 2) involves one acetyl-coenzyme-A C2 unit and three
malonyl-coenzyme-A Cs units, it is clear that the 8-CH3 of 2 originates from one initial
acetic acid unit. As a result, significant deuterium abundance discrepancies are observed
for the 8-CH3 and 9-CH3 groups in 1and 2.
It should be stressed that such considerations, which are based solely on a few
preliminary measurements, would warrant further confirmation starting from many
different samples from various origins. Work in this direction is currently in progress.
6 ACKNOWLEDGMENT
The author thanks G. J. Martin for fruitful discussions.
References

2
3
4

5
6
7
8

9
10

11
12
13
14

A. Rossmann and H.-L. Schmidt, in: Frontiers in Flavour Science, P. Schieberle, K.-H.
Engel (eds), Deutsche Forschungsanstalt fur Lebensmittelchemie (Publisher) Garching,
Germany, 2000, pp 97- 10 I.
W. Meier-Augenstein, J. Chromatogr. A , 1999,842, 35 1.
H.-L. Schmidt, A. Rossmann and R. A. Werner, in: Flavourings, E. Ziegler, H. Ziegler
(eds) Wiley-VCH. Weinheim New York Chichester, 1998, pp. 539-594.
D. Joulain, N. Blaison, R. Laurent, G. Mangeot, C. Senemaud and J.-M. Belin, oral
communication, Annual Meeting of the SociCtC Franqaise de Chimie, Nice, Sept. 5-8,
1988.
Nippon Oxigen Co., Ltd., Japan Kokai Tokkyo Koho (priority 04/07/1998).
R. Guilly and J.-L. Brazier, Rivista Italiana EPPOS, Special issue, 1991, 249.
B. Salmon, Doctorate Thesis, University of Nantes (1996).
D. Joulain, The flavor of tropical fruit juices. Annals of the 23rd IFU Symposium,
Havana (Cuba), 7- 13 February 2000, pp. 191-205
G. J. Martin, M.-L. Martin, F. Mabon and J. Bricout, J. Am. Chem. SOC., 1982, 104,
2658.
D. Joulain, C. Fuganti, M. Mendoza and G. Zucchi, , i<Biogeneration of raspberry
ketone and related compounds D, oral communication, IFT Annual Meeting, New
Orleans, 26 June 1996.
M. Nozaki, Y. Washizu, N. Suzuki and T. Kanisawa, in: BioJlavour 95, P. Etievant, P.
Schreier (eds.), INRA Editions, Paris, 1995, pp. 255-260
K. Hor, C. Ruff, B. Weckerle, T. Konig and P. Schreier, J. Agric. Food Chem., 2001,
49, 21.
C. F. Culberson, W. L. Culberson and A. Johnson, in Chemical and Botanical Guide to
Lichen Products, Second Supplement, St. Louis, (1977)
M. Yamazaki, M. Matsuo and A. Shibata, Chem. Pharm. Bull., 1965,13, 1015.

FRAGRANT ADVENTURES IN MADAGASCAR

The analysis of fragrant resin from Canarium madagascariense

Robin Clery
Natural Products Research, Quest International, Ashford, Kent. TN24 OLT

1 INTRODUCTION
Natural Products Research in the fragrance industry is increasingly about the discovery of
new scents rather than the identification of natural ingredients. To discover new natural
scents you have to search for them in their native environment. One of the most unique
botanical environments on earth is the rainforest of Madagascar where 80% of the 1200
plant species are endemic. It was the combination of this and the number of plants that are
grown and extracted for their perfume in Madagascar that lead us to choose this location
for a fragrance prospecting expedition. One of the most remote relics of original rainforest
is the Montagne d'Ambre in the far north of Madagascar. Its name even suggests there may
be some historical connection with fragrant materials.
Several species were found which were of fragrance interest, notably Canarium
madagascariense, a close relative of elemi which exudes a very fragrant resin. The resin
from this species has been examined by GC-olfactometry and GCMS and compared to that
of elemi. Elemi gum is collected from several species but mainly Cluzonicum (Miqu)
A.Gray, (syn. Ccommune Wild) in Malaysia and the Philippines. The resinoid and
essential oil are important fragrance ingredients valued for both their odour and fixative
properties'.2. The resin of Cmadugascariense is not traded as a fragrance ingredient,
although we heard evidence that it is used locally both as incense and medicially.

2 METHODS
2.1

Sample collection

Samples of the resin from C.madagascariense were collected in the Montagne D'Ambre
region of Madagascar in October 2000 with the collaboration of E.S.S.A (Ecole
SupCrieure des Sciences Agronomiques) and A.N.G.A.P(l'Association Nationale pour la
Gestion des Aires ProtCgCes). The resin which is a clear golden to dark brown colour as it
is exuded from wounds in the bark of the tree dries to a white crystalline mass. The fresh
resin has an interesting and unusual fragrance similar to frankinsence but with strong

Natural Products and Essential Oils

93

citrus, lime and juniper top notes. A sample of elemi (C. Euzonicum) from the in-house
reference collection was selected for comparative analysis.

2.2 Extraction and analysis

The resin was extracted with a 80:20 mixture of hexane:acetone, filtered and concentrated.
A second sample was dissolved in ethanol, filtered and concentrated. Both samples were
analysed by GC-MS and GC-olfactometry.
GC-olfactometry was carried out using an HP5890 fitted with an OPTIC@ injector and
equipped with an effluent splitter, a custom built smelling port and an odour response
button to record the detection of an odour electronically.
Column: HP5 25m x 0.32mm x 0.52um
Carrier: He @ 40cdsec
Temp. program: 50" 3"/min. 280"(5min)
GC-MS analysis was accomplished using a Varian3400 GC coupled to a Finnigan ITS40
ion trap mass spectrometer.
Column: HP ultra2 (Hewlett Packard), 50m x 0.2mm 0.33pm,
Injector: 270"
Carrier: Helium at 1.6ml/min
Temp. program: 50" 2"/min. 270"
Scan range: 35 - 450 amu
Identifications were made by reference to both in-house and commercial libraries of mass
spectra 3 , with confirmation by relative retention indices (RRI).
Comparative GC-FID analyses were carried out using an HP6890 equipped with an FID
and an OPTIC0 injector.
Column: HP5 25m x 0.2mm x 0.33um
Carrier: Hydrogen @ 40cdsec
Temp. program: 50" lO"/min. 280"(5min)
Relative retention idices (RRI) were calculated on this HP5 column using n-alkanes from
C7 (heptane) to C30 (triacontane).

3 RESULTS

3.1

GC-olfactometry

A typical annotated chromatogram is shown in figure 1. The description of the odour

associated with each peak is recorded and the odour response chromatogram is annotated
with the descriptions. The majority of peaks had the familiar odours of the
monoterpenoids. A less familiar odour, but one consistent with the rich woody odour of the
resin itself was recorded at 15.0 min. From the odour descriptions obtained by GColfactometry it is possible to ascertain which of the components present make an important
contribution to the fragrance of the original sample.

Advances in Flavours and Fragrances

94

E
Eo

1
L

woody pinene

__ ..-2
7.480
--.______
.

8.871
8.988

Ih

9.292

-9.526

w s
3

r n m!

10.275

oily terpenic
-7 10.860
-

__ ___

i l l 126

-. _l_.
Il
__..__ -.

terpenic + cineole

- oily
- .t-E p n__ i c-

= 12 352

cresolic I phenolic
-

bergamotI linalool
PI

..

___

14.066

resinous woody typical

--_- - -.____
- --715 013

mothballs I camphor
-__

_ _ -J 16 526

/ 17.257
i-

camphor woody
0

medicinal

>-

117 706

+
6

18 113
18 258

stong woodyterpenid

16.873

-17.395
- 17 768
181361

18 296

19.393

I.
L

Figure 1

Annotated GC olfactogram of Canarium madagascariense resin.

95

Natural Products and Essential Oils

3.2

GC-MS results

Having completed the olfactory assessment, the sample was analysed by GC-MS, with
particular attention being paid to those areas of the chromatogram for which interesting
odours had been recorded. The peak at RT 15.0min was identified as cis p-menth-2-en-1-01
and a trace of the trans isomer was found at RT 15.5min co-eluting with another
compound. Although not readily available, this material has been previously reported as
having a woody, sandalwood, minty, lime odour somewhat similar to 4-terpineol. This
description is consistent with that from GC-olfactometry and lead us to conclude that it
makes an important contribution to the overall fragrance of the resin. The reference sample
of elemi was also analysed by GC-MS and the full results are given in Table 1.
Chromatograms of the extracts from the two species are given in Figure 2. The
significant differences between the two species are the presence of elemol, elemicin and
limonene in elemi oil and the greater proportion of p-cymene and terpinolene in
C.madagascuriense extract. Both samples contain alpha and beta phellandrenes and
previous anlayses have indicated higher levels of phellandrenes in elemi4 but as these
materials are not particularly stable the proportion in the oil has more to do with the age,
treatment and history of the resin or oil sample than the botanical source.

Table 1 Composition of extract of C.madagascariense and Cluzonicum

RRI
930
938
95 1
954
977
982
992
1009
1015
1021
1029
1033
1034
1037
1039
1049
1063
1072
1074
1091
1101
1121
1123
1129
1147
1140
1142
1151
1155
1174
1176

Compound name
alpha-thujene
alpha-pinene
alpha fenchene
camphene
sabinene
beta-pinene
myrcene
alpha-phellandrene
3-carene
alpha-terpinene
para-cymene
limonene
beta-phellandrene
1,8-cineole
cis-ocimene
trans-ocimene
gamma-terpinene
trans-sabinene hydrate
para-cresol
terpinolene
linalol
fenchol
cis-2,8-menthadien-1-01
cis-p-menth-2-en-101
1,3,8-~-menthatriene
trans-p-menth-2-en-1-01
limonene epoxide
trans-beta-terpineol
camphor
cis-beta-terpineol
borneol

Canarium
Cmadagascariense
%RPA
0.60

10.70
0.12
0.16
0.06
0.55
0.75
5.48
tr
2.76
21.74
2.00
5.00
0.05
0.03
tr
1.62
0.10
tr
14.65
0.06
tr
0.10
0.12
0.10

0.90
0.18

Elemi
C.luzonicum
%RPA
0.07
0.28
tr
3.35
0.14
0.60
8.01
0.04
0.17
1.44
59.40
tr
tr
0.28
0.26
0.12
0.29
0.52
0.06
0.04
0.02
tr
0.06
tr
0.02
0.02
0.02

Advances in Flavours and Frugrances

96
1184
1190
1197
1204
1210
1214
1226
1231
1238
1250
1257
1264
1295
1305
1311
1345
1347
1359
1359
1385
1403
1407
1431
1446
1472
1499
1536
1561
1564
1610
1666
1668
1670
1688

terpinen-4-01
para-cymen-8-01
alpha-terpineol
gamma-terpineol
cis-sabinol
trans-piperitol
cis-carve01
nerol
trans-carve01
carvone
linalyl acetate
piperitone
thymol
carvacrol
terpinene hydrate
delta-elemene
trans-piperitol acetate
terpinyl acetate
alpha-cubebene
alpha-copaene
beta-copaene
methyl eugenol
caryophyllene
trans-alpha bergamotene
humulene
germacrene D
delta-cadinene
elemicin
elemol
guaiol
gamma eudesmol
beta-eudesmol
valerianol
bulnesol
Total identified

0.54
0.6 1
6.5 1
0.0 1
0.10
0.05

tr
0.08
0.16
0.14
tr
0.5 1

0.24
0.07
1.78
0.19
0.02
0.07
0.05
0.06
0.06

0.05
tr

0.02
0.10
0.45
0.09
tr

0.02
0.12
0.12
0.26
0.2 1

0.23
0.1 1

0.18
0.24

0.1 1
0.17
4.00
15.00
0.24
0.17
0.35
0.24
0.06

77.8

99.0

%RPA data from integrated FID chromatogram. Identities from GCMS. tr - trace <0.01%rpa
RRI on HP5 column using n-alkanes.

Table 1 Continued Composition of extract of C.madagascariense and C.luzonicum

4 CONCLUSIONS

Although C.madagascariense is in the same botanical family and genus as elemi

(Cluzonicum),it is clear from both the olfactory data and the analytical data that the odour
and composition of the resin from these two species is distinctly different. This is the first
detailed analysis of the resin from C.madagascariense and provides the information from
which a new perfumery accord has been created that replicates the fragrance of this resin
from the rainforest of Madagascar.

Natural Products and Essential Oils

97

2.785

3.932

alpha pinene

p6m
7.137

7.458

alpha terpineol

%
%

11%

7.679

9.125

L;iIg2

I-

10.347

10.514

10 930

11.251
11.412
11.690

12.277

11.153
11.248

11.708
11.806

12.579

12.868

13.478

i
,
,
,
,
Figure 2 Comparison of extracts of C.luzonicum and C.madagascariense

98

Advances in Flavours and Fragrances

The discovery of this fragrant resin and the many other scented flowers, leaves, and barks
that were found in the forests of Madagascar during our expedition illustrate the
importance of such ventures in the discovery and creation of new fragrances. Only by
travelling to regions rich in unusual flora with a team of fragrance and botanical experts
can one successfully discover such a variety of new scents. Through the identification of
the plants and the analysis of the extracts, we can discover the secrets of the natural
chemistry of these fragrant materials. With advances in modern technology and the skills
of the creative perfumer it is possible to bring these scents from the rainforest to the
consumer without the need to remove any plants from their native habitat.

References
1

3
4

S. Arctander Perfume and Flavour materials of Natural Origin, Arctander, 1960 p221
K. Bauer, D. Garbe and H. Surburg, Common Fragrance and Flavour materials, 3'd
Edn, Wiley-VCH, 1997, p187
R.P. Adams, Identification of essential oil components by Gas Chromatography /
Mass Spectroscopy, Allured, 1995
B.M. Lawrence, The composition of Elemi oil, P e r f & FZav,1980, 5 (l)Feb/March, 57.

THE EFFECT OF MICROGRAVITY ON THE FRAGRANCE OF A MINIATURE

ROSE, OVERNIGHT SCENTSATION ON SPACE SHUTTLE (STS-95)

Braja D. Mookherjee, Subha Patel, Weijia Zhou2

Research and Development, International Flavors and Fragrances, Inc., 1515 State
Highway 36, Union Beach, New Jersey, 07735
2Wisconsin Center for Space Automation and Robotics, University of Wisconsin, 2346
Engineering Hall, 1415 Engineering Dr., Madison, Wisconsin, 53706

1 INTRODUCTION
Everybody knows that gravity governs the universe. It is due to gravity that we stand,
fruits fall and leaves flutter to the ground. NASA has sent man to the moon and into space
for the last 30 years to study the effect of micro gravity on the physiology of astronauts.
During previous shuttle missions, seeds and tubers were grown in space, but to our
knowledge, no one has ever sent a living flowering plant into space to study the effect of
microgravity on the production of the aroma constituents of a flower.

Figure 1 Oak tree

The plant, (Figure I), is the only living entity that grows against the effects of gravity,
a phenomenon known as gravitropism. With this in mind, it would be very interesting to
know how a flowering plant responds to microgravity. NASA provided us with the
opportunity to study the effects of microgravity on the aroma of a flower on STS-95 Space
Shuttle Mission.

100

Advances in Flavours and Fragrances

Figure 2 Astroculture unit

2 EXPERIMENTAL

To send a living plant into conditions of microgravity, one needs to design a proper
chamber, which could hold the plant in space and provide all of its biological requirements.
This chamber, known as the ASTROCULTURETMunit, (Figure 2), was developed and
built by Dr. Zhou and his associates of WCSAR. Despite the large appearance of this unit,
the plant-growing chamber is only 5 inches by 8 inches in dimension.
This chamber is equipped with LED lighting for photosynthesis, a substrate medium
(arcelite), a reservoir containing nutrient solution, as well as humidity and temperature
controls. This unit provides all of the earthly growth requirements of a living plant. Once
the chamber was ready, much careful consideration was given to the choice of flowering
plant to be sent into space.
To send a flower to space, the following criteria had to be met:
1. Since we are a Flavour and Fragrance Company, the plant must have a flower
with an aesthetically pleasant and suitable aroma for making the perfume.
2. The plant has to fit in the growing chamber which is 5 x 8.
3. The flowers life span has to be 10 days from bud to bloom.
4. The plant must bloom during October-November, 1998.
So, what was that unique flower which would fulfil all of these requirements?

Figure 3 Hybrid tea-roses in garden

101

Natural Products and Essential Oils

After much thought, a rose plant was selected to be sent into space because of the
following reasons:
1. The rose is the queen of all flowers. Every person on this earth adores the rose.
2. It is the national flower of the United States of America, established by
President Ronald Reagan in 1986.
Being an American involved in research on an American space flight, what could be
better than to send Americas national flower, the rose, into space? However, traditional
rose plants are at least 2 to 4 feet high and cannot be accommodated in the plant chamber,
(Figure 3). For this reason, we looked for another kind of rose i.e. a miniature rose.

Figure 4 Miniature roses in IFF Greenhouse

But another problem came to our hand. Most miniature roses do not have any aroma.
Therefore, we searched for a miniature rose with a pleasant aroma. After screening many
miniature roses, we found a hybrid miniature rose, known as Overnight Scentsation,
(Figure 4),not only because of its beauty but also for its sensational aroma. This rose has
an intense green rosy aroma, which is not found in regular roses. Now we had to grow this
rose so that it blooms in the cold months of October/November.

Figure 5 IFF Greenhouse

102

Advances in Flavours and Fragrances

Figure 5 shows the special IFF greenhouse where light, humidity and temperature are
controlled by computers. We have grown this miniature rose in this special IFF
greenhouse, where they bloom throughout the year.
Now that the ASTROCULTURETMunit has been built, the rose has been selected and
grown in the greenhouse, only one more important aspect of the project remains to be
solved, that is how to analyse the aroma of the living rose in space.

Figure 6 SPME Living Flower Sampling Technology

To do that, another unique IFF technology was used, Living Flower technology. Let
me describe this very simple but unique method. In 1992-1993, we developed a simple
technique to analyse a living flower without picking, touching or harming the flower in any
way. A device known as an SPME needle containing a glass fibre coated with liquid
absorbent is placed above the flower petals without touching the flower. Depending upon
the intensity of the aroma of the flower, the needle is kept there for 0.5 - 2 hours. The
molecules surrounding the flower are absorbed onto the needle and introduced into an
instrument called a gas chromatogram - mass spectrometer for subsequent analysis (Figure
6 )Now, we had to adopt this sampling method in the ASTROCULTURETMunit.

Figure 7 SPME sampling unit

Natural Products and Essential Oils

103

A precise metal holder as shown in this picture, containing eight SPME needles was
developed so that the rose could be analysed over four consecutive days. This sampling
device was fitted within the ASTROCULTURETMunit in such a way that when the rose
bloomed the SPME needle would extend above the rose (Figure 7).
After devising this analytical method, we performed preliminary experiments with the
miniature rose in this AstrocultureTMunit on earth.

Figure 8 Ground Simulation experiment

This is the Ground Simulation experiment where the rose was analysed from bud to
full bloom over four consecutive days, (Figure 8). This method of analysis was taught to
Astronaut Mr. John Glen, who actually performed this experiment in space. Now lets look
at the comparative analysis of the living rose (Overnight Scentsation) in the green house
and in the AstrocultureTMunit.

Figure 9 Comparative Analysis of living rose (Overnight Scentsation) in NORMAL SOIL

& ASTROCULTURE UNIT (average of 3 days)

We found that the total amounts of the volatile contents produced by the flower in the
AstrocultureTMunit have decreased by a factor of 3 when compared to the total amounts of
the volatile contents produced by the flower in the green house, (Fig. 9).
Figure 10 shows the differences in the constituents of the rose grown in the
greenhouse and in the AstrocultureTMunit.

104

Advances in Flavours and Fragrances

Figure 10 Comparative Analysis of living rose (Overnight Scentsation) in NORMAL SOIL

& ASTROCULTURE TM UNIT (average 3 days)
This shows that the flower's aroma has changed in this AstrocultureTMunit. We may
ask why? Answer is simple. We have changed the growing conditions of the flower,
particularly the environmental atmosphere. We are of opinion that like a human being,
when the flower is placed in a closed chamber, it is under stress and the physiology of the
plant has probably changed, hence the aroma is also changing. Please remember that we
have to compare this ground simulated AstrocultureTM unit experiment with the space
experiment.

Figure 11 Rooting tray with rose plant and 2 buds

Everything is ready for the space experiment. Now we are waiting for the final launch
day on October 29'h 1998. This rose plant with 2 rose buds was placed in the plant-growing
chamber of the ASTROCULTURETMunit, and the SPME needle box was positioned in
such a way that as the rose bloomed, the SPME needle could be positioned above the rose
petals. As you can see, the stem of the rosebud was tied with a thread on a metal pole so
that the rosebud does not bend or fall due to the tremendous thrust of the launch. By the
way, this is the only work I have done, (Figure 11).

Natural Products and Essential Oils

105

Figure 12 Shuttle launch

Finally, the much awaited launch day came! On the 29'h of October, 1998, this shuttle
containing the rose plant with two rose buds was launched, (Figure 12). Unless you
personally see the launch, you probably cannot imagine the tremendous thrust the shuttle
produced at the moment of launching. I literally prayed to God to keep the rose bud intake
on the plant! And God granted my prayers!

Figure 13 Rose is blooming-video picture

This is the actual video picture of the rose after 2 days in space. As you can see in the
picture, the rose has bloomed in space and was analysed by John Glenn and other
astronauts for four consecutive days by SPME needles as seen here, (Figure 13).

Advances in Flavours and Fragrances

106

Figure 14 Shuttle Module returning from space

After nine days in microgravity, the rose came back to the earth in this module (Figure
14).

Figure 15 Astroculture

TM

unit

Figure 15 shows laboratory assistants opening the box to take out the rose plant after 9
days in space. You can probably imagine the intense tension of our minds at this moment!

Figure 16 Flowers when returned from space

Natural Products and Essential Oils

107

The box is opened, and the rose plant is alive. Figure 16 is a picture of the rose as it
came from space. You can see that the flowers are fully open.

Figure 17 Close-up of the flowers from space

Figure 17 is a close-up picture of blossoms as grown in space. Careful examination
revealed that more new shoots and leaves had been produced. The flower also smelled
good.
All eight SPME needles were analysed by GC-MS. The Space data was compared
with our ground simulation experiment. Now let us see the results.

Figure 18 GC/MS profile of space-rose and earth-rose

108

Advances in Flavours and Fragrances

Looking at figure 18, the top is the space rose - GCMS profile and the bottom is that
of the ground simulated experiment in earth. One can easily see the differences in the
constituents of the Space rose verses Earth rose.
Let us now see the quantitative differences in the constituents of both, space and earth
rose.

Figure 19 Total volatiles of ground experiment and space experiment

The very first thing we found was that the total amount of volatiles produced by the
flower in space has been significantly reduced when compared to the amount of volatiles
produced by the flower in the AstrocultureTMunit on earth, (Figure 19). Although it has not
been scientifically determined why the Space rose produces less aroma, one could
speculate that based upon the increased growth of the foliage, that more nutrients were
available to the lower branches than to the top of the plant where the flower grows. Or
under microgravity - just like human beings, the plant was under stress. Human beings
loose muscle, the flower lost the aroma! Now let us see the differences in constituents of
the both flowers.

Figure 20 Individual components

Natural Products und Essential Oils

109

Figure 20 shows the changes in the constituents of both roses. The green note
constituents such as cis-3-hexenyl acetate, hexyl acetate and rose oxide have been
drastically reduced. The heart of the rose constituents' viz. phenyl ethyl alcohol, citronellol,
geraniol, and methyl geranate have been increased. At the same time, non-polar
hydrocarbons also increased.

Figure 21 Results of space flower experiment

Figure 21 summarises our findings. From these results, I would like to predict that
plants probably would grow more in other terrestrial places like Mars, where the gravity is
less than that on earth, as long as proper plant growth conditions are met. The flower will
also provide an aesthetic aroma but this will be different that experienced on earth.
This experiment also poses a very important question to my mind. Since the plant
grows more and produces a different but more pleasant aroma under micro gravity, there
must be another force that governs the life of the plant. Could it be God or Gravity?

I would like to leave this question to all the distinguished

Natural Product Chemists to solve this natural mystery.
Although we may not solve the mystery, it brought happiness in
the minds of people around the world. Last September,
Shiseido, the famous company of Japan introduced a new
fragrance, utilising the Space Rose fragrance, called ZEN. I am
sure you know that ZEN means, ultimate Bliss, (Figure 22).

Figure 22 Shiseido 's Zen fragrance

Acknowledgements
We would like to thank all NASA officials for providing this opportunity to work with
WCSAR. In this respect, we would also like to thank Mr. Robert Meyers, Mr. Sam Durst
and Mr. Perry Sandstrom at WCSAR who assisted in building the Astorculture Unitm.
Last but not least, we would like to thank our colleague, Mr. Robert Trenkle for devising
the original method of SPME.

Organic and Bioorganic Chemistry

AMBERGRIS FRAGRANCE COMPOUNDS FROM LABDANOLIC ACID AND

LARIXOL

Aede de Groot
Laboratory of Organic Chemistry, Wageningen University, Dreijenplein 8, 6703 HB
Wageningen, The Netherlands

1 INTRODUCTION
Labdanes form a large group of diterpenes, consisting of a substituted decalin system, with
a substituted side chain at C9'. The Cg side chain may contain various functional groups as
can be seen from labdanolic acid (1) and larixol (2), but often a tertiary methyl-vinylalcohol functionality at C13 as in larixol is encountered.

labdanolic acid (1)

larixol (2)

Arnbrox@

Arnbrox@derivatives Ambroxenes

A6-Ambra-oxides

Figure 1
Labdane diterpenoids are available from nature and have been used frequently as
starting material for the synthesis of other natural products. Most of the chemistry of
labdanes has been developed for the preparation of Ambrox@ and Ambrox@-like
compounds (see figure 1)2.
Since ancient times, ambergris has been one of the most highly valued perfumery
materials3. Ambergris is a metabolic product of the sperm whale (Physeter macrocephalus
L.), which accumulates as concretions in the gut. (-)-Ambrox@(lo),the commercially most
important constituent of natural ambergris, was recognised as the prototype of all
ambergris odourants4. For this reason, various synthetic routes to (-)-Ambrox@ and its
racemate have been developed5. The price of (-)-Ambrox@ is relatively high, which
induces a continuing search for new syntheses6 of this compound and its derivatives,
preferably starting from cheap, abundantly available natural labdanes. Sclareol is
nowadays the industrially used starting material for the preparation of (-)-AmbroxB7.
Labdanolic acid (1) and larixol (2) are both available from labdanum gum7 and larix

Advances in Flavours and Fragrances

114

terpentine respectively, but have found little use as starting material in syntheses of
Ambergris odour compounds. In our laboratory, we have investigated possible applications
of these two labdanes.

2 AMBROX~

2.1 Ambrox@from labdanolic acid

Labdanolic acid (1) is the main component (ca. 40%) in the acidic fraction of the n-hexane
The oxidative degradation of the Cg sideextract of Cistus Zudaniferus L. (Ro~k-rose)~.
chain of labdanolic acid can in principle provide suitable synthons for the synthesis of (-)Ambrox@(10). The degradation of labdanolic acid or its methyl ester is not an easy task
because the carboxyl group is the only available functional group in the side chain. Lead
tetraacetate is the most commonly used decarboxylating reagent and moderate yields of the
acetate were obtained. We have investigated the iododecarboxylation of labdanolic acid
as the most significant step in its conversion into AmbroxO (Scheme 1).

Labdanolic acid

Scheme 1
The commercial extract from Cistus Zuduniferus L. was obtained by steamdistillation of the twigs and leaves of the plant, or by treatment of the plant with hot neutral
or alkaline water. However, these conditions easily caused dehydratation of the tertiary
hydroxyl group, which resulted in a mixture of labdenic acids. To prevent this dehydration
the dried twigs and leaves of the Cistus Zuduniferus L. were soaked with n-hexane. The
hexane solution was separated from the solid material and after evaporation of the solvent
a sticky labdanum gum was obtained. The gum was dissolved in ether and extracted with
NaOH solution. The basic solution was acidified, extracted, and after concentration a crude
labdanolic acid was obtained. The purification of this crude labdanolic acid appeared to be
difficult, but after conversion of the c8 tertiary alcohol group into its acetate 3, the
purification of the acetate proved to be relatively easy. The iododecarboxylation of 3 was
achieved by irradiation with a lOOW tungsten lamp and the iodide 4 was obtained in a
good yield of 76%.

fl 6 fl

Organic and Bioorganic Chemistry

OH

PHcc14,
IBDA,12
l00W

L
AcCI, DMA

115

76%

35-45 m%

KOtBu
THF
74%

1. Q
2. LiAlH 4
60%
Ambrop

Scheme 2
Because iodide 4 is not a stable compound it was treated directly with KOtBu in THF
at ruum temperature to cause dehydroiodination and hydrolysis of the acetate at Cg which
resulted in compound 5 in a yield of 74%. Ozonolysis of the double bond in 5 and
reduction of the intermediate ozonides gave the methyl ketone 6, which immediately
cyclised to sclareol oxide 7. Sclareol oxide was ozonolysed and reduced directly to afford
diol 9 in 60% overall yield from 5. Stirring of diol 9 in nitromethane in the presence of p toluenesulfonic acid gave Ambrox@(10) in 87% yield (Scheme 2).
An obvious way to circumvent the unstable sclareol oxide was the synthesis of 9 via a
Criegee rearrangement13 of the ozonide of 5, but this approach was unsuccessful since the
methyl ketone 6 was obtained as the result of a normal ozonolysis of 5. To prevent the
interference of the alcohol at Cg during the ozonolysis, the group was again protected as its
acetate and now the ozonolysis and Criegee rearrangement worked well (Scheme 3). After
reduction of the intermediate with LiAlH4, diol 9 was obtained in 89% overall yield from
compound 11. This diol was transformed into (-)-Ambrox@ as described in Scheme 2.
However, the shortest route was found when the dehydrohalogenation of iodide 4 is
performed with KOtBu in refluxing THF, or in DMSO at room temperature. The
intermediate alkene 5 isomerised in situ to 12 (Scheme 3), and ozonolysis of the double
bond followed by reduction of the intermediate ozonides with sodium borohydride gave
diol9 in 92% overall yield from 12.

KOtBu, THF

reflux
78%

@?' l:

92%
$aBb

12

Scheme 3

10

116

Advances in Flavours and Fragrances

In this way it was demonstrated that Ambrox@can be obtained in a short procedure of

4 steps in 47% overall yield starting from the acetate of labdanolic acid. The main
drawback of the method is the necessary application of iodine in the decarboxylation
reaction. When a cheaper procedure can be found for this reaction, labdanolic acid (1) will
become a good alternative starting material for the industrial preparation of Ambrox@.

2.2 Ambrox@-likecompounds from larixol

For the bicyclic part of the labdanes to be of use in the synthesis of other natural products
the Cg side chain has to be shortened. Several studies to oxidise the side chain have been
reported in the literature for larixol (2) and epimanool (14)14,but many do not use cheap
oxidants, are irreproducible or do not contain an easy work-u . Also, some special
oxidation procedures, which have been developed for sclareol (18)' , do not proceed in the
same way for larixol due to the absence of the hydroxyl group at c8 in the latter This c8
hydroxyl group participates in intermediates, which are formed during the oxidation of
sclareol, and such participation is not possible in epimanool (14), larixol (2), and its acetate
13. Later, the presence of the secondary hydroxyl group at Cg in the latter two compounds
and the double bond at CSmay cause selectivity problems. A modification of the method of
Ogino et d6,
using anhydrous permanganate in the presence of a phase-transfer catalyst,
was investigated for the oxidation of the labdanes 2 and 13, 14, 18-20, and 27 that are
depicted in Scheme 414f.

ds"c

3e&M
C
;&y 2

@; qJjY
+

trio1

1, R = OH; larixol
13, R = OAc; larixyl acetate
14, R = H; epi-manool

rnethylketone

yield of methyl ketones

15, R = O H
16, R = O A c
17, R = H

68%
72%

67%

3 eq. KMnO, , TEBAC

rnethylketone

18, R = H; sclareol
19, R = OH
20, R = OAC

21, R = H
22, R = OH
23, R = OAC

enol ether

24,R=H
25, R = OH
26, R = OAC

3 eq. KMnO, , TEBAC

CHpC17, 0-3 'C

56%
28

27

3 eq. KMnO, , TEBAC

CH2C17,rl. ) ) )

68%

Scheme 4

yield of en01 ethers

29

86%
74%
90%

117

Organic and Bioorganic Chemistry

The oleoresin of the Venice larch terpentin of the Larix decidua Miller (L. eurupaea
D. C . ) consists mainly of larixol, epi-manool and larixyl acetate 1317. To facilitate
purification, the extract was hydrolysed to give larixol as the main constituent, which could
be obtained in pure form via crystallisation from cyclohexane. The most apolar compound
in the residual mother liquor is epi-manool", which could be isolated by chromatography.
Compound 19 was obtained from larixol in 64% overall yield, after epoxidation of the
exocyclic double bond with oxone and reductive opening of the epoxide using LiAlH4.
Compound 279c was synthesised from larixol in 73% overall yield by oxidation with
pyridinium chlorochromate (PCC) to the C6-ketone, followed by base catalysed
isomerisation of the exocyclic double bond using methanolic sodium methoxide.
The oxidation of the labdanes 2 and 13, 14, 18-20 and 27 was investigated using
standard methods with 1.5 or 3 equivalents of permanganate, and the results of the latter
method are summarised in Scheme 4. From these results it became clear that generally
good to high yields of single products could be obtained. In addition, the influence of the
substituent at CS on the product formation became evident. When an exocyclic double
bond was present at Cg a reasonable selective oxidation of the double bond in the side
chain could be achieved to give the methyl ketones in good yields. When a hydroxyl group
was present at C S ,this group had a strong tendency to react with the methyl ketone in the
side chain, and cyclic enol ethers were isolated as the main reaction products in high yield.
These enol ethers were usually not very stable and with many oxidation systems further
oxidation took place to afford compounds with a shortened side chain at Cg. The
application of sonication accelerates the oxidation appreciably, and shortened the reaction
time from 14 to 2 hours, but the products and the yields were not affected. The ratio
between triols and methyl ketones was pH dependent, with alkaline pH favouring the
triols.
The oxidation of the CS side-chain of larixol or its acetate provided suitable synthons
for the synthesis of Ambrox@-likecompounds. 6 3-Hydroxy Ambrox can be considered to
be a key intermediate in such syntheses and the conversion of larixol into this intermediate
has been investigated first as is indicated in scheme 5.

2R=OH
13 R = OAC
30 R = OTBDMS

15R=OH
16 R = OAC

31, R = OTBDMS

36

6a-hydroxy-Arnbrc%
40

R = OAC
R = OMe

Scheme 5
An obvious way to achieve further breakdown of the obtained methyl ketones to a
functionalised two carbon side chain is the Baeyer Villiger oxidation. This reaction was
investigated
for
three
derivatives,
containing
an
hydroxyl
group,
a

118

Advances in Flavours and Fragrances

tert.butyldimethylsilyloxy group or an acetate at Cg (Scheme 5). It is known that methyl

ketones give acetates upon Baeyer Villiger oxidation, but under these conditions double
bonds may be epoxidized as well. Reduction of the acetate and the epoxide moiety should
be possible in one step to give the triol 36, which may be cyclised to 6a-hydroxy Ambrox@
40.
When methyl ketone 15 was treated with rn-CPBA in dichloromethane at room
temperature the expected epoxy-acetate 34 was formed along with the acetal 32 in an
almost equal yield (Scheme 6). Apparently, epoxidation of the double bond is faster than
the Baeyer Villiger oxidation of the methyl ketone and a subsequent reaction of the
epoxide with the carbonyl group gives the acetal 32. A similar mixture of compounds 33
and 35 was obtained in the Baeyer Villiger oxidation of the TBDMS protected compound
31 with rn-CPBA. On the other hand, when the acetate 16 was treated with m-CPBA the
Baeyer Villiger product 39 was obtained in a good yield of 84% (Scheme 6).

15, R = OH

31, R = OTBDMS

32,R = OH
33, R = OTBDMS

WR=OH
35 R = OTBDMS

36,R = O H
LiAIHp
62%

6a-hydroxy-Arnbrc%

16

39

40

Scheme 6
The epoxy-acetates 34 and 39 were treated with LiAlH4 to reduce the epoxide to the
desired tertiary hydroxyl group at C8, and also simultaneously reduce the acetates.
However, the reduction of these two epoxy-acetates gave the cyclic ether 38 as the main
product with the desired triol 36 being obtained only as a minor component. Obviously the
acetates are reduced first, and then the alkoxide in the side-chain attacks the epoxide to
form the cyclic ether 38. In contrast the reduction of compound 35 with LiAlH4 proceeded
well and gave a 62% yield of diol 37. The triol 36 could be cyclised in 64% yield to
6a-hydroxy-Ambrox@ 40 after stirring in nitromethane in the presence of p toluenesulfonic acid. Although incidental transformations in this sequence proceeded in
good yield, no overall high yield conversion of the methyl ketones to 6a-hydroxy
Ambrox@could be achieved and therefore other routes were investigated.

Organic and Bioorganic Chemistry

119

oxone

LiAIH4

84%

Larixol

94%

41

42R=OH
43, R = OAC

43

6u-hydroxy Sclareol

42

4R=OH
45,R = OAC

36

Scheme 7
The two major difficulties in the former route were connected with the lack of
selectivity in the Baeyer Villiger oxidation and in the reduction of the resulting epoxyacetates. To avoid these problems, the exocyclic double bond at CP, in larixol was first
converted into a hydroxy-methyl group by oxidation with oxone followed by reduction of
the epoxide 41 with LiAlH4 to give 6a-hydroxy-sclareol 42 (scheme 7). In this case no
competitive formation of cyclic ethers was observed. The C6-hydroxyl group in 42 could
be protected selectively as its mono acetate 43, and both compounds were oxidised with a
catalytic amount of OsO4 and an excess of NaI04 to afford the aldehydes 44 and 45
respectively, as the major products in high yield. When the aldehydes were reduced with
LiAlH4 triol 36 was obtained in 94% yield.
Two other routes starting from 6a-hydroxy-sclareol 42, based on the palladium
catalysed elimination or isomerisation of acetates in the side chain, followed by ozonolysis,
were also investigated (scheme 8). Protection of the secondary and the tertiary hydroxyl
groups in 42 as acetate groups could be achieved with acetyl chloride in N,Ndimethylaniline to give triacetate 46 in high yield. When 46 was treated with a catalytic
amount of palladium acetate in the presence of triphenylphosphine at 100 OC, a mixture of
three dienic acetates 47, 48 and 49 in a ratio of 2:1:2 was obtained respectively in over
90% yield. Ozonolysis of this mixture gave an aldehyde, which was reduced and
saponified to give triol 36 in a good yield.

47

46

60-70%

96%

from the mixtur r:

50

49

48

PdC12

:$&

52

51

Scheme 8

36

Advances in Flavours and Fragrances

120

Triacetate 46 could also be isomerised in nearly quantitative yield to the allylic acetate
5019 (Scheme 8). Ozonolysis of the double bond in the side chain gave methyl ketone 51,
which was transformed into triacetate 52 by Baeyer Villiger oxidation, again in a good
83% yield. Complete reduction of 52 with LiAlH4 gave triol 36 in 85% yield. The same
triol can also be obtained from triacetate 52 by saponification.
When the results, obtained in the conversion of larixol into 6a-hydroxy-Ambrox@are
compared, it can be concluded that the route described in Scheme 7 gives the highest yield
in the shortest number of steps. On the other hand, the second route of scheme 8 is the
easiest one to perform and has our preference.

2.3 Other Ambrox* derivatives

6a-Hydroxy Ambrox@40 was used as the key intermediate for the preparation of a number
of simple Ambrox@derivatives. Treatment of 40 with Jones reagent in acetone produced
6-0xo-Ambrox~53 in 9 1% yield. Conversion of 40 to its acetate 54 was achieved by acetic
anhydride and DMAP in pyridine and dichloromethane in 94% yield. The methyl ether 55
was obtained in 91% yield after treatment of 40 with sodium hydride and methyl iodide.
When 40 was treated with p-toluenesulfonic acid in benzene under Dean Stark conditions a
non-separable mixture of three compounds 56, 57, and 58 was formed in a ratio of 2: 1: 1.
The structures of 56 and 57 were confirmed by independent synthesis (vide infra), the
structure of compound 58 is still tentative. This pleasant smelling mixture was separated on
a GC-MS apparatus and on a GC-sniff apparatus, and all three compounds showed a
pleasant smell, with the A5-alkene 56 as the most interesting one.

@SOH,benzene
Dean Stark

80%
ratio2: 1 : 1
55

59

71%

56

57

58

Scheme 9
This A5-alkene 56 was produced in a more selective way by conversion of 40 to
mesylate 59 and treatment of this mesylate with magnesium iodide in toluene or by
treatment with lithium bromide and lithium carbonate in NJV-dimethylformamide (DMF)
at 120C. This afforded the A5-alkene 56 in about 70% yield, with preservation of its
pleasant smell.

121

Organic and Bioorganic Chemistry

A selective synthesis of A6-Ambroxene 57 proved to be more difficult and laborious.

Several approaches were investigated, and ultimately the one via ring closure of allylic
alcohol 61 proved to give the best results (Scheme 10). Three reaction sequences were
investigated for the conversion of larixol into diketone 29'Odand the best one was to first
oxidise the side chain in larixol with KMnO4 to methyl ketone 16 (see Scheme 4), followed
6 and isomerisation of the double bond. In this way
by oxidation of the hydroxyl group at c
diketone 29 was obtained in 56% overall yield. However, in our hands the Bayer-Villiger
reaction of 29 to 60 could be accomplished only in a moderate 30% yield, although several
attempts were undertaken to improve this reaction*'. The carbonyl and acetate group in 60
could be reduced in moderate yield in one operation using LiAlH4. The cyclisation of the
obtained diol could be achieved after acidification of the reaction mixture to give the A6ambroxene 57 with the correct configuration at Cs, and this result encouraged us to look
for a better synthesis of the intermediate diol61.

Larixol
2

29

61

60

57

Scheme 10
The oxidative breakdown of the side chain to an ethylene hydroxy moiety can be
accomplished in several ways as was described in schemes 7 and 8. Selective protection of
the primary hydroxyl group in the side chain of trio1 36, followed by oxidation of the
secondary hydroxyl group at cg and elimination of the tertiary hydroxyl group at CS, gave
the enone 63 in 67% overall yield2' (scheme 11).

36

#yo
0

66

6-0x0-sclareolide

@
0

53

@J0:y.
0

65

6-oxo-Arnbr&

@
57
A'-Ambroxene

Scheme 11
A selective synthesis of A6-Ambroxene 57 from enone 63 now proved to be possible
in three simple steps. Reduction of the carbonyl group at Cg with DIBAH in toluene at low
temperature, deprotection of the hydroxyl group in the side chain of 64 and acid catalysed
cyclisation afforded A6-Ambroxene 57 in high yield. The first two reactions could also be

122

Advances in Flavours and Fragrances

carried out in the reversed sequence, and the second two could be carried out in one pot.
Oxidation of trio1 36 with PCC led in one step to 6-0x0 sclareolide 66. Deprotection of the
hydroxyl group in 63 led to cyclisation to 6 0 x 0 Ambrox@5322(seeScheme 11).
The approach described in Scheme 11 seemed suitable for the selective synthesis of
other A6 Ambrox@-like compounds as wellge. The allylic hydroxyl group at c6 can be
abstracted easily and the resulting mesomeric carbocation can be intercepted by the
hydroxyl group in the side chain. The enones 27 and 29,obtainable in a few reaction steps
from larixol, are good intermediates to try out the possibilities of this approach for the
synthesis of several
substituted A6-tricyclic tetrahydropyranyl ethers (A6-Ambraoxides) (Schemes 12-14).
The enone 27 could be synthesised easily from larixol by oxidation of the hydroxyl
group at c6, followed by base catalysed isomerisation of the exocyclic double bond to the
conjugate endocyclic position'" (Scheme 12). Reduction of the carbonyl group at c6 often
requires protection of the hydroxyl group in the side chain, this was the case in the
reduction of 27. On the other hand, the acid catalysed deprotection of the hydroxyl group
in the side chain often proceeded along with cyclisation as was demonstrated in the
synthesis of compound 68. Deprotection of the highly hindered TBDMS group in 67 with
TBAF could not be accomplished.
PBDM S

@
I

2)
l)TBDMSCI,92%
DIBAH
toluene, -7(F'C,

27

29

@\

HF, CH$N

'

OH

67

98%

68

70

69

80%

71

Scheme 12
Oxidation of the side chain in 27,gave diketone 29 (see also Scheme 10) in which the
non-conjugated carbonyl group in the side chain could be manipulated selectively.
Addition of methyl lithium to 29, protection of the hydroxyl group in the side chain, and
reduction of the carbonyl group at c6 afforded the allylic alcohol 70.However, when this
compound was subjected to the combined deprotection cyclisation reaction, the dienealcohol 71 was isolated as the only product and no cyclisation was observed.
Similar behaviour was observed with the C 13-monomethyl analogues (Scheme 13).
Selective reduction of the carbonyl group in the side chain of 29 gave an unseparable
mixture of the C 13- mono methyl compounds 72.A similar reaction sequence consisting of
protection of the side chain, reduction of the C6-carbonyl group, and deprotection provided
again a mixture of C13-mono-methyl diene-alcohols 74a and 74b. The reduction of both
carbonyl groups in diketone 29 could be carried out simultaneously using DIBAH, and
now the intermediate diols 75 and 76 could be separated by chromatography. Aqueous acid
treatment of the separate isomers did give some of the cyclised product 77 and 78

123

Organic and Bioorganic Chemistry

respectively, but the elimination reaction was the predominant

the dienes were unsuccessful.

Attempts to cyclise

4:1

Lb$t

aq,HCI,
B20

pfjYH

+ :@

Ho

76

30: 1

74b

78

Scheme 13
Finally, the cyclisation of the C13 unsubstituted analogue was investigated (Scheme
14). For its preparation, the side chain of larixol had to be shortened to methyl ketone 15,
followed by an iodoform reaction to yield the acid 79.Reduction of the corresponding ester
and selective protection of the primary hydroxyl group in the side chain afforded 80.

fl

OC H3

Larixol
2

-0TB

d$

1) KI, 12, KOH, 56%

KMn04b

1) L i A I b , 93%

2) TBDMSCI, 94%

@cH*N2,98%

79

15

DMS

@OmDMS

DIBAH, 87%

1) PCC, 82%

fl
-

1) HF, C H g N

2) SiO2, 88%

2) NaOMe, 78%

80

81

82

83

Scheme 14
Allylic alcohol 82 could be prepared in the usual way by oxidation of the hydroxyl
group at C g , isomerisation of the double bond, and reduction of the enone 81. Deprotection
and acid catalysed cyclisation of 82 again resulted in just dehydration to the diene-alcohol
83 with no cyclised product being observed23.

Advances in Flavours and Fragrances

124

So it turned out that the successful cyclisation of the A7, 6P-hydroxyl compound 61 to
a five membered cyclic ether, e.g. A6-Ambroxene 57, was not matched by an equally
successful cyclisation of the corresponding homologues to the six membered cyclic ethers,
e.g. the A6-Ambra-oxides. In all but one of the latter cases, elimination of the 6P-hydroxyl
group to the corresponding A6,A8-dienes was observed as the major reaction.

References and notes

(a) Conolly, J. D.; Hill, R. A. Dictionary of Terpenoids, Chapman & Hall: London,
1991, vol. I, pp. xxxii. (b) Conolly, J. D.; Hill, R. A. Dictionary of Terpenoids,
Chapman & Hall: London, 1991, vol. 11, pp. 677-736.
(a) Ruzicka, L.; Janol, M. M. Helv. Chim. Acta 1931, 14, 645-648. (b) Ruzicka, L.;
Engel, L. L.; Fischer, W.H. Helv. Chim. Actu 1938, 21, 364-370. (c) Ruzicka, L.;
Seidel, C. F.; Engel, L. L. Helv. Chim.Acta 1942, 25, 621-630. (d) Ohloff, G. Helv.
Chim. Acta 1958, 41, 845-850. (e) Corey, E. J.; Sauers, R. R. J. Am. Chem. SOC.
1959, 81, 1739-1742. (f) Vlad, P. F.; Koltsa, M. N.; Sibirtseva, V. E.; Kustova, S. D.
J. Gen. Chem. USSR (Engl. Transl.) 1980, 50, 162-171. (g) Vlad, P. F.; Ungur, N. D.
Synthesis 1983, 216-219. (h) Leite, M. A. F.; Sarragiotto, M. H.; Imamura, P. M.;
Marsaioli, A. J. J. Urg. Chem. 1986, 51, 5409-54 10. (i) Decorzant, R.; Vial, C.; Naf,
F. Tetrahedron 1987,43, 1871-1879. (i) Zahra, J.-P.; Chauvet, F.; Coste-Manikre, I.;
Martres, P.; Perfetti, P.; Waegell, B. Bull. SOC. Chim.Fr. 1997, 134, 1001-1024. (k)
Muller, M.; Schroder, J.; Magg, C.; Seifert, K. Tetrahedron Lett. 1998, 39, 46554656.
(a) Ohloff, G. The Fragrance of Ambergris, in Fragrance Chemistry; Theimer, E.
T. Ed.; Academic Press. New York, 1982, pp. 535-573. (b) Kraft, P.; Bajgrowicx, J.
A.; Denis, C.; Friter, G. Angew. Chem. Int. Ed. 2000,39, 2980-3010.
(a) Ohloff, G.; Giersch, W.; Pickenhagen, W.; Furrer, A.; Frei, B. Helv. Chim.Acta
1985, 68, 2022-2029. (b) Escher, S . ; Giersch, W.; Niclass, Y.; Bernardinelli, G.;
Ohloff, G. Helv. Chim. Acta 1990, 73, 1935-1947.
(a) Cambie, R. C.; Joblin, K. N.; Preston, A. F. Aust. J. Chem. 1971, 24, 583-591. (b)
Gonzilez-Sierra, M.; Rdveda, E. A.; Lbpez, J. T.; CortCs, M. J. Heterocycles 1987,
26, 2801-2804. (c) Koyoma, H.; Kaku, Y.; Ohno, M. Tetrahedron Lett. 1987, 28,
2863-2866. (d) Christenson, P. A. Tetrahedron 1988, 44, 1925-1932. (e) CosteManikre, I. C.; Zahra, J. P.; Waegell, B. Tetrahedron Lett. 1988, 29, 1017-1020. ( f )
Buchi, G.; Wiiest, H. Helv. Chim. Acta 1989, 72, 996-1000. (g) Sell, C. Chem. Ind.
1990, 516-520. (h) Mori, K.; Tamura, H. Liebigs Ann. Chem. 1990, 361-368. (i)
Snowden, R. L.; Linder, S. M. Tetrahedron Lett. 1991, 32, 41 19-4120. (j)Snowden,
R. L.; Eichenberger, J.-C.; Linder, S. M.; Sonnay, P.; Vial, C.; Schulte-Elte, K. H. J.
Org. Chem. 1992, 57, 955-960. (k) Snowden, R. L.; Eichenberger, J.-C.; Giersch,
W.; Thommen, W.; and Schulte-Elte, K. H. Helv. Chim. Acta 1993, 76, 1608-1618.
(1) Barrero, A. F.; Altarejos, J.; Alvarez-Manzaneda, E. J.; Ramos, J. M.; Salido, S.
Tetrahedron 1993,49, 6251-6262. (m) Martres, P.; Perfetti, P.; Zahra, J.-P.; Waegell,
B.; Giraudi, E.; Petrzilka, M. Tetrahedron Lett. 1993, 34, 629-632. Barrero, A.F;
Alvarez-Manzaneda, E.J.; Altarejos, J.; Salido, S.; Ramos, J.M. Tetrahedron 1993,
49, 10405-10412. (n) Kutney, J. P.; Cirera, C. Can. J. Chem. 1997, 75, 1 136-1150.
(a) Decorzant, R.; Vial, C.; Naf, F.; Whitesides, G. M. Tetrahedron 1987, 43, 18711879.
(b) Frhter, G.; Bajgrowicz, Jerzy A.; Kraft, P. Tetrahedron 1998,54,7633-7703.

Organic and Bioorganic Chemistry

8
9

10

11

12

13
14

15

16
17

125

(a) Cocker, J. D.; Halsall, T. G.; Bowers, A. J. Chem. SOC. 1956, 4259-4262. (b)
Cocker, J. D.; Halsall, T. G. J. Chem. SOC. 1956, 4262-4271. (c) Tabacik Wlotzka,
C.; Mousseron, M.; Chafal, A. Bull. SOC.Chim. Fr. 1963, 2299-2306. (d) Overton,
K.H.; Renfrew, A.J. J. Chem. SOC. 1967, 931-935. (e) de Pascual Teresa, J.;
Urones, J. G.; San Feliciano, A. An. Quim. 1972, 68 B, 727-732. (0 de Pascual
Teresa, J.; Urones, J. G.; Basabe, P.; Aubanell, F. H. An. Quim. 1979, 75, 335-340.
(g) Vlad, P. F.; Koltsa, M. N.; Mironow, G. N. Russian Chem. Bull. 1997, 46, 855873.
(a) Sandermann, W.; Bruns, K. Tetrahedron Lett. 1965, 42, 3757-3760. (b)
Sandermann, W.; Bruns, K. Chem. Ber. 1966,99,2835-2841.
(a) Beatrice M.F.Lagne1, Christophe Morin, Aede de Groot, Synthesis, 2000, 19071916.
(b) Marjon Bolster, Ben J.M. Jansen and Aede de Groot, Tetrahedron, 2001, 57,
5657-5662.
( c ) Marjon Bolster, Ben J.M. Jansen and Aede de Groot, Tetrahedron, 2001, 57,
5663-5679.
(d) Marjon Bolster, Beatrice M.F. Lagnel, Christophe Morin, Ben J.M. Jansen and
Aede de Groot, Tetrahedron, 2001, 57, 8369-8379. (e) Marjon Bolster, Ben J.M.
Jansen and Aede de Groot, Tetrahedron, 2001, submitted.
(a) de Pascual Teresa, J.; Urones, J. G.; Montafia Pedrero, A.; Basabe Barcala, P.
Tetrahedron Lett. 1985,26, 5717-5720. (b) Deluca, M. E.; Seldes, A. M.; Gros, E.
G. Helv. Chim. Acta 1986, 59, 1844-1850. (c) de Pascual Teresa, J.; Urones, J. G.;
Montaiia Pedrero, A.; Basabe Barcala, P. Tetrahedron 1986, 42, 4563-4576. (d)
Urones, J. G.; Basabe, P.; Marcos, I. S.; Gonzdez, J. L.; JimCnez, V.; Sexmero, M.
J., Lithgow, A. M. Tetrahedron 1992, 48, 9991-9998. (e) Lithgow, A. M.; Marcos,
I. S.; Basabe, P.; Sexmero, M. J.; Diez, D.; Gbmez, A.; Estrella, A.; Urones, J. G.
Natural Product Lett. 1995,6, 29 1-294.
The commercial extract from Cistus ladaniferus L. was a gift of Destilaciones
Bordas Chinchurreta, S.A., Sevilla (Spain).
Concepcibn, J. I.; Francisco, C. G.; Freire, R.; Hernandez, R.; Salazar, J. A.;
Suirez, E. J. Org. Chem. 1986,51,402-404.
Schreiber, S. L.; Liew, W.-F. Tetrahedron Lett. 1983,24,2363-2366.
(a) Hosking, J. R. Chem. Ber. 1936, 69, 780-785. (b) Schenk, H. R.; Gutmann, H.;
Jeger, 0.; Ruzicka, L. Helv. Chim. Actu 1952, 35, 817-821. (c) Schenk, H. R.;
Gutmann, H.; Jeger, 0.; Ruzicka, L. Helv. Chim. Actu 1954, 37, 543-546. (d)
Sandermann, W.; Bruns, K. Chem. Ber. 1966, 99, 2835-2841. (e) Grant, P. K.;
Weavers, R.T. Tetrahedron 1974, 30, 2385-2395. (f) de C6u Costa, M.; Tavares,
R.; Motherwell, W. B.; Marcel0 Curto, M. S. Tetrahedron Lett. 1994, 35, 88398842. (g) Morin, C.; Nedjar, N. Tetrahedron Lett. 1996,37,4705-4706.
(a) Barrero, A. F.; Altarejos, J.; Alvarez-Manzaneda, E. J.; Ramos, J. M.; Salido, S.
Tetrahedron 1993, 49, 6251-6262. (b) Barton, D. H. R.; Taylor, D. K.; Tse, C.-I.
Tetrahedron Lett. 1994, 35, 9505-9508. (c) Vlad, P. F. Russ. J. Phys. Chem. (Engl.
Transl.) 1992, 66, 467-468. (d) Martres, P. P.; Perfetti, P.; Zahra, J.-P.; Waegell,
B.; Giraudi, E.; Petrzilka, M. Tetrahedron Lett. 1993, 34, 629-632. (e) Friiter, G.;
Bajgrowicz, J. A.; Kraft, P. Tetrahedron 1998, 54, 7633-7703. (f) Schroder, M.
Chem. Rev. 1980,80, 187.
Ogino, T.; Mochizuki, K. Chemistry Lett. 1979,443-446.
(a) Wienhaus, H.; Pilz, W.; Seibt, H.; Dassler, H.-G. Chem. Ber. 1960, 93, 26252630. (b) Mills, J. S. Phytochemistry 1973, 12, 2407-2412. (c) Sandermann, W.;
Bruns, K. Chem. Ber. 1966,99,2835-2841.

126

18

19

20
21
22

23

Advances in Flavours and Fragrances

(a) Norin, T.; Ohloff, G.; Willhalm, B. Tetrahedron Lett. 1965, 39, 3523. (b) Bruns,
K. Tetrahedron 1969, 25, 1771-1775. (c) Barrero, A. F.; Sanchez, J. F.; AlvarezManzaneda, E. J.; Muiioz Dorado, M.; Haidour, A. Phytochemistry 1993,32, 12611265.
(a) Oehlschlager, A. C.; Mishra, P.; Dhami, S . Can. J. Chem. 1984, 62, 791. (b)
T a m m , Y.; Yamada, Y.; Ochiai, H.; Nakajo, E.; Yoshida, Z. Tetrahedron 1984,
40, 1791. (c) Coste-Manicre, 1. C.; Zahra, J. P.; Waegell, B.,Tetrahedron Lett.
1988,29, 1017-1020.
Herlem, D.; Ouazzani, J.; Khuong-Huu, F. Tetrahedron Lett. 1996,37, 1241-1244.
This enone 63 is also an important intermediate in the synthesis of crotomachlin
and 8-epi-crotomachlin. Herlem, D.; Khuong-Huu, F; Kende, A S . Tetrahedron
1994,50, 2055-2070.
The structure of the 6-0x0-Ambrox' obtained in this way was identical to the one
obtained by oxidation of 6-hydro~y-Ambrox~.
Several syntheses of the saturated analogues have been reported. (a) Ruzicka,
L.;Seidel, C.F. Helv. Chim. Acta 1950, 33, 1285-1296. (b) Sandermann, W.; Bruns,
K. Chem. Ber. 1966, 99, 2835-2841. (c) Ohloff, G.; Giersch, W.; Schulte-Elte,
K.H.;Vial, C. Helv.Chim. Acta 1976, 59, 1140-1157. (d) Cambie, R. C.; Hayward,
R.C.; Palmer, B.D.; Robertson, J.D.; Rutledge, P.S.;Woodgate, P.D. Aust. J. Chem.
1982, 35, 1699-1707. (e) Vlad, P. F.; Ungur, N. D. Synthesis 1983, 216-219.. ( f )
Coste-Manicre, I. C.; Zahra, J. P.; Waegell, B. Tetrahedron Lett. 1988, 29, 10171020.. (g) Urones, J. G.; Basabe, P.; Marcos, I. S.; Gonziilez, J. L.; JimCnez, V.;
Sexmero, M. J., Lithgow, A. M. Tetrahedron 1992,48,9991-9998.

THE SYNTHESIS OF FRAGRANT CYCLOPENTANONE SYSTEMS

H. C. Hailes
Department of Chemistry, University College London, 20 Gordon Street, London, WC 1H
OAJ, UK

1 INTRODUCTION
Cyclopentanones have long been recognised as having attractive olfactory properties and
we have been particularly interested in developing new synthetic routes to these molecules
and related analogues. A class of interesting 5-membered ring compounds includes
calythrone 1 (Figure 1), a naturally occurring cyclopentene- 1,3-dione P-triketone, with a
floral odour, reminiscent of cis-jasmone. The structurally related compound 2-butyl-4,5dimethylcyclopent-4-ene- 1,3-dione 2 was first identified3 during the synthesis of bovilide4.
Whilst 2 was noted to possess a buttery jasmine odour, the olfactory properties of related
analogues was unknown.
The methyl jasmonates5, the (+) trans-isomer 3 and (k)epi-isomer 4 shown, are key
well known natural products used widely in the formulation of many perfumes. Methyl
epijasmonate 4 possesses superior olfactory properties compared to trans-methyl
jasmonate 3, and the epi-isomers have also been shown to have greater biological
activities.677The unnatural analogue methyl dihydrojasmonate 5 is a constituent of famous
fragrances8 and magnolia ketone 6 has also been reported to possess fragrant properties.

fie
0

Caty-thrme1

2 B utyl-4,5dinethylcycbpent4enel,%dione2

Methyl dhydrqasmmate 5

Figure 1

Magndia ketone 6

Structures 1-7

Methyl jasmonate3

Methyl epijasmnate4

CisB-(2-oxopopyl)-2-(pen t-2Zenyl)
cyclopentml-me 7

128

Advances in Flavours and Fragrances

With our interests in the fragrance properties of new cyclopentanones and with a
desire to investigate new synthetic routes to these compounds we focussed our attention on
the synthesis of 2 and analogues, methyl epijasmonate 4 and the magnolia ketone analogue
7, and finally investigations into short syntheses of both enantiomers of trans-methyl
jasmonate 3 and trans-methyl dihydrojasmonate 5 . The results from these investigations
are described below.

2 RESULTS AND DISCUSSION

2.1 2-Alkylated Cyclopentene 1,3-Diones1

During the synthesis of bovilide, which is found in green tea leaves," 2-butyl-4,5dimethylcyclopent-4-ene- 1,3-dione 2 was identified as a side product3, Investigations into
the synthesis of 2 and analogues were performed for potential use in perfumery
applications. Two strategies were used and are outlined below: firstly the synthesis of 2
and related analogues via a 2-ylidenebutenolactone to identify the compounds with the
strongest olfactory properties; secondly, the synthesis of 2 using a Pauson-Khand and
selective oxidative strategy.

1 . NaOMe

EK)

2.2 M H-O'C
76%

a%

K,C03, n<Bu),l\Br

n-butyl/pentyl/hexyl
bnmide

9 0

OAOEt

Scheme 1

>&.
0

*
0

Synthesis of 2-alkylated cyclopentene 1,3-diones

The synthetic route shown in Scheme 1 was initially explored, due to previous
reports on the rearrangement of ylidenebutenolactones2.'2 and our aim of generating
1,3-dione 9. The reaction
analogues via the key intermediate 4,5-dimethylcyclopent-4-enebetween 2,3-dimethylmaleic
anhydride and ethoxycarbonylmethylenetriphenylphosphorane in toluene gave the butenolactones 8a and 8b in 76% yield, and an isomeric
ratio of E:Z of 2:3.l3>I4The geometric mixture of isomers was not problematical due to the
rearrangement in the next step. Accordingly, the mixture of 8a and 8b was treated with
sodium in methanol and then dilute HCl to effect the ring opening and closure, followed by
hydrolysis of the ester and decarbox lation to generate 4,5-dimethylcyclopent-4-ene1,3dione 9, in an overall yield of 83%.' We found that the use of methoxide, rather than
ethoxide, led to enhanced yields in this system.
Problems are often encountered in the selective C-alkylation of 1,3-dicarbonyl
compounds due to competing self condensation and 0-alkylation reactions. We initially
wished to access quantities of analogues of 2, both mono- and dialkylated materials, for
fragrance analysis purposes. Several alkylation procedures were explored including the
use of alumina impregnated with sodium ethoxide,16however, this led to the formation of
self condensation products only.
The use of a biphasic procedure using
tetrabutylammonium bromide as a phase transfer catalyst17 proved more promising and
using bromobutane led to the formation of 2 and 10 in 57% and 11% yields respectively.
Attempts to improve the yield of the monoalkylated compound in particular were not

129

Organic und Bioorganic Chemistry

successful, however it enabled access to these materials. Therefore, using the same
procedure with pentyl- and hexylbromide the corresponding analogues (11-14) were
generated. Analysis revealed that compound 11 possessed a much weaker jasmine odour
than 2, and that 13 was odourless. By comparison, the dialkylated compounds, 10, 12, 14
had weaker buttery odours. Thus 2-butyl-4,5-dimethylcyclopent-4-ene1,3-dione 2 still
possessed the strongest odour characteristics, and therefore a novel alternative route to 1
was sought to avoid the non-selective low yielding alkylation procedure.
The Pauson-Wand reaction is a powerful synthetic method to construct five
membered rings using allenic compounds together with alkyne cobalt c ~ m p l e x e s . ' ~This
~'~
would enable the formation of the five membered ring, however, oxidative cleavage would
be required to reveal 2. The route explored is outlined below in Scheme 2.
1. n(Bu)sSnH,AIBN,90C

15

80%
16

Scheme 2

Use of the Pauson-Khand to generate 2-butyl-4,5-dimethylcyclopent-4-ene-1,3dione 2

Dicobalthexacarbonylbut-2-ynecomplex was prepared from but-2-yne. 19720 Undec6 - y n - 5 - 0 1 ~was

~ ~prepared
~~
in 87% yield from the reaction between the anion of 1-hexyne
and n-valeraldehyde at low temperature. This was stored under nitrogen to avoid aerial
oxidation to the corresponding ketone, undec-6-yn-5-one, which possessed a strong
pineapple odour. Undec-6-yn-5-01 was then heated with tributyltin hydride and ALBN to
give the hydrostannylation product which was directly converted into the mesylate, and
underwent elimination to form 5,6-undecadiene 15.22.23The Pauson-Khand reaction was
then performed using N-methylmorpholine N-oxide monohydrate and 4-pentylidene-5butyl-2,3-dimethylcyclopent-2-enone16 was readily formed in 80% yield.''
The selective oxidation of the trisubstituted exocyclic alkene was initially
attempted using catalytic osmium tetraoxide/sodium periodate
however,
despite exploring a range of conditions low yields were achieved. The use of an
ozonolysis procedure together with an azo-dye (Sudan Red 7B26)to indicate completion of
the trisubstituted alkene bond cleavage successfully generated 2 in an isolated yield of
87%. Overall, using one synthetic route we were able to assess the olfactory properties of
several analogues and have prepared the most interesting compound using an alternative
synthesis.
2.2 Methyl Epijasmonate 4 and Cis-3-(2-Oxopropyl)-2-(pent-2Z-enyl)-cyclopen~n-lone 727
Methyl epijasmonate 4, in particular the (+)-enantiomer is known to have numerous
biological function^.^
Despite the ready epimerisation of epijasmonate to the
thermodynamically more stable trans-isomer 3, with its much greater odour characteristics
it still has potential for use in fragrance applications for example via a pro-fragrance
strategy. Furthermore, we were interested in the fragrance properties of analogues of

130

Advances in Flavours and Frugrunces

magnolia
We therefore focussed on the development of an efficient synthetic
route, to (+)-methyl epijasmonate 4 and (+)-cis-3-(2-oxopropyl)-2-(pent-2Z-enyl)cyclopentan-1 -one 7 (the unsaturated, epi-analogue of magnolia ketone), that could also be
adapted to potentially generate enantioselective samples.
One of the most efficient methods of establishing the cis-stereochemistry on the five
membered ring is via a Diels-Alder reaction. Previous routes have been reported to the
jasmonates using cyclopentenone and a Diels-Alder strategy including Torii s2 (A,
Scheme 3) who used butadiene and subsequently generated a hemiacetal to establish the
stereochemistries. A patent also describes the synthesis of methyl trans-jasmonate using
2-alkoxybutadiene (B,Scheme 3). The key cycloaddition steps were not high yielding
because regio-isomers were generated together with isomeric

Scheme 3

Previous routes to the jasrnonates using a cycloaddition reaction

The difficulties that can be encountered with such a Diels-Alder reaction using
butadienes have been investigated with the reaction proceeding to give a mixture of
cycloadducts including double bond shifted isomerised product^.^' Other synthetic
approaches reported to epijasmonate include a free radical approach to establish the cisstereochemistry by K n ~ c h e l ~and
,
M ~ n t f o r t s and
~~ Be~tmanns~
enantioselective
~
syntheses using chiral synthons with subsequent removal of the chiral directing group. We
wished to develop a general route for the construction of cis-substituted cyclopentanones,
but using readily available starting materials (see scheme 4).

Scheme 4

Strategy to synthesise epijasmonate using isoprene

The use of isoprene as a diene was ideal due to its low cost, and compared to 2alkoxybutadienes is less prone to polymerisation under Lewis acid conditions.
Furthermore, such a strategy would allow the direct generation of 6 . The reaction between
cyclopentenone and isoprene was initially explored using aluminium chloride as the Lewis
acid catalyst, however, the best reaction regioselectivity we observed was 5: 1, paraisomer: rneta-isomer in 49% yield (literature, ratio 10:1, 76% yield).33
An alternative, more selective procedure was required, and we noted that Grieco
has reported acid catalysed ionic Diels-Alder reactions in high yields and regioselectivities

Organic and Bioorganic Chemistry

131

using the ethylene ketals of cyclohexenones as dieneophiles, in 4.0-5.0 M solutions of

lithium perchlorate together with camphor sulfonic acid (CSA).37 Accordingly, the
ethylene ketal of 2-cyclopenten- 1 - 0 n e ~was
~ reacted with isoprene under these conditions
and readily gave the required adduct 17 in excellent yields (see Scheme 5). The
regioselectivity was 9 5 5 for 17a:17b (para-isomer: meta isomer) by 'H NMR.
Fortunately, the minor regioisomer was readily removed using column chromatography to
give 17a in 85% isolated yield. We noted that this reaction has also been reported but in
lower yields by Vankar et al, who has performed the reaction with lithium erchlorate in
nitromethane (73% yield) or using Nafion-H as an acid catalyst (68% yield).3 8

81
+

&+

4M
I mUCIo4-Et,0
oPhgA

91Yoyield

17b
Ratio955

TBDMS

1. Ozone,M e S
2. PbP=CHCH2CH3

67%
18

Scheme 5

19

>
20

Synthesis of the key intermediate 20

The direct ozonolysis of 17a to cleave the double bond resulted in the formation of
a complex mixture of products. It was evident that the ketal was also being cleaved during
the ozonolysis, so it was necessary to deprotect the ketal and reduce and protect the
resulting ketone to ensure that no epimerization would take place during the remaining
synthetic steps. Deprotection of the ketal was carried out using dilute hydrochloric acid
(2.7 M) at room temperature to give 18. We noted that if this reaction was performed at
higher temperatures migration of the double bond occurred yielding a mixture of isomers.
The reduction of the ketone was carried out using several reducing agents, however, our
preferred method was using cerium (111) chloride heptahydrate4' with sodium borohydride
which generated the diastereoisomeric alcohols in 9 1% yield and a ratio of 10:1 by 'H
NMR. Since one diastereoisomer was formed in much greater quantities than the other it
eased characterisation of the compounds in subsequent steps because at no stage was it
possible to readily separate the diastereoisomers. Indeed, it was not necessary since both
diastereoisomers would provide the required epi-compound after a late stage oxidation.
After protection of the alcohols in 81% yield, the silyl ether 19 was then reacted
with ozone treatment followed by a reductive work-up (dimethyl sulfide) to afford the
methyl ketone intermediate. The Wittig coupling was carried out using the unstabilised
ylid (supernatant solution), generated in situ from n-propyl phosphonium bromide and
sodium hexamethyldisilylazide, to which was added the freshly prepared aldehyde. The
use of such 'salt free' conditions is known to give rise to predominantly 2-alkenes, and the
key intermediate 20 was formed in 67% over two steps.4'
The conversion of 20 into cis-3-(2-oxopropyl)-2-(pent-2Z-enyl)-cyclopentan-1-one
6 was readily achieved (see Scheme 6). The deprotection of 20 with TBAF, and oxidation
of the resulting alcohol under mild conditions utilising freshly prepared Dess-Martin
periodinane4* gave 7 in 79% yield over 2 steps. The material possessed a sweet floral
odour.

132

Advances in Flavours and Fragrances

Alternatively the methyl ketone 20 was converted to the corresponding acid using a
sodium hypobromite in a dioxanelwater mixture for the haloform reaction. This
simultaneously led to the removal of the silyl protecting group in 89% overall yield.
Methylation of the acid afforded the methyl cucurbate diastereoisomers in 92% yield and
finally, oxidation of the cucurbate was carried out, again using Dess-Martin periodinane, to
readily generate methyl epijasmonate 4 in a yield of 78% over the 3 steps.
OTBDMS

1 . TBAF

2.D ~ s - M a l t h

20
1. NaOBr, dioxane

3.Dess-Martin oxidation

Scheme 6

Synthesis of methyl epijasmonate 4 and cis-3-(2-oxopropyl)-2-(pent-2Zeny1)-cyclopentan-I-one 7

This methodology was able to generate methyl epijasmonate and cis-3-(2Oxopropyl)-2-(pent-2Z-enyl)-cyclopentan1-one in overall yields of 29%, and we were
interested to establish whether the route could be adapted to generate enantioenriched
samples. Whilst it is envisaged that the ionic Diels-Alder goes via an oxygen allyl cation
which is stabilised by the lithium ion, it was not known whether a homochiral ketal or the
allyl cation generated would have any facial directing effects on the approaching diene.
Accordingly, the dibenzyl ketal 21 was prepared from cyclopentenone and (2S,3S)-1benz yloxy-4-( 2,4-cyclohexadienylmethoxy)butane-2,3-dio1.43~44When this was reacted
with isoprene in the presence of lithium perchlorate and 1% CSA unfortunately the two
diastereoisomers 22a and 22b were formed in equal q ~ a n t i t i e s . ~ ~However,
diastereoselectivites have been reported when performing ionic Diels-Alder reactions with
chiral acetals under alternative conditions, for example using TiC12(iP~0)2.~~
Although we
did not investigate this further it is possible that this approach could lead to the selective
formation of products.

21

Scheme 7

22a

22b

Homochiral ketal approach

2.3 Enantioselective Syntheses of Methyl Dihydrojasmonates 5 and Methyl

Jasmonates 346
Methyl trans-jasmonate and methyl trans-dihydrojasmonate are widely used in the
formulation of many perfumes.* There is keen interest in the enantioselective syntheses of
these compounds and several syntheses have been reported. For example, (-)-3a and (+)-

Organic and Bioorganic Chemistry

133

3b have been prepared using an enantiopure precursor and Posner's Pummerer

rearrangement strategy4' and the dihydroj asmonates (-)-5a and (+)-5b using a solid-liquid
asymmetric phase transfer catalysis where a range of e.e.s was obtained depending on the
phase transfer catalyst and reaction conditions ~tilised.~*
However, many of the syntheses
reported are multistep and therefore overall low yielding.
With the intense interest in this class of compounds there is a continuing desire to
develop versatile, short, selective syntheses. We were interested in developing a strategy,
which involved the conjugate addition of chiral phosphonamide anions to a,P-unsaturated
carbonyl compounds with subsequent removal of the chiral template. Additions using
~
chiral phosphonamide anions have been reported by., for example, H a y n e ~ ?Hanessian
who used Cz-symmetrical Fhosphonamides," Hua" and Denmark who used 1,3,2oxazaphosphorane 2-oxides' . Hua et a1 has reported the addition of extended lithium
anions of chiral phosphonamides to several cyclic a$-unsaturated carbonyl compounds
such as 2-cy~lopentenone.~'The products generated were subsequently converted into psubstituted acids and aldehydes with enantioselectivities of 28-98% e.e. The selectivities
observed depended on the enone, the phosphonamide used and N-alkyl group." As shown
in Scheme 8 the use of phosphonamides derived from (lR,2S)-ephedrine led to the
formation of products in approximately 70% d.e. However, more variable selectivities
were observed for those prepared from ( lR,2R)-norpseudoephedrine with an N-'propyl
group (22-88% d.e.).5'
1. mBuli, THF, -78OC

73% e.e.

Scheme 8

Hua's use of chiral phosphonamides to generate a P-substituted aldehyde"

The introduction of an alkyl group at the a-position on the enone system is likely to
enhance facial stereodifferentiation, with chelation of the lithium ion with the
cyclopentenone on the O=P-0 face of the phosphonamide (due to the N-alkyl group
present).
This would result in the formation of adducts in overall higher
diastereoselectivities than reported by Hua and would provide rapid access to the
jasmonates.
1,3,2The chiral phosphonamides, (2R,4S,5R)-3,4-dimethyl-5-phenyl-2-(2-propenyl)oxazaphospholidin-2-one 23, and (2S,4S,5R)-3,4-dimethyl-5-phenyl-2-(2-propenyl)-l,3,2oxazaphospholidin-2-on 24, were used since Hua had reported similar selectivities with
both diastereoisomers which we hoped to achieve. These were readily prepared from 2propene-1-phosphonyl dichloride and (lR,2S)-ephedrine in a 1:1 ratio and were readily
separated by flash column chromatography as a solid and an
Our initial studies
focused on the use of the enone, 2-pentyl-2-cyclopenten- 1-one.
Deprotonation of the phosphorous template 23 and addition of 2-pentyl-2cyclopenten-1-one led to the formation of 25 as the major addition product together with a
minor diastereoisomer in 80% yield (Table 1). The use of template 24 similarly led to the
formation of 26 and a minor diastereoisomer in 86% yield. The separation of the
diastereoisomeric products was investigated but not successful.

134

Advances in Flavours and Fragrances

After the highly selective reaction with 2-pentyl-2-cyclo enten-1 -one the reaction
was then performed using 2-(2-pentynyl)-2-cyclopenten-bone5' which generated 27 and
28 together with minor diastereoisomers in 76% and 83% yield, respectively.
r

25

23

Phosphonamide

Cyclopen tenone

Major Diastereoisomer

minor dhstereciswner

Major: minor
Yield
diastereoisomeric ratio

25

26

92:8

83%

Ph
Me

24

Table 1

28

The use of templates 23 and 24 and selectivities observed in additions to asubstituted cyclopentenones

The high reaction selectivities observed using both cyclopentenones, compared to

those reported by Hua who observed approximately 70% d.e.s, can be rationalised by
considering the higher facial stereodifferentiation obtained due to the presence of an alkyl
chain at the a-position (Figure 2). Thus, as shown the favoured reaction involves addition
onto the Si face of the a-substituted cyclopentenone and the presence of the alkyl group at
the a-position enhances the selectivity due to unfavourable steric interactions encountered
on additions to the Re face.
Efficient removal of the template was critical to the success of this synthetic strategy,
and initial experiments with 25 and 26 explored the use of both oxidative conditions
(hydrogen peroxide) and reductive conditions (dimethyl sulfide). The later generated 3-

Organic and Bioorganic Chemistry

135

0x0-2-pentanylcyclopentylethanal in 62% yield whilst the use of hydrogen peroxide gave

a mixture of inseparable compounds, A stepwise oxidation of the aldehyde with
subsequent methyl ester formation was carried out generating (+)-5b in only 20% yield

rrinor diastereoisomer

from 26.

Figure 2

The use of template 23 and 24 facial stereodiflerentiations

A more direct procedure was required, and the ozonolysis was then carried out in a
2.5 M solution of sodium hydroxide in methanol.55 This directly afforded (-)-methyl
dihydrojasmonate (-)-5a in 60% yield from 25 (Scheme 9) and (+)-5b in 57% yield from
26 in 91% e.e. and 85% e.e. respectively.
Due to the presence of the alkyne moiety in 27 and 28 care had to be taken during the
ozonolysis to avoid oxidation of the alkyne. Whilst again ozonolyses under reductive
conditions was considered, the corresponding aldehyde was generated in 39% yield.
Therefore, the reaction was carried out as before but was closely followed by tlc analysis to
indicate maximal product formation and the reaction was rapidly quenched. The
corresponding methyl esters were then isolated in 39-40% yield and subsequent Lindlar
reduction afforded (-)-3a and (+)-3b in 92% yield and 90% e.e. and 84% e.e., respectively.

27 R = CH~CHCHCHZCH~

I . Ozone, -78OC, Cti&h

2.5.M NaOH in MeOH
40%
**-\

C02Me

2. H2, Lindlar catalyst

92%

k2?
(->a

Scheme 9: Removal of phosphonamide templates

90% e.e

136

Advances in Flavours and Fragrances

In summary, we have used the 1,4-addition of chiral 2-propenylphosphonamide

anions to a-substituted cyclopentenones in high diastereoselectivities to selectively
generate (-) and (+)-methyl dihydrojasmonate (5a and 5b) and (-) and (+)-methyl
jasmonate (3a and 3b) via a short synthetic route. Such a synthetic strategy could be used
to readily synthesise further jasmonate analogues.

3 CONCLUSION
The synthesis of compounds 2, 3a, 3b, 4, 5a and 5b and 7 have been described. The
methodology developed has enabled the synthesis of new compounds for fragrance
analysis as well as investigating new synthetic routes to important extensively used
fragrance compounds. The synthetic routes described can also be used in the preparation
of further analogues and have several advantages over previously reported methods.

Acknowledgements
We are grateful to Bush Boake Allen Ltd for funding this research and providing a
studentship to M. Hashim Javaid.

References
1.
2.
3.

4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.

A.J. Birch and P. Elliott, Aust. J. Chem., 1956, 9, 95.

D.R. Gedge and G. Pattenden, J. Chem. SOC., Chem. Commun., 1978,880.
J. Janes, Personal communication, Bush Boake Allen Ltd., 1995.
K. Takahashi, T. Somega, S. Muraki and T. Yoshida, Agric. Biol. Chem., 1980, 44,
1535.
E.P. Demole, E. Lederer and D Mercier, Helv. Chim. Actu, 1962,45, 675.
E.P. Demole, in Fragrance Chemistry, Ed. Theimer, E. T., Academic press, New
York, 1982, 349.
M.H. Beale and J.L. Ward, Nat. Prod. Rep., 1998,533.
G. FrAter, J.A. Bajgrowicz and P. Kraft, Tetrahedron, 1998,54,7633.
C.J.L. Celli, Chem. Abstr., 1973,78, 147433.
H.C. Hailes, B. Isaac and M.H. Javaid, submitted for publication.
H. Horita, T. Hara, A. Sannai and T. Fujimori, Agric. Biol. Chem., 1985,49, 3601.
N.G. Clemo, D.R. Gedge and G. Pattenden, J. Chern. SOC., Perkin Trans. 1 , 1981,
1448.
R.A. Massy-Westropp and M.F. Price, Aust. J. Chem., 1980,33, 333.
C.F. Ingham, R.A. Massy-Westropp, G.D. Reynolds and W.D. Thorpe, Aust. J.
Chem., 1975,28,2499.
R.A. Massy-Westropp and M.F.Price, Aust. J. Chem., 1981,34, 2369.
B.C. Ranu and S. Bhar, J. Chem. SOC., Perkin Trans. 1 , 1992,365.
D.F. Taber, D.M. Gleave, R.J. Herr, K. Moody and M.J. Hennessy, J. Org. Chem.,
1995,60, 2283.
0. Geis and H.-G. Schmalz, Angew. Chem., Znt. Eng. Ed., 1998,37,9 11.
M. Ahmar, F. Antras and B. Cazes, Tetrahedron Lett., 1995,36, 4417.
I.U. Khand, G.R. b o x , P.E. Pauson, W.E. Watts and M.I. Foreman, J. Chem. SOC.,
Perkin Trans I , 1973,977.
K. Nonoshita, K. Maruoka and H. Yamamoto, Bull. Chem. SOC. Jpn., 1992,65, 541.
Y. Colas, B. Cazes and J. Gore, Bull. SOC. Chim. Fr., 1987, 165.

Organic and Bioorganic Chemistry

23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.

137

T. Konoike and Y. Araki, Tetrahedron Lett., 1992,33,5093.

G. Majetich, J. Defauw and C. Ringold, J. Org. Chem., 1988,53, 50.
T.K.M. Shing, J. Chem. SOC., Chem. Commun., 1986,49.
T. Veysoglu, L.A. Mitscher and J.K. Swayze, Synthesis, 1980,807.
H.C. Hailes, B. Isaac and M.H. Javaid, Tetrahedron, in press.
C.J.L. Celli, M. Plattier and P.J. Teisseire, Chem. Abstr., 1973,79,91675.
S. Torii and H. Tanaka, J. Org. Chem., 1975,40,462.
Hasegawa Perfumery Ltd., Japanese Patent JA-0022 152, March 1977.
T. Kitahara, K. Mori, M. Matsui, M. Iwamoto, Y. Takagi and Y. Warita, Agric. B i d .
Chem., 1982,46,1369.
H.O. House and G.H. Rasmusson, J. Org. Chem., 1963,28,31.
F. Fringuelli, F. Pizzo, A. Taticchi, T.D.J. Halls and E. Wenkert, J. Org. Chem., 1982,
47,5056.
P. Knochel and H. Stadtmiiller, Synlett, 1995,463.
F.-P. Monforts, I. Gesing-Zibulak, W. Grammenos, M. Schneider and K. Laumen,
Helv. Chim. Acta, 1989,72, 1852.
H.J. Bestmann, S. Qrschbaum and G.J. Roth, Synlett, 1997, 618.
P.A. Grieco, J.L. Collins and S.T. Handy, Synlett, 1995, 1155.
Commercially available, or see E.W. Garbisch, J. Org. Chem., 1965,30, 2109.
R. Kumareswaran, P.S. Vankar, M.V.R. Reddy, S.V. Pitre, R. Roy and Y.D. Vankar,
Tetrahedron, 1999,55, 1099.
J.-L. Luche and A.L. Gemal, J. Org. Chem., 1979,44,4187.
G . Stork and K. Zhao, Tetrahedron Lett., 1989,30,2173.
D.B. Dess and J.C. Martin, J. Org. Chem., 1983,48,4155.
H.C. Hailes, B. Isaac and M.H. Javaid, unpublished results.
E.A. Mash and C.J. Flann, Syn. Commun., 1988, 18, 391.
T. Sammakia and M.A. Berliner, J. Org. Chem., 1994,59,6890.
H.C. Hailes, B. Isaac and M.H. Javaid, Tetrahedron Lett., in press.
G.H. Posner, E. Asirvatham and S.F.Ali, J. Chem. SOC. Chem. Commun., 1985,542.
T. Perrard, J.C. Plaquevent, J.-R. Desmurs, and D. Hkbrault, Org. Lett., 2000,2, 2959.
M.R. Binns, R.K. Haynes, A.G. Katsifis, P.A. Schober and S.C. Vonwiller, J. Am.
Chem. SOC., 1988,110,541 1.
S. Hanessian, A. Gomtsyan, and N. Malek, J. Org. Chem., 2000,65, 5623.
D.H. Hua, R. Chan-Yu-King, J.A. McKie and L. Myer, J. Am. Chem. Soc., 1987,109,
5026.
S.E. Denmark and J.-0. Kim, J. Org. Chem., 1995,60,7535.
D.B. Cooper, C.R. Hall, J.M. Harrison and T.D. Inch, J. Chem. Soc., Perkin Trans I ,
1977, 1969.
G. Buchi and B. Egger, J. Org. Chem., 1971,36,2021.
J.A. Marshall and A.W. Garofalo, J. Org. Chem., 1993,58,3675.

DESIGNING DAMASCONE- AND IONONE-LIKE ODORANTS

Philip Kraft
Givaudan Dubendorf Ltd, Fragrance Research, Uberlandstrasse 138, CH-8600 Dubendorf,
Switzerland. Fax: (internat.) +41 (0)1/824 29 26. E-mail: philip.kraft@givaudan.com

1 INTRODUCTION

Put on your red dress

and slip on your high heels
and some of that sweet perfume
it sure smells good on you ..."
J. Gill, 'My, My, My'

When it comes to smelling 'sweet' or 'powdery', ionones are the perfumery materials of
choice. Discovered as early as 1893 by Tiemann and Kriiger? the ionones became central
to perfumery and fragrance chemistry, and thus inspired the synthesis of numerous analoga
and derivatives. In a broader sense, edamascenone and other damascones, Is0 E Super',
Koavone@,Timberol' and also Georgywood@can be considered ionone analoga, too.
Figure 1 places these odorants in a common t i m e l i ~ ~ e . ~

a-lp- ion one

(1893)

p-damasc(en)one

"Is0 E Super Plus"

--.--..._-_.._--_..
, (1990) 0.005 ng/L

Koavone
(1982)

(1967)

"seco-p-ionone"

Is0 E Super '

Timberol

Georgywood

(1962)

(1972) 500 ng/L

(1979)

(1996) 0.015 ng/L

Figure 1 The evolution of ionones, damascones and their analoga and derivatives

139

Organic and Bioorganic Chemistry

The ionone analoga in Fig. 1 cover a broad range of odour notes from floral
damascenone) via animalic (Timberol@)to woody characteristics (Georgywood*). In the
syntheses of these odorants, ionones served as structural or chemical templates, and thus
principles of rational design are well illustrated in this family.
2 SECO-IONONES

Acyclic compounds of the pseudoionone series were long known to possess no distinct
odour characteristics. Therefore, Jitkow and Bogert4 first varied the substituents on the
cyclohexenyl ring; and they stated in 1941 that the typical violet odour of pionone does
critically depend on a trimethyl-substituted cyclohexenyl ring. Of these three methyl
groups, two have to be adjacent to the 3-0x0-1-butenyl chain, so in position 1 and 5
according to the trivial numbering in Fig. 2. Accordingly, two bulky hydrophobic groups
flanking the 3-0x0- 1-butenyl substituent seemed necessary. But then again, was a
cyclohexenyl ring really crucial?
In 1962, Sestanj' imaginarily cut out ring atom C-2 and C-3 of pionone, and
proposed 5-isopropyl-6-methy1-3,5-heptadien-2-one
as an interesting target molecule that
could retain the odour characteristics of the parent pionone. He synthesised this 1(2),3(4)diseco-pionone starting from 3-isopropyl-4-methyl- 1-pentyn-3-01, and indeed found the
compound to possess an odour similar to that of pionone.' The physical data were also
reminiscent to those of pionone; however, as one could expect the boiling point of 104108"C/10 mm was significantly lower: The derived seco-ionone was more diffusive, while
close in odour to its parent compound.

1962
Sestanj

1983
IFF

Koavone
odor of Isoraldeine

Figure 2

The seco-concept in the design of ionones

The idea of Sestanj was taken up in the early 1980s by chemists of IFF, who chose
Isoraldeine@(8-methylionone) as parent structure. Employing ethyl methyl ketone instead
of acetone in the classical ionone synthesis, 8-methylionone had also been discovered by
Tiemann in 1893.6 It is more intense than the genuine ionones, possesses a very fine odour
reminiscent of violets with a slightly woody-vetiver tonality, and became therefore the
most popular violet odorant in perfumery.
Dissecting the bonds between C-2(C-3) and C-4(C-5), transposing the C-5 methyl
group to C-7, and hydrogenating the endocyclic double bond, leads from Isoraldeine* to
Koavone*, which was still quite close in odour to Isoraldeine* and 10-methylionone. This
odour similarity was also rationalized by a superposition analysis on c o m p ~ t e rone
; ~ of the
first applications of molecular modelling in the flavour and fragrance industry.

140

Advances in Flavours and Fragrances

3 CYCLO-IONONES
Omitting 'unnecessary' atoms, while retaining all structural features that code a given
odour note, hrnishes more volatile and more flexible molecules. These can povide more
diffkivity and impact in a perfume. OAen however, more substantivity is desired, and this
on the contrary requires higher molecular weight, besides lower thresholds that make up
for the lower vapour pressure.
Introducing additional rings in a molecule is an interesting strategy to devise more
substantive odorants. In addition to increasing the molecular weight, a ring system can
rigidify the molecule; thus allowing better insight into the receptor geometry. This insight
can then inspire hrther structure modifications, and thereby guide the synthesis of more
potent odorants.
There's quite
a bit of steric
bulk here, so
are the two
double bonds
electronically
necessary 7

methyl group
be included In
the ring ?

stronger, why not

this methyl group,
can it be included,
too 7

Figure 3

Designing cyclo-ionones at the desktop and on the computer

Fig. 3 presents the initial thoughts and molecular-modeling calculations that preceded
the synthesis of 1- {bicyclo[6.4.0]dodec-1(8)-en-10-yl)ethan-l-one.8The steric bulk of the
gern-dimethyl and the 8-methyl group of Isoraldeine' was mimicked by fusing a sixmembered ring. The remaining 5-methyl group was then found to be best superimposed by
a cyclooctene ring. This simplified the synthetic access by Diels-Alder reaction of but-3en-2-one with bis(methylene)cyclooctane, speculating that the double-bonds were more of
conformational than electronical importance. Indeed, the designed target compound
showed the typical fruity-woody, violet odour of p-ionone.*
1- {Bicyclo[6.4.0]dodec-I ($)-en- 10-y1)ethan-1-one was, however, not only interesting
for giving an idea about the molecular dimensions of the ionone receptor, but also as a
structural link to the woody family, e.g. Is0 E Super@.An interesting new reaction* led
regio- and distareoselectively to a potent woody-ambery odorant, reminiscent of
Georgywood@ and Is0 E Super@(Fig. 4). In the presence of Wilkinson's catalyst, 4methylenespiro[2.7]decane reacted with 3-methylbut-3-en-2-one to provide (r-9,c- 10)- 1(9,l O-dimethylbicyclo[6.4.0]dodec-1(8)-en-lO-yl) ethan- 1-one, which possesses an intense
woody-ambery odour with an excellent threshold of 0.01 ng/L air. By introduction of just
two methyl groups one can switch the odour note from typical violet-like to woodyambery. Further derivatives* provided additional insight into the structure-odour
correlation of these two odour notes.

141

Organic and Bioorganic Chemistry

r

Figure 4

Synthesis of (r-9,c-10)-1-{9,1 O-dimethylbicyclo[6.4.O]dodec-I(8)-en-10yllethan-1-one - a powerful odorant reminiscent of Is0 E Super@

4 ISO E SUPER@AND GEORGYWOOD~

Is0 E Super@,as a better quality of Isocyclemone E@was called later by IFF, was for
instance used in 18% in ))Trksor(<(LancBme, 1990), in 25% in ))Fahrenheit(((Dior, 1988),
in 35% in >)DCclaration<((Cartier, 1998), and even in 43% in nFeminitC du Boiw
(Shiseido, 1992)? It is synthesized by Diels-Alder reaction of myrcene with 3-methylpent3-en-2-one and subsequent acid-catalyzed cyclization as outlined in Fig. 5 .

ISO E Super@
threshold 500 ng/L

myrcene

Industrial synthesis of Is0 E Super@

Figure 5

However, GC-olfactometry at Givaudan showed that not the main product of Fig. 5 ,
but a ca. 5% constituent determines the characteristic woody-ambery odour of the
commercial material. While the theshold of the main compound of Is0 E Super@is about
500 ng/L, this minor constituent possesses an odour threshold of only 0.005 ng/L; and thus,
became known as Is0 E Super Plus at Givaudan internally.

5
(2.22)
48

\
8a
1

Is0 E Super Plus

threshold 0.005 ng/L

Figure 6

(1.50,
1.57)

2.20)

(5.44)

(1.71.
2.08)

Structure elucidation of the powerful impurity of Is0 E Super@

By reaction of commercial Is0 E Super@with peracetic acid and repeated column

chromatography, this powerful minor constituent could be isolated: and INADEQUATE
as
well as NOEDIFFexperiments led to the assignment of the structure depicted in Fig. 6. This
structure was proved by a directed synthesis starting from a-ionone. Its formation in the

142

Advances in Flavours and Fragrances

industrial synthesis of Is0 E Super@can be rationalised by acid-catalysed rearrangement of

the endocyclic double bond during the cyclisation step.
However, the structure of the powerfbl minor constituent turned out to be too difficult
to be produced on an industrial scale; and therefore, numerous derivatives were
synthesised, aiming at an easier access. Georgywood@(Fig. 7) turned out to be the best of
these derivatives. Its industrial synthesis by Diels-Alder reaction of homomyrcene with
methyl isopropenyl ketone is straightforward, and it was introduced into perfumery in
1996. >>GoldenMoments(< (P. Presley, 1999) with 5% of Georgywood is a recent
example for its use in perfumery.

Georgywood

Is0 E Super Plus

Superimposing Georgywood@on the lead structure Is0 E Super Plus

Figure 7

In order to arrive at the same spatial orientation of the substituents, the configuration
of the acetyl-bearing carbon of Georgywood@ is inverted. As is depicted in Fig. 7,
Georgywood@superimposes much better on Is0 E Super Plus than one would expect
bearing their two-dimensional structural formulae in mind.
Quite recently, Erman and coworkers at Millenium were able to increase the content
of Is0 E Super Plus by employing phosphoric acid in methanol. It is, however, still a
long way to the production of commercial Is0 E Super Plus at affordable price, and
Georgywood@constitutes a benchmark difficult to beat.
5 TIMBEROL~
A diastereomeric mixture of tetrahydroionols only possesses a very weak and
uncharacteristic odour. But when the side chain is elongated by just two carbon atoms, a
surprisingly powerful powdery, woody-floral odorant of animalic, steroid-type tonality
results, which has been successfully introduced into perfumery as Timberol@by Dragoco.
Though actually Timberol@ has not much in common with Ambrox@ in terms of
odour, the motivation for its synthesis may have very well been the design of a
9( lo), 12(13 -diseco-8-desmethyl Ambrox*, as sketched in a report by Brunke, Rojahn and
Warnecke. However, the 9-methyl-Timberol, which structurally resembles Ambrox@
even closer, is by far weaker than Timberol@,and smells only floral-sweet and fatty; so has

143

Organic and Bioorganic Chemistry

nothing in common with the odour of Ambrox@.In addition, the C5-stereochemistry of the
most active stereoisomer of Timberol@,which recently has been introduced into perfumery
as Dexnorlimbanol by Firmenich, does not match the configuration of Ambrox@

OH

TimberoI
Dexnorlimbanol
(Firmenich)
odorous principle

(Dragoco)
powdery, woody,
floral , an ima Iic

A possible idea behind the synthesis of Timberop, the

design of a 9(10), 12(13)-diseco-8-desrnethyl Arnbrox@?

Figure 8

But even though in terms of odour and stereochemistry this seco-design was not
successful, it may have helped to find the way to a new powerhl woody-floral, animalic
odorant, which -with respect to the weak and uncharacteristic tetrahydroionol- probably
no one would have expected.

6 THE DAMASCONE FAMILY

Like with the ionones, the first structure modifications in the damascone family were
systematic permutations and modifications of the cyclohexenyl s~bstituents.'~
And as it is
summarised in Fig. 9 quite a few of these analoga found their way into perfumery.

a-da m a scon e
(Firmenich)

S-damascone
(IFF)

isoda mascon e
(Dragoco)

apple, nutty, woody

fru ity-wood y,
da mascon e- Iike

Givescone
(Givaudan)

(Weyerstahl, 1996)

(Quest)

Figure 9

The damasconefamily so far

However, all these analoga possess rather different odour notes, and a- and
damascone as well as edamascenone remained benchmarks difficult to beat. The
importance of a-damascone is also documented in the elaboration of a stereoselective

Advances in Flavours and Fragrances

144

synthesis of (S)-(-)- a-damascone, which could be carried out on industrial ~ c a l e . That

~
this elegant process, which utilises (lR,2S)-(-)-N-isopropylephedrine as a chiral auxiliary
in an enantioselective protonation, still has not been employed technically, might be due to
the fact that both antipodes smell relatively similar. The @)-(+)-isomer is certainly more
apple-like, and also possesses an unpleasant cork-stopper nuance, but both enantiomers
share the same main odour characteristics, and their thresholds differ just by a factor of
70.
It should therefore still be possible to find more powerful damascone odorants, for
which one would expect a higher chiral discrimination on the receptor. If a chiral odorant
is perfectly complementary to the chiral proteinogenic receptor site, its enantiomer should
be odorless, or weaker by at least a factor of 1000.
OMgBr
BrMg4

15% aq.
H2S04

fruity, damascone- and tobacco-like, with

aspcets of saffron, ionones, and camphor

Figure 10

p-damascone

The seco-damascone patented by Takasago

In 1978, even before the introduction of Koavone (Fig. 2), 2-methyl-3-isopropylhepta2,5-dien-4-one, a 2(3),4(5)-diseco-~damascone,had already been described in a patent of
Takasago. l5 Grignard reaction of diisopropyl ketone with p-(3-butyn-2-olato)bis(magnesiurn bromide) furnished 5-isopropyl-6-methylhept-3-yne-2,5-diol,
which was transformed
into the target structure by means of a Rupe rearrangement (Fig. lo). This secodamascone possesses fruity, damascone- and tobacco-like odour characteristics, which are
indeed quite close to the parent /%damascone. However, 2-methyl-3-isopropylhepta-2,5dien-4-one was never introduced into perhmery.
7 SUBSTITUTED OCT-2-EN-4-ONES
We found this interesting seco-damascone by GC-olfactometry as a ca. 1.1% trace
constituent of a very complex, crude reaction product. At first, no reasonable structure
could be proposed for this trace constituent on the basis of the chemistry involved or the
GCMS data.16 Thus, a ca. 370 pg sample was isolated by preparative GC, but HMBC and
NOESY NMR spectra did not allow the unambiguous assignment of the structure. While
there was no doubt about a hepta-2,5-dien-4-one skeleton with one tetrasubstituted double
bond bearing a methyl and an isopropyl group, the relative positions of the latter remained

145

Organic and Bioorganic Chemistry

unclear. Besides 2-methyl-3-isopropylhepta-2,5-dien-4-one,

5,6,7-trimethylocta-2,5-dien4-one was therefore also a possible structure for this trace constituent. When we
synthesised both compounds, it turned out that the trace constituent of the reaction mixture
was indeed the 2-methyl-3-isopropylhepta-2,5-dien-4-one
that had been patented by
Takasago (Fig. 10). Yet, when we finally had also the 5,6,7-trimethylocta-2,5-dien-4-one
in hands, we found that this also possessed a damascone odour, and in fact was much
stronger and by far superior to that of the trace constituent.
1. NaNH2/
liq. NH3
2. Me1

OEt
1. LDA

2. BrMg+,,

Figure 11

Synthesis of (2E,5E/Z)-5,6,7-trimethylocta-d,
5-dien-4-one

Figure 1 1 details our synthesis of 5,6,7-trimethylocta-2,5-dien-4-one,

which
commenced with the preparation of 1-ethoxypropene from 2-chloro- 1,l -diethoxyethane.
After treatment with sodium amide in liquid ammonia, the resulting sodium acetylide was
alkylated with iodomethane. Addition of 1 -ethoxypropene to methyl isopropyl ketone in
the presence of boron trifluoride diethyl etherate furnished in 59% yield the ethyl 2,3,4trimethylpent-2-enoate. This was converted in the next step to the target compound by a
Grignard reaction with prop- 1 -en- 1 -ylmagnesium bromide and in-situ trapping of the
resulting dienone as lithium enolate.I6 (2E,5E/Z)-5,6,7-Trimethylocta-2,5-dien-4-one
possesses a very powerful and diffusive, fruity-rosy, pdamascone-like odour, reminiscent
of apples, plums, raisins, and other dried fruits. Contrary to what one would expect with
the previous seco-structures in mind, these interesting odour characteristics are however
mainly due to the (2E,SZ)-configurated cis-isomer, while the (2E,SE)-configurated transisomer is almost not detectable for some people.
weaker, but still
very diffusive,
fruity -rosy,
reminiscent of
raspberries and
dried fruits

fruity-rosy,
damascone,
reminiscent of
raspberries and
dried fruits,
bit agrestic

linear, intense
and longlasting,
fruity-rosy,
damascone-like,
reminiscent
of apples
and dried fruits

fruity, intensely
of apples, but
also reminiscent
of dried fruits,
with a celeryand jasmone-like side,
and agrestic aspects

Figure 12

Some derivatives with odour descriptions

146

Advances in Flavours and Fragrances

Therefore, 5,6,7-trimethylocta-2,5-dien-4-onecan not be considered a secodamascone, anymore, and it thus was very exciting to explore the structure-odour
correlation of related oct-2-en-4-ones. In Figure 12, four analoga are shown, together with
their respective odour descriptions. Even the partially hydrogenated derivative still
possesses a fruity-rosy odour, reminiscent of raspberries and dried fruits. In the dienone
analogs, the strength of the (5E)-isomers increases when the C6-substituent becomes more
bulky, and finally the (SE)-isomers become even stronger than the (5Z)-isomers.
discovered by serendipity, still
However, (2E,52)-5,6,7-trimethylocta-2,5-dien-4-0ne,
constitutes the most powerful and most pleasant damascone-like odorant of this series. As
with the powerful impurity in Is0 E Super@,serendipity therefore continues to play an
essential part in fragrance chemistry.
References

1
2
3
4
5

6
7
8
9
10
11
12
13
14
15
16

J. Gill, On Johnny Gill, Motown Record Company, Los Angeles, 1991, 530 025-2,
track 6.
F. Tiemann, P. Kriiger, Ber. Dtsch. Chem. Ges. 1893,26,2675-2708.
M. Gautschi, J.A. Bajgrowicz, P. Kraft, Chimia, 2001,55,379-387.
O.N. Jitkow, M.T. Bogert, J. Am. Chem. SOC.,
1941,63, 1979-1984.
K. Sestanj, Croat. Chem. Acta 1962,43,211-217.
F. Tiemann (Haarmann & Reimer GmbH), Ger. Pat 75120, 1893.
C.B. Warren, W.E. Brugger, G.S. Zander, Chem. Ind. London 1983,1,36-38.
P. Kraft, Synthesis, 1999,695-703.
C. Nussbaumer, G. Frater, P. Kraft, Helv. Chim. Acta, 1999,82, 1016-1024.
P. Kraft, J.A. Bajgrowicz, C. Denis, G. Frater, Angew. Chem. Int. Ed.
2000,39,2980-30 10.
M. Erman, M. Williams, P. Whelan, C. Cardenas, Perfum. Flavor. 2001,26
(MarcWApril), 16-2 1.
E.-J. Brunke, W. Rojahn, H.-U. Warnecke, Dragoco Rep. 1980,199-203.
G. Friter, J.A. Bajgrowicz, P. Kraft, Tetrahedron 1998,54,7633-7703.
C . Fehr, Enantioselective Protonations in Fragrance Synthesis, in Chirality in
Industry 11,eds., A.N. Collins, G.N. Sheldrake and J. Crosby, John Wiley & Sons
Ltd., Chichester, 1997, Chapter 16,335-35 1.
T. Kobayashi, H. Sugiyama, H. Tsuruta, T. Yoshida (Takasago Perfumery KK),
JP 1323532,1978; Chem. Abstr. 1980,93, 149797; AN 1980549797; DW 1985-32.
P. Kraft, C. Denis, W. Eichenberger, Eur. J. Org. Chem. 2001,2363-2369.

CREATION OF FLAVOURS AND THE SYNTHESIS OF RAW MATERIALS INSPIRED

BY NATURE'

M. L. Dewis' and L. Kendrick2

'IFF, R&D, 1515 Highway 36, Union Beach, NJ 07735, mark.dewis@iff.com
2BBA, Blackhorse Lane, Walthamstow, London, E l 7 5QP, UK

1 INTRODUCTION

1.1 The concept of Generessence

History - The flavour industry was born in the 1800s with the availability of herb and spice
extracts, manufacture of simple chemicals and the distillation of simple essential oils. A
flavour at this time would contain a mix of an extract (violet leaves), essential oils (coriander)
and simple chemicals (vanillin).
In the earlier part of the 1900s developments in the chemical industry meant that new
flavour chemicals were becoming available. At this time very little was known about the
chemical composition of flavours in nature so flavour chemicals were used simply because
they happened to smell of a particular food product.
In the mid 1900s flavour companies targeted useful chemicals in terms of odour and made
every possible variation. Aldehyde C16 is a good example of a component isolated from this
type of research. Also at this time the first serious attempts were made to analyse the chemical
composition of food. Chemists relied on bucket chemistry and identification of just one
chemical was a major undertaking. One important chemical discovered in this way was
raspberry ketone, the structure of which was initially kept a closely guarded secret. The
invention of G C M S opened up the possibility of detailed analysis of the flavour of natural
food. However attempts to recreate flavours from these first analyses was a failure due to the
lack of fine detail although this work did give rise to the discovery of many novel materials
such as hexenols and pyrazines. Flavours became recognisable and the modern flavour
industry was born.
Traditional flavours are based around key components with some occasional addition of
new compounds found in analyses. These flavours do not however represent the true
compositions of the original food, so as a result may be unbalanced and inflexible. For
example, they will work well in the application for which they were designed, but may not be
automatically used in another without modification.

148

Advances in Flavours and Frugrunces

Generessence@flavours- a new approach

The concept of Generessence@ is to re-create as exactly as possible the flavour as nature
intended.
Only component volatile compounds found in the target food are allowed in the flavour
formulation. This enforced discipline forces the flavourist to be truly creative and not rely on
traditional flavour ingredients.
Once the qualitative part is correct the quantitative part follows and the result is a truly
natural flavour character which is not distorted by dosage, is water and / or oil soluble and can
be applied to all applications.

2 INSTRUMENTATION

2.1 GC
2.1. I Non-polar. The extracts were analysed on a HP6890 GC with splitlsplitless injection
and a Flame Ionisation Detector (FID). lpL of extract was injected onto a SGE BPX-5
chromatographic column (50m x 0.22 mm ID x 0.25pm film thickness) in the split mode. The
injection temperature was 220C. The GC oven was run using temperature programming, with
an initial temperature of 40C held for 0.5 minutes ramped at 3"C/min to a final temperature
of 290C held for 5 minutes. The detector temperature was 320C.
2.1.2 Polar. The extracts were also injected onto a HP 6890 GC with splitlsplitless injection
and flame ionisation detector (FID) fitted with a SGE BP21 chromatographic column (50m x
film thickness) using the same injection and detection techniques. The
0.22 mm ID x 0 . 2 5 ~
GC oven was run using temperature programming with an initial temperature of 40C held for
0.5 minutes ramped at 3"C/min to a final temperature of 240C and held for 20 minutes.
All data was collected and stored on the Labsystems Atlas data system.

2.2 Mass Spectroscopy

For analysis on both BPX-5 phase and BP 21, the sample was introduced via a Hewlett
Packard 5890 GC into a Micromass Prospec magnetic sector mass spectrometer. GC oven
conditions were the same as the conditions outlined above. The mass spectrometer was
operated in EI mode scanning from m/z 500 to m/z 33 @ 1 second per scan. Spectra were
analysed on the OPUS data system using the Bush Boake Allen in-house library and the
commercial Wiley6 and NIST libraries.

2.3 Infra Red

Run as a neat liquid between sodium chloride plates on a Perkin-Elmer System 2000 FTIR
spectrophotometer.

Organic and Bioorganic Chemistry

2.4

149

NMR

Run in deuterochloroform solution on a JEOL GSX-270 FTNMR spectrometer with deuterium

lock and 5mm 'W 13C dual probe. The 'H spectrum was referenced against TMS as internal
reference and the 13C spectrum was referenced against the central peak of the
deuterochloroform triplet as 77.0ppm. Chemical shifts are quoted in ppm. (6).

2.5 Headspace
Headspace was collected onto tubes containing 160mg Tenax TA resin.
Thermally desorbed at 220C for 5 min in a Perkin Elmer ATD 400.
ATD internal cold trap at -3O"C, trap fired at 250C for 0.8 min and sample passed to GC
column via a fused silica transfer line.
2.6 Sulphur detector

Sievers 350B SCD. Probe fitted to HP5890 FID. Sample split 1:1 between SCD and FID.

3 COMPONENT IDENTIFICATION
Quantified components were identified by comparison of their mass spectra and linear
retention time indices' with authentic reference compounds. Authentic reference compounds
were synthesised as required and confirmed by spectral methods ('H-NMR, 13C NMR, IR,
MS) .
4 SCOPE OF DISCUSSION
A total of 193 Generessence@analytical projects have been undertaken. For the purpose of this
discussion and by way of example, we shall outline just a few based on three target samples,
beef, mango and hazelnut.

5 SAMPLE PREPARATION
5.1

Mango

Pulp and skin of organoleptically selected mangos (434g) were homogenised and extracted
with dichloromethane for five hours using several small soxhlet extractors. The combined
extracts were initially concentrated down on a rotary evaporator. High boiling and non-volatile
matter was removed using a bulb to bulb distillation apparatus, collecting the volatiles in a
dry-ice cooled collection bulb. The distillate was further concentrated by blowing down with
a stream of nitrogen gas.

150

Advances in Flavours and Fragrances

5.2 Grilled beef

5.2.1 Headspace. Slices of lean sirloin beef were cooked under an electric grill for
approximately 3 minutes each side until medium-well done. The hot meat was immediately
placed inside a specially designed glass apparatus with integral condenser, preventing
moisture from entering the Tenax traps and allowing meat flavoured condensation to be
collected for future extraction. Headspace was collected on each slice for 15 minutes after
which the sample was replaced with a freshly grilled slice. This process was repeated to give
a total sampling time of 4 hours.
5.2.2 Steam distillate. Slices of the grilled sirloin beef were cooked as described above. The
meet was minced and steam distilled in a Nickerson and Likens apparatus* for 4 hours with
dichloromethane as solvent. The extract was concentrated using an air condenser.
5.2.3 Liquid liquid extract of condensation. Condensation collected during the headspace
sampling was found to have a strong meaty odour. As such this was extracted into
dichloromethane and analysed by GC-MS.

5.3

BBQ beef

5.3.1 Headspace. Half-inch thick slices of lean sirloin beef were barbecued on a charcoal
garden barbecue until medium-well done. Each freshly cooked steak was placed in the
collection apparatus described above. The headspace was collected for a total of 5 hours on
successive steaks, each steak being replaced after 20 minutes to maintain the freshly cooked
aroma.
5.3.2 Steam distillate. Slices of the BBQ sirloin beef were cooked as described above. The
meat was minced and steam distilled in a Nickerson and Likens apparatus2 for 4 hours with
dichloromethane as solvent. The extract was concentrated using an air condenser.
5.3.3 Liquidliquid extract of condensation. As with the Grilled beef the condensation
collected during headspace sampling was extracted into dichloromethane and analysed by GC-

MS.
5.4 Beef and onion
Cubed beef and diced onions were cooked together to identify the composition of the flavour
of the combined materials. The analyses included a headspace sample collected during the
final minutes of cooking and a sample obtained by simultaneous distillation extraction of the
cooked mixture.
Cubes of beef were browned on all sides and mixed with lightly sautCed onions, covered
with water and simmered (covered) for approximately three hours. Headspace samples of the
cooking vapours were collected onto Tenax through a small condensate trap to control
moisture. At the end of the cooking time the mixture was steam distilled into pentane-ether
using a Nickerson and Likens apparatus2. The extract was concentrated by air condenser.

Organic and Bioorganic Chemistry

151

Additionally, this analysis provided an opportunity to characterise novel flavour

components generated from the precursor ingredients of the two food substrates.
5.5

Hazelnuts

The hazelnuts used for this work were chosen from a wide selection available by a panel of
flavourists.
5.5.1 Untoasted hazelnuts (steam distillate). The nuts were shelled, chopped and steam
distilled into dichloromethane for 2 hour using a Nickerson and Likens apparatus*. The
resulting extract was concentrated using an air condenser and analysed by GC, GC-MS and
GC fitted with a sulphur detector.

5.5.2 Untoasted hazelnuts (headspace). Freshly shelled hazelnuts were roughly chopped,
placed in a shallow dish and enclosed within a Rilsan bag. Headspace was collected for 3.5
hours.
5.5.3 Toasted hazelnuts. Headspace and steam distillation were prepared as above after
toasting the hazelnuts in an oven for 8 minutes at 24OoC
6 HIGHLIGHTS OF THE VOLATILE COMPONENTS IDENTIFIED
6.1 Mango
A total of 262 components were identified and quantified, 7 of which were a direct result of
the sulphur analysis. The analysis contained many terpene hydrocarbons, a notable
observation was ocimene being the major component in the Alphonso mango whilst
terpinolene was the major component in the Philippine mango. In both cases the level of the
other terpene hydrocarbon was negligible. All analyses contained a wide range of straight and
branched chain saturated aliphatic ether esters, whilst cis-3-hexenyl esters, 3-hydroxy ethyl
esters, methyl ketones, (E)-2-unsaturated aldehydes and lactones were also observed.
Comparison of the trace obtained by normal detection and using the sulphur detector is
shown in figures 1 and 2.

152

Figure 1

Figure 2

Advances in Flavours and Fragrances

153

Organic and Bioorganic Chemistry

Component

CAS

557-48-2
(E)-2-(Z)-6-nonadienal
64- 19-7
acetic acid
107-92-6
butyric acid
928-96-1
(Z)-3-hexenol
3338-55-4
(Z)-ocimene
3779-6 1- 1
(E)-ocimene
141-7 8-6
ethyl acetate
105-54-4
ethyl butyrate
2,5-dimethyl-4-0~0-4,5-dihydrofuran-3-yl 114099-96-6
bu tyrate
4-acetoxy-2,5-dimethyl-2,3-dihydrofuran-3- 4 166-20-5
one
4077-47 -8
4-methoxy-2,5-dimethyl-3(
2H)-furanone
586-62-9
terpinolene
methyl thioacetate
2-(methylthio)ethanol
methyl thiobutyrate
methyl (methylthio) acetate
ethyl (methylthio) acetate
ethyl 3-mercaptobutyrate

1534-08-3
527 1-38-5
2432-5 1- 1
16630-66-3
4455-13-4
156472-94-5

Recent reports in
nature
Chung et a14

Retention
index3
765
187
397
457
636
650
200

New identification

400
1024
85 1

Stojanova et al,
Menon et al,
Verzera et
Karagiand
Song et a1'
Weenen'
Werkoff et a19
New
identificationl o

669
693
294
450
495
517
593
642

Table 1 A selection of signijicant or interesting components in mango

6.2 Beef (combinationof the analyses of BBQ beef, grilled beef and beef and onion.)
A total of 528 components were identified and quantified in the 8 analyses, including 220
components that were sulphur and /or nitrogen containing.
A wide range of alkyl pyrazines was identified, along with a variety of 2-acetyl 5 and 6
ring-hetercycles. Sulphur components included trithianes, dithiolanes, 2-methyltetrahydro
furan-3-thiol, alpha mercapto ketones, alkyl and ally1 sulphides, disulphides and higher
sulphides, and particularly of interest furfuryl propyl disulphide' .
Among oxygen containing moieties a wide range of gamma and delta-lactones, branched
long chain saturated chain aldehydes, (Z)-2-aldehydes, C5 to C 12 (E)-2-aldehydes, C7 -C11
dienals and 2,3-diketones were identified. Aliphatic aldehydes C3 -C18 with C14 - C18
aldehydes constituting a large proportion of the volatile components. A number of acids were
also noted.

'

154

Advances in Flavours and Fragrances

Component
Cis,cis-2,4,6-triethyl-1,3,5-trithiane
2,5-dimethyl-3-ethylpyrazine
2,5-dimethylpyrazine
2-acetyl-2-thiazoline

CAS
53897-58-8
13360-65-1
123-32-0
29926-41-8

2-acetylpyrrole

1072-83-9

2-acet ylthiazole

24295-03-2

2-ethyl-4-methyl-1 ,3-dithiolaneI6
2-methyltetrahydrofuran-3-one
2-rnethyltetrahydrof~ran-3-thiol~
3-mercaptobutan-2-one
3-mercaptopentan-2-one
4-acet yl-gamma-butyrolactone
bovolide
decan-3-one
diallyl disulphide
dimethyl disulphide
dimethyl trisulphide
dipropyl disulphide
dipropyl trisulphide
furfuryl propyl disulphide
methional
methyl mercaptan
methy lpyrazine
ThialdineI6
(E)-2-(E)-4-decadienal
trimethylpyrazine

17564-27-1
3 188-00-9

Table 2

40789-98-8
67633-97-0
29393-32-6
774-64-1
928-80-3
2 179-57-9
624-92-0
3658-80-8
629-19-6
6028-6I - 1
252736-36-0
3268-49-3
74-93-1
109-08-0
638-17-5
25 152-84-5
14667-55 - 1

Recent reports in
nature
New identification
Choi et all2
Zehentbauer et a1 and
Ong et a t 3
Mahatheeranont et
Chung et a14 and Naf
et alls
New identification
Chung et a14

Naf et all
Naf et alls

New identification

Chungi8
Naf et all5

Retention
index*
1253
693
5 19
729
683
636
757
414
516 563
424
51 1
756
1143
795
697
34 1
583
723
950
1008
5 22
25
428
83 1
936
609

A selection of significant or interesting components in Beef

6.3 Hazelnut
A total of 336 components were identified and quantified. Unsurprisingly, it was observed that
pyrazines, pyridines and pyrroles were identified in the toasted hazelnuts. A range of aliphatic
alcohols and aldehydes, furans, 2,4-dienals, (E)-2-aldehydes, methyl substituted C-6 and C-7
saturated and unsaturated ketones, as well as methyl alkyl ketones were found in both cooked
and uncooked hazelnuts.

Organic and Bioorganic Chemistry

Component
Pyrazines
2,3-dimethylpyrazine
2,5-dimethylpyrazine
trimeth ylpyrazine
2-ethyl-3-methylpyrazine
5-methyl-6,7dihydrocyclopentapy razine
Saturated ketones
2-methylhexan-3-one2'
3-methylheptan-4-one2'
5-methylhexan-2-one
5-methy lheptan-3-one
Unsaturated ketones
2-methyl-(Z)-4-hexen-3-one
2-methyl-(E)-4-hexen-3one2'
5-hepten-2-one
5-methyl-1 -hepten-4-one

155

CAS

59 10-89-4
123-32-0
14667-55-1
15707-23-0
23747-48-0

7379- 12-6
15726-15-5
1 10-12-3
54 1-85-5
50396-99- 1
50396-90-2
67 14-00-7
8 1925-80-6
5090- 16-4
20859-10-3

5-methyl-(Z)-2-hepten-4-one
5-methyl-(E)-2-hepten-4-one
5-methyl-(E)-3-hexen-2-one
3-penten-2-one
Aldehydes
2,4-decadienal
hexanal
3-methylbutanal
Sulphur containing
2-(methylthio) acetaldehyde
4-( methylthio) butan-2-one
4-(methylthio) pentan-2-one
Others
2-acetyl pyridine
methyl pyrrole-2-carboxylate
octan-2,3-dione

Table 3

Recent reports in Retention Toasted

nature
index
523
519
609
610
757

Froehlich and
Schreier"
Ross et a]*'
Valeroet aIz2
Baser et

434

New
Identification
New
Identification
Kay~er*~
New
Identification
El-Sahart~~~
New
Identification

438

530
45 1
542

484
494
523
595
600

103070-07-1
102322-83-8
5 166-53-0
625-33-2

Blanch26
Shiota2'
Chung et aI4

530
578
500
333

2363-88-4
66-25- 1
590-86-3

Song et a12*
Valero2'
Valero2'

936
403
24 1

23328-62-3 Werkhoff et a12'

34047-39-7 Werkhoff et a126
143764-28-7 Werkhoff et
1122-62-9
1193-62-0
585-25-1

368
604
633

Chung et a14

649

Chevance"

589

A selection of significant or interesting components in Hazelnut

Untoasted

156

Advances in Fluvours and Frugrunces

6.4 Sensory evaluation

Where practicable components were identified as target compounds for synthesis by GC
~ d o u rEvaluation
.~~
of synthesised components was undertaken globally by an expert panel of
flavourists. Evaluations were performed in water, salt water or sugar water at a concentration
consistent with the threshold of the material. Results were presented on a hedonic scale, with
evaluation in application undertaken on selected candidates.

7 SYNTHESIS OF REFERENCE COMPONENTS

Authentic reference components were synthesised by established synthetic procedures:
3-mercaptopropyl propionate, 2mercaptopropyl propionate, ethyl 4-mercaptobutyrate, ethyl 3-mercaptobutyrate, 2,4,6triethyl- 1,3,5-trithiane, 2-ethyl-4-methyl- 1,3-dithiolane, octan-2,3-dione, 4-(methylthio)
butan-2-one, 4-(methylthio) pentan-2-one7 4-methyl-3-hepten-2-one,5-methyl-3-hepten-2one, 6-methyl-(E)-3-hepten-2-one,3-methylheptan-4-one, 5-hepten-2-one, 2-methyl-(Z)-4hexen-3-one, 2-methyl-(E)-4-hexen-3-one, 3-mercaptopropyl acetate, 3-hydroxypropyl
thioacetate, I72,4-trithiolane,thialdine, ethyl 2-methyl-2-pentenoate, 3-methylheptan-4-one7 2methylhexan-3-one, 4-acetyl-gamma-butyrolactone, bovolide, furfuryl propyl disulphide, cis1-pentenyl formate, 3-(allylthio) propanal, 12-methyltridecanal. Components were isolated by
distillation, chromatography or recrystallisation and verified by spectral analysis.

2,5-Dimethy1-4-oxo-4,5-dihydrofuran-3-y1butyrate,

7.1 Compounds identified and synthesised from analysis of the volatile components in
mango32

2,5-Dimethyl-4-0~0-4,5-dihydrofuran-3-y1
butyrate was identified for the first time in nature
and confirmed by comparison with an authentic standard. This material has been found to be
useful in a range of fruity flavours in addition to mango, imparting a more natural or less
candy note to the creation.
Key sulphur component:
The sulphur trace showed a component in the same place as a mango tropical odour was
observed by a GC-odour experiment. This component MW 148 was therefore identified as a
synthetic target.
From the analytical data collated from the Mango Generessence@analysis, retention time,
odour, GS-MS, GC-IR it was not possible to explicitly define a synthetic target. Therefore
four structures 1-4 were proposed (figure 3).

Organic and Bioorgunic Chemistry

Ethyl 3-mercaptobutyrate 1

2-Mercaptopropyl propionate 3

157

Ethyl 4-mercaptobutyrate 2

3-Mercaptopropyl) propionate 4

Figure 3
Two of these structures possess the 1,3-0xygen, sulphur o d o ~ r o p h o r so
e ~commonly
~
observed
in compounds with a tropical odour or natural occurrence. Thus these were the preferred
candidates, although all four were synthesised to validate the identification beyond dispute.
Comparison of analytical data with the authenticated synthetic standard34confirmed the
active nature identical target molecule as ethyl 3-mercaptobutyrate 1lo.

7.1.1Synthesis of the target molecule 1. Initially addition of sodium hydrogen sulphide to

ethyl crotonate in aqueous sodium bicarbonate solution was attempted, but successful reaction
conversion was not accomplished at ambient or 5OoC. Reaction of ethyl crotonate with
hydrogen sulphide in hexane was found to be unsatisfactory, with or without aluminium
chloride addition. It was found that reaction of sodium hydrogen sulphide with ethyl crotonate
in aqueous sodium bicarbonate solution was moderately successful under slight pressure
(scheme 1).
The isolated yield of ethyl 3-mercaptobutyrate 1 by this method was typically only 3%
and the product ratio was clearly variable. To improve the yield, reduction of dimer 6 was
undertaken (scheme 2) which was accomplished in moderate yield and made the overall yield
of product from ethyl crotonate 15%.
Once the target compound had been confirmed as the key ingredient in the mango
analysis, organoleptic evaluation, formulation trials and application tests confirmed that the
compound was indeed a key mango and tropical component in both flavour and fragrance
applications.
At this time steps were taken to improve the synthesis both in terms of ease and yield.
Attempts to form 6 using the chemistry described were unsatisfactory as large quantities of the
monosulphide 5 were always formed. It was found that addition of sulphur to the reaction
mixture enabled formation of disulphide 6 as the major product in good conversion (ca. 70%)
and in moderate yield (ca. 65% of theory). After reduction of 6 to the target mercaptan 1 the
overall yield of distilled material was 29%.

Advances in Flavours and Fragrances

158

Figures show % in final reaction mixture prior to work-up

1 - 10%

Scheme 1

Scheme 2
As a 'spin off' from this analysis a useful intermediate was made en-route to target 2. This
compound ethyl 4-(thioacetoxy) butyrate" has also been found of use as an artificial
flavouring ingredient.

7.2 Compounds identified and synthesised from analysis of the volatile components in
beef'
Although not specifically discussed in this work a Generessence@analysis of fried onion has
been undertaken as part of the program. The results below demonstrate some of the
differences in the synthesised components identified when beef is cooked with onion rather
than in isolation.
The following component was identified in beef, beef and onion and fried onion:
octan-2.3 -dione.
The following components were identified in beef, but not in beef and onion or fried onion:

Organic und Bioorganic Chemistry

159

4-acetyl-y-butyrolactone,cis-bovolide.
The following components were identified beef and onion and fried onion, but not in beef
2-ethyl-4-methyl- 1,3-dithiolane, 2,4,6-triethyl- 1,3,5-trithiane.
The following components were identified beef and onion but not in beef or onion:
furfuryl propyl disulphide, 3-(allylthio) p r ~ p a n a l ~ ~
The following component was identified beef, beef and onion but not in onion:
Thialdine.
Of the above mentioned components, furfuryl propyl disulphide, 2-ethyl-4-methyl- 1,3dithiolane and 2,4,6-triethyl- 1,3,5-trithiane are believed to be newly identified components in
nature.

7.3 Compounds identified and synthesised from analysis of the volatile components in
hazelnut3
Two of the compounds synthesised for use in hazelnut application, namely 6-methyl-(E)-3were newly identified in nature. One of
hepten-2-one and (E/Z)-2-methyl-4-hexen-3-0ne~~
these 2-methyl-(E)-4-hexen-3-oneis covered by a Haarmann and Reimer Patent and
although found to be organoleptically viable was not used in final creations.
6-Methyl-(E)-3-hepten-2-one, and 5-methyl-3-hepten-2-one were made by standard aldol
chemistry. (E/Z)-2-methyl-4-hexen-3-onewas synthesised by reaction of 2-bromopropane
with 2-butenal under Grignard conditions followed by oxidation of the resultant 2-methyl-4hexen-3-01 with pyridinium chlorochromate to furnish the required product. The previously
identifiedz4 5-hepten-2-one was made by ethyl acetoacetate condensation with crotyl bromide
followed by hydrolysis and decarboxylation.
4-(Methylthio) butan-2-one and-4-(methylthio) pentan-2-one were prepared my addition
of methyl mercaptide to the corresponding unsaturated ketone.
Octan-2,3-dione (also found in the analysis of beef and in many other Generessence@
analyses) was found to be an interesting component for use in the nut, cereal, meat and baked
categories.
Several of the ketones outlined in figure 3 were found to be cheaper than 5-methyl-(E)-2hepten-4-one FilbertoneTMand of key importance to the flavour. One component, 6-methyl(E)-3-hepten-2-one was indeed found to be essential to a good natural topnote flavour.
Removal of either this component or 5-methyl-(E)-2-hepten-4-onegreatly reduced the
organoleptic authenticity of the final flavour, particularly when used in application.
8 ACKNOWLEDGMENTS

The authors acknowledge the work of the following former Bush Boake Allen employees who
undertook some of the work described in this communication. Mr. R. Belanger, Mr. D. J.
Edwards, Dr. A-D. Fortineau, Dr. B. 0. Isaac, Mr. I. R. Smith, Mr. L. Trinnaman, Mr. N.
Unwin, Dr. G. E. Usher and Mr. A. Yusuf.

160

Advances in Flavours and Fragrances

References and Notes

1
2
3
4
5

6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34

35
36
37

Ethyl esters
J. Chromatogr. 1966,21, 1-5.
Ethyl esters, BPX5, 50m
J. Agric. Food Chem. 2001,49( l), 192
J. Essent. Oil Res. 2000, 12(4),438. J. Essent. Oil Res. 2000, 12(4),431. J. Agric. Food
Chem. 2000,48(7), 2903.
Vitis, 2000, 39(2), 7 1.
J. Agric. Food Chem. 1998,46(9), 3721.
J. Agric. Food Chem. 1996,44(lo), 3291
J. Agric. Food Chem. 1998,46(3), 1076
This material is the subject of a filed patent application.
This material is the subject of a filed patent application.
Han'guk Yongyang Siklyong Hakhoechi, 1996,25(3),478.
J. Cereal Sci. 1998,28(1), 81. J. Agric. Food Chem. 1998,46(6),61 1 and 2282.
J. Agric. Food Chem. 2001,49(2), 773.
Flavour Fragrance J. 2000, 15(5), 329.
Two diastereoisomers were identified.
Flavour Fragrance J. 1995, 10(4),243.
J. Agric. Food Chem. 1999,47(7), 2690.
J. Food Sci. 1990,55(l), 176
Subject of H&R Patent, DE 3525604.
J. Nat. Prod. 1996,59(1),49.
Chromatographia, 2000,52(5/6), 340.
J. Essent. Oil Res. 1999, 11(2), 253.
Flavour Fragrance J. 1998, 13(3), 209.
Adv. Food Sci. 1998,20(5/6), 198.
J. Agric. Food Chem. 1998,46(lo), 4283.
J. Agric. Food Chem. 1993,41(1l), 2056.
Flavour Fragrance J. 1999, 14(6), 383
Adv. Food Sci. 1996, 18(1/2), 19.
J. Agric. Food Chem. 1999,47(12), 5151.
HP-6890/Multitrax/Aromatrax, Multi-dimensional GCBniffer.
Furaneyl butyrate identified in analysis of Alphonso mango. Sulphur components
identified in Pakistani Mango
D. J. Rowe and B. Tangel, Perfumer and Flavourist, 1999,24,36-44
Ethyl estersBPX-5 642, BP 21 825; m/e 148M+(30), 42( loo), 69,61,74,55; IR (cm-1):
2980,2930,2907,2567, 1733, 1451, 1414,1373,1351, 1298, 1229, 1177, 11 15, 1090,
1035;'H NMR (270MHz, 6):4.16, 2H, t, 3.38,1H, irreg. Septet, 2.51-2.66, 2H, m,1.83,
IH, d, 1.38,3H, d, 1.27, 3H, t; I3C NMR (67.8MHz, 6):171.0,60.4,45.8, 31.2,24.7, 14.1.
Beef analysis includes a variety of different analyses, for the purpose of this discussion it
includes Grilled beef, Barbequed beef and fried and onion.
Tentative identification, along with 3-(propenylthio) propanal.
Components identified in analysis of either or both toasted or untoasted Hazelnuts

Flavours/Foods

NEW RESULTS ON THE FORMATION OF IMPORTANT MAILLARD AROMA

COMPOUNDS

Peter Schieberle and Thomas Hofmann

German Research Centre for Food Chemistry, Lichtenbergstrasse 4, 85748 Garching,
Germany; E-mail: Peter.Schieberle@Lrz.tum.de

1 INTRODUCTION
Flavours produced during thermal food processing, such as baking or roasting, are
universally known and are widely appreciated by man. Because usually the raw materials
do have rather bland flavours, it is quite obvious that the aroma compounds must be
formed by a chemical degradation of odourless precursors present in the food. One of the
first scientists who gave a more general idea on the reaction generating this type of food
aromas was Louis Camille Maillard. In his early studies* he reacted amino acids in the
presence of reducing carbohydrates and was able to generate different odours besides
brown colour. Because of the pioneering work performed later by Amadori3, Heyns4 and,
in particular, by Hodge, the first stages of the complex reaction cascade, today known as
the Maillard-reaction (MR), are basic knowledge in food chemistry and, in particular, the
primary reactions have been introduced in many textbooks.
One of the first comprehensive reaction schemes was published by Hodge about 50
years ago and named after him as the Hodge-scheme (Figure 1). It is suggested that, first,
the carbohydrate, e.g., an aldose, reacts with an amine to yield the glycosylamine, which
forms an 1,2-eneaminol and yields either the 3-deoxyosone by elimination of water or
undergoes a rearrangement into the Amadori-compound. This, in turn should form the 2,3enediol which may eliminate the amine, thereby generating a methyl a-dicarbonyl
intermediate, which is known today as the 1-deoxyhexosone. The deoxyosones will either
eliminate a further molecule of water or, alternatively, the carbohydrate skeleton will be
cleaved.
Although the Hodge scheme is frequently used to explain the early stages of the
Maillard reaction, it does not take into consideration that Amadori compounds and
deoxyosones do mainly exist in the less reactive cyclic conformation6 and, furthermore, it
does not give detailed information on the formation of volatile flavour compounds in the
advanced stages of the reaction.
Systematic sensory evaluations performed on thermally processed Maillard-type
mixtures have shown that the amino acid reacted clearly determines the type of aroma
generated. E.g. proline was characterized as a precursor of cereal, cracker-like aromas,
leucine generates chocolate-type odours and cysteine produces flavour impressions similar
to those of processed
This principle of generating flavours by heating mixtures of
amino acids and carbohydrates is today widely used in the manufacturing of the so-called
process flavourings.

Advances in Flavours and Fragrances

164

N-Substituted glycosylrmine

'Hzo

4 20

!lfd<

H~F-N<

F=U

?-OH

FHOH
FHOH

* CHOH
C-OH

CHOH

CH2

CHOH

I)

fH3

fH3

[-OH

c=o

f=O

e,
h=O

+IOH

GHOH

CHOH

C-OH
!--OH

2,3-Enediol

+Aainc

CHI
-ARiR@)

5-Hylroxymethyl2-furaldehyde

=!@+

3 -0eoxy hexosone
intermediate

1,2-Eacaminal

C-OH
&-OH

H =O

Hf=O
f=O

!H

CHOH
Amadori
compound

Aldohexose + Amiae

Methyl
a-dicrrbonyl
intermediate

+hint

C-methyl reductones
and a-dicarbonyls
IPyruvaldehyde, rcetol,
diacetyl, hydroxydiacetyll

Figure 1. The so-called "Hodge-scheme" of the Maillard reaction'

However, most of these products are made by an empirical rather than a systematic
approach. A main reason for this is the fact that often neither the key aroma compounds
evoking the characteristic aromas of processed foods nor the odours generated in processed
Maillard mixtures have been systematically characterized. However, to optimise either the
flavour formation in a food itself or in a reaction flavour, this knowledge is a prerequisite
to focus the investigations on reaction pathways leading to the formation of only the key
aroma compounds.
The application of an Aroma Extract Dilution Analysis on flavour extracts followed by
quantitations and a calculation of odour activity values (OAV: ratio of concentration to
odour threshold) has been proven as a useful approach to detect the most odour-active
compounds in foods or model systems". Among the odourants identified in processed
foods by the OAV approach, some compounds generated by Maillard-type reactions have
gained much interest due to their characteristic flavour qualities and, also, their high odour
potencies. Examples are the following: In fresh wheat bread crust' ','' as well as in freshly
popped
2-acetyl-1-pyrroline and 2-acetyltetrahydropyridine were characterized as
key aroma compounds. On the other hand, studies on Maillard-type model mixtures
containing the amino acid proline have shown that both intensely roasty, popcorn-like
smelling odourants are formed as the key aroma
The caramel-like smelling 4-hydroxy-2,5-dimethyl-3(2H)-furanone(4-HDF) has been
identified as an important aroma compound in many processed foods based on a
combination of analytical and sensory approaches (Table l)'27'7-2'.
Furthermore, 4-HDF
has also been characterized as a key odourant in sugar containing model systems
independently from the amino acid r e a ~ t e d ~ * - ~ ~ .
Another group of important food odourants are the Strecker aldehydes which are
generated by an oxidative decarboxylation of amino
Due to their low odour
thresholds, in particular, methional, 2- and 3-methylbutanal, phenylacetaldehyde and
methylpro anal have been identified as important contributors to several food
1,1 .17,19. Model reactions performed recently have, however, shown that besides the
Strecker aldehydes, the corresponding flavour-active Strecker acids may also be generated.

'

Flavours/Foods

165

Food
Conc. ( p g k g )
OAV
Roasted coffee
131000
13100
Cooked beef meat
9075
908
Rye bread c r u d 2
43 10
3 15
French fries
259 1
111
Wheat bread crustI2
1920
140
Popcorn2o
100
1370
Barlev malt2
820
60
a
Odour activity values were calculated by dividing the concentration by the odour threshold in
the respective matrix.

Table 1. Odour activity values of 4-hydroxy-2,5-dimethyl-3(2H)-furanone

in several
foods
A great number of comprehensive reviews on the Maillard reaction have previously
been published, e.g.27-32.While some papers focused exclusively on the mechanistic
aspects of the MR, a few recent papers also gave valuable information on the correlation
between the MR and food flavour corn pound^^^-^^. Because the knowledge on the
formation mechanisms of the above mentioned aroma compounds has been much extended
within the recent years, the intention of the following chapter is to discuss the formation
pathways of some Maillard-type aroma compounds established as key food aroma
compounds mainly based on results obtained in the authors lab.

GENERAL ASPECTS OF THE MAILLARD REACTION

In Figure 2, the reactions taking place in the primary stage of the MR are displayed based
on Hodges scheme, but taking into consideration mainly the cyclic forms of the respective
intermediates. The Schiff base (2 in Figure 2) formed from a hexose and an amino
compounds must not necessarily cyclise into the glycosylamine (3 in Figure 2). The key
intermediate is probably the 1,2-eneaminol (4 in Figure 2). This compound has three
possibilities to react further. Either an oxidation may take place leading to an a-0x0 imine,
which after hydrolysis, forms the hexosone (6 in Figure 2). Compound 4 may also
eliminate water, yielding the 3-deoxyosone after hydrolysis of the a-0x0 imine formed.
The third possibility is the Amadori rearrangement yielding either the Amadori compound
(7 in Figure 2) or the 1-deoxyosone (8 in Figure 2) after &elimination of the amino
compound.
None of the reaction intermediates shown in Figure 2 are stable and, therefore, the socalled advanced Maillard reaction will take place during further heat treatment. In
general, the following types of reactions must be considered:
dehydration reactions maintaining the carbohydrate skeleton
retro-AZdoZ reactions leading to fission products
AZdoZ-type reactions of fissions products previously formed
substitution of oxygen by nitrogen or sulphur
redox reactions
Strecker reaction

166

Advances in Flavours and Fragrances

11

(1 : carbohydrate; 2: Schiffbase; 3: Glycosylamine; 5 : 3-Deoxyosone; 6: Glucosone;

8: I-Deoxyosone; 7: Amadori product)

Figure 2. Primary reaction pathways of the Maillard reaction

To illustrate the two first mentioned reactions in more detail, the formation of furfural
(I in Figure 3) from the 3-deoxyosone of a pentose or norfuraneol (I1 in Figure 3) from the
1-desoxyosone is shown in Figure 3. Furthermore, a retro-Aldol cleavage of different adiketones is exemplified in Figure 4 yielding various fission products, e.g., 2-oxopropanal
or gl yoxal.

Figure3. Formation of fu?$ural ( I ) and norfuraneol (11) from the 3- and the I deoxyosone of pentoses

Flavours/Foods

167
n

R
O

CH,

I1

OH

I11
Figure 4. retro-Aldol cleavage of a glycosone ( I ) or the 1- (II) and the 3-deoxyosone
(III), respectively
3 FORMATION OF 4-HYDROXY-2,5-DIMETHYL-3(2H)-FURANONE
(4-HDF)

4-Hydroxy-2,5-dimethyl-3(2H)-furanone(4-HDF) is not very stable and can easily be

adsorbed by the food matrix or by silica during column separation and work-up
manipulations. So, an important prerequisite in performing studies on the formation of
such flavour compounds is an exact method for quantitation. Previously we have
e ~ t a b l i s h e dthat
~ ~ a stable isotope dilution assay using a carbon- 13 labelled isotopomer of
4-HDF (100 % labelling) as the internal standard is the method of choice and is, therefore,
used throughout our experiments. The differentiation between the labelled internal
standard [I3C2]-4-HDFand the flavour compound is done by mass chromatography (Figure
5).
It is well known that rhamnose, a 6-deoxyhexose, is an effective precursor of 4hydroxy-2,5-dimethyl-3(2H)-furanone. As shown in Figure 6, its formation is easily
possible from the cyclic form of the 1-deoxyrhamnosone. However, rhamnose is not
present ubiquitously in foods and, compared to glucose or fructose, its concentration is
generally much lower.
In a study on the precursors of 4-HDF in wheat bread crust or bakers yeast,
respectively, we could earlier identify fructose- 1,6-biphosphate and glucose-6-phosphate
as effective precursors of 4-HDF2'. As shown in Figure 7, acetylformoin, formed by
elimination of phosphate from the 1-deoxyhexosone-6-phosphate,is suggested as the key
intermediate. Its disproportionation (self-oxidation) immediately results in 4-HDF
formation.

Advances in Flavours and Fragrances

168
f

m l z : 129
A = 75489

A = 63448

Total ion current

&
I

No.

by a stable isotope
Figure 5. Quantification of 4-hydroxy-2,5-dimethyl-3(2H)-furanone
dilution analysis in combination with mass chromatography
H A

H3C' X 0 O

cHH 3

OH

16

f'

H: f&
0

H3C

H,O

Figure 6. Formation of 4-hydroxy-2,5-dimethyl-3(2H)-furanonefrom the 1 -deoxyrhamnose

wH
I- 4

OH

OH

H20

OH

+CH3
OH

OH

acetylformoin

Figure 7 . Proposed formation of 4-hydroxy-2,5-dimethyl-3(2H)-furanone

from glucose4-phosphate via acetylformoin as the key intermediate2'

169

Flavours/Foods

In a very recent model study on the formation of 4-HDF from acetylformoin we found that
the presence of proline much increased the yields of the flavour compound from
a~etylformoin~~.
It is assumed that acetylformoin reacts in its cyclic form yielding the
Schiff base with proline. Dehydration and decarboxylation then yields 1-pyrroline and 4HDF in a Strecker-type reaction34.
In a further model experiment, we recently compared the effectiveness of ribose,
glucose and rhamnose in forming 4-HDF in the presence of the amino acid cysteine. The
results indicated (Table 2) that generally increasing the pH also increased the yields of 4HDF independently from the carbohydrate used. However, it is interesting to note that 4HDF, a six-carbon compound, can be formed from a pentose. This gave us the idea that an
Aldol-type reaction of fission products, e.g., 2-oxopropanal and 2-oxopropanol may also
generate the 4-HDF via 2,5-dioxo-3,4-dihydroxyhexaneas the key intermediate (Figure 8).
In this type of reaction no reduction (disproportionation) of an intermediate is necessary.
Quantitative model studies on 2-oxopropanal/2-oxopropanolmixtures confirmed this
assumption, thereby corroborating the reaction pathway. The formation was favoured at
higher values and about 1.1 mol % of 4-HDF were formed at pH 7.035.
Carbohydrate
Ribose
Glucose
Rhamnose
a

Amount (pg) formed at pH

<

0.8
2.3

19
80
19800

208

788
1140
79860
The carbohydrate (10 mmol) and cysteine (3.3 mmol) were reacted at 145C for 20 min in
phosphate buffer (100 mL; 0.5 mol/L; pH 5.0).

Table 2.
Influence of the p H on the formation of 4-hydroxy-2,5-dimethyl-3(2H)furanone (4-HDF)from carbohydrates (Hofmann and Schieberle, unpublished results)a

by an Aldol reaction of
Figure 8. Formation of 4-hydroxy-2,5-dimethyl-3(2H)-furanone
2 - oxopropanal and 2-oxop ropanol (hydroxyace tone)

Advances in Flavours and Fragrances

170

From these studies the question arose whether and to what extent the 4-HDF is formed
from hexoses, such as glucose or fructose, either via acetylformoin or via carbohydrate
fission products.
By application of a so-called carbohydrate modul labelling experiment36 we could
recently show that in the formation of flavour compounds in the MR very often different
pathways are operating and, that water is an important modifier: while under dry-heating
conditions, 4-HDF was formed from proline and glucose only via the C-6 pathway (Figure
7), in the presence of water about 62 % of the 4-HDF was found to be formed from
carbohydrate fission products. Because in both model reactions acetylformoin showed a
similar isotopic pattern as the 4-HDF34,2-oxopropanol is obviously not the only C-3
intermediate. Based on the results obtained34, 1,3-dihydroxy-2-propanonewas suggested as
another very efficient intermediate in 4-HDF generation. A partially labelled 4-HDF was
formed from a 1+1 mixture of labelled and unlabelled glucose and, furthermore, partially
labelled acetylformoin was also detected34, the following reaction pathway is obvious: 2oxopropanal is condensed with dihydroxyacetone yielding a C-6 compound which forms
acetylformoin upon elimination of water (Figure 9). This is finally reduced in a Strecker
reaction with proline.
In general, the data indicates that the reaction conditions used in the Maillard reaction
may not only change the yields of certain aroma compounds but may also change the
formation pathways leading to the flavour compound under investigation.

T
OH

OH

OH

H2O

Figure 9. Formation of 4-hydroxy-2,S-dimethyl-3(2H)-furanoneby a retro-Aldol

reaction of 2-oxopropanal and dihydroxyacetone via acetylformoin as the key
intermediate34

I71

Flavours/Foods

4 FLAVOUR CONTRIBUTION AND FORMATION OF 2-ACETYLTETRA-

HYDROPYRIDINE
Hunter et al.37 were the first who described 2-acetyltetrahydropyridine as volatile
constituent of a thermally treated proline/dihydroxyacetone mixture. Later on the
compound was identified as an important odourant in wheat bread crust" and p ~ p c o r n ' ~ .
As indicated in Table 3, the odourant is the most odour-active among the 3 aroma
compounds eliciting a roasty, popcorn-like aroma in popcorn. In a model study on a
thermally treated proline/glucose mixture, we have recently show that the same three
odourants also contributed to the aroma of the model mixture thereby confirming this
amino acid as the precursor of the three flavour compound^'^.
By application of the carbohydrate modul labelling approach36 we have recently
shown that a 1+1 mixture of 2-acetyltetrahydropyridine and [ 13C3]-2-acetyltetrahydropyridine is formed from a 1+1 mixture of glucose and ['3Cs]-gl~~ose34.
This experiment
corroborated our previous data indicating that a ring opening of proline occurs during
ACTHP formation incorporating 3 carbon atoms from a carbohydrate fission pr~duct'~.

Odourant

Cone. ( c l g k )

Odour threshold

2-Acetyltetrah ydrop yridine

1207
0.054
2-Acetyl-1-pyrroline
57
0.0073
2-Propionyl- 1-pyrroline
21
0.010
a
OAVs were calculated by dividing the concentrations by the odour thresholds.

OAV"
2235 1
7808
2100

Table 3. Odour activity values (OAV) of selected roasty, popcorn-like smelling odourants
in ~ o p c o r n ' ~
In a previous paper we had suggested that 1-pyrroline, which may be formed by a
reaction of proline with a~etylformoin~~,
reacts with 2-oxopropanol in an Aldol-type
reaction to yield 2-( 1-hydroxy-2-oxo-3-propyl)pyrrolidine(Figure 10) which after
enolisation and water elimination directly yields the ACTHP. This reaction pathway has
been proven by model experiment^^^ and about 1 mol % of ACTHP was formed when 1pyrroline and 2-oxopropanal were reacted.
However, the following labelling experiments showed that 2-oxopropanol is not the
only intermediate in ACTHP formation. Synthesis of the 1-[ '3C1]-labelled Amadori
compound of proline and glucose (100 % label) and its degradation by a thermal treatment
in water generated a 1+1 mixture of l3C*-ACTHP(labelled in the methyl group) and
unlabelled ACTHP34.Contrary, only 30 % of the precursor 2-oxopropanol was labelled.
So, an additional pathway generating ACTHP must exist. Based on the obtained results it
could be concluded that acetylformoin is also a key intermediate in ACTHP formation. Its
Strecker-type reaction with proline, cyclisation, hydrolytic ring opening and elimination of
2-oxopropanal yields 2-( 1-hydroxy-2-oxo-3-propyl)pyrrolidine (Figure 11). This
intermediate was recently synthesized by us and proven to be a key intermediate in
ACTHP formation3*.

172

Advances in Flavours and Fragrances

L
7

Figure 10 Formation of 2-acetyltetrahydropyridinefrom I -pyrroline and 2-oxopropanol

(hydroxy -2 -propanone)

H,C

OHo

OH

wcH3

___*

H,C-C-CHO

OH

CH,

Figure 11. Formation pathway leading

2-acetyltetrahydropyridine

from

acetylformoin

and

proline

to

5 AROMA COMPOUNDS FORMED BY THE STRECKER REACTION

Once hexosones or deoxyosones as well as fission products of carbohydrates with an a diketo structure have been formed, a Strecker reaction will compete with the primary
reaction pathways shown in Figure 1. The Strecker reaction starts with the formation of an
imine from the amino acid and an a-dicarbonyl compound (Figure 12). The azavinylogous
B-ketoacid formed (I) easily decarboxylates yielding intermediate 11 which is finally
hydrolysed into an a-aminoketone and the Strecker aldehyde (V). Although about 20
amino acids may occur naturally in foods, only the following six amino acids are important
precursors of food odourants. The reason is that the aldehydes formed have quite low
odour thresholds:
Valine
+ 2-Methylpropanal
Leucine
-+ 3-Methylbutanal
Isoleucine
-+ 2-Methylbutanal
Methionine
+ Methional

Flavours/Foods

173

Phenylalanine
Alanine

+
-+

Phenylacetyldehyde
Acetaldehyde

By application of analytical studies in combination with sensory experiments, Strecker

aldehydes have been characterized as contributors to many food aromas, such as wheat
bread crust", roasted coffeeI7 or French fries", and recently we have found (Table 4 ) that
four Strecker aldehydes, namely 2- and 3-methylbutanal, methional and methylpropanal
were among the eight most important odourants of a special barley malt2'.
Very often also the corresponding acid, in particular, 2- and 3-methylbutanoic acid and
phenylacetic acid, additionally contribute to the overall aromas of foods and it is generally
believed that the acids are formed by oxidation from the respective aldehydes. However, a
very recent experiment using labelled phenylacetaldehyde clearly showed that this
assumption is wrong39.To shade some light on the formation of the acids in course of a
Strecker reaction, first, the time course of the formation of phenylacetaldehyde and
phenylacetic acid in a glucose/phenylalanine mixture was studied. The results showed that
the formation of both compounds started immediately with heating, but with increasing the
reaction time, the formation of the acid was favoured39.Further model studies indicated the
crucial role of free oxygen in the formation of, in particular, the acid (Table 5). In the
absence of oxygen (models A; Table 5) always the aldehyde formation was favoured.
However, in the presence of oxygen (models B; expts. 1 und 3; Table 5 ) acid formation
was favoured. Addition of copper ions (expt. 2; Table 5) enhanced this effect.

I11

Figure 12.

IV

Strecker degradation of L-phenylalanine initiated by 2-oxopropanal

Odourant
3-Methylbutanal
2-Methylbutanal
Methional
2-Methylpropanal

Conc.
(cl g k )

7500
6900
35
4000

Odour threshold
(pgkg starch)
32
53
0.27
56

OAV

235
130
130
70

Table 4. Concentrations and odour activity values (OAVs) of selected key odourants in
barley rnalC1

Advances in Flavours and Fragrances

174

Expt.
1
2
3
4
a

a-Dicarbonyl
2-Oxopropanal
2-0xopropanal~
Glyoxal
3-Deoxyosone

PA (pmol/mmol)
model Ab
model B"
11.1
10.2
10.0
n.a.
9.2
8.3
8.4
7.8

PAA (pmol/mmol)
model Ab
model B"
4.7
20.4
27.1
n.a.
2.2
12.2
1.5
3.3

A solution of the a-dicarbonyl (1 .O mmol) and L-phenylalanine ( 1 .O mmol) in phosphate buffer (10 mL;
0.1 mol/L; pH 7.0) was refluxed for 60 min in a closed vial.
Oxygen was absent.
Oxygen was present.
The reaction was performed in the presence of oxygen and copper (11) ions (0.05 mmol CuS04).
n.a., not analysed.

Table 5. Amounts of phenylacetaldehyde (PA) and phenylacetic acid (PAA) generated

from L-phenylalanine in the presence of diflerent a-dicarbonylsa - Influence of
oxygen and copper ions
Based on these results, the formation pathway shown in Figure 13 was proposed. The
first steps of the Strecker reaction are the same as shown in Figure 12. However,
intermediate I11 contains an eneaminol structure which has recently been shown to be
susceptible to air oxidation4'. The oxidation step, which may be catalysed by traces of
metal ions, leads to an a-0x0 imine (V in Figure 13) which, after enolisation is hydrolysed
to yield phenylacetic acid.

IV

[ H3p..
1
""q)

__*

H3C

VI

Figure 13. Proposed reaction pathway leading to the formation of phenylacetaldehyde

(IV) and phenylacetic acid (VI)from 2-oxopropanal and L-phenylalanine
Interestingly, if the Strecker degradation of phenylalanine was catalysed by the 3desoxyosone, compared with 2-oxopropanal, the acid concentration was not as
significantly enhanced (expt. 4; Table 5 ) . This fact can be explained by the reaction
pathway shown in Figure 14. Contrary to 2-oxopropanal, the 3-deoxyosone exists
preferably as a cyclic ketone. This intermediate may also form a Schiff base with the
amino acid leading to dehydration and decarboxylation. However, the 3-amino-

175

Flavours/Foods

dihydropyranone ring structure is not susceptible to oxidation, thus favouring the formation
of the aldehyde.

Figure 14. Proposed mechanism for the preferred formation of phenylacetaldehyde from
the 3-deoxyosone and L-phenylalanine
In a very recent study4' we showed that the presence of oxygen generally enhances the
yields in a Strecker reaction, in particular, when Amadori products were reacted (Table 6).
However, besides being more effective precursors, the Amadori products also favoured the
generation of aldehydes compared to binary reaction mixtures containing glucose and
phenylalanine (Table 6).
Exp.
I"
IIb
IIIC
a

ARP-Phe
PA
0.6
5.5
13.8

GlcPhe
PAA
0.3
3.0
7.6

PA
0.4
1.4
2.6

PAA
0.2
1.8
3.9

The precursors ( 1 mmol each) were heated (100C) in phosphate buffer (10 mL; 0.5 mol/L; pH 7.0) for
120 min in a closed vial under an atmosphere of argon.
Argon was replaced by air oxygen.
Heating was performed in an air atmosphere and in the presence of copper (11) ions (0.05 mmol CuS04).

Table 6 . Concentrations (pmol/mmol) of phenylacetaldehyde (PA) and phenylacetic acid

(PAA) generated upon thermal treatment of either N-(I -deoxy-D-fructosyl)-Lphenylalanine (ARP-Phe)or glucose/phenylalanine (GMPhe)
This data led to the proposal of a reaction pathway yielding a "Strecker" aldehyde in a
reaction different from the classical Strecker reaction. It is assumed that the Amadori
compound (I in Figure 15), an a-aminoketone, is directly oxidised into the corresponding
a-iminoketone (I1 in Figure 15). Hydrolysis of this intermediate may yield the glucosone
(I11 in Figure 15) thereby re-generating the amino acid. However, cyclisation of the
oxidized Amadori compound (IIb in Figure 15) followed by a decarboxylatiod
dehydration yields intermediate IV which is finally hydrolysed into phenylacetaldehyde.

176

Advances in Flavours and Fragrances

6 CONCLUSIONS

In general, the results of our studies show that the presence of a Maillard-type flavour
compound in a food does not tell us much detail about the way it was formed. So, from the
chemical standpoint, it is not very helpful to focus simply on carbohydrates and amino
acids as the precursors in the MR. The most effective intermediates and reaction pathways
have to be discovered, because the traditional published ways for the Maillard reaction
cannot be regarded as the one and only way that a food flavour compound can be formed.
In particular, if one intends to optimise the yields of a Maillard-type aroma either in a
food or a reaction flavouring the following systematic concept may be useful:
Characterise the key Maillard-type odourants in foods
Increase knowledge on the flavour potential of a food by quantitative correlations of
precursors present in the food and amounts of flavour compounds formed thereof by a
thermal treatment
Investigate the formation of single odourants in model system by means of labelling
experiments and synthesised intermediates
Enhance the yields of key aroma compounds by either enhancing the precursor/
intermediate concentrations in the food itself or by reacting precursors outside the food

Ho
0

m
l

Ila

YO

Figure 15. Formation of phenylacetaldehyde via an oxidative degradation of N-(I -deoxyD-fructosyl-L-phenylalanine (ARP-Phe)
References
1.
2.
3.
4.
5.
6.

7.

8.

L.C. Maillard, C.R. Hebd, Seances Acad. Sci., 1912,154,66.

L.C. Maillard, C.R. Hedb, Seances Mem. SOC.Biol., 1912,72,559.
M. Amadori, Atti. R. Accad. Naz. Lincei. Mem. C1. Sci. Fis. Mat. Nat., 1931, 13,72,
K . Heyns and A. Noack, Chem. Ber., 1962,720
J. E. Hodge, J. Agric. Food Chem., 1953, 1,928.
H. Weenen and S.B. Tjan, In: Flavour Precursors, R. Teranishi, G.R. Takeoka and
M. Giintert (eds.), ACS Symposium Series 490, American Chemical Society,
Washington DC, 1992,217.
W. J. Herz, R. S . Schallenberger, Food Res., 1960, 25,491.
G . MacLeod, M. Sayyadain-Ardebik, Crit. Rev. Food Sci. Nutr., 198 1 , 14, 309.

Flavours/Foods

9.
10.
11.
12.
13.
14.
15.

16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.

I77

M.J. Lane and H.E. Nursten, In: The Maillard Reaction in Food and Nutrition, ACS
Symposium Series 215, Amercian Chemical Society, Washington DC, 1983, 141.
P. Schieberle, In: Characterization of Foods-Emerging Methods, A. Goankar (ed.),
Elsevier, 1995,403.
P. Schieberle and W. Grosch, Z. Lebensm. Unters. Forsch., 1985,180,474.
P. Schieberle and W. Grosch, Z. Lebensm. Unters. Forsch., 1991,192, 130.
P. Schieberle, J. Agric. Food Chem., 199 1,39, 114 1.
P. Schieberle, J. Agric. Food Chem., 1995,43,2442.
P. Schieberle, In: The Maillard Reaction in Food Processing, Human Nutrition and
Physiology, Advances in Life Sciences, Birkhauser, Basel, 1990, 187.
D.D. Roberts and T.E. Acree, In: Thermally Generated Flavours, T.H. Parliament,
M.J. Morello, R. McGorrin (eds.), American Chemical Society, Washington DC,
1994,71.
P. Semmelroch and W. Grosch, Lebensm. Wiss. Technol., 1995,28,310.
H. Guth and W. Grosch, Lebensm. Wiss. Technol., 1993,26, 171-177.
W. Wagner and W. Grosch, Lebensm. Wiss. Technol., 1997,30, 164.
P. Schieberle, In: Flavour Precursors, R. Teranishi, G.R. Takeoka and M. Guntert
(eds.), American Chemical Society, Series 490, Washington DC, 1992, 164.
B. Fickert and P. Schieberle, J. Agric. Food Chem., 2001, submitted.
T. Hofmann and P. Schieberle, J. Agric. Food Chern., 1995,43, 2187.
T. Hofmann and P. Schieberle, J. Agric. Food Chern., 1998,46, 2721.
T. Hofmann, P. Munch and P. Schieberle, J. Agric. Food Chern., 2000,48,434.
A. Strecker, Ann., 1862, 123, 363.
A. Schonberg and R. Moubacher, Chem. Rev., 1952,50,261.
F. Ledl and E. Schleicher, Angew. Chemie Int. Ed. Engl., 1990,29, 565.
V. Yaylayan, In: Trends in Food Science and Technology, Elsevier, 1990, 20.
F. Ledl, In: The Maillard Reaction in Food Processing, Human Nutrition and
Physiology, Advances in Life Sciences, Birkhauser, Basel, 1990, 19.
H. Weenen, J. Kerler and J.G.M. van der Ven, In: Flavours and Fragrances, K.A.D.
Swift (ed.), The Royal Society of Chemistry, 1997, 153.
R. Tress1 and D. Rewicki, In: Flavour Chemistry - Thirty years of progress, R.
Teranishi, E.L. Wick and J. Hornstein (eds.), Kluwer AcademicPlenum Publishers,
New York, 1999,305.
G. Rizzi, In: Flavour Chemistry - Thirty years of progress, R. Teranishi, E.L. Wick
and J. Hornstein (eds.), Kluwer Academic/Plenum Publishers, New York, 1999,335.
A. Sen, P. Schieberle and W. Grosch, Lebensm. Wiss. Technol., 1991,24,364.
T. Hofmann and P. Schieberle, In: Flavour 2000 - Perception, Release, Evaluation,
M. Rothe (ed.), Eigenverlag, Bergholz-Rehbriicke, 2001,3 1 1.
T. Hofmann and P. Schieberle, In: Heteroatomic flavour compounds in foods.
Proceedings of the ACS Symposium August 2000, Washington DC, American
Chemical Society, Washington DC, 2001, submitted.
T. Hofmann and P. Schieberle, J. Agric. Food Chem., 2001, submitted.
I.R. Hunter, M.K. Walden, J.R. Scherer and R.E. Lundin, Cereal Chem., 1969, 46,
189.
T. Hofmann and P. Schieberle, J. Agric. Food Chem., 1998,46, 2270.
T. Hofmann, P. Munch and P. Schieberle, J. Agric. Food Chem., 2000,48,434.
T. Hofmann and P. Schieberle, In: Flavour Science - Recent Development, A.J.
Taylor and D.S. Mottram (eds.), The Royal Society of Chemistry, 1996, 182.
T. Hofmann and P. Schieberle, J. Agric. Food Chem., 2000,48,4301.

OUT OF AFRICA : THE CHEMISTRY AND FLAVOUR PROPERTIES OF THE

PROTEIN THAUMATIN

Steve Pearce & Hayley Roth2

Maverick Innovations Ltd, The Technology Centre, Station Road, Framlingham, Suffolk,
IP13 9EZ
2Britannia Natural Products Ltd, 2, Millbrook Road, Birkenhead, Wirral, CH4 1FL

1 INTRODUCTION
In 1848 a British army surgeon, WF Daniell, working in Western Africa noticed the local
use of a fruit to sweeten fermented palm wine. Further investigation revealed that part of
this fruit was intensely sweet and so he collected specimens for identification and
classification by Kew Gardens in London. Upon return to England Dr Daniell presented a
paper to the Pharmaceutical Society of Great Britain where he referred to this fruit as the
Miracle Fruit of Western Africa. Kew later named the fruit Thaumatococcus daniellii.
Since the death of Dr Daniell in 1865 Thaumatococcus daniellii has been largely
ignored until the 1970s when extensive research and product development led to the
identification and isolation of the key sweet component, a protein now known as
Thaumatin.
2 PRODUCTION

Thaumatococcus daniellii grows freely below the canopy of the secondary rainforest of
Western Africa and is effectively
a weed that grows on the forest
floor. It is widely distributed
throughout the region growing in
Ivory Coast, Nigeria, Ghana and
Cameroon. Dr Daniell referred to
it growing in Soudan this is an old
name for part of Southern Nigeria
and not to be confused with the
Republic of Sudan we know
today.

Figure 1 Thaumatococcus
daniellii fruit on the floor of the
rainforest

Flavou rs/Foods

179

After collection of the fruit the high protein aspect or Aril is cut out and frozen before
shipping to the UK for extraction.

Figure 2 Cutting out the arils from Thaumatococcus daniellii fruit in Ivory Coast
Once in the UK the production operation commences with washing and then aqueous
extraction of the protein from the Aril. The liquor is then concentrated by reverse osmosis
before freeze drying to produce the effectively pure protein, Thaumatin.

Figure 3 Concentration and purijication of Thaurnatococcusdaniellii extract

180

Advances in Flavours and Fragrances

Figure 4 Pure Thaumatin

3 BIOCHEMICAL STRUCTURE AND PHYSICAL PROPERTIES OF

THAUMATIN
Thaumatin is composed of a mixture of closely related proteins known as the thaumatins.
The predominant protein is Thaumatin I (TI) followed by Thaumatin I1 (TII, maximum
45%).

Both TI and TI*are proteins of 207 amino acids, with a molecular weight of 22,209
and 22,293. This is quite small for a protein but it is similar in size to the caseins in milk
(i.e. casein 23,000), to the enzymes trypsin (23,800) and papain (20,900). It is, however
considerably larger than insulin (1 1,466) and the polypeptide hormones.
The sequence of amino acids is shown in figure 1. All the common amino acid is
alanine. The N-terminal ammonium group is not acetylated or converted into a pyrrolidene
carboxyl group, this latter feature being often found in extracellular proteins such as
immunoglobulins, venoms or hormones. The C-terminal carboxyl group is also
unsubstituted by an amide group. This rare event is occasionally found in venoms and
hormones.
The relative abundance in TI and T ~respectively
I
of the basic amino acids arginine (12
or 13 residues) and lysine (1 1 residues) plus the weakly basic asparagine (10 or 8 residues)
and glutamine (4 or 5 residues) compared to the acidic amino acids, aspartic acid (1 2 or 13
residues) and glutamic acid (6 residues) is responsible for the high overall isoelectric point
of around pH 11. This isoelectric point is high compared to most proteins indicating the
relatively dense accumulation of the basic residues at the hydrophilic surface of the
protein. Protamines and histomes have high isoelectric points but they have much higher
content of arginine (70-80% in protamines, TI has 8.5%) and have a much lower molecular
weight (circa 5000).

Flavours/Foods

3.1

181

Tertiary Structure of The Thaumatins

A striking feature of the thaumatins is the 8 disulphide bridges which give a crosslinked
network of amino acid chains thus conferring unusual and extremely useful stability to heat
and pH denaturation.
Disulphide bridges are common in proteins that travel or operate in extracellular space
as digestive enzymes, immunoglobulins and milk proteins. For example, lysozyme has 5
disulphide bridges, comprising 7.2% of the protein (almost identical to TI: 7.4 % cysteine),
trypsin has 6 and bovine serum albumen, 17 disulphide bridges.
The presence of Cys-Cys at positions 158 and 159 in the sequence is another
frequently encountered structural element (e.g. human serum albumen) which forms the
basis of linking three chain segments close together again conferring stability.
The basic amino acids lysine and arginine and the dicarboxylic amino acids aspartic
acid and glutamic acid also contribute to the tertiary structure through salt linkages.

3.2

Quaternary Structure

TI and TI^ have only 1 polypeptide chain and thus normally have no oligomeric forms or
sub units and hence no regular quaternary structure. The type of amino acid residue and
their linkages and sequence suggest ready enzymolysis and digestion by human gastric and
intestinal proteases and peptidases. This has been confirmed in vivo by rat digestibility and
human volunteer studies. (Data on File at Britannia Natural Products)

3.3

Solubility

Thaumatin is freely soluble in cold water too greater than 20% solution (200 mg/ml)
whereas normal levels of use are 0.00005 - 0.0005 % by weight.
In alcohols such as ethanol and isopropyl alcohol, 12% solutions can be obtained in
60% alcohol and prehydration in a little water will allow solubility in up to 90% alcohol.
Gels can be formed at high concentrations of Thaumatin in alcohol. Solubility is good (at
least 10%)in aqueous glycerol, propylene glycol and polyols such as sorbitol (all 60-80%
in 40-20% water). Thaumatin is not suitable in typical organic and aprotic solvents.

3.4

Calorific Value

The extreme effectiveness of Thaumatin allows practical use levels to be effectively noncalorific. Although per gram it has the normal calories of a protein (4.1 cal/g, or 17 kJ/g),
a typical use as a flavour modifier would have no measurable calories.

SAFETY AND LEGISLATION

Thaumatin is metabolised like any other protein and is recognised as completely safe. It is
a naturally occurring protein extracted from a fruit and meets all the legislative
requirements as a natural food ingredient. It is also recognised and certified as meeting the
requirements of both Kosher and Halal authoritys.
Generally speaking there is world-wide approval for Thaumatin as a natural food
ingredient and more specific approval for use a sweetener and flavour enhancer. In the
USA Thaumatin is designated FEMA GRAS (Generally Recognised As Safe) and has the

Advances in Flavours and Fragrances

182

FEMA Number 3732. It is widely approved FEMA and recognised by the FDA for about
30 applications in food and beverages.
In Europe approval is given for use both as a sweetener and as a flavour enhancer with
the designation E 957.
In Japan Thaumatin is a natural ingredient approved for general use in food and
pharmaceuticals.
The FAOWHO Joint Expert Committee on Food Additives (JECFA) have declared
Thaumatins allowable daily intake (ADI) as not specified. In so doing recognising and
acknowledging the safe non-toxic character of the protein.
4.1

Oral Care

Thaumatin has a very important property which makes it extremely useful for oral care
products (as well as sweet confectionery), it is non-cariogenic .i.e. it does not promote the
cause of dental caries. Thus it has proved a very popular ingredient in oral hygiene
products.
5 ESSENTIAL PROPERTIES
Many of the interesting flavour and food properties of Thaumatin can be traced back to its
unique sweetness profile. Recognised as the sweetest natural substance Thaumatin has a
slow onset of sweetness followed by a very elongated taste and sweetness profile.

Sucrose

Thaumatin

12
10

2
0

Figure 5

Sweetness profile of Thaumatin

FlavourdFoods

183

5.1 Sweetness Intensity

Thaumatin is well known as the sweetest Natural substance and is indeed mentioned in the
Guinness Book of Records as such, being 6000 times sweeter than a 1% sucrose solution.
The sweetness intensity measurement does in fact alter with the concentration of the
comparison solution. So, comparing Thaumatin with a 10% solution of sucrose may give a
figure closer to 3000 times its sweetness.
5.2

Synergy with other Intense Sweeteners

Thaumatin shows excellent synergy with other intense sweeteners and polyols typically
contributing say 1% or 2 % of total sweetness to the combination but resulting in a total
overall perceived sweetness increase of 10%to 20%.

5 !

Thaumatin (ppm)
Figure 6 Thaumatin sweetness synergy
Best results are seen with Sweeteners where the further enhancement of the sweetener
system occurs through masking of the bitter and metallic after taste contributed by
products such as Saccharin for example.
5.3

Flavour Modification

Thaumatin can show significant flavour modification over a wide range of concentration
levels and a variety of applications.
Essentially the slow onset of sweetening effect and its longevity are the key factors in
this flavour improvement and enhancement. Beware however that the wrong characters are
not enhanced in a complex mixture thus leading to a loss of performance.
At levels above 30 ppm, when used on its own, Thaumatin will exhibit a faint
liquorice taste and slight cooling effect in the mouth. When working with Thaumatin one
should be aware of this as the liquorice perception can be quite longstanding and may
interfere with later work and evaluation of Thaumatin s effects in other formulations.

184

Advances in Flavours and Fragrances

Thaumatin (ppm or mg/Kg)

0.2 - 0.5

Effect
Enhances Feed Flavours
and Masks Bitterness

0.5 - 5.0

Modifies Flavour &

Enhances Aroma
Prolongs Flavour &
Intensifies Sweetness
Masks Extreme
Bitterness & Contributes
Sweetness

30 - 200

50- 150

Table 1

Examples
Animal Feed & Pet Food,
Saccharin Masking &
Synergy
Flavourings & Coffee
Chewing Gum & Hard
Candies
Pharmaceutical & Oral care
Products

Broad application levels for flavour enhancement

Flavour Characteristic

Normal Perception
Threshold (%)
0.00006

Peppermint
Menthol
0.00006
Coffee Extract
0.01
Chocolate Flavour
0.0048
Milk Base
0.048
Vanilla essence
0.0048
Orange Essence Oil
0.0048
0.0024
Apple Aroma
Chicken Extract
0.072
Beef Extract
0.0045
Lemon Essence Oil
0.02
Strawberry Flavour
0.00024
Mustard Seed Oil
0.0006
Tuna Fish
0.0048
Perilla
0.0024
Table 2 Examples offlavour enhancement

Reduction of Threshold
with 0.5 ppm Thaumatin
10
3-5
3-4
2
2
2
2
2
1.5 - 2
2
2 -8
1.5 - 2
2
1.5 - 2
4

Successful applications of this flavour enhancement effect can be seen with Mint
flavours particularly in oral care and chewing gum confectionery. From Table 2 above it
will have been noted that there is a significant reduction in the threshold of perception
when Thaumatin is applied with mint characters including peppermint oil, spearmint oil
and menthol. This can enable lower volumes of mint oils and flavours to be used in the
application or as is often the case with chewing gum to prolong the perception of flavour in
the product.
5.4

Mouth Feel improvement with Thaumatin

A very interesting and notable fact is that of the improvement of mouth feel as well as the
enhancement of flavour in many applications. This is very noticeable and useful in the
application in low fat dairy products. Often sensed as being watery and lacking in body a
small addition of Thaumatin to a low fat formulation can give the perception of a much
greater fat content. By adding Thaumatin to the flavour system or fruit preparation that

185

Flavou rs/Foods

may be used to flavour a yoghurt for example can modify the perception of the creaminess
and fat content of the product base and add an enhanced flavour effect.

Apparent Fat Content (MouthFeel Enhancement) with

Thaumatin Using a Low Fat Strawberry Yoghurt Sweetened with
Aspartame & Ace-K
41
39
37
w 35
a
t.;i 33
31
29
27
25

v
A

.R

Thaumatin (ppm)
Figure 7

Thaumatin effects on mouth feel

Figure Seven also serves to demonstrate the importance of finding the correct level for
the addition of Thaumatin in each application. There are no exact rules for the use of
Thaumatin except to say that each product is different and that Thaumatin will have an
optimum level of performance in every application.
Here we can see that around 2 ppm of Thaumatin gave the optimum mouthfeel
enhancement in this particular combination of low fat dairy product and strawberry
flavour. Increasing the level of Thaumatin does not necessarily lead to improved
performance. It is most likely here that the panel has started to focus on the negative
aspects now being improved and brought forward through the increased addition of
Thaumatin. For example, an aspect of the strawberry flavour the panel find unpleasant (say
butyric notes) may now being brought forward.
These effects, then, can be used to advantage in product applications such as Vitamin
and Mineral Tables, OTC Products, Nutraceuticals, Active Pharmaceutical Preparations,
Soya Peptides and masking of ingredient degradation products.

5.5 Bitterness Masking with Thaumatin

The late onset of sweetness and the ability of Thaumatin to enhance flavours means that it
can also demonstrate the ability to mask the aftertaste and bitterness of materials. A
particularly useful benefit when working with both Pharmaceutical and Nutraceutical
ingredients.
The benefit of using Thaumatin in Soya containing foodstuffs is demonstrated in
Figure 8.

186

Advances in Flavours and Fragrances

Bitterness Masking of Thaumatin against lo00 ppm of Soya

Peptides

Thaumatin (ppm)

Figure 8

Bitterness masking of Thaumatin

A good example of this in a successful application is the use of Thaumatin in

combination with a Lemon Flavour in beverage systems. This combination of Lemon Oil
based flavour in an aqueous environment exposed to light and air will almost certainly lead
to oxidation of the aldehydes present to give a cardboard like off-note. This character is
easily detected and will give the final beverage an unacceptable short shelf life often of
only a few weeks.
Demonstration of the Masking Effect of Thaumatin in Diet
Lemon Drink Storage Trial

Figure 9

O$-note flavour masking with Thaumatin

Fla vou rs/Foods

187

Using Thaumatin in the sweetener system works by masking these off-notes.

Although it does not prevent the oxidation of the Citral, by preventing the detection of the
degradation products it enables the product to be perceived as acceptable and thus
lengthens the shelf life.

6 APPLICATIONS
6.1

Flavours

The use of Thaumatin at low levels (typically less than 2 ppm at use level) in flavour
formulations will greatly enhance the immediate impact and the longevity of perception as
well as having particular synergistic effects with specific individual components.
Especially noteworthy is the impact with Mint oils and Menthol.

6.2

Flavour Enhancers

Whilst having a flavour enhancement impact of its own Thaumatin has been shown to
have a major synergism with Monosodium Glutamate (MSG) allowing the reduction of
MSG by 80% without noticeably lowering the perceived flavour enhancement effect.

6.3

Chocolate

Thaumatin has the right combination of effects to make a major impact on the flavour of
Chocolate. Firstly, there is the ability to enhance the creamy nature of dairy products and
boost the apparent fat content thus giving a smoother mouthfeel to the product. Add to this
the masking of the bitter notes associated with chocolate and the overall improvement is
significant.

6.4

Cereals

The use of Thaumatin to reduce the sugar content of childrens cereals is a relatively new
application but of course one with great merit. Technically the Thaumatin can be either
dusted on (which may affect the appearance of the cereal) or be included in the formulation
before cooking and extrusion. Thaumatin has great heat stability and may be extruded with
other materials without loss of performance.
Not only can the sugar level be reduced but the cereal flavour is enhanced with the
further benefits including : Improved mouthfeel when eaten with low fat milk; further
flavour improvement of fruit pieces; greater appeal for chocolate flavour cereal as bitter
notes are masked and creamy chocolate flavour is improved. This combination also tends
to make the cereal more appealing to children.
6.5

Pet Food and Animal Feed

The use of Thaumatin in both pet food and animal feed is fairly well explored. It does
appear that Thaumatin does however possess the ability to be appealing to both cats and
dogs an unusual property in a pet food ingredient. It has been shown that pets prefer food
with Thaumatin and can detect doses measured in parts per billion.

188

Advances in Flavours and Fragrances

Thaumatin is widely used in animal feed applications as a powerful natural sweetener and
flavour enhancer.

6.6

Tea and Coffee

The main components of both Tea (Caffeine and Tannic acid) and Coffee (Caffeine) are
extremely bitter and astringent so one would expect Thaumatin to improve palatability by
masking this bitterness. However, Thaumatin has a great impact when milk is added
especially noticeable with coffee of course. Thaumatin also has a significant role to play
with the latest trends in iced tea and coffee drinks and particularly with those that are dairy
based, contributing to all aspects of the product through masking bitter notes to improving
mouthfeel and creaminess and enhancing added flavour character.

6.7

Alcoholic Beverages

Thaumatin works well in alcoholic beverages at levels around 5 ppm it noticeably

smoothes flavour characters such as gin and at higher levels contributes to sweetness. In
complex mixtures, Martini and Pimms like products, that include herbal extracts
Thaumatin can help in bringing out and reinforcing character such as orris root or
In low alcohol products such as shandys or fruit juice containing new age beverages
for example Thaumatin can help give the impression that there is a higher concentration of
alcohol present.

6.8

Over the Counter Medicines and Pharmaceuticals

Thaumatin has proved extremely useful in masking long-lasting bitter aftertastes of active
materials in oral medicines particularly those used in cough remedies and pain relief.
Reasonably high levels (220 - 500 ppm) of Thaumatin may need to be used but will
prove effective with problem products such as Paracetamol, Codeine and the intensely
bitter and metallic after taste left with B vitamins for example.
The flavours that tend to be used in these products, mint, menthol, aniseed
wintergreen etc, also readily lend themselves to enhancement with Thaumatin.
Each formulation will have to be examined for the optimum level of Thaumatin of
course since the overall effect will depend on the total active ingredients and the sweetener
system being employed.
Again the non-cariogenicity of Thaumatin gives it further advantage in childrens
medicines.

7 CONCLUSION
Thaumatin is a very complex and intensely sweet protein, natural and safe in use, with
excellent characteristics for use in flavours.
Whilst possessing a flavour of its own at levels above 30 ppm, when it reveals the
cooling, liquorice mouthfeel, at lower levels it displays great synergy with flavours and the
ability to enhance the impact of a variety of flavour components and foods.

Flavours/Foods

189

Acknowledgements & Bibliography

The practical work referred to above has been mostly carried out by Hayley Roth and
Steve Pearce at Britannia Natural Products Limited and represents a summary of our
current understanding of the use of Thaumatin in flavour work. However, we are indebted
to those that provided the sound basis to work upon.
These people include our current colleagues at Britannia, who have provided great
support and encouragement, as well as the former employees of the Talin Company and
Tate & Lyle who carried out much of the early research and generated data on potential
applications for Thaumatin. Much of that work is referred to below.
Please contact either of us for further information or help and advice on how to use
Thaumatin in your food and flavour formulations.
For further reading and a more detailed historic and technical treatise on Thaumatin we
have provided the following list of references.
W F Daniell, Katemfe, or the miraculous fruit of Soudan. Pharm J 14, p158 (1855).
J M Dalziel. In: A Dictionary of the Economic Produce of the Malay Peninsula. Crown
Agents for the Colonies (1935).
J M Dalziel. The Useful Plants of West Tropical Africa. Flora of West Tropical Africa 3,
1:8 1, London, Crown Agents (1937).

G E Inglett and JF May, Tropical plants with unusual taste properties. Econ Bot 22: 4
p326 (1968).
M A Adansi. Indigenous plants in Ghana with taste-modifying properties or sweetening
principles. Ghana Jnl Agric Sci 3 p 207 (1970).
G E Inglett. Natural and synthetic sweeteners. Hon Science 5:3 p139 (1970).

H van der We1 and K Leove. Isolation and characterisation of Thaumatin I and 11.
Eur J Biochern 31 p22 1 ( 1972).
H van der Wel. Thaumatin, the sweet-tasting protein from Thaumatococcus daniellii
Benth. Olfaction and Taste IV p226 (1972).
0 Korver, M van Gorkom and H van der Wel. Spectrometric investigation of Thaumatins
I and 11, the sweet-tasting proteins form Thaumatococcus daniellii Benth. Eur J Biochem
35 p554 (1973).

R G Cagan. Chemostimulatory protein. A new type of taste stimulus. Science181

p32 (1973).
J N Brouwer, G Hellekant, H van der We1 and Y Zotterman. Electrophysiological study of
the gustatory effects of the sweet proteins rnonellin and Thaumatin in monkey, guinea pig
and rat. Acta Physiol Scand 89 p550 (1973).

190

Advances in Flavours and Fragrances

H van der Wel. Thaumatin and monellin, the intensely sweet-tasting proteins from
Tropical African fruits. Collog Ges Biol Chem 25 p235 1974).

H van der Wel. J N Brouwer and H De Iongh. Sweet-tasting and sweetness inducing
proteins from fruits of African plants. Nut Synth Zusatzs Nahr Menschen Znt Symp 154
( 1974).
H van der Wel. T C van Soest and E C Royers. Crystallisation and crystal data of
Thaumatin I, a sweet-tasting protein from Thaumatococcus daniellii Benth FEBS letter 56
p316 (1975).

J D Higginbotham. Natural protein sweeteners. Chem and Znd 20thMarch p262 (1976).
G Hellekant. On the gustatory effects of monellin and thaumatin in dog, hamster, pig and
rabbit. Chem Senses Flavour 7 p97 1976).
G Hellekant, D Glaser, J N Brouwer and H van der Wel. Gustatory effects of miraculin,
monellin and thaumatin in the Sanguinus Midas Tarmarin monkey studied with
electrophysiological and behavioural techniques. Acta Physiol Scand 97 p24 1 (1976).
H van der We1 and W M Bel. Effect of acetylation and methylation on the sweetness
intensity of Thaumatin I. Chem Senses Flavour 2 p21 (1976).
J A Morris. Sweetening agents from natural sources. Lloydiu 39:1 p25 (1976).
W M Nicol. The Serendipity berry of Katemfe - natural protein sweeteners from West
Africa. Nigerian Field 16:4 p150 (1976).
J D Higginbotham and C A M Hough. In: Sensory Properties of Foods (G G Birch and
K J Parker, eds). Applied Science Publishers, London P129 (1977).

H van der Wel, H Wiesrma and J N Brouwer. Tritium labelling of Thaumatin I, a sweettasting protein from Thaumatococcus daniellii Beth, by reductive methylation J Labelled
Compounds and Radiophamaceuticals 14 p375 ( 1 978).
H W Hsu. D L Vavak. L D Slatterlee and G A Miller. A multi-enzyme technique for
estimation of protein digestibility. J Food Sci 425 p1269 (1977).
C A M Hough and J A Edwardson. Antibodies to thaumatin as a model of the sweet taste
receptor. Nature 271 p383 (1978).

R J Summerfield. B H Most and M Boxall. Tropical plants with sweetening properties:

Physiological and agronomic problems of protected cropping. 2. Thaumatococcus daniellii,
Econ Bot 32 p321 (1978).
H van der We1 and W J Bel. Structural investigation of sweet-tasting proteins thaumatin
and monellin by immunological studies. Chem Senses Flavour 3:l p99 (1978).

191

Flavou rs/Foods

H van der We1 and K Arvidson. Qualitative psychophysical studies on the gustatory effect
of the sweet-tasting proteins thaumatin and monellin. Chem Senses Flavour 3:3 p291
(1978).

K J Parker. Alternatives to sugar. Nature 9 p271 (1978).

D Glaser, G Helekant, J N Brouwer and H van der Wel. The taste responses in primates to
the proteins thaumatin and monellin and their phylogenetic implications. Folia Primatol
29:l p56 (1978).

R B Iyengar, P smith, F van der Wel, J van Brouwershaven, P Ravestein, G Richters and P
D van Wassenaar. The complete amino acid sequence of the sweet protein Thaumatin I.
Eur J Biochem 96 p193 (1979).

I C Onwuerme, B E Onochie and E A Sofowora. Cultivation of T daniellii the sweetener.

World Crops May/June p106 (1979).
J D Higginbotham. In: Development in Sweeteners I (C A M Hough, K J Parker, and A J
Vlitos, eds), Applied Science, London, p87 (1979).

J D Higginbotham. The Sweet Challenge. Spectrum 164 p7 and in French: Fruits 35: 11
p709 (1980).
J D Higginbotham. TALIN, a novel sweet protein from Thaumatococcus daniellii. In:
Health and Sugar Substitutes (B Guggenheim, S Karger, eds), Basel, p172 (1979).

T Ochi. Properties and applications of thaumatin, a new natural protein sweetener. New
Food Ind 2 2 5 pl(1980).
H vander Wel. Physiological action and structure characteristics of the sweet-tasting
proteins thaumatin and monellin. TIBS May p 122 (1980).

H van der We1 and W J Bel. Enzymatic properties of the sweet-tasting proteins thaumatin
and monellin after partial reduction Eur J Biochem 104:2 p4 13 ( 1980).
X D du Villard, H van der Wel, and J N Brouwer. Enhancement of the perceived sweetness
of sucrose in the art by thaumatin. Chem Senses Flavor 5 2 p93 (1980).
Anon. Proposed voluntary standards for thaumatin. Food Additive Federation News,
Japan 57 p 5702 (1980).
S Ohashi, Natural sweetener:
(1981).

Thaumatin.

Food Sanitation Research 31, Japan p83

T Ochi. Sweetening agents in chilled desserts. Thaumatin as a natural sweetening agent.

New Food Industry 23:2 p 13 ( 1981).
J D Higginbotham, M G Lindley and J P Stephens. Flavour enhancing properties of
TALIN protein sweetener (thaumatin). In: The Quality of Food and Beverages Vol 1. (G
E Inglett and G Charalambous, eds). Academic Press Inc, New York, p91 (1981).

192

Advances in Flavours and Fragrances

H van der Wel. The influence of chemical modifications of the sweet-tasting proteins
thaumatin and monellin on the sweetness intensity and character. In: 'The Quality of Food
and Beverages'. Vol 1. (G E Inglett and G Charalambous, eds). Academic Press Inc, New
York. P113 (1981).
S A Leach and R M Green. Reversal of fissure caries in the albino rat by sweetening
agents. Caries Res 15 p508 (1981).

L Edens, L Heslinga, R Klok, A M Ledeboer, J Man, M Y Toonen, C Visser and C T

Verrips. Cloning of c-DNA encoding the sweet tasting plant protein thaumatin and its
expression in Escherichia coli, Gene 18 pl (1982).
J D Higginbotham, D J Snodin and J W Daniel. TALIN protein sweetener, toxicology and
safety assessment. Abstracts of 1982 Annual Conference, Society of Toxicology. The
Toxicologist 2:1 p 176 ( 1982).
K K Eaton, J D Higginbotham, D J Snodin, J W Daniel and D R Stanworth. Oral and
inhalant allergenicity of TALIN protein, a water soluble protein, Abstracts of XI
International Congress on Allerology London 17-22"d October ( 1982).
S Ohashi and T Ochi. Flavour enhancing effect of the protein sweetener thaumatin on
fermented milk. New Food Industry 24 p5 (1982).

H van der Wel. Structural modification of sweet proteins and its influence on sensory
properties. Chem Ind 19 (1983).
J D Higginbotham, D J Snodin, K K Eaton and J W Daniel. Safety evaluation of thaumatin
(TALIN protein). Fd Chem Toxic 21:6 p 185-823 (1983).
J D Higginbotham. Recent developments in non-nutritive sweeteners. In: 'Developments
in Sweeteners I1 (T J Grenby, K J Parker and M G Lindley, eds). Applied Science, London
(1983).

J P Stephens. TALIN protein - a versatile new ingredient. Food March p12 (1983).
Twenty-seventh Report of the Joint FAO/WHO Expert Committee on Food Additives,
World Health Organisation Report Series 696 (1983).
H van der Wel, R G Iyengar, J van Brouwershaven, P D Van Wassenaar, W J Be1 and F J
G van der Ouderaa. Assignment of the disulphide bonds in the sweet-tasting protein
Thaumatin I. Eur J Biochem 144 p41-45 (1984).
L Edens, I Bom, A M Ledeboer, J Maat, M Y Toonen, C Visser and C T Verrips.
Synthesis and processing of the plant protein thaumatin in yeast. Cell 37 p629-633 (1984).
A M Ledeboer, C P Verrips and B M M Dekker. Cloning of the natural gene for the
sweet-tasting plant protein thaumatin. Gene 30 p23-32 (1984).

FlavourdFoods

193

A M de Vos, M Hatada, H van der Wel, H Krabbendam, A F Peerdeman and Sung-Hou

b m . Three-dimensional structure of Thaumatin I, an intensely sweet protein. Proc Nut1
Acad Sci USA Biochemistry 82 p1406-1409 (1985).
A Mackenzie. J B Pridham and N A Saunders. Changes in the sweet proteins (Thaumatin)
in Thaumatococcus daniellii fruits during development. Phytochemistry 24:ll p25032506 (1985).
H Rogatka, M Naim and U Zehari. Evidence for taste persistence of sweeteners in baboon
monkeys: A proposed experimental model. Chemical Senses 10:3 (1985).
S Ohashi and T Ochi. The flavour-enhancing effect of the natural sweetener thaumatin.
New Food Industry 27:3 (1985).

J Higginbotham. TALIN revisited Food Sept. 1986. P39-40.

J D Higginbotham. TALIN protein, thaumatin. In: Alternative Sweeteners (L OBrien
Nabors and R C Gelardi, eds). Marcel Dekker, New York (1986).
J Foster, Just a Spoonful of TALIN.. . Pig Farming November 1986, p24-25.
T H Grenby and M G Saldanna. Studies on the Inhibitory Action of Intense Sweeteners on
oral Micro-organisms relating to Dental Health Caries Res, 20 p7-16 (1986).

B J C Cornelissen, R A M Hooft, van Huijsduijen and J F Bol. A Tobacco Mosaic virusinduced Tobaccus Protein is Homologous to Thaumatin. Nature 321 29 May p531-532
(1986).
M Cusack, R J Beynon and P S Rogers, Cysteine Proteinase Activity in preparations of
Thaumatin. Biochem, SOC, Trans, 15 p880 (1987).

M Richardson, S Valdes-Rodiguez and A Blanco-Labra. A possible function for

Thaumatin and a TMV-Induced protein suggested by homology to a maize inhibitor.
Nature 327 4 June p432-434 (1987).
J Foster. Sweet Enough to Drink. Pig Farming Jan 1988.
M Witty. Thaumatin 11, a Simple Marker Gene for use in Plants. Nucleic Acids Res. 17
No 8 p3319 (1989).

STABILITY OF THIOLS IN AN AQUEOUS PROCESS FLAVOUR

Chris Winkel, Paul B. van Seeventer, Hugo Weenen and Josef Kerler
Quest International, Huizerstraatweg 28, 1411 GP Naarden, the Netherlands

1 INTRODUCTION
Process flavourings, which are generally prepared by thermal treatment of an aqueous
solution of reducing sugars and amino acids, show considerable flavour instability in
aqueous solution. To overcome this instability, process flavours are generally spray-dried.
The resulting powder has sufficient flavour stability, if temperature and humidity do not
get too high. The stability in water is so limited however, that a great deal of effort in
planning is required, to prevent significant flavour loss, before the product is spray-dried.
Especially thiols, which are essential for a good meaty character in savoury process
flavours, are particularly unstable. This urged us to look in more detail at flavour instability
of a model thiol containing process flavour in aqueous solution.
Thermal treatment of a solution of ribose and cysteine results in the formation of several
specific character impact components with a meaty overall impression [Hofmann, 1995a and
1995b; Schieberle and Hofmann, 19961. The use of cysteine in combination with a pentose is
responsible for the formation of very potent sulphur containing meat flavour compounds, such
as 2-methyl-3-furanthi01 (MFT), 3-mercapto-2-pentanone (3MP), 2-mercapto-3-butanone
(MB), and 2-furfurylthiol (FFT)[Hofmann, 1995a and 1995bl. Some of these compounds
were similar to those identified in beef [Gasser and Grosch, 19881 and were recently
quantified in several types of heated meat [Kerscher and Grosch, 19981. In the study
described here the ribose + cysteine based process flavour reported by Hofmann was used
(1995a&b), as a model for meaty process flavours. As thiols are known for their
reactivityhnstability such as oxidation and involvement in nucleophilic or radical reactions,
we assumed that the degradation of these compounds is a likely cause of the limited shelf-life
or flavour instability of savoury process flavours.
The interaction of thiol and disulphide flavour compounds with food components has
been studied by Mottram (1996), which indicated a.0. that these compounds undergo redox
reactions with proteins. Hofmann (1996) and Guth (1995) reported model studies on the
oxidative stability of odour-active thiols and disulphides, respectively. Eisenreich ( 1994;
1995) has reported studies on antioxidative activities of volatile sulphur-containing
heterocyclic compounds confirming their reactivity in free radical oxidation reactions.
The study described in this paper was designed to determine the (in)stability of the
most important character impact components in an aqueous solution of the model process
flavour described by Hofmann (1995a&b), and to investigate why some of the character

Flavours/Foods

195

impact components are unstable. A more extensive study has been published recently by
Seeventer (2001).
2 EXPERIMENTAL PROCEDURES

2.1 Materials
2-Methyl-3-furanthiol and 2-mercapto-3-butanone (10% solution in triacetin) were obtained
from Oxford Chemicals (Cleveland, U.K.), 2-furfurylthiol, maltol, and D(-)-ribose from
Aldrich (Steinheim, Germany), L-cysteine and 2,5-dimethyl-3-methylfuranthiol
from Quest
International (Naarden, The Netherlands)
2.2 Analytical instruments
A Fisons 8000 series gas chromatograph type 8130, equipped with a HP-5 capillary
column (50 m x 0.32 mm; 1.05 pm film), a flame ionisation detector (FID), and a Fisons
AS 800 autosampler was used for GC analysis. The GC was run with an injector
temperature of 225 "C and detector temperature of 250 "C. The oven temperature was
programmed from 65 "C to 120 "C at 3 " C h i n and from 120 to 250 "C at 40 "C/min.
Chrom-Card for Windows (Fisons Instruments) was used for data handling. HRGC mass
spectrometry was performed using a Finnigan MAT TSQ 70 mass spectrometer in the
electron impact mode at 70 eV. Analysis of cysteine using capillary electrophoresis was
performed with a Hewlett Packard 3D CE.

2.3 Procedure for the preparation of the aqueous model process flavouring
A solution of 180 mmol D-ribose and 60 mmol L-cysteine in 1800 ml phosphate buffer (0.5
M, pH 5.0) was heated in a two litre autoclave from room temperature to 130 "C in 10 min
and was kept at 130 "C for 20 min, followed by rapid cooling with tap water.

2.4 Storage of the aqueous model process flavouring

Portions of 100 ml were stored in the dark at 50 "C in closed bottles both under air and argon
atmosphere. At certain time intervals, after addition of maltol as the internal standard, the
aqueous reaction mixtures were extracted with dichloromethane-diethylether 7/3 v/v (20 ml).
The extract was dried (Na*SO4), carefully concentrated to a volume of 2 ml at room
temperature and at a pressure of 200 mbar using a rotary evaporator, and finally concentrated
under a stream of argon to 1.0 ml and subjected to HRGC analysis.

2.5 Storage of reconstituted model process flavourings

A stock solution in ethanol of 2-methyl-3-furanthio1, 2-mercapto-3-butanone, and 2furfurylthiol was added to an aqueous phosphate buffered solution (0.5M; pH 5.0) to obtain a
final concentration of 50 pM of each of these components. This is a concentration
approximately 10 fold higher than found in the aqueous model of Hofmann (1995a and
1995b). Either ribose or/and cysteine were added in various concentrations, typically in the
range of 25-250 mM. Volumes of 5 ml of these solutions were stored in the dark at 50 "C in
closed glass containers (15 ml) under an air atmosphere. At certain time intervals, after

196

Advances in Flavours and Fragrances

addition of maltol as the internal standard, the aqueous solutions were extracted with
dichloromethane-diethylether 7/3 v/v (2.0 ml). The extract was dried (Na2SO4) and analysed
by HRGC.
All plots of relative area against storage time showed a linear decrease, indicating zeroorder or pseudo-zero order kinetics for the compounds investigated. In all storage experiments
described in this report, zero-order models were used based on visual assessment and the
coefficient of determination (R2) obtained from regression analysis. The relative decrease in
the starting concentration was calculated by linear regression, using the least squares method.

2.6 Deuterium exchange experiment with 2-methyl-3-furanthiol and bis(3-furanyl)

disulphide
'H-NMR was used to follow the deuterium exchange of 2-methyl-3-furanthiol and bis(3furanyl) disulphide at room temperature in CH30D solution containing DCl during several
days.

3 RESULTS AND DISCUSSION

3.1 Flavour stability of the aqueous model process flavouring
The stability in an accelerated storage test at 50 "C of some important flavour compounds in
the aqueous model process flavouring [Hofmann, 1995a and 1995b; Schieberle and
Hofmann, 19961 is shown in table 1 (entry 1). 2-Methyl-3-furanthiol (MFT) was found to
be the most unstable component. Furthermore, an interesting observation was the fact that
there was almost no difference between storage of the aqueous model process flavouring
under air or argon atmosphere (data not presented here), probably due to an anti-oxidative
effect of the matrix. The anti-oxidative properties of the matrix components cysteine and
products formed in the Maillard reaction are well known. Entry 2 shows the results
obtained from a storage experiment of a reconstituted mixture of flavour compounds in
aqueous solution, using concentrations which are approximately tenfold higher than in the
original process flavour. The concentrations were increased to allow more facile analysis of
the components of interest.
We anticipated that the presence of residual cysteine andor ribose could have an effect
on the stability of some of the character impact components of the model process flavour. The
effect of additional cysteine and ribose on the stability of the flavour impact components was
therefore investigated in the reconstituted mixture (Table1, entries 3-5). Residual cysteine and
ribose concentrations were determined in a freshly prepared model process flavour, and were
found to be 15 mM and 25 mM, respectively.
Entries 3 and 4 of table 1 show the results obtained when the reconstituted mixture was
exposed to accelerated storage conditions with addition of ribose and cysteine respectively.
Comparison of entry 2 with entry 3, shows that ribose addition has no effect on stability. A
stabilising effect of the addition of cysteine was clearly present (entry 4). Comparison of entry
4 with entry 5 shows that, when both ribose and cysteine are present (entry 5), an even larger
stabilising effect on MFT and the furanones can be observed, than with cysteine alone.

FlavourdFoods

197
I'

2b

3b

4b

Sh

model proces
flavouring

no addition

+ 250 mM

+ 150mM

+ 150mM

ribose

cysteine

cysteine

+ 250 mM
ribose
Compound

Decrease in concentration (70per day)

2-mercapto-3 -butanone (MB)

< 20

> 90

>90

< 30

< 30

2-methyl-3-furanthi01 (MFT)

< 60

>90

>90

< 80

< 50

2-furfurylthiol (FIT)

n.d.'

>90

> 90

< 50

< 50

According to the procedure of Hofmann (1995a and 1995b): a phosphate buffered (0.5 M, pH 5.0) solution of D-ribose (100 mM) and L
cysteine (33 mM) was heated in an autoclave from room temperature to 130 "C in 10 min and was kept at 130 "C for 20 min. Character
impact components were formed in the order of concentrations of around 5 pM.
h50FM of each character impact component present in aqueous phosphate buffered solution (0.5 M; pH 5.0).
Not determined.
A

Table 1. Stability at 50 "Cof jive character impact components of the aqueous model process
flavour (entry I ) and reconstituted models (entry 2-5): influence of ribose and
cysteine on the decrease in concentration (% per day).
Because of the stabilising effect of cysteine on the model process flavouring, the stability
of the reconstituted aqueous mixture with character impact components in concentrations of
50 pM was investigated at various cysteine concentrations. Figure 1 shows the stability of 3mercapto-2-butanone (MB), 2-furfurylthiol (FFT), and 2-methyl-3-furanthiol (MFT') at
various cysteine concentrations. Interestingly the decrease in concentration of MFT versus the
cysteine concentration has a clear minimum, which is different from what is observed for
MB and FFT. The stability of MFT shows an optimum at a cysteine concentration of around
50 mM.
decrease in concentration
(% per day)

60
40
MB
201

25

50

- s

o !
0

100
125
[cysteine] (mM)

75

150

175

I
200

Figure 1. Stability of a reconstituted model process flavouring: influence of [cysteine] at pH

5.0 (0.5M phosphate).
Based on these results we suggest that the observed stabilisation of MFT by additional
ribose (see table 1; entries 4 and 5 ) was caused by the decrease in the cysteine concentration
due to the formation of the thiazolidinecarboxylic acid derived from cysteine and ribose. The

Advances in Flavours and 'Fragrances

198

actual concentration of cysteine after addition of ribose (250 mM) will be lower than 150
mM, as a result of the formation of this thiazolidinecarboxylic acid.
The results shown in figure 1 suggest that the mechanism of degradation of MFT in the
presence of cysteine is slightly different from that of MB and FFT.Apart from an oxidative
degradation pathway via (mixed) disulphide formation, another pathway leading to nonvolatile or polar degradation products by reaction of MFT with a.0. cysteine may be possible,
which would account for the decreased stability of MFT at higher cysteine concentration.
Evidence for the exceptional reactivity of MFT came from the storage of a mixture of
only the three thiols MB, FFT,and MFT in phosphate buffered solution (0.5 M; pH 5.0) in the
absence of cysteine. The amount of MB and FFT at the beginning was equal to the total
amount of both thols and (mixed) disulphides at the end of storage. This was absolutely not
the case for MFT. After storage a large amount of MFT was not detected, neither as thiol nor
as (mixed) dmlphide, using GC-analysis. When the same experiment was performed with
2,5-dimethylfuranthiol(DMFT) instead of MFT, the total amount of DMFT at the beginning
was equal to the total amount of both DMFT and (mixed) disulphides at the end of storage
(data not presented).
SH

'

SH

'

,SH

Dimers type 5-C?

Dimers type 5-S?

Trimers and polymers ?

Figure 2. Proposed mechanismfor polymerisation of 2-methyl-3-furanthiol

We think this can be explained by the greater ability of MFT to undergo

oligomerisation/polymerisation.We know that M l T easily undergoes polymerisation, when
stored neat at ambient temperature, as some of the commercial samples already contain
considerable amounts of polymeric material. Therefore, we hereby propose a hypothetical
mechanism for the polymerisation of MFT as shown in Figure 2, which can explain these
observations. Protonation at the 2-position leads to an electrophilic species, which would
easily react with a nucleophile. As the thiol group of MFT is a good nuclophlle, protonation
would catalyse dimerisatiodpolymerisation.Apart from the reaction of MFT with itself, other
thiols, when present in a large enough concentration, could react as well, e.g. cysteine. This
mechanism could explain the different effect of the cysteine concentration on the stability of

199

FLavours/Foods

MFT in comparison to other thiols such as MB and FFT (Figure 1). In the case of DMFT,
polymerisation at the 5-position would be more difficult due to steric hindrance caused by the
5-methyl group.
If the left part of the mechanism is of practical importance, coupling with other thiols
would be expected as well. Since in the reconstituted mixture experiments no loss of the other
thiols was observed, we must conclude that this part of the proposed mechanism is not
significant.
3.2 Deuterium exchange experiment with MFT and bis(3-furanyl)disulphide
To confirm our hypothesis of the degradation mechanism of MFT in aqueous solution, a
deuterium exchange experiment with MFT was performed. Using 'H-NMR, some protondeuterium exchange was observed at the 4-position, but to a much larger extent at the 5position (fig. 3) when MFT was kept in CH30D solution containing DC1 at room temperature.
This is in agreement with the preference of electrophilic substitution at the position next to the
hetero-atom of aromatic heterocycles. However, except for the formation of the
corresponding disulphide, no evidence for the formation of dimers was found, as was
expected according to the mechanism in figure 2. An explanation for this could be that in the
model NMR experiment a methanolic solution had to be used instead of an aqueous
phosphate buffered (0.5 M; pH 5) solution.

4-position: H/D-exchange of 10%

5-position: HID-exchange of 85%

(thioVdisulfide 5-611)

Figure 3. Deuterium exchange experiment with 2-methyl-3-furanthiol in CHjlODDCl during

64 days at room temperature.
To obtain deuterium exchange data for the 2-positon of 3-furanthiols, the synthesis of 3furanthiol was synthesised (Hofmann, 1995a) according to figure 4. Deuterium exchange
experiments with t h s compound could provide evidence for the proposed mechanism of the
degradation of MFT (fig. 2). Some difficulties were encountered during the isolation of 3furanthiol. The purity of the isolated compound was not satisfactory ('H-NMR), possibly
caused by the inherent instability of 3-furanthiol.

3-furanthiol

Figure 4. Synthesis of bis(3-furanyl)disulphide.

bis(3-furanyl) disulfide

Advances in Flavours and Fragrances

200

However, bis(3-furanyl) disulphide was isolated in a pure form and was used for the
deuterium exchange experiment. The high exchange rate at the 2- and 2-positions was
remarkable, when compared to that of the other positions (fig. 5). The preference for
exchange at the 2- and 2-positions is the result of the directing and activating effect of the Satom linked to C-3 and C-3. These results suggest that 2-protonation is facile in MFT, and is
possibly a crucial step in the degradation of MFT as shown in figure 2. Addition of
nucleophiles to the 5-position of protonated MFT will most likely proceed via an ionic
mechanism. Nevertheless, other mechanisms ( e g radical mechanism) could play a role in the
degradation of MFT.
Bis(3-furanyl)disulfide: deuterium exchange in CKOD/DCI (8 days, RT):
2,T-position: H/D-exchange of 71%
55-position: H/D-exchange of 10%

Figure 5. Deuterium exchange experiment with bis(3-furanyl) disulphide.

CONCLUSIONS

The data presented indicates that in meat process flavours based on cysteine and ribose, 2methyl-3-furanthiol is the one of least stable character impact components. The instability
is not due to disulphide formation, but appears to result from electrophilic coupling
reactions, resulting in products which cannot be detected by GC. H-D exchange showed
that H-5 exchanges much faster than any other proton in MFT, suggesting that the
reactivity of the 5-position is responsible for the rapid decomposition of MFT. This is in
agreement with the lower stability of MFT in comparison with 2,5-dimethyl-3-furnathiol.
Residual cysteine stabilises MFT and other thiols.
References
Eisenreich, J.P.; Shibamoto, T. Antioxidative activity of volatile heterocyclic compounds. J.
Agric. Food Chem., 1994,42, 1060-1063.
Eisenreich, J.P.; Wong, J.W.; Shibamoto, T. Antioxidative activities of furan- and
thiophenethiols measured in lipid peroxidation systems and by tyrosyl radical scavenging
assay. J. Agric. Food Chem., 1995,43,647-650.
Gasser, U.; Grosch, W. Identification of volatile flavour compounds with hgh aroma values
from cooked beef. Z. Lebebnsm. Unters. Forsch. 1988, 186,489-494.
Guth, H.; Hofmann, T.; Schieberle, P.; Grosch, W. Model reactions on the stability of
disulphides in heated foods. J. Agric. Food Chern., 1995,43,2199-2203.
Hofmann, T.; Charakterisierung intensiver Geruchstoffe in KohlenhydraKysteinModelreaktionen und Klikung von Bildungswegen. Thesis, Technische Universitut Miinchen,
1995a.
Hofmann, T.; Schieberle, P. Evaluation of the key odourants in a thermally treated solution of
ribose and cysteine by aroma extract ddution techniques J. Agric. Food Chem., 1995b, 43,
2 187-2194.
Hofmann, T.; Schleberle, P.; Grosch, W. Model studies on the oxidative stability of odour
active thiols occurring in food flavours. J. Agric. Food Chem., 1996,44,251-255.

Flavours/Foods

20 1

Kerscher, R; Grosch, W. Quantification of 2-methyl-3-furanthiol, 3-mercapto-2-pentanone,

and 2-mercapto-3-pentanone in heated meat. J. Agric. Food Chem., 1998,46, 1954-1958.
Mottram, D.S.; Szauman-Szumski, C.; Dodson, A. Interaction of tho1 and disulphlde flavour
compounds with food components. J. Agric. Food Chern., 1996,44,2349-2351.
Schieberle, P.; Hofmann, T. Untersuchungen zum EinfluB von Herstellungsparametem auf
den Aromabeitrag intensiver Aromastoffe in CysteidKohlenhydrat- Reaktionsmischungen.
LebensrnitteZchernie, 1996,50, 105-108.
Seeventer, P.B. van; Weenen, H.; Winkel, C.; Kerler J. Stability of thiols in an aqueous
process flavour. J. Agric. Food Chern., 2001,49,

HIGH IMPACT AROMA CHEMICALS

D J Rowe
Oxford Chemicals Limited, North Gare, Seaton Carew, Hartlepool, TS25 2DT, UK

1 INTRODUCTION
The demand for high impact chemicals has been driven by the twin engines of increased
consumer sophistication in the market for flavours and by improvements in the analytical
techniques needed to identify character impact molecules. Consumers are no longer
content simply with apple and orange but look to passion fruit and mango and the
availability of the fruit itself means that flavour must closely match the real thing.
Fortunately, developments in the flavour and fragrance industry have gone hand-in-hand
with advances in the chemical sciences. From the nineteenth century, which saw the
identification and synthesis of key materials such as cinnamaldehyde and vanillin, to more
recent decades, advances such as the so-called hyphenated techniques, in particular GCMS and GC-Olfactometry, have enabled flavour chemists to identify the compounds
present in natural materials. Some of these compounds, though present at only trace levels,
are key contributors to the odour and flavour of natural materials. This has been augmented
recently by the use of Solid Phase Microextraction (SPME) to capture the aroma chemicals
at source, such as the IFF Living Flower@ and Living FlavourBf technologies,
Givaudans Taste-TrekTMstudies on aroma chemicals emitted by plants in the rain forest
canopy and IFFBBAs Generessence@ programme. Many of the materials identified are
the high impact aroma chemicals which will be discussed in this article.
The term high impact aroma chemical is one which many of us can understand but
for which there is no dictionary definition. I will set four key criteria which, for the
purpose of discussion, will constitute a high impact aroma chemical ;
Low odour threshold. This is an obvious feature but there is no absolute definition of
low to which we can turn! So for the purpose here, I have set low odour threshold to be
less than ten parts per billion (10 ppb, or 10 in lo9).Some apparently odorous compounds
fall out by this definition; for example 2,3,5-trimethylpyrazine 1 has an odour threshold of
around 1000 ppb and cannot be considered high impact. However the 2-alkoxy-3alkypyrazines such as 2-methoxy-3-methylpyrazine2, which has an odour threshold of
only 5 ppb, would constitute a high impact material by our definition.

Flavours/Foods

203

Figure 1

Contrasting pyrazines

Character impact. The material should have recognisable character, even at the low
levels that such a material would be used at. This criterion eliminates many esters, such as
ethyl 2-methylbutyrate 3, which has an odour threshold of only 0.1 ppb, but at low levels
has only a vague fruitiness, which may be pleasant but not characteristic,

Figure 2

Esters, odourous but low impact?

Desirable character impact. Although many chemicals have a powerful odour this
odour is not always a desired one! For example 2,4,6-trichloroanisole 4 is a highly odorous
metabolite of a fungus that attacks paper. However it is highly improbable that this can be
considered a high impact aroma chemical as it is unlikely that a flavourist or perfumer
will have a brief to recreate the aroma or taste of mouldy books!

CI

Figure 3

An unwanted odour

This is of course a matter of context, as many aroma chemicals are repellent when
neat or in high concentration, but in the correct context contribute to the desired effect. For
example the nature of 4-mercapto-4-methyl-2-pentanone5 can be gleaned from its
common name of Cat Ketone; however, it is also a key component of sauvignon grape2.
Depending on the context it can be used to re-create the bouquet of a fine Cabernet
Sauvignon or the atmosphere of where the local alley cats have marked their territory!

204

Advances in Flavours and Fragrances

SH

Figure 4

Chat ou vin?

Available to the flavourist. There are three aspects here:

Regulatory. The material should not be forbidden in the context it is to be used; for
flavour use the material should be nature identical, preferably FEMA GRAS. Natural status
may be important, especially for the USA where the nature identical category is lacking.
Stability. Many materials can be manufactured but have insufficient shelf-life to be
useful in a finished flavour. For example cis-3-hexenal 6 is a key aroma chemical emitted
by cut grass and other vegetation; it has a low odour threshold (0.25 ppb) and a powerful,
desirable green character. Unfortunately it is chemically reactive, and readily rearranges to
the more stable conjugated form, giving the more familiar trans-2-hexenal 7. Even the
half-life of the isolated material may not be enough to guarantee a useful level of
stability. The rearrangement is prototropic and hence catalysed by both acid and base; the
rate will be increased by a factor of ten for each pH unit away from its optimum stability
point.

Figure 5

Rearrangement of hexenals

Economics. Many high impact chemicals are relatively expensive, which reflects the
small market volume and the difficulties associated with manufacturing and handling such
materials. However, the material must be commercially available at a price which enables
a flavourist or perfumer to add value to their formulation by its use; if the material is
captive, the internal costings must not be prohibitive. In short, despite high prices, the
high impact of these materials gives more fizz for your buck. An example of this can be
seen from garlic chemistry. The major component of garlic oil is ally1 disulphide 8 (2propenyl disulphide). The isomeric 1-propenyl disulphide 9 is also present (cis- and transforms) but whereas the former is readily synthesised, and hence cheap and readily
available, no suitable route for large-scale preparation of the latter exists at present.
Laboratory syntheses have been reported, but the costs of material made in such a way
mean that any advantage in the flavour is outweighed by a vast increase in costs; the added
value is insufficient.

Fla vou rs/Foods

205

Figure 6

Contrasting garlic disulphides

For the purpose of this discussion, high impact chemicals are those which fulfil these
criteria. Such compounds are powerful materials, highly active at low levels; their uplifting
effect makes them the Viagra@of the Flavor and Fragrance Industry.

1.1 A note on identification of High Impact Aroma Chemicals

It was noted in the introduction that new analytical techniques have enabled chemists to
identify aroma chemicals down to lower and lower levels. It is a tribute to the remarkable
ingenuity of chemists of the 19hand early 20thcenturies that it was possible to identify any
materials, when the only techniques available were elemental analysis by combustion,
molecular weight determination by freezing point depression / boiling point elevation, and
proof by synthesis. At the risk of using a clich6, people today have it easy! As can be
seen on the time line below, the identification of many powerful aroma chemicals took
place during the sixties and seventies, as GC and GCMS became available.

PURIFICATION BY DISTILLATION
& RECRYSTALLISATION; FEW
REFERENCE COMP0UNDS;kgr OF
MATERIAL NEEDED

1892 ALLYL DISULPHIDE

BASICS OF ORGANIC CHEMISTRY
ESTABLISHED; REFERENCE
COMPOUNDS AVAILABLE; 100gr

1900

1 9 0 9 METHYL SULPHIDE
1925 METHYL MERCAPTAN

UV SPECTROSCOPY
1 9 5 7 METHYL PROPYL DISULPHIDE
I R SPECTROSCOPY, PAPER &
COLUMN CHROMATOGRAPHY

l g 5 O - r
1964 2-METHYLTETRAHYDROFURAN-3-ONE

NMR SPECTROSCOPY, GAS

CHROMATOGRAPHY; mgs
FT NMR;,

1969 2-ISOBUTYLTHIALOLE
1 9 7 1 THIOMENTHONE 1973 2-ACETYLPYRAZINE
1974 TETRAHYDROQUINOXALINE

GCMS; pgs

GC-OLFACTOMETRY

1982 GRAPEFRUIT MERCAPTAN

SOLID-PHASE
MICROEXTRACTION; ngs

zooo

2000 3-MERCAPTO-2-METHYLPENTAN-1-OL
2 0 0 1 2,4,6-TRITHIAHEPTANE

Figure 7 A Time-Linefor high impact aroma chemicals

206

Advances in Flavours and Fragrances

2 USES
High Impact Chemicals have been identified in many foodstuffs and have many
applications, to the extent that a simple listing would make dull reading. However, many
key flavour notes are associated with high impact aroma chemicals, as illustrated by the
flavour wheel below (figure S), running from the sweet and fruity notes through to
savoury and alliaceous. The range of chemical functionalities is also wide, with sulphurcontaining molecules particularly prominent. The order in the wheel is not random; the
East - West division is largely between sweet and savoury. Clockwise from truffle to
vegetable, the materials are formed by biogenesis in plants, whereas most of the
remainder, in the Southern part of the wheel, are Maillard products formed in cooked
foods. The materials in the Eastern sector of the wheel, from green, grassy to smoky,
are of interest to the perfumer as well.

Figure 8

A Flavour Wheelfor high impact aroma chemicals

Fluvou rs/Foods

2.1

207

Green, grassy

Here the traditional molecules are the hexenyl compounds, the so-called Cg wound
compounds, produced by all green tissues as a consequence of the peroxide defence
mechanism, where linoleic acid acts as trap for peroxy radicals and is in turn cleaved.
As noted above, the true high impact chemical in this group is the unstable cis-3-hexenal 6,
the initial cleavage product of linoleic acid. The stable aroma chemicals trans-2-hexenal 7
(leaf aldehyde) and cis-3-hexenol 10 (leaf alcohol) are widely used; with odour thresholds
of 17 and 70 ppb respectively, they are borderline cases as high impact chemicals, though
their distinctive character is in their favour. A fresh greenness is also associated with the
more odorous 2-isobutylthiazole 11 (odour threshold 3 ppb). This molecule is released by
tomato vine and has both tomato and more general green (string bean, geranium leaf)
character, especially on dilution.

A.

HO--10

11

Figure 9
2.2

Green, grassy compounds

Herbaceous

This is a very wide-ranging concept with a large overlap with the fragrance area. A series
of thioesters, typified by sec-butyl 3-methylbut-2-enethioate 12 are found in galbanum oil,
and the key odourants of coriander are long-chain unsaturated aldehydes such as trans-2dodecenal 13.

208

Advances in Flavours and Fragrances

H

13

12

Figure 10 Molecules with herbaceous odours

2.3

Spicy

Spicy can refer to both strongly-flavoured or hot. For the former, an excellent example
is the fenugreek lactone or sotolone 14, which has an intense and very persistent curry
aroma - its persistence includes a remarkable tendency to linger on the skin should one
have accidental contact with it! For hot or piquant we have capsaicin 15; whilst its
effect is as a sensation via the trigenimal system, the intensity of its effect on the tongue
(or any other part of the body it has contact with!) is sufficiently intense to consider it to be
a high impact chemical!

14

15

Figure 11 Spicy chemicals

2.4 Fruity, ester-like

The esters are obvious candidates here, but as noted above whilst many have low odour
thresholds (ethyl butyrate 1 ppb, ethyl isobutyrate 0.1 ppb, ethyl 2-methylbutyrate 3 0.1
ppb, ethyl hexanoate 1-3 ppb), they lack the character which would make them truly high
impact. Instead we may illustrate fruity notes with what may be the ultimate high impact
aroma chemical, p - 1-menthen-8-thiol 16, the Grapefruit mercaptan. This has the
remarkably low threshold of lo- ppb, and retains its character even at low levels. At
high concentrations, the molecule simply has a sulphurous, almost rubbery odour common
to mercaptans, and requires dilution to at least 0.001% before the fresh grapefruit uice
character can be recognised. As an illustration of what an odour threshold of 10-.? ppb
means, one tonne of grapefruit mercaptan could odourise 100,000 km3 of water, a Great
Lakes volume of liquid!

Fkuvours/Foods

209

3
16

Figure 12 Fruity molecules

2.5

Tropical

This is one of the most important areas for high impact aroma chemicals. Analysis of
passionfruit and durian has shown the presence of many powerful sulphur compounds, a
large number of which were included in FEMAs GRAS 18 list in 1998. Possibly the best
known is tropathiane, 2-methyl-4-propyl- 1,3-0xathiane 17,(odour threshold -3 ppb) ; 3mercapto-l-hexanol 18 and a number of acylated derivatives were included in FEMAs
GRAS 18 list, as were thioesters such as the thiohexanoate 19 and thioisovalerate 20.

17

18

H3csiby

19

20

Figure 13 Sulphur compounds for tropical notes

2.6

Blackcurrant

This is a very popular flavour in Europe associated with many health-related products
(nutraceuticals or functional foods) and with alcoholic drinks (Cassis liqueur, and added
as a cordial to some spirits). The key material in blackcurrant is 2-methoxy-4-methyl-4butanethiol 21; it is also a key component contributing a fruity flavour to olive oil3. Two
other materials have been used to recreate the rather catty note of blackcurrant; pmenthathiolone 22, the main odour-active ingredient of Buchu leaf oil, and 4-mercapto-4methyl-2-pentanone 5 , the Cat Ketone mentioned earlier;

210

Advances in Fluvours and Fragrances

21

22

Figure 14 Catty mercaptans for blackcurrant

2.7

Vegetable

This is obviously rather a large category! A compound of major importance is the

ubiquitous dimethyl sulphide 23 (DMS, methyl sulphide, odour threshold -3 ppb). When
pure, this has a clean, crisp sweetcorn odour. Some material on the market lacks this note
and has unpleasant, sulphurous, rotten cabbage odours; GCMS on such material has shown
the presence of dimethyl disulphide and methyl ethyl sulphide. Other powerful compounds
for vegetable note are 3-methylthiopropanal 24 (methional, odour threshold 0.2 ppb) and
its homologue 3-methylthiobutanal 25. On a more specific note, we should mention 2isobutyl-3-methoxypyrazine26,the bell pepper, main character impact compound found
in green or bell peppers, with its very low odour threshold of 0.002 ppb.

23

24

25

26

Figure 15 Molecules for vegetable aromas

2.8

Nutty

Nuttiness is associated with pyrazines; whilst it is part of th characte of almost all

pyrazines, it is particularly associated with the higher pyrazines such as
methyldihydrocyclopentapyrazine 27 (Maple lactone pyrazine) and 5,6,7,8tetrahydroquinoxaline 28 (THQ). 2-Acetylpyrazine 29 is very reminiscent of popcorn;
whilst its odour threshold is rather high at 62ppb, its persistent character earns it
membership of the high impact club. Some pyrazines are present in the raw nut,
whereas others are formed in roasted nuts by the Maillard reaction between amino-acids

21 1

Fluvours/Foods

and sugars. This also generates furans, such as 5-Methylfurfural 30, with an almond,
marzipan aroma.
0

27

29

28

30

Figure 16 Nutty aroma chemicals

2.9

Caramel

Caramelisation occurs on the heating of sugars and carbohydrates. This is the typical
Maillard reaction sequence, which generates furans with caramel odours. The ubiquitous
hydroxydimethylfuranone 31 has a sweet, cotton-candy aroma and a low odour threshold
of 0.04ppb. 2-Methyltetrahydrofuran-3-one 32 (coffee furanone) is less odorous but has a
very pleasant, sweet caramel character.

31

32

Figure 17 Molecules for caramel odours

2.10 Woody, smoky
Guaiacols are very important in this area. 4-Ethyl- and 4-methylguaiacols, 33 and 34, have
rather phenolic, medicinal odours with threshold value of 90 and 50 ppb respectively, but
more important is 4-vinylguaiacol 35 (2-rnethoxy-4-vinylpheno1,
MVP). This has a spicy,
clove-like smokiness particularly associated with smoked ham, and a low odour threshold
of only 3 ppb. It is also available in a natural form.

212

Advances in Flavours and Fragrances

OH

33

34

35

Figure 18 Guaiacolsfor smoke flavours

2.11 Roasted, burnt

This sector is associated with cooked food; in this and the following sectors the high
impact chemicals are those produced in the Maillard Reaction. These are secondary
products formed from reaction of primary Maillard products such as Strecker aldehydes
with sulphur sources such as cysteine. Shieberle has used the term advanced or
extended Maillard products for such materials. For roasted and burnt notes derivatives of
furfuryl mercaptan 36 are paramount. The mercaptan itself, with an odour threshold of
O.OOSppb, was the first high impact aroma chemical to be identified. It exhibits one of the
classic phenomena associated with high impact chemicals, the change in the nature of the
odour with concentration. At low concentrations (0.01-0.5ppb) the material has a roasted,
coffee aroma, becoming burnt and sulphurous in the range 1-l0ppb. The neat material has
no coffee odour, only an unpleasant oily smell resembling gasoline. Derivatives of furfuryl
mercaptan tend to be somewhat less odorous; the disulphide 37 (dithiodimethylenedifuran)
is much less obnoxious, and the mixed disulphide furfuryl methyl disulphide 38 has a
pleasant sweet coffee (Mocha) aroma; the latter has an odour threshold of 0.04ppb.

37

36

38

Figure 19 Fu$uryl mercaptan derivatives for cofiee and roasted notes

2.12 Meaty, beefy

This is the province of 2-methylfuran-3-thiol 39 (MFT) and its derivatives. The thiol, its
disulphide 40, mixed disulphide 41 and thioether 42 have all been found in beef; the odour
threshold of the disulphide has been reported as being as low as 2 x
ppb, but our

213

Fla vours/Foods

experience of working with these materials indicates that this odour threshold may be due
to residual thiol. MFT itself has an initially rather chemical odour, becoming more meaty
on dilution. The disulphide has more recognisable character, a rich aged beef, prime rib
aroma. The GRAS 19 thioether 42 has more roasted character.
Other mercaptans have beef character. 3-Mercapto-2-butanone 43 and 3-mercapto-2pentanone 44 are commonly found in beef Maillard reactions; the latter has an odour
threshold of 0.7ppb.

40

39

+
41

42
SH

43

44

Figure 20 Sulphur compounds for beef

2.13 Other Meats

Whilst 2-methyl-3-furanthiol is important in other meats as well as beef, in particular pork,

other high impact chemicals also occur. Mercaptopropanone dimer 45 has an intense
chicken broth odour and the unsaturated aldehyde truns-2-truns-4-decadienal46 is very
reminiscent of chicken fat. The latter has been implicated in the observation that whereas
2-methyl-3-furanthio139 has been found in chicken, its intensely beefy disulphide 40 is not
found or is at a much lower level. It has been proposed that this be due to oxidants being
scavenged by unsaturated aldehydes such as 46 and hence not being available for the
oxidation of 39 to 40. 2,5-Dimethylfuranthiol 47 has been reported to be present in
chicken, but other work has failed to confirm this. The authors experience of the this
material is that it is more prone to oxidation than 2-methyl-3-furanthio1, and this may be
preventing its detection. When a sample of neat 47 is left in the laboratory exposed to the
air it rapidly becomes cloudy due to droplets of water formed as the by-product of aerial
oxidation. A compound with excellent pork character is pyrazineethanethiol 48. This has
not yet been reported in nature, and may be an analytical quirk. Since vinylpyrazine 49

214

Advances in Fluvours and Fragrances

has been found in pork (and other meats), and pork is rich in sulphur compounds, including
hydrogen sulphide, it is difficult to see how 48 cant be formed!
OH

AY
45

OH

A
46

47

49

48

Figure 21 Pork chemistry

Lamb character is associated with two acids, 4-methyloctanoic 50 and 4methylnonanoic acid 51;whilst these have the higher odour thresholds of other carboxylic
acids, their sharp-fatty aromas give them at least honorary membership of the high impact
club!

F0
50

OH

51

Figure 22 Lamb acids

2.14 Bouillon, HVP
The aroma chemical associated with this is 4-methylthiazole-5-ethanol 52 (Sulfurol).
However, it has a reported odour threshold of over 10,00Oppb, so is scarcely a high impact
chemical. It is also a well-known phenomenon that apparently identical batches of sulfurol
have different odours, with the desirable meaty note not always present. A possible
candidate for this impurity is 2-methyltetrahydrofuran-3-thiol 53. This is an intensely
savoury molecule, with brothy, casserole, boiled meat notes and even allium overtones.

Fluvours/Foods

215

Its carbon, oxygen-sulphur framework is actually the same as that in sulfurol; it may be a
degradation product or a by-product formed during the synthesis of sulfurol.

52

53

Figure 23 Savoury, bouillon compounds

a
Figure 24 Structural relationship between sulphur01 and 2-methyltetrahydrofuran-3-thiol

2.15 Fatty
Not at first sight the most desirable of characters, but fattiness is key character in
foodstuffs (as those forced to eat low-fat or reduced fat foodstuffs know to their cost) in
terms of both flavour and mouth-feel. Aldehydes have very fatty notes, in particular
trans-2-nonenal 54 and trans-2-trans-4-decadienal46; the latter is reminiscent of chicken
fat and has an odour threshold of 0.07 ppb. A molecule with great potential in this area is
12-methyltridecanal 55. This tallowy material is found in beef fat and appears to
originate from micro-organisms in the rumen of cattle4. It is absorbed by the gut as
plasmalogens, and released only when the beef is heated over a long period, e.g. stewing;
briefly roasting the meat does not release this chemical. Hence, with the use of this
material we have the potential to create a boiled or stewed beef flavour well differentiated
from roasted or fried beef.

216

Advances in Flavours and Fragrances

H

54

46

55

Figure 25 Fatty aldehydes

2.16 Cheesy, rancid
Cheesiness, desirable or otherwise, is often associated with acids, but these have quite
high odour thresholds, e.g. valeric acid 56, which has a nauseating sweaty-cheesiness at
high concentration, but also the mercifully high odour threshold of 3000 ppb! However,
such is the character of these that the impact is greater than the odour threshold might
imply. Unsaturated acids such as trans-2-hexenoic acid 57 have more powerful, acrid
odours; several trans-2-enoic acids (trans-2-hept, oct- and non-enoic acid) were included
on the GRAS 19 list. Simple thioesters such as methyl thiobutyrate 58 and methyl (2methy1)thiobutyrate 59 also have an intense cheesy-sweet-fruity odour;

H3C\

An

58

59

Figure 26 Molecules for cheesy, rancid notes

2.17 Mushroom, earthy
Here we have a classical high impact aroma chemical, l-octen-3-01 60, with an odour
threshold of only lppb and very characteristic of mushroom. However, this is not the
whole story as the related l-octen-3-one 61 has a threshold some two hundred times lower
at only O.OSppb! This has a very fresh wild mushroom aroma. It has also been identified as
a powerful odorant in materials as diverse as elder flower, raspberry and chocolate.
Earthiness is also associated with some pyrazines, especially 2-methyl-3methoxypyrazine 2.

Flavou rs/Foods

217

60

61

Figure 27 Compoundsfor mushroom and earthy aromas

2.18 Truffle
The black truffle is perhaps the most select member of the fungal food family. Whilst it
contains familiar volatiles such as 1-octen-3-01 60, the key character impact material is
bis(methy1thio)methane (2,4-dithiapentane, truffle sulphide) 62. This has the very powerful
earthy-alliaceous aroma associated with the truffle; also present is tris(methy1thiomethane)
(3-methylthio-2,4-dithiapentane,
methylidynetris(methy1 sulphide), manxane) 63, with an
aroma more reminiscent of the white truffle. Also very recently identified in white truffle
is the isomer of 63, 2,4,6-trithiheptane 64.

/vS\

/svsvs\

sYs
s\

62

64

63

Figure 28 Trufle volatiles

2.19 Garlic
Garlic is rich in sulphur compounds, especially allyl compounds; indeed, the commonly
used term allyl for prop-2-enyl derives from allium sativum or garlic. The major
component of garlic oil is allyl disulphide 8, with the mercaptan 65 and higher sulphides
such as the trisulphide 66 and mixed disulphides such as 67 and 68 also present. Ally1
methyl disulphide 68 is particularly pungent and has been detected at unexpectedly high
concentrations in the breath of garlic eaters.

ms\sMmSH
8

65

67

Figure 29

Aroma chemicalsfor garlic

68

Advances in Flavours and Fragrances

218

2.20 Onion
As with garlic, onion is high in sulphur compounds, but mostly these are saturated
compounds such as the methyl and propyl sulphides 69 - 73. These have less harsh,
'sweeter' notes compared to the ally1 compounds. Recently two new highly odorous
mercaptans were identified in onion6, 3-mercapto-2-methylpentan- 1-01 74, an onion- and
leek-like material with an odour threshold of 0.15 ppb, and 3-mercapto-2-methylpentanal
75, more pungent and meaty, with an odour threshold of 0.95 ppb.

mS\S/CH
m s \ S / s \ C H ,

-S/CH3

69

71

70

72

73

74

75

Figure 30 Old and new molecules from onion

Some of the higher allium sulphides may be formed by chemical transformations;
when unsymmetrical disulphides are treated with base, 'disproportionation' takes place to
form a statistical mixture of the symmetrical and unsymmetrical disulphides. This is
presumed to occur via nucleophilic attack on the disulphide bond, as shown below. Traces
of thiols are excellent catalysts for this reaction, which gives the possibility of this being
facilitated by cysteine in foodstuffs and flavours;

HIGH pH

R/s\s/R'

R/s\s/R'

,/S\s/R

R,/s\s/R'

HIGH [Nuc']

"";)
/\
L

/R'
S

i"
R/s

s/

-s/R'

R'

I
s\Ra

%Figure 31 Disproportionation of disulphides

FlnvourdFoods

219

Trisulphides are even more sensitive, rapidly forming a symmetrical mixture of di-,
tri-, and tetra-sulphides at pH 9, presumably by a similar pathway;

HIGH pH (>9)

R/s\s/s\R

R/S\s/R

HIGH [Nuc]

+
R /s\s/s\R

+
R/S\s/S\s/R

Figure 32 Disproportionation of trisulphides

3 THE STINKOPHORE; STRUCTURE-ODOUR RELATIONSHIPS IN HIGH
IMPACT AROMA CHEMICALS
Very few structure-odour relationship studies have been conducted in the flavour area, in
part because the usage of materials in this area is dominated less by activity and more by
the issue of nature-identical (and/or natural). There is little value in designing the worlds
truffliest molecule if, in the end, it cannot be used. We might also comment that nature
has done rather well in making high impact chemicals herself anyway! However, there are
some conclusions that we can draw. The first is the Tropical olfactophore, the 1,3oxygen-sulphur relationship. We see this in many powerful tropical, fruity and vegetable
aroma chemicals;

A=H, SCH3, ring

B=H, CH3, Acyl, absent if carbonyl
R,, R2 =H, alkyl,
R3 =H, alkyl, ring
R4=H, CH3, ring, OR
R5=H, absent if carbonyl

Figure 33 The tropical olfclctophore

220

Advances in Flavours and Fragrances

SH

OH

'ip"

Figure 34 Molecules showing the tropical olfactophore

The high frequency with which we see this functionality probably reflects the
biosynthetic route, i.e. Michael addition of a sulphur nucleophile to an a$,-unsaturated
carbonyl compound followed by further transformations (reduction, acylation, alkylation)
at both the sulphur and oxygen functionality's.

Figure 35

Biosynthesis of the tropical olfactophore

Interconversion between functionalities on the oxygen and sulphur takes place readily.
When S-acetyl-3-mercaptohexanal (76) is reduced with sodium borohydride, the 0-acetyl
alcohol 78 is produced along with the expected S-acetyl compound 77, and smaller
amounts of the de- and di-acylated products 79,SO;

Flavou rs/Foods

22 1

76

77

78

79

80

Figure 36 An unexpected acyl transfer

This OAc-SAC Shuffle takes place due to the facile formation of the 6-membered
transition state;

Figure 37 The OAc-SACshuffle

222

Advances in Flavours and Fragrances

We have a more specific olfactophore for the catty, blackcurrant area, where the three
key compounds all have a tertiary (1,l -Dimethyl) mercaptan;

d3

A=H, SCH3, ring

B=H, CH3, Acyl, absent if carbonyl
R3 =H, alkyl, ring
R,=H, CH3, ring, OR
R5=H, absent if carbonyl

Figure 38 The catty olfactophore

+fCH3

Figure 39 Molecules showing the catty olfactophore

This leads us to the possibility of building a molecule smelling more like cats urine
than anything nature has produced, should we so desire.. .
A more tentative olfactophore is the 1,2-0xygen, sulphur relationship in meaty,
savoury compounds, usually formed as advanced Maillard products;

A=ring, absent if carbonyl

B=H, CH3, Acyl, SR
R1=ring, absent if carbonyl
R,, R2, R3=ring, H, alkyl

Figure 40 The savoury olfactophore

Flavours/Foods

223

SH

Figure 41 Molecules showing the savoury olfactophore

4 THEFUTURE
There are three areas where developments are continuing. The first is in synthetic
chemistry; materials that are interesting but too expensive for use at present may become
available at a useable price due to the discovery of a viable synthetic route. The cycle of
discovery, synthesis, and manufacture with falling prices has occurred since the 19h
century work on cinnamaldehyde and vanillin. A second area involves further analytical
work, which may be in examination of new exotic foodstuffs or re-evaluation of familiar
materials. For example, the cat ketone 5 was recently found to be an unexpectedly
important odourant in grapefruit7. We are now able to go further down in the noise on a
gas chromatography and identify materials at lower levels but even higher odour
thresholds. For example dimethyl selenide (81) has been detected in garlic and garlic
breath. Handling and manufacture of such materials may require glove-box techniques
more familiar from radiation chemistry! The third area that may have an important effect
on the usage of high impact aroma chemicals is research being carried out on delivery
systems. A number of very interesting aroma chemicals are also highly reactive and have a
short half-life in normal formulations. Systems which can trap and release such materials
could enable their use for the first time; examples include the related 2-acetyl- 1-pyrroline
82 (basmati rice) and 2-acetyltetrahydropyridine 83 (bread crust);

224

Advances in Flavours and Fragrances

+
5

82

..,.i-:1
0

83

Figure 42 Future high impact aroma chemicals?

5 A NOTE ON ASSOCIATIONS
Whilst the human nose is an unsophisticated instrument compared with that of some
animals, it remains a more powerful organ that we sometimes realise. It appears to have a
hot line to the brain, and our ability to associate odours with people and with places is
well known. The impact of some of the materials discussed in this paper makes them
very effective in this and some of the associations that people have made when shown
these materials are greengrocers for 2-isobutylthiazole, presumably due to its tomato
notes, the cinema for 2-acetylpyrazine (via its popcorn odour!) and fields at 6a.m. for
1-octen-3-one. Some twenty-five years ago the commentator had a job starting early in the
morning and would go out to pick mushrooms in an adjacent field. Some associations are
very personal and depend on very individual circumstances; whereas most commented on
the blackcurrant, fruity aroma of 4-methoxy-2-methy1-2-butanethio1,
to a colleague with a
six month old baby, it was wet diapers.

6 WHY?
A final question that may be asked is simply why are we so sensitive to these aroma
chemicals?
The functions of our senses of smell and taste are threefold;

1. To findattract a mate
2. To findidentify food
3. To avoid toxins
It is actually quite difficult clearly explain our responses in these simple terms.
1. To finaattract a mate

Flavours/Foods

225

This is probably the least relevant here! Whilst the extreme sensitivity of insects to sex
pheromones is well studied, the fact that these aroma chemicals are found in foodstuffs is
something of a complication. Sexual attraction between an animal and its food is unlikely
to be a successful evolutionary strategy.

2. To findhdentify food
At first sight this is the obvious explanation! However, most of the high impact
aroma chemicals are formed only when food is cooked. Since the cooking of food is of
very recent provenance in evolutionary terms, it is unlikely that we have evolved any
physical features to respond to this. Some high impact materials are found in fruits, but
again, the chemical may only be released when the fruit is actually being eaten. For
example ally1 disulphide and the other allium sulphides are only released only when the
tissues of the garlic clove have been damaged, and a cow certainly doesnt smell of roast
beef!
3 . To avoid toxins

There are three main sources of toxins;

Those present in the environment
Those produced as an organisms waste
Those produced by the decomposition of food.
It is this latter area which gives a clue that this may be the cause of the response to
these high impact chemicals. The group of compounds to which we have the greatest
sensitivity is mercaptans, which are produced by the decay of cysteine and methionine in
proteins. This may be the origin of our response to the simple materials such as hydrogen
sulphide, methyl mercaptan and simple alkyl thiols. The enhanced response to mercaptans
such as 2-methyl-3-furanthi01 and p-menthene-8-thiol may simply be that these happen to
trigger the receptors more easily. This is a coincidental response and not a specific
design. To use an analogy from pharmaceutical chemistry, morphine happens to fit our
endorphin receptors in the brain with great efficacy, but it is not suggested that we have
evolved to develop morphine addiction!
Our response to these molecules appears to have a very primitive origin; we have yet
to meet an individual with specific anosma to these materials. The ability to respond to
chemicals in our surroundings, is the primary sense. We now differentiate taste and smell,
but to the simplest organisms it is as one. Even the simplest and most primitive organisms,
the prokaryotic bacteria and archea, have this sense. This leads us to a fascinating
possibility; there is much evidence that life evolved in a sulphur-rich environment, where a
sulphur compound would be a nutrient or a toxin, depending on concentration. Does our
love for roast beef and truffles have its ultimate origins in the days when the only course
on the menu was the primordial soup?

CONCLUSION

As the great 17h Century philosopher Rene Decartes might have said, odorato ergo sum.

Advunces in Fluvours and Fragrances

226

References
Whenever possible odour thresholds quoted are those in Flavor- Base 98,
Leffingwell & Associates, Canton, Georgia, USA
P. Darriet, T. Tominaga, V. Lavigne, J. N. Boidron and D. Dubourdieu, Flav Fragr J ,
1995,10,395
H . Guth and W. Grosch, Fat Sci Technol, 1991,93 335
R. Kerscher, K. Nurnberg, J. Voigt, P. Schierberle and W. Grosch, J Agric Food
Chem, 2000,48,2387
U. Jorgensen, M. Hansen, L. P. Christensen, K. Jensen and K. Kaack, J Agric Food
Chem, 2000,48,2376
S. Widder, C. S. Luntzel, T. Dittner and W. Pickenhagen, JAgric Food Chem, 2000,
48,418
A. Buettner and P. Schieberle, J Agric Food Chem, 1999,47, 5 189
E. Block, X.-J. Cai, P. C. Uden, X. Zhang, B. D. Quimby and J. J. Sullivan, Pure &
Appl. Chem., 1996,68,937
S . Mahatheeranont, S. Keawsa-ard and K. Dumri, JAgric Food Chem, 2001,49,773

Subject Index

Acetic acid, 7, 10
Acetone, 7, 10
ACTHP, 171
formation of, 171
Alcohols
1-alkyl, 15-24
2-alkyl, 15-24
relationship between odour &
chemical, 15-24
Aldehydes, 2 15
Amadori compound, 163,165,175
Ambergns, 113
Ambrox, 113-124
Amoore theory, 3,4, 5
Androstenone, 4, 5, 7
Anethole, 72
Angelica lactone
alpha, 39,42-45
beta, 42-45
gamma, 42-45
gc of, 43-45
Anosmia, 4, 5, 6
Aroma chemicals
high impact, 202-225
Beef
analysis of, 153-154
BBQ, 150
grilled, 150
with onion, 150
Beer
analysis of, 32
Bergamot oil, 76
Bitter orange oil, 76, 78
Bourgeonal, 11
Buchu leaf oil, 209
Canarium luzonicum, 92-98
Canurium madugascariense, 92-98
gc mass spectrum of, 95
gc-olfact ometery, 93
Capsaicin, 10,208
Carbonic acid, 10
Cat ketone, 203, 209

Chicken, 2 13
Chromatography
liquid
analysis of acids in hops, 35
micellar electrokinetic (MEKC), 35
Cineole, 7
Cistus laduniferus, 114
Citrus oils, 74
coumarins in, 74
germacrenes in, 74
methoxyflavones in, 75
psoralens in, 75
sesquiterpenes content of, 74
Civetone, 12
Clary sage oil, 67
Compararative molecular field analysis,
15-24
Cyclopentanones, 127-136
Cysteine, 196- 199
Damascenone, 85,138
Damascone, 138, 143-146
Desorption
liquid, 29
thermal, 29
Detector
sulphur, 149
Diketones
alpha, 39-42
Dimethyl selenide, 223
Dimethyl sulphide, 2 10
Distillation
hydro, 65
steam, 65
Electronic nose, 8
Elemi, 92,95, 96
Essential oils, 57-83
adulteration of, 68-79
chemical composition differences,
58-67
isotope measurement in
stable, 69, 70

228

unstable, 69
standardisation, 68-79
Ethanol, 7
Extraction
accelerated solvent (ASE), 27
gum phase (GPE), 27
solid phase (SPE), 27
solid phase micro (SPME), 27, 102105,202
sorptive, 27-39
headspace (HSSE), 28,30
stir bar (SBE), 28-39
in analysis of beer, 32
in analysis of tea, 30
in analysis of yoghurt. 30
ultrasonic WE), 27
Flavour wheel, 206
Furanone
4HDF, 167-170
Galaxolide, 12
Galbanum oil, 207
Garlic, 204,217
Gas chromatography, 40-44
chiral, 78
olfactometry, 93
Generessence, 147-148
Georgywood, 138, 141- 142
Geranium oil, 58
Guaiacol, 21 1
Hazelnut, 151
analysis of, 154-155
4-HDF, 164-170
formation of, 167- 170
Helional, 9, 10
Hexenal, 88-89,204,207
Hexenol, 207
Hodge-scheme, 163-165
Humulones, 35

Advances in Flavours und Fragrances

stable, 69-70, 84-91, 167

unstable, 69
Kallmann syndrome, 6
Koavone, 138,139
Labdanes, 113
Labdanolic acid, 114-116
Lamb, 214
Larixol, 116-120
Lavandin, 58
Leaf alcohol, 207
Leaf aldehyde, 207
Lemon oil, 73, 77, 78, 80
Lime oil, 76
Madagascar, 92
Magnolia ketone, 127
Maillard reaction, 45, 163-176, 196,21021 1,212
Mango, 149
analysis of, 151- 153
Matrices
fatty, 30
Melissa officinalis, 64, 65
Mentha pulegium, 64
1-Menthen-g-thiol, 208
Menthol, 7, 10
Methyl dihydrojasmonate, 8, 127, 132136
Methyl epijasmonate, 127-132
Methyl jasmonate, 127, 132-136
MFT, 194-200,212
Microgravity
effect on fragrance of rose, 99-109
Monosodium glutamate, 6
Mouthfeel, 184-185
Mushroom, 2 16
Musk, 11
Musk ketone, 12

NMR
Ionone, 139- 140
beta, 139
Iso-E-super, 138-142
Isohumulones, 35
Isomerisation, 44
Isoraldeine, 139, 140
Isotopes
deuterium exchange, 196, 199-200

nitrogen, 45-53
of pyrazines, 45-53
site-specific natural isotope
fractionation (SNIF-NMR), 85
Norlimbanol, 12
Odour sensation
subjectivity of, 3- 13

Subject Index

Odour threshold, 202

Olfactophore
in flavour area, 21 9-223
Onion, 2 18
Origanum majorana, 65
Passion fruit
carbon isotopic deviation in, 86
Pauson-Khand reaction, 128-129
Peppermint oil, 61, 77
Pineapple
carbon isotopic deviation in, 86
Piperin, 10
Piperonal, 9, 10
Polidy, 60
Polydimethylsiloxane, 27
Pork, 2 14
Principle component analysis, 19-23
Process flavours, 194-201
flavour stability in, 196-198
Propanol, 7
Pyrazines, 39,202-203,2 10-211,216
analysis of substitution pattern, 4553
Pyridine
2-acetyltetrahydro, 17 1- 172
Raspberry ketone, 87-88
Receptor
olfactory, 9
OR17-40, 9
Resinoids
lichen, 89-9 1
Rose
hybrid tea, 99- 109

229

Sotolone, 208
Space shuttle, 99
Strecker acids, 164
Strecker aldehyde, 164
Strecker reaction, 165, 169, 170, 172-175
aroma compounds formed by, 172175
Structure activity relationship, 3-14, 215
quantitative, 15-24
Structure odour relationship, 219-223
Tagetes minuta, 58,61
Tautomerism
keto-enol, 40-42
Thaumatin, 178-189
calorific value, 181
solubility of, 181
structure of, 181
sweetness of, 182-183
use in oral care, 182
Thaumatococcus daniellii, 178-189
Thiols, 15-24,209-2 10
1-alkyl, 15-24
2-alkyl, 15-24
in process flavours, 194-20 1
relationship between odour &
chemical structure, 15-24
Thyme oils, 63
Timberol, 12, 138, 142
Trigeminal, 6, 7, 11
Tropathiane, 209
Truffle, 217

Umami, 6
Vanillin, 9

Safrole, 9
Sandelice, 5
Salvia officinalis, 58

Yarrow, 60