Flavours and Fragrances (Woodhead Publishing Series in Food Science, Technology and Nutrition)

Flavours and Fragrances

Flavours and Fragrances Edited by

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

WOo D H E A D

P uB LI s H I N G L I MI TE D

Cambridge England

Published by Woodhead Publishing Limited, Abington Hall, Abington, Cambridge CB 1 6AH, England www.woodheadpublishing.com The proceedings of the 1997 RSUSCI International Conference on Flavours and Fragrances held 30 April-2 May in Warwick, UK First published by The Royal Society of Chemistry 1997 Reprinted by Woodhead Publishing Limited 2005

0 Woodhead Publishing Ltd, 2005 The authors have asserted their moral rights This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. Reasonable efforts have been made to publish reliable data and information, but thc authors and the publishcr cannot assume responsibility for the validity of all materials. Neither the authors nor the publisher, nor anyone else associated with this publication, shall be liable for any loss, damage or liability directly or indirectly caused or alleged to be caused by this book. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming and recording, or by any information storage or retrieval system, without permission in writing from the publisher. The consent of Woodhead Publishing Limited does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from Woodhead Publishing Limited for such copying. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library

ISBN- 13: 978-1-85573-780-8 ISBN-10: 1-85573-780-9 Printed in the United Kingdom by Lightning Source U K Ltd

Preface

This book is a compilation of the majority of the twenty one papers presented at the 1997 RSCfSCI flavours and fragrances conference at Scarman House, University of Warwick. The aim of the meeting was to bring together scientists from both industry and the academic world, who have a common interest in the chemistry of flavours and fragrances. The subject matter was intentionally broad, covering areas such as chemoreception, analytical techniques, essential oils, the synthesis of flavour and fragrance materials in the laboratory, clean efficient syntheses on a manufacturing scale, and the formation of flavours both in the cooking process and using biotechnology. The book is divided into the same sections as the original meeting. The meeting was spaced over two and a half days and saw speakers and delegates from all comers of the world exchanging ideas and information. Special thanks go to the Conference Secretariat Elaine Wellingham for helping me organise the conference on behalf of the Biological & Medicinal Chemistry Sector of the RSC and the Fine Chemicals Group of the SCI. The staff at Scarman House, University of Warwick, deserve a mention for their very professional and smooth running of such a wonderfid conference facility. Thank you also to all of the speakers and co-authors who firstly, submitted their manuscripts so promptly, and secondly, for co-operating with the editor! A very big thank you goes out to my team of proof readers, Jamie Mankee, Dave Munro, Lucy Swift, and Kim Yarwood. Their work is invaluable, and they probably know the manuscripts ‘off by heart’ by now! Finally, thank you to the RSC for agreeing to publish this book.

Contents

Chemoreception and Structure-Activity Relationships Toward a Rational Structure-Function Analysis of Odour Molecules: The Olfactory Receptor TM4 Domain Michael S. Singer, Gordon. M.Shepherd

3

The Impact of Recombinant DNA-Technology on the Flavour and Fragrance Industry Andreas Muheirn, Alex Hausler, Boris Schilling, Konrad Lerch

11

The Design and Synthesis of Novel Muguet Materials Karen J. Rossiter

21

Aura of Aroma@:A Novel Technology to Study the Emission of Fragrance from the Skin Braja D.Mookerjee, Suba M. Patel, Robert W. Trenkle, Richard A . Wilson

36

Dependence of Intensity of Musk Odour on the Energy Gap between Frontier Molecular Orbitals Mikhail. Y. Gorbachov

48

Essential Oils and Analytical Derivatized Cyclodextrins in Enantiomer GC Separation of Volatiles Carlo Bicchi, A . D 'Arnato, V. Manzin

57

Production, Chemistry and Sensory Properties of Natural Isolates Mans H. Boelens

70

An Odour Sensing System for Use in Measuring Volatiles in Flavours and Fragrances Using QCM Junichi Ide, Takarnichi Nakarnoto, Toyosaka Moriizurni

87

The Aromatic Resins: Their Chemistry and Uses David A . Moyler, Robin A. Clery

96

...

Flavours and Fragrances

Vlll

Studies Towards Structure Determination of Substituted Pyrazines Michael Zviely, Alexander Kern, Igal Gozlan, Ron Frim

116

Flavours Generation of Taste through (Redox) Biocatalysis Corja Laane, Ivonne Rieljens, Huub Haaker, Willem van Berkel

137

The Maillard Reaction in Flavour Formation Hugo Weenen, J. Kerler, J.G.M. van der Ven

153

Relationships between Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles AndyJ Taylor, R. S. T. Linforth

171

Organic Chemistry Synthesis and Odour Properties of Chiral Fragrance Chemicals Tetsuro Yamasaki

185

In Search of Nascent Musks ..... Or Not! Walter C. Frank

196

Synthesis and Application of Thiocarbonyl Compounds Shuichi Hayashi, S. Hashimoto, H. Kameoka, K. Sugimoto

209

Heteropolyacids and Related Compounds as Catalysts for Fine Chemicals Synthesis Ivan V. Kozhevnikov

222

Subject Index

237

Chemoreception and Structure-ActivityRelationships

Toward a Rational Structure-Function Analysis of Odour Molecules: The Olfactory Receptor TM4 Domain Michael S. Singer and Gordon M. Shepherd SECTION OF NEUROBIOLOGY AND INTERDEPARTMENTAL NEUROSCIENCE PROGRAM, YALE UNIVERSITY SCHOOL OF MEDICINE, 236 FMB, 333 CEDAR STREET. NEW HAVEN, CT 06510,USA

1

INTRODUCTION

The family of olfactory receptor proteins (ORs) is currently the best candidate site for odour transduction in the vertebrate olfactory system. The first evidence for this was the identification of ORs as members of the G protein-coupled receptor (GPCR) superfamily, as predicted by an odour-induced adenylate cyclase activity”’ and supported by blockade of odour responses by the GPCR antagonists propranolol and a t r ~ p i n e Further .~ support came from estimates that mammals possess 1000 different OR subtypes, consistent with the capacity to detect many thousands of odour~.~-’ Direct evidence followed from two expression systems, which showed selective responses by rat subtype OR5 to two aldehydes of the lily of the valley class, lyral and lilia1.8‘9 Finally, computational analysis of the OR family has pointed to 5-10 amino acid residues likely to interact with odour molecules. Variations in these residues were postulated to account for different odour preferences across OR subtypes.”

-

Knowledge of ORs and their interactions with odour molecules obviously has important implications not only for basic research on olfactory transduction but also for practical applications in the fragrance industry. Given this potential importance, we have initiated studies of the mechanisms of odour molecule-OR interactions at the molecular level, using new advances in computational analysis of gene families. This paper focuses on OR proteins at the amino-acid level, as the basis for developing a rational structurefunction analysis of odour molecules. It is oriented toward organic chemists and perfhers in order to illustrate how data on ORs may inform the practice of fragrance and flavor chemistry. The OR fourth transmembrane domain (TM4), which has attracted interest in several studies, serves as a model for this discussion.

2 THE STRUCTURAL MOTIF OF G PROTEIN-COUPLED RECEPTORS Figure 1 shows the predicted tertiary structure of OR proteins, which is likely to be conserved across different species and subtypes. Like other GPCRS, ORs are predicted to have seven a-helical transmembrane domains (TMs), identified as hydrophobic spans of

4

Flnvours and Fragrances

18-25 amino acids. The amino terminus is extracellular, and the carboxyl terminus is intracellular," Conserved cysteines in the first (El) and second (E2) extracellular loops are likely to form a disulfide bond. Molecular models, correlated mutation analysis, and positive selection moments have pointed to an odour-binding pocket (A) formed by TM3, TM4, TM5 and TM6 (reviewed in ref. 10). The third intracellular loop (B), between TM5 and TM6, is believed to be the principal site of G protein activation. The extracellular loops (C) may serve multiple functions: binding odour molecules; docking olfactory binding proteins, or guidin axons of olfactory receptor neurons to their glomerular 8-13 targets in the olfactory bulb.

3 SEQUENCE DIVERSITY: VARIATIONS ON THE GPCR MOTIF Prior to the identification of ORs, several workers reasoned that amino acids in the odour binding site would vary across subtypes, in order to accommodate diverse odour molecule^.^^'^ This prediction was fulfilled by TM3, TM4, and TM5 of the rat ORs isolated by Buck and Axel: and yielded the first clue that TM4 may interact with odour molecules. Ben-Arie et a[.'' noted diversity in TM4 and TM5 of human ORs, labelling these domains and the loop between them (E2) an OR hypervariable region.I6 Both observations compared favorably with the related P-adrenergic receptor (PAR), where an ._ TM5 and TM6." agonist binding

Figure 1 Schematic model of an olfactory receptor, with extracellular surface at top. Transmembrane a-helices are shown as numbered cylinders, interhelical loops as lines. Three functional regions are marked: (A) predicted odour-binding pocket,(B) predicted site for G protein coupling, and (C) extra-cellular loops. Asterisk (*) indicates predicted disulfide bond between extracellular loops.

Toward a Rational Structure-FunctionAnalysis of Odour Molecules

5

4 POSITIVE DARWINIAN SELECTION Sequence diversity, as cited above, is dificult to interpret due to random mutations, which accumulate in noncritical areas of proteins. However, the significance of this diversity can be tested more rigorously at the nucleic acid level. This is carried out by measure of positive Darwinian selection, an evolutionary process which favors amino acid diversification over conservation. Under positive Darwinian selection, the relative rate of nonsynonymous mutations, which encode different amino acids, out aces the ?'-I9 This relative rate of synonymous mutations, which encode the same amino acid. process occurs in at least two classes of immune system molecules, where it is notably restricted to the antigen recognition sites. Ngai et of.*' reasoned that by analogy with immune system molecules, ORs would show positive Darwinian selection restricted to the odour binding site. Their analysis of ORs from the catfish subfamily 32 pointed to TM3 and TM4 as domains likely to interact with odour molecules.

5 MOLECULAR MODELS Evidence that rat subtype OR5 is activated preferentially by lyral or lilial' prompted us to build three-dimensional models of OR5 and other subtypes.2' These were based on bacteriorhodopsin coordinates,22 which have served as a template for several GPCR model^.^^-^' The models were further constrained by methods which predict the limits and relative rotations of transmembrane cr-heli~es.~'-~~ The OR5 model enabled us to search for candidate odour-binding sites by calculating interaction energies for a panel of odour ligands at various locations and orientations in the protein structure, a process referred to as docking. Lyral showed optimal interactions in a pocket bounded by TM3, TM4, TM5, TM6 and TM7, which showed considerable similarities with the PAR binding pocket. Three critical residues were identified in TM4: histidine 155, alanine 156, and histidine 159. Notably, these three residues were situated on the sanie side of the TM4 helical structure, near the extracellular part of the receptor. The importance of the TM4 residues, particularly His 155 and His 159, has been supported by subsequent models. Bajgrowicz and Broger2' built another model of OR5 based on an earlier PAR model, employing different modelling and docking software. The binding pocket was similar to that of Singer and Shepherd2' and indicated important roles for both His 155 and His 159. 6 CORRELATED MUTATION ANALYSIS The limitations of computer models and well documented difficulties of GPCR structure prediction and determination have prompted efforts to correlate protein sequence and receptor function directly, without the intermediate step of structural informati~n.~''~~ These efforts include correlated mutation analysis (CMA),28"2 which has proven useful for GPCRs such as opsins and muscarinic acetylcholine re~eptors.~'CMA compares several subtypes from the same receptor family, scanning the sequences for pairs of residues which remain conserved or mutate in tandem. Residues which behave in this way are said to show correlated mutations. They are believed to be functionally or structurally interdependent, based on the theory that one mutation compensated for the

Flavours and Fragrances

6

deleterious effects of the other (see review in ref. 10). When we applied the method to rat OR sequences, 13 out of 310 residues were significantly correlated.33s34Most of these residues were concentrated in the upper half of TM4, TM5 and TM6, consistent with the predictions of Buck and Axel4 and Ben-Arie et al.," as well as the PAR binding pocket." The correlated residues also matched positively selected residues in the catfish sequences of Ngai er aZ.20Furthermore, there was a striking correspondence between the CMA data and the previously published computer models. The three residues identified by CMA in TM4 were exactly those found in the OR5 model: His 155, Ala 156, and His 159. Thus, evidence for the importance of these residues converged from several directions: amino acid hypervariability, overlap with the PAR binding pocket, positive Darwinian selection, two independent structural models, and the structure independent method of CMA. 7 HELICAL WHEEL STRUCTURE The secondary structure of TM4 can be modelled as a canonical or-helix (see ref. 35). Two important features of TM4 can be seen in the helical wheel diagram in Figure 2, which shows the orientations of residue side chains viewed en face, from outside the cell. His 155 and His 159 point approximately in the same direction as Ser 148 and Thr 152. The side chain polarity of all four residues, as well as possible cationic states in the histidines, provides strong evidence that they are oriented toward the receptor cleft (cf. refs. 36, 37), and probably into an odour binding site (cf. ref. 38). In contrast, other residues in TM4 are decidedly hydrophobic, consistent with placement in the lipid or helical interface. Exceptions are Ser 143, whose near cytosolic location may relieve it from the constraints of membrane hydrophobicity, and Thr 153, an amino acid which is known to be stable at helical interface^.^^ Thus, canonical (a-helix modelling of TM4) supports the hypothesis that His 155

Figure 2 Helical wheel diagram of TM4, viewed from outside the cell. Amino acids are based on rat subrype OR5 (ref: 8). Circles denote polar amino acids: filled circles, histidines; open circles, serines and threonines. Dotted lines indicate nonpolar amino acids. Note similar orientations ofpolar residues.

and His 159 could contribute to an odour-binding pocket. Simulations of a-helical domain interactions have recently been carried out with reasonable accuracy, as shown for glycophorin A.40 This approach may render it feasible to model interactions between odour molecules and TM4 at the atomic level.

Toward a Rational Structure-Function Analysis of Odour Molecules

I

Crabtree4' cited evidence that mercaptans, characteristically detected at low thresholds, bind to a transition metal center such as Cu(1). The arrangement of the histidines suggests that they could help to coordinate such an ion. 8 DUAL-HISTIDINE SUBSITE AND ODOUR DETERMINANTS

The biophysical properties of histidine make it possible to predict how His 155 and His 159 might bind odour molecules. The side chain of histidine contains an imidazole ring, capable of forming hydrogen bonds, electrostatic interactions, and van der Waals interactions. The imidazole ring shifts readily between protonated and deprotonated states due to its pK of 6.5, very near physiological pH. These properties make histidine quite versatile in its range of possible non-covalent bonds. Two histidines, such as His 155 and His 159, may also act cooperatively. This model is consistent with the predictions of Kosower,I4 who listed several pairs of amino acids postulated to interact with particular functional groups on odour ligands. Two histidines, for instance, were predicted to interact best with aldehyde or nitrile groups. We have further postulated that the "dualhistidine subsite" on ORs can bind an array of different polar groups or nonpolar regions with graded affinities, as shown in Figure 3. Collectively, these polar group and nonpolar regions may be called determinants. In this view, most odour molecules can be defined as an assembly of 2-5 determinants in a specific geometric arrangement. 9 SUBSITE-DETERMINANT MODEL

The consensus model for the odour-binding pocket of OR5, which has arisen out of the computational approaches described above, is shown in Figure 4. The dual histidine subsite can be seen on TM4. At least 3 more potential subsites were identified: a dual serine subsite (Ser 246 and Ser 249) on TM6; Ile 199 and Phe 206 on TM5; and Phe 322 on TM3.I' We postulate that each of these subsites binds specific determinants over an affinity spectrum as shown for the dual histidine subsite in Figure 3. The dual serine subsite, for instance, contains two hydroxyl groups predicted to be 5A apart, and may thus bind hydroxyl groups and mines with high affinity, via two hydrogen bonds.38 The phenylalanine and isoleucine residues may form surface-specific van der Waals forces or n-n orbital interactions with alkyl or phenyl determinants. These considerations help to define the fundamental biophysical units that mediate OR activation, which appear to be the elementary interactions between the determinants in an odour molecule and the subsites in the OR pocket.' The overall interaction between an odour molecule and an OR would then represent the sum of complementary interactions between 2-5 subsites and 2-5 determinants. Variations in the amino acids at each subsite would modulate preferences for particular determinants; for example, dual histidines for aldehyde, dual serines for hydroxyl, phenylalanines for phenyl, and aliphatic amino acids for alkyl determinants. The relative positions of subsites in the pocket, constrained by the protein structure, would provide the basis for discrimination of isomers and enantiomers. Steric hindrance and electrostatic repulsion would also shape OR specificity.

a

Flavours and Fragrances

alaenyde R-CHO

alkane R-CH3

W

I

deterrnmanl

Figure 3 Potential interactions for difSerent Jirnctional groups (determinants) at the dualhistidine subsite (His 155 and His 159). Above, the dipole moment of the aldehyde carbonyl bond would allow for interaction with histidines as shown. Lower affinities would apply for an 0-H bond (hydroxyl) and N-H bond (amine) based on smaller dipole moments. The C-H bond of the alkane would not interact due to its lack of dipole moment. Below, hypothetical summary of the subsite afjinities for several determinanfs. Orher subsires would have drfferent aflnity spectra. The overall odour specificity of the receptor would represent a function of ihese spectra and the relative locations ofthe determinants.

Figure 4 End face view of the predicted odour-binding pocket in the OR5 receptor, based on the studies reviewed in this paper.

Toward a Rational Structure-Function Analysis of Odour Molecules

9

We postulate that the possible permutations of subsites in -1000 receptors would be sufficient to encompass the vast odour space that mammals, including humans, can discriminate. An important feature of our model is that ORs show specificity for combinations of determinants, rather than individual determinants or particular odours. The model should provide new insights into the biophysical mechanisms underlying odour perception, and help to guide fragrance and flavor chemists to a new level of structure-function analysis.

Acknowledgments We are grateful to Charles Greer, Emmanouil Skoufos, Donald Engelman, Robert Crabtree, Laerte Oliveira, Bob Bywater, Wilma Kuipers, Chris Sander, and Heinz Breer for helpful discussions, and to Gerrit Vriend for discussions and WHATIF software. This work was supported by grants to GMS from NIDCD and NASA, NINM and NIDCD through the Human Brain Project. MSS is supported by the Yale University MSTR References 1. U. Pace, E. Hansky, Y Salomon and D. Lancet D, Nature, 1985,316,255-258. 2. R. Reed, Neuron, 1992,8,205-209. 3. S . Firestein and G. M. Shepherd, NeuroReport, 1992; 3,661-664. 4. L. BuckandR. Axel, Cell, 1991,65,175-187. 5. D. Lancet, Ann. Rev. Neurosci., 1986,9,329-355. 6. G. M. Shepherd and S . Firestein, J SteroidBiochem. Molec. Biol., 1991,39,583-592. 7. D. Lancet, E. Sadovsky and E. Seidelmann, Proc. Natl. Acad. Sci., 1993,90,37153719. 8. K. Raming, J. Krieger, J. Strotrnann, L. Boekhoff, S. Kubick, C. Baumstark and H. Breer, Nature, 1993,361,353-356. 9. H. Kiefer, J. Krieger, J. D. Olszewski, G. Von Heijne, G. D. Prestwich and H. Breer, Biochem., 1996, 35, 16077- 16084. 10. G. M. Shepherd, M. S. Singer and C. A. Greer, Neuroscientist, 1996, 1,262-271. 1 1. U. Gat, E. Nekrasova, D. Lancet and M. Natochin, Eur. J. Biochem., 1994,225,11571168. 12. A. L. Hughes and M. K. Hughes, J. Molec. Evol., 1993,36,249-254. 13. M. S. Singer, G. M. Shepherd and C. A. Greer CA, Nature, 1995,377, 19-20. 14. E. M. Kosower, 'Molecular Mechanisms for Sensory Signals: Recognition and transformation', Princeton UP, Princeton, NJ, 1991. 15. N. Ben-Arie, D. Lancet, C. Taylor, M. Khen, N. Walker, D. H. Ledbetter, R. C a n T o m , K. Patel, D. Sheer, H. Lehrach and M. A. North, Hum. Mol. Gen., 1993,3, 229-235. 16. D. Lancet and N. Ben-Arie, Curr. Biol., 1994, 3,668-674. 17. C. D. S trader, I. S . Sigal and R. A. F. Dixon., FASEB J., 1989; 3, 1825-1832. 18. A. L. Hughes and M. Nei, Proc. Nail. Acad. Sci. USA, 1989; 86,958-962. 19. A. L. Hughes and M. Nei, Nature, 1988,335, 167-170.

10

Flavoitrs and Fragrances

20. J. Ngai, M. M. Dowling, L. Buck, R. Axel and A. Chess, Cell, 1993,72,657-666. 21. M. S. Singer and G. M. Shepherd, NeuroReport, 1994, 5, 1297-300. 22. R. Henderson, J. M. Baldwin, T. A. Ceska, F. Zemlin, E. Beckmann and K. H. Downing, Mol. Biol., 1990,213, 899-929. 23. S. Trumpp-Kallmeyer, J. Hoflack, A. Bruinvels and M. Hibert, J. Med. Chem., 1992, 35,3448-3462. 24. M. F. Hibert, S. Trumpp-Kallmeyer, A. Briunvels and J. Hoflack, Mol. Pharmacol., 1991,40,8-15. 25. D. M. Engelman, T. A. Steitz and A. Goldman, Ann. Rev. Biophys. Chem., 1986, 15, 321-353. 26. P. Cronet, C. Sander and G. Vriend, Prot. Eng., 1993,6,59-64. 27. J. Bajgrowiez and C. Broger, in 'Flavours, Fragrances and Essential Oils. Proceedings of the 13th International Congress of Flavours, Frangrances and Essential Oils', ed. K. H. C. Baser, AREP, Istanbul, 1995. 28. L. Oliveira, A. C. M. Paiva, G. Vriend, J. Comp. Aid. Molec. Des., 1993; 7,649658. 29. W. Kuipers, L. Oliveira, A. C. M. Paiva, F. Rippmann, C. Sander, G. Vriend, C. G. Krus, 1. van Wijngaarden, and A. R 1 Jzerman, in 'Membrane Protein Models: Experiment, theory, and speculation', ed. J. Findlay, BIOS, Oxford, 1 996. 30. L. M. Gregoret and R. T. S auer, Proc. Natl. A cad. Sci. USA, 1993,90,4246-4250. 31. U. Gdbel, C. Sander, R. Schneider and A. Valencia. Prot. Struct. Func. Gen., 1994, 18,309-3 17. 32. E. Neher, Proc. Natl. Acad Sci. USA, 91, 1994,98-102. 33. M. S. Singer, L. Oliveira, G. Vriend and G. M. Shepherd, Receptors and Channels, 1995,3,89-95. 34. G. Vriend, J. Mol. Graph., 1990, 8, 52-56. 35. C. Chothia, Ann. Rev. Biochem., 53, 1984,537-512. 36. D. Donnelly, J. P. Overington, S. V. Ruffle, J. H. A. Nugent and T. L. Blundell, Protein Sci., 1993,2,55-70. 37. D. Donnelly, J. B. C. Findlay and T. L. Blundell, Receptors and Channels, 1994, 2, 61-78. 38. M. S. Singer, Y. Weisinger-Lewin, D. Lancet and G. M. Shepherd, Receptors and Channels, 1996,4, 141-147. 39. K. R. MacKenzie, J. H. Prestegard and D. M. Engelman, Science, 1997,276,131-133. 40. P. D. Adams, D. M. Engelman and A. T. Briinger, Prot. Struct. Func. Gen., 1996, 26,257-261. 4 1 . R. Crabtree,J. Inorg. Nucl. Chem., 1977,40, 1453. 42. C. D. Strader, T. M. Fong, M. P. Grazianoand, M. R. Tota, FASEBJ., 1995,9,745754.

The Impact of Recombinant DNA-Technology on the Flavour and Fragrance Industry Andreas Muheim, Alex Hausler, Boris Schilling and Konrad Lerch* GIVAUDAN-ROW RESEARCH LTD., 8600 DUBENDORF, SWITZERLAND

1 INTRODUCTION Biotechnology, of which recombinant DNA technology is an important sub-discipline, has a long tradition in the production of food and flavours. Man first started to apply microbes around 3500 BC for the production of wine, beer, bread and many other food articles that became an indispensable part of our daily diet’. At the beginning such fermentations were carried out on a rather empirical level and only in the 19th century the scientific basis was laid by the discoveries of Louis Pasteur. Isolation and controlled cultivation of microbes became possible and about 20 years a o these techniques also found application in the K production of various flavour chemicals . In the early 1970’s, recombinant DNA technology emerged and soon started to become a significant part of today’s biotechnology. Immediate impacts of this new technology were observed in pharmaceutical research. Consequently, the first genetically engineered product, human insulin produced by bacteria, entered the market in 1982. Since then, more than 33 new drugs produced with recombinant DNA technology have been registered worldwide. In addition, 284 biotech drugs were in development in 1996, representing a three-fold increase since 1 9893. A similar change has been initiated in the food industry by the introduction of the FLAVRSAVR tomato in 1994 as the first genetically engineered whole food. Today, recombinant DNA technology has definitely found its way into the food industry, underlined by more than 3,600 transgenic field trials carried out by 1995. So far, 18 genetically engineered agricultural products have been approved for commercialization4. These include plants such as corn, cotton, soybeans and potatoes with improved pathogedpest resistance, herbicide tolerance and food quality’. The application of recombinant DNA technology in the flavour and fragrance industries is less advanced than in the pharmaceutical and food industries. Nevertheless, first products involving recombinant DNA technology in one or the other way have been commercialized. Today, recombinant DNA technology has also become an important part of the research activities of flavour and fragrance companies. The following chapters

12

Flavours and Fragrances

illustrate areas of the flavour and fragrance industry that will be increasingly influenced by the use of recombinant DNA technology.

2 PRODUCTION OF AROMA CHEMICALS

2.1 Natural aroma chemicals Nature is a rich source of aroma chemicals of which several thousand have been identified and chemically synthesized. With the on-going trend towards natural flavours, aroma chemicals were increasingly required to be of natural origin. Separation techniques such as extraction and distillation of natural materials are successfully used in our industry. When these could not be achieved at economic costs, enzymatic or microbial conversions are used instead. Today, fermentative processes are employed to produce many aroma chemicals such as various aliphatic and aromatic acids (e.g. 2-methylbutyric acid and phenylacetic acid), different esters (e.g. ethyl-2-methybutyrate, methylanthranilate) and lactones (e.g. y- and &decalactones). For the production of high impact chemicals such as P-damascenone, methional or ynonalactone none of the above mentioned techniques could be reasonably applied so far. These chemicals are generally found in very small quantities in plant materials making their recovery an expensive endeavour. As no microbial or enzymatic conversions are evident, other approaches are needed. In nature, these aroma chemicals are formed by specific yet sometimes unknown pathways. However, some have been recently elucidated, as is the case for hraneol and yn~nalactone~.’.With the help of recombinant DNA technology the corresponding genetic information from the original source can be isolated and subsequently transferred into a suitable host strain. This allows an efficient microbial production of natural aroma chemicals. An application where we have explored this strategy is the production of cis-3hexenol*. Natural cis-3-hexeno1, also referred to as leaf alcohol, and its esters are of high demand as they are widely used in various fruit flavours. Traditionally, cis-3-hexenol is isolated from mint terpene fractions. In the plant, cis-3-hexenol is formed from linolenic acid via the hydroperoxide and cis-3-hexenal (Figure 1). As peppermint oil fractions could not satisfy the global need for natural cis-3-hexeno1, an enzymatic route starting out from linolenic acid was established’. The fatty acid is oxidized to the hydroperoxide using e.g. soya flour containing lipoxygenase. The conversion of the hydroperoxide to cis-3-hexenal is achieved by using a fruit source such as guava that was found to contain high activities of the hydroperoxide lyase. Reduction to the leaf alcohol is finally performed by yeast cells. The drawback of this enzymatic transformation, though independent of peppermint oil fractions, is the rather large amount of fruit that has to be processed. The lyase was shown to be the rate limiting factor in the enzymatic conversion of linolenic acid to cis-3-hexenol.

The Impact of Recombinant DNA-Technology on the Flavour and Fragrance Industry

noc-

_

-

13

Linolenic add

O2

1

Lipoxygenase

OOH

i

Hydroperoxide lyase

0

["'I -o n

j

Alcohol dehydrogenase

t

ur3-hexenol

Figure 1 Biosynthetic pathway of cis-3-hexenol in plants. Three enzymes are involved in the production of cis-3-hexenol starting from linolenic acid. Various plant materials serve as sources for these enzymes allowing a conversion of the acid to cis-3-hexenol at industrial scale. As in such a reconstituted production system the hydroperoxide lyase is the rate-limiting factor, the gene. coding for this enzyme was heterologously expressed in yeast cells to yield a highly active lyase material.

We have therefore purified the linolenic-acid-hydroperoxide-lyase from banana plants, allowing the determination of four independent, internal amino acid sequences. Degenerate oligonucleotides and resulting PCR-fragments helped to isolate the structural gene for the lyase from a banana cDNA library''. Interestingly, the DNA sequence shared 44% identity with the sequence of allene oxide synthase. The latter enzyme is important in the biosynthesis of methyl jasmonate, another important key flavour impact chemical. The heterologous expression of the lyase gene helped to overcome the drawbacks of the enzymatic production route supplying highly active lyase material. As can be seen in Figure 2, higher amounts of cis-3-hexenol have been produced in the presence of the recombinant yeast cells compared to homogenized bananas. To unify all three enzymes involved in the formation of cis-3-hexenol in yeast, we also cloned and coexpressed the lipoxygenase gene. This generated an even more efficient system to produce cis-3-hexenol. In addition, it should be pointed out that lipoxygenases and lyases with different specificities have been described that are involved in the degradation of fatty acids". Heterologous expression of such genes would allow the production of other important flavour chemicals such as 1-octene-3-01 or 2,6-nonadienal.

14

Flavours and Fragrances

a :

b

.I

I:

Figure 2 GC analysis of the reaction mixtures in which the hydroperoxide of linolenic acid was incubated with a) homogenized banana material and reducing yeast cells and b) with recombinant yeast expressing the lyase gene and additional reducing yeast cells. The recombinant yeast process yielded higher amounts of cis-3-hexenol with no formation of trans-2-hexenol as side product.

Finally, applications based on our cloning of the lyase gene are not limited to microbial systems. The gene could also be transferred into plant hosts resulting in an increased formation of cis-3-hexenol/ cis-3-hexenal upon maceration of the plant. Ultimately, recombinant DNA technology could be used to enhance and to alter the flavour profile of fruits and plants by overexpressing key metabolic enzymes. As examples, strawberries high in furaneol or especially green smelling apples in which the lyase gene would be overexpressed can be imagined.

2.2 Tasty peptides Tasty peptides have been found in various food products such as meat, cheese, fish and y o g h ~ r t s l * ' ~In~ .order to improve or to boost flavours with such specific tasty peptides, they were so far synthesized either chemically or enzymatically. Both strategies, however, are not feasible for a large-scale and commercial production of tasty peptides due to high production costs14. We have therefore investigated the heterologous production of peptides using recombinant yeast strains. Several model peptides were chosen, among them an octapeptide known as beefy meaty peptide (BMP). This peptide was found in meat and was suggested to enhance the taste of beef gravy". Figure 3 shows the recombinant DNA technology approach used to produce the BMP in yeast cells. Linking the genetic information of the octapeptide to the yeast mating pheromone a-factor in a suitable expression vector allowed the secretion of the desired peptide into the culture medium. From the culture filtrate, the peptide can be easily recovered and used in a semi-

The Impact of Recombinant DNA-Technology on rhe Flavour and Fragrance Industry

15

purified form. Alternatively, intracellular accumulation of the peptide offers the possibility 16 to generate specially flavoured yeast extracts . This approach can be seen as a further improvement of yeast strains that have previously been engineered to contain high content of 5’-nucleotides (IMP, GMP).

Meat (flavor)

Peptides

L

/Yeast expression\

Purification

5 ’ -GAAGCTGAAGCTAAGGGTGACGAAGAATffTTGGCTTGA3

. . . . .a-factor..LysGlyAspGluGluSerLeuAla

Figure 3 Protease digestion of food proteins results in the formation of tasty peptides. Their organoleptic character and amino acid sequence can be determined after purification. As a new production strategy to obtain such peptides, yeast cells were transformed with an engineered yeast secretion vector. Upon induction, these cells started the synthesis of a fusion protein consisting of the desired peptide (in our case BMP) and the yeast a-factor. The desired peptide was then cleaved during the secretion process and could be easily recovered from the culture fluid in a semi-purified form.

3 PLANT ENGINEERING 3.1 Improvement of the flavour and fragrance profiles Plants are a major part of our daily diet and due to their smell and taste, .are established sources for raw materials used in the flavour and fragrance industry. More than 3,000 different essential oils have been analyzed and many of them are utilized in the creation of fine fragrances or serve as starting materials for the isolation and modification of chemicals”. As an illustration, 36,000 metric tons of d-limonene are extracted annually from citrus oils’*. Other commercially used examples include 1-carvone, geraniol and also menthol, the latter with an annual sales volume of about 2 billion USD. The way plants are nowadays industrially improved was dramatically changed by the new possibilities of using recombinant DNA technology. Presently, huge efforts are undertaken, for example, to increase the content and quality of fatty acids in oil crop

16

Flavours and Fragrances

plants". Genetically engineered rapeseed underwent most field trials after potato, and in 1995 an engineered canola crop with high laurate was commercialized20*z'. Commercial examples of genetically engineered plants used in the flavour and fragrance industry are not yet known, but the example of a transgenic Pelargonium plant, commonly referred to as lemon geranium, can illustrate the potential of recombinant DNA technology22. In this example, the titre of geraniol was increased 4-fold and that of citronellol by 13-fold in the transgenic plant as compared to the wild-type. To optimally design such higher yielding plant species, an improved understanding of metabolic pathways as well as of the post-harvest biochemical reactions are required. DNA sequencing programs elucidating total plant genomes are expected to simplify the cloning of important gene sequences. 3.2 Safety aspects

The safety of genetically modified organisms has been assessed by the FDA23and the EC. Nevertheless, the public acceptance of products from transgenic plants is still rather low. This is especially the case in Western Europe where major concerns are expressed in Germany and Austria. On the other hand, little consumer reactions to genetically engineered food products have been observed in the USA. However, a recent European study showed that genetic engineering was ranked as a potential food risk similar to that of artificial food coloring and, interestingly, much safer than food irradiation or pesticide residuesz4.

4 NEW ENZYMES

The majority of industrial enzymes are used today in food preparations and in fabric care products. Both markets represent roughly a 160 million USD turnover annually2s. Therefore development of new enzymes is targeted mostly at these two segments. These enzymes also find limited applications in the flavour and fragrance industry. Examples include proteases for the peneration of food protein hydrolysates and lipases for the production of natural esters 6'27. In the past, many enzymes involved in the generation of flavours or flavour precursors have been characterized. It is a well known fact that during the post-mortem aging various hydrolyzing enzymes are released within the meat. This results in the formation of flavour precursors that are characteristic to the type of mea3'. The use of such enzymes in the flavour industry is limited as they are not available at reasonable costs. For the time being it does not appear that enzyme manufacturers will produce them due to the rather small market. The advent of recombinant DNA technology, however, has now great1 facilitated their large scale production rendering it feasible for flavour companies as welJ9. With the availability of such enzymes more authentic meat, cheese and other flavour mixtures could be generated. A first development in this direction can be seen in the area of enzyme-modified cheese. Treating milk proteins with commercially available proteases often results in bitter products. Though the occurrence of the bitter peptides has been extensively studied, screening of various commercially available proteases and mixtures thereof was needed to prepare pleasant, non-bitter flavours. As cheese flavours are the product of microbial

The Impact of Recombinant DNA-Technology on the Flavour and Fragrance Industry

17

activities, extracellular enzymes from various starter cultures have been characterized. Interestingly, some of them were shown to have debittering activities3'. Such proteases removing off-flavours are of great interest to flavour industries which have started to clone and express the corresponding genes3'. This opens new avenues to making for example cheese flavours more characteristic and intensive. In recent years, new classes of enzymes such as oxidases (peroxidases and polyphenoloxidases) have been introduced for food and detergent applications. Oxidases are also important for the production of many different flavours. As an example, the oxidative degradation of amino acids yields various flavour aldehydes. L-amino acid oxidase (LAO) deaminates various amino acids, resulting in the formation of the corresponding keto acids which after decarboxylation yield flavour aldehydes (Figure 4). We have recently cloned the LAO gene from the filamentous fungus Neurospora crassa and overexpressed it homologously in the parent host3*. Alternatively, flavour aldehydes can also be produced by decarboxylation of amino acids and deamination by a rnonoamine oxidase. We have purified a novel monoamine oxidase (MAO) from Aspergillus niger and cloned the structural gene33. This FAD containing enzyme oxidizes various amines such as phenethylamine and methylthiopropylamine to the corresponding aldehydes. The gene coding for M A 0 was heterologously expressed in Escherichia coli. Incubating the above mentioned amines with protein extracts of such induced E. coli cells resulted in the formation of rnethional and phenylacetaldehyde. The broad substrate specificity makes this enzyme also attractive for the generation of various other flavour aldehydes. L- Amino Acid R NH2-CH-COOH

4

-7-

H202 + NH3

I

C-COOH

I1 I1 0

NH2-cH2

H20 + 0 2

LAO

R

I

coz

(L-Amino Acid Oxidase)

4

MA0 (Monoamine Oxidase

H202 + NH3

Decarboxylase

R

I -T b

Keto Acid

Figure 4

R

Decarboxylase

I

H20 + 0 2

Amine

Flavour aldehydes such enzymatically formed from for LAO and M A 0 have allowing the formation of precursor.

CHO

Flavour Aldehyde

as methional or phenylacetaldehyde were the corresponding amino acid. The genes coding been functionally expressed in microbial hosts flavour aldehydes when fed with the required

18

Flavours and Fragrances

5 MOLECULAR OLFACTION AND TASTE A rather high impact of recombinant DNA technology can be expected in the field of molecular olfaction and taste. The first putative olfactory receptors were cloned in 199134. Since then the understanding of olfactory receptors and their signal transduction mechanisms has been drastically increased. It became widely accepted that olfactory receptors belong to the G-protein coupled seven transmembrane receptor family3’ that represent 60% of the targets for all drugs sold today. A possible interaction of odorants with a heterologously expressed mammalian receptor has been suggested in the case of lilialTMand lyralTM36. Such ligand-receptor models form a broad and scientific basis for the pharmaceutical industry to find new drugs targeting diseases such as AIDS, cancer or arthritis. Research in the field of molecular olfaction and taste has benefited a lot from the techniques and know-how developed for the discovery and screening of pharmaceutical drugs. Structureodor relationship studies, for example, have been widely applied in the search for novel fragrance molecules37338,but the lack of a three-dimensional structure of an olfactory receptor, so far, has hampered the efforts to model and study the odorant-receptor interactions. Thus, many recent efforts have been directed towards functional expression of olfactory receptors. A specific receptor can be expressed at the surface of a cell (Figure 5). These receptors are linked via ordinary signal transduction mechanisms to a reporter gene signaling receptor binding of a potential odorant. Such engineered screening systems are widely used in the pharmaceutical industry to test low molecular weight drugs.

0 0,

oaa o.

odorant

signal

enzyme

binding

transdudion

activity

Figure 5 Recombinant yeast cells expressing olfactory receptors at the cell surface can be used to screen a mixture of odorants. Upon specific odorant binding the intracellular signal transduction cascade is activated stimulating a reporter gene. The resulting enzyme activity correlates to the binding strength of the odorant and can be easily monitored. In contrast to olfaction, much less is known about taste receptors. Nevertheless, a simplified screening system for bitter and sweet compounds has been established. RuizAvila er 0 1 . ~isolated ~ receptors from the bovine tongue papillae, added recombinant Gprotein gustducin or transducin and incubated this reconstituted tongue receptor system in the presence of GTP-y-S and potential tastants. Trypsin digestion followed by Western blot analysis indicated if interaction between tastant and receptor occurred. This system which

The Impact of Recombinant DNA-Technology on the Flavour and Fragrance Industry

19

so far is only applicable to bovine and not to human, not only allows the molecular screening of potent tastant, but more interestingly also of taste enhancing or blocking agents. In summary, screening system based on gustatory and olfactory receptors are feasible and will certainly become part of future investigations in flavour and fragrance industries. 6 OUTLOOK The flavour and fragrance industry is now at the point where the pharmaceutical industry was 20 years ago with respect to recombinant DNA technology. At that time, recombinant DNA technology entered the pharmaceutical research without much notice. The benefit of this technology was then clearly seen with the rather sudden emergence of first products and has thus developed to become an essential part of the research and production of new drugs. However, it can be foreseen that the flavour and fragrance industry will go through a similar phase until the commercial benefit of recombinant DNA technology is clearly recognized. It is already evident that recombinant DNA technology will become an important tool for the discovery and production of cheaper and novel flavour and fragrance chemicals. Furthermore, the technology is essential to ultimately advance our understanding of olfaction and taste. This will lead to the discovery of novel odor and taste modifying compounds, changing the way flavour and fragrance compositions will be formulated in the future. References

‘ ’ ’ * 10

11 12

13

14

‘5

16

Praeve P. e t a / . ,In: Fundamentals of biotechnology, Weinheim; Deerfield Beach, FL , 1987,1. Janssens L. eta/., Production of flavours by microorganisms, Proc. Biochem. 1992,27, 195. Facts & Figures, PhRMA facts April 1997. James C. and Krattiger A.F., Global review of the field testicg and commercialization of transgenic plants. ISAA Briefs No. 1 lsBBB Ithaca, NY, 1996,31. Beck C.I. and Ulrich T., Biotechnology in the food industry. Bioflechnology 1993,11, 895. Zabetakis 1. et a/.,The biosynthesis of 2,5-dimethyl-4-hydroxy-2H-furan-3-one and its derivatives in strawberry. In: Flavour science. 8th Weurman Symposium, 1996,90. Tress1 R. eta/., Formation of y- and Glactones by different biochemical pathways. In: Flavour science. 8th Weurman Symposium, 1996,141. Patent pending, Givaudan Roure. Gautier A. eta/., Firmenich patent 951215 UPC 940418. HBusler A. and Schilling B., Future impact of recombinant DNA technology on the production of natural aroma chemicals. Proceedings of the 5th Wartburg Symposium, Germany, 1997,in press. t!atanaka A., The fresh green odor emitted by plants, Food. Rev. Int. 1996,12,303. Aristoy M.C. and Toldra F., Isolation of flavour peptides from raw pork meat and dry-cured ham. In Food flavours: generation, analysis and process influence, Elsevier Science B.V., 1995. Mojarra-Guerra S.H. et a/., Isolation of low-molecular-weight taste peptides from Vacherin Mont d’Or cheese. J. Food Sci. 1991,56,4. Gill 1. et at., Biologically active peptides and enzymatic approaches to their production. Enz. Microb. Techno/.1996,18,162. Spanier A.M., BMP: a flavor enhancing peptide found naturally in beef. Its chemical synthesis, descriptive sensory analysis, and some factorsaffecting its usefulness. In: Food flavors: generation, analysis and process influence, Elsevier Science B.V., 1995,1365. Patent pending, Givaudan Roure.

20 l7 18

I’ 20

21

22

23

24

25

26

27 28

29

30

31

32 33

35

37

38

39

Flavours and Fragrances

Cheetham P.S.J., The flavour and fragrance industry In: Biotechnology: the science and the business, Harwood Academic Publishers, 1991, 26, 481. Nonino E.A., Where is the citrus industry going? Perfumer 8 Navourist 1997, 22, 53. Murphy D.J., Engineering oil production in rapeseed and other oil crops. TlBTECH 1996, 14, 206. Goy P.A. and Duesing J.H., From pots to plots - genetically-modified plants on trial, Bioflechnology 1995, 13, 454. Liu K. and Brown E.A.. Enhancing vegetable oil quality through plant breeding and genetic engineering. f o o d Technology 1996, 11, 67. Pellegrineschi A. et a/., Improvement of ornamental characters and fragrance production in lemon-scented geranium through genetic transformation by Agrobacterium rhizogenes. Bioflechnology, 1994, 12 ,64. Hallagan J.R. and Hall R.L., Safety assessment of flavour ingredients produced by genetically modified organisms. ACS Symp., 1995, 605, 59. Hoban T.J., Consumer acceptance of biotechnology: an international perspective. Nature Biotechnology 1997, 15, 232. Wrotnowski C.. Unexpected niche applications for industrial enzymes drives market growth. Genetic engineering news 1997, 17, 14. Lieske B. and Konrad G., Protein hydrolysis the key to meat flavouring systems. Food Rev. Intern. 1994, 10, 287. Lecointe C. etal., Ester synthesis in aqueous-media in the presence of various lipases, Biotechn. Lett.1996, 18, 869. Spanier A.M. and Miller J.A., Role of proteins and peptides in meat flavor. ACS Symp. Ser. 1993, 528, 78. Headon D.R. and Walsh G.. The industrial production of enzymes. Biotechnology advances 1994, 12,635. Izawa-N et a/., Debittering of protein hydrolysates using Aeromonas caviae aminopeptidase, J. Agr. Food, 1997,45, 543. Quest Int. BV, Unilever patent EP 565172 A l . Niedermann D. and Lerch K., Molecular cloning of the L-amino-acid oxidase gene from Neurospora crassa, J. Biol. Chem. 1990, 265, 17246. Schilling B. and Lerch K., Cloning, sequencing and heterologous expression of the monoamine oxidase gene from Aspergillus niger, Mol. G. Genet. 1995, 247, 430. Buck L. and Axel R., A novel multigene family may encode odorant receptors - a molecular basis for odor recognition, Cell 1991, 65, 175. Shepherd G. eta/., Olfactory receptors - a large gene family with broad affinities and multiple functions. Neuroscientist 1996, 2, 262. Raming K. etal., Cloning and expression of odorant receptors. Nature, 1993, 361, 353. Bajgrowicz J. and Broger C., Molecular modelling in design of new odorants: scope and limitations. Proceedings of the 13th Congress of flavours, fragrances and essential oils,1995.3,1 Rossiter K.J., Structure-odor relationships. Chem. Rev. 1996, 96, 3201. Ruiz-Avila L eta/., Coupling of bitter receptor to phosphodiesterase through transducin in taste receptor cells. Nature 1995, 376, 80.

-

The Design and Synthesis of Novel Muguet Materials K. J. Rossiter QUEST INTERNATIONAL, ASHFORD, KENT TN24 OLT, UK

1 ABSTRACT

A qualitative structure activity approach, involving the hybridisation of two fragments found in two different known muguet odourants, led to the identification of several novel muguet smelling 3-alkoxypropan-1-01s. Several analogues of the best muguet lead, 2,2,7trimethyl-4-oxaoctan-1-ol,were synthesised and assessed in an attempt to better understand the structural requirements for a mupet odour. The effect of the methyl substitution pattern at C1, C2 and C3, and replacement of the ether linkage by an olefinic linkage, a triple bond and a methylene group were investigated. It was found that C1 and C3 should ideally be unsubstituted and that C2 should be dimethylated. Floral character was retained when the ether oxygen atom was replaced by a saturated or olefinic carbon. The muguet note appeared to be more pronounced in the 3-alkoxypropan-1-01 and the trans-alkenol, with the floral odour of the cis-alkenol and the fully saturated analogue shifting more in the direction of rose.

2 INTRODUCTION Since flowers from the muguet (lily of the valley) plant are very small and difficult to extract, it is impossible to produce a commercially viable blossom oil from this plant. Therefore, perfumers have to rely on synthetic substitutes, such as cyclamen aldehyde (l), Bourgeonal@(2), Lilial@(3), hydroxycitronellal (4), Lyral@( 5 ) and Dupical@(6) to create this odour type (Figure l), none of which have yet been identified in the rnuguet flower. In fact the odour of these materials is somewhat heavier than that of the living flower, which is soft and quite rose like, but because of their widespread use in perfiunery, the consumer and also the perfumer now use the term muguet to describe odours which are similar to these aroma chemicals rather than to the odour of the flower. The aldehydic muguet ingredients were among the first to be discovered and are still highly valued in perfumery today. Since all of the early muguet odourants were aldehydes the presence of the aldehyde functional group was originally believed to be a prerequisite for a muguet odour. However, today there are also a number of alcohols which can be used to create a muguet odour and these include Majantol@(7), Florosa Q@ (S), famesol (9),

22

Flavours and Fragrances

(1)

(2)

Cyclamen aldehyde

Bourgeonal@ (Quest Int.)

Lilial@ (Givaudan)

LyraP ( I F F )

Dupical@ (Quest Int.)

(3)

I

OH (4)

Hydroxycitronellal

Figure 1 Synthetic aldehydic muguet odourants

2,6-dimethylheptan-2-01 (lo), and Mayol@ (1 1) (Figure 2). Farnes01'**~~ and dihydrofarne~ol~'~ have actually been found in the plant oil. The two classes of muguet odourants have quite different odour profiles. For example, Figure 3 shows the odour aspects which are significantly different for seven representative muguet materials. The three 3-(p-alkylphenyl)-propanals, cyclamen aldehyde (I), Bourgeonal@ (2), and Lilial@ (3), are more fruity with melon aspects; the two alcohols, Florosa Q@(8) and Mayol@(1 l), tend to have prominent herbal, lavender, pine and sandalwood notes; while the two materials which contain both an alcohol and aldehyde hnctional group, hydroxycitronellal (4) and Lyral@(9, are perceived as sweet. Pelzer et al.' studied the odour profiles of 181 substances possessing a typical muguet scent and found that the carbonyl compounds, in addition to the muguet aspects, exhibited lime blossom notes.

LO" q

OH

\

(7)

(8)

Majantol@ (Wacker)

(9)

Florosa Q@ (Quest lnt.)

Farnesol .OH

f (10)

Q

2,6-Dimethylheptan-2-01

Mayol@ (Firmenich)

Figure 2 Synthetic alcoholic muguet odourants

(11)

23

The Design and Synthesis of Novel Muguet Materials

Florosa Q H cyclamen aldehyde Lyral

1

hydroxycitronelkl

/

SWEET

1

H Lilial

Figure 3 Diflerences in odour profiles of muguet odourants

In more recent years, attention has been focused towards the discovery of nonaldehydic muguet materials. This is because of the instability associated with certain materials from this chemical class. Common side reactions include oxidation to the corresponding carboxylic acid and, in some cases, the corresponding lower homologue ketone; reaction with alcohols to form the corresponding acetals; aldol condensations; and the formation of aldehyde trimers (Figure 4). In general, the more hostile the product base the less stable the fragrance ingredient. For example, aldehydic ingredients are notoriously unstable in antiperspirant formulations which are strongly acidic due to partial hydrolysis of active antiperspirant agents such as aluminium chlorohydrate.

oxidation

/

m

Figure 4 Chemical instability of aldehyde group

acetal formation

Flavours and Fragrances

24

The histogram in Figure 5 shows the relative organoleptic stability of five muguet ingredients in an aerosol antiperspirant. Two factors are considered when determining the organoleptic stability of an ingredient in a product base. The first is the degree to which the ingredient covers the base and the second is the persistence of performance over time. Thus in Figure 5 the taller the bar the more stable and better performing the ingredient is. The poor to moderate organoleptic performance of Lilial" (3), hydroxycitronellal (4),and Lyral" ( 5 ) is predominantly attributed to the chemical instability of the aldehyde group. For example, our studies have shown that the percentage of Lilial" remaining in an ethanol based roll-on antiperspirant after 4 weeks at 37°C is only 15%. Lilial" reacts with the ethanol in the base to form the corresponding diethyl acetal and undergoes oxidation. Bourgeonal" (2), on the other hand, performs well for a combination of reasons. Although Bourgeonal" reacts with the ethanol to form the corresponding diethyl acetal, it does not appear to undergo autoxidation. Consequently, the level of Bourgeonal" remaining after 4 weeks at 37°C is typically 65% as op osed to only 15% for Lilial". Another reason for the superior performance of Bourgeonal IS that it is a very potent material being 2 to 4 times stronger than Lilial" at the same concentration.' Florosa Q" (S), the only alcoholic muguet odourant in Figure 5, performs well because it is chemically very stable. Ether groups, apart from the possibility of peroxide formation, are relatively inert, and although possible side reactions of the alcohol group will depend upon the specific formulation of the antiperspirant, there is relatively little it could do other than acid catalysed dehydration, and this usually requires fairly high temperatures or lower pH. Consequently the level of Florosa Q" in, for example, an ethanol based roll on antiperspirant after 4 weeks storage at 37°C remains unchanged. The good as opposed to very good rating is a sensory effect.

8.

Organoleptic Stability

7

Hydroxycitmnellal Lilial

Lyml

Bourgmna Florosa

Figure 5 Organoleptic performance in an aerosol antiperspirant (3 7 ° C 12 weeks)

25

The Design and Synthesis of Novel Muguet Materials \

(7)

Majantol@ (Wacker)

/

(12) Reference 6

(13) Quest Research Material

Figure 6 Incorporation of the 2,2-dimethylpropan-l-oIJi.agment 3 DESIGN, SYNTHESIS AND ODOUR EVALUATION OF NOVEL MUGUET INGREDIENTS 3.1 Lead Identification The aim of the present study was to find a novel muguet odourant which is stable and which could potentially be prepared cost effectively on an industrial scale. It is clear from the above that the best chances of finding such a material lies within the alcohol chemical class. From a list of known muguet odourants it was observed that a few members (for example, 7, 12 and 13 in Figure 6) had one structural feature in common, the presence of a 2,2-dimethylpropan-l-ol fragment (14). One easy way of incorporating this fragment into a molecule is by the reductive cleavage of 5,5-dimethyl-ly3-dioxanes (IS), which in turn can be easily prepared from a carbonyl compound and the readily available and inexpensive 2,2-dimethylpropane-1,3-diol(Scheme 1). On a laboratory scale the most convenient way of reductively cleaving an acetal or ketal is by the use of reducing agents such as BH3 in tetrahydrofuran' or a combination of LiAIH4 and AIC13 in diethyl ether.' The latter reduces cyclic acetals in virtually quantitative yields and as such was the procedure adopted for this work. On an industrial scale, however, the preferred cheapest and safest rocess would be hydrogenolysis in the presence of an acid and a metal catalyst?*"*

P'

The target compounds 16 contain the 2,2-dimethylpropan-l-ol fragment connected via an ether linkage to the carbon skeleton of the starting carbonyl compound. The ether oxygen atom provides an area of high electron density at position four and as such may fulfil a similar role to the double bond and the phenyl ring often found at this position (see compounds 1,2,3,5,6,7 and 8).

26

Flavours and Fragrances

a

Known muguet odourant Two desiredfragments

Starring materials used to incorporare HO above fragments in target hydroxy ether

target compound

Figure 7 Design of novel muguet odourants using the hybrid approach The criteria for selection of the carbonyl substrates was not only availability and price, but also, and more importantly, upon the similarity of their carbon skeleta to those of known muguet odourants. For example, the tricyclic fragment in Dupical@( 6 ) can be attached to the 2,2-dimethylpropan-I-o1 fragment (13) by the use of pketotricyclo[5.2.1 .02.6]decane (17). The resulting hydroxy ether (18), in addition to containing two fragments found in known muguet odourants, is also of a similar size to Dupical@(Figure 7). By using this qualitative structure activity approach it was anticipated that the chances of finding materials of the desired odour type would be increased. Ten 2,2-dimethyl-3-alkoxypropan1 -ols, which had been identified as target molecules using this hybrid approach, were prepared and their organoleptic properties assessed. A few of the bigger molecules, such as compound 18, had woody type notes, but the majority (70%) were floral in character and, although rose appeared to be the main floral aspect, three out of the seven floral odourants were also described as having muguet connotations (compounds 19-2 1, Figure 8).

The Design and Synthesis of Novel Muguet Materials

27

(20) Novel 2,2-dimethyl-3-oxa-propan-l-ol

(12)

(21) Reference 12

Known muguet odourant

Figure 8 Novel 2,2-dimethyl-3-oxa-propan-l-ols possessing muguet odours

3.2 Lead Optimisation Pelzer et al.’ have suggested that the substitution pattern, particularly around the functional group, is an important criterion for alcohols and aldehydes to smell of muguet (Figure 9). Pelzer’s analysis of 73 alcohols produced, not only the distance constraints shown in Figure 9, but also the following rules: 1) C1 should be substituted by one to three alkyl groups, ideally three, provided that the alcohol group is not overshadowed too strongly by steric hindrance.

2) C2, C5, and to a lesser extent C6 and C7 should be substituted by a single alkyl group, ideally a methyl group. Dimethylation generally has a detrimental effect on the fragrance impression. 3) Where a double bond is present it should preferably be at C4 or C6. A double bond between C3 and C4 or, to a lesser extent between C2 and C3 generally has a negative influence on the odour.

Cl-C4 = 3.2 f Cl-Cg = 4.0 f Cl-Cg = 5.0 f Cl-Rl =4.7* Figure 9 Pelzer ’s muguet fiagment for alcohols

0.3A O.3A 0.3A 0.4A

28

Flavours and Fragrances

3.2.1 Variation of the Methyl Substitution Pattern. In order to investigate the effect of introducing a methyl group at C1 for the 3-alkoxypropanol family, compound 19 was oxidised to the corresponding aldehyde (23) which was then subjected to a methyl Grignard reaction to yield 3,3,8-trimethyI-5-oxanonan-2-01(24) (Scheme 2).

(23)

Scheme 2

An easy way of changing the methyl substitution pattern at C1, C2 and C3 is to use a range of different 1,3-diols in the formation of the 1,3-dioxane precursors (Scheme 1). Therefore three analogues (compounds 25-27) of the best muguet lead (1 9) were prepared by reacting 3-methylbutanal with 2-methylpropane-l,3-diol, propane- 1,3-diol and 2methylpentan-2,4-dioI (hexylene glycol), all of which are cheap and readily available. In the case of hexylene glycol, where the diol is unsymmetrical, cleavage of the two different C - 0 bonds will give rise to two different products (27). The odour properties of these four hydroxyethers along with that of the parent compound 19 are listed in Figure 10. By comparing the odours of these five compounds, it was concluded that in this series of alkoxyalcohols dimethylation at position two is very important for a good muguet odour. Replacement of the methyl groups by hydrogen leads to a gradual decrease in the muguet character and the introduction of a fruity note. For example, compound 25, which has only one methyl group at position two, has a slightly more fruity and much less intense odour than 19. Removal of both methyl groups leads to complete destruction of the muguet character and a strengthening of the fruity note (26 and 27). The organoleptic properties of the individual isomers of 27 formed from the ring opening of 4,4,6-trimethyl2-(2-methylpropyl)-l,3-dioxanewere assessed by gc-sniffing and found to be very similar. Introduction of a methyl group at position one, whilst retaining dimethylation at C2, resulted in a very weak muguet odourant 24. In fact these findings contradict the aforementioned rules of Pelzer el ~ 1 They . ~ state that dimethylation at C2 generally has a detrimental effect on the fragrance impression, whereas in this series the dimethyl analogue 19 is the best with respect to both odour quality and intensity. They also suggest that C1 should be substituted by one to three alkyl groups, ideally three, provided that the polar group is not overshadowed too strongly by steric hindrance. From this statement one might expect 3 alkyl groups to be better than 2, which in turn are better than one. However, in this study the converse was found to be true with the secondary alcohol 24 being less intense than the primary alcohol 19. It is possible that the presence of both the methyl group at C1 and the dimethyl substitution at C2, in combination, lead to steric hindrance of the alcohol group and that this is the reason why 24 is much weaker than 19. Thus the rules of Pelzer and co-workers do not appear to apply to 3-alkoxypropanols. However, since they do not list all of the alcohols used in their study, it is impossible to tell whether or not they included ethers in their data set, or whether they restricted their set to compounds where the only functional group is hydroxyl.

The Design and Synthesis of Novel Miiguet Materials

29 ( I 9) Muguet, rose

(24) Very weak, muguet, rose, fruity

To+oH

( 2 5 ) Weak, fruity, rose, muguet

(26) Weak, fruity, woody T

O

A

o

H

TomoH+ Ton OH

(27) Fruity, green, herbal, woody

Figure 10 The effect of the methyl substitution pattern on the odour of 3-alkoxypropanols 3.2.2 Replacement of the Ether Linkage by a Double Bond. It is generally known within the perfiunery industry that the introduction of unsaturation in a molecule can lead to an increase in odour inten~ity.'~ Therefore the effect of replacing the ether linkage by an olefinic linkage was investigated (Figure 11). Since this structural change introduces rigidity into the molecule, one would expect, as is often the case with alkene geometric isomers, that cis- and trans-2,2,7-trimethyloct-4-en1-01 would have different odours. Molecular modelling was used to predict whether either of the isomers of 2,2,7trimethyloct-4-en- 1-01 (28) would have a muguet odour. 3.2.2.1 Modelling Work. Florosa Q@ was chosen as the muguet standard for the modelling work for two reasons. Firstly, it is relatively rigid conformationally and thus relatively easy to model, and secondly, the two isomers of Florosa Q@ are known to have different organoleptic properties. Sommer and GUntertl4 have reported that only the cisisomer, in which the hydroxyl and isobutyl groups are both equatorial, is the actual fragrance carrier. Our own findings suggest that both isomers have a muguet odour but that the cis-isomer is much more potent than the trans- (Figure 12). The two isomers of Florosa Q@ were separated by column chromatography and a series of dilutions in ethanol prepared. Forty nine people were asked to smell the dilutions from smelling strips, waiting a few minutes after the strips had been dipped for the ethanol to evaporate. They started with the lowest concentration and worked their way up to the most concentrated. They noted when they could first detect an odour, thus providing a measure of the relative odour threshold of the two isomers. Nine people could not smell the trans-isomer at the highest concentration (10% w/v) and two could not smell the cis-isomer. The results for those people who could detect an odour (40 and 47 for trans- and cis-Florosa Q@ respectively)

(19)

(284

Figure 11 Replacement of the ether group by a double bond

I

(2W

30

Flavours and Fragrances

I

fl

x'""

, ;uI

trans-Isomer

cis-Isomer Strong muguet

Weak muguet / 8 % r i c

I6

I2 6 4

0.005 0.01

0.05

0.1

0.5

I

5

0.005 0.01

10

0.05

0.1

0.5

I

5

10

lconcl at which fin1 smelt

lconcl at which Rnt smelt

Figure 12 Dgerences in odour thresholds of cis- and trans-Florosa Q" are displayed as histograms in Figure 12. From the spread of data it can be seen that the trans-isomer is first detected at higher concentrations than the cis-isomer (i.e. it is less potent). The assumption was made that compounds will have a good muguet odour if they can adopt a conformation that closely resembles that of cis-Florosa Q". The alkoxypropanols are very flexible molecules and thus 2,2,7-trimethyl-4-oxaoctan-1-01 (19), which has a very similar carbon skeleton to that of Florosa Q@,can easily adopt relatively stable cyclic-type conformations (within 1-5 kcal/mol from the energy of the straight chain conformer, 19b) which resemble that of either trans- or cis-Florosa Q@(19a and 19c respectively) (Figure 13). In these conformations the distance between the two

A

A cis-Isomer, srrong muguer 0-0Distance = 4. I7A

trans-Isomer, weak-odourless 0-0 Distance = 3.49A

a

OH OH

( I 9a) 0-0 Distance = 3.53A

(19b)

( 19c)

0-0 Distance = 4.96A

0-0 Distance = 3.96A

Emrgy=-I 18.6(-125.2) kcal/mol

Energy = -120.8 (-129.9) kcallmol

Energy = - I 19.8 (-129.2) kcallmol

Figure 13 Comparison of the conformational arrangement of 19 with that of Florosa Q" (The cyclic-type conformations of 19 were tined to the cis and trans forms of Florosa Q" using the SYBYL MULTI-FIT tool. Heats of formation were calculated using the PM3 (and AMl) methods in MOPAC.)

31

The Design and Synthesis of Novel Muguet Materials

w Figure 14 Top: superimposition of 19a and cis-Florosa Q". Bottom left: trans-2,2,7trimethyloct-4-en-1-01. Bottom right: cis-2,2,7-trimethyloct-4-en1-01. oxygen atoms are virtually identical and the isobutyl group is located in a similar position in space. It is postulated that the muguet aspects of 19 are associated with the cyclicconformation (19c). The similarity between 19c and cis-Florosa Q" is clearly seen in the top of Figure 14, where the two molecules have been superim osed (the green molecule is 19c, and the molecule coloured by atom type is cis-Florosa Q . Turning now to the alkene derivative, 2,2,7-trimethyloct-4-en-1-01 (28). Although both trans- (bottom left, Figure 14) and cis-2,2,7-trimethyloct-4-en-l-ol (bottom right, Figure 14) can adopt the postulated active cyclic-type conformation, only the trans-isomer has the isobutyl group located in a very similar position to that of 19c and cis-Florosa Q@ (top, Figure 14). It was thus predicted that trans-2,2,7-trimethyloct-4-en-I-ol would have a muguet, rose odour and that cis-2,2,7-trimethyloct-4-en-I-ol would not.

4

3.2.2.1 Synthesis of trans- and cis-2,2,7-trimethyloct-4-en-l-ol.Trans-2,2,7trimethyloct-4-en01 was successfully prepared using the route outlined in Scheme 3. This process uses Claisen chemistry to ensure trans-c~nfiguration'~ of the double bond. 2Methylpropanal was treated with vinyl Grignard reagent to yield 5-methylhex-1-en-3-ol' (29). The vinyl alcohol (29) was reacted with 3-methylpropanal to give predominantly the trans-isomer of 2,2,7-trimethyloct-4-enal(30). The aldehyde was reduced to the corresponding alcohol (28a) using sodium borohydride. The isomeric composition of the product was 95% trans and 5% cis (capillary gc, rpa).

32

Flavours and Fragrances 1-

Cis-2,2,7-trimethyloct-4-enolwas prepared in five steps (scheme 4). 4-Methylpent1-ene was brominated to give 1,2-dibrom0-4-methylpentane,which on subsequent dehydrobromination using the method of Dehmlow and ThieserI6 yielded 4-methylpent-I yne. The latter was coupled with the THP ether of 3-bromo-2,2-dimethylpropan-l-ol according to the procedure of Schwarz and Waters". Deprotection yielded 2,2,7trimethyloct-4-yn- 1-01 (3 l), which was hydrogenated under Lindlar conditions to give the desired cis-alcohol (28b). For comparative purposes the fully saturated analogue, 2,2,7-trimethyloctan-1-01 (32) was also prepared by catalytic hydrogenation of trans-2,2,7-trimethyloct-4-en-l-ol. The organoleptic properties of 2,2,7-trimethyl-4-0xaocten-l-ol(19), trans-2,2,7trimethyloct-4-en- 1-01 (28a), cis-2,2,7-trimethyIoct-4-en-1-01 (28b), 2,2,7-trimethyloct-4yn-1-01 (31) and 2,2,7-trimethyloctan-l-ol(32) were compared (Figure 15). The odours of these samples were assessed as 1% and 10% solutions in diethyl phthalate by a panel of perfumers and fragrance chemists. The organoleptic purity of each sample was checked by gc olfactometry.

(31)

Scheme 4

The Design and Synthesis of Novel Muguet Materials

33

(19) Muguet, rose

(28a) Muguet, rose, woody

(28b) Rose, geraniol-like,rnuguet

(31) Weak, minty (minty note from allene impurity)

(32) Rose,geraniol-like, muguet

Figure 15 Replacement of the C-0 linkage and its effect on odour

All of the materials were floral in odour, except for 2,2,7-trimethyloct-4-yn-l-ol (31) which was described as weak and minty. This note was shown, by gc olfactometry, to be due to the presence of a trace amount of the corresponding allene. The odour descriptors given to the remaining four compounds varied significantly from subject to subject thus making it very difficult to draw definite conclusions about the effect of structure on the muguet character of these alcohols. The trans-isomer 28a and the alkoxyalcohol 19 were generally described as having a muguet odour with some rose notes, which was consistent with the modelling predictions. The trans-alkene also possessed some woody character. The two materials which were described by the widest range of descriptors were the fully saturated alcohol and the cis-alkenol. Some assessors perceived these as being very muguet like whilst others perceived them as being definitely rose and geraniol like. The fact that these two compounds presented the greatest problems with regard to subjectivity, could possibly be explained by the fact that their structures do not satisfactorily meet the requirements for a muguet odour. The saturated compound 32 is also more conformationally mobile, thus allowing it to adapt to fit a wider range of receptors. Flexible molecules tend to have complex odour profiles and as such are often described using a number of odour descriptors. These molecules can adopt a large number of energetically favourable conformations each of which may be responsible for triggering a different odour response. For example, Yoshii et investigated the stable conformations of (R)-ethylcitronellyl oxalate and found that the most stable compact conformations fitted their previously published benzenoid musk model. One of the stable conformations partially resembled one conformer of S-citronellol, a rose odourant. They concluded that these conformations could be responsible for ethyl citronellyl oxalates main odour quality (musk) and its secondary odour quality (rose) and that other notes, such as woody and fresh, might be explained by further conformational comparisons with other structure-odour models. Rigid compounds on the other hand usually have a well-

34

Flavours and Fragrances

defined odour which can be described using only one or two words (odour descriptors). They are thus easy to classify and are also relatively easy to model. It is for these two reasons that structure-odour relationship studies have been concentrated on odour groups such as ambergris, bitter almond, musk and sandalwood rather than odour groups such as floral, fruity and green. A recent review on the field of structure-odour relationships is provided by Rossiter.”

4 CONCLUSIONS

1) Several novel muguet-smelling 3-alkoxypropan- 1-01s were identified using a hybrid qualitative structure activity relationship approach.

2) In the 3-(3-methylbutoxy)propan-1-01 series it was found that dimethylation at position two is very important for a good muguet character. Successive replacement of the methyl groups by hydrogen leads to a gradual decrease in the muguet character and the introduction of a fruity note. This seems to contradict earlier SAR models. 3) Floral character is retained when the ether oxygen atom is replaced by a saturated or olefinic carbon. The muguet note appears to be more pronounced in the 3alkoxypropan- 1-01 and the trans-alkenol, whereas the cis-alkenol and the fully saturated analogue appear to be more rose-like. 4) Common problems encountered with the development of structure-odour relationships

were exemplified by this work. These include the subjectivity of odour, the importance of organoleptic purity, and the dilemmas associated with the modelling of conformationally flexible molecules. Acknowledgements

I would like to thank the following colleagues from Quest International. Ian Payne for the organoleptical stability data of muguet ingredients, Steven Rowland for analysis of the breakdown products of fragrance ingredients in antiperspirant base, and Kerry McInerney and Anne Richardson for the sensory profiling work. I would also like to thank Joe Metcalfe (Oxford University) and Karin Rose (Portsmouth University) who, during undergraduate placements at Quest International, helped with the synthesis of the 3-alkoxypropan-1-01s and the determination of the relative odour thresholds of the ‘8 Florosa Q isomers respectively.

References

1. 2. 3.

M. Boelens, H. J. Wobben and J. Heydel, Perfum. Flavor., 1980,5 (6), 2. D. P. Anonis, Perfum. Flavor., 1987, 11 (6), 31. E. J. Brunke, F. J. Hammerschmidt, F. Rittler and G . Schmaus, SOFW J., 1996 122 (9), 593.

The Design and Synthesis of Novel Muguet Materials

4.

5.

6. 7. 8.

9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20.

35

H. Surburg, M. Guentert and H. Harder, “Recent Developments of Flavour and Fragrance Chemistry”: Proceedings of the 3rd International Haarmann & Reimer Symposium, VCH Publishers, Weinheim, Germany, 1993, p. 103. R. Pelzer, U. Harder, A. Krempel, H. Sommer, H. Surburg and P. Hoever, “Recent Developments of Flavour and Fragrance Chemistry”: Proceedings of the 3rd International Haarmann & Reimer Symposium, VCH Publishers, Weinheim, Germany, 1993, p. 29. C. G. Cardenas, H. M. Hoffmann and B. J. Kane, Perfum. Flavor., 1993,lS (l), 11. H. I. Bolker and B. I. Fleming, Can. J. Chem., 1974,52,888. E. L. Eliel, V. G. Badding, and M. N. Rerick, J. Am. Chem. Soc.,1962,84,2371. W. L. Howard, J. H. Brown Jr., J. Org. Chem., 1961,25, 1026. P. A. Gorin, J. Org. Chem,, 1959,24,49. E. F. M. Ghenassia and A. J. Lakodey, PC UK PRODUITS CHIMIQUES UGIINE KUHLMANN, European Patent 0092 463 Al, published 26/10/83 C. Anselmi, M. Centini, M. Mariani, A. Sega and P. Pelosi, J. Agric. Food Chem., 1992,40,853. P. A. Edwards and P. C. Jurs, Chem. Senses, 1989,14 (2), 281. H. Sommer and M. Giintert, Haannan & Reimer Contact, 1993,59, 9. P. Vittorelli, T. Winkler, H. J. Hansen and H. Schmid, Hefv. Chim. A m , 1968, 51, 1456. E. V. Dehmlow and R. Thieser, Tetrahedron, 1986,42 (13), 3568. M. Schwan. and R. M. Waters, Synthesis, 1972,2,567. F. Yoshii, S. Hirono and I. Moriguchi, Quant. Struct. Act. Relat., 1994, 13, 144. F. Yoshii, S. Hirono, Q. Liu and I. Moriguchi, Chem. Senses, 1992, 17,573. K. J. Rossiter, Chem. Rev., 1996,96 (8), 3201.

Aura of Aroma@:A Novel Technology to Study the Emission of Fragrance from the Skin Braja D. Mookerjee”, Suba M. Patel, Robert W. Trenkle and Richard A. Wilson INTERNATIONAL FLAVORS & FRAGRANCES INC.. I5 15 HIGHWAY 36. UNION BEACH, NJ 07735, USA

1. INTRODUCTION It is a common belief that people perceive a fragrance by the relative volatility of its components. Thus, a fragrance is described as having a topnote, a middle note, and a bottom note. A new technology has now developed to prove that this is not so. A fragrance is actually perceived by the diffusion of molecules which is an inherent property of the compounds and is independent of their molecular weight, boiling point, and odour threshold or odour value. This technology is called Aura of Aroma@ and has been trademarked by IFF. By means of this technology we have shown, for the first time that, depending upon the fragrance, various skins may or may not have a significant effect on the emission of fragrance molecules. Both the technology and the results will be discussed in detail. 1.1 What is “Aura of Aroma@?”

Figure 1 shows an actual photograph of an eclipse of the sun taken from National Geographic magazine. When the sun is totally eclipsed by the moon, the surrounding glow is called the “Aura.” Similarly, if we consider a drop of fragrance, just as in the case of the sun in the photo, the molecules surrounding the drop form an Aura of that particular fragrance. It is a common belief that a fragrance smells layer by layer. For example, one smells first the most volatile components, called the “Top Note”, then one smells the “Middle Note” which contains the components with boiling points in the middle range, and finally one perceives the highest boiling molecules which constitute the “Bottom Note”. In reality, this is not the case. When a drop of fragrance is placed on the skin, several selected molecules from the lowest boiling to the highest boiling, irrespective of their molecular weights, boiling points, and vapour pressures, come into an “Aura,” eventually hit our nose, and give our first impression of that particular fragrance. The composition of this Aura is dependent on a characteristic property of each fragrance molecule known as its “diffusivity.”

*Correspondence to: Dr. Braja D. Mookerjee.

Aura of Aroma": A Novel Technology KO Study the Emission of Fragrance from the Skin

Figure 1. Eclipse of the sun. (Taken from National Geographic magazine)

Figure 2. Dendrobium superbum orchid

31

38

Flavours and Fragrances

1.2. What is “Diffusivity?”

Diffusivity is the inherent property of a compound to emit its molecules into the air. One compound is said to be more diffusive than another if its molecules tend to pass into the air to a greater extent than those of the other compound. This is called “relative diffusivity.” Diffusivity is independent of boiling point, molecular weight, odour threshold, or odour value. This phenomenon is true not only for a fragrance placed on the skin but also for a living flower’s fragrance which forms an Aura. This Aura is actually due to the aroma molecules which are constantly being produced by the living flower and are coming off the surface of the petals.

2 AURA OF AROMA@THE EVOLUTION OF LIVING FLOWER@ TECHNOLOGY. Indeed, the Aura of Aroma@ technology evolved naturally from the Living Flower@ technology. In this connection, we have now developed a novel technology to capture the aroma molecules surrounding the flower petals without touching the flower or any part of the plant. This is called SPME, Solid-Phase Micro-Extraction. To our knowledge, no one has ever before applied this technique to the analysis of the Living Flower@aroma. Figure 2 shows the Dendrobium superbum orchid, now growing in our greenhouse, which comes from New Guinea and is unique in the world because of its raspberry-like fruity odour. The SPME needle, which is nothing more than a 2-3 mm glass fibre coated with a highboiling liquid adsorbent, is placed in close proximity to the flower without touching it and is kept there for a period of from 112-1 hour depending on the odour strength of the blossom. The aroma molecules around the petals are adsorbed onto the fibre which is then analysed by GUMS to give the aroma profile of that particular flower. The data in the first column of Table 1 shows the composition of the living orchid fragrance, which represents more than 98% of its aroma profile. When we reconstituted this fragrance we discovered that it lacked the diffusivity of the Dendrobium orchid. We then analysed the headspace volatiles of this recreated living orchid fragrance. This study revealed the very interesting phenomenon that the headspace composition of a living flower fragrance consists of several highly diffusive molecules the nature of which is independent of their molecular weight, boiling point, and odour strength. For example, benzyl acetone, with the lowest molecular weight did not increase at all in the headspace, whereas linalool which has a higher molecular weight than benzyl acetone, dramatically increased more than 15 times. Benzyl acetate increased 25 times, while 2-tridecanone, a so-called Middle Note component based on its molecular weight and boiling point, increased 275 times. Oxyphenylon, another Middle Note component and the compound responsible for the characteristic raspbeny odour decreased substantially in the Aura, while high boiling compounds like 2-pentadecanone and ethyl myristate were still

39

Aura of Aroma": A Novel Technology to Study the Emission of Fragrance from the Skin

Mol.

ComDound

Benzyl Acetone Benzyl Acetate Linalool Oxyphenolon 2-Tridecanone 2-Pentadecanone Ethyl Myristate

m

Topnote Topnote Topnote Middle Note Middle Note Bottom Note Bottom Note

148 150 154 164 198 226 256

Living Orchid Fragrance %! 0.02 0.20 2.20 1 1.90 0.02 69.00 14.80

Aura %! 0.03 5.20 34.10 1.70 5.50 33.50 8.50

Table 1. Aura of Aroma of Living Flower (Dendrobium superbum orchid).

substantially present in the Aura. When we recreated this Aura and added it back to the living orchid fragrance at a level of lo%, the fragrance now became very diffusive. 3 AURA OF AROMA@ON SKIN. Now, we want to know what happens when a true fragrance is applied to the skin. In other words, we want to study the Aura of the fragrance. When a fragrance is applied to the skin of a woman, a natural Aura of Aroma@surrounds her body. We want to study this Aura of Aroma@on skin. Now, I want to describe the technique we use to study this Aura. We have developed a very simple yet elegant technique in which 10 microliters of perfume is applied to a clean inner forearm of a woman. Immediately a small glass globe is placed over it and sealed by contact with the skin. The same SPME needle as used for Living Flower studies is inserted through a septum in the top of the vessel and positioned so that the fibre tip is approximately 1 cm above the sample area. The needle is kept in place for a period of 1/2-1 hour following which it is immediately analysed by G C M S to determine the composition of the fragrance Aura. In this connection, we would like to mention that, to our knowledge, nobody has ever used the SPME method to study the release of fragrance from skin. We have done this work at least three years ago and disclosed it in 1996. This novel technique was described in a recent trade journal called Spray Technology & Marketing' . The first fragrance we studied was the classical fragrance Shalimar created by Guerlain in 1925 (Table 2).

40

Flavours and Fragrances

w

ComDonent Limonene Linalool Linalyl Acetate Ethyl Vanillin Coumarin Methyl Ionone Musk Xylol

Topnote Topnote Topnote Middle Note Middle Note Middle Note Bottom Note

30.0 1.7 9.9 0.2 1.7 1.1 Trace

Aura on Skin

w 20.4 17.9 21.6 1.6 7.8 2.1 0.3

Table 2 Comparison of Aura of Shalimar on Skin & Fragrance Oil. You can see that limonene, the most volatile component, which constitutes 30% of the fragrance oil, is only present at 20% of the Aura, whereas linalool, which is only 1.7% in the oil increases eight times in the Aura. Similarly, Linalyl acetate also doubled in the Aura. Ethyl vanillin, a high boiling chemical which was used for the first time in Shalimar to the extent of 0.2% in the oil, dramatically increased eight times in the Aura. Similarly, methyl ionone, another high boiling compound used for the first time in Shalimar, doubled in the Aura. Most interesting is the musk xylol, the highest boiling compound which is only in trace amounts in the oil, was 0.3% of the Aura. This shows that the Aura of a fragrance is composed not only of highly volatile molecules but also that both low-boiling and higher boiling compounds are highly diffusive. Next, we studied a very successful 1990’s fragrance, Amarige, introduced by Givenchy (Table 3). The Aura studies showed that Top Note constituents like linalool, benzyl acetate and styrallyl acetate increased in the Aura, but, at the same time, medium-boiling components like cashmeran, which could not even be detected in the fragrance oil, showed up in the Aura to the extent of 0.5%. Similarly, other high-boiling compounds like bacdanol, cedramber, and is0 E super doubled in the Aura. At the same time, Hedione, a very renowned and widely used fragrance chemical, is not diffusive at all, and, in reality,

ComDonent Linalool Benzyl Acetate Styralyl Acetate Cashmeran Bacdanol Hedione Cedramber is0 E super Ambrox Benzyl Salicylate Muskalactone

m Topnote Topnote Topnote Middle Note Middle Note Middle Note Middle Note Bottom Note Bottom Note Bottom Note Bottom Note

1.7 4.9 1.2 0.0 0.2 29.9 1.5 7.1 0.2 32.5 0.9

Aura on Skin

U.hu 17.9 22.7 9.7 0.5 0.5 4.9 4.9 12.1 0.1 1.1 0.4

Table 3 Comparison of Aura of Amarige on Skin & Fragrance Oil.

Aura of Aromam: A Novel Technology to Study the Emission of Fragrance from the Skin

ComDound Ethyl Linalool Linalyl Acetate Floralozone Cyclogalbaniff Dihydro Myrcenol Linalool Limonene beta Ionone Polysantol is0 E Super Ambrox Hedione Galaxolide Tonalid

Increase or Decrease

yoin Qid

!%dAua

inAm

0.7 10.4

2.8 36.0 0.3 0.7 10.6 11.5 1.4 6.2 0.4 4.8 0.4 5.0 0.6 0.1

4 times 3 times 3 times 3 times 2 times

0.1 0.2 5.8 7.7 4.5 2.5 0.2 4.8

0.7 25.9 5.5 3.3

41

1.5 times 0.3 times 3 times 2 times

same 0.5 times 0.2 times 0.1 times 0.03 times

Table 4 Comparison of Aura of Unisex Fragrance on Skin (Ihr) & Fragrance Oil can be considered to function primarily as a diluent. However, so-called Bottom Note compounds like Ambrox and musklactone, which is the common name for cyclopentadecanolide, both among the highest boiling of fragrance chemicals, showed their presence in the Aura in appreciable amounts indicating that they are very diffusive molecules which play important roles in the first impression of Amarige. Next, we studied a highly successful unisex commercial fragrance (Table 4). as we expected, ethyl linalool and linalyl acetate are enhanced in the Aura due to their high difisivity. Interestingly, middle-boiling range compounds like floralozone, cyclogalbaniff

ComDound Aldehyde AA Methyl Phenyl Acetate Ethyl Linalool Diphenyl Ether Cyclogalbaniff Methyl Ionones Ethyl Acetoacetate is0 E Super Ambrox Hedione Cyclopentadecanolide Galaxolide

w 0.05 0.02 5 .oo 0.01 0.20 2.50 1.80 2.60 0.20 18.00 4.80 14.00

Increase or Decrease 'YOin Aura 0.80

0.20 30.00 0.04 0.90 9.50 3.40 2.10 0.10 2.40 1.20 0.90

ir.uhl3

16 times 10 times 6 times 4 times 4 times 4 times 2 times

same 0.5 times

0.13 times 0.25 times 0.07 times

Table 5 Comparison of Aura of Feminine Fragrance on Skin (1 hr) & Fragrance Oil

42

Flavours and Fragrances

and beta ionone all increased by a factor of three in the Aura. At the same time, highboiling compounds like polysantol, is0 E super, and Ambrox were present in the Aura in appreciable quantities due to their diffusivity. Once again, please note the poor diffusivity of Hedione as well as Galaxolide and Tonalid. The next perfume we studied was a very successful modem feminine fragrance also introduced in the 90’s (Table 5). In addition to two very volatile chemicals like Aldehyde AA and methyl phenyl acetate, highly diffusive middle-boiling range compounds like ethyl linalool, diphenyl ether, cyclogalbaniff, and methyl ionone, together with high-boiling yet very diffusive materials like iso E super, Ambrox, and Cyclopentadecanolide are responsible for this unique feminine note. Please observe that Hedione does not influence the Aura due to its nondiffusivity. Finally, we studied a very sensuous and highly successful woman’s fragrance, recently launched in both Europe and the US.(Table 6). It is very easy to see from the data in Table 6 why this fragrance is so diffusive. Three extremely diffusive chemicals which cannot even be detected in direct analysis of the perfume oil, Lolitol, Passionfruit Compound, and Methyl Octin Carbonate, are readily seen in the Aura. In addition, several other very diffusive molecules have been used including Givaudan’s Givescone, IFF’S Floralozone, akha and beta Damascone, cis Jasmone, and Undecavertol as well as the already discussed Ethyl Linalool, Linalyl Acetate, Methyl Ionones, and Cyclopentadecanolide.

ComDound

Lolitol Passion Fruit Comp’d. Methyl Octin Carbonate Givescone Floralozone alpha Damascone beta Damascone Ethyl Linalool Undecavertol Linalyl Acetate cis Jasmone Methyl Ionone Cyclopentadecanolide Galaxolide

M

0.10 0.01 0.10 0.04 1.40 0.30 2.00 0.10 2.00 0.90 8.00

O ‘ A 0.10 0.0 1 0.10 1.20 0.10 0.50 0.20 6.00 1 .oo 7.80 0.20 4.60 0.20 1 .oo

Increase or Decrease iILAU3

- times - times - times 10 times 10 times 5 times 5 times 4 times 3 times 3 times 2 times 2 times 0.25 times 0.13 times

Table 6 Comparison of Aura of Woman’s Fragrance on Skin (lhr) & Fragrance Oil

Aura of Aroma": A Novel Technology to Study the Emission of Fragrance from the Skin

43

From these studies we have proved beyond a doubt that the first impression of a fragrance is not only due to the highly volatile so-called Top Note chemicals, as once believed. Actually, it consists of a combination of highly volatile, middle-boiling range molecules, as well as high molecular weight high-boiling compounds possessing a quality called high diffusivity. Table 7 summarises our finding with regard to the aroma molecules, from the lowestboiling to the highest boiling, which appear simultaneous1y in the Aura of any fragrance in which they are present. Relatively higher molecular weight sulphur compounds like 8puruMenthanethiol and Passionfruit Compound play key roles in many fragrances at very low levels not only due to their extremely low odour thresholds but also because of their very high diffusivities.

Extremely Diffusive Fragrance Molecules *(Molecular Weight) Passionfruit Compound (160)* Methyl Octin Carbonate (168) 8-puruMenthanethio1 ( I 70)

Diffusive Topnote Molecules Aldehyde AA ( I 38) Lolitol (144) Linalool ( I 54) Dihydro Myrcenol(lS6) Styralyl Acetate (1 64) cis-Jasmone ( 164) Ethyl Linalool(l68) Diphenyl Ether (170) Linalyl Acetate (196)

Diffusive Middle Note Molecules Coumarin (146) Ethyl Vanillin (166) Floralozone ( 190) a & PDamascones ( 192) lonones & Methyl lonones (192/206) Cyclogalbaniff (198) Cashmeran (206) Cedramber (236)

Diffusive Bottom Note Molecules Cedrene.Cedrol (204/222) a & PSantalol(220) Polysantol(222) Patchouli Alcohol (222) is0 E Super (234) Ambrox (236) Cyclopentadecanolide (240)

Highly Used NON-Diffusive Molecules Hedione Benzyl Salycilate Galaxolide Tonalid

Table 7

44

Flavours and Fragrances

At the same time, various so-called Middle Note compounds ionones, Cashmeran, and Cedramber as well as Bottom sesquiterpenic materials such as Cedrol, Santalol, and very high compounds like Ambrox and Cyclopentadecanolide appear in which constitute the first real impression of any fragrance.

like Ethyl Vanillin, the Note compounds like boiling amber and musk the Aura of fragrances

Compounds like Hedione, Benzyl Salicylate, and Galaxolide, which are used throughout the fragrance industry in ton quantities, are relatively non-diffusive. In other words, they play little role in the Aura of a fragrance. Now, one could easily imagine that a creative perfumer, selecting from compounds mentioned in Table 7, could easily create a long lasting and, at the same time, highly diffusive characteristic fragrance as he or she desires. 4. EFFECT OF SKIN ON EMISSION OF FRAGRANCE It is commonly believed that the smell of a fragrance differs greatly from skin to skin. Therefore, we studied whether the fragrance composition changes depending upon the skin to which it is applied. At IFF we have set out to study the emission of fragrance from skin. We have utilised the Aura technology for the study of this interesting phenomenon. Now, the question comes: what kind of skin should we study? A number of investigators have performed these studies using the skin of various Caucasian people, and they stressed the importance of fat and moisture content of the skin. We think that these are important considerations, but, we have selected the skins from a global perspective. In other words, we have selected people from different parts of the world. We selected a professional Indian lady in her mid ~ O ’ Sa, Jamaican lady of the same age who works in a clerical position, a light-skinned Caucasian girl in her mid ~ O ’ S and , a managerial Caucasian lady in her mid 50’s who has never married. Each of these women were first placed on a bland diet and asked to maintain their normal working conditions without too much physical activity. They were also asked to clean their skin by washing with non-fragranced soap. The first perfume that we selected for our skin effect studies was the unisex fragrance the Aura of which we discussed earlier (see Table 4). Two drops of the fragrance were applied to the forearm skin of each lady and the Auras were collected for one hour starting immediately after application and then analysed. These experiments were repeated twice. Table 8 show the composition of fragrance from each skin. One could easily see that among the so-called highly volatile components there is essentially no change from skin to skin. The Limonene composition varies from 9-13%; the second major component, Linalool, varies from 14-15%; the major constituent, Linalyl

45

Aura of Aroma": A Novel Technology ro Study the Emission of Fragrance from the Skin

Woman % SM(45 yr)

ComDonent cis 3-Hexenol

Limonene Dihydro Myrcenol cis 3-Hexenyl Methyl Carbonate Linalool Benzyl Acetate 4-Terpinenol Ethyl Linalool (2-10 Aldehyde Carveol Citronellol Ally1 Amy1 Glycolate Citral Linalyl Acetate alpha Terpinyl Acetate Citronellyl Acetate

w 0.03 10.70 9.50 0.30 14.10 0.40 0.10 1.80 0.05 0.40 0.50 0.20 0.20 30.90 0.30 0.10

Woman % Woman % CP(47 yr) PAM(36 yr)

Woman % CB(53 yr)

Jamaican

Caucasian

Caucasian

0.04 9.30 10.30 0.30 15.40 0.40 0.10 2.00 0.05 0.40 1.oo 0.30 0.20 3 1.OO 0.30 0.10

0.03 13.30 9.30 0.30 13.90 0.40 0.10 1.80 0.05 0.40 0.30 0.20 0.10 32.0 0.30 0.10

0.03 10.10 9.20 0.30 13.80

0.40 0.10 1.80 0.04 0.40 0.50 0.20 0.20 3 1.30 0.30 0.10

Table 8 Emission of Unisex Fragrance from Different Skins (Applied two drops to skin and immediately collected for 1 hour.) Acetate, basically did not change at all. Similarly, none of the minor components changed appreciably. Table 9 shows the composition of the Auras for the Middle and Bottom Notes. Once again, there is basically no change from skin to skin. The slight changes observed in the case of beta Ionone, Coumarin, and Tonalid for one of the Caucasian ladies is considered within the range of experimental error. Since this fragrance #1 was introduced as a unisex fragrance we also studied the effect of skins of five different male lab workers with ages between 20 and 55. One was Jamaican, one was Indian, and the rest were Caucasian. These results were compared with those obtained for the four women previously described. Within the limits of experimental error, essentially no difference was observed among the composition of the so called lower-boiling components of this fragrance from men to women. This was also found to be the case for the higher-boiling components. It is obvious that different skins do not significantly affect the emission of this particular fragrance. The next perhme we studied for the effect of skin on the emission of fragrance was another women's fragrance which was first introduced into the U.S. in 1994. In this case, together with the four women we also included one young male.

46

Flavours and Fragrances

Woman YO Woman YO Woman YO Woman Yo CP(47 yr) PAM(36 yr) CB(53 yr) SM(45 yr) Component Neryl Acetate Geranyl Acetate cis Jasmone Coumarin Cyclogalbaniff Floralozone Caryophyllene beta Ionone Pol ysantol Lilial Helional Kharismal is0 E Super Ambrox Galaxolide Tonalid

.h&j!j

Jamaican

1 .oo 1.60 0.10 0.10 0.20 0.20 0.60 4.10 0.20 0.60 1.60 1.20 2.00 0.10 0.20 0.10

1

.oo

Caucasian 1.10 1.70 0.10 0.05 0.10 0.10 0.70 1.90 0.10 0.10 0.10 0.50 1.10 0.10

1.60 0.20 0.10 0.30 0.20 0.60 3.60 0.20 0.30 0.20 1.20 1.70 0.10 0.20 0.10

Ca_ucasian 1 .oo

0.10 0.03

1.60 0.20 0.10 0.30 0.20 0.60 4.20 0.20 0.50 0.20 1.40 2.30 0.20 0.20 0.10

Table 9 Emission of Unisex Fragrance from Different Skins (Cont’d) (Applied two drops to skin and immediately collected for 1 hour.) The Auras were collected from the five subjects for one hour immediately after application of the fragrance. As can be seen in Table 10, there is no significant change in the composition of the initial Auras from all of these five skins. Woman YO

Woman YO

Woman YQ

Woman %

(45 Yr)

(47 yr)

(36 yr)

(53 yr)

Jamaican

Caucasian

Caucasian

L2xUskQ

2.80 7.40 20.30 1.80 6.60 ISO 0.80 2.10 I 1.40 13.10 15.00

0.80

3.30 4.20 20.30 I .80 5.20 0.90 1.10 2.50 14.10 15.80 16.50 0.60

2.10 4.60 18.10 I.60 5.70 1.10 0.90 2.30 12.70 14.60 18.10 1.oo

0.20 1 I.20

12.90

2.80 4.80 24.90 I.70 3.40 1.10 1.30 3.10 14.20 14.00 13.40 0.50 0.30 10.70

m

€Qmw.nm Dihydro Myrcenol Phenyl Ethyl Alcohol Benzyl Acetate Ethyl Linalool Citronellol Phenyl Ethyl Acetate Linalyl Acetate Dimethyl Octyl Acetate Citronellyl Acetate Geranyl Acetate gamma Methyl lonone Lilial Bacdanol iso E Super Cyclopentadecanolide Galaxolide

TN TN

TN TN TN TN TN TN TN TN

MN MN BN 6N BN BN

3.40 5.90 20.50 I .90 6.40 I.20 0.90 2.10 11.20 12.70 15.0 0.80 0.20 12.20

TN = Topnote, MN = Middle Note, BN = Bottom Note Table 10 Emission of Women’s Fragrance from Different Skins (Applied two drops to skin and immediately collected for 1 hour.)

Man % (20yr)

0.30 12.50

47

Aura of Aroma@:A Novel Technology to Study the Emission of Fragrancefrom the Skin

(45 Yr)

Woman % (47 yr)

Woman % (36 yr)

Woman % (53 yr)

Ilu!ian

LimaiGm

Caucasian

Caucasian C-aucasian

Woman %

ComDonent

Dihydro Myrcenol Phenyl Ethyl Alcohol Benzyl Acetate Ethyl Linalool Citronellol Phenyl Ethyl Acetate Linalyl Acetate Dimethyl Octyl Acetate Citronellyl Acetate Geranyl Acetate gamma Methyl Ionone Lilial Bacdanol is0 E Super Cyclopentadecanolide Galaxolide

TN TN

TN TN

Man % (20Yr)

0.50

0.60 0.50

1.10

TN TN

TN

TN TN TN

m MN BN BN BN BN

0.30 4.00 9.30 22.70 2.80 1 .so

38.60

5.20 17.60 3.70 53.20 1.70

0.40 5.20 10.80 24.50 2.80 1.50 35.60 0.80 0.80

2.40 I1.50 3.40 2.20 56.20 1.90 2.40

1.60 3.90 15.40 2.90 2.30 48.10 I .so 1.90

TN = Topnote, MN = Middle Note, BN = Bottom Note Table 11 Emission of Women’s Fragrance from Different Skins (Applied two drops to skin, after 45 minutes, collected for 1 hour.) Since this particular fragrance lasts longer on the skin than the unisex fragrance, in another experiment we collected the Auras after waiting for a period of 45 minutes following the application of the fragrance. Table 1 1 shows that irrespective of the skin, after 45 minutes none of the skins retained any of the lower-boiling components, however, for the higherboilers, there are some significant differences. From these results, we can conclude that, for the two fragrances we studied which are of totally different composition, there were basically no changes observed in the initial Auras of the fragrances. However, after waiting approximately one hour, the feminine fragrance shows some differences in the composition of the higher-boiling constituents of the Auras. Therefore, we may assume that skin may have some effect depending on the composition of the fragrance. To really establish the effect of skin, obviously we must do more studies on various fragrances.

5. ACKNOWLEDGEMENT We would like to thank Mr. Gene Grisanti, Chairman and CEO of IFF, and Ms Lisa DelBeccaro, Director of Advertising and Public Relations, IFF Inc., for permitting us to present and publish this paper. References 1. “IFF Announces New Method for Living Flower@Analysis”. Spray Technology & Marketing, p26. Edited by Michael N. SanGiovanni, Published by Industry Publications, Inc., Fairfield, New Jersey, October, 1996.

Dependence of Intensity of Musk Odour on the Energy Gap between Frontier Molecular Orbitals M. Yu. Gorbachov INSTITUTE OF CHEMISTRY OF AS, MD 2028 KISHINEV, REPUBLIC OF MOLDOVA

1. INTRODUCTION

The musk odour is found within four very different chemical families. These are macrocyclic compounds, in particular lactones and ketones, aromatic nitrocompounds, aromatic benzenoids, such as acyl tetralins and indanes, and a few steroids’. The commercial importance of musk odourants, the fact that musk is an odour facet which is well defined, and also that the aromatic musks are fairly ridgid has resulted in a large number of studies into the relationship between structure and musk odour. Some, on the basis of the observation that there are subtle odour differences between the different families, have assumed that there is more than one musk receptor, and searched for correlations within a given group. Others have tried to speculate on molecular parameters common to all groups. One such study was that of Bersuker et al.* They found that two independent molecular fragments with special geometrical and electronic characteristics were required for a compound to smell of musk. The first fragment consists of a polar group (CO, NO, OH) whose electronegative heteroatom is situated symmetrically, and is at a distance of 6.7 k 0.5 A with respect to two methyl (or methylene) groups, the distance between the later being 2.5 f 0.5 A. The second fragment includes two other methyl (or methylene) groups situated at a distance of 5.5 f 0.5 8, from each other. Both of these conditions are true provided that there are no bulky substituents close to the functional group making the later sterically unaccessible. These two structural fragments are called hereafter the active fragments I and I1 respectively. Their presence in the investigated molecular structure results in its musk odour. In the absence of either or both fragments the musk odour disappears. These fragment rules allow a qualitative description of the influence of insignificant changes of molecular stucture on the presence (or absence) of musk odour. However, the problem of the dependence of the musk odour intensity remains unsolved.

2. RESULTS AND DISCUSSION This paper describes the first part of the present work on understanding the relationships between chemical structure and musk intensity. Twenty compounds including representatives from the three main musk families have been investigated (Figure 1). All

Dependence of Intensity of Musk Odour on the Energy Gap between Frontier Molecular Orbitals

0.00

49

* -0.03

0.0 1

o.oo*

0.00

-0.18

-0.08

H

0.02 0.0 1 0.00

(4)

(3)

0.00

*

0.00

0.02*

(6)

C 0.15*

0.00 oo.o*

-0.24

0.02

*o.oo (7) 0.00 oo .w,

H

(9)

b

C

0 -0.22

Figure 1 The charges on the activity fragments' atoms in the compounds under investigation

50

Flavours and Fragrances

*o.o I

*-0.0 1

0.14i "\\

-0'39

1

"O2W 0.00

*o.oo

-0.25

(13) 0.02*

0.00

*-0.01

i

o,C),o \ c=@

P *-0.0 1

-0.23

0.00

0.00

(14)

-0.3 1

/

0.02 0 . 0 (15) *o.oo 2 0 C 0.02

o-02Qc=@

-0.27

0.02

*-0.05

Figure 1 Continued

-0.27

*-0.05

Dependence of Intensity of Musk Odour on the Energy Gap between Frontier Molecular Orbitals

51

of these compounds possess a musk odour of varying intensities. The compounds 1-9 have a weak musk odour, and the compounds 10-20 have a moderate to very strong musk odour (see the work in ref.2). The active fragments I and I1 for each compound are shown in Figure 1 by means of circles and asterisks respectively. Some geometrical and electronic characteristics of the compounds are presented in Table 1. Table 1 Some geometrical and electronic characteristics of the organic molecular systems 1-20.

p-dipole moment, Q-complete charge on all carbon and heteroatoms of fragments I and 11, El-energy of HOMO, EZenergy of LUMO The geometrical parameters and the dipole moments (p) were calculated by means of the MM2 method3. In Table 1, R1 is the distance between the electronegative heteroatoms and the methyl (or methylene) groups of the active fragment I. R2 is the distance between the two methyl (or methylene) groups of the active fragment 11. The distance between the methyl (or rnethylene) groups of the active fragment I changes very little for the twenty compounds under investigation. The electronic parameters, which are also resented in Table 1 and in Figure 1 were calculated on the basis of the CND0/2 method (the atomic charges, the energy levels, and the energy gaps). The relative calculation error of this method is small for different sequences of organic compounds, belonging to different structural classes’.’.

B

52

Flavours and Fragrances

A comparison of the dipole moments and the distances for the active fragments I and I1 (Figure 1 and Table 1) shows that these molecular descriptors alone can not be used to descriminate between the compounds 1-9 with the weak musk odour, and the compounds 10-20 with the moderate or strong musk odour. Both the atomic charges on the individual atoms of the two active fragments and the energy levels for the HOMO (El) and the LUMO (E2) are also incapable of classifying these two classes of musk compounds. However, the value of Q (Table I ) which is the complete charge on all the carbon and heteroatoms of the fragments I and I1 may be a useful discriminating feature. The compounds with a weak musk odour have negative Q values, which are less or equal to -0.21 (all the charges in Table I are given in the electronic charge units). For moderate and strong musks the negative Q values are more negative than -0.21.

Table 2 The dependence of Appel’s Intensity of the musk odour on the values of AE. Compound

Energy (e.V.)

El

Exaltolide Ethylene Brassylate -Ambrettolide Versalide Musk Ambrette Musk Ketone Musk Xylene

- 13.237 -13.071 -1 2.414 -1 1.232 -12.779 - 12.39 1 - 1 2.852

A1

AE

E2 4.763 4.417 4.643 3.205 0.503 0.0 19 -0.460

I

18.000 17.488 17.057 14.437 13.282 12.410 12.312

7

6 6

5 4 3 3

The cycloketones 16 and 17, and the nitro musks 19 and 20 possess the maximum negative Q values, and thus are the strongest musks in this data set. For compounds 1 and 3, their interaction with with an odourant bioreceptor is via their respective bromine and chlorine atoms. These atoms have relatively large van der Waals radii and hence relatively small negative atomic charges. Consequently these compounds have low Q values and thus a weak musk odour (together with the sulphur containing macrocyclic ether 2). However, the value of Q taken alone is not enough for a complete quantitative description of the musk odour intensity. Let us consider the compounds shown in Table 2. These are all strong musk odourants. The quantum chemical calculation shows they have much the same value of Q (Q = -0.31). Nevertheless the Appel’s intensity (AI) of their musk odour is different. It can be seen from Table 2 that for the compounds under consideration there is a clear cut dependence of the values of A1 on the values of the energy gap (AE) between the above mentioned energies E2 and El. An increase in the value of AE by e.V. leads to an increase in the corresponding value of A1 by one unit. Thus a conclusion may be drawn. The musk odour intensity appears to be dependent upon both the complete atomic charge (Q) on the carbon and heteroatoms of the two active structural fragments and the energy gap (AE) between the two frontier molecular orbitals (the HOMO and the LUMO). It seams reasonable to introduce the term 5, to describe the intensity of the musk odour for all compounds under investigation.

C=

IQIAE

Dependence of Intensity of Musk Odour on the Energy Gap between Frontier Molecular Orbitals

I I

53

I

Where Q is the absolute value (module) of Q. 6 is the product of Q I and AE. In the last column of Table 1 are given the values of E, for the compounds 1-20. The moderate to strong musks (10-20) have 5>3.00; for the weak musks (1-9) E,<3.00. For the case when Q is constant, the musk odour intensity will depend only on AE (the compounds in Table 2).

I I

3 CONCLUSION It is clear that the term E, can be used to predict the intensity of the musk odour compounds containing the active molecular fragments I and I1 provided that there are no bulky substituents near the atoms of these fragments. The influence of different bulky groups on the musk odour intensity will be the subject of the next work. References 1. M.G.I. Beets, ‘Molecular Structure and Organoleptic Quality’, Society of Chemical Industry, London, 1957. 2. I.B. Bersuker, A.S. Dirnoglo, M. Yu. Gorbachov, P.F. Wad, M.Pesaro, New. I. Chem., 1991, 15,307. 3. U. Burkert, N.L. Allinger, ‘Molecular Mechanics. ACS Monograph 177’, American Chemical Society, Washington, 1982. 4. ‘Semiempirical Methods of Electronic Structure Calculation, Part A: Techniques’, Ed. by G.A. Segal, Plenum Press, New York, London, 1977. 5 . M. Yu. Gorbachov, PhD Thesis, University of Rostov-on-Don, 1986.

Essential Oils and Analytical

Derivatized Cyclodextrins in Enantiomer GC Separation of Volatiles C. Bicchi, A. D'Amato and V. Manzin DPARTJMENTO DI SCIENZA E TECHNOLOGIA DEL FARMACO, VIA PIETRO GIURIA 9.1-10125 TORINO,ITALY

1 INTRODUCTION Cyclodextrins (CDs), also known as cycloamiloses, cycloglucanes or cyclomaltooligoses, derived from the enzymatic degradation of the starch by cyclodextrin glycosyltransferases from either Klebsiella pneumoniae or Bacillus macerans. CDs are a homologous series of non-reducing cyclic oligosaccharides made up of six or more (a)-D-glucopyranose units linked by a-1,Cglycoside bonds. The number of monosaccharide units varies from six to twelve but only the homologues with six, seven or eight units are commercially available. CDs were first isolated by Villiers in 1891, and characterized as cyclic oligosaccharides in 1904 by Schardinger: their original name was actually Schardinger's dextrins. At the end of the forties, Freudenberg and Cramer discovered the CD property of forming molecular inclusion complexes'.

Figure 1 Structures of a-,p and ycyclodextrins.

58

Flavours and Fragrances

The basic concepts of inclusion can easily be summarized through the two following sentences: - the first one from Cramer’s 1952 article2 “....From this we have discovered a relationship between molecular shape and biochemical activity. Here the CD molecule exerts its efect solely as a result of its geometrical shape, not by means of functional groups. However, only those reagents that can enter the molecular cavity can be considered. Thus the striking parallels with the lock and key relationships of biochemical processes are revealed here ... . ,,, - the second from Saenger‘s 1980 article3 on the thermodynamics of inclusion complexes “..._the inclusion process does not depend primarily on the (chemical) nature o f the guest molecule ....,, The use of CDs in GC is more recent: in the early sixties, attempts were made to exploit their inclusion capability to separate branched from linear isomers or geometric isomers. The first separation of enantiomers was obtained by Sybilska and Koscielski in 1983, who separated a- and 0-pinene, the corresponding pinanes and 6-3-carene, with a column packed with underivatized a-CDs. The first capillary column applications were in 1987, with the almost contemporary work of Juvancz’ and Schurig6*’. Over the following two or three years, several groups began to work in the field, in particular Armstrong, Bicchi, Grob, K8nig, Mosandl, Sandra, Schomburg and Venema‘s groups. At the same time, the first applications in the aroma, perhmes and essential oil field were develo ed in particular by the German groups of Konig, Mosandl, Nitz, Schreier, and Werkhoff 8- 14. The research dealt with and still deals with two main topics: - development of a theoretical model to explain CD separation mechanisms in GC; although several groups have dealt with this topic, an entirely satisfactory model to design a racemate resolution has yet to be found, - development of new CD derivatives with universal enantioselectivity and evaluation of their chromatogaiphic performances. 2 CD DERIVATIVES FOR ENANTIOSELECTIVE GC

The authors’ studies with CDs have mainly dealt with the synthesis of new derivatives, with an enantioselectivity as wide as possible to separate most of the usual volatile racemates. CDs are generally diluted in polysiloxane, as first proposed by Schurig mainly because of a wider range of operative temperatures, the inertness and efficiency of columns prepared by high-temperature silylation, the small CD amounts necessary to prepare columns, shorter analysis times, the possibility of tuning of column polarity by using different diluting phase and of determining thermodynamic parameters involved in enantiomer dis~rimination”~~. Since the authors’ studies mainly compare the GC performance of columns with different characteristics, three types of reference tests are generally used: 1) the Grob test to evaluate the column chromatographic quality, 2) a chiral testi4, that is a mixture of standard racemates with widely different structures to evaluate the column enantioselectivity quickly (Figure 2), and 3) a number of mixtures of standards, again with different structural characteristics, to evaluate column enantioselectivity in depth. Each new phase or column is tested with a total of 200 racemates.

Derivatized Cyclodextrins in Enanriomer GC Separation of Volatiles

1R 70

29 90

30 110

59

4p

min

C

IJO

Figure 2 CGC pattern of the chiral test of a 2,3-dimethyl-6-t.bu~ldimethylsilyl-y -cyclodextridOV-l701 column. Several series of CD derivatives were synthesized in the authors’ laboratory: each series was designed with the aim of increasing enantioselectivity, solubility in polysiloxanes and column stability. Figure 3 reports the chronological sequence of the groups of synthesized CD derivatives. The first series of CD derivatives consisted of 2,3,6-trimethyl- and 2,6-dimethyl-3trifluoroacetyl-a-, -p-, and -y-CDs14. These CD derivatives were chosen to evaluate the influence of the size of the CD mouth and of the polarity of the substituents in position 3 on the GC enantiomer separation and, last but not least, because they were among the most widely used at that time. With these derivatives, all diluted in OV-1701, the influence of several variables on column performance and on enantiomer separation was investigated; in particular CD quantity and percentage and how they influence the column’s minimum operating temperature, stationary phase film thickness, column length and conditionin temperatures and their influence on enantiomer separation and column stability over time’ *I6. One of the most noticeable phenomena was that columns prepared with all the CD derivatives investigated, after conditioning, changed their performances over time, although to different extents. In particular they increased their minimum operative temperatures, losing their ability to separate highly volatile racemates16. This was perhaps because of either a CD recrystallization or a separation of the CD from the diluting phase, where it was dispersed rather than solubilized. New amorphous or low-melting-point CD derivatives, more soluble in polysiloxane, were then synthesized. To replace trimethyl-p- and -y-derivatives, alkylated derivatives, in particular 2,3,6-tripentyl-, 2,6-dipentyl-3-methyl-, and 2,6-dimethyl-3-pentyl-P- and -yCDs were evaluated. Dimethyl-pentyl-P- and -y-CDs offered the hoped-for characteristics, both in terms of column reproducibility and consistency, and of enantioselectivity”.

9

Flavours and Fragrances

60

n = 1,2

RI,

R3 =

CH3

RI. R2 = CH3, CH3CO CH2CH3

/

R3 = t.butylhmethylsily1

\

\

R, = thexyldimethylsilyl

Figure 3 Chronological sequence of the groups of synthesized CD derivatives In the same way, 3-(2-oxopentyl)- and 3-[2-oxopentyl-( 1,1,1 -trifluoro)]-2,6-dimethylor 2,6-dipentyl-P- and -y-CDs were then synthesized to replace dimethyl-trifluoroacetylCDs. but were only partially successful". In 1991, Mosandl first described the use in GC of a series of CDs asymmetrically substituted in position 6 with the t-butyldimethylsilyl (TBS) group''. These derivatives really represented a new generation of CD derivatives specific for GC, since they combine high enantioselectivity with good solubility in polysiloxanes and high column stability. Both the 2,3-dimethyl-6-TBS-P- and -y-derivatives showed a very interesting enantioselectivity, but when diluted in apolar polysiloxanes (SE-30, PS-347.5) the resulting columns had a minimum operative temperature of 90°C, thus preventing their use with highly volatile racemates. A fiuther series of CD derivatives was then developed to widen their range of applications in apolar diluting phases. In order to obtain derivatives with low melting points and, most important, with better solubility in apolar polysiloxanes at low temperatures, longer alkyl chains, namely ethyl and pentyl groups, were introduced in positions 2 and 3. The 2,3-diethyl-6-TBS-P- and -y-derivatives offered the hoped-for characteristics, successfully replacing the 2,3-dimethyl-6-TBS-denvatives in terms of both minimum operative temperature and enantioselectivity 20 . Incidentally, the 2,3-diethyl-6TBS-P-CD is the derivative which showed the widest enantioselectivity of those synthesized in the authors' laboratory to date.

61

Derivatized Cyclodextrins in Enanriomer GC Separation of Volariles

At present, the thexyldimethylsilyl group is under evaluation instead of the tbutyldimethylsilyl group, as a substituent in position six. This group should again increase CD solubility in polysiloxanes and their temperature stability and inertness. The intent is to analyse high boiling and polar racemates*'. Table 1 lists the racemates which were separated with resolutions better than 1.5, with the different series of CD derivatives synthesized in the authors' laboratory.

Table 1 Racemates separated with resolutions better than 1.5 with the different series of CD derivatives synthesized in the authors' laboratory. A: trimethyl-CDs; B: 2,6dimethyl-3-tr$uoroacetyl-CDs;C: 2,6-dimethyl-J-pentyl-CDs; D: 2,3-diace&l-6t-butyldimethylsilyl-CDs; E: 2,3-dimethyl-6-t-butyIdimethylsilyl-CDs; F: 2,3diethyl-6-t-butyldimethylsilyl-CDs. Compounh

A

Limonene 3-0ctan0l cis-Hexenyl-2-methylbutyrate Menthol Isomenthol Neomenthol Neoisomenthol Sesquicineol Hydroxycitronellal Isopinocampheol y-Pentalactone y-Hexalactone y-Heptalactone y-Octalactone y-Nonalactone y-Decalactone y-Undecalactone y-Dodecalactone GHexalactone GHeptalactone 6-Octalactone 6-Nonalactone &Decalactone 6-Undecalactone &Dodecalactone Linalol Linalyl acetate Citronellyl acetate

X X

B

C

D

E

F

x x x x x x x x x x x x x x x x x x x x x x X x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x x X

X

x x x x x x x x x x x x x x x X

X

x x X x x x x X x X X

x x x x x x X

x x

X X

x x X X

Flavours and Fragrances

62

Compounds

A

B

C

D

E

F

1 -Phenylethanol Ethyl 2-phenylbutanoate 4-Isopropyl- 1-methyl-7-oxabicyclo-[2,2, Il-heptan-2-one 3-Oxocineol Menthone Isomenthone 4-Terpineol Carvone 1 -(3-Methylbutenyl)-4-rnethylcyclohexa-3-enyl methyl ketone Camphor Bomyl acetate Borneo1 Isobomeol Isobomyl isobutyrrate Patchouol Massoia decalactone Massoia dodecalactone a-Terpineol 1,4-Dimethyl-7-oxabicyclo[2,2,1 lheptan-2-.ol rruns-2-(2-Butylcyclopropyl)acetic acid Menthyl acetate cis-Rose oxide trans-Roseoxide Citronellol Tetrahydrolinalol 5,9,9-Trimethyl-2-oxa[4,5]spirodeca-3 -ene Methyl cis-2-(2-hexylcyclopropyl)acetate Methyl cis-2-(2-heptylcyclopropyl)acetate Methyl cis-2-(2-octylcyclopropyl)acetate cis-2-(2-Ethylcyclopropyl)ethanol cis-2-(2-Butylcyclopropyl)ethanol cis-2-(2-Pentylcyclopropy1)ethanol cis-2-(2-Hexylcyclopropyl)ethanol cis-2-(2-Heptylcyclopropyl)ethanol cis-2-(2-Octylcyclopropyl)ethanol

x x x

x x x x

x x x x x x x

x x x x x x x x

x x x x x x x

x X x x x x x x

X X

X X X X X X

x X x x x x x

X X

x x X

x x x X x x x x x x x x x x

X X X

x x X

X X

x x x x X

truns-2-(2-Butylcyclopropyl)eth~0l ~runs-2-(2-Pentylcyclopropyl)ethanol truns-2-(2-Hexylcyclopropyl)ethanol rruns-2-(2-Heptylcyclopropyl)ethanol truns-2-(2-Octylcyclopropyl)ethanol Ethyl P-hydroxyhexanoate Linalyl propanoate Styrallyl acetate

X

x X

x x X x x x x X

x x x x x x x x x x x x x x x x x x x x x x x x

x x x x x x x x x x x x x x x x x x x x X x x x X

X

x x x x

63

Derivarized Cyclodexrrins in Enanriomer GC Separation of Volatiles

A

Compounh

~ans-2-(2-Pentylcyclopropyl)aceticacid rrans-2-(2-Hexylcyclopropyl)acetic acid trans-2-(2-Heptylcyclopropyl)acetic acid Methyl fruns-2-(2-butylcyclopropyl)acetate cis-2-(2-Butylcyclopropyl)acetic acid cis-2-(2-Pentylcyclopropyl)acetic acid cis-2-(2-Hexylcyclopropyl)acetic acid 1-0cten-3-01 Ethyl 2-methyl butyrate cis-Nerolidol mans-Nerolidol a-Pinene p-P'inene Pulegone $-hone cis-y-Irone cis-a-lrone rrans-a-hone

B

C

D

E

F

x x x x x X X

x x x x x x x x x X

x x

X

X X X

X

x x X X X X X

X

x x x X

x x

Figure 4 reports the GC patterns of the best separations of linalyl acetate obtained with each of the groups of CD derivatives described above. This racemate was taken as a marker of enantioselectivity in the authors' laboratory, in consideration of the importance of lavender essential oil and of the difficulty of separating it with the first groups of derivatives. a) b)

20 90

30

30

110

80

10 80

20 90

20 90

min 'C

Figure 4 CGC patterns of the best above separations of linalyl acetate obtained with each of the digerent groups of CD derivatives above described. a) trimethyl-& -CD; b) 2,6-dirnethyl-3-pentyl-yCD; c) 2,3-diacetyl-6-TBS-yCD; d) 2,3-diethyl6-TBS-PCD.

Flavours and Fragrances

64

3. CD APPLICATIONS IN THE ESSENTIAL OIL AND AROMA FIELDS Multidimensional GC (MDGC) and GUMS or GC/FT-IR are of course the methods of choice when the enantiomeric excess (ee) of a component must be determined in a complex mixture. MDGC actually makes it possible to isolate (and sometimes also to accumulate) the peak of the racemate under investigation through the first column, and to transfer and re-analyse it automatically in the second chiral column. On the other hand, GUMS or GCRT-IR are complementary or alternative to MDGC, because they can give selective detection through diagnostic ions, or through absorptions specific to the structures under investigation. But the ee of a racemic component in a complex mixture can also be determined through specific methods. One of them is based on varying the diluting phase. The diluting phase plays a fundamental role in enantiomer ~eparation*~+~' because a) the analytelcyctodextrin hodguest interaction is thermodynamically driven: the enantiomer separation can therefore be improved by reducing the analysis temperature with a diluting phase as apolar as possible; b) Venema first showed that a diluting phase/CD interaction competitive with the analyte/CD hostlguest interaction is also e ~ t a b l i s h e d ~ c)~the ; polarity of the resulting mixed stationary phase is conditioned by a different diluting phase. Figure 5 shows that a racemate can be separated depending on the diluting phase: with 2,3-diacetyl-6-TBS-P-CD, linalyl acetate is base-line separated with PS-086 as diluting phase, but not separated at all with OV- 1 70 1 .

OV- 170 1

_20_ 90

PS-086

min 'C

I

10 80

20

mln

90

oc

Figure 5 CGC patterns of linalyl acetate analysed with 2,3-diacetyl-6-TBS-PCD with OV-1701 and PS-086 as dilutingphases.

Derivatized Cyclodextrins in Enantiomer GC Separation of Volatiles

65

Different diluting phases contribute differently to analyte retentions, because of the different analyte/diluting phase chromatographic interaction. The combination of retention data (or better retention indices) from different columns, coated with the same CD derivative but dissolved in different diluting phases, makes it possible to identify and determine the analyte ee in the complex mixture Without time consuming clean-up procedures or expensive MDGC or GC/MS GC/FT-IR techniques2'. This approach is complementary to that proposed by Konig which combines the chromatographic data obtained from more than one column coated with different CD derivatives but with the same diluting This approach was used, for instance, to determine linalol, terpinen-4-01 and a-terpineol ee in a lemon oil. Figure 6 reports the CGC pattern of a lemon oil and the location of the peaks of the three components investigated. 2,6Dimethyl-3-pentyl-P-CD was used as chiral selector, and PS-347.5(polymethylsiloxane), PS-086 (polymethylphenylsiloxane, -1 2% phenyl), and OV-1701 (polymethylphenylcyanopolysiloxane, -7% cyanopropyl, -7% phenyl), as diluting phases. Figure 6,at the bottom, reports those parts of the chromatograms recorded with the three columns, where the three racemates in question elute. With the column coated with PS347.5 as diluting phase, one of the enantiomers of linalol coeluted with another component, while terpinen-4-ol and a-terpineol, which are minor or trace components, eluted very close to major components. With PS-086, linalol enantiomers were fully separated from the oil components, but a major component overlapped one of the 4terpineol enantiomers. With OV-1701,a-terpineol and terpinen-4-01 were well separated, but one of the enantiomers of linalol coeluted with a component of the oil. It thus appears evident that there is a serious risk of errors in ee determination due to peak-overlapping; and that risk can be strongly reduced by using at least two columns. The second method concerns the characterization of a complex mixture through the ee determination of some of its optically active components28.It is quite often reported that an essential oil or, more often, several essential oils from different species of different genera, are characterized by the ee of a chiral component measured with a dedicated CD GC column'*. But in many cases, origin and authenticity of an essential oil is best evaluated through the simultaneous determination of the enantiomer abundances of several of its optically active components, hopefully, in a single GC run. Recent advances wider enantioselectivit of the "last generation" of CDs; availability of data base, on enantiomeric separationJ9; increased knowledge of the role of the diluting phase - have made it possible to progress from the "one column for one racemate" approach to the "one column for one problem" approach. A typical application is with peppermint oil which can be commercially characterized through the ees of some of its more representative optically active components, including a-pinene, P-pinene, limonene, menthone, isomenthone, menthol, isomenthol, pulegone, menthyl acetate. These racemates can be separated successfully and simultaneously with a 30% 2,3-dimethyl-6-f-butyldimethylsilyl-P-CD diluted in PS-086 (Figure 7). With this CD/polysiloxane combination, however, one of menthyl acetate enantiomers coelutes with one isomenthol enantiomer. An unequivocal ee determination of these components in peppermint oil can successfully be achieved with multidimensional GC (MDGC) using a first column to separate these components in their racemic form. The lower part of the figure shows the CGC pattern of a commercial sample of peppermint essential oil.

-

Flavours and Fragrances

66

0

n yl

- r

.'"

0 ' 5 0

Ph

L-L

L

1.0

10

20 $0

3p rio

rn C

Figure 6 At the top: CGCpattern of a lemon oil and the location of the p e a h of linalol, terpinen-4-01 and a-terpineol. Column: 2,6-dimethyl-3-pentyl-pCD/PS-34 7.5. A t the bottom: parts of the chromatograms where linalol, terpinen-4-01 and citerpineol elute. Columns: a) 2,6-dimethyl-3-pentyl-PCD/PS-347.5, b) 2,6dimethyl-3-pentyl-PCD/PS-086, c) 2,6-dimethyl-3-pentyl-PCD/OV-1701 [reprinted from Journal of Chromatography, 666, Bicchi G., 'Cyclodextrin derivatives in the gas chromatographic separation of racemic mixtures of volatile compounds - VII', pgs. 137-146, 1994 with the kind permission of Elsevier Science - NL, Sara Burgerhartstraat 25, 1055 KV Amsterdam, The Netherlands]

67

Derivatized Cyclodextrins in Enantiomer GC Separation of Volatiles

b

C

w

80

20

3.0

mmbr

80

100

110

"c

Figure 7 a) CGC pattern of the separation of a-pinene (I), Ppinene (2), limonene (3), menthone (4), isomenthone (S), menthol (6). isomenthol (7), pulegone (8); b) CGC pattern of the separation of menthyl acetate (9); c) CGC pattern of a commercial sample of peppermint essential oil. Column: 30% 2,3-dimethyI-6-t-butyldimethylsilyl-~Clldiluted in PS-086 @om the reference [28] with the permission of the copyright owner).

4 CONCLUSIONS In conclusion, derivatized cyclodextrins represent a milestone for enantiomer separation through GC, in particular in the flavour and fragrance field. Their success can be summarized in numbers: more than 700 articles published since 1987, over 50% of them dealing with flavours and fragrances, over 500 racemates separated in the field;

Flavours and Fragrances

68

unlike the previous enantioselective stationary phases, with which over 80% of racemates required diastereomerisation, 95% of the racemates separated with CD derivatives were separated without derivatization. References 1 . V. Schurig and H.-P. Nowotny, Angew. Chem. Int. Ed. Engl., 1990, 29, 939 and references cited therein. 2. F. Cramer, Angew. Chem., 1952,64, 136. 3. W. Saenger, Angew. Chem., 1980, 92, 343; Angew. Chem. Int. Ed. Engl., 1980, 19, 344. 4. D. Sybilska and T. Koscielski, J. Chromatogr., 1983,261,357. 5. 2. Juvancz, G. Alexander and J. Szejtli, J. High Res. Chromatogr. Chromatogr. Commun, 1987,10,105. 6. V. Schurig and H.-P. Nowotny, Proceedings of Advanced in Chromatography 1987, A. Zlatkis (Ed.), Berlin, 8-10 Sept. 1987. 7. V. Schurig and H.-P. Nowotny, J. Chromatogr., 1988,441, 155. 8. A. Mosandl, J. Chromatogr., 1992,624, 267 and related references. 9. A. Mosandl, Kontakte (Darmstadt), 1992,3,38 and related references. 10. P. Werkhoff, S. Brennecke, W. Bretschneider, M. Giintert, R. Hopp and H. Surburg, 2. Lebensm. Utters. Forsch., 1993, 196, 307 and related references. 11. 'P.Werkhoff, S . Brennecke, W. Bretschneider, Chem. Mibobiol. Technol. Lebensm., 199 1, 13, I29 and related references. 12. C. Bicchi, V. Manzin, A. DAmato, P. Rubiolo, Flavour Fragr. J., 1995, 10, 127 and related references. 13. H.-P. Nowotny, D. Schmalzing, D. Vistuba, V. Schurig, J. High Res. Chromatogr. Chromatogr. Commun, 1989,12,383. 14. C. Bicchi, G. Artuffo, A. DAmato, A. Galli and M. Galli, J. High Resolut. Chromatogr., 1991, 14, 301. 15. C. Bicchi, G. Artuffo, A. DAmato, A. Galli and M. Galli, Chiralify 1992, 4, 125. 16. C. Bicchi, G. Artuffo, A. D'Amato, A. Galli and M. Galli, J. High Resolut. Chromatogr., 1992,15,655. 17. C. Bicchi, G. Artuffo, A. D'Amato, V. Manzin, A. Galli and M. Galli, J. High Resolut. Chromatogr., 1992, 15, 7 10. 18. C. Bicchi, A. DAmato, V. Manzin, A. Galli and M. Galli, , J. High Resolut. Chromatogr., 1995, 18,295. 19. H.-G. Schmarr, A. Mosandl and A. Kaunzinger, J. Microcol. Sep., 1991,3,395. 20. C. Bicchi, A. D'Amato, V. Manzin, A. Galli and M. Galli, J. Chromatogr. A , 1996, 742, 161. 21. C. Bicchi, A. D'Amato, V. Manzin, A. Galli and M. Galli, inpreparation. 22. C. Bicchi, G. Artuffo, A. DAmato, V. Manzin, A. Galli and M. Galli, J. High Resolut. Chromatogr., 1993,16,209. 23. C. Bicchi, A. DAmato, V. Manzin, A. Galli and M. Galli, J. Microcol. Sep., 1995, 7, 326. 24, W.M. Buda, K. Jacques, A. Venema, P. Sandra, Fresenius J. Anal. Chem. 1995,352, 679.

Derivatized Cyclodextrins in Enantiomer GC Separation of Volariles

69

25. C. Bicchi, A. DAmato, V. Manzin, A. Galli and M. Galli, J. Chromatogr. A, 1994, 666, 137.

26. W.A. Kijnig, B. Gehrcke, D. Icheln, P. Evers, J. Donnecke, and W. Wang, J. High Resolut. Chromatogr., 1992, 15,367. 27. W.A. Kijnig, A. Kriiger, D. Icheln, T.Runge, J. High Resolut. Chromatogr., 1992, 15 184. 28. C. Bicchi, A. DAmato, V. Manzin, P. Rubiolo, Flavour Fragr. J., 1997, 12, 55.

29. B. Koppenhofer, Chirabase GC, Universitiit Tubingen.

Production, Chemistry and Sensory Properties of Natural Isolates Mans H. Boelens BOELENS AROMA CHEMICAL INFORMATION SERVICE (BACIS), GROEN VAN PINSTERERLAAN 21,1272 GB HUIZEN. THE NETHERLANDS

1. INTRODUCTION

Natural isolates encompass distilled essential oils, cold-pressed oils, extracted oleoresins, concretes and absolutes, plant - or animal exudates, and tinctures. The main aspects of the production of these isolates are the preparation of the plant material, the isolation methods, the yields, the economics and the quality control. The preparation of the plant material may involve for example, harvesting, threshing, chipping, drying, grinding, hydrolysis and fermentation. During these preparations various volatile compounds, such as unsaturated aliphatic alcohols and aldehydes, mono- and sesquitetpene oxides, coumarins, polyfunctional benzenoids and irones, can be formed. Citrus oils can be produced by expression, such as the Italian pellatrice and sfumatrice methods, the American Brown oil extractor, and the FMC apparatus. The biogenesis of limonene in citrus oils will be discussed. Leaf, seed and flower oils are manufactured by steam distillation, hydrodistillation and hydrodiffusion. The production of mint oil, bitter orange leaf & flower oil and rose oil will be shown. The chemistry of the olfactively character impact compounds in these oils will be demonstrated. Biogenesis of the following characteristic compounds in Mentha species is shown diagramatically: menthone and menthol in peppermint and commint, carvone in spearmint, menthofuran in watermint, pulegone in penny royal. Other natural isolates, such as oakmoss absolute, treemoss absolute and labdanum gum are manufactured by solvent extraction. Others such as rose and bitter orange flower oils can be produced by supercritical fluid carbon dioxide extraction. The formation of volatile compounds, such as amberoxide etc., by acidic photochemical oxidation on the labdanum plants will be discussed. Modem continuous distillation and extraction processes are practised today. Yields and economics of volatile naturals will be commented on. Ranges and anomalies in the yields are noticed. Scope and limitations of the economics, such as raw material, capital, energy and labour costs will be discussed. The physico-chemical standards for many volatile natural isolates oils have been published. The analyses of rose oils, bitter orange flower oils and nagarmotha oil by modem spectroscopic techniques will be discussed. The odour intensities of the oils were determined.

Production, Chemistry and Sensory Properties of Natural Isolates

71

1.1 Importance of natural isolates

Essential oils are mixtures of volatile compounds isolated from plant and animal materials. It has been published that about 350,000 different plant species exist’, and that from these plant species approximately 60,000 should be medicinal plants, and about 17,500 (5%) should be aromatic plants2. The practical use of medicinal plants is estimated at a number of 10,000. About 300 different plant species are used for the production of essential oils for the food, flavour and fragrance industry. The annual world production of volatile oils is estimated at around 50,000 tons3. This production, however, could be up to 100,000 tons with a value of about 1 billion $, based on the figures of the world cons~mption~’~. More than 50% of the quantity of all the essential oils are citrus and mint oils. Particularly, it should be noted that sweet orange oil is produced in thousands of tons. Apart from the production of volatile natural oils, 250,000-300,000 tons of turpentine is produced, from which about 100,000 tons are used for the production of terpenoids for the flavour and fragrance industry. Excellent reviews about the production of essential oils have been written by Meyer-Wamod4, Amaudo5 and Lawrence6. Flores and Segredo’ published on the citrus oil recovery during juice extraction. Boucard and Serth” wrote about a continuous steam stripping process for the distillation of essential oils. Boelens et a/.” published about ten years work on the hydrodiffusion of oils. MoylerI4 reviewed thoroughly ten years work of carbon dioxide extracted oils. With respect to the production of essential oils the following subjects will be discussed: the preparation of the plant material, the isolation methods, yield and economics, and quality control. 2. PREPARATION OF THE PLANT MATERIAL 2.1 General

Various pre-preparations of the plant materials are often necessary before the essential oils can be isolated. Sometimes the desired volatile products are even not present as such in the fresh material, therefore hydrolysis (mosses) or fermentation t’orris root) is necessary. 2.2 Harvesting

All plant material must be harvested before the volatiles can be isolated. The harvesting may simply involve picking of fruits or flowers, or cutting branches of the trees. In practice, however, modem apparatus have been developed, as for example for the cutting of the flowertops of lavandula species. Leaves of plants can, after damaging during harvesting, produce a series of volatile compounds, as for instance (Z)-3-hexenol (leaf alcohol) and (E)-2-hexenal (leaf aldehyde) in freshly mown grass by enzymatic lipoxidation of linolenic acid16. 2.3 Threshing

It will be clear that seeds and fruits must be threshed before the oil can be isolated. Special apparatus have been developed for the threshing of, for instance, umbelliferous fruits, such as aniseed (see grinding).

12

Flavours and Fragrances

FORMATION OF GREEN ODOURANTS

C2H5(CH=CHCH2)3(CH2)6COOH

Linolenic acid

v

A-Jy (Z)-3-hexenal H

A

d

(Z)-bhexenol

(E)-a-hexenal

FORMATION OF HAY-LIKE ODOURANTS DURING DRYING Hydrolysis of ortho-Coumaric glucoside

o-Coumaric-glucoside

o-Coumaric acid

Coumarin

Production, Chemistry and Sensory Properties of Natural Isolates

13

2.4 Drying

Often the plant material is dried before the volatiles are produced. Some herbs, as for example mint plants, are only partly dried before production of the oil. Others, for example spices, are more thoroughly dried as a type of conservation before the oils are isolated. During damaging and drying of the plant material, again chemical reactions may occur, as for instance with the enzymatic conversion of the glycoside melitoside into glucose and (Z)-coumaric acid, which easily cyclizes to coumarin with a characteristic hay-like odour. This reaction occurs during the drying of grass and in some labiate species, like spike lavender and rosemary". 2.5 Grinding

Some fruits and seeds are sometimes ground before isolation of the volatiles as for instance with some umbelliferous fruits. Often the yields of the oil increase after grinding of the h i t s , especially with wet grinding directly into the apparatus.

2.6 Chipping Woody plant material, such as cedarwood, sandalwood etc., are chipped before steam distillation. The chipping of cedarwood before steam distillation has been described by Boucard and Serth". 2.7 Hydrolysis

In some plant material the volatiles are not present as such but are formed after hydrolysis of the less or non-volatile compounds. An example of these materials are oak and treemoss, which contain non-volatile depsides. These depsides are polyfunctional dimeric benzene derivatives, which hydrolyse into monomers after treatment with hot water or steam, or in an enzymatic process. Atranorin is an odourless depside. which occurs in oakmoss. This depside is converted by hydrolysis into methyl beta-orcinyl carboxylate, the olfactively character impact compound of oakmoss extract, and odourless atranylic acid, which by decarboxylation affords the volatile compound atranol. 2.8 Fermentation Some volatiles, as for instance the irones from orris root, are only obtained after fermentation of the dried roots. During this fermentation process, hydrolysis and oxidation may occur.

3. ISOLATION METHODS 3.1 General

The methods for the production of essential oils have been modernised during the last 15 years. Modem continuous distillations and extractions have been introduced during the last decade. The following isolation methods will be treated in more detail: expression of citrus fruits, steam distillation of labiate oils, hydrodistillation of flower oils, hydrodiffision of leaf oils, solvent extraction of mosses, and supercritical carbon dioxide extraction of flower concretes.

Flavours and Fragrances

74

OH

OH

H20

+

* CH,O@COOH

HO-@COOH

H+ BARBATIC ACIO

-cop t

m

L 8-ORCINYLCARBOXYUTE

8-PRCINYL METHYL ETHER

8-PRCINOT,

Formation of volatiles from an oakmoss depside

OH

")

(chloro)-Al"ORIN

R

R

(chloro)-ATAMTYL ACIO

( c o r o )-Al"OL

Formation of volatiles from an oakmoss depside

Oicmol Monomethylelner Us H

MOnOChlOrWfClnOl monomelPvl

einer

Everninic acid melhyl esler R= COOCH,

R,= H . R,= CI. R,= H MOnOChlorWrCinOldmelhyl elherR,s CH,. R,= H . R,= CI

Evarninic acd elhyl esler R= COOC,H, p-Orcmol mOnOmelnyl elher R,= CH,, R,= H p-orcmoearbory~eaod meinyl

OU

Ro*cooc*n,

esw R,= H. R,= COOCH,

CI

CH,

MonOChlorWI?IelliniCacid elhyl ester

R=H Moncchloroevetnlnicacid elnyl

ester R- CH,

Hemalommc and melnyl esler: R,o H. R,= COOCH, Hematommr m a ethyl es~er. R,I H. R,= COOC,H,

ochiomrclnol monomelhyl einei

Chlorrwlranol. R,= CI. R,= H

Degradallon products of lhe deprider in 0akmol)s

Production, Chemistry and Sensory Properties of Natural Isolates

75

3.2 Expression For the expression of citrus fruits there are mainly four methods in use, the Italian pellatrice method, the sfumatrice method, the American Brown oil extraction and the FMC Corporation process. The methods will be discussed more in detail. One should notice that expressed oils always contain a non-volatile residue, which can vary in concentration from 2 to 7%. The dominant monoterpene in all citrus oils is (+)-(6R)-limonene, which accounts for 95% of cold pressed sweet orange oil. Limonene is most probably formed durin growing and ripening of the fruit via mevalonate, geranyl- and (-)-(3R)-linalyl pyrophosphate18.18. 3.2.1 Pellatrice method. The pellatrice expression of citrus fruits concerns the abrasion of the surface of entire fruits. The fruits are rotated against an abrasive surface of a moving Archimedes' screw. During this movement the oil cells burst and the oil is released with and washed away with a water spray. An oil-water emulsion is obtained and the oil is isolated by centrifugal separators. The advantages of the method are: good yields and a quality oil that contains more oxygenated compounds. Some drawbacks may be a darker oil and slightly more residue. 3.2.2 Sfurnatrice method. Before treatment the peel and pulp are separated from the fruits, and the peel hardened in a lime bath. With the sfumatrice method the oil is isolated from the peel by a ribbed roller pressing and a water spray*. The oil water emulsion is centrifuged. Advantages of the method are: the pulp is separated, a lighter product is obtained, and less residue is produced. Drawbacks are the peeling and lime treatment, and no optimal yield. 3.2.3 Brown oil extractor. In the Brown method the whole fruit is used. The fruits move on a bed of rollers covered with needles, and a water spray removes the oil-water emulsion. There are drying rollers and a solid eliminator for the solid materials. The oil is centrifuged. Advantages are the low solid content and a water recycle, a drawback is the relatively high capital costs'. 3.2.4 FMC Corporation method. The most ingenious apparatus for the expression of citrus fruits is the FMC apparatus. More than 50% of all citrus oils are isolated by this method. The method is based on the whole fruit extraction principle. The recovery of the oil occurs during juice extraction'. During the extraction cycle the components of the apparatus interact to separate the various parts of the fruit instantaneously. The citrus oil glands burst and release their oil when the peel is deflected by the pressure created between the cup fingers during the extraction cycle. Recycle water is introduced during extraction, through a special ring located at the upper cup, to capture the oil. The oil is finally isolated by centrifuging. Advantages of the method are: fully automatic, minimum labour costs, juice and oil production. Minor drawbacks are: grading of the fruits, high capital costs (leasing is possible), yields up to 85% of the oil content. 3.3 Distillation Another important method of isolation is the distillation of essential oils. One can distinguish batch or continuous steam distillation, hydrodistillation and hydrodiffusion. A modem approach to essential oils distillation from the herb has been described by Denny9. The different distillation methods will be discussed below. 3.3.1 Steam distillation. Steam distillation is featured by the fact that the plant material is extracted by direct steam (produced in the still) or by indirect steam. The still often has a grill at the bottom with the plant material sometimes is in a perforated basket. Steam distillation is used for the production of labiate leaf and flower oils, laurel leaf oil, eucalyptus leaf, bitter orange leaf oil and umbelliferous h i t oils etc. For yields of the steam distillation of umbelliferous fruit oils see Table 5 . The bulk of essential oils are, apart from the expression of citrus oils, still manufactured by steam distillation. 3.3.2 Hydrodistillution.Hydrodistillation is mostly carried out with flowers, e.g. bitter orange flower, rose or jasmine. The flowers are in a perforated basket and are heated in 2 - 3 times their

76

Flavours and Fragrances

BIOCHEMICAL FORMATION OF LIMONENE IN CITRUS FRUITS (CH3)2C=CH(CH2)2C(CH3)=CHCH,OPP Geranyl pyrophosphate

v (CH,),C=CH(CH2),C(OPP)(CH3)CH=CH2 Linalyl pyrophosphate

A

v A Limonene

FORMATION OF CHARACTERISTIC ODOURANTS IN MENTHA SPECIES

6

6 v

'

3

A \ 4S-(-)-LIMONENE C6 oxidation

C3 oxidation

A

A

A I

O ' A

D-Pulegone

-OH

A

L-Menthol

-0

n L-Menthone

(Mentha pulegium) (Mentha awensis 8 Mentha piperita)

A,

O \ d

wo v /-=

A \

Menthofuran

L-Carvone

(Mentha aquatica)

(Mentha spicata)

71

Production, Chemistry and Sensory Properties of Natural Isolates

weight of water with indirect steam (from outside the still). A volume of water equal to the weight of the flowers is distilled. Yield of the separated oils is in general below 0.1% and the distillate water is saturated with the more soluble oxygenated derivatives (see Analysis). 3.3.3 HydrodiJkion. Hydrodiffision is carried out with low pressure steam (< 0.1 bar) replacing the volatiles from the intact (uncomminuted) plant material by osmotic action. In the hydrodiffusor the low pressure steam flow goes, according to the law of gravity, from the top through the vegetable load down to the condenser at the bottom. The isolation sequence of the volatile components is determined to a great extent by their water solubilities. As a consequence, the condensate water is more or less saturated with the polar constituents of the oil". Test results of a hydrodiffusor are shown in Table 1.

Table 1 Test Results of Schmid Hydrodi@sor LS 500 Compared with Hydrodistillation (according to Schmid Hydrodiffusion SA, Switzerland, May 198 1) Product (Origin)

Hydrodifision Time (hr) Yield (%)

Cistus leaves (France) Cistus leaves (Spain) Lavender (France) Lavandin (France) Cumin fruits (Poland) Caraway fruits (Poland)

8

8 0.5 0.5 4

4

Hydrodistillation Time (hr) Yield (%)

16

0.13 0.15 0.73 1.7 5.0 3.6

16 1

0.04 0.05 0.75

1

1.4

12 10

3.7 4.5

3.4 Extraction

A third method for the isolation of essential oils is the extraction of plant material, which for instance can be solvent extraction, subcritical liquid carbon dioxide and supercritical fluid carbon dioxide extraction. One has to keep in mind that with every type of extraction a certain amount of non-volatile compounds will be extracted. 3.4.1 Solvent expaction. Solvent extraction can be carried out in two types, namely by percolation and by immersion. In percolation the solvent runs through the raw material. In immersion the solvent covers the plant material completely, the solvent moving from the bottom to the top. A wide range of solvents are in use, such as alkanes, haloalkanes, benzenoids, ethers, ketones etc. The most usual solvent is hexane. 3.4.1.1. Oakmoss and Treemoss extracts. Yields of the extraction of oakmoss lichen with various solvents and transesterification of the depsides is shown in Table 2. Oakmoss and treemoss (Pine tree) absolutes were produced and the odour intensities determined by a group of observers. Also, the absolutes were analysed by gas chromatographic mass spectrometric techniques. The concentrations of the main constituents are shown in Table 3.

Table 2 Yields of Extraction and Transesterijication of Oakmoss Lichen Solvent

Method

hexane dichloromethane benzene acetone methanol benzenehethanol

extraction extraction hydrolysidextraction hydrolysidextraction alcoholysis/extraction transesterification

Yield (%)

2 5

7 10 15 10

Flavours and Fragrances

78

Table 3 Main Constituents in Oakmoss Absolute and Treemoss Absolute

Oakmoss absolute

Compounds

Treemoss absolute

(%I Methyl 2,4-dihydroxy-3,6-dimethylbenzoate

3-Chloro-2,6-dihydroxy-4-methylbenzoate 2,6-Dihydroxy-4-methyIbenzaldehyde Cembrene Methyl 2,4-dihydroxybenzoate

57

47 10 5 2 0.5

10 6

I 0.7

Oakmoss absolute and treemoss absolute were olfactively evaluated by a group of perfumers. The odour of oakmoss absolute was qualitatively preferred over the odour of treemoss. The treemoss absolute has a burnt, pyrogenic odour. A group of 35 observers made a paired comparison intensity test between oakmoss absolute and treemoss absolute in O.Ol%, 0.1% and 1% solution. The results are shown in the graph. 3.4.1.2. Labdanum extracts. A practical combination of extraction and distillation is used for the production of labdanum gum, labdanum oil and labdacist oil. Labdanum gum is produced by extraction of the plant exudate from the twigs and leaves of Cisrus 1udanifr.Us (Cistrose) with suitable solvents in a yield of 4 - 6 % on fresh plant material. Labdanum oil is processed from labdanum gum by hydrodistillation in a yield of 0.05 - 0.15% on fresh plant material. Labdacist oil is manufactured by esterification of labdanum gum with ethanol, followed by high vacuum distillation in a yield of I - 2% on fresh plant material. The main constituents of labdanum oil and of labdacist oil are shown in table 4.

Table 4. Main Constituents of Labdanum oil and Labdacist oil Compounds

Labdonum oil

Lobdacist oil

(%I Monoterpenoids (20) Sesquiterpenoids (1 0) 3-Phenylpropanoic acid Ethyl 3-phenylpropanoate Ethyl labda(e)noates Am beroxide

12

1

20 22 4

3 < 0.5 6

< 0.5

0.4

30

< 0.1

The market price of labdanum oil is about $ 71 O/kg and of labdacist oil about $ 18Okg. Labdanum oil and labdacist oil were olfactively evaluated by a group of perfumers. The odour of labdanum oil was qualitatively preferred over the odour of labdacist oil. The odour character of labdanum oil was warmer and more amber-like. A group of 42 observers made a paired comparison intensity test between labdanum oil and labdacist oil in 0.01, 0.1 and I % solution. As can be seen from the graph the odour of labdanum oil was significantly stronger than that of labdacist oil. 3.4.2 Subcritical liquid carbon dioxide extraction. The subcfitical liquid carbon dioxide extraction is carried out at 50 to 80 bar and a temperature of 0 to 10 C. Moyler14 has published in detail about this extraction. The most practical extraction of this type is with hop cones (fruit cones of Hurnulus lupulus L.). The yield of a steam distilled oil of hop cones is about 0.5%, whereas the yield of the carbon dioxide extraction is ca 12%, due to the fact that non-volatile polyfunctional diterpenes (humulones) are soluble in carbon dioxide. About fifty essential oils obtained by liquid carbon dioxide extraction are commercially available.

79

Production, Chemistry and Sensory Properties of Natural Isolates

MAIN ACIDS M LABDANUM GUM

CINNACIIC A C I U

PIIENYLUIOPANOIC A C I I )

+

cool1

COOll

(JJy I

LAIIDANOLIC

ncrn

LAIIIXNIC ACIIIS

LABDANUM RAW MATERIALS FOR AMBER SCENTS

&cooll /

I

LADDANOLIC A C I l l

LAIIIJENIC A C I I ) S

($ /

,

AMOBI(

OXIDE

AMIIIIA

OXIDE

AMllEll

KETAL

80

Flavours and Fragrances

PAIRED COMPARISON INTENSITY TEST BETWEE

160 -

7 1

,40-

-0-

TREEMOSS AES

120-

0

:>-:

100-

I-

v, z

80-

W

40

-

2o

-1

0'

SIGNIFICANCY LEVEL 14.1 I

I

I

I

I

-2.0

-1.5

-1.o

-0.5

0.0

LOG (CONCENTRATION)

PARED COMPARISON INTENSITY TEST BETWEEN LABDANUMAND LABDACIST OIL

/

c

/'

z 100 i 5t z

,

50

/

SIGNIFICANCY LEVEL 16 4

0

I

I

I

I

-2.0

-1.5

-1.0

0.5

LOG (CONCENTRATION)

I

0.0

81

Production, Chemistry and Sensory Properties of Natural Isolates

3.4.3 Supercritical fluid carbon dioxide extraction. The supercritical fluid carbon dioxide extraction is carried out at pressures over 80 bar and in general with temperatures above room tem~erature'~. The supercritical fluid extraction (SFE) has come more into practice during the last five years. One can extract rather fresh plant material (labiates) with this method. One may also make first a solvent extraction (e.g. hexane) to prepare a concrete, and subsequently carry out a SFE-extraction, as for instance with flower concretes. Some analyses of examples of bitter orange flower concrete and rose concrete are shown in Table 6 and 7. 4 YIELDS AND ECONOMICS 4.1 General

Yields and economics are imported for successful production of an essential oil. 4.2 Yields

One can sometimes find quite a range of yields for one and the same essential oil. This range may be due to several reasons, for example climate, soil, isolation method etc. More often, however, published yields are too optimistic and not reproducible. Table 5 demonstrates the variation in the yields of umbelliferous fruit oils as collected by MoylerI4. In general the yield of carbon dioxide and of ethanol extractions are higher than those of steam distillation, these differences are mainly due to the fact that the extracts contain nonvolatile residue.

Table 5 Yielh ("h)of UmbeIIijerousfruitoils Steam distillation

Angelica Anis Caraway Carrot Celery Coriander Cumin Fennel Parsley

0.3-0.8 2.1-2.8 3 -6 0.2-0.5 2.5-3.0 0.5-1 .O

2.3-3.6 2.5-3.5 2.0-3.5

Liquid Fluid Ethanol C02-extract. CO2-extract. extract.

3 7 3.7 1.8 3 1.5 4.5 5.8 3.6

3.3

15 20 3.3 13

12 15

20

4.3 Economics

The production costs of an essential oil involve the raw material costs, capital costs, labour costs, and energy costs. The raw material costs comprise the plant material, solvents etc. The capital costs are the investments (leasing), depreciation and interest. Labour costs concern working, maintenance and quality control costs. Because many aromatic plants grow in developing countries and they are used as raw materials in industrialised countries there is often a controversy between financial (capital/profit) economy and social (work) economy'. Raw materials costs may be relatively high with flower oils (also working hours). Certain apparatus (CO2 extraction, FMC apparatus) have high capital costs. Steam and hydrodistillation have high energy costs. CO2 extraction and hydrodiffusion need more labour costs.

82

Flavours and Fragrances 5 QUALITY CONTROL

5.1 General

The quality control of essential oils may concern the physicochemical standards, the chromatographic and spectroscopic analysis, and sensory analysis. 5.2 Physicochemical Standards

The physicochemical standards can comprise acid, alcohol, carbonyl, and ester number; the solubility in ethanol/water of various concentrations, specific gravity, optical rotation, refractive index, freezinglcongealing and flash point, moisture content and the evaporation residue. A lot is known about the physicochemical properties of essential oils. One can find physicochemical standards in publications of the International Organization for Standardization (ISO), Essential Oil Association of the United States (EOA), in the Food Chemicals Codex (1996-IV), Monographs of the Research Institute Fragrance Materials (NFM), the Pharmacopoeias (EP, BP, USP, DAB, etc.) and in the published country standards (AFNOR, DIN etc.).

5.3 Analysis

5.3.1 Rose and Bitter Orange Flower Oil~'~.Analyses of essential oils are carried out using the most modem gas chromatography and up to date spectroscopic techniques. For gas chromatographic analyses high resolution, high precision fused silica capillary columns are in use. Some examples of these analyses are given in Table 6,7 and 8 showing the headspace and oil analyses of rose flowers and of bitter orange flowers. Table 6 Headspace Analysis of Rosa Damascena: Variation in Composition after Picking Compounds

Monoterpene hydrocarbons 2-Phenylethanol Citronellol Geraniol Nerol Phenylethyl acetate Eugenol Methyleugenol cis-Rose oxide trans-Rose oxide

Living

Picked

("/.I

("/.I

28

57 2 2

40 8

2.5 2.5 3

+

0.5 i

+

+ + 1 2 2 1

0.5

These analyses were carried out in the rose fields near lsparta in central Southern Turkey by Dr. Robin Clery, Quest International, Ashford, Kent, UK.

Production, Chemistry and Sensory Properties of Natural Isolates

83

Table 7 Chemical Composition ofRose Oils Compounds

Hydrodistilled

Carbon dioxide extracted

(%I Citronellol Geraniol Nerol 2-Pheny lethanol Rose oxides Methyl eugenol beta-Damascenone

30 18 9 2

8 4

0.5 2 0.015

2 67 0.15 0.7 <0.005

From these analyses it is clear that a significant variations exist in the concentrations of the main constituents for rose headspaces and oils. Notably the concentration of 2-phenylethanol can vary from 2 - 67 %. An odour intensity comparison test between hydrodistilled rose oil and fluid carbon dioxide extracted rose oil was carried out. The oils were presented in 0.5,0.75 and 1% solution to 36 observers in all possible paired combinations. The intensity score ploned against the logarithm of the concentration is shown in Figure 1. The hydrodistilled rose oil is about 2 to 3 times more intense than the carbon dioxide extract. Table 8 Chemical Composition ojBitter Orange Flower Oils Compounds

Hydrodistilled

Monoterpene hydrocarbons Linalyl acetate Linalool Nitrogen compounds Sesquiterpene alcohols

Carbon dioxide extracted

fi)

&)

38

28 24 35 2

4

38

0.5 4

2

The odour value, quality, and intensity of hydrodistilled and fluid carbon dioxide extracted orange flower oils were determined by a group of five perfumers. The oils were evaluated neat and in I% and in 0.1% solution. The carbon dioxide extracted oil was preferred over the hydrodistilled oil. The odour intensity of the carbon dioxide extract of the orange blossom concrete was about twice of that of the hydrodistilled oil. Another nice example of a modem analysis of an essential oils is 5.3.2Nugarmotha oil. that of Nagarmotha oil. Nagarmotha oil is a rather unknown oil, which is isolated by steam distillation from the roots of an Indian plant, named Cyperus scariosus. The oil has a warm woody odour character, resembling that of vetiver oil. The oil consists mainly of sesquiterpenoid hydrocarbons, epoxides, alcohols and ketones. The oil was analysed on a 50 M capillary fused silica column with DB-5(HP-5, SE-54)as stationary phase and on an identical column connected with a 2 M capillary fused silica column with carbowax 20M as stationary phase. The oil contained about 50 sesquiterpenoids, which were identified with mass spectrometry. The sesquiterpene hydrocarbons, for instance alpha-gurjunene (24%) and rotundene (6%), have the same retention times on both columns; the sesquiterpene epoxides, such as caryophyllene oxide (4%), show a small shift in their retention times, where as the sesquiterpene alcohols and ketones, like aristolone (3%), show a distinct shift in their retention times. Thus a combination of capillary

84

Flavours and Fragrances

1-

INTENSITY COMPARISON TEST BETWEEN ROSE OILS

140 160

k---

-.-

HYDROMSTILLEO ROSE OIL *-CCARBONMOXlDe ROSE EXTRACT

120 -

80 60 -

100

40 20

o

~

-0.35

. -0,30

I

,

,

,

l

-0,25

,

l

,

-0.20 -0.15 -0.10 LOG (CONCENTRATION)

Figure I .

l

,

-0,05

l 0.00

,

l

,

Production, Chemistry and Sensory Properties of Naiural Isolates

85

gas chromatographic columns can be used to identify and quantify the different functional groups in sesquiterpenoids.

6. CONCLUDING REMARKS The preparation of plant material for the production of volatile natural isolates has been discussed. Because various chemical reactions may occur during these preparations, the natural isolate does not completely represent the volatiles which are present in the living plant. The production of volatile natural isolates by expression, distillation and extraction was discussed. These production methods have improved during the last decade, mainly due the introduction of continuous processes. The biochemical formation of limonene in citrus oils was discussed. It seems probable that there exists specific enzymatic genes for the biochemical formation for one chemical compound in different plant species, as for instance for limonene. Yields and economics of the isolates with different technical methods were shown. The yield, and as a consequence the economics, can vary significantly depending on the isolation method used. The yields of expressed and extracted isolates are higher than the yields of distilled isolates due to the fact that the former contain non-volatile compounds. The production, the chemical analysis and the sensory properties of rose oil, bitter orange flower oil, treemoss absolute, oakmoss absolute, labdanum oil, labdacist oil and nagarmotha oil were discussed. The chemical composition and the sensory properties of a natural isolate are strongly dependent on the production method used. Essential oils often do not exactly represent the composition of the volatiles, which are present in the living plant material. Particularly the chemical composition of steam distilled oils differ significantly in terms of the composition of the volatiles in the plant material, this again is mainly due to the production method. References 1.

H.K. Airy Shaw, A Dictionary of Flowering Plants and Ferns, by J.C. Willis, 7th ed., The University Press, Cambridge, 1966.

2. B.M. Lawrence, Is the Development of an Essential Oil Industry, in Malaysia a Viable Commercial Opportunity ?, pp. 187-204. Essential Oils 1992-1994, Allured Publ. Corp., Carol Stream, IL USA, 1995. 3. N. Verlet, Commercialization of Essential Oils and Aroma Chemicals. A Manual on the Essential Oil Industry, Editor K. Tuley De Silva, UNIDO, Vienna, Austria, 1995.

4. B. Meyer-Warnod, Natural Essential Oils. Extraction Processes and Application to Some Major Oils. Pert Flm., 9, 1984(April/May), 93-103. 5. J.-F. Arnaudo, Le Gout du Nature1 - The Taste of Nature. Private Publication BIOLANDES AROMES, 23, Villa Marie-Justine, 92 100 Boulogne (France), 1992- 1996. 6. B.M. Lawrence, The Isolation of Aromatic Materials from Natural Plant Products 57-1 54.

in: A Manual on the Essential Oil Industry, Editor K. Tuley De Silva, UNIDO, Vienna, Austria, 1995. I. J.H. Flores and Guillermo T. Segredo, Citrus Oil Recovery During Juice Extraction.

86

Flavours and Fragrances

Per- Flav., 1996(May/June), 21, 13-15. 8.

Philip E. Shaw, Citrus Essential oils. Pet$ Flm., 1979 (DecIJan), 3,35-40.

9. E.F.K. Denny, The modem approach to essential oil distillation from the herb. Proceedings of the International Conference on Essential Oils, Flavours, Fragrances and Cosmetics, 161- 165. Beijing, China 9-13 October 1988. Steam Distillation of the Subcutaneous Essential Oils, 1988. 10. G.R. Boucard and R.W. Serth, A Continuous Steam Stripping Process for the Distillation of Essential Oils. Pet$ Flav., 1991 (MarcWApril), 16, 1-8. I I . Mans H. Boelens er al., Ten Years of Hydrodiffusion of Oils. P e r - Flav., I990 (Sepu'Oct), 15, 11-14.

12. J.-P. Bats er al., Continuous Process for Oakmoss Extraction. Perf Flav., 1990 (Nov/Dec), 15, 15-16. 13. Mans H. Boelens, Formation of Volatile Compounds from Oakmoss. Per- Flav., 1993 (JadFeb), 18,27-30. 14. D.A. Moyler er al., Ten Years of Carbondioxide Extracted Oils. Proceedings 12th ICEOFF

Vienna, Austria, October 1992. 15. Mans H. Boelens and Harrie Boelens, Differences in Chemical and Sensory Properties of

Orange Flower and Rose Oils Obtained from Hydrodistillation and from Supercritical C 0 2 extraction, Per$ Flav., 1997 (MaylJune), 22,3 1-35. 16. T. Galliard. Biochemistry in Wounded Plant Tissues, Ed. G. Kahl, de Gruyter, Berlin, New York, 1978, 155-201. 17. H. van Genderen, L.M. Schoonhoven, and A. Fuchs, Chemisch-Ecologische Flora van Nederland en Belgie, KNNV Uitgeverij, Utrecht, 1996,22. 18. R. Croteau, Biochemical and Molecular Genetic Aspects of Monoterpene Formation,

Proceedings of 27th International Symposium on Essential Oils, Sept. 8- I I, 1996 WienNienna, Austria (to be published 1997). 19. A . Yuba er al., Limonene synthase from Perillafru~escens,Proceedings of the 27th International Symposium in Essential Oils, Sept. 8-1 I , 1996 - WienNienna, Austria (to be published 1997).

Acknowledgement The author is greatly indebted to Dr. Arantxa Bordas of Destilaciones Bordas Chinchumeta, Dr. Ir. Jean-Francois Arnaudo of Biolandes Technologies, Dr. Robin Clery of Quest International, Dr. Jose Flores of FMC Corporation, Dr. Ir. Piet Traas and Dr. Pieter de Valois of Quest International for their valuable information, photographs, slides, sensory and GC-MS analyses.

An Odour-Sensing System for Use in Measuring Volatiles in Flavours and Fragrances Using OCM Junichi Ide', Takamichi Nakamoto2 and Toyosaka Moriizurni*

' KAWASAKI RESEARCH CENTRE, T. HASEGAWA CO.. LTD., JAPAN FACULTY OF ENGINEERING, TOKYO INSTITUTE OF TECHNOLOGY, TOKYO, JAPAN

1. ODOUR-SENSING OVERVIEW

Odour-sensing systems are required in many fields ranging from the food, drink, and cosmetic industries to the environment and others. Human sensory tests, now utilised to discriminate odours in such fields are inevitably affected by variations in the inspector's state of health and mood at the time of the assessment. Therefore, objective evaluation methods are desired for the fields mentioned above.

Gas chromatography (GC), a method for quantitatively and qualitatively analysing gas components, has been used to study odours. Whilst being an excellent method for gas analysis, this method requires sample pre-treatment, and relatively long timespans for measurement. Also, the results obtained by this method sometimes differ from those obtained from human sensory tests. Therefore, the development of a rapid and reliable odour-sensing system is highly desirable to overcome such problems. The present authors have developed an odour sensing system, using a QCM (Quartz Crystal Microbalance) array and neural-network attern recognition', by which whiskey aromas,perfumes, and flavours can be identified2*! A system for measuring many samples involving an automatic sampling stage was M h e r developed and applied to flavour discrimination4.

2. RECENT DEVELOPMENTS While many types of gas sensors have been studied, the selectivities of such sensors have been found to be insufficient. It is said that an odour may be distinguished by the output pattern of many receptors having partially overlapping characteristics in the olfactory system5. Therefore, it would be useful to determine the output pattern from a plurality of different sensors so that selectivity can be raised. Characteristics of the hitherto main odour sensors are listed in Table 1.

Reference Gardner'

Sensor Semiconductor or metal oxide

Persuad '

Conducting polymer

Moriizumi & Nakamoto'

QCM

Advantages -Comparatively insensitive to humidity. -Most popular one. -Easy to integrate. -Used at room temp. -Diversity in sensing film -Used at room temp.

Disadvantages -Slight degree of freedom. -Work at high temp. -New technology. -Sensitive to humidity -Sensitive to humidity

In using our QCM gas sensor, various characteristics can be easily changed by selecting different films suitable for the target odours. Also, a QCM gas sensor is cheaper than any other gas sensor. Recently, commercial instruments using semiconductor metal oxides or conducting polymers have been developed in the United Kingdom and France.

3. PRINCIPLE AND PROCEDURE 3.1 Basic Principle

The olfactory system of a living body was mimicked in developing our odour sensing system. In recent years, a better understanding of human olfaction has been achieved, leading to improvements in system design. Several researchers have employed a plurality of gas sensors with pattern recognition techniques, to obtain outputs to distinguish between gases. Hence, QCM's were used instead of olfactory cells, and neural networks or multivariate analysis was used instead of the olfactory nervous system. 3.2 The QCM System

A quartz resonator sensor is composed of a quartz plate with electrodes on both sides. The resonator with sensing film coated over the electrodes can work as the sensor (Figure 1). As odourant molecules are adsorbed onto the film, the resonance frequency decreases. After desorption, it recovers. This phenomenon is called the mass loading effect. The shift in frequency being proportional to the total mass of adsorbed odourant molecules. In this study, 10 MHz AT-cut quartz resonators were used. The sensor response of the QCM

Qas Sensing

Film

I 1 1

E lectroda

Ouartr Plate

Time

Figure 1 A Quartz-Resonator Sensor.

( sec.

I

Figure 2 Response to Ethanol Gas.

89

An Odour Sensing System for Use in Measuring Volatiles in Flavours and Fragrances

coated with ethyl cellulose is shown in Figure 2. When ethanol vapour is adsorbed onto the film the frequency decreases. The frequency than recovers when dry air is supplied, thus restoring the sensor. In Figure 1, when the shift in frequency is 300Hz,the output of the sensor would be 300Hz.

3.3 Procedure A schematic diagram of our system is shown in Figure 3. the system is composed of a vapour supplying system, a sensor cell and an electronic portion. A sample vapour is injected into the sensor cell where eight sensors were installed to obtain data. Dry air is then supplied to refresh the sensors. After recovery of responses, the next sample vapour is injected to repeat the process.

;I

IEI\,M

Figure 3 Diagram of Odour-Sensing System Using an Auto-Sampling Stage.

Crll

Figure 4 Photograph of Automatic Sampling System: (a) Whole System, (b) Details of Syringe Needles.

90

Flavours and Fragrances

In our present system, liquid odour samples are poured into vials which are sealed with rubber stoppers and positioned. Moving a mechanical stage along XYZ axes, two syringe needles are positioned and made to penetrate the vials through the rubber stoppers. In figure 4(b), three solenoid valves attached to the XYZ stage move along with the syringe needles. Upon penetration, controlled dry air, at a rate of 120ml/min, is supplied through one needle, thus pushing the odour though the other and into the sensor cell. Then, with the vial now empty, dry air is supplied in the same manner to cleanse and restore the sensors and tubes.

4. EXPERIMENTAL (ODOUR SEPARATION) The data obtained may be said to be eight dimensional, since eight sensors are used in the present study. Thus, PCA erincipal Component A n a l y ~ i s ) ~a , dimensional reduction technique is used, converting the results into two visual dimensions. One can then study the pattern differences among the odour samples. the data is analysed after the normalisation of : where XI is the signal of the I-th sensor

k’/C,”=,X,

The following experiments on odour separation for flavour and fragrance identification were performed. (1) 4 typical essential oils, (2) 5 citrus essential oils, (3) 20 typical essential oils. The sensing films used were gas chromatographic stationary phase materials, cellulosic materials and lipid films. 4.1 Separation of Four Typical Essential Oils The four essential oils were; ginger, clove, Valencia orange, and peppermint oil. All have typical aromas that can be easily distinguished by an ordinary person. All samples were measured eight times to assure reproducibility of data. The films used are shown in Table 2. Figure 5 shows a scatter diagram of the four essential oil samples having different aromas. The four samples showed great differences. 22 126261

Table 2 Films Used in Sensor Array.

f. Clove

i‘

Xo

I 2 3

Z1 72x1

4

5 6 5

Peppermint

R

GIIII inaterial Dioleyl phospbatidyl seriu Cholesterol Pertluorinated bilayer Lecithin Diet,hyleneglycol si1ccinat.e Sph.yngoul,yelin Acet.yl cellnlose Ethyl cellulose

Vnlencia Ornngc

Figure 5 Scatter Diagram of Four Typical Essential Oils.

ClassiGcat.ion lipid sterol synthesized lipid lipid

GC:

lipid cellulose cellulose

91

An Odour Sensing Systemfor Use in Measuring Vola!iles in Flavours and Fragrances

Table 3 Films Used in Sensor Array.

T

Filni niat.erial 1 Dioleyl phosphat,idyl seriu 210.

Sphiugoniyelin Diole.yl phosphatidyl choline Lecit.biu Pertluoriiiakd bilayer Sphyugoniyeliu 33%# Cbolest.crol67% Spbyugoiuyelin 50% Lecitbiu50% Ethyl cellulose

2

3 4

I

Lime

Valencia orange+.

.if'

5 8

'. +-

7

R

ClavsiGcariou lipid lipid lipid lipid syutbesized lipid lipid srerol lipid lipid cellulose

Calabria orangc

-2

Figure 6 Scatter Diagram of Citrus Essential Oils.

4.2 Five Citrus Essential Oils Using the films listed in Table 3, the output patterns of five citrus essential oils were measured. The oils were; grapefruit, lemon, lime, Calabria orange, and Valencia orange. These oils are not readily distinguished from each other by the average individual. Procedures were the same as those described for the first experiment. Results by PCA are shown in Figure 6. All samples were measured four times. The points for each group were close to each other, showing that their aromas were similar. As can be seen from the figure, each sample can be distinguished from the others. Lime was furthest removed from the other four samples. Lemon was also separated from the three samples of grapefruit, Calabria orange, and Valencia orange. The two orange oils were closely related, but still separated. Coefficient of variation indicates that experimental error was 1.2%. Table 4 Essential Oils Used in the Experiment. C:lasiIicat.iou Citms Floral Auis

Miut Resiu Riist.is Woody Cirroiiella Medical Spic.v

Figure 7 Scatter Diagram of 20 Typical Essential Oils.

Esseuriol oil Bwganiout Ylaug yhng, Canauga Fenuel. Star anis: Caraway Majorani, Spanlint., Peppermint. Galbauuin Eucalypt.us, Roseniaiy, Thyme Ceder wood, Sandal wood

Leiiiougrass

Wiut.ergreeu Nut.tneg, Csrdaniou. (:assia

92

Flavours and Fragrances

4.3 Twenty Typical Essential Oils

Using the films listed in Table 3, the output patterns of twenty typical essential oils (Table 4) were measured. Included were essential oils having citrus, floral, mint, and spicy notes etc. Their scatter diagram developed by PCA with a variance of 97% with respect to the two major principal axes shown in Figure 7. Ail samples were measured four times. It was found that the twenty essential oils were fairly well separated and that they were clearly distinguishable.

For twenty samples, each being measured four times, the total time taken was 320 minutes. 5 . SENSING FILM (CHARACTERISATION) An odour-sensing system, using an automatic sampling stage was developed. Since the

response time of the QCM sensors was fast, it was found to be considerably efficient in measuring numerous samples. The various features of QCM become available, by the selection of the appropriate films suitable for the object odour. Thus, the characterisation of sensing films is essential, and well suited films can be selected from the many kinds available. Sensing films were first grouped, using sensor responses obtained for typical aroma compounds. Next, sensor responses using the material parameters were predicted. 5.1 Samples

As the thic.kness of the coating films were different, the data was analysed after .the nonnalisation of AFs f 4=, where is the frequency change due to the adsorption of sample vapour and 4cis the frequency change due to coating a film on the QCM. The samples tested were compounds with the typical functional groups listed in Table 5. They were aliphatic hydrocarbon, aromatic hydrocarbon, ester, ketone, alcohol, acid, and base. The samples vapour pressure was not high and ranged between 1 and 30 mmHg at 18°C. There were two samples in each functional category, with the number of carbon atoms differing by one. The film materials tested are shown in Table 6. Most of these films are typically used in odour-sensing systems.

Aliphatic Hydrocarbons Aromatic Hydrocarbons Esters Ketones Alcohols Acids Bases

1. Heptane 3. Benzene 5. Propyl Acetate 7. 2-Pentanone 9. Propanol 1 1. Acetic Acid 13. Pyridine

2. Octane 4. Toluene 6. Butyl Acetate 8. 2-Hexanone 10. Butanol 12. Propionic Acid 14. 2-Methylpyridine

An Odour Sensing Sysiernfor Use in Measuring Volatiles in Flavours and Fragrances

Table 6 Film Materials tested. Classifications Lipids Sterol Celluloses Cyclodextrins GC (Polar) GC (Slightly polar) GC (Non-polar) Others

93

No. of Films 11 1 3 5 7 3 5 2

5.2 Film Materials (Classification)

Results of the film properties as analysed by PCA are shown in Figure 8, where the fourteen dimensional data was reduced to two dimensions. The films with similar characteristics were found to be grouped close to each other. Most of the polar GC stationary phases were located in the positive value region of the first principal component, while most of the non-polar or those that were slightly polar were located in the negative region. Furthermore, the films of phosphatidyl lipids and cyclodextrins were located in the centre, the two being separated by the third principal axis (Not shown). The results suggest that the 37 film materials could be classified into mainly four groups such as polar GC materials, non-polar GC materials, cyclodextrins, and lipids. Since lipids are located between polar GC and non-polar GC stationary phase materials, their characteristics could not be realised using GC films with different polarities. 5-

i

- T - I l - m 1 7 -

.GC

0

iar)

T--

15

29

Lipids

CC (polar)

I

Figure 8 Scatter Diagram of 37 Typical Sensing Film Materials. 5.3 Frequency Shift Prediction

Prediction of the frequency shifts was attempted using multiple regression analysis. The material parameters used for the analysis are listed in Table 7. The stepwise method was used in the regression analysis to select predictors form the twelve parameters listed in Table 7.

94

Flavours and Fragrances

Table 7.List of Aromatic Sample Parameters. 1. Molecular weight 3. Refractive index 5 . Dipole moment 7. Heat of combustion 9. Solubility parameter 11. Kovat's index (OV-101)

2. Boiling point 4. Dielectric constant 6. Coefficient of viscosity 8. Heat of vaporisation 10. Kovat's index (PEG 20M) 12. Parameter 10 / Parameter 1

A prediction model for each sensing film material was made. The predicted values versus the corresponding measured values for the PEG 200 and squalane are illustrated in Figure 9(a,b). PEG 200 is located on the far right of the scatter plot (Figure 8), whereas squalane is located on the left side of the same plot (Figure 8). Thus, the characteristics of these two films is quite different. The predictors 8, 10, and 12 were selected for PEG 200, with the multiple coefficient of determination R2 being 0.948. The predictors 5, 6, 7, and 9 were selected for squalane, with R2 being 0.980.

0

200

400

Measured Value

6W

Measured Value

Figure 9 Plot of Predicted vs Measured Frequency Shifts for 14 Samples of Sensing Films (a) Material 23: PEG 200, (b) Material 3 1: Squalane.

6. CONCLUSIONS The authors have developed an odour-sensing system using an automatic sampling stage. Due to the shorter response time of QCM sensors, this system was efficient for measuring many samples. QCM gas sensors are cheaper than any others. Their characteristics can easily be changed by selecting the appropriate different sensing films. In classifying the sensing film materials, it is suggested that the film materials should be grouped into the four categories; polar GC materials, non-polar GC materials, cyclodextrins, and lipids.

An Odour Sensing System for Use in Measuring Volatiles in Flavours and Fragrances

95

Prediction of sensor response was made possible with the use of regression analysis, and the material parameters of the vapours. Mapping of the combined sensor output pattern and the human sensory tests is viewed to be a worthwhile future research target. Further detailed, and extensive work on film selection and stability enhancement is required. References 1. T. Nakamoto and T. Moriizumi, “Odour sensor using quartz-resonator array and neural network pattern recognition”, Proc. IEEE Ultrasonics Symp., Chicago, IL, USA., 1988,613-616. 2. T. Nakamoto, A. Fukuda, and T. Morizumi, “Identification capability of odour sensor using quartz-resonator array and neural-network pattern recognition”, Sensors and Actuators B, 1, 1990,473-476. 3. T. Nakamoto, A. Fukuda, T. Morizumi, and Y. Asakura, “Improvement of identification capability in an odour-sensing system”, Sensors and Actuators B, 3, 1991, 221-226. 4. J. Ide, T. Nakamoto, and T. Moriizumi, “Development of odour-sensing system using an auto-sampling stage”, Technical Digest of the 4th Int. Meet. on Chemical Sensors, 1992,414-417. 5. K. Kurihara, T. Nomura, M. Kashiwayanagi, and T. Kurihara, “Transduction mechanism of chemical sensors in biological systems and its application to artificial membrane sensors”, Tech. Digest, 4th Int. Conf. Solid-state Sensors and Actuators (Transducers ‘87), Tokyo, Japan, June 2-5, 1987,569. 6. H.V. Shurmer, and J.W. Gardner, “Odour discrimination with an electronic nose”, Sensors and Actuators B, 8, 1992, 1- 11. 7. K.C. Persaud, and P. Travers, Intelligent Instrumental Computers, Elsevier Science Publishing, N. York, 1991, 147-154. 8. K. Ema, M. Yokoyama, T. Nakamoto, and T. Moriizumi, “Odour-sensing system using a quartz-resonator sensor array and neural network pattern recognition”, Sensors and Actuators B, 18, 1989,289. 9. W.R. Dillon, and M. Goldstein, Multivariate Analysis Methods and Applications”, J. Wiley and Sons, Inc., N.York, 1984.

The Aromatic Resins: Their Chemistry and Uses David A. Moyler' and Robin A. Clery2

' H. E. DANIEL LIMITED, TUNBRIDGE WELLS, KENT TN2 3EY, UK QUEST INTERNATIONAL, ASHFORD, KENT TN24 OLT, UK

1 INTRODUCTION Aromatic resins have a long and safe history of use in the formulation of Flavours and Fragrances. Their effective contribution to the perceived functional performance and enjoyment of consumer products justifies their position as key ingredients. The importance of the use of balsams in modem creative perfumery has been published by Boelens [l]. This current paper gives an insight into the chemical composition by GCMS, standardised relative retention indices, odour and flavour profiling and applications of ten of the most popular aromatic resins. 1.1

The Aromatic Resins, Botanical names and Origins

Galbanum Olibanum Myrrh Opoponax Labdanum Elemi Copaiba Peru Benzoin

Tolu

Ferula galbanifua, Boswellia sp Commiphora sp. Commiphora erythraea Cistus ladaniferus Canarium sp. Copailera sp. Myroxylon pereirae Styrar benzoin Myroxylon sp.

Iran East Africa East Africa East Africa Spain Philippines South America South America East Indies West Indies

The gum resin exudates obtained from plant materials have a very long history of use in the Flavour, Fragrance and Pharmaceutical Industries.

The Aromatic Resins: Their Chemistry and Uses

97

Their use was first referenced in the Old Testament in Genesis, 37 : 25. They were incorporated into early fragrances such as Kyphi, by the ancient Egyptians [2]. Harvesting of the gums is illustrated in the old [3] engraving shown here. Such collecting and subsequent hand sorting is still carried out today in many of the producing areas. These gums are a totally natural renewable botanical source which provide a living for many native populations, many of which live in Third World Countries [4].

Figure 1 Historical Engraving of the Harvesting of Frankincense Gum 1.2 Terminology

There is considerable conhsing overlap about the terminology used to describe the gums and their extracts. Table 1 is an attempt to clarify some of the definitions, but even this does not prevent the historical misuse and accepted custom and practice of certain names being used for these extracts.

98

Flavours and Fragrances

Table 1 Some Terminology usedfor Natural Aromatic Gum Extracts. Type

Definition

Example

Gum

Semi-solid natural aromatic material.

Benzoin, Elemi gums

Balsam

Natural raw material exuded from a tree or plant.

Peru balsam

Extract

Concentrated products obtained from treatment of a natural product.

Hot alkali extract of Labdanum leaves, twigs

Resin

Natural: formed by oxidation of terpenes in the botanical. Prepared: oleoresins with the essential oil removed.

see balsam

Oleoresin

Extracts of botanicals with solvent removed. Usually food materials.

Vanilla oleoresin

Resinoid

Extracts of botanicals with solvent removed, viscous and dark coloured. Usually perhme materials.

Myrrh resinoid

Resinol

Selective extracts with solvent removed, mobile at ambient temperature, light in colour. Perfume materials.

Myrrh resinol

C02 Extract

High pressure carbon dioxide extract at low temperature with features of both the absolute and essential oil.

Olibanum C 0 2 (Frankincense) 151

Absolute

Alcohol soluble selective extract, the resinous components removed.

Galbanum absolute

Essential oil

Gum that has been water, steam or

Copaiba oil (vacuum distilled)

vacuum distilled as appropriate.

A collection of Resinols, Resinoids, Absolutes, C02 extracts and Distilled Essential oils are available commercially, all exhibiting a natural balsamic richness, each with its own individual characteristics. The resin range is equally applicable for Fragrances, Flavours, Aromatherapy and Pharmaceuticals; their true to nature characteristics finding uses in products needing balsamic notes and good cost in use performance.

99

The Aromatic Resins: Their Chemistry and Uses

2

RESIN EXTRACTS

Commercial Resin Extracts are available as the products of several different extraction and distillation technologies. Table 2 Commercially available Resin Extracts Gum Resinoid, Galbanum Y Olibanum Y Myrrh Y Opoponax Y Labdanum Y Elemi Y Copaiba Y Peru Y Benzoin Y Tolu Y

Resinol, Y Y Y

Absolute, Y Y Y

Y Y Y Y Y Y Y

Y Y Y

CO oil, Y Y

Y

Essential oil Y

Y Y Y Y Y Y Y

Y

They are all manufactured to the highest standards of the Industry fiom carefully selected prime source gums and offered as high quality extracts. 3

GALBANUM

The Galbanum plant Ferula galbaniflua, is a member of the Urnbellifrue ( umbrella like ) family of plants which grow wild in Iran and surrounding areas. When lifted fiom the ground, the exposed roots are cut by hand with a knife and a soft, ( so called Levant ) gum resin is exuded, which is very viscous having the consistency of fresh honey. The commercial gum contains woody root splinters, sand, gravel, plant fibres and water. Galbanum odour is powerful, herbaceous, green-woody like bell peppers, salad leaves or hyacinth flowers. The characterising components are a group of trace thiocompounds, undecatrienes and pyrazine derivatives whose chemistry has been studied and reported in detail [ 6,7,8 1. The main green note being;

2-(sec-butyl) 3-methoxypyrazine The main thio compound being;

?-

a \ 0

sec-butyl3-methyl-2-butenethioate

100

Flavours and Fragrances

Table 3 Composition of Galbanurn Oil Component Tricyclene alpha Pinene Camphene beta Pinene Sabinene delta 3 Carene beta Myrcene Limonene cis beta Ocimene trans beta Ocimene Terpinolene EZ 1,3,5 Undecatriene EE I ,3,5 Undecatriene other 98 components

DB wax retention index 1016 1021 1068 1111 1124 1142 1159 1193 1231 1251 1286 1456 1471

--

Area % 2.3 9.6 0.2 53.7 1.4 13.1 2.9 1.8 2.3 1.3 0.5 0.5 0.2 10.2 100.0

3.1 Galbanurn Resinoid is a brown viscous liquid with thecharacteristic green odour supplemented by a resinous, balsamic backnote. It is a good fixative and blends much better than artificial resins because its odour does not “grow or fade” in product.

3.2 Galbanurn Resinol is pale coloured and easily poured but has all of the attributes of fixation and performance of the Resinoid, with just the coloured sticky gum components replaced. The performance strength and retention is readily appreciated in cosmetic creams, cream soaps, soaps, detergents and talc fragrances. 3.3 Galbanurn Absolute is a mobile almost colourless liquid which is hller bodied and much more substantive than the oil that it can be used to replace. The odour is less terpenic than that of the oil, a fact confirmed by GLC analysis. It is more intense and cost effective than the oil, closely resembling the volatile headspace of the gum itself. 3.4 Galbanurn COz is a pale yellow liquid of similar properties and performance to the absolute. The liquid gum is usually adsorbed onto a solid inert substrate and packed into a column before extraction with carbon dioxide. When supercritical conditions are used and a resinoid produced, the extract has to be fractionated at near critical conditions to obtain the extract with the characteristics of the absolute. The efficient separation of the undesirable resins plays an important part in the processing.

Galbanurn oil is often water distilled rather than steam distilled to give a purer, fresh, green note of true fidelity.

3.5

101

The Aromatic Resins: Their Chemistry and Uses

3.6

Applications Fragrances:- All galbanums find extensive use in modest amounts in the formulation of chypres, fougeres, pines, also giving natural leafy character to many florals. For instance, it is indispensable in hyacinth bouquets and fragrances of the spring fresh type. Flavours:- The fresh green leafy characteristics of galbanum has uses in vegetable, fruit and flavours for salad dressings. There are few or no applications to Aromatherapy or Pharmaceutical products. 4

OLIBANUM

The gum which exudes from ducts or incisions made in the bark of the Olibanum or Frankincense tree Boswellia sp, a member of the Burseraceae family, is a golden coloured “tear” shaped solid which melts on warming. It can be pyrolysed to give the rich aromatic incense so traditionally characteristic of churches. There are several varieties of the tree including principally B. carterii and B. jereana but about a dozen others as well [ 9 3. Quality variations can only be judged after extraction, as the gums are physically almost indistinguishable. Olibanum odour is dry, slightly green, fresh-balsamic and tenacious. Its characterising components are the Incensoles, a series of terpenoids, consisting of the alcohol, acetate and oxide [10,11,12];

H

Incensole Table 4 Composition of Olibanum Oil (B. jereana) [ 131

Component alpha Pinene alpha Thujene Camphene beta Pinene Sabinene delta 3 Carene beta Myrcene alpha Phellandrene L (-) Limonene alpha Terpinene gamma Terpinene para Cymene is0 Thujone beta Caryophyllene Terpinen 4 01 delta Cadinene

DB wax retention index 1021 1027 1068 1111 1124 1142 1159 1168 1193 1200 1245 1269 1419 1596 1604 1753

area YO 27.0 13.5 0.5 1.o 4.1 0.8 8.7 1.7 16.0 0.6 0.5 4.1 1.2 4.2 1.2 1.6

102

Flavours and Fragrances

2693

0.5 12.8 100.0 4.1 Olibanum Resinoid is a dark brown viscous liquid with the familiar freshbalsamic note of the gum. It is an excellent fixative in creams and candle products, where alcohol solubility does not feature. It gives a long lasting balsamic freshness which is difficult to achieve with other resins. Incensole others I32 components

--

4.2 Olibanum Resinol is the pale coloured easily pourable more concentrated version of the Resinoid and is stronger and of greater staying power. This ease of use saves time in compounding and avoids the need to heat in order to mobilise before weighing, so avoiding the possibility of any pyrolysis or burning, as the resins are not good conductors of heat.

4.3 Olibanum absolute is a mobile pale oily liquid which is similar in consistency, only a little thicker than the oil. The odour compared to the oil is richer, less terpenic and this can be readily confirmed by the GLC analysis which shows much lower monoterpene hydrocarbon levels and substantially more Incensole [ 5 ] . The absolute can be molecular distilled and contains solvents like isopropyl myristate, but these should be clearly described in any product literature. Olibanum C 0 2 this commercially available extract is a mobile oil with properties similar to the molecular distilled absolute. The oil is a fractionated supercritical carbon dioxide extract. 4.4

Olibanurn oil can be steam distilled or a true water distilled oil, produced under the optimum conditions from the grade of gum that has been selected to give the oil of the best odour. This is best described as diffusive, fresh apple peel, rich and sweetly balsamic. 4.5

Applications Fragrances: Olibanums blend with spices, Orientals, ambers, and are very good citrus cologne modifiers, imparting balsamic freshness. Flavours: For reasons of solubility the oil and absolutes are preferred in the limited flavour uses for Olibanum. 4.6

5

MYRRH

The Myrrh tree Commiphora sp. is a member of the Bursuraceae family, which grows in North Eastern Africa and Southern Arabia [3]. It gives a dark hard resinous exudate which has been used since Biblical times for early fragrances and as a preservative for mummification. The commercial gum is made up of “tears” and rounded masses with a red - brown colour, which can sometimes have a dusty appearance. Myrrh odour is warm, balsamic, sweet and a little spicy; distinctly different from the other gums. The characterising components are the Curzerene and Lindestrene family of odiferous furanosesquiterpenes [ 14,15,16] .

103

The Aromaiic Resins: Their Chemistry and Uses

Lindestrene c 1 [15] Table 5

Lindestrene c 2 [15]

Curzerene

Composition of Myrrh Oil

Component

DB wax retention index

para Cymene delta Elemene alpha Copaene Bourbonene Alloaromadendrene delta Cadinene Curzerene Lindestrene C1 [15] Lindestrene C2 [15] Curzerenone Furanogermacrene other 108 components

1269 1498 1504 1510 1611 1753 1802 2044 2052 207 1 2546

--

Area 'YO 1.4 2.8 3.9 8.3 5.4 2.6 12.7 23.4 6.4 1.9 3.0 28.2 100.0

5.1 Myrrh Resinoid is a clear red - brown soft resin with the intense, rich, spicy balsamic note of the gum. It finds uses as a fixative, especially in fine soaps, where its warm aromatic lasting power can readily be appreciated.

Myrrh Resinol has the same fragrance performance attributes as the resinoid but is lighter in colour and being a pourable extract it is easier to use. It can contain diethyl phthalate to fluidise it, and its solubility is much better in this solvent than any other. 5.2

5.3 Myrrh Absolute is a mobile slightly viscous liquid which has been separated from top quality aromatic gum. Compared to the oil, it has more depth and staying power associated with the higher content of Curzerene derivatives, which is confrmed by GLC analysis [5]. Myrrh absolute can be processed by molecular distillation and contains di ethyl phthalate. 5.4

Myrrh C02 is a red-brown oil with similar properties to the molecular distilled

oil. Myrrh oil can be steam or a true water distilled oil, made from selected gum, under the conditions shown to give oil of the best odour. The odour is described as being intensely warm spicy balsamic in character and a sweetness which increases to a deep aromatic dryout which is difficult to duplicate from other resins. 5.5

-

Flavours and Fragrances

104

As the density of this oil is close to 1.0, some difficulties can be experienced in collecting fractions of the oil as the waterloil condensate collects. Applications Fragrances: All of the Myrrhs can be used in small amounts in heavy floral perfumes, Orientals, fougeres and woody balsamic bases. Flavours: The burning taste of Myrrh is used to good effect in oral hygiene products and in a famous brand of fantasy flavoured liqueur.

5.6

6

OPOPONAX

Although from the same Commiphora family as Myrrh, Opoponax gum Commiphora erythrea var. Glabrascens, has characteristics of its own. The crude exudate comes in regular nut sized lumps which are dark red in colour. Like many of the other gums, it also comes from the “Horn of Africa “ and is harvested and graded in the same way. Opoponax odour is sweetly aromatic, with a tenacious animalic, woody root - like note, not found in Myrrh. IFR4 recommends that the level of extracts and distillates be restricted to 0.6 % in consumer products and the crude gum itself is not used. The main characterising components are the Santalenes and Bisabolenes [ 16,17,18], the principal isomers being :-

alpha Santalene

alpha Bisabolene.

Table 6 Composition of Opoponax Oil Component HP 5 retention index cis beta Ocimene 1036 trans beta Ocimene 1049 cis alpha Bergamotene 1415 alpha Santalene 1420 trans alpha Bergamotene 1436 epi beta Santalene 1449 trans beta Farnesene 1458 beta Santalene 1462 AR Curcumene 1483 Curzerene 1496 alpha Bisabolene (cis and trans) 1504 beta Bisabolenes 1509 gamma Bisabolene 1515 * 1902 * current 2090

Area YO 0.4 26.2 0.9 20.1 5.2

0.8 0.6 0.5 1.1 0.6 14.2 1.6 1.3 0.6 0.13

105

The Aromatic Resins: Their Chemistry and Uses

* *

211

other

0.09 0.12 25.56 100.00

2168 2280

research

--

components

Opoponax Resinoid is a dark brown viscous liquid with the characteristic sweet spicy note of the gum. Its fixative properties can be utilised in cosmetic cream, talc and candle fragrances to give the warmth of a leather type note. 6.2 Opoponax Resinol is the lighter coloured pourable version of the Resinoid, in a more concentrated form. It has retained all of the depth of sweet spicy note and can be used as a replacement for the resinoid in all its applications. Opoponax Absolute is a mobile pale oily liquid with the complete volatile 6.3 odour profile of the Bisabolenes from the gum, in an alcohol soluble form which is ideal for incorporation into fragrances and colognes. Opoponax absolute can be molecular distilled and contains isopropyl myristate. 6.4

Applications Fragrances: All Opoponax extracts find limited use in p e r f b e bases of the woody, oriental, leather, chypre and fougere types. Flavours: The wine residue and vegetable soup like notes of opoponax may find some use in flavours.

LABDANUM

7

The small wild shrub Cistus ladangerus, which grows in the coastal regions of the Western Mediterranean, gives a dark red - brown resinous exudation from its stems and twigs. The commercial Labdanum gum however, is obtained as a total extract of the leaves and twigs when forcibly treated in alkaline boiling water. Gum floats on the waters surface and is skimmed off and dried [ 19 1. Figure 2

Flowchart of the Extracts of Labdanum

CISTUS LADANIFERUS PLANTS (twigs & leaves)

I hydrodistil

t CISTUS oil

I

I

non-polar

semi-polar

t

t

CISTUS resin

I

LABDANUM resin

I

ethanol

AMBREINE

I

t

t LABDANUM md.

-

v

LABDANUM gum

I

molecular distil

CISTUS nbr

hcxane

hot alkali

hydrodistil

t

1 dly

t

a

LABDANUM oil

LABDANUM concrete ethanol

LABDANUM absolute

I06

Flavours and Fragrances

Labdanum odour is sweet, herby - balsamic, amber animalic, rich and long lasting. The characterising components are sesquiterpenic and of the families of Ledol and Ambrox [20,21,22,23].

Ledol

Ambrox

Table 7 Composition of Labdanum Cistus Oil Component Furfural Tricyclene + alpha Thujene alpha Pinene Camphene Benzaldehyde Sabinene beta Myrcene alpha Phellandrene + delta 3 Carene alpha Terpinene para Cymene Limonene + beta Phellandrene 1, 8 Cineole cis beta Ocimene trans beta Ocimene gamma Terpinene Terpinolene Linalol alpha Copaene beta Caryophyllene Ledene delta Cadinene Ledol Viridiflorol Ambrox (three isomers [23]) Other 220 components

HP 5 Retention index

830 932 940 95 5 96 I 977 99 1 1007 1018 1027 1032 1036 1040 1049 1066 1093 1102 1376 1423 1495 1524 1590 1601 1756

--

Area YO

0.5 1.2 40.8 6.0 0.9 0.5

0.5 0.3 0.4 1.9 1.9 2.9 0.2 1.1 1.o

1.8 0.5 0.2 0.2 1.6 0.7 2.62 1.07 0.16 3 1.05 100.00

Labdanum Resinoid is a thick, high viscosity, almost black gummy resin, which can not be decolourised by simple methods using citric acid and charcoal. It is an excellent fixative and is frequently used in coloured soaps and detergents. 7.1

I07

The Aromatic Resins: Their Chemisrry and Uses

Labdanum Resin01 is a light pourable extract which can be utilised in the same 7.2 way as the resinoid but with the advantage of being pale and non discolouring, applicable even to white toilet soaps. Labdanum Absolute an almost colourless, mobile oil which captures all of the volatile aromatic components of Labdanum in a form which is ideal for incorporation into fine fragrances and colognes. The odour is intense and captures all of the Ambrox components to give a good cost in use. Labdanum absolute can be molecular distilled and contain isopropyl myristate. 7.4

Applications Fragrances: All Labdanums give a cost effective, deep, rich, leathery amber note, which has a wide range of fragrance applications across the perfumery range. Their odour performance is excellent in cosmetic cream, soap, detergent and highly acidic consumer products. Flavours: Labdanum is used in some flavours for tobacco.

8

ELEMI

The natural oleoresin gum which exudes from Canarium sp., a member of the sub-group of Terpentines grown in the Philippine Islands. There are several varieties of the tree all being related to the same family as Olibanum and Myrrh. The crude Elemi gum is a semi-solid, off-white wax like mass which often contains splinters of wood, bark, dirt and earth. Quality variations of Elemi are best judged after distillation of a sample and assessment of the odour and analytical characteristics of the oil produced. Elemi odour is very fresh, terpene like, peppery, citrus with a woody balsamic backnote. The characterising components are the monoterpenes, especially phellandrene and the sesquiterpene alcohol Elemol [24].

A I

v

n alpha Phellandrene

-

\\/\/\(OH Elemol

Flavours and Fragrances

108

Table 8 Composition of EIemi Oil

Component alpha Pinene alpha Thujene beta Pinene Sabinene beta Myrcene alpha Phellandrene Limonene 1, 8 Cineole beta Phellandrene cis beta Ocimene gamma Terpinene trans beta Ocimene para Cymene Terpinolene trans Sabinene hydrate Caryophy Ilene Terpinen 4 01 alpha Humulene Methyl Eugenol Elemol Elemicin Other 82 components

DB wax retention index 1021 1027 1111 1124 1159 1168 1193 1214 1220 1231 1245 1251 1269 1286 1462 1590 1596 1669 2002 2056 2210

--

Area YO 0.4 0.1 0.3 5.9 0.6 11.7 65.0

2.5 1.6 0.5 0.3 0.3 1.4 0.3 0.2 0.05 0.5 0.01 0.2 2.8 1.8 3.54 IOO.00

Elemi Resinoid is a pale yellow semi-solid mass with the typical fresh woody resinous note of the botanical. It is a good fixative for use in soaps and detergents, especially considering its modest cost. 8.1

Elemi Resinol has the odour and performance of the Resinoid but in a light coloured easily pourable form. 8.2

Elemi Absolute is a waxy solid which can be made by extraction of the gum 8.3 with alcohol in which it is not very soluble; or as a colourless oil by molecular distillation at low temperature and pressure. The absolute has a desirably lower content of terpenes than the oil, which improves stability and shelf life in finished product. Elemi oil is colourless and has all of the desirable characteristics of the botanical in an easy to use soluble form. 8.4

8.5

Applications Fragrances: Elemi is very useful as a topnote freshener in many compositions, especially the white florals.

109

7’he Aromatic Resins: Their Chemistry and Uses

Flavours: Elemi blends well with other essential oils when used in flavours. 9

COPAIBA

The large wild tree Copaifera sp. of which there are several species, grows in Brazil and other Countries in the northern part of South America. The natural oleoresin flows from cavities which are made in the trees trunk. The balsam is sometimes hazy due to the presence of water andor resinified material. Copaiba balsam odour is faint, slightly woody and peppery and only moderately tenacious. The main components are sesquiterpene hydrocarbons, principally beta Caryophyllene, but include Copaenes, Bergamotenes, Cubebenes and Elemenes.

beta Caryophyllene Table 9 Composition of Copaiba Balsam Oil Component delta Elemene alpha Cubebene alpha Copaene beta Cubebene beta Elemene cis alpha Bergamotene trans alpha Bergamotene beta Caryophyllene alpha Humulene Other 35 components

DB wax retention index 1451 1464 1491 1521 1536 1571 1584 1596 1669

_-

Area % 1.8 0.9 8.8 0.5 0.6 1.2 12.7 41.6 5.8 26.1 100.0

9.1 Copaiba Resinoid is a brown viscous liquid with a mild-woody, spicy odour which is moderately long lasting. Copaiba Resinol retains the odour and performance characteristics of the balsam in a clear, mobile pale brown extract which makes it more applicable than the balsam as a modifier and blender, to light coloured soaps and toiletries. It is used in detergent and industrial fragrances. 9.2

9.3 Copaiba oil is a clear colourless oil which is high vacuum distilled directly from the balsam. The optimum conditions use the minimum of heat necessary to complete the processing, in the minimum “still” residence time. This oil captures all of the fruity,

I10

Flnvours and Fragrances

spicy character of the balsam in an alcohol soluble form which is ideal for incorporation into cologne formulations. Applications Fragrances: The copaibas do find some use in fragrances, but are not particularly good fixatives. They are blenders and modifiers that are best used where they can harmonise with ylang, heliotropine, vanillin, the ionones etc. They are used in pine, woody, sweet floral and spice fragrances where their low cost offers some advantages to the creative perfumer in lower cost formulations. 9.4

Flavours: The bitter and slightly irritating flavour of Copaiba do not make it a popular ingredient. 10

PERU

The Central American tree Myroxylon pereirae, now commercially grown in El Salvador rather than Peru, gives a gum exudate which is a dark red - brown, viscous but pourable liquid. When left to evaporate, Peru balsam does not dry to a sticky residue like some other gums, but stays oily and mobile. Peru balsam odour is the reference balsamic note; rich, sweet, soft and long lasting with a vanilla - like dryout. Its characterising components are cinnamic derivatives and vanillin [1,19]. IFRA recommends that the level of extracts and distillates be restricted to 0.4 % in consumer products and the crude balsam itself is not used [25].

Table 10 Composifion of Peru Balsam Oil Component Benzaldehyde Benzyl Alcohol Methyl Benzoate Benzoic Acid Methyl Cinnamate Vanillin Cinnamic acid Nerolidol b Nerolidol a Benzyl benzoate cis Cinnamyl Benzoate

HP 5 retention index 96 1

1035 1091 (1 200)* 1379 1391 (1500)* 1534 1564 1762 1930

Area % 0.03 1.6 0.02

(2.0) 0.01 0.4 (0.1)

0.2 2.2 86.2 0.0 1

111

The Aromatic Resins: Their Chemistry and Uses

0.37 0.01 6.0 0.004 0.0 1 0.006 0.01 -0.82 100.00 * Retentions vary with quantity of acid due to reaction on the column and area % for these . . acids are not.

Benzyl cis Cinnamate trans Cinnamyl Benzoate Benzyl trans Cinnamate cis Cinnamyl cis Cinnamate cis Cinnamyl trans Cinnamate trans Cinnamyl cis Cinnamate trans Cinnamyl trans Cinnamate Other 16 components

1947 066 2088 2312 2403 2428 2560

Peru Resinoid is a viscous but mobile extract which captures the warmth and 10.1 almost food - like sweetness of the balsam. It is an excellent fixative in soaps, shampoos, hair conditioners and other toiletries, where it gives a substantive and appealing blend with most aromachemicals. Peru oil is distilled directly from the balsam at high vacuum using the minimum 10.2 of heating, rather than steam processing, which does not efficiently liberate the oil from this exudate. It is an amber coloured liquid with a true balsamic, rich vanilla sweet and slightly spicy note. Peru oil is one of the few ingredients which exhibits balsamic base, middle and topnote characteristics when incorporated into perfiunes and colognes. 10.3 Applications Fragrances: The odour of Peru is one of the longest lasting of all of the resinoids, it also shows very good stability in difficult products containing bleaches, high alkalinity and medicaments. Besides the obvious use in oriental and amber bases for fragrances, the Peru balsam note makes an important contribution to all of the long lasting sweet florals.

Flavours: There are some uses in the flavouring of tobacco products. Aromatherapy and Pharmaceuticals have virtually no uses for Peru balsam. 11

BENZOIN

Sumatra Benzoin is a natural balsamic resin exuded from the small tree Styrax benzoin, which grows extensively in Sumatra and Malaysia. ( A different, considerably more expensive grade from Siam is also used ). Incisions are made through the bark or after peeling off small areas of bark to expose the viscous balsam .The balsam flows and solidifies on exposure to the air. The gum is commercially available in many grades, from the best light coloured “almonds” to the dusty sweepings, which contains dirt, sand, wood splinters etc. The odour of Benzoin is best described as sweet, balsamic and vanilla like, with a slightly bitter taste. The characterising components are benzyl alcohol, the benzoresinyl benzoates and cinnamates together with some vanillin. These components do have a certain antioxidant effect in products, but a discoloration should be expected in formulations that contain alkalis, anthranilates, and indole.

112

Flavours and Fragrances

0-H 0

Benzyl Alcohol

Benzyl Benzoate

Vanillin

Table 11 Composition of the Volatile Fraction of Benzoin Resin Component Benzaldehyde Benzyl Alcohol Methyl Benzoate Ethyl Benzoate Benzoic Acid cis Cinnamic Alcohol trans Cinnamic Alcohol Methyl Cinnamate Vanillin Ethyl Cinnamate Cinnamic Acid Nerolidol b Nerolidol a Benzyl Benzoate trans Cinnamyl Benzoate Benzyl trans Cinnamate cis Cinnamyl trans Cinnamate Other 83 components * seePeru

HP 5 retention index 96 I 1035 1091 1170 (1200)* 1259 1300 1379 1391 1462 (1500)* 1534 1564 1762 2066 2088 2403

Area % 0.08 41.8 0.08 0.04 (2.5) 0.01 0.05

0.05 1.4 0.8 (2.7) 0.01 0.4 46.6 0.1 0.9 1.5 1.08 100.00

Benzoin Resinoid is a dark brown thick liquid which is very popular in soaps 11.1 and detergents because of its cost effectiveness. Benzoin Resinol is a lighter coloured ambient pourable extract, which is a 11.2 convenient form for compounders to add to fragrance compounds. The usual application of Benzoin is in soaps and detergents, where it gives a pronounced fixative effect and a lasting sweet balsamic note on the skin and on clothes that have been laundered. Benzoin Absolute is extracted from the gum with hot alcohol and is a solid 11.3 once the solvent is evaporated. Benzoin absolute has a good preservative effect on many formulations. It is not viable or appropriate commercially to make a distilled oil or molecular distilled Benzoin, as this natural is a low cost material.

113

The Aromatic Resins: Their Chemistry and Uses

11.4

Applications Fragrances: The relatively low cost and large scale production of the gum and its extracts, make it useful in soap and detergent fragrances. However, its best performance is in neutral and acidic products if the effect of the free benzoic acid is utilised in addition to the resinous notes. Flavours: Despite having a slightly bitter character, it can be used in flavours for its balsamic vanilla notes. Pharmaceuticals: Traditionally used in Friars balsam as a decongestant and respiratory disinfectant when inhaled in steam. 12

TOLU

The tall tree Myroxylon balsamum sp. is native to the jungles of the northern parts of South America, particularly Colombia and Venezuela. It is a solid brown mass which fractures like flint when a lump is hit or dropped. There is some confusion in the trade as to qualities of the balsam, probably due the several different species that are collected. Tolu odour is sweetly balsamic, faintly floral and spicy, with an undertone of vanillin [2, 191. It is similar to benzoin, but the odour comparison is less vanilla like and more cinnamic than benzoic in character. This is reflected in the composition of the characterising components, the cinnamates [26,27]. 0

0 I

/Pv-?-- 0

w

4

k2

I

0

v

Cinnamyl Cinnamate

-

\/

Benzyl Cinnamate

Table 12 Composition of the Volatile Fraction of Tolu Resin

Component Benzaldehyde Benzyl Alcohol Methyl Benzoate Ethyl Benzoate Benzoic Acid cis Cinnamic Alcohol trans Cinnamic Alcohol Methyl Cinnamate Vanillin Ethyl Cinnamate Cinnamic Acid Benzyl Benzoate trans Cinnamyl Benzoate Benzyl trans Cinnamate

HP 5 retention index 96 1 1035 1091 1170 (1200)* 1259 1300 1379 1391 1462 (1500)* 1762 2066 2088

Area % 0.09 43.4 0.06 0.03 (0.1 1) 0.0 1 0.05 0.02 0.16 1.o (1.43) 50.7 0.1 0.5

Flavours and Fragrances

114

cis Cinnamyl trans Cinnamate Other 44 components * see Peru

2407

--

1.5 0.84 100.00

12.1 Tolu Resinoid is a dark brown viscous resin with good fixative and preservative properties.

Tolu Resinol is a brown pourable liquid, much lighter and more mobile than the traditional dark brown resinoid that it supersedes. It is an excellent fixative, with the cinnamon like spicy floral character being useful in oriental and floriental fragrance types. 12.3

Applications Fragrances: Like benzoin, it has a lasting resinous note in soaps and detergents but shows its full strength in neutral and more acidic product applications. Flavours: Tolu is used in the flavouring of tobacco products. Pharmaceutical: Tolu is used as a decongestant in cough linctus and syrup formulations. 13

EXPERIMENTAL

The identification of the components listed in the tables is by GCMS using as Finnigan instrument fitted with a 30 metre DB 5 capillary column and an Ion trap detector. Calibrated and standardised capillary columns of HP 5, DB 1 and DB wax according to the methodology published in “The Analyst” (London) [ 13,28,29,30,3 11 , by the Royal Society of Chemistry Essential oils analysis sub-committee, in a Unicam 610 GLC instrument fitted with an FID detector, were used for the relative retention indices studies. The current authors are members of this RSC / EO committee, DAM has been so for the past 26 years, RAC has joined this year. 14

ACKNOWLEDGEMENTS

The authors would like to thank:- Nick Moss, Martin Buckby, Herbie Femandes, Neil Owen and Ray Lucas for their advice and valuable practical support. 15

CONCLUSIONS

The cost effective contribution to consumer products made by the extracts of the Aromatic resins has long been appreciated by their users. Their aroma and fixative properties make an important contribution to the perceived enjoyment of perfumes, cosmetics, toiletries, soaps, detergents, household, oral hygiene, pharmaceuticals, tobacco, savoury, beverage, bakery, confectionery and aromatherapy products.

115

The Aromaric Resins: Their Chemistry and Uses

16

REFERENCES

[ 1 ] H. Mans Boelens, D. de Rijke, H.G. Haring, Perfumer & Fluvorist, 6,7- 13, (198 1). [ 2 ] B. Hephrun, “Ancient Perfiunery“, Proc. Aroma Sym. , AANA I NORA paper 1, New York, U.S.A.(1994). [ 3 ] N. Groom.,”Frankincense and Myrrh, a study of the Arabian Incense trade”, pub. Longman, 280 pp, ISBN 0-582-76476-9, (1981). [ 4 ] J. Verghese, Perfumer & Fluvorist, U (I), 1-12, (1988). [ 5 ] D.A. Moyler, N.E. Moss, M. Buckby, Proc. 13th. ICEOFF, Istanbul, “Olibanum and Myrrh, the Golden oils”, (1995). [ 6 ] J.W.K. Burrell, R.A. Lucas, D.M. Michalkeiwicz, R. Riezebos, Chemistry & Industry, 1409-10 ( 1970). 3837-8, (1971). [ 7 ] idem. Tetrahedron letters, [ 8 ] B.A. Mc Andrew, D.M. Michalkeiwicz, Dev. FoodSci. vol18, pub. Elsevier, ed. Lawrence, Mookhejee, Willis. ISBN 0-444-42964-6,573-85, (1988). [ 9 ] H. Klein, H. Obermann, Proc. 7th. Cong. Ess.Oils, 400, (1979). [lo] G. Vernin et af.,Dev. Food Sci. vol24, pub. Elsevier, 51 1-42, ed. Charalambous, ISBN 0-444-88246-4 (1989). [I 13 B.M. Lawrence, The Essential Oils,1976-8, pub. Allured, 3 1-2, ISBN 0-93 1710-03-0. [121 P. Maupetit, Perfumer & Flavorist, 9 (6), 19, (1 984). [ 131 Analytical Methods Committee, The Analyst (London), 1089, (1 993). [141 P. Noble, Ph.D. thesis, Univ. Wurzburg, Germany, (1 980). [ 151 R.A. Wilson, B.D. Mookejee, Proc. 9th. Cong. Ess. Oils, 4,1 - 10, (1 983). [ 161 B.M. Lawrence,The Essential Oils, 1981-87,84-6, pub. Allured, ISBN 0-93 1710-17-0. [17] F. Delay, G. Ohloff, Helv. Chim. Acta., 62,369-77, (1979). [181 A. Maradufu, Phytochem., 2,677-80 (1 982). [191 S. Arctander, Perfume and Flavor Materials of Natural Origin, pub. Allured (1 962). [20] B.M. Lawrence, The Essential Oils,1981-87, pub. Allured, 24-5,93-6. ISBN 0-93 1710-17-0. [21] D.de Rijke, R.Ter Heide, H. Mans Boelens, Proc. 8th. IEOC, Cannes, France, paper 179, (1980). [22] G. Ohloff,Scent and Fragrances,pub. Springer - Verlag, 181-4, (1994). [23] K.Tajima, et al., Proc. 13th. ICEOFF, Istanbul, 2 17-24, (1995). [24] B.M. Lawrence, The Essentialoils, 1981-7, pub. Allured, 109-1 10. ISBN 0-93 1710-17-0. [25] D.L. Opdyke, Food Cosmetics Toxicology, la,95 1-3, (1974). [26] I. Wahlberg, M-B Hjelte, K. Karlsson, C.R. Enzell, Acta Chem. Scand., 2 , 3 2 8 5 (1971). [27] K.J.Harkiss, P.A. Linley, Planta Med., 15,61, (1979). [28] Analytical Methods Committee, The Analyst (London), 262-73 (1 980). [29] Analytical Methods Committee, The Analyst (London), U,448-55, (198 1). [30] Analytical Methods Committee, The Analyst (London), 1339-60, (1984). [31] Analytical Methods Committee, The Analyst (London), 1125-35, (1988).

s,

m,

m, me, m,

Studies towards Structure Determination of Substituted Pyrazines

’,

Michael Zviely Alexander Kern’, Igal Gozlan2 and Ron Frim2

’ FRUTAROM LTD, HAIFA. ISRAEL

* TAM1 LTD, HAIFA, ISRAEL

Introduction Pyrazines are materials obtained in Maillard reactions 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 synthesize 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 modem analytical instruments, the exact structures were not defined. For example, a mixture of isomers of methyl-methoxypyrazines was defined in an inexact manner, namely 2-methoxy-3(5 or 6)-met hylpyrazine. The problem of the exact structure of several commercially significant pyrazines arises while 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 - m e t h y l - & ~ ~ r 6-methoxy-pyrazine and 2-methyl-&~~h-n-propoxy-pyrazine. Let us look at several examples taken fiom commercial catalogues. In one, the item appears as: “2-Methoxy-3(5/6)-methylpyrazine”

The CAS (Chemical Abstracts Service) number mentioned is [68378-13-21, which represents only the following possibilities:

Studies towards Structure Determination of Substituted Pyrazines

117

Namely, 2-methoxy-3(or 5)-methylpyrazine In another company’s catalogue, a similar product appears as: “Mixture of 2-Methoxy-3-methylpyrazineand 2-Methoxy-5-methylpyazinealso named 2 or 5 or 6-Methoxy-3-methylpyrazine (Mixture of Isomers)”

-75-85%

-1 525%

negligible%

The declared product in the catalogue consists of 2-rnethoxy-3-methylpyazine (-75-85%) and 2-methoxy-5-methylpyazine(-1 5-25%) and 2-methoxy-6-methylpyrazine (negligible %). Another example taken from a catalogue is: “2-sec-Butyl-3(5/6)-methoxypyrazine”

The cited CAS number is [24168-70-51,which belongs only to the 2,3- isomer:

I The last illustration from a series of cases is a product referred to as: “Mixture of 2-Methylthio-3methylpyrazine and 2-Methylthio-5-methylpyrazine”

118

Flavours and Fragrances

and declared as the following mixture: 2-Methylthio-3-methylpyrazine (-80%) arid 2-Methylthio-5-rnethylpyrazine (-20%) urld 2-Methylthio-6-methylpyrazine (negl.%). The CAS number named is [59021-03-31, which belongs to (Methylthio) methylpyrazine:

and a similar mixture has a CAS 2-methyl-3(or 5)-(methyIthio)pyrazine:

number

[68378-12-11.

which

is

As demonstrated by these cases, there is confbsion in outlining the exact structure o f these and other pyrazines, and the intention of this presentation is to try and determine the exact location of the substituents. Several works concerning the structure of substituted pyrazines by means of GC were published in the literature,' Other published works describe the structural elucidation of substituted pyrazines using 60 MHZ 'H-NMR.' Based on our observations, "H-NMR, especially a weak field machine of 1.4 Tesla (v 60 M H z ) is not sufficiently unambiguous in defining the exact structure of these molecules. The following 'H-NMR simulated spectra taken in v 400 MHz machine exemplify the problem;

'H-NMR Simulated Spectra of 5- and 6-Methvl-2-methoxv~yrazines

* Spcclra done by Advanced Chemistry Dcvelopment Inc -IHNMR SomVarc

119

Studies towards Structure Determination of Substituted Pyrazines

In a work published by Lutz gtd,3they claimed that the product of the direct chlorination of 2-methylpyrazine was shown to be a mixture of 2-chloro-3-methylpyine and 2-chloro-6-methylpyrazine. This work describes the synthesis of 2-chloro-6-methylpyrazine from 2-hydroxy-6-methylpyrazineusing phosphorus oxychloride:

2-Hydroxy-6-methylpyazineitself is prepared in this work by a condensation of pyruvaldehyde bisulfite adduct and glycineamide hydrochloride. The structure of the molecule is proved by several methods, i.e. IR, W. 'The method which was used by us to elucidate the structures of these rather simple but puzzling molecules was the application of nitrogen NMR spectroscopy, more specifically '%-NMR spectroscopy.

In general, nitrogen NMR spectroscopy has been well described in review literature4 to show the importance of nitrogen chemistry and the potentialities and difficulties of NMR work with the two magnetic nuclei, 14N and '%. Both have rather low magnetogyric ratios (y), and so rather low NMR sensitivity, of about one-tenth that of 13C. '% has spin Vi but low natural abundance (0.365%) and negative NOE factors (y being negative), and may be slow to relax. 14N is nearly 100% abundant but is quadropolar (I = l), though with a relatively small quadrupole moment (Q = 0 . 0 1 7 ~ 1 mZ) 0 ~ ~so that broadening may not be excessive, but nevertheless can complicate the structure elucidation. Nuclear ProDerties of Hvdrogen. Carbon and Nitrogen Isotopes NUCLEUS

SPIN NUMBER

:H

:H

lg)C

;' N

YN

%

1

%

!h

3 0.015% 0.411

2 1.108% 0.673

1 3 99.635% 0.193

2 0.365% -0.271

0.009

0.016

0.001

0.001

125.7

36.1

-50.556

No. OF EIGENSTATES 2 NAnmAL ABUNDANCE 99.985% GYROMAGNETIC 2.675 RATIO y 10' (T's") RELATIVE SENSITIVITY 1 FREQUENCY (RF) v m 500 (at 1 1.744Tesla)

Several procedures were used in elucidating the structures, especially Insensitive Nuclei Enhancement by Polarization Transfer (INEPT) and Heteronuclear Correlated 2 0 M R experiments. The apparatus used had a magnetic field of 11.744 T, v -50.556 MHz. The solvent was &-DMSO, and the standard was CH3I5N0z.

I20

Flavours and Fragrances

Results and Discussion The first sample was constituted mainly of 2-chloro-3-methylpyrazine,accompanied by another isomer (5- or 6- methyl-) as a minor impurity: 1

1

1

g"XC' y CH3

2-Chloro-5- or 6- methylpyrazine

2-Chloro-3-methylpyrazine

The predominate molecule is 2-chloro-3-methylpyrazine,in accordance to the following spectral data as seen by the INEPT spectrum:

INEPT Spectrum of 2-Chloro-3-methvlpvrazine(maior) and Isomer (minor) t

The signal at 6 -19.07 ppm; d, Jz 12.2 Hz was attributed the nitrogen atom in position 4 and the signal at 6 -35.5 ppm; d, J2 12.2 Hz was attributed to the nitrogen atom in position 1 . These coupling constants are suitable for 2 bonds distance: 1

The minor signal data at 6 -36.95 ppm; J3 4.4 Hz is attributed to NI; the coupling constant value fits for the 3 bonds distance of the following structure:

Studies towards Structure Defermination of Substituted Pyrazines

121

’The other signal, at 6 -16.6 ppm; t, Jz 7.7 Hz, which belongs probably to N4 has a J magnitude which fits the 2 bonds distance of the following structure: I

“INf H

Y

H

This molecule will be discussed later, in the sample which contains the pure isomer. The assignments of the nitrogen atoms positions were done according to the 2D N-H correlation spectrum as following: The nitrogen signal at & -19.07 ppm relates to the signal of the aromatic hydrogen at position 5 with SH 8.4 ppm, and to the methyl group at position 3 with 6~ 2.5 ppm; so this nitrogen is at position 4; The nitrogen signal at & -35.5ppm relates & to the signal of the aromatic hydrogen at position 6 with SH 8.2 ppm, so this signal belongs to the nitrogen at position 1. The minor signal data in the 2D spectrum are as following: The signal at 6~ -36.95 ppm which relates&o to the methyl group in 6~ 2.4 ppm, thus belonging to the nitrogen atom at position 1, and the signal at 6~ -16.6 ppm which relates to the aromatic hydrogen atoms 3 and 5, with a chemical shift of SH 8.4 ppm

122

Flavours and Fragrances

2D N-H Correlation Spectrum 2-Chloro-3-methvlpvrazine(ma-ior)and Isomer (minor)

123

Studies towards Structure Determination of Substituted Pyrazines

The second sample contains the isomer which is the topic of this work, namely 2-chloro-5- methylpyrazine or 2-chloro-6-methylpyrazine. 1

1

2-Chloro-5- or 6- methylpyrazine

Following is presented the INEPT spectrum of this molecule:

INEPT SDectrum of 2-Chloro-5- or 6-methvl- pvrazine

r7

I

i

Ji, -15

1

2-Chlorc-5- or E rnethylpyrazine

11,, -20

,

,

,

,

,

,

.

-2s

,

, -30

,

,

+,

--.- _-

The INEPT spectrum shows the following signals: At 6 -37.3 ppm; J p 6.97/2 Hz structure:

= 3.49 Hz a signal which is attributed to NI in

the following

1

The coupling constant fits for 3 bonds distance. The signal at 6 -16.6 ppm; dd. Jz 6.95 Hz, fits to N-4. It is split into a doublet of doublets by the hydrogen atoms on C-3 and (2-5. This structure is supported by the 2D N-H correlation spectrum, in which nitrogen atom no. 4 (at 6 -16.6 ppm) is related to the aromatic hydrogen atoms on C-3 and C-5 (at 6~ 8.4 ppm), and nitrogen atom no. 1 (at 6~ -37.3 ppm) is related to the methyl hydrogen atoms (at 6" 2.4 ppm).

124

Flavours and Fragrances

2D N-H Correlation SDectrum 2-Chloro-5- or 6-methvl- Dvrazine

I l l -, , E

, , r

n

,,,,,I,, J .

Ln

, ,

,,,,,

LD

, , r.

,!-0

m

nLI-. 3 L

I

125

Studies towards Structure Determination of Substituted Pyrazines

If the correct stucture was 2-chloro-5-methylpyrazine,the 2D N-H correlation spectrum would have to show a relation between nitrogen no. 4, the hydrogen atom on '2-3, and the methyl group at C-5. N-1 should relate only to the aromatic hydrogen atom no. 6 as shown here:

But no relation exists in the 2D N-H correlation spectrum, between N4 and H3. The third sample to be analyzed was 2-methoxy-6-methylpyrazine (99% pure by GC).The lNEPT spectrum shows 2 signals, one at 6 -37.65 ppm; d (J = 19.5 Hz), attributed to nitrogen atom no. 4, and the other at 6 -100.54 ppm; d (I = 4.9 Hz), attributed to the nitrogen atom at position 1 . INEPT Spectrum of 2-Methoxv-6-rneth~l~wazine

--300

7 200

100

0

-1 0

Y

-200

The 2D N-H correlation spectrum of the sample shows a relation between the nitrogen absorption at 6~ -37.65 ppm and 2 aromatic hydrogens at & 7.95and 8.05 ppm and no methyl group. The second relation is between the nitrogen at & -100.54 ppm and the methyl group hydrogens at & 2.4ppm.

Flavours and Fragrances

I26

2D N-H Correlation SDectrum of 2-Methow-6-methylpvrazine

0

\

P

0

Studies towards Structure Determination of Substituted Pyrazines

I27

These relations are consistent with 2-methoxy-6-methylpyrazine.The structure of 2methoxy-5-methylpyrazine would require a relation between a nitrogen, one aromatic hydrogen, one methyl group, and an additional relation between the other nitrogen and one aromatic hydrogen. These relations do not exist in the N-H correlation spectrum, therefore the correct structure is 2-methoxy-6-methylpyrazine.

The fourth sample contains 2-methyl-3-propoxypyrazine (9Ph purity by GC): 1

INEPT SDectrum of 2-Methyl-3-oroDoxvpyrazine PI

c N

!

i

i -3c

?s:

-40

-60

-50

The DEPT spectrum shows the following signals:

-70

-EC

-s; I

I

For nitrogen atom no.1 at 6 -28.45 ppm; JZ 10.4 Hz. This coupling constant fits to the 2 bonds distance as shown in the following drawing: 1

For N-4 at 6 -90.8 ppm; Jz 6.97 Hz,the coupling constant fits for the 2 bonds distance as shown in the following drawing: 1

128

Flavours and Fragrances

The 2D N-H correlation spectrum show relations between N-l (at 6~ -28.45 ppm) and the methyl group on C-2, and between the aromatic hydrogen atom on C-6 (at I ~ H-8 ppm). The other correlation seen in this 2D spectrum is between nitrogen atom no. 4 (at 6~ -90.8 ppm) and the aromatic hydrogen on C-5 (at 6~ -8 ppm) only. 2D N-H Correlation of 2-Methvl-3-pro~oxvp~azine

Studies towards Structure Determination of Substituted Pyrazines

129

'The next sample to be studied was of 2-methyl-6-propoxypyrazine (99% purity by W):

4

'The INEPT spectrum of the sample is shown here: INEPT SDectrum of 2-Methyl-6-propomwazine

The INEPT spectrum shows the following signals:

For N-1 at 6 -99.2 ppm;

J3

3.5

Hz;and for N-4 at 6 -37.4 ppm; dd, Jz 14.0/2 = 7

Hz; The 2D N-H correlation spectrum shows relation between N-1 (at & -99.2 ppm) and the methyl group on C-2 (at SH 2.2 ppm), and a correlation between N-4 (at 6~ -37.4 ppm) and the aromatic hydrogen atoms on C-3 and C-5 (at & -8 ppm).

1

4

130

Flavours and Fragrances

2D N-H Correlation Spectrum of 2-Methvl-6-propoxypyazine

131

Studies towards Structure Determination of Substituted Pyrazines

After establishing these methods to elucidate problematic pyrazine structures, another problem was solved, namely the quantitative composition of 2-acetyl-3,6-dimethyIpyrazine and 2-acetyl-3,5-dimethylpyrazine, which was not clear until now. The GC shows 2 signals with 70:30 area, but it is not known which peak belongs to which molecule - 3,Sdimethyl- and 3.6-dimethyl-. Thus, a mixture consisting of 2-acetyl-3,6-dimethylpyrazine and 2-acetyl-3,5-dimethylpyrazine was analyzed using the above techniques, in order to determine the molecules composition in the mixture. 0

0

In the following INEPT spectrum four different peaks are seen, representing the two pairs of nitrogen atoms belonging to the two molecules: INEPT Spectrum of the Mixture

The signals at 6 343.38 ppm and 332.52 ppm, are attributed to the nitrogen atoms 1 and 4 respectively of 2-acetyl-3,6-dimethyIpyrazine.The signals at 6 339.99 ppm and 335.95 ppm, correspond to the nitrogen atoms 1 and 4 respectively of 2-acetyl-3,Sdimethylpyrazine. The precise structure confirmation is shown by the heteronuclear correlated 2D NMR spectrum of the mixture, as presented here;

132

Flavours and Fragrances

Heteronuclear Correlated 2D NMR Spectrum of the Mixture

Studies towards Structure Determination of Substituted Pyrazines

133

The assignments of the nitrogen atoms positions were done according to the 2D N-H correlation spectrum as following: The nitrogen signal at SN 343.41 ppm relates to the signal of the aromatic hydrogen at position 5 with 6~ 8.45 ppm, and to the methyl group at position 3 with & 2.6 ppm; so this nitrogen is at position 4; The nitrogen signal at 6~ 332.61 ppm relates&o to the signal of the methyl group at position 6 with &I 2.5 ppm, so this signal belongs to the nitrogen at position 1. These results are only consistent with 2-acetyl-3,6-dimethylpyrazine: 0

2-AcetyI-3,Wimethylpyrazine

The nitrogen signal at 8~ 339.99 ppm relates only to the signals the two methyl groups at positions 3 and 5 with 6~ 2.5, 2.6 ppm; so this nitrogen is at position 4. The nitrogen signal at SN 335.99 ppm relates only to the signal of the aromatic hydrogen at position 6 with 6~ 8.3 ppm, so this signal belongs to the nitrogen at position 1. These results are only consistent with 2-acetyl-3,5-dimethyIpyrazine: 0

2-Acetyl-3.5-dimethylpyrazine

Lastly, the area of the two aromatic hydrogen signals at 6" 8.3 and 8.45 ppm is in a 1:2 ratio, so the major isomer is 2-acetyl-3,6-dimethylpyrazineand the minor isomer is 2-acetyl-3.5- dmethylpyrazine.

To summarize this work, the following points were concluded: The exact structure of several commercially important pyrazines was determined, thus revealing the mystery of these molecules; 0

0

The second group of the bi-substituted pyrazines is at position 6 and not 5; An easy method to decide upon the exact structure of substituted pyrazine molecules, i.e. 15-Nitrogen NMR spectroscopy was established;

I34

Flovours and Fragrances

Acknowledgments

We thank Bella Lapid for locating all the information from the Chemical Abstracts Service, and Ella Dagan for operating the NMR machine. A special gratitude to Janet Elion for helping in the rewriting of this article.

References 1. G.M. Nakel and L.V. Haynes,J.Agr.

Food Chem. 20,682, 1972; D.T. Stanton and P.C. Jurs, Anal. Chem. 61, 1328, 1989; S. Mihara and H. Masuda, 1. Chromatogr, 402, 309, 1987. 2. A.F. Bramwell, G. Riezebos and R.D. Wells, Tetrahedron Letters 2489, 1971; A.F. Bramwell and R.D. Wells, Tetrahedron 28, 4155, 1972. 3. W.B. Lutz, S. Lazarus, S. Klutchko and R.l.Meltzer, J. Om.Chem. 29,415, 1964; 4. W. von Philipsborn and R. Miller,&gew. Chem. Int. Ed. Engl, 25, 383, 1986. 5 . File ‘Registry’ at 13 Mar 1997, American Chemical Society (ACS).

Flavours

Generation of Taste through (Redox) Biocatalysis Colja Laane*, Ivonne Rietjens, Huub Haaker and Willem van Berkel DEPARTMENT OF BIOCHEMISTRY, DREUENLAAN 3.6703 HA WAGENINGEN, THE

NETHERLANDS

INTRODUCTION In nature, enzymes play a key role in the generation of taste. A wide variety of enzymes are believed to be involved in the natural production of all the thousands of sensoric molecules known. to-date. In most cases several enzymatic steps, sometimes in combination with chemical reactions, are required to produce just a single compound. Despite common metabolic pathways, the biosynthesis of each sensoric molecule seems to require its own set of unique enzymes and many such metabolic routes have to run parallel to arrive at a given aroma profile. With this in mind it is not surprising that the metabolic puzzle of flavour biogenesis is extremely complex and far from being elucidated, and that little is known about the nature of the hundreds of different enzymes involved. Typically, enzymes are divided into the following classes: hydrolases, isomerases, lyases, transferases, ligases and oxidoreductases. Each class performs a range of essentially similar type of reactions and is further subdivided into more specific enzymes catalysing in general only one type of conversion of a limited number of substrates. Several approaches have been followed to use enzymes, either in an isolated form, or present in their natural habitat within a microorganism or plant cell, in the production of flavouring preparations (1, Figure 1).

*To whom correspondence should be addressed. Tel: +3 1.3 17.482868 Fax: +3 1.3 17.484801 E-mail: Colja.Laane (9alg.bc. w a u l

138

Flavours and Fragrances

SC'BSTRATE(S)

v

Enzyme

--

BIOCATALYST Microorganism

Plant cell

I

I

J

Figure 1 Biotechnologicnl upproachcs to the yrudicctiori uf naturalfkuvours The general approach involves the biocatalytic conversion of either single substrates or mixtures of compounds by these biocatalysts to one specific top note, a mixture of sensoric compounds (e.g. building block), or a multifunctional ingredient which delivers, in addition to taste, other attributes to the final consumer product. In the case of top notes single substrates (precursors) are commonly applied, while for building blocks and multifunctional ingredients, a tailored substrate mixture is required. One of the prime reasons to explore biocatalytic routes for the production of flavours is the consumer preference for all-natural food ingredients. In contrast to chemical synthesis, bioconversions are in general more specific and generate only those isomers which are found in nature. Furthermore, biocatalysts are able to operate under very mild conditions and possible waste products are inherently biodegradable. Another important factor which has sparked the use of biocatalysts for the production of bio-flavours is the enormous technical advances made since the 80's to overproduce, design and redirect metabolic fluxes at will, by genetic, protein, and metabolic pathway engineering. For each class of enzymes, in particular for oxidoreductases, the current advances and scope of these technologies in tailoring the taste of food will be highlighted and discussed.

HY DROLASES By far the best studied class of enzymes are the hydrolases. They include enzymes which act on ester (lipases, esterases), ether (ether hydrolases), glycosyl (carbohydrases), peptide, or amide bonds (proteases, peptidases) by splitting C-0 and C-N bonds in the presence of water. The same enzymes can also be used to form these bonds in media containing little water. In addition, less well characterised hydrolases exist for halide, P-C, P-N, and S-N bonds.

I39

Generation of Taste through (Redox)Biocatalysis

Of all the enzymes used todate in commercial applications, more than 90% belong to the class of hydrolases. Typical bulk applications include detergent formulations, food & feed, as well as the treatment of pulp and paper. In food systems hydrolases are supplemented, in addition to the often already present endogenous hydrolases, to further control taste, texture, appearance, nutritional value, and processing tolerance of the food product. Table 1 lists a few examples of hydrolases, which are specifically applied to generate taste, either in situ within a given food product, or in vitro to produce a topnote or building block. Type of hydrolase

Purpose

Application example

Ref.

Lipases

Release fatty acids/alcohols (Trans)esterification Chiral resolution

Building block (e.g. cheese) Topnotes (e.g. tactones) Topnotes (e.g. I-menthol)

2 3 43

Formation peptidedamin0 acids Peptide synthesis

Savoury building block

6

Aspartame

Release of glycosylically bound flavours Release of sugars

Beverages (e.g. wine, fruit juice) Mail I ard products

6

Liberation of S-GMP/S-IMP

Building block (e.g. taste enhancer)

7

Proteases and peptidases Carbohydrases

~~

Esterases

As far as we know only C-0 and C-N splitting/synthesisinghydrolases are being applied for flavouring purposes, although interesting applications can be envisaged for each of the other types of hydrolases.

Up until recently all flavour generating hydrolases were designed for other purposes. By trial and error some of them proved to be suited for the in situ or in vitro production of flavouring preparations. This situation is bound to change since the recent introduction of FLAVOURZYME from NOVO (8). This hydrolase mixture consists essentially of proteases and peptidases and is specifically created to breakdown proteins extensively into amino-acid based savoury building blocks (e.g. HVP-replacers). It can be expected that, due to the rapid progress in genetic engineering, more hydrolases will become available in the near future which are specifically designed for flavouring purposes.

ISOMERASES Relatively little is known about this class of enzymes. In general they catalyse geometric or structural changes within one molecule. According to the type of isomerism they may be called racemases, epimerases, (cis-trans) isomerases, and tautomerases.

Type of isomerase

Purpose

Application example

Ref.

Glucose isomerases Interconversion of glucose into fructose

Sweetener

9

Protein disulphideisomerases

Cross-linking proteins

Texturiser which may affect mouthfeel/taste

10

Double-bond isomerases

Changing the double bond position

Topnotes (e.g.(iso)eugenol)

Cis-trans isomerases

Cis-trans isomerisation

Topnotes

An interesting example could be a double bond isomerase which is able to convert eugenol into isoeugenol. Eugenol is readily available from clover, but cannot be converted directly by lipoxygenase into vanillin, while isoeugenol is hardly available but readily reacts with lipoxygenase to form vanillin (see oxoreductases). That an eugenoMsoeugenol converting isomerase may exist in nature arises from the observation that in certain Lemon grass oils both compounds are present.

LYASES Lyases catalyse the reversible cleavage of C-C, C-0, C-N and other bonds by elimination to produce double bonds, or vice versa, catalyse the addition of groups to double bonds (1 1). The type of reaction catalysed can be decarboxylation (decarboxylases), dehydration (dehydrases), deamination (deaminases), cyanohydrin cleavage (oxynitrilases), aldol cleavage (aldolases), and a, 0, or y elimination (PLP-, pyridoxal phosphate-containing, lyases). In cases where the reverse reaction is the more important, or the only one to be demonstrated, 'synthetase' may be used in the name. Lyases already have a relatively long history of safe use in various food systems. Wellknown examples include lyases which hydrolyse polysaccharides (12) such as pectin (pectin lyases), alginate (alginate lyase ) and xanthan gums (gelrite lyase). They are mainly added to reduce viscosity in food products, or for the production of specific oligomers with interesting chelating and physiological properties. These lyases are, however, not intentionally used for flavouring purposes.

Generation of Taste through (Redox) Biocaralysis

141

Table 3 shows some type of lyases which are targetted specifically for flavouring applications.

Type of lyase

~~

C-S lyases

C - 0 lyases

C-C lyases (splitting)

C-C lyases (forming)

Application example

Purpose

-

~

Preparation of thiol flavours: -p-mentha-8-thiol-3-one -furfury1 mercaptane -methane thiol from methionine -dially1thiosulphonate -thiopropionic acid derivatives

Topnotes (e. g.coffee, cheese, garlic, beverages and roasted beef aroma)

Dehydration: -debittering of citrus limonoids Hydration: -a-terpineol from limonene -step in methylketone formation Decarboxylation: -4-vinyl phenols from cinnamates -acetoin from a-acetolactate -step in 2-phenylethanol synthesis Removal of CN-group: -benzaldehyde Splitting hydroperoxyfatty acids: -step in (Z3)-hexenol synthesis Aldol condensation: -step in furaneol synthesis Deamination: -step in 2-phenylethanol synthesis

13 14 15 16

Beverages

17

Topnote beverages Blue Cheese

19

Phenolic building block

20

18

Beer off-flavour removal 21 Topnote beverages

19

Topnote beverages

19

'Green leave' topnote

22

Topnote strawberry

23 -

Topnote beverages

19 -

~

C-N lyases

Ref.

A new and interesting development is the microbial production of a whole range of sulphur top notes using dimethylsulphonium propionic acid (DMSP) from algae as a precursor (16). DMSP is an easily extractable osmoregulator in various algae (e.g. Ufva species) and is present in concentrations of 0 3 2 % of dry weight. Figure 2 depicts the type of conversions which can be achieved.

142

Flavours and Fragrances

CH, CH,-S-CH,-CH,-COO

r

c

H

-

Desulfobacteria

I - - - CH,-S-CH,-CH,-COOH I

T

(CH,),S

CH,SH

AL~LESTERS

Methanobacteria HS-CH,-CH,-COOH

______+

ALKYLESTERS

Figure 2 DMSP-derived sulphur top notes. DMSP, dimethylsulphonium propionic acid; MTPA, methyl thiopropionic acid; MPA, methyl propionic acid; DMS,dimethyl sulphide. In bold are the characier impact compounds.

It was shown that in a single step DMSP can be demethylated by several Desulfbbacteria to methylthiopropionic acid (MTPA). In turn, MTPA can be further demethylated to mercaptopropionic acid (MPA) by selected Methanobacteria. In both cases the substrates serve as the sole carbon and energy source and complete conversions are achieved. The desired natural sulphur flavours can subsequently be obtained by mild esteri fication of the DSMP derivatives. In addition microbes have been isolated which are able to generate dimethyl sulphide (DMS) directly from DMSP. The enzyme involved has been called DMSP-lyase and is partially characterised (24). It remains to be established whether there are micro-organisms which are capable of producing methanethiol directly from MTPA. A nice example of the participation of two lyases in the production of a flavour is the conversion of I-phenylalanine into 2-phenylethanol by Saccharomyces cerevisiae (1 9). The first step involves a deaminase yielding phenylpyruvic acid, which is subsequently converted by a decarboxylase to generate phenylacetaldehyde. In turn, phenylacetaldehyde is reduced to the desired beverage topnote 2-phenylethanol by an oxidoreductase.

TRANSFERASES Transferases are enzymes transferring a group, for example, the methyl group or a glycosyl group, from one compound (generally regarded as donor) to another compound (generally regarded as acceptor). In many cases, the donor is a cofactor, carrying the group to be transferred. On paper many transferases should be involved in the in vivo biosynthesis of flavours. A typical example includes the formation of isoamyl butyrate, a fruity, banana like character

Generation of Taste through (Redox) Biocatalysis

143

impact flavour. In nature this compound is probably formed from butanoyl-CoA and isoamyl alcohol with a specific acyl CoA transferase (Figure 3). For the in vitro production of bioesters the use of transferases seems to be to complicated and expensive due to the requirement of cofactors. In those cases lipases and esterases in their synthetic mode are more appropriate (see hydrolases).

Figure 3 Possible transferase-catalysed route to fruity bioesters Transglycosylases can also be applied to the production of flavours, especially as a tool to protect sensitive groups selectivily during a multi-step chemohiosynthetic procedure. Such an approach could be applied, for example, in the synthesis of flavouring phenolic compounds to prevent polymerisation during one of the steps (25).

LIGASES Ligases are enzymes catalysing the joining of two molecules with concomitant hydrolysis of the pyrophosphate bond in ATP or a similar triphosphate. In nature these enzymes must be involved in the biosynthesis of flavours. For example, the formation of butanoyl-CoA, the precursor for isoamyl butyrate (see figure 3) is catalysed by a ligase. For the in vitro production of flavours however, these enzymes seem less suitable due to the requirement of an expensive cofactor (e.g. ATP). Although sophisticated ATP-regenerating systems have been designed, it is not likely that isolated ligases will be applied in the near future for flavouring purposes.

OXIDOREDUCTASES To this class belong all enzymes catalysing oxidoreductions. The substrate oxidised is regarded as hydrogen or electron dosw. The classification is based on 'donor:acceptor oxidoreductase'. The recommended name is 'dehydrogenase', wherever this is possible; as an alternative 'acceptor reductase' can be used. 'Oxidase' is used only where 0 2 is an acceptor. Classification is difficult in some cases, because of the lack of specificity towards the acceptor. Oxidoreductases, also known as redoxenzymes, are heavily involved in flavour generation. Not only in the formation of desirable flavours, but also in the generation of off-flavours. Interest in applying redoxenzymes for flavouring purposes is clearly growing. Especially the use of oxidases, which do not require an expensive cofactor, has made much progress in recent years. It can, however, be expected that reductases will find a home in flavour research as well.

144

Flavours and Fragrances

For the sake of clarity oxidases and reductases will be treated seperately. In table 4 examples are listed of oxidases which are specifically applied for flavouring purposes. Type of oxidase

Purpose

Phenol oxidases

Enzymatic browning

Application example

Ref.

Fruits, beverages

26

Hydroperoxidation: -step in aldehyddalcohol synthesis -step in Gdecalactone synthesis -aromatic aldehydes (e.g.vanillin)

Various topnotes Topnote (e.g.butter flavour) Beverages, deserts, etc.

22 19 27

Glucose oxidases

Acidification

Cheeses

Sulphydryl oxidases

Thiol off-flavour removal

Beverages (e.g. UHT-milk)

Alcohol oxidases

Production of acetaldehyde or long-chain aldehydes

Topnotes beverages, fruits

Aldehyde oxidases

Off-flavour scavenging

Soybean extracts

l i 0

Vanillyl alcohol oxidase

Phenolic derived flavours

Phenolic topnotes and building blocks

I ::

Peroxidases

Specific oxidised flavours Demethylation of amines

Topnote (e.g. nootkatone) Topnote

~

Lipoxygenases

I I 1

28 28 36

7 29

(e.g.methylanthranilate)

Phenol oxidases belong to the so-called monooxygenases, since they catalyse the incorporation of one atom of oxygen into the substrate. Phenol oxidases are present in all fruit and vegetable products and could affect the flavour by the well-known browning reaction (26). The enzymatic browning reaction involves the hydroxylation of plant phenolic compounds into the dihydroxy derivative, which is subsequently oxidised by the same enzyme to the quinone form. In a non-enzymatic reaction these highly reactive quinones react with, among others, amino acids to generate coloured pigments. In this way tasty phenolic compounds are probably lost and at the same time flavouring reactive intermediates could be generated. Lipoxygenases are dioxygenases and are well-known ‘off flavour generators in nature. Their potential in the in vitro production of natural flavours has long been recognised. Typical examples include the conversion of polyunsaturated fatty acids into various short to medium chain aldehydedalcohols (22). or into S(-)-6-decalactone (19, butter flavour). Well-known fatty-acid derived flavouring aldehydes/alcohols include (E2)-pentenal (green apple), (E2)-hexenal (leaf aldehyde), (Z3)-hexenol (leaf alcohol), (E2,E6)-nonadienal (cucumber), I-octen-3-one (field mushroom), (1 ,Z5)-octadien-3-one (geranium leaves), ( 1,E3,ES)-undecatriene (balsamic), and ( 1 ,E3,ZS,Z8)-undecatetraene (seaweed). In all cases the first step involves the hydroperoxidation of a specific polyunsaturated fatty acid

Generation of Taste through (Redox)Biocaralysis

145

(1). Depending on the degree of unsaturation and the regioselectivity of the lipoxygenase different hydroperoxy compounds are formed, from which the above mentioned flavours can be derived by subsequent enzymic reactions. For (Z3)-hexenol,linolenic acids are used as a substrate, while for most of the other aldehydedalcohols higher unsaturated fatty acids are required. For the production of the chiral lactone, linoleic acid is the fatty acid of choice. The production of (Z3)-hexenol has recently been commercialised using plant homogenates (e.g. alfalfa sprouts, green peppers) which are relatively rich in hydroperoxide lyase; the second enzyme required to split the hydroperoxide fatty acid into smaller fragments. Finally, (Z3)-hexenol was obtained using the reductive enzymatic power of baker's yeast (35). Very recently, Givaudan-Roure has cloned and overexpressed the hydroperoxide lyase from banana in bakers yeast and has developed a 'single step' process which produces (Z3)-hexenol in a relatively high yield (see relevant Chapter in this book).

Following the same principles the other aldehydeslalcohols can be produced. For the production of the Clo-lactone a different strategy has to be followed after the formation of the hydroperoxide. It involves the fermentative P-oxidation of the hydroperoxide intermediate by the yeast Pichia etchellsii, and the subsequent cyclisation of 5hydroxydecanoic acid to the corresponding S(-)-&decalactone. In figure 4 the routes to (Z3)-hexenol and S(-)-&decalactone are depicted.

COOH

Linolenic acid Lipoxygenase

-

'OH

COOH

Linoleic acid

~

Lipox ygenase

COOH

1

COOH

-

Hydroperoxide lyase

'OH

-OH

1

Pichia @-oxidation)

nCHO H*

1

Reductase

a.Ikyz:Eii

temperature

nCHPoH 0

(Z3)Hexenol

S(-)-&Decalactone

Figure 4 Two examples of the use of lipoxygenase in the production of fatty-acid derived flavours

As has been shown by Quest International (27) lipoxygenases also accept a relatively broad spectrum of phenolic compounds as substrates. Of interest to the flavour industry are the lipoxygenase-catalysed conversions of isoeugenol and coniferyl benzoate from Siam resin to vanillin (4-15% yield, Figure 5). At present the commercialisation of these biotransformations is hampered by the fact that isoeugenol is not readily available and that coniferyl benzoate is difficult to handle in a reactor.

146

HO

Flavours and Fragrances

9

fi =fi - CH3

H3C0

H

H

\O / P c=c-c-0-c H H H

H3C0

Lipoxygenase

Isoeugenol

'0

H,CO

Vanillin

\ /

Conifery lbenzoate

Figure 5 Lipoxygenase-catalysed production of vanillin Glucose oxidase is mainly added to food systems to remove traces of glucose (e.g. treatment of egg white prior to dehydration), oxygen (e.g. prevention of discolouration andor oxidationj, or to produce hydrogen peroxide in situ for textural or antimicrobial purposes (28). Another application could be the in situ production of a (weak) acid in Mozzarella and cottage cheeses to enhance its typical soft acid flavour. Sulphydryl oxidases are capable of oxidising sulphydryl groups to disulphides in the presence of oxygen. Industrially, the potential of this enzyme lies in its ability to ameliorate the cooked sulphury off-note which is formed during the UHT treatment of milk (28), and in texturising food systems by rearrangement of the disulphide linkages between proteins (36). The former application has been demonstrated on a pilot-scale using immobilised enzyme columns (28), while the latter application remains to be proven. Sulphydryl oxidase from Aspergillus is one of those enzymes which has recently become available at sufficient quantities by genetic modification (36), and this will undoubtly facilitate the quest for (novel) commercial applications. In general aldehydes are more potent flavours than their alcoholic counterpart. Hence alcohol oxidases are interesting enzymes for the in vitro production of flavouring preparations. Typical examples include the use of methanol oxidase from Pichia, Hansenula, and Cundida (7) for the production of natural acetaldehyde from ethanol. This enzyme is being induced during growth on methanol. At the end of the logarithmic growth phase cells are harvested and incubated with ethanol. In this way concentrations of about 1.5% natural acetaldehyde can be achieved, which can be concentrated further to the desired application level. From yeast to yeast the substrate specificity of the alcohol oxidase is different. Hence this procedure can also be used to convert other alcohols, such as hexenol and other long-chain alcohols, to the corresponding aldehyde (29).

Generation of Taste through (Redox) Biocatalysis

147

Aldehyde oxidases may be applied to improve the aroma profile of certain food products by the conversion of undesirable aldehydes to their corresponding acid in the presence of oxygen. In this way the beany off-flavour from soybean extracts has been 'neutralised successfully (30). As an alternative to alcohol and aldehyde oxidases the corresponding dehydrogenases could in principle be used. A severe drawback, however, is that these dehydrogenases require the expensive cofactor NAD(P)+ instead of (cheap) oxygen as an electron acceptor. Although various sophisticated NAD(P)'-cofactor regenerating systems have been designed and substantial cost reductions have been realised in this way, it is evident that in commercial applications oxidases are preferred over their dehydrogenase counterpart. A special type of alcohol oxidase is vanillyl alcohol oxidase (VAO) from Penicillium simplicissimurn. Recently, it has been shown by us that this stable, flavin-containing enzyme has a very broad substrate specificity and readily converts para-substituted phenols into interesting flavour precursors or flavouring compounds (3 1,32, Figure 6 ) . Apart from natural vanillin and coniferyl alcohol, different vinylphenols (e.g. para-vinylguaiacol) can be produced from cheap raw materials and oxygen as an electron acceptor. VAO renders itself also for the generation of flavouring building blocks. To that end a natural mix of phenolic compounds (e.g. creosote) could be treated with VAO to enrich it with a range of vinylic and aldehydic flavouring substances.

Figure 6 Reactions catalysed by vanillyl alcohol oxidase (VAO)

148

Flavours and Fragrances

Peroxidases are enzymes which use hydrogen peroxide instead of oxygen as an electron acceptor. These Fe-heme containing enzymes occur widely in nature and can be applied to remove excess hydrogen peroxide in certain food products (e.g. cold sterilisation of milk in regions lacking refrigeration), to generate powerful antimicrobial agents such as HOSCN in milk (e.g. lactoperoxidase), to texturise food systems by cross-linking polysaccharides andor proteins, or as a biocatalyst to produce specific flavours (33). The latter two applications are relatively new and form part of our research programme in Wageningen (34). Enzymes currently under investigation for the in vitro production of bioflavours are horseradish peroxidase, soybean peroxidase, and several so-called microperoxidases. These microperoxidases (MPs) consist of a ferriprotoporphyrin IX covalently attached to a small residual oligopeptide chain. They can be prepared by mild proteolytic digestion of, for example, cytochrome c, and exhibit both cytochrome P450 type of oxygen transfer reactions as well as peroxidase-type of reactions (37). Depending on the number of amino acids in their oligopeptide chain MPs are denoted MP-8, MP-I I , etc. Reactions catalysed by (micr0)peroxidases include sulphoxidation, N-demethylation, oxidation and hydroxylation. A typical example of an interesting N-demethylation reaction is the conversion of methyl-N-methylanthranilate from citrus into the concord grape topnote N-methylanthranilate (34, Figure 7). Both horseradish peroxidase, soybean peroxidase and MP-8 were shown to catalyse this reaction in the presence of hydrogen peroxide.

CH3 I NH

Methyl N-methylanthanilate

NH-2

I

Methyl anthranilate

Figure 7 Peroxidase-catalysed demethylation of methyl-N-methylanthranilate Although reductases play an important role in the in vivo synthesis of various flavours (see for example figure 4) little attention has been paid to this type of biocatalysts. In most cases whole microbial or plant cells are used to perform a bioreductive reaction due to the requirement of (expensive) cofactors. Typical examples include the reduction of certain double bonds in terpenes by plant cells (l,38), the reduction of Massoi lactone to R(+)-& decalactone by basidiomycetes and S. cerevisiae (39), and the bakers yeast catalysed reduction of ketones into (chiral) alcohols (40). Recently, we started to explore the reductive power of the extreme thermophile microorganism Pyrococcusfuriosus. This organism grows rapidly around 100 OC and was found to convert a whole range of aliphatic and aromatic acids into their corresponding aldehydes and alcohols (41). The reduction of the acid to the aldehyde is catalysed by a W-

Generation of Taste through (Redox) Biocatalysis

149

containing ferredoxin aldehyde oxidoreductase, while the subsequent reduction of the aldehyde to the alcohol is catalysed by a broad spectrum thermophilic alcohol dehydrogenase. Currently, we are optimising the reduction of C3-CI2 acids using starch as a cheap substrate for the organism. In addition this organism is capable of reducing ketones to (chiral) secondary alcohols (Figure 8).

0 II

CHs-(CH2)"-C-CH3

0 II R-(CHZ)~-C-CH~

OH I > CH3-(CH2),-CH-CH-j

OH I

> R-(CHz),-CH-CHS

Figure 8 Bioreductive reactions catalysed by Pyrococcus furiosus. n = l - 1 6 and R=aromatic Whether this organism is able to reduce double bonds remains to be established. The advantage of using such an extreme thermophile is the possibility to scavenge the volatile products on line during fermentation. It is furthermore envisaged that the extra costs associated with the heating of the fermentor is outweighted by the fact that the process does not require expensive sterilisable equipment.

MULTIFUNCTIONAL INGREDIENTS The trend to design natural food ingredients which deliver in addition to taste: other attributes to the final product is clearly emerging. Traditionally, multifunctional ingredients are obtained by mixing several ingredients together, of which each has its own functionality. In this way many multifunctimal ingredient mixes have been created, which affect at least two of the following characteristics: taste, appearance, texture, nutritional value, shelf-life and process tolerance. By using enzymes, either as such, or in combination with microorganisms, multifunctional ingredients can be produced from relatively simple raw materials in one step. At Quest International several of such fermentative processes have been, or are being developed (42.43). A typical example is depicted in figure 9 and involves the production of a breadhacker improver which speeds up the dough processing time significantly and clearly has a positive effect on the taste and texture (e.g. volume) of

I50

Flavours and Fragrances

the final product (44). A tailored enzyme cocktail consisting of amylases and proteases is fermented in a controlled fashion together with a wheat flour fraction and selected strains of lactobacilli andor yeast to obtain the desired functionalities.

Wheat flour fraction

-

I

Enzymes, yeast, lactobacillus

+; Fermentor

1

BREAD IMPROVER with taste enhancing effects

Figure 9 Schematic representation of a process which yields a multifunctional bread improver. DSP is downstream processing, including removal of liquid and (spray)drying Another multifunctional ingredient, which is currently on the market, is prepared by fermenting either tomato solids, meat extracts or whey in the presence of lactic acid bacteria. The final ingredient after drying delivers viscosity (e.g. dextrans) as well as, for example, tomato taste to end products such as ketchups and soups. Alternatively, lactics grown in the presence of meat bone extracts deliver a stabilising and thickening ingredient with a meaty taste. Similarly, shelf-life extenders with a tailored taste profile can be produced by growing a bacteriocin-producing Pediococcus in the presence of whey.or corn steep liquor. Yet another multifunctional ingredient involves the production of a butter flavour which also provides texturising properties by fermenting Streptococcus diacetylactis in the presence of pasteurised milk or whey. Finally, cheese building blocks with multiple functions can be obtained by growing methyketone-producing Penicillium species in the presence of whey. The way of processing is such that the yield of the food-fermentation process is very high. Typically, all the raw materials present during fermentation are dried and hence end up in the products (42,43).

Generation of Taste through (Redox) Biocatalysis

151

CONCLUDING REMARKS Tailoring taste and other functionalities in food products through biocatalysis is still in its infancy. In practice only a limited number of biocatalysts, mainly enzymes and microorganisms, are intentionally being used to affect in one way or another the taste of foods. The scope, however, of biocatalysts is enormous and it can be expected that through genetic, protein and metabolic pathway engineering novel biocatalysts will soon become available, which are specifically designed for flavouring purposes. Especially for redoxenzymes we foresee great opportunities since, unlike most other enzymes, they are capable of tailoring not only the taste, but also all the other relevant functionalities in foods.

ACKNOWLEDGEMENT The authors wish to thank Mr. M. Fraaije for drawing some of the figures.

REFERENCES 1. R.G. Berger, ‘Aroma Biotechnology’, Springer-Verlag, Berlin, 1995. 2. E.W. Seitz, J.Dairy Sci., 1990,73,3664. 3. U. Antczak, J. G6ra, T. Antczak and E. Galas, Enzyme Microb. Technol., 199 1, 13, 589. 4. K.H. Engel, 1n:’Flavour Precursors’, R. Teranishi, G.R. Takeoka, M. Gunther (eds.), ACS Symp ser 490, ACS Wash DC, 1992,21. 5. Y.Watanabe and T. Inagaki, Jap patent 79101487, 1979. 6. P. Christen and A. L6pez-Munguia, Food Biotechnol., 1994,s. 167. 7. H.Stam, M. Hoogland and C. Laane, 1n:’Food Fermentations’, B. Wood (ed.), 1997, in press. 8. G. Budolfsen, K. Hansen, P. Hvass and P.M. Nielsen. Wopatent 9425580, 1994. 9. W.J. Quax, Trends in Food Sci. & Technol., 1993, 4, 3 1. 10. J.S. Weisman and P.S. Kim, Nature, 1993,365, 185. I 1. M.J. Van der Werf, W.J.J. Van den Tweel, J. Kamphuis, S. Hartmans and J.A.M. De Bont, Trends in Biotechnol., 1994,12,95. 12. R. Pickergill, 3. Jenkins, G. Harris, W. Nasser and J. Robert-Baudouy, Structural Biology, 1994, 1,717. 13. P. Van der Schaft, I. Van Geel, G. De Jong, N. Ter Burg, 1n:’Trends in Flavour Research’, H. Maarse, D.G. Van der Heij (eds.), Elsevier, Amsterdam, 1994,437. 14. R.C. Lindsay and J.K. Rippe. 1n:’Biogeneration of Aromas’, T.H. Parliament, D. Croteau (eds.), ACS Symp ser 3 17, ACS Wash DC, 1986,286. 15. J.R. Whitaker and M. Mazelis, 1n:’Food Enzymology’, P.F. Fox (ed.), Elsevier, London, 1991,2,479. 16. T.A. Hansen and M.J.E.C. Van der Maarel, WO patent 94/26918 and 96/03518. 17. J.P. Van der Lugt, 1n:’Proceedings 6th Neth. Biotechnol. Congress’, R.M. Buitelaar (ed.), 1996, p. W26. 18. K.R. Cadwallader,R.J. Bradd0ckandM.E. Parish, J.FoodSci., 1992,57, 241. 19. P.S.J. Cheetham, Tibrech., 1993, 11,478.

152

Flavours and Fragrances

20. I. Dugelay, Z. Gunata, J.C. Sapis, R. Baumes and C. Bayonove, J.Agric.Food Chem., 1993,41,2092. 2 I . J.C. Slaughter and F.G. Priest, 1n:’Food Enzymology’, P.F. Fox (ed.), Elsevier, London, 1991,1,61. 22. J,C. Villettaz and D. Dubourdieu, 1n:’Food Enzymology’, P.F. Fox (ed.), Elsevier, London, 1991,2,427. 23. C-H. Wong, F.P. Mazenod and G.M. Whitesides, J.0rg. Chem., 1983,48,3493. 24. M.J.E.C. Van der Maarel. W. Aukema and T.A. Hansen, Ferns Microbiol. Lett.,1996, 143, 241. 25. R.G. Berger, ‘Aroma Biotechnology’, Springer-Verlag, Berlin, 1995, 101. 26. A.M. Mayer and E. Harel, 1n:’Food Enzymology’, P.F. Fox (ed.), Elsevier, London, 1991, 2, 373. 27. P.A. Markus, A.L.J. Peters and R. Roos, EU patent 542284, 199I . 28. P.F. Fox and M.B. Grafferty, 1n:’Food Enzymology’, P.F. Fox (ed.), Elsevier, London, 1991, 1, 253. 29. W.D. Murray and S.J.B. Duff, Appl. Microbiol. Biotechnol., 1990,33,202. 30. N. Takahashi, R. Sasaki and H. Chiba, Agric. Biol. Chem., 1979,43, 2557. 31. M.W. Fraaije, C. Veeger and W.J.H. Van Berkel, Eur. J. Biochem., 1995,234,271. 32. M.W. Fraaije, F. Drijfhout, G. Meulenbelt and W.J.H. Van Berkel, 1n:’Proceedings 6th Neth. Biotechnol. Congress’, R.M. Buitelaar (ed.), 1996, p.PH1. 33. D.S. Robinson, 1n:’Food Enzymology’, P.F. Fox (ed.), Elsevier, London, 1991,2, 399. 34. M.J.M. Van Haandel, C. Laane and I.M.C.M. Rietjens, Biotech. Bioengineering, in preparation. 35. Pernod Ricard, FR patent 8912901 and EU patent 481 147. 36. J. Maat, W. Musters, H. Stam, P.J. Schaap, P.J. Van der Vonderwoort, J. Visser and J.M. Verbakel, US patent 5,529,926. 37. A.M. Osman, J. Koerts, M.G. Boersma, S . Boeren, C. Veeger and I.M.C.M. Rietjens, Eur. J. Biochemistry, 1996,240, 232. 38. F. Drawert and R.G. Berger, In:’45 Disk Tagung, FK der Ernahrungsindustrie Hannover’, 1987,93. 39. P.H. Van der Schaft, N. Ter Burg, S. Van den Bosch and A.M. Cohen, Appl. Microbiol. Biotechnology, 1992.36, 7 12. 40. J. Heidlas et af.,Eur. J. Biochem., 1988, 172, 633. 4 1. E. Van den Ban, C. Laane and H. Haaker, unpublished results. 42. C. Laane, Voedingsmiddelentechnologie,1995,23,24. 43. C. Laane, Int. Food Ingred., 1996,1,23. 44. A.A.G. Van Duynhoven, J.N. Visser and M. Hoogland, EU patent 93309163.9, 1993.

The Maillard Reaction in Flavour Formation Hugo Weenen, J. Kerler, and J. G. M. van der Ven QUEST INTERNATIONAL, HUIZERSTRAATWEG 28,141 1 GP NAARDEN, THE NETHERLANDS

1

INTRODUCTION

An important part of thermal flavour generation is the reaction between amino acids

and reducing sugars, first described by Maillard’. It is not really a single reaction, but a cascade of reactions, the outcome of which is dependent on precursors and reaction conditions. Well known name reactions which are part of the Maillard reaction include Amadori rea~~angement”~, Heyns rearrangemen?94and Strecker degradati~n’.~. From a sensory point of view the Maillard reaction is important because intense aromas are developed, and browning occurs. According to an EC directive and IOFI guidelines, flavourings which are obtained by thermal treatment of a reducing sugar and a nitrogen source (amino group) are considered process flavourings. In the flavour industry process flavourings are generally generated by heating aqueous solutions of mixtures of: amino acids and reducing sugars. The resulting brown product mixtures are then generally spray dried. Process flavourings are often developed by empirical means, i.e. the reaction conditions are optimized using organoleptic evaluation. However, much effort has been spent on analysing the reaction products and on trying to understand reaction mechanisms. The first report on aroma formation by the Maillard reaction is probably by W. Ruckdeschel in 1914’. A more elaborate study on flavour formation by the Maillard reaction was carried out by Herz and Schallenberger in 1960’. They described the formation of a series of well known aromas such as chocolate, popcorn, bakery and potato from reactions between single amino acids and glucose. In that same year a patent was published by Morton et al.9 in which meat aroma formation is claimed from the reaction of cysteine, with pentoses and hexoses. In the 1960’s a great number of patents and publications followed, demonstrating the effectiveness of the Maillard reaction in the industrial production of flavourings”. Another important milestone in the history of Maillard chemistry was certainly the publication of the Hodge scheme in 1953”. The Hodge scheme indicates the interconnections of all the important reactions and intermediates in the Maillard reaction, including Amadori rearrangement products and deoxyosones. The scheme is still frequently used, but is somewhat complicated, and needs some revisions (vide inpa).

I54

Flavours and Fragrances

Moreover, it includes only very few flavour substances, and it is therefore not particularly useful, when assessing the role of the Maillard reaction in flavour formation. The development of capillary GC and HRGC-MS has made it possible to identif a great number of volatiles in thermally processed foodsI2 and process flavourings 13-17. However, there is no simple correlation between the relative concentration of these volatiles and their contribution to the overall flavour. Recently developed GC olfactometry techniques such as Charm Analysis18 and Aroma Extract Dilution A n a l y ~ i s ' ~have ' ~ ~ overcome this problem, and have shown to be very useful tools to deal with the apparent complexity of thermally generated flavours (vide inja). As mentioned before, the Maillard reaction is in reality a sequence of reactions, resulting in the formation of reactive intermediates. These will react further to other reactive intermediates, and finally form volatile flavour substances, compounds with taste and mouthfeel, and brown pigments. A considerable amount of effort has been spent to prepare these reactive intermediates and study their reactivity, in order to improve our understanding and control of the Maillard reaction (vide inja). Flavour substances which are formed by the Maillard reaction can be classified in two classes, from a chemical point of view: 1. Products which are formed after condensation of deoxyosones, e.g. furfural, 2,5dimethyl-4-hydroxy-3(2H)-furanone etc. These products are generally formed in relatively high yields. 2. Products which are formed after carbohydrate fragmentation. Such products are always formed in relatively low yields (
2 DETERMINATION OF CHARACTER THERMALLY PROCESSED FOODS

IMPACT

COMPONENTS

IN

Aroma character impact components can be determined by either Charm Analysis", I9,2 I .22 or Headspace Dilution Analysis''. These Aroma Extract Dilution Analysis methods are all based on the evaluation of aroma compounds by GC olfactometry, using stepwise decreasing concentrations. This way a dilution will be obtained, in which only a small number of components are detected by GC ~ l f a c t o m e t r y ~ ~ ' ~ ~ . These components can then be quantified, which is often done using stable isotope dilution analysis. Based on the aroma value concept of Rothe and Thomas2', their odour activity values can then be determined (OAV, ratio of the concentration divided by the odour threshold). Aroma components with OAV's higher than 1 contribute to the overall flavour of a foodstuff or flavouring. The higher the OAV, the more it is expected to contribute to the overall flavour. Important aroma compounds derived from the Maillard reaction in thermally treated foods such as meat, bread, coffee, cocoa, chocolate, sesame, popcorn, French fries, tea, and fish are summarised in Table 1. In addition, their descriptors and odour thresholds (in water) are shown, as well as the foods in which they have been identified

The Maillard Reaction in Flavour Formation

155

by means of aroma extract dilution analysis (AEDA), static headspace gc/olfactometry (GC/O-H), or on the basis of their odour activity values (OAV's). The character impact compounds shown in Table 1 include substances containing sulphur, oxygen andor nitrogen, and their contribution to the overall flavour is strongly dependent on the type of food and the mode of processing. The flavour of cooked meat is a typical example of a Maillard type aroma, which is formed from sugars, amino acids, peptides, proteins, lipids and nucleotides. Boiling rather than roasting of meat favours the formation of odour active thiols, such as 2methyl-3-furanthiol (l), 2-furfurylthiol (2). 3-mercapto-2-pentanone (1) and 2,5dimethyl-3-furanthi01 (4)31*33*34. They are generated either by reactions of sugars with cysteine or by thiamine breakdown. Compounds 1 and 2 contribute considerably to the aroma of boiled beef, whereas 3 has been found to be important for both boiled beef and hicke en'^''^. Thiol 4 has only once been reported as an important contributor of cooked chicken34, but since then, 4 has never again been identified in the volatile fraction of thermally treated meat. Recent studies have shown that in boiled chicken35 and stewed beet6, furfuryl thiol (Z), hydrogen sulfide (6)and methanethiol (I)are the only thiols which contribute significantly to the overall flavour. This is probably due to the higher stability of 2 compared to 1 and 3 in terms of oxidation, and other

reaction^'^. An important sulphur containing compounds in cooked beef and chicken, which is not derived from cysteine or thiamine, is the Strecker aldehyde methional (9)31933,41. In addition, the typical Maillard products 4-hydroxy-2,5-dimethyl-3(2H)-furanone(HDF, La), and 3-hydroxy-4,5-dimethyl-2(5H)furanone (sotolon, U ) ,as well as the Strecker aldehydes 2- and 3-methylbutanal (fi,fi), methylpropanal and acetaldehyde (19) are responsible for the caramel-like, seasoning-like and malty character of thermally prepared beef and ~ h i c k e n * ~ . ~ ~ * ~ ' . The composition of meat flavour depends strongly on how meat is prepared. Most of the thiols, particularly 2-methyl-3-furanthiol (1)and its oxidation product bis-(2methyl-3-furyl)-disulfide(2), belong to the character impact compounds of boiled beef. 2-Ethyl-3,5-dimethylpyrazine(24), 2,3-diethyl-5-methylpyrazine and 2-acetyl-2thiazoline are typical key odorants of roast beef3. The aroma of coffee has been studied quite extensively, resulting in more than 700 reported volatile components. However, Semmelroch and G r ~ s c hwere ~ ~ able to simulate the aroma of Arabica and Robusta coffee brews with reconstituted aromas consisting of 23 odorants exhibiting high OAV's. Besides 2-furhrylthiol (2), which is the most important flavour substance of coffee, the aroma of Arabica coffee is characterised by the Maillard products HDF (La), sotolon (U),methanethiol (j), 2,3butanedione 2,3-pentanedione ( X ) ,2-ethyl-3,5-dimethylpyrazine(B), and 2,3diethyl-5-methylpyrazine(Xi), as well as the Strecker aldehydes methylpropanal and 2(3)-methylbutanal (J5J6)43. These compounds are responsible for the caramel-like, sulfiry, roasty, earthy, buttery and green odour notes of the Arabica coffee. In Robusta coffee, which has a more roasty and earthy character than Arabica coffee, the pyrazines U and 2 are more important, whereas the furanones U and U contribute to a lesser extent to the overall flavouP3.

(u),

(m)

(a),

(x),

(n),

malty, cocoa-like, green malty, fruity, cocoa-like

3-methylbutanal

15.

16.

2-ethyl-4-hydroxy-5-methyl-3(2H)- caramel-like, sweet furanone (EHMF) 3-hydroxy-4-methyl-5-ethyl-2(SH)- seasoning-like

0.35 (30)

7.5 (52)

20 (43), 1.15 (52)

0.001 (54), 0.3 (53)

furanone (sotolon)

furanone (abhexon) 2-methylbutanal

14.

13.

12.

seasoning-like

B. I I.

I.O(3l)

3-hydroxy-4,5-dimethyl-2(5H)-

Compounds containing oxygen

10.

5.

10 (52). 100 (53)

roasty, popcorn-like, burnt

2-acetyl-2-thiazoline

6. 7. 8. 9.

0.7 (17) 0.018 (17) 0.2 (26) I0 (27) 0.00002 (28) 0.004 (29), 0.01 (73) 0.2 (30)

caramel-like, strawberry-like

sulfury, catty meaty, sweet, sulfury sulfury, putrid sulfury, eggy, putrid meaty, sulfury cooked meat-like cooked potato-like

3-mercapto-2-pentanone 2,5-dimethyl-3-furanthiol methanethiol hydrogen sulfide bis-(2-methyl-3-furyI)disulfide 2-methyl-3-(methyldithio)furan methional

3. 4.

0.007 ( I 7) 0.01 (17),0.12(26)

Odour threshold in water [pg/kg]

4-hydroxy-2,5-dimethyl-3(2H)furanone (HDF)

meaty, sweet, sulfury roasty, sulfury

Compounds containing sulphur 2-methyl-3-furanthiol 2-furfurylthiol

A.

I. 2.

Odour description

Compound

beef (42), chicken (39, coffee (38), french fries (47), bread (50) beef (26,33,42), chicken (35). chocolate (51). cocoa (51), coffee (38), french fries (47), bread (50,60), french fries (47). tea (20). beer (691

coffee (52), chocolate ( 5 I), cocoa ( 5 I )

beef(26,3 1,33,42), chicken (35), coffee (43), beer (58), popcorn ( 5 8 ) , bread (50,58),chocolate ( 5 I), sesame (59), french fries (47), tea (20) beef (26,33,42), chicken (35). coffee (43), french fries (47), tea (20), chocolate ( 5 I), cocoa ( 5 I ) coffee (43)

beef (32,33), chicken (34.35). pork (36), coffee (37,38) beef (26,33,34), chicken (34,35), pork (36), coffee (37,38), sesame (39), popcorn (40) beef (33,34,45), chicken (33,4 I ) chicken (34) beef (26,42), chicken (35), coffee (43), cod (44) chicken (27.35) beef (32-34,45), chicken (34,4l), tea (20) cocoa (51). chocolate (51) beef (26,32,33,42,45,46), chicken (34,35), pork (36), coffee (37,43), french fries (47), potato chip (30), fish (44,48,49), bread (50) beef (26,3,42,46), chicken (34,35), sesame (39), boiled trout (48)

Detected inb

Important aroma compounds derivedfrom the Maillard reaction in various thermally treatedfoodsa

No.

Table 1

-

23 5

b

x

2z

2

m Ln

acetaldehyde

2,3-butanedione

2,3-pentanedione Compounds containing nitrogen 2-acetyl- 1 -pyrroline

6-acety ltetrahydropyridine 2-ethyl-3,S-dimethylpyrazine

18.

19.

20.

21. C. 22.

23. 24.

earthy, roasty

earthy, roasty earthy, roasty roasty, sweet, nutty

2-ethenyI-3,S-dimethylpyrazine 2-ethenyl-3-ethyI-S-methyl-

pyrazine 2-acety lpyrazine

26.

27. 28.

29.

62 (63)

1.0 (31), 0.09 (41)

0.4 (30), 0.4-5.0 (56)

sesame (39), popcorn (66), bread (60)

bread (50,60), rice (65), popcorn (40,66), Sesame (39), beef (32,33), french fries (47) bread (50,60), popcorn (40,66) beef (31,42), chicken (41), coffee (37,43), sesame (39), bread (50,60,67), french fries (47), cocoa ( 5 I), chocolate (5 I), sesame (59), popcorn (60) bread (50,67), french fries (47), cocoa (5 I), chocolate (5 I), popcorn (60) beef (31,42), chicken (41). coffee (37,43), french fries (47), cocoa (5 I), chocolate (5 I), sesame (59), popcorn (60), bread (67) coffee (68) coffee (68), french fries (47)

beef (42). chicken (35). bread (50), chocolate (51), coffee (38), french fries (47), tea (20) beef (32,42), cocoa (5 I), chocolate ( 5 l), bread (50,60), french fries (47), coffee (37,61), tea (20) beef (26,42), chicken (39, coffee (38), french fries (47), fish (44) beef (26,33), bread (50,60), coffee (43). fish (44,48,49), chocolate (5 I),cocoa (5 I) fish (44,48,49), coffee (38), bread (50)

Detected inb

The sensory significance of the aroma compounds have been evaluated on the basis of the aroma value concept, that means on the basis of either aroma extract dilution analysis (AEDA), static headspace gdolfactometry (GCO/H), or stable isotope dilution analysis (IVA). Detected and evaluated in foods by AEDA, GCO/H, or IVA.

earthy, roasty

0.1 (62)

roasty, popcorn-like, breadlike roasty, cracker-like earthy, roasty 1 .O (63), 1.6 (64) 2(31),0.16(41)

30 (57)

2-3 (56), 15 (57)

25 (26), 15-120 (56)

4 (30)

0.1-2.3 (56), 0.7 (49)

Odour threshold in water [pgkg]

buttery, green

buttery

solvent-like

25.

b:

':

pheny lacetaldehyde

17. honey-like, sweet, flowery

Compounds containing oxygen (continued) methylpropanal malty, fruity, pungent

B.

Odour description

Compound

Important aroma compounds derivedjom the Maillard reaction in various thermally treated foodsa(continued)

No.

Table 1

I . VI

I58

Flavours and Fragrances

The character of cocoa and chocolate is also strongly affected by both a fermentation and a heat treatment step during processing. Their flavour is mainly characterised by lipid oxidation and Maillard reaction products. Recently, Schnermann and Schieberlesl identified 2-ethyl-3,s-dimethylpyrazine 2-ethyl-3,6-dimethylpyrazine 2,3-diethyl-S-methylpyrazine 3-methylbutanal phenylacetaldehyde and sotolon (2) as important aroma compounds of cocoa and chocolate. In addition, 2-methyl-3-(methyldithio)furan (&), an oxidation product of 2-methyl-3furanthiol and methanethiol, was described with a high flavour dilution factor. This meat-like disulfide has a low odour threshold in water, and illustrates that a limited number of Maillard reaction products are present in a great number of sometimes unrelated thermally treated foodstuffs. The flavour of cereal products, especially of bread, has been extensively studied, and the results have been recently reviewed76. The pleasant roasty character of wheat bread crust and popcorn depends on the formation of 2-acetyl-I-pyrroline (22) and 2acetyltetrahydropyridine (2). Both compounds, which are not important in wheat bread crumb and rye bread, are generated by a reaction of proline (2and 23) andor ornithine with reducing sugars or sugar breakdown products". The amount of formed in bread is strongly dependent on the amount of yeast used for the baking process 76 . Besides lipid oxidation products, the caramel-like and malty aroma of both wheat and rye bread is determined by HDF and the Strecker aldehydes methional (I),3-methylbutanal and methylpropanal French fries are a popular food with a characteristic pleasant aroma that has been described as deep-fried, boiled potato-like, earthy, malty, and ~ a r a m e l - l i k e ~Using ~. AEDA and GC/O-H, (E,E)-2,4-decadienal, methional (9), the pyrazines 2, and HDF methanethiol (I),3-methylbutanal (14).and 2,3-butanedione were evaluated as key odorants of French fries47.

(z),

(a),

(u),

(a),

(u),

(a)

a

(u),

a, (a)

(u),

3

(u), (u).

IMPORTANT AROMA COMPOUNDS IN PROCESS FLAVOURINGS

Meatlike process flavours require cysteine in combination with pentoses and hexoses as starting materials. Although the most important role of cysteine seems to be to generate hydrogen sulfide, it cannot be simply replaced by hydrogen sulfide. The special role of cysteine can be attributed to its inhibitory effect on browning, as well as to the formation of potent thiols exhibiting very low odour thresholds. Precursor studies have shown that the thiol group is introduced by the reaction of a carbonyl group with hydrogen sulfide17.For example, 2-furfurylthiol(2 in Table 2) is formed by the reaction of furfural with hydrogen sulfide, whereas furfurylalcohol under the same conditions did not result in considerable amounts of 2. The character im act compounds of cysteine based process flavourings have been extensively ~tudied'~.'-~',and are shown in Table 2. 2-Methyl-3-furanthiol (I),2-furfurylthiol (2), 3-mercapto-2-butanone (l), 3(2)mercapto-2(3)-pentanone (4,I),hydrogen sulfide (Y),methanethiol (G),ethanethiol 2-acetyl-2-thiazoline and S-acetyl-2,3-dihydro-I ,4-thiazine (fi) have been identified as im ortant flavour substances in reactions of cysteine with ribose, glucose, and r h a m n o ~ e ~ ' ~Accurate ~. quantitative data has shown that cysteine and ribose compared to glucose or fructose are a better precursor system for the compounds l-Ii7. In addition, in riboselcysteine models the thiophene analogues of . I and 2, namely 2-

(u),

(u),

I59

The Maillard Reaction in Flavour Formation

methyl-3-thiophenethiol(6) and 2-thenylthiol (9),contribute considerably to the overall flavour. Compounds 1,2, and 4, as well as 2-methyl-4,5-dih dro-3-furanthiol ( I )have 7Y also be identified as important thiamine breakdown products . Important aroma compounds, which are exclusively formed in the presence of glucose, are 2-( 1-mercaptoethyl)furan (la) and 2-( 1-mercaptoethyl)thiophene (U). They are responsible for the sulfury, burnt character of thermally treated cysteine/glucose model systems. In contrast, 5-methyl-2-furfurylthiol (lQ)and 5methyl-2-thenylthiol (11)only belong to the potent odorants of cysteinelrhamnose process flavourings. The yields of most of the important thiols (see Table 2) formed in the Maillard reaction are strongly dependent on the H,S concentration, and therefore, also on the amount of cysteine added to the reaction system. Aroma compounds, which are not dependent on cysteine, but contribute significantly to the overall flavour of cysteine derived process flavourings are 4hydroxy-2,5-dimetyl-3(2H)-furanone 4-hydroxy-5-methyl-3(2H)-furanone 2-ethyl-4-hydroxy-5-methyl-3(2H)-fUranone 3-hydroxy-4,5-dimetyl-2(5H)furanone and 3-hydroxy-6-methyl-2(2H)-pyranone (2.5)70’71.Compounds 2, 23, and 2 are preferably formed from cysteine in the presence of rhamnose, and therefore, are responsible for the more caramel- and seasoning-like character of these process flavourings7’. In contrast, HMF (22) is more important for the flavour of ribose/cysteine model systems, whereas the contribution of sotolon to the aroma of process flavourings is less influenced by the type of sugar. In thermally treated solutions of proline and glucose, 2-acetyl- 1-pyrroline (y), 6acetyltetrahydropyridine 2-acetylpyridine and HDF (a)have been evaluated as important aroma compounds by means of charm analysis16. The characteristic roasty and popcorn-like odour qualities of this process flavouring can be attributed to the nitrogen containing compounds 2 l and which have also been described as character impact compounds of various thermally treated cereal products76.

(a),

(a),

(z),

(a),

(a)

(a),

(a),

a,

4

REACTIVE INTERMEDIATES

4.1 Amadori and Heyns Rearrangement Products (ARP’s and HRP’s)

The first event in the reaction of an amino acid and reducing sugar is the condensation of the amino group of the amino acid and the carbonyl of the sugar. This results in an imine or enamine, which cannot be isolated, but which reacts further via an Amadori (in case of aldoses) or Heyns (in case of ketoses) to relatively stable amino ketoses and aldoses, which can be isolated. ARP’s have been detected in a great number of foodstuff^^^-^', which include dried fruits, vegetables, cocoa beans, tea, tomato paste, tobacco, beet molasses, malt, and beer. He ns rearrangement products (HRP) have only been identified in liver and liquorice Yg. Apparently foodstuffs which have undergone heat treatment contain A m ’ s and HRP’s. Studies on flavour formation from ARP’s and HW’s were carried out by van den Ouweland et d7’.A large number of typical Maillard volatiles were identified. Comparison of glucose/glycine and fructose/glycine rearrangement products, showed that volatile formation is significantly different for ARP’s and HRP’s.

sulfury, putrid sulfury, catty

3-mercapto-2-butanone

3-mercapto-2-pentanone

2-mercapto-3-pentanone

2-methyl-3-thiophenethiol

3.

4.

5.

6. 7. 8. 9. 10. I I. 12. 13. 14. 10 (27)

0.2 (26)

sulfury, putrid

I.0(31) 1.25 (17)

sulfury, putrid roasty, popcorn-like

roasty,popcorn-like

methanethiol

ethanethiol

2-acetyl-2-thiazoline

5-acetyl-2,3-dihydro- I ,4-thiazine

17.

18.

19.

0.00002 (28) 0.042 (1 7) 0.048 ( I 7) 0.049 (1 7) 0.022 ( 17) 0.038 ( I 7)

meaty, sulfury meaty, sulfury meaty, sulhry sulfury, roasty sulfury, roasty sulfury, roasty sulfury. burnt sulfury, burnt sulfury, burnt sulfury, egg-like, putrid

0.02 ( I 7)

0.7 ( I 7)

0.01 (17), 0.12 (26) 3.0(17)

0.007 ( 17)

Odor threshold [pgkg] in water

16.

15.

2-methyl-4,5-dihydro-3-furanthiol bis-(2-methyl-3-furyI)-disulphide 2-thenylthiol 5-methyl-2-furfurylthiol 5-methyl-2-thenylthiol 2-( I -mercaptoethyl-furan 2-( I -mercaptoethyl)thiophene 2-methyltetrahydrothiophene-3-one hydrogen sulfide

sulfury, roasty, coffee-like

2-furfurylthiol

2.

sulfury, catty

meaty, sulfury. sweet

Compounds containing sulphur 2-methyl-3-furanthioI

A.

I.

Odour description

Compound

Flavour impact components of processjlavouringsa

No.

Table 2

cysteine/ribose (70,72), cysteine/glucose (7 I), cysteinel rhamnose (7 I), glutathione/ribose (72). thiamine (72) cysteinelribose (70,72), cysteine/glucose (7 I), cysteinel rhamnose (7 I), glutathione/ribose (72), thiamine (72) cysteinc/ribose (70), cysteine/glucose (7 I), cysteinel rhamnose (7 1) cysteindribose (70,72), cysteine/glucose (7 I), cysteind rhamnose (7 l), glutathione/ribose (72), thiamine (72) cysteine/ribose (1 7). cysteine/glucose (1 7), cysteinel rhamnose ( I 7) cysteine/ribose (70) thiamine (72) cysteine/ribose (70,72), glutathionelribose (72) cysteine/ribose (70). cysteine/glucose (7 1) cysteinelrhainnose (7 I ) cysteine/rhamnose (71) cysteine/glucose (7 I ) cysteine/glucose (71) cysteine/ribose (70.72) cysteine/ribose (70). cysteine/glucose (7 I), cysteinel rhamnose (7 1) cysteine/ribose (70). cysteinelglucose (7 I), cysteinel rhamnose (7 1) cysteine/ribose (70), cysteine/glucose (7 I), cysteind rhamnose (71) cysteinehibose (70), cysteinelglucose (7 I), cysteine/ rhamnose (7 1) cysteine/ribose (70), cysteine/glucose (7 I), cysteinel rhamnose (71)

Detected inb

4

0

w

a a

b

5

n

2 52

8

-

f. a,b in Table1

Compounds containing oxygen 4-hydroxy-2,5-methyl-3(2H)thiophenone 4-hydroxy-2,5-dimethyl-3(2H> furanone (HDF) 4-hydroxy-5-methyl-3(2H)-furanone (HMF) 2-ethyl-4-hydroxy-5-methyl-3(2H)furanone (EHMF) 3-hydroxy-4,5-dimethyl-2(5H)furanone (sotolon) 3-hydroxy-6-methyl-2(2H)-pyranone Compounds containing nitrogen 2,3-diethyl-5-methylpyrazine 2-acetyl- 1-pyrroline 6-acetyltetrahydropyridine 2-acetylpyridine cysteine/glucose (71), cysteinelrhamnose (71) prolinelglucose ( I 6) prolinelglucose (16) proline/glucose (16)

1.0 (31), 0.09 (41) 0.1 (62) 1.0 (63), 1.6 (64) I9 (63)

earthy, roasty roasty,popcorn-like roasty, burnt caramel-like roasty, caramel-like

seasoning-like

seasoning-like

(74) 20 (43), 1. I5 (52)

caramel-like, sweet

cysteinelribose (70). cysteinelglucose (7 I), cysteine/rhamnose (7 I ) cysteine/rhamnose (7 1)

cysteine/rhamnose (71)

8500 (17), 23000

0.3 (53),0.001 (54) 15.0 (17)

cysteine/ribose (70), cysteine/glucose (7I), cysteinel rharnnose (7I), proline/glucose (16) cysteine/ribose (70)

10 (52), 100 (53)

cysteine/glucose (7I )

Detected inb

caramel-like,strawberrylike caramel-like, burnt chicory

24.0 (I 7)

[pgkg] in water

Odor threshold

caramel-like, sweet, meaty

Odour description

Flavour impact components ofprocessflavouringsa(continued)

Compound

e2

3

s.

9

2

3

3

-_

f:5

F

a

162

Flavoirrs and Fragrances

Scheme 1 : Bicyclic structures of 3-deoxyglucosone

0

o+oH

’

‘CH,OH

OH

HO & ~ ~ o H ’ z 3

Ho*H

5

OH

4

O H

HO

&cH20H

HO

7

8

The Maillard Reaction in Flavour Formation

163

In addition they found differences in the composition of volatiles, depending on whether APR’s were heated dry or in aqueous solution. Heinzler and Eichner” investigated volatiles formation from fructosealanine, under roasting conditions. The main aroma components formed were 2-acetylpyrrole, 5-methylfurfural and 2-acetylfuran. Other flavour substances included pyrazines, pyrroles, pyridines and h a n s . Unfortunately neither van de Ouweland et al. nor Heinzler and Eichner compared their results with the breakdown of the corresponding unreacted amino acid and sugar mixtures. Huyghues-Despointes” el al. and Keyahani et ~ 1investigated . ~ ~ the formation of volatiles from proline-glucose ARP and phenylalanine-glucose ARP using pyrolysidGCNS analysis. They compared the ARP’s with unreacted glucose, phenylalanine and a glucose + phenylalanine mixture. Clear differences were observed between ARP’s and the corresponding amino acidglucose mixtures. Unfortunately the results are difficult to compare with solution Maillard chemistry, as many of the products are quite different from what is obtained in an aqueous process flavouring. 4.2 Thiazolidine Carboxylic Acids

When cysteine is heated in the presence of a reducing sugar, ARP’s are not formed but instead thiazolidine carboxylic acids are formed as intermediate^^^'^^. These thiazolidine carboxylic acids are relatively stable, but decompose at their melting pointsg4.When heated to their melting points, 4-alkylthiazolidinecarboxylic acids were found to give 2,4-dialk 1-2,3-dihydrothiophenes and 2,4-dialkylthiophenes in yields varying from 0.16-18% L . The anionic form of these thiazolidinecarboxylic acids is apparently relatively stable, and has been proposed to inhibit browning and flavour developments5. The reactivity of the thiazolidines can be enhanced by the use of reactive sugars and buffer salts (low pH). The inhibiting effect of the thiazolidine pathway can apparently be avoided by using A m ’ s + cysteineg4. 4.3 Deoxyglycosones

The formation of deoxyglucosones from ARP’s was demonstrated by Beck et a1.86 using diaminobenzene as a trapping agent. ARP’s from glucose and maltose were found to give mainly 1- and 3-deoxyglucosones, but also 4-deoxyglucosone. The role of the deoxyglucosones in the Maillard reaction in foods and the human body has been very well reviewed by Led1 and Schleicher”. The structure of 3-deoxyglucosone was studied by Weenen and TjanSgvg9, who showed that 3-deoxyglucosone in aqueous solution exists as a mixture of five major relatively stable bicyclic isomers, with no free carbonyl moiety present (Scheme I). When heated under acidic conditions, intramolecular condensation takes place, resulting in relatively high yields of 4(hydroxymethyl)fuhralg8. When heated with asparagine methylated pyrazines are formedg8.Surprisingly, the yields of these methylated pyrazines were lower from 3deoxyglucosone, than from glucose or fructose, suggesting that carbohydrate fragmentation in the Maillard reaction is more efficient from other intermediates (see below).

Flavours and Fragrances

164

Scheme 2: Proposed formation of a-dicarbonylhgments

-

1. Formation of glyoxal

sugar, HRP

RA

glycolaldehyde b -.

[Ol

glyoxal

2. Formation of pyruvaldehyde

AR P

-b

RA glyceraldehyde-aa adduct

\

pyruvaldehyde

1-Done. 3-Done

3. Formation of diacetyl and 2J-pentanedione

1-Done

1. isom. A 2. RA

RA : retroaldolisation isom: isomerisation aa: amino acid

HCHO diacetyl

-b

2,3-pentanedione

165

The Maillard Reaction in Flavour Formafion

Figure 1: Formation of alpha-dicarbonyl compounds f?om carbohydrates and alanine (reprinted with permission h m ref 92. Copyright 1997 Elsevier Science Ltd)

I2O

2

80

-P

60

e 8 0

e .BU

40

I

m

.c

-a 20 m 0

glyoxal

pyruvaldehyde

w diacetyl

Ipentanedione

I66

5

Flavours and Fragrances

CARBOHYDRATE FRAGMENTATION

Products formed in the Maillard reaction from carbohydrate fragments include pyrazines, thiazoles, carbocylic compounds and other heterocyclic compounds, and are always formed in relatively low yields”. Carbohydrate fragmentation seems to be the yield determing step in these reactions, and therefore deserves special attention. When using cysteamine as a trapping agent for carbonyl containing intermediates, Weenen and Tjan” found that hydroxyacetone was formed from glucose and fructose, but not from 3-deoxyglucosone. They explained this by postulation that hydroxyacetone is formed by P-cleavage from isomerised 1 -deoxyglucosone. More recently Weenen and Apeldoom used 1,2-diaminobenzene as a trapping agent for a-dicarbonyl compounds, and concluded that carbohydrate fragmentation in the Maillard reaction takes place by retro-aldolisation of deoxyglycosones, the starting sugars (or their amino acid adducts) and possibly also of ARP’s and HRP’s9’. These results are in contrast with earlier conclusions based on pyrazine formation studies, but upon closer inspection are in agreement with the results of these experiments”. Four a-dicarbonyl containing fragments were detected: glyoxal, pyruvaldehyde, diacetyl and 2,3-pentanedione (Figure 1). As expected 3-deoxyglucosone was a particularly efficient precursor of pyruvaldehyde, but to our surprise xylose was also a relatively good precursor of pyruvaldehyde. This may be related to the expected lower stability of deoxypentosones. The formation of glyoxal can best be explained as resulting from retroaldolisation of the starting sugar or the corresponding amino acid adduct. The resulting glycolaldehyde should then be oxidised to give glyoxal, for example via a Strecker-like mechanism (Scheme 2). If HRP’s cleave via retroaldolisation, C-3 fragments are formed with only one carbonyl, i.e. glyceraldehyde or hydroxyacetone (+ amino acid adducts). Hydroxyacetone can isomerize to glyceraldehyde, and glyceraldehyde can generate pyruvaldehyde, upon p-elimination of H20. A summary of the proposed mechanisms for the formation of glyoxal, pyruvaldehyde, diacetyl and 2,3-pentanedione from normal sugars, ARP’s, HRP’s, 1- / 3-deoxyosones is given in Scheme 2. 6

DISCUSSION

Aroma character impact components can be determined by either Charm Analysis’’, l9,2 1.22 , or Headspace dilution analysis2’. These Aroma Extract Dilution Analysis methods have been applied to thermally treated foods and to process flavourings. Results so far indicate a large degree of similarity between thermally treated foods and process flavourings, as expected, with as notable exceptions, S-acetyl-2,3-dihydro- 1,4thiazine and 2-( 1 -mercaptoethyl)furan, which have only been detected in process flavourings so far. Research on Amadori rearrangement products, Heyns rearrangement products, thiazolidines and deoxyglycosones is fragmentary, but shows the importance of these reactive intermediates for the course of the Maillard reaction. Results so far suggest that fragmentation occurs from the dexoyglycosones, ARP’s and HRP’s, but also from the starting sugars, or their amino acid addition products. The implications for the overall scheme of flavour formation in the Maillard reaction are summarized in Scheme 3.

Flavour Substances

+

Scheme 3: The Maillard reaction in flavour formation (Reproduced with permission from ref. 92. Copyright 1997 Elsevier Science Ltd)

Ffavotir Substances

168

Flavours and Fragrances

References

I. 2.

L.C. Maillard, C.R. .4cad. Sci. Ser., 1912.2, 154. M. Amadori, Atti. R. Accad. Naz.. Lincei. Mem. CI. Sci. Fis. Mat. Nat., I93 I , 13, 72. 3. V.A. Yaylayan and A. Huyghues-Despointes, Crit. Rev. Food Sci. Nutr., 1994, 34(4), 32 1. 4. K. Heyns and H. Noack, Chem. Ber., 1962,720. 5. A. Strecker, Ann. 1862, 123,363. 6. A. Schonberg and R. Moubacher, Chem. Rev., 1952,50,261. 7. W.Z. Ruckdeschel, Z. ges. Brauw. 1914,37,430 + 437. 8. W.J. Herz and R.S. Schallenberger, Food. Res., 1960,25,491. 9. I.D. Morton, P. Akroyd and C.G. May, G.B. Patent 836,694, 1960. 10. G. MacLeod and M. Seyyedain-Ardebik, Crir. Rev. Food Sci. Nutr., 1981, 14,309. 11. J.E. Hodge, J. Agric. Food Chem., 1953, 1, 928. 12. H. Maarse, C.A. Visscher, L.G. Willemsens, L.M.Nijssen and M.H. Boelens, ‘Volatile Compounds in Food. Qualitative and Quantitative Data.’ Supplement 5 to the Sixth Edition. TNO Nutrition and Food Research, Zeist, The Netherlands, 1994. 13. R. Tressl, B. Helak, N. Martin, and E. Kersten, ‘Thermal Generation of Aromas’, American Chemical Society, Washington, DC, 1989, 156. 14. M. Guntert, J. Bruning, R. Emberger, M. Kopsel, W. Kuhn, T. Thielmann and P. Werkhoff, J. Agric Food Chem., 1990,38,2027. 15. P. Schieberle, Z. Lebensm. Unters. Forsch, 1990, 191,206. 16. D.D. Roberts and T.E. Acree, ‘Thermally generated flavors’, ACS Symposium Series 543, 1994, 7 1. 17. T. Hofmann, PhD Thesis, Technical University of Munich, 1995. 18. T.E. Acree, J. Barnard and D.G. Cunningham. Food Chem., 1984, 14,273. 19. F. Ullrich and W. Grosch, Z. Lebensm. Unters. Forsch., 1987, 184,277. 20. H. Guth and W. Grosch, Flavour Fragrance J., 1993,8, 173. 21. W. Grosch, Flavour Fragrance J., 1994,9, 147. 22. W. Grosch, Trends Foods Sci. Technol., 1993,4, 68. 23. T.E. Acree, ‘Flavor Science - Sensible Principles and Techniques’, American Chemical Society, Washington, D.C., 1993. 24. P. Schieberle, ‘Characterization of Food-Emerging Methods’, Elsevier, Amsterdam, 1995. 25. M. Rothe and B. Thomas, Z. Lebensm. Unters. Forsch., 1963, 119,302. 26. H. Guth and W. Grosch, J. Agric. Food Chem., 1994,42,2862. 27. E.L. Pippen and E.P. Mecchi, J. FoodSci., 1969,34,443. 28. R.G. Buttery, W.F. fladdon, R.M. Seifert and J.G. Turnbaugh, J. Agric. Food Chem., 1984,32,674. 29. P. Schieberle, personal communication. 30. D.G. Guadagni, R.G. Buttery, and J.G. Turnbaugh, J. Sci. Food Agric.. 1972, 23. 1435. 3 1. C. Cerny and W. Grosch, Z. Lebensm. Unters. Forsch., 1993, 196, 123.

The Maillard Reaction in Flavour Formation

32. 33. 34. 35. 36. 37. 38. 39. 40. 4 1. 42. 43. 44. 45. 46. 47. 48. 49. 50. 5 1. 52. 53.

54.

55. 56. 57. 58. 59. 60. 6 1. 62. 63. 64. 65. 66. 67. 68.

169

U. Gasser and W. Grosch, Z. Lebensm. Unters. Forsch., 1988, 186,489. R. Kerscher and W. Grosch, Z. Lebensm. Unters. Forsch., 1997,204, 3. U. Gasser and W. Grosch, Z. Lebensm. Unters. Forsch., 1990, 190,3. J. Kerler and W. Grosch, Z. Lebensm. Unters. Forsch., 1997, submitted for publication. U. Gasser and W. Grosch, Lebensmittelchemie, 1991,45, 15. I. Blank, A. Sen and W. Grosch, Z. Lebensm. Unters. Forsch., 1992, 195,239. P. Semmelroch and W. Grosch, Lebensm. Wiss. Technol., 1995,28,3 10. P. Schieberle, ‘Progress in Flavour Precursor Studies’, Allured Publishing Corporation, Carol Stream, USA, 1993, 343. P. Schieberle, J. Agric. Food Chem., 1991,39, 1 141. J. Kerler, PhD Thesis, Technical University of Munich, 1996. J. Kerler and W. Grosch, J. FoodSci., 1996,61 (6), 1271. P. Semmelroch and W. Grosch, J. Agric. Food Chem., 1996,44, 537. C. Milo and W. Grosch, J. Agric. Food Chem., 1995,43,459. H. Guth and W. Grosch, Lebensm. Wiss. Technol., 1993,26, 171. C. Cerny and W. Grosch, Z. Lebensm. Unters. Forsch., 1992, 194,322. R. Wagner and W. Grosch, Lebensm. Wiss. Technol., 1997,30, 164. C. Milo and W. Grosch, J. Agric. Food Chem., 1993,41,2076. C. Milo and W. Grosch, J. Agric. Food Chem., 1996,44,2366. M. Rychlik and W. Grosch, Lebensm. Wiss. Technol., 1996,29,515. P. Schnermann and P. Schieberle, J. Agric. Food Chem., 1997,45, 867. P. Semmelroch, G Laskawy, I. Blank and W. Grosch, Flavour Fragrance J., 1995, 10, 1. W. Grosch, G. Zeiler-Hilgart, C. Cerny and H. Guth, ‘Progress in Flavour Precursor Studies’, Allured Publishing Corporation, Carol Stream, USA, 1993, 329. A. Kobayashi, ‘Flavor Chemistry - Trends and Developments’, ACS Symposium Series 338, Washington DC, 1989. S . Fors, ‘The Maillard reaction in Foods and Nutrition’, ACS Symposium Series 215, Washington DC, 1988. J.C. Leffingwell and D. Leffingwell, Perfum Flavor, 1991, 16,2. I . Blank, A. Sen and W. Grosch, ‘ASIC, 14th Colloque, San Francisco, 1991. P. Schieberle, Lebensmittelchemie, 1993,47, 15. P. Schieberle, Food Chem., 1996,55 (2), 145. P. Schieberle and W. Grosch, 2. Lebensm. Unters. Forsch., 1987, 185, 11 1 . H. Holscher, PhD Thesis, University of Hamburg, 1991. R.G. Buttery, L.C. Ling, B.O. Juliano and J.G. Turnbaugh, J. Agric. Food Chem., 1983,31, 823. R. Teranishi, R.G. Buttery, and D.G. Guadagni, ‘Geruchs- und Geschmacksstoffe’, Verlag Hans Carl, Nilmberg, 1975, 178. R.G. Buttery and L.C. Ling, J. Agric. Food Chem., 1995,43, 1878. R.G. Buttery, L.C. Ling, B.O. Juliano, Chem. Ind., 1982,958. P. Schieberle, J. Agric. Food Chem., 1995,43,2442. P. Schieberle and W. Grosch, Z. Lebensm. Unters. Forsch., 1994, 198,292. M. Czemy, R. Wagner and W. Grosch, J Agric. Food Chem., 1996,44,3268.

I70

Flavours and Fragrances

69. P. Schieberle, Z.Lehensm. Un1er.s. Forsch.. 1991, 193, 558.

70. 7 1. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88.

89. 90.

91.

T. Hofmann and P. Schieberle, cJ. Agric. Food Chem., 1995,43,2 187. T. Hofmann and P. Schieberle. -J. Agric. Food Chem., 1997, 45, 898. U. Gasser. PhD Thesis. Technical University of Munich, 1990. R. Tress1 and R. Silvar, J. Agric. Food Chem.. 1981,29, 1078. R.G. Buttery, G.R. Takeoka, G.E. Kramnier and L.C. Ling, Lehensm. Wiss. Technol.. 1994,27, 592. T. Hofmann and P. Schieberle. J. Agric. Food Chem., 1996,44, 25 1. W. Grosch and P. Schieberle, FoodRev. In[., 1997, submitted for publication. K.O. Herz and S.S. Chang, Advcin. Food Res., 1970, 18, 1. I.D. Morton, P. Akroyd, and C.G. May, U K Patent 836694, 1960. G.A.M. Ouweland, H.G. Peer and S.B. Tjan. ‘Flavor of foods and beverages’, G. Charalambous (ed.), Academic Press, 1978, 13 1. M. Heinzler and K. Echner, Z. Lehensm. Unlers. Forsch., 1991, 192,445. R. Wittman and K. Echner, Z. Lehensm. Unters. Frosch. 1989,188,212. A. Huyhues-Despointes, V.A. Yaylayan and A. Keyhani, J. Argric. Food Chem., 1994,42,25 19. A. Keyhani and V.A. Yaylayan, J. Agric. Food Chem., 1996,44.223. K.B. de Roos. In: Flavor precursors, R. Teranishi, G.R. Takeoka and M. Guntert (eds.), American Chemical Society, 1992. 203. G.P. Rizzi. A.R. Steinle and D.R. Patton. In: Food Science and Human Nutrition, G. Charalambous (ed.), Elsevier, 1992, 73 1. J. Beck, I;. Led1 and T. Severin, Lehensm. Un/ers. Forsch.. 1989, 188, 118. F. Ledl, E. Schleicher. Angew. C‘hemie In[. Ed. Engl., 1990, 29, 565. H. Weenen and S.B. Tjan, In: Flavor Precursors, R. Teranishi, G.R. Takeoka and M. Guntert (eds.), ACS Symposium Series 490, American Chemical Society, Washington. 1992, 2 17. H. Weenen and S.B. Tjan, In: Trends in flavor research, H. Maarse and D.G. van der Heij (eds.), Elsevier Science B.V., Amsterdam, 1994, 327. H. Weenen and W. Apeldoorn, In: Flavour Science: Recent developments, A.J. Taylor and D.S. Mottram (eds.). Royal Society of Chemistry, Cambridge, 1996,211. H. Weenen, S.B. Tjan, P.J. de Valois. N. Bouter, A. Pos and H. Vonk, In: Thermally generated flavors, Maillard, microwave and extrusion processes, T1i.H. Parliment, M.J. Morello and R.J. McGorrin (eds.), ACS Symposium Series 543, America1 Chemical Society, Washington DC, 1992, 142.

Relationship between Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles A. J. Taylor and R. S . T. Linforth DEPARTMENTAL OF APPLIED BIOCHEMISTRY AND FOOD SCIENCE, UNIVERSITY OF NO’ITINGHAM, SUITON BONINGTON CAMPUS, LOUGHBOROUGH LEI 2 SRD, UK

1. INTRODUCTION

For many years, flavour scientists have studied the relationship between the flavour perceived by consumers and the qualitative and quantitative aspects of the mixture of flavour chemicals that cause perception in humans. Broadly, research has fallen into two distinct areas. The first area has studied single flavour chemicals, particularly those that produce a sensation of sweetness. The sweetening power of individual compounds has been determined by sensory analysis and then related to the amount of compound administered, to derive a quantitative relationship. These mathematical relationships are often referred to as the Psychophysical Laws.’,’ The second area has studied foods (or model systems) which contain complex flavour mixtures and attempted to derive relationshipsbetween the flavour composition and the perceived flavour of the sample. To implif) the problem, some workers have tried to quantifjl individual flavour compounds and relate the concentration of these individual compounds to flavour characteristics e.g. the gassy note of a flavour might be related to the amount of hexanal or the total amount of the Ca aldehydes and alcohols. While this approach reduces the complexity of the ?roblem, it has been criticised as being too simplistic (see for example Bootp ). Since some flavour molecules can interact with more than one sensor in the nose (or in the mouth and nose in some instances), there is some evidence to suggest that the relationship between perception and the amount of a compound administered, depends also on the concentrations of other compounds present in the flavour. Recent developments in analytical methodology have allowed the concentration of flavour volatiles to be measured in the nose during eating4 at concentrations around 10ppbv which brings the sensitivity into a range where many compounds have their odour rhreshold values and which is therefore useful to flavour scientists. Our laboratory has developed an interface for an Atmospheric Pressure Ionisation Mass Spe~trometer~’~ and this provides a measure of volatile concentration with the time of eating (typically 20 to 60 seconds). The data produced by this technique, now include a temporal dimension as concentration of the volatile changes with time. The purpose of this paper is to consider how these data might be correlated with measurements of sensory perception using simple food systems and simultaneous Time-Intensity sensory measurement combined with real time in-nose measurements of volatile concentration. Some background to the problem is

172

Flavours and Fragrances

presented initially, followed by preliminaty experimental data obtained t?om gelatin gels and then from starchlsucroseconfectionery. L.1 Psychophysical Laws

The relationship between sweetening power and concentration of the sweetening agent was first investigated last century with the results laying the foundation for psychophysics, a discipline that has been recently reviewed by Hoppe.’.’ This work has led to a number of “Laws” which relate the stimulus (S) applied (i.e. the concentration of the flavour compound) to the response (R) or perceived flavour intensity by assuming there is a mathematical relationship between R and S. The relationship proposed by Stevens (also known as the Power Law) has been widely used. Stevens Law

R=aSb

(1)

‘The exponent b can be obtained by subjecting people to a series of stimuli of different concentrations,determining the response and then plotting Ln R against Ln S. The slope of the line is b and the intercept is a. The value of b varies with experimental conditions and with the range of concentrations used according to Hoppe.’ Stevens obtained a value of 1.3 for b in the case of sweetness associated with sucrose.’ 1.2 Correlation of composition with sensory properties

The correlation of sensory properties with the composition of complex flavours is subject to the limitations explained in the Introduction. Nevertheless, several studies have derived empirical relationships by collecting data and subjecting them to various mathematical treatments. This approach ignores any possible mechanisms for olfaction, nor does it consider interaction between volatiles in a direct way although the maths may indicate such interactions. This chapter is not intended as a complete review of the literature but mention should be made of the work on Cheddar cheese flavour by Bacremont & Vickers’ as well as work on tea by Togari and coworkers.’ This latter paper is a good example of what can be achieved by correlation of volatile composition with sensory properties, Togari and co-workers analysed many samples of tea for total volatile composition imd then correlated these data with sensory attributes for the samples. They attempted to describe sensory attributes like “flowery” in terms of the mixture of compounds that correlated (positively and negatively) with that attribute. They described the rationale to their work clearly in their paper and stressed that a sufficiently wide range of samples must be analysed to obtain good quality data. An example of the results they obtained is shown below. Sweet floral = -0.059l[pentanal] -0.671[2-heptanone] + 0.562[linalooI] +0.693[2-phenylethanol]+0.713[jasminelactone] - 0.134 It is not clear how widely applicable this method is although, presumably, it would need to be validated for each food product. In particular, is the method capable of discriminating subtle differences in flavour (and their relationship to compositional changes) or is it just a crude indicator of quality? Further studies will no doubt answer these questions.

Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles

I73

Delahunty" and co-workers studied the volatile compounds present in the buccal lieadspace during eating of cheeses and attempted to correlate the amounts of volatiles present over the time of eating with the perceived quality of the cheeses. Samples were obtained over four time ranges (0 to 15; 0 to 30;0 to 45; 0 to 60 sec) which produced a cumulative trace of volatiles with time due to the experimental method employed. Delahunty reported that straight buccal headspace analysis allowed prediction of sensory attributes but time course studies improved predictive ability. This finding suggests that the time course of release does play a role in flavour perception although the relationship is not clear. 1.3 Mechanisms of olfactory perception

Recently, understanding of the structure of olfactory receptors and their arrangement both in the nose and within structures of the brain (reviewed by Shepherd at the Warwick meeting) has given the potential to model the relationship between the interaction of a volatile with the receptor and the subsequent sensory signal. This helps in the design of appropriate neural network models which could show the type of neural signal that might be produced by any given mixture of volatiles. However, this information is still some distance away from the verbal description of flavour perception on which we rely for consumer assessment of a food product. During the first session of the meeting, the problems that humans encounter when trying to describe odours by use of appropriate words were discussed. Van Toller had demonstrated in his presentation that people were more able to describe verbally, shapes or colours rather than flavours and it was suggested that human recognition of flavour might be more akin to recognition of human faces i.e. it is easy to recognise people by their faces but extremely difficult to describe verbally the features that distinguish one face from mother. This assertion suggests that the pattern of volatiles sensed by the nose (and other sensors) may be the key to recognition and, therefore, that attempts to relate sensory and compositional data should follow this route. This is especially significant when it is considered that most flavours are due to a mixture of chemicals and, while there are certain character impact compounds, they rely frequently on the presence of other compounds to produce a good flavour sensation. Booth and co-workers3 have described an interesting approach to explain the recognition of strawberry aroma. Their data suggests that recognition depends on the combination of two separate olfactory stimulation patterns and their work presents a unique approach to the whole subject of sensory recognition. 1.4 Rationale for measuring in-nose concentration

The rationale for measuring the concentration of volatiles in the expired air of humans during the consumption of food has been discussed in detail e l ~ e w h e r e ~and ~'~'~ only a brief overview is given here. Basically, there are three types of volatile extraction und analysis, each of which has its merits. Determination of the fotal volatile profile of a food is often achieved by steam distillation or solvent extraction of the food followed by GC-MS. Similarly, the volatiles above a food, the hedpace, can be determined under a range of conditions of temperature, water content etc. either at equilibrium or under dynamic conditions. However, neither of these analyses takes into account the changes that take place on eating and it has been shown that, in many foods, the volatile profile measured in expired air from the nose (nosespace or b r e d by breath) is very different

174

Flavours and Fragrances

tom the headspace profile in both qualitative and quantitative terms

.~.. 'J.'*

Current thinking

is that the volatile profile measured by nosespace analysis must be very close to the profile

experienced by the olfactory epithelium and thus, if the relationship between volatiles and perception is to be studied, then this seems the logical type of instrumental analysis to perform. 1.5 Measurement of in-nose or in-mouth concentration

Previous used trapping of volatiles on adsorbents like Tenax, followed by GC-MS analysis to record the volatiles released over periods of time during eating. Using MI-MS, with the interface developed at Nottingham, it is possible to monitor 12 to 20 ,volatiles simultaneously in real time and provide a record of the way that the concentrations of volatiles in the nose or mouth change during eating and after swallowing. In our nosespace method, the subject inserts a soft plastic tube into one nostril and air is sampled by a second capillary tube that fits into the soft plastic tube at right angles. This arrangement allows subjects to breathe normally with no great hindrance. Since the MI-MS samples tidal air from the subjects, the inspired air contains only those compounds present in the laboratory air while the expired air contains products of metabolism from the lungs (e.g. acetone) as well as volatiles released in the mouth and rransported retronasally to the nose. The traces obtained show breath-by-breath release md Figure 1 indicates the behaviour of two different volatiles (ethyl butyrate and ethanol) released from a gelatidsucrose gel as well as the presence of a liver metabolite (acetone) in the expired air of a healthy male volunteer

Ethyl Butyrate 14 01

loo

* 0

Ethanol

,100

I

*

~

1I

I_

0 1

3

~

~ 1340

1380

1380

1400

l
Time

1440

1480

14 80

Figure 1. API-MS traces showing acetone release (a metabolic product from the liver) which monitors the breathing pattern of the subject, along with the release of ethyl hutyrntr and ethanol. The volatiles were contained in a gelatinhcrose gel that was placed in the mouth at 13.2 min, chewed and then swallowed at 14.0 min.

Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles

175

Each trace shows peaks that correspond to individual inhalations and exhalations. The release of ethyl butyrate is irregular with jagged profiles present on each breath and little ethyl butyrate persists after swallowing. It is thought that ethyl butyrate shows this behaviour because it is released principally from fresh fiacture surfaces formed when the gel is chewed. Each bite is accompanied by a mouth movement that pushes a small amount of air out of the mouth and into the nose by the retronasal route. Ethanol however, is released in a more regular way. The hypothesis is that it is released from the gel into the ciqueous phase by a combination of melting and solubilisation and then enters the air phase liom the liquid phase. This also explains why it persists in the expired air after swallowing as, although the gel has been swallowed, some of the volatile is still present in mouth in the saliva phase. The data in Fig 1 can be transformed into volatile release curves by plotting the peak heights against time which yields a simpler curve that is easier to interpret. Since data collection is rapid, many replicates can be run and the data either averaged or pooled. 1.6 Objectives and limitations

To study the relationship between in-nose concentration and sensory perception, gelatin gels containing either isoamyl acetate (IAA), benzaldehyde or no volatile were prepared. Although these gels contained only one aroma volatile (and hence represented cm atypical food situation) they did allow some basic testing of hypotheses to determine what sort of correlation was appropriate for temporal data. Gels were eaten by panellists trained in the technique of sensory Time Intensity analysis and these data, as well as innose Concentration were collected and then compared using simple psychophysics and a basic damping theory. Subsequently, mint flavoured sweets were used for hrther cxperiments. 2. MATERIALS AND METHODS Gelatin gels were prepared at concentrations of 10 or 20% with different amounts (0, 0.5 or 2Yo) of Sunflower oil to give different flavour release characteristics. Aspartame was iised as the sweetener (0.0625 g/L in all samples) and benzaldehyde and isoamyl acetate were added at 8mgkg. The gelatin solution was prepared, held at 5OoC. the requisite inount of oil added and the sample homogenised to mix the phases. Samples with no oil iddition were also homogenised. Gels were cooled and stored at 4OC before use. Gel !;amples (about 0.4g) were given to trained Time Intensity panellists who recorded the intensity of flavour on a scale marked from “not at all” to “extremely” after reference to a standard gel sample. Alternatively, Polo mints (Nestle; normal and extra strong) were given to trained panellists who sucked them for one minute, then chewed and swallowed them after 2 minutes. For both gels and mints, expired air was collected from one nostril iit 50 d m i n and introduced into the API-MS (Micromass, Altrincham, UK) via the interface designed at Nottingham. By this means, simultaneous Time Intensity and breath by breath data were obtained. 3. RESULTS 3.1 Time-Intensity and breath by breath release from gelatin gels

Both Time Intensity (TI) and breath by breath (3B) analyses were normalised so that the maximum peak height for each sample was set at 100%. This step allowed easier comparison of the individual data sets but removed some of the quantitative information.

176

Flavours and Fragrances

Because the purpose of this study was to test whether TI and 3B data sets could be correlated using simple procedures, comparative data were sufficient to test the hypotheses. Both sets of data could then be analysed to determine standard TI parameters like the time at which maximum intensity (TRYX) occurred.

120

4

80

E -2 C

60

40

20

0

0

10

20

30

50

40

60

70

80

90

100

Time (sec)

Figure 2. Time-Intensity curves for one panellist consuming six dgerent gelatin gels containing only benzalakhyde. The panellist correctly identified a hidden control containing no benzaldehyde

0

10

20

30

40

50

60

70

80

90

100

Time (sec)

Figure 3. Breath by breath release of benzaldehyakfrom the six gelatin gels. Data were obtained simultaneouslywith the Time Intensiv data. The peak height of each breath has heen ploned to show the release profile with time.

111

Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles

The normalised TI traces obtained by one panellist from six different samples containing benzaldehyde are shown in Figure 2. The gel composition was varied so it might provide different rates of volatile delivery and, although there are some differences in the profiles in Fig. 2, they do not follow a logical sequence i.e. there is no pattern of changing volatile release with changing gel properties. The relative release profiles for benzaldehyde in the nose of the panellist for the same samples are shown in Fig 3 and, it is evident that they are also very similar. This supports the idea that the pattern of release in Fig 2 is not actually different, despite differences in gel composition. These two sets of data are directly comparable as they were obtained simultaneously. The same procedure was carried out for gels containing isoarnyl acetate instead of benzaldehyde with similar results (data not shown). 3.2 Relating Time Intensity and breath by breath concentration with psychophysics

Psychophysical laws have previously been applied by measuring the sensory response to samples administered at different times, separated by periods of many minutes. To test whether psychophysics could relate the TI and breath by breath (3B) data obtained over the short periods experienced during eating, the traces for a single gel sample containing benzaldehyde were plotted together and the profiles compared. The breath by breath trace showed a rapid increase in concentration to a maximum value with a slower (but still fairly rapid) decline. The TI trace showed an initial increase that was very similar to the 3B trace but then decayed at a much slower rate. Examination of all six pairs of TI and 3B traces showed similar patterns.

0

10

20

30

40

50

60

70

80

90

100

Time (Sec)

Figure 4 Simultaneous Time Intensiv PI) and breath by breath (3B) tracesfor a single (relatingel containing benzalakw

If psychophysical principles are applied to the data in Fig 4, then it is clear that, for the same stimulus intensity, there can be different sensory intensities depending on the time at which the stimuli were applied. For instance, a stimulus of 40% evoked a 40% sensory perception at 8 seconds but a sensory perception around 85% at 20 seconds. Normally,

I78

Flavours and Fragrances

psychophysics are only applied to the maximum intensity values. Initially, this suggested that psychophysics was inappropriate for data which spanned both sides of the maximum intensity values. However, the traces in Fig 4 suggest that perception depends not only on the stimulus but also on the previous history of stimulation. A simple way to express this previous history is to express the data on a cumulative basis, a method that had proved usefbl in relating volatiles in-mouth to the sensory qualities of cheese.16 By adding each point to the previous point, the data are not only smoothed, but the position of any one point depends on the pattern of previous points. 'The data from Fig 4 were replotted on a cumulative scale in Fig 5

0

10

20

30

40

50

en

70

en

m

r

m

(-1

Figure 5. Cumulative traces obtained simultaneously for Time lntensity and breath by breath releasefrom a gelatin sucrose gel containing isoamyl acetaie

To determine whether psychophysical laws could link these curves (Fig 5), two methods were used. The first assumed Steven's Power law and plotted the natural logarithm of the Stimulus (3B values) against the natural logarithm of the Response (TI values) which should yield a straight line, the slope of which is b (see equation (1)). .4lthough the initial part of the curve was linear, the slope varied between 0.61 and 1.14 for the six samples (mean value 0.81 SD 0.21) and later points exhibited non-linear behaviour. It has been reported that Steven's Law is sometimes applicable only over a particular concentration range. Similarly if the person assessing the flavour becomes habituated to a particular flavour after a period of time, it is unlikely that Stevens Law will be applicable past that time. Psychophysics has been used mainly for taste perception measurements where habituation is not so rapid compared to olfactory perception. A plot of LnS versus Cumulative TI (R) also failed to deliver a straight line suggesting that Fechners law was also not applicable, The second method used the Solver fbnction in Microsoft Excel to obtain the best lit of the data from which values of the exponent (b) were obtained for all six samples. 'Table1 shows the values obtained for the six gels containing isoamyl acetate and benzaldehyde. The values in Table 1 were difficult to interpret as the amount of variation that might be expected in experiments like these was not easy to estimate. There was also the question as to whether the low values ( for example 1.13 and 1.47) should be discarded iis outliers, either due to poor fit of the two sets of data or due to variation in analysis. A

179

Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles

check on the quality of fit for all six curves showed that the outlying values were not associated with poor fit quality. The conclusion was that the psychophysical laws were difficult to apply to this type of data and did not provide consistent results from the cumulative traces. Table 1. Valuesfor the exponent b in Stevens Law (R=as6) where R is the perception und S the stimulus.

Sample 1

Sample 2

Sample 3

Sample 4

Sample 5

Sample 6

Compound

Mean f SD

Isoamyl acetate Benzaldehyde

1.47 2.71

2.30 2.57

2.60 1.13

2.55 2.47

1.76 1.63

2.59

2.2 1

2.37

(0.48) 2.15 (0.63)

.3.3 Relating Time Intensity and breath by breath traces using damping theory

Since the psychophysical results were inconclusive, an alternative approach was considered. The basis for this approach was that the TI curve followed the 3B trace as breath concentration increased but, thereafter, perception decreased at a slower rate than the 3B curve, possibly due to the difference in the binding and dissociation constants of the volatiles to the sensors. This behaviour is analogous to the situation when a car wheel hits a hole in the road. Initially, the wheel moves at a speed dependent on how hard the wheel hit the hole (the compression stimulus) but the dampers or shock absorbers then come into play and slow the return rate of the wheel to the road. There are various degrees of complexity in describing damping theory with the ability to adjust both compression and rebound damping. In this case, compression damping seemed to be minimal and rebound damping was the major effect. Using a simple damping equation, the 3B data were taken and matched to the actual TI curve using the Solver knction in Microsoft Excel. A typical set of traces is shown in Fig 6. Curves were fitted for all six samples and the values for K1 to JS3 calculated (Table 2). The knction used is shown below.

K1 x [TI +(K2 x previous TI)] +K3

(2)

Table 2. Valuesfor the three constants 6 1 , K2 and K3) in the simple damping theory used tofit the 3B traces to the TI curvesfor the six gelatin samples with isoamyI acetate

Sample #

9

10

11

12

13

16

K1

0.71 0.45 6.1

0.42 1.82 -0.17

0.25 3.64 -2.24

0.35 2.42 -1.27

0.45

0.34 2.49 -1.1

1(2 1<3

-

1S O

1.91

Flavours and Fragrances

180

100

I

J+36 actual -=-Actual TI +Theoretical

80

TI d a m e

40

20

0 0.

20

40

60

80

100

Time (see)

Figure 6. Actual breath release and Time Intensity curves for isoamyl acetate in a gelatin gel with the curve calculated using simple damping theory. Ahhouj$ the theoretical curves generated, fitted the observed data well, the K values obtained showed no logical trend or pattern. This may be because the model used was too simple and a more elegant pharmacokinetic model might be more appropriate. However, these preliminary experiments were only designed to test the obvious, simple hypotheses before considering more complex solutions. Neither Stevens Law nor simple damping theory could provide an obvious relationship between the TI and 3B data. One potential explanation is that there is too much variability in the data collected in these preliminary experiments to see clear patterns. For instance, did the panellists produce consistent TI data under unfamiliar laboratory conditions and what is the variability of the concentration in-nose due to different Ixeathing and chewing patterns? Secondly, flavour is usually perceived and judged by people over periods of minutes, whereas here, the changes were being followed on a second by second basis and it is not clear whether humans can discriminate odours this quickly. Humans can certainly sense odours rapidly but perhaps quantitative assessment is a slower process. The work described here shows clearly the difficulty of designing suitable experiments to investigate this problem. 3.4 Relationship between menthol concentrations in-nose and perceived flavour intensity

Much of our previous work on volatile release in-mouth has used mint flavours as they are readily available in a variety of commercial confectionery products that provide different rates of flavour release. Mint flavour is readily recognised by panellists and the confectionery used was more acceptable than the single volatile gelatin gels. However, some caution needs to be exercised as mint flavour is due to a combination of menthol, rnenthone (plus other compounds) and ideally all components contributing to mint flavour should be measured. Menthol also interacts with trigeminal and olfactory receptors which may confuse the perceptual signal produced. Despite these limitations, preliminary

181

Sensory Time-Intensity Measurements and In-nose Concentration of Volatiles

experiments were carried out. Mint flavoured sweets (Polos; two different strengths) were given to Six panellists who ate them while recording Time Intensity data. Air from the nose was sampled simultaneously and the amounts of menthol calculated by comparing peak heights to a calibration curve. Since Power law relationships seemed to be inappropriate for temporal data, the TI curves were analysed to determine the maximum Intensity recorded by each panellist and the intensity after 1 minute in mouth for each of the two strength mint sweets. The relevant menthol concentration was determined for each of these sensory values and the data plotted (Fig. 7). 1600

1400

1200

1000

;

.B

p 'E

2 600

400

200

0 0

1ooO

2000

3Ooo

4ooo

5ooo

Menthol breath concentration (ppbv)

Figure 7 . Plot of in-nose menthol concentration against perceived mint flavour intensity for 6paneIIists. Sensory and in-nose &ta were collected simultaneously

For all panellists, there was a clear linear relationship between the concentration of menthol in-nose at a particular time and the intensity of mint flavour perceived at that time, although each panellist showed a different response. It is interesting that the relationship k r these data is linear and shows much greater consistency than the data obtained by psychophysical and damping theory calculations.

I82

Flavours and Fragrances

4. CONCLUSION These preliminary experiments have attempted to correlate in-nose concentration with sensory perception using different data analyses. The ability t o measure in-nose concentrations during eating allows us to study whether psychophysics, damping theory o r other models are appropriate but further, more refined work, is needed t o make progress on this long standing and difficult problem.

Acknowledgements The assistance of Dr Wendy Brown (Institute of Food Research, Reading UK) in supplying the trained TI panellists and the gelatin gels is gratefully acknowledged. Research into flavour release at Nottingham University is funded by the Biotechnology and Biological Sciences Research Council, the Ministry of Agriculture, Fisheries and Food under the LIM< initiative and by a consortium of companies including Firmenich, Micromass and Stable Micro Systems.

References 1. K. Hoppe and J.H.A. Kroeze in “Aroma: Perception. Formation. Evaluation” Proceedings of the 4th Wartburg Aroma Symposium, Eds.M. Rothe & H-P Kruse, Eigenverlag Deutsches lnstitut Fuer Ernaehrungforschung 1995. 2 . K. Hoppe in Proceedings ofthe 5th Wartburg Aroma Symposium Eds.M. Rothe & H-P Kruse, in press, 1997 3. D.A. Booth, R.P.J. Freeman and M.S. Kendal-Reed, in “Aroma: Perception, Formation, Evaluation” Proceedings of the 4th Wartburg Aroma Symposium, Eds.M. Rothe & H-P Kruse, Eigenverlag Deutsches Institut Fuer Ernaehrungforschung 1995 1. R.S.T. Linforth, K.E. lngham and A.J. Taylor in “Flavour Science: Recent Developments”, Royal Society of Chemistry, pp36 1-368, ISBN 0 85404 702 6 5 , R.S.T. Linforth and A.J. Taylor UKPatent Application 9615303.6 5. A.J. Taylor and R.S.T. Linforth UK Patent Application 9615304.4 7 . S . S . Stevens, Percept. Psychophys., 1969, 6 , 302. 8. C. Dacremont and Z Vickers, J Food Sci., 1994, 59, 981. ‘9. N. Togari, A. Kobayahi and T Aishima, Food Research In/,, 1995, 28, 485 10. A.J. Taylor, 1996, Crit. Rev. in FoodSci Nutr. 36, 765 I1. A.J. Taylor and R.S.T. Linforth, 1996, T r e d in Food sci. Technol. 7 , 444. 12. AJ Taylor and RST Linforth in Proceedings of the 5“ Wartburg Aroma Symposium, Eds M Rothe and H-P Kruse, in press, 1997 13. K.E. lngham, R.S.T. Linforth and A.J. Taylor, 1995, Flavour Fragrance J 10, 15. 14. K.E. lngham, A.J. Taylor, F.F.V. Chevance and L.J. Farmer in “Flavour Science: Recent Developments”, Royal Society of Chemistry, pp 386-391; ISBN 0 85404 702 6 15. R.S.T. Linforth and A.J. Taylor, 1993, Food Chem. 48, 1 15 16. C.M. Delahunty, J.R. Piggott, J.M. Comer and A Paterson, J. Sci Food Agric. 1996, 71, 213. 17. J.P. Roozen, and A. Legger-Huysman, in Aroma perception, formation and evaluation, Procedngs of the 4th Wartburg Aroma Symposium, M. Rothe and H-P. Kruse. Eds. PotsdamRehbrucke, Deutsches lnstitut her Emahrungsfomhung, 1995, 627.

Organic Chemistry

Synthesis and Odour Properties of Chiral Fragrance Chemicals Tetsuro Yamasaki CENTRAL RESEARCH LABORATORY. TAKASAGO INTERNATIONAL CORPORATION, 1-4-11, NISHI-YAWATA, HIRATSUKA-CITY, KANAGAWA, 254, JAPAN

1 INTRODUCTION Olfactory receptors belong to the rhodopsin superfamily, i.e., they are peptides consisting of L-amino acids. Thus, fragrance science frequently deals with the phenomenon of chirality. The character and intensity of fragrance chemicals often relates to their stereochemistry. Stereoisomers differ in their qualitative odour characteristics as well as in their quantitative odour properties. The optical purity and diastereomeric composition of a diastereomeric mixture of chiral fragrance chemicals can also influence the odour greatly. Enantiodifferentiationhas been recognised as an important principle of odour perception. Many fragrance chemicals are chiral, i.e. they contain one or more asymmetric carbon atoms and exhibit optical activity. However, it is often difficult to obtain large quantities of optically pure chiral fragrance materials. Synthetic chemicals are often racemic, or have a specific ratio of stereoisomers that is related to either the optical purity of the starting material, or the stereoselectivity of the reaction. Materials of natural origin generally have a characteristic distribution of stereoisomers due to their stereoselective biogenetic formation mechanisms. The growing awareness of the importance of chirality in fragrance chemistry has resulted in considerable efforts to develop asymmetric syntheses of chiral fragrance chemicals.

2 ASYMMETRIC SYNTHESIS OF L-MENTHOL In the course of developing chiral flavour and fragrance chemicals on an industrial scale, the process for the asymmetric synthesis of Lmenthol(9) has been established by use of the optically active Rb-BINAP catalyst (Diagram 1)'. The I-menthol process is currently running at over 1000 tons per annum. The key step in the synthesis of I-menthol (9) is the Rh-(5')-BINAP catalysed asymmetric isomerization of diethygeranylamine (1) to (R)-citronella1diethylenamine (3). In the asymmetric isomerization of allylic mines to enamines, there is an interesting correlation between the chirality of the B N A P ligand, the geometric configuration of the starting dlylic amines, and the configuration at C-3in the enamine product (Scheme 1)2. Treatment of trans-allylic amines with Rh-Q-BINAP yields almost enantiomerically pure

186

Flavours and Fragrances

(S)-BINAP

(R)-BINAP

Diagram 1 (R)-enamines. The same reaction with Rh-(R)-BINAP gives (5')-enamines. For example, diethylgeranylamine (1) with Rh-(5')-BINAP gives (R)-(+)-citronella1 diethylenamine (3), and Rh-(R)-BINAP gives (5')-(-)-citronella1diethylenamine (4),respectively. On the other hand, the corresponding cis-allylic amines with the same pair of catalysts yields a pair of the antipodes. For example, diethylnerylamine ( 2 ) with Rh-(9-BINAP gives the (5')-(-)isomer (4),and 2 with Rh-(R)-BINAP gives the (R)-(+)-antipode (3), respectively. The C6-C7 double bond remained intact to the conditions of the asymmetric isomerization reaction. The optically active enamines 3 and 4, can be converted to (R)-(+)-citronella1 (5) and (5')-(-)-citronella1 (6) respectively, by hydrolysis under mild aqueous acid conditions. It is of importance that each enantiomer of the citronellal produced in this manner is of a higher enantiomeric purity (typically 98-99 % e.e.) than the equivalent substance obtained from a natural source. The enantiomeric purity of natural (R)-(+)-citronella1 (5) is at best 80% e.e. and for the (8-(-)-antipode (6) at best 89% e.e.3 A series of enantiomeric pairs of citronella1 derivatives have been prepared from optically pure 5 and 6 , and their olfactory properties have been compared4. Generally, the molecules prepared from 6 showed better odour characteristics. Cyclization of (R)-(+)-citronellal(5) by the carbonyl ene reaction yields I-isopulegol (8) with a diastereomeric excess of greater than 98% (Scheme 2). The C-3 methyl stereocenter in 5 guided the stereochemical course

. Rh-(S)-BINAP

1 Diethylgeranylamine

(R)-(+)-Citronella1 diethylenamine

5 (R)-(+)Citmnellal

Rh-(S)-BINAP

2 Diethylnerylamine

4 (S)-(-)-Citronella1 diethylenamine

Scheme 1

6 (S)-(-)-Citronella1

187

Synthesis and Odour Properties of Chiral Fragrance Chemicals

5

0 I-lsopulegol

(R)Citronellal

i

A 9 /Menthol

Scheme 2

via the chair transition state intermediate 7 in which the C-3 methyl group, the C-1 aldehyde carbonyl group, and the C6-C7 double bond are all equatorial. The result is the formation of I-isopulegol(8) with all of the ring substituents equatorial in orientation. Hydrogenation of the double bond in 8 completes the synthesis of I-menthol (9). 3 CHEMISTRY AND ODOUR PROPERTIES (2S, 4R)-Rose oxide (I-cis-rose oxide)

The diastereomeric mixture of rose oxide is assumed to be responsible for the floral green top note of Bulgarian rose oil. Rose oxide has four stereoisomers. The chromatographic separation of them with the use of multi-dimensional GC, elucidated that the (4R)diastereomers occur more abundantly in nature, and also that (2S,4R)-rose oxide is responsible for the odour characteristics. The diastereomer composition of rose oxide influences the odour greatly. The composition of the (2S,4R)- vs (2R,4R)-diastereomers in Bulgarian rose oil is 75 : 25, which is the best in nature. Mixtures of the diastereomeric rose oxide's have been identified in white wine, cognac, grapes, tea oil, rose oil, citronella oil, geranium oil, and tropical fruit juices3. The stereoselective synthesis of all four stereoisomers of rose oxide has been achieved starting from optically active citronellol (10) (Scheme 3f. Stereochemistry at the 4-methyl group was controlled by the chirality of the 3-methyl group in citronellol(l0). For example, (4R)-rose oxide is obtained from (5')10 and the (4S)mtipode from (R)-10, respectively. Relative stereochemistry at the 2position is controlled by the choice of catalyst for the cyclization of 13. Acid catalysed cyclization of (9-13 gives (2S,4R)-rose oxide (I-cisrose oxide) (14) with a 95 % cisselectivity, where as Pd-(9-BINAP gives a 1:1 mixture of (2R,4R)-isomer and 14. The (2R,4R)-rose oxide (I-rrans-rose oxide) is obtained by column chromatographic separation of this mixture. Similarly, (2R,4S) and (2S,4S)-rose oxide (d-cis and d-trans rose oxide,

188

poH opoH Hop Flavours and Fragrances

CH,COOOH~

!

i) AI(~-P~OL

OH

Cy-10 (S)-Citronellol

(S)-12

(S)-11

-Po" I

Pd(OAc)*

Catalyst

Cntalvst Pd-( R )-EINAP Pd-( S )-BINAP

(S)-13

H

(2s) 7 0 / /(2 30 R 53 / 47 9515

)

6 - 1 14 (2S,4R)-Rose oxide

Scheme 3 respectively) have been prepared from (R)-10. (2S,4R)-rose oxide (14) has an odour threshold value of 1400 to 1/300 of the other isomers (Table 1). The odour character of

Table 1 Odour of the stereoisomers of rose oxide Configuration (2S, 4R) (2R, 4 4 (2R, 4s) (2S, 4s)

Threshold* Odour Description 0.5 Floral-green; clean, light, rose, green, diffusive, strong 160 Floral-green; green herbal (minty), fruity 50 Herbal-green-floral; hay green, earthy, heavy 80 Herbal-green-floral; fruity, herbal rose, citrus (bitter peel) * Odour threshold values Lppb][WN]

14 is greatly superior over the other isomers in terms of both quality and quantity. (2S,4R)-rose oxide (14) is also biodegradable. In a biodegradability test, 14 readily disappears, the (2R,4R)-isomer slowly decreases but still remains after 4 weeks, and the (4S)-isomers are not digested at all. It is interesting that 14 has better odour properties and higher safety potentials.

The head space gas chromatographic analysis elucidated that the odour of white cyclamen (Cyclamenpurpuruscens) is attributable to (3S)-(6E)-dihydrofarnesol(I 6). The odour character of natural 16 is floral green,white cyclamen, muguet, lily, fresh, clean and much better than that of the unnatural (3R)-(6E)-enantiomer (17), which has balsamic nuance.

189

Synthesis and Odour Properties of Chiral Fragrance Chemicals

16 (ZE.6E)iarnesol

16 (6E)-(3S) :Floral green, whlte cyclamen, muguet. Illy, fresh, clean

17 (6E)43R) : Weak, balsamic, floral,

Scheme 4

Ru (R)-BINAP catalysed asymmetric hydrogenation of all-trans-famesol(l5) gives 16, and hydrogenation with (Q-BMAP gives 17, respectively (Scheme 4). The remote double bonds at the C6 and C10 positions remain intact during the asymmetric hydrogenation. Compounds 16 and 17 have a similar range of antibacterial activity against a considerably wide range of micro-organisms, which includes body malodour related micro-organisms (Table 2). Compound 16 is not as active as antibacterial pharmaceuticals, but is active enough under the concentration of fragrance composition, and inhibits the growth of bacteria which are responsible for body malodour. Table 2 Antimicrobial activity of dihydrofamesol isomers [mcg/M 11

Compound (6E)-(3S)-Dihydrofamesol (6@-(3R)-Dihydrofmesol Test micro-organisms;

An 50

Pa Sa Cm Bs Pac 25 25 25 25 <8 50 25 25 25 25 <8 An :Aspergillus niger Pa :Pseudomonas aeruginosa Sa :Streptococcus aureus Cm : Corynebacreriurnminutissirnum Bs :Bacillus subtilis Pac : Propionibacteriurnacnes Mf :Malasseziafirfir

Mf <400 <400

(R)-Styrallyl acetate

The catalytic hydrogenation of acetophenone 18 (Rl=H, R2=CH3) yields the racemic styrally alcohol 19 (Rl=Me, R2=CH3). Esterification of the racemate 19 gives racemic styrallyl esters 20 (Rl=H, R2=Me/Et, acetate R3=CH3, propionate R3=C2H5, and isobutyrate R3=i-C3H7). These are widely used as floral green raw materials reminiscent of gardenia or strawberry (Scheme 5). Noyori et a t . have reported that the asymmetric hydrogenation of 18 (R 1=H, ==Me) with the Ru-BINAP-ethylenediamine-KOH-isopropanolsystem gave the optically active styrallyl alcohol 19 in 87-94% e.e.

I90

Flavours and Fragrances

RpJJR*

OH t

KOH

ta

P i

'Ph

RTRz / 19

84-94%ee

OCO R, Rl

RsCOOH H*

20

Scheme 5 Table 3 Odour of enantiomeric styrallyl esters Common odour description

R2

R1 H

Me

R3 Me

H

Me

Et

H

Me

n-Pr

H

Me

i-Pr

H H

Me Me

n-Bu i-Bu

H 4-Me

Et Me

Me Me

4-Me 4-Me

Me Me

Et n-Pr

4-Me

Me

i-Pr

4-Me0

Me

Me

Floral, green, reminiscent of gardenia, fresh, herbal Floral green, reminiscent of gardenia, strawberry, apple Floral fruity, apple, pineapple, strawberry Floral green, reminiscent of strawberry, pineapple, herbal Floral fruity, strawberry Floral fruity, reminiscent of strawbemy Floral (rose) with green (twig) Floral (rose) with fresh green Floral (rose) with green, fruity Floral (rose) with strawberry, pear character Floral (rose) with strawberry, pineapple, green Floral (rose, lilac) with balsam

Difference of tone of enantiomers (R) (s) Fruity, green Sharp, metallic

Fruity

Thin, light, metallic Fatty, green

Light, fruity

Green

Heavy Fruity

Fruity, fatty Sharp, green

Metallic, bitter Light, fresh, fruity Pineapple Honey

Flat Metallic

Fruity

Balsamic, fatty

Flat

Flat

Clean, green

Peach Fatty

The alkylphenones 18 were asymmetrically hydrogenated to yield compounds 19. Subsequent esterification of 19, yielded pairs of the enantiomeric styrallyl ester analogs 20. Differences in the odour characteristics between the enantiomers of styrallyl esters (Rl=H, ==Me) are not very great. For instance, they belong to the same category in odour classification, but differ from each other in odour characteristic tones and nuances (Table 3). Generally speaking, the (R)-enantiomers resemble the racemates with more fruity, clean, and light nuances. On the other hand, the (9-enantiomers have more green, sharp, and metallic nuances, and are somewhat different from the racemates.

191

Synthesis and Odour Properties of Chiral Fragrance Chemicals

(R)-Matsu takeol Matsutake, (Tricholoma matsutake), is an alluring and expensive mushroom in Japan. Japanese people are very fond of it not only for its taste but also for its scent. The odour of matsutake is assumed to be due to (R)-matsutakeol(24). The enantiomeric mixture of (R) and (3-24 has been identified in a large number of mushrooms and each has a characteristic distribution of enantiomers. Optical distribution of (R) and (S)-24 in matsutake is 94 : 6, and in the well known truMe (Tuber melanosporum) is 96 :4'. Utilising asymmetric hydrogenation with Ru-BINAP catalyst as the key step for introducing chirality in the molecule, both enantiomers of optically active matsutakeol (24) have been prepared in 99 % e.e. (Scheme 6). The enantiomerically pure (R)-(-)matsutakeol(24) has the typical odour of the matsutake mushroom, but the (3-(+)antipode has a weak, dirty and somewhat green mushroom odour (Table 4).

HZIRU-BINAP

L C O O M e CSHll

21

CSHll L C O O M e

99Xee

22

L CSH11 L 23

s m -

C d l l 24

Matsutakeol

Scheme 6 Table .4 Odour characteristics of the enantiomers of matsutakeol Configuration (34 (39

Threshold* Odour description 0.01 Strong mushroom like Herbaceous, musty 0.1 * Odour threshold value [ppm] [ W N ]

(5')y-Lactones Enantiomeric mixtures of y-lactones (27) having C5, C6 and C7 side chains are found in a large number of fruits, for example, peach, apricot, mango, passion fruit and strawbeny, each with a characteristic distribution of enantiomers*. The (R)-isomers are always present more abundantly in nature than the corresponding (S)-isomers. Both enantiomers of y-lactones (27) have been prepared in 99 % e.e. via the corresponding optically active y-hydroxyesters (26). These were in turn prepared by the asymmetric hydrogenation of the corresponding y-ketoesters (25) (Scheme 7)9. Preliminary odour evaluation in water showed that both enantiomers belong to the same category and have a similar value as fragrance chemicals (Table 5). Evaluation in flavour bases elucidated that they have different characteristics. For example, (9-27 has a light and sweet top note which has freshness, and thus may be preferable in fruity bases, The (R)-isomers have strong, thick, and heavy tones which may be suitable for different purposes. As (3-27 are minor components in nature, one could envisage having the

192

Flavours and Fragrances

possibility to improve the odour properties of fruity bases with the use of optically pure (9-27.

]

0

OR1

0

25

H' ___)

RP

26

27

Scheme 7 Table 5 Odour characteristics of enantiomers of y-lactones

CSHll C6H13 C7H15

50 2 10 * Odour threshold value [ppb] [ W N ]

0

0 -&R HZIRU-BINAP

50 2 5

howR MCPBA

-

95 97Xee

29

20

30

Scheme 8 Table 6 Odour characteristics of enantiomers of &lactones

Cyclopentanone R C4H9 C6H 13

(9* 30 50

(R)* (9* 100 30 500 50 * Odour threshold value [ppb] [ W N ]

&Lactone (R)* 10 50

(.!+&Lactones

Asymmetric hydrogenation of 2-alkylidenecyclopentanones(28) gives optically active 2-alkylcyclopentanones (29) in 95-97 % e.e. (Scheme 8)"). 2-Pentylcyclopentanone (R=C4H9) and 2-heptylcyclopentanone (C6H13) have a jasmine like floral odour. Odour strength and odour properties of the (S)-2-alkylcyclopentones (29) are superior to those of the corresponding (R)-enantiomers (Table 6). The Baeyer-Villiger oxidation of 29 gives b-lactones (30) with a retention of chirality. &Lactones (30) have a creamy and coconut like odour. As in the case of the y-lactones, evaluation in flavour bases showed that (9-30 has a cleaner and lighter tone and it would be suitable for fruity bases. Again, the corresponding (R)-isomers were heavy with fatty notes, and had bodies that may work better in milk, butter and roast flavours.

193

Synthesis and Odour Properties of Chiral Fragrance Chemicals

Super Santalex Sandalwood oil is a very important essential oil. However, production of sandalwood oil is decreasing due to the restricted felling of the sandalwood tree. The restricted felling is due to increasing environmental conservation pressures. Thus the scarcity of sandalwood oil has resulted in considerable efforts to develop new sandal chemicals. In the 1940's, it was found that the reaction of camphene (3 1) and guaiacol (32) with an acidic catalyst such as boron trifluoride, followed by hydrogenation with b e y - N i at 180°C under 120-180 atm gave a viscous oil which had a strong sandalwood oil odour (Scheme 9). This oil (Santalex T T M ;Takasago) is a mixture of many components. In 1964, Demole" clarified that 3-rrans-isocamphylcyclohexanol(34) is the principal odour of Santalex TTM. The stereoselectivesynthesis of (34) was disclosed in 1980 by G. K. Lange and coworkers. This was achieved by reducing the 3-keto derivative (38) (Scheme 10) with a bulky alkyl borohydride'*. However, this reagent is too expensive for production on an industrial scale. Analysis by GC sniffing coupled with other structure determination elucidated that Santalex 7contains only 8 % of 3-rrans-isocamphylcyclohexanol(34),and also that the biggest peak in the GC spectrum is the odourless 4-trans isomer (37). Although also odourless, we paid attention to the second biggest peak, that of 3-cisisocamphylcyclohexanol(35) with the intention of turning it into 34. Stereoselectivehydrogenationof 38 with a Ru-triphenylphosphineethylenediamine-KOH-isopropanolcatalyst systemI3yields 34 with a stereoselectivityof 95%. By utilising this reaction, the content of 34 in Santalex TTM has been increased. Catalytic dehydrogenation of Santalex FMover a Copper-Chromiumcatalyst converted

-

31 CamPhene

3-trans :Strong sandal

32 Guaiacol

35

33 36 OH 4 4 s : Odourless

I+

37 4-trans : Odourless

Others

Scheme 9

I94

Flavours and Fragrances

OH

34 3-trans : Strong sandal

Copperchromate

0

-Y 38 H ~/R U C I~[PP~?], Ethylenediamine KOHllPA

35 3-15s: Odourless

34 3-trans : Strong sandal

Scheme 10 both 34 and 35 to 38. Stereoselective hydrogenation of the mixture with the above mentioned catalyst system gave Super Santalex. The level of the principal odour component 34 is increased from .8% in Santalex T T M to 24% in Super Santalex. Super Santalex also has a more diffusive and much cleaner odour properties than Santalex 'FM.

4 DISCUSSIONS There are two major merits with chiral fragrance chemicals. One of them is the improvement of the odour properties. In general, a specific stereoisomer of a chiral fragrance chemical will have better olfactory properties than the other stereoisomers in terms of their qualitative odour characteristics as well as their quantitative odour properties. The optically purer compounds compounds may have much enhanced olfactory properties when compared to the less pure compounds or the racemates. On the other hand, optical purity may be proportional to the production cost. From an economical standpoint the ratio of odour characteristics against price should be always taken in account. Another merit of chird fragrance chemicals is an improvement of safety potentials. Asymmetric synthesis can produce nature identical chiral fragrance chemicals, which are often considered to be safe. In another case, a specific stereoisomer is responsible for the odour properties of the stereoisomeric mixture, and the other isomers are essentially odourless oi' only have a decreasing value of the odour character (e.g., I-cis-rose oxide, (3S)-(6E)-dihydrofamesol, (R)-matsutakeol, trans-3-isocarnphyIcyclohexanol,I-muscone, etc). In such cases, one can reduce environmental pollution with the use of a small amount of the pure stereoisomer. In many cases, stereoisomers have similar toxicological properties and thus reduction of the amount used may improve safety potentials. In some cases, a specific stereoisomer may have a higher safety potential when in comparison with the other isomers, an example of this is I-cis-rose oxide which is biodegradable but the

Synthesis and Odour Properties of Chiral Fragrance Chemicals

I95

corresponding d-isomers are not, Safety and environmental issues will be more important in the future. Demerit of optically active fragrance chemicals is obvious, they are more expensive than the corresponding racemates. An improvement of synthetic technology is indispensable for future reductions of price. Although, BINAP is a very powerful tool for the industrial production of chiral fragrance chemicals, it is necessary to add new tools for asymmetric synthesis, such as chiral catalysts for reactions such as the Aldol condensation, Diels-Alder reaction, and so on. Improvements in the chemistry of asymmetric synthesis as well as development of production technology will be the key to the future of chiral fragrance chemicals. Acknowledgements

I would like to thank Mr. T. Yamarnoto and co-workers for the synthesis of chiral fragrance chemicals, and Mr. M. Emura for his analytical support and usehl suggestions. The author also would like to express his thankful gratitude to Dr. H. Kumobayashi and Dr. N. Sayo for the useful suggestions and discussions in the BINAP chemistry and Dr. H. Tsuruta and Dr. T. Kanisawa for their continuous support and useful discussions. References I. 2.

3. 4. 5.

6. 7 8. 9. 10. 1 I.

12.

13.

Recent reviews: K. C. Nicolau and E. J. Sorensen, 'Classics in Total Synthesis', VCH, Weinheim, 1996, pp. 343-379. a) K. Tani, T. Yamagata, S. Otsuka, S. Akutagawa, H. Kumobayashi, T. Taketomi, H. Takaya, A. Miyashita, and R. Noyori, J. Chem. Soc., Chem. Commun., 1982,600. b) K. Tani, T Yamagata, S. Akutagawa, H. Kumobayashi, T. Taketomi, H. Takaya, A. Miyashita, R. Noyori and S. Otsuka, J. Am. Chem. Soc., 1984,106,5208. P . Werkhoff, S. Brennecke, W. Bretschneider, M. Gijntert, R. Hopp and HSurburg, Proceedings of 12th. International Congress of Flavours, Fragrances, and Essential Oils (Post congress volume), October 4-8, 1992, Vienna, Austria, p. 64. T. Yamamoto, H. Matsuda, T. Ohmoto, A. Shimada and T. Sato, Proceedings of 12th. International Congress of Flavours, Fragrances, and Essential Oils, October 4-8, 1992, Vienna, Austria, p.442. H. Matsuda, H. Yamamoto and T. Kanisawa, Proceedings of 13th. International Congress of Flavours, Fragrances, and Essential Oils, October 15- 19, 1995, Istanbul, Turkey, p. 8 5. T. Ohkuma, H. Ooka, S. Hashiguchi, T. lkariya and R. Noyori, J. Am. Chem. SOC.,1995, 11,7,2675. Unpublished data. A.Bernreuther, N. Christoph and P. Schreier, J. Chromutog., 1989,481,363. T. Ohkuma, M. Kitamura and R. Noyori, Tetrahedron Lett., 1990,31,5509. T. Ohta, T.Miyake, N. Seido, H. Kumobayashi, S. Akutagawa and H. Takaya, Tetrahedron Lett. 1992,33,635. a) E. Demote, Helv. Chim. Acfu., 1964,47, 1766. b) E. Demote, Hefv. Chim. Acfa., 1969, 57,2065. G,K. Lange and K. A. Bauer, Proceedings of Fragrance Flavor Subst., Proc. Int. Haarman Reimer Symp., 1980, p. 1 1 1. T. Ohkuma, H. Ooka, T. lkariya and R. Noyori, J. Am.Chem. SOC.,1995,117, 10417.

In Search of Nascent Musks... Or Not! Walter C. Frank UNION CAMP TECHNOLOGY CENTRE, PO BOX 3301, PRINCETON, NJ 08543-3301, USA

1 ABSTRACT The musk aroma has been valued by perfumers for what seems an eternity. Its characteristic odor is one of the few that can stand alone, unadorned by supportive scents that are required to round out the wide variety of other odor types. In addition, the substantivity and stability of most musks have made them a sought after material for performance based products. In the past one hundred years, a veritable army of researchers from all of the major fragrance houses have attempted to identify the structural sub-units required to provide the ultimate musk aroma chemical. These studies have encompassed everything from the strictly empirical to the QSAR approach and everything in between. The on-going requirement for even higher level performance based aroma chemicals has made the family of musks the most studied of all aroma chemicals, as customers continue to push for better cost performance in finished fragrances. Early and serendipitous empirical studies provided the nitro musks, a cost effective and powerful group of musks. The discovery of these materials fuelled research in the 1960's and 1970's which led to a new class of synthetic musks, the polycyclics. As nitro musks fell into disfavor, polycyclics became the cost effective workhorse in performance based products, and made inroads into the fine fragrance market as well. While an extensive effort was carried out in macrocyclic musks, their high cost of manufacture predicated their only significant use within fine fragrances. With the advent of molecular modeling software and more powerful computers, structure activity and conformational analysis were blended with empirical methods to identify the latest round of super musks. This presentation will focus on the tradition of musk manufacture at BBA and its commitment to be a premier musk aroma chemical manufacturer. After a long history in the manufacture of musk xylene, BBA turned its process chemists and engineers loose in the field of polycyclics and first identified patented technology for the manufacture of Abbalide' and then Tetralide@. Based on the accumulated understanding from these successfil projects, they began a search for a polycyclic "supermusk". Combining empirical and computer based methodologies, they discovered a novel potent musk aroma chemical. Its improved odor performance compared to existing polycyclics sparked process research related to the formylation of aromatic ring systems, a key structural sub-

197

I n Search of Nascent Musks... Or Not!

unit. From this investigation arose a number of novel and commercially applicable technologies to carry out this transformation. 2 INTRODUCTION

Since the dawn of perfumery, musk has been both an integral and much sought after component of fragrance mixtures. The structural identification and synthesis of both natural and synthetic musk compounds have peaked the curiosity of chemists for well over one hundred years. In a very early report of a synthetic sequence to a musk odorant in 1878, Kelbela mentioned the weak musk odor of trinitro-para-cymene. However, Ib Baur elevated synthetic musk discovery to a new level shortly after his discovery in the late 1880’s of the trinitro derivative of rneru-methyl-tertiary-butylbenzene.In a brief span of time, he synthesized and patented the products which we know today as “Musk Xylene” (l), “Musk Ambrette” (2), and “Musk Ketone” (3). Until recently, these three materials comprised a very large portion of the total commercial musk market.

? OMe

1

2

3

Additional discoveries in the synthetic musk field were subdued until the late 1940’s when Carpenter and Easte? reported the first nitro-free benzenoid musk, 2,4-ditertiary-butyl-5-methoxybenzaldehyde. This sparked the most enduring and encompassing era of synthetic musk research in the fragrance industry as chemists and their employers each sought to claim their piece of history and profitability. And so, the field of polycyclic musks was spawned. By the 1960’s, a host of familiar musk products like PFW’s Phantolid@ (4) and Tonalid@ (9, IFF’s Celestolide@ (6), and Givaudan’s Versalide@ (7) were on the aroma chemicals scene. In the latter 1960’s, another significant milestone occurred with the identification of isochromans as musk odorants. The best example of this type is IFF’s Galaxolide@(8). During this time of both empirical approach and musk odor theories, 3 process chemists were rapidly responding to the call for cost effective syntheses of these exciting materials. As Friedel-Crafts chemistry became a better appreciated technique, it evolved into a workhorse for chemists in our field. No more effectively used than in the area of cyclialkylation, researchers such as Beets, Bruson and Kroeger, Ipatieff, and Wood used this chemistry to prepare indane and tetralin precursors to a wide range of new polycyclic musks. The discovery that, under Bronsted and/or Lewis acid catalysis, neohexene and/or

198

Flavours and Fragrances

2,3-dimethyl- 1-butene reacted with para-cymene derivatives to form hexamethyltetralins4 led to cost-effective, large scale production of the highly substantive musk Tonalid

I

(5).

6

5

4

@

8

3 DISCUSSION We at Union Camp and Bush Boake Allen have always been proud of our process skills, and our industry history is rich with the results of our need to manufacture high quality products in the most cost-effective manner. In this paper, I would like to convey our efforts in musk related research over the past 15-20 years. It is a story which, through perseverance, takes us from innovative process chemistry on the two major polycyclic musks of our time into new, more powerful polycyclic musks, and back again into process research. These latter discoveries came from a combination of empirical and structure based approaches with the additional caveat of minimizing process cost from the onset. They paralleled and complemented the renewed interest of others in our industry 5d such as Fehr,5a Willis and Zazula,sb Suzukamo and S a k i t ~ , ’Helmlinger ~ and Pesaro, Frater,

Christenson,

et.,5f whose work on lipophilicity and functional group

related structural limitations both contributed to the general understandingSg and identified new, more powerful polycyclic musks.

199

In Search of Nascent Musks... Or Nor!

3.1 Process Development and Manufacture of Abbalide

@

Our initial work in the 1980's was directed toward lower cost manufacture of indanes and indanepropanols, key intermediates in the preparation of our isochroman musk, Abbalide@(8). The original IFF process' for its counterpart, Galaxolide@is shown below (Scheme 1).

Scheme I

BBA sought to exploit her process skill capabilitities advantageously to reduce overall manufacturing costs. Quickly, Ferber and Goddard' first developed novel cyclialkylation technology to pentamethylindane from alpha-methylstyrene and 2methyl-Zbutene with phosphoric acid (Scheme 2). This process led to overall cost reductions in raw materials, catalyst, and capital employed.

+

t.

AICI, I CH,C4 Vacuum

*

HO

Scheme 2 This was followed by another novel BBA process by Ferber, for the preparation of indanepropanols which utilized refluxing solvent to remove heat effectively (Scheme 2). Reductions in capital cost and yield gains were the direct consequence of this improvement in process technology.

200

Flavours and Fragrances

3.2 Sacrificial Olefin Methodology for tbe Manufacture of Tetralide' In the mid-1980's. we shifted our emphasis toward preparation of alkyl substituted indanes and tetralins, intermediates in polycyclic musk manufacture. One target of this effort was Tetralide' (5), our counterpart to Tonalid', which was prized for its very substantive behavior on cloth. To gain market entry, we required a technological edge, and we once again turned to process chemistry to provide it. Acetylhexamethyltelrahydronaphthalene (Tetralidem (5)) can be and has been prepared in a number of synthetically ( Figure 1 ).

and involves two distinct synthetic steps, shown retro-

a 9

a Figure 1 Retro-synthetic Analysis of 6-Acetyl-I, 1.2.4.4,7-hexarnethyl-1,2,3,4tetra hydronaphthalene

Most of the research in our industry has centered on the synthesis of 1,1,2,4,4,6hexamethyltetrahydronaphthalene (HMT, 9). The preparation of HMT can be envisaged from para-cymyl cation and an electron rich C, olefin. The para-Cymyl cation electrophile has been prepared in a number of ways, most expediently from para-cymene itself via Lewis acid catalyzed hydride abstraction (Figure 2).

In Search of Nascent Musks... Or Not!

20 1

AICI, cation source

*+

Figure 2 para-Cymyl Cation Generation via Hydride Abstraction

Initially, we identified technology which provided operating improvements to existing methods." Our application of quaternary ammonium salts to existing Sumitomo technology4 improved both the consistency and yield of the process (Figure 3).

+

/

R-X

AIC Q uatI,

*

+

*+

Figure 3 Improvements to Existing 2,3-Dimethyl-l-butene Based Technology But we recognized the need to prepare para-cymyl cation more eficiently using olefin alternatives to organic halides or excess C6 alkene. (Figure 4). A series of olefins

+

& x t \ Or

H+

I

Figure 4 The Reaction Properties of Sacrificial Olefins

& /

202

Flavours and Fragrances

which we discovered provided the necessary reaction properties of both strong basicity to compete effectively for proton versus c6 olefin and poor nucleophilicity to inhibit its participation in a parallel cyclialkylation.’

Further, they did not, when protonated,

compete with para-cymyl cation for alkylation by c6 olefin, probably due to steric constraints realized in the olefin’s approach. The best of these, 2,4,4-trimethyl-2-pentene, delivered excellent yields of HMT from readily available diisobutylene (Figure 5 ) , when used in conjunction with either neohexene or 2,3-dimethyl-l-butene. The story is even

Figure 5 The Sacrificial Olefin Process for the Manufacture of HMT more satisfying when one realizes the interconnectedness of the cyclialkylating olefin progeny neohexene with its “sacrificial olefin” progenitor. BBA utilized this “sacrificial @ olefin” technology to manufacture Tetralide . Other concurrent Friedel-Crafts studies led I2 to, among other things, a better understanding of indaneketralin interconversions. 3.3 The Search for More Intense Polycyclic Musks The synthetic understanding gained during our studies on indane and tetralin synthesis and interconversion begged the question as to whether other “super musks”, as yet unidentified, might be prepared using our technologies. So, in the late 1980’s, BBA expanded her efforts directed toward discovery of more powerful musk aroma chemicals. Concurrent research camed out by Feh? on the impact of lipophilicity in polycyclic musks supported our belief that stronger musks might still be found, but his work went in a different direction. Our preliminary attempts to exploit recently gained process understanding gave rise to structures with reasonable musk odor, but not with the 13 intensity and/or character desired. What they did generate was a host of materials for which we had strong sensory comparisons to fundamental structure. As we viewed the amassed information, we identified structural traits which we believed would give a better musk. Although some of 5 this contradicted earlier findings about functional group accessibility, we continued, in part due to the fact that the product could be made from reasonable cost, readily available 14 raw materials. The result was a superior polycyclic musk (FHMI, 11). The material was

203

In Search of Nascent Musks... Or No?!

m

as substantive as Tetralide , but with a far more natural and pleasing musk character than any other polycyclic musks. Its intensity was estimated to be four times that of Abbalide

m

and twice that of Tetralidem. The corresponding nitrile (12) was also an effective, stable, musk aroma chemical.

15

11

12

As might be anticipated from the structure, the other homologously substituted indanes had characteristic musk aromas.I6 In some of our early preparations of FHMI, we noticed the appearance of an additional gc peak with time. This turned out to be the formate ester (13) brought about by an unanticipated rearrangement of the aldehyde (Scheme 3). The formate ester was, surprizingly, a pleasant yet slightly more volatile 16 musk.

13 Scheme 3 As a result of this unexpected finding, we subsequently identified a number of 17 indane and tetralin formate ester musks.

3.4 Development of Formylation Technology for Alkylated Aromatic Rings However, the task of cost effective formylation with alkylated aromatics, our major challenge in commercially preparing FHMI, returned us to our milieu of process organic chemistry. The most effective reagent, afpha,afpha-dichloromethyl methyl ether,’ was not commercially available. Benzylic oxidation did not provide the required selectivity . Other formylation reagents were either too hazardous to contemplate using in an industrial setting, given our company’s environmental awareness, or non-reactive for other than activated aromatic systems (Figure 6 ) . It had been known for some time that aldehydes reacted with chloroform under alkaline conditions to form trichloromethylcarbinols, such as the precursor to “Rosone”. We reckoned that if we could prepare the corresponding trichloromethylcarbinol (1 4), it would easily collapse in base to give FHMI (Figure 7).

204

Flavours and Fragrances

&

High yield not readily available

C12CHoMec TiCl,

All benzylic positions oxidized to some degree

&+"0"

Reagent questions

Figure 6 Aldehyde Introduction Methodologies

d

H /

7 OH-

d

C

c

1

3

-

RosoneTM

/

14

Figure 7 Formation and Decomposition of Aryltrichloromethylcarbinols The recent literature was dotted with examples of trichloromethylcarbinols as intermediates to alpha-hydroxy acids and esters," but none of these gave commercially suitable routes to the desired carbinol.

205

In Search of Nasceni Musks... Or Not!

The suggestion of a possible answer came, ironically, from much earlier pesticide research. In the preparation of DDT, chlorobenzene reacts with chloral under acid catalysis to yield the product diary1 trichloroethane (Scheme 4).

/o

c1

0 +

CI,CAH

-

-

H+

H+

DDT

Scheme 4 The proposed intermediate is an aryltrichloromethylcarbinol analogous to the one which we desired. The issue became one of how to selectively protect the intermediate from a second arylation under the reaction conditions. We targeted Lewis acids as potential catalysts for two reasons. First, we knew, based on reports, that Bronsted acids wouldn't provide adequate selectivity. Secondly, we hoped that the trichloromethyl group would coordinatively stabilize the Lewis acid:carbinol complex (15). We knew that we

15

needed to reduce the acidity of the reaction or suffer the consequence of diarylation because of the inherent electron richness of our polyalkylated substrates. Secondly, we were concerned about Lewis acid catalyzed alkylbenzene rearrangements. In this regard we utilized the hemiacetal of chloral to both enhance overall substrate solubility and reduce Lewis acidity. We were pleasantly surprised with high selectivities to the carbinol, especially when titanium tetrachloride was used as the Lewis acid.20 Subsequent treatment of the carbinol with base provided high yields of the corresponding aldehyde2' (Figure 8). The reasonable cost coupled with the commercial availability of chloral makes this route a practical manufacturing option for the formylation of alkyl substituted aromatic compounds.

206

Flavours and Fragrances

CHO I

CHO

92%

I

OMe

I

I

I

OMe

OMe

I

I

53%

70%

92%

Reaction Conditions:

-15"C, 6 hours, hydrocarbon solvent, chloral or its hemiacetal added to TiCI, : substrate complex

Figure 8 Chloral Based Formylation Results

4 SUMMARY AND CONCLUSIONS

In summary, I have highlighted research efforts camed out at Bush Boake Allen/Union Camp in support of musk aroma chemicals. It has been complementary to work carried out by others in our industry. I have tried to show not only the importance of solid process development research, but its creative synergism with product related science. As a consequence we have provided highly cost effective processes to a range of commercial musks, a series of new musk aroma chemicals as well as a more complete understanding of the musk structure/ odor relationship, and commercially viable formylation technology for alkylated aromatics. This research has provided a sound foundation for BBA's continued exploration of new performance based products and her customers.

in Search ofNascent Musks...Or Not!

207

ACKNOWLEDGEMENT The author would like to acknowledge the analytical and technical contributions of J. Yakupkovic, J. Malloy, L. Nielson, D. McMahon, B. Strasser, and Richard Veazey from Union Camp Technology Center, and similar efforts from L. Trinnaman and N. Unwin from Bush Boake Allen. He is deeply indebted to the skilled laboratory support of D. Miller, J. Mahurter, and Z. Sun, without whom the majority of this work would not have been accomplished. The author also would like to recognize the managerial support of both Union Camp Corporation and Bush Boake Allen during the entirety of this project. Finally, this publication is dedicated to the life and memory of Dr. Gerald J. Ferber of Bush Boake Allen, whose friendship, dedication and encouragement continue to be greatly appreciated.

References 1. W. Kelbe, Annalen, 1881, 210, 54; W. Kelbe, Berichte, 1881, 14, 1240. 2. M. S. Carpenter and W. M. Easter, Jr., (The Givaudan Corporation) U.S. Patent 2,450,878 (1948); M. S . Carpenter, W. M. Easter, Jr., and T. F. Wood, J. Org. Chem., 1951, 16,59. 3. T. F. Wood, ‘Fragrance Chemistry: The Science of the Sense of Smell’, E. T. Theimer, Ed., Academic Press, Inc., New York, 1982, Chapter 14, 509; M. G. J. Beets, ‘Structure-ActivityRelationships in Human Chemoreception’, Applied Science Publishers, Ltd., London, 1978. 4. T. F. Wood et al., J. Org. Chem., 1963, 28, 2248.; G. Bern and E. H. Polak, (Polak’s Frutal Works, Inc.) U.S. Patent 2,851,501 ( 1958 ); T. F. Wood and G. H. Goodwin, (The Givaudan Corporation) U.S. Patent 3,246,044 (1966); S . E. Meakins, (Quest International B.V.) European Patent Application 0 393 742 A1 (1990); T. T. Wood and E. Heilweil, (The Givaudan Corporation) U.S. Patent 3,856,875 (1974); H. Sato, et al. (Sumitomo Chemical Company, Ltd.) U.S.Patent 4,284,828 (1981); R. L. Cobb, (Phillips Petroleum Company) U.S. Patent 4,551,573 (1985); No inventor listed, (Nitto Riken Kogyo), Jpn Kokai Tokkyo Koho 82 40 420 (1982). 5. a) C. Fehr, et al., Helv. Chim. Acta, 1989, 72, 1537; b) B. J. Willis and T. J. Zazula, (Fritzsche, Dodge and Olcott, Inc.) U.S. Patent 4,476,040 (1984); c) G . Suzukamo and Y.Sakito, (Sumitomo Chemical Company, Ltd.) U.S. Patent 4,767,882 (1988); d) D. Helmlinger and Mario Pesaro, (The Givaudan Corporation) U.S. Patent 4,634,548 (1987); e) G. Frater, et al., Flavours. Fragrances, and Essential Oils, Proceedings of the 13’h International Congress of Flavours, Fragrances and Essential Oils, Istanbul, Turkey, 15-19 October 1995, 3, 151; f ) P. A. Christenson, et a]., (BASF K and F Corporation) International Patent Publication WO 90/08533 (1990); g) This list is certainly not all inclusive as many others have contributed to the overall musk effort. I apologize for any that might have been inadvertently leA unmentioned. 6. L. G. Heeringa and M. G. J. Beets, (International Flavors and Fragrances, Inc.) U.S. Patent 3,360,530 (1967). 7. G. J. Ferber and P.J. Goddard, (Bush Boake Allen, Ltd.) U.S. Patent 4,440,966 (1 984).

208

Flavours and Fragrances

8. G. J. Ferber, P. J. Goddard, and R. S . Holden, (Bush Boake Allen, Ltd.) U.S. Patent 4,5 15,990 (1985). 9. J. Stofberg, et al.,( Polak’s Frutal Works, Inc.) U.S. Patent 3,278,621 (1966); S. J. Kahn, (Universal Oil Products) U.S. Patent 3,379,785 (1968); D. Davidson and R. M. Lusskin (The Trubek Labs) US. Patent 3,045,047 (1962); V. N. Ipatieff, H. Pines, and R. C. Oldberg, J. Am. Chem. SOC., 1948,70,2123. 10. W. C. Frank and D. M. Miller, Bull. Chem. SOC. Jpn., 1993, 66, 125; W. C. Frank, (Union Camp Corporation) U. S. Patent 4,877,911 (1989); W. C. Frank, (Union Camp Corporation) U. S. Patent 4,877,914 (1989). 1 1. W. C. Frank, (Union Camp Corporation) U. S. Patent 4,877,910 (1989); W. C. Frank, (Union Camp Corporation) U. S. Patent 4,877,912 (1989); W. C. Frank, (Union Camp Corporation) U. S. Patent 4,877,913 (1989); W. C. Frank, (Union Camp Corporation) U. S. Patent 4,877,915 (1989); W. C. Frank, (Union Camp Corporation) U. S. Patent 4,877,916 (1989), W. C. Frank and D. Miller, Proceedings ofthe 12th International Congress of Flavours, Fragrances, and Essential Oils, 1992,465. 12. W. C. Frank, (Union Camp Corporation) U. S. Patent 5,087,785 (1992). 13. W. C. Frank, (Union Camp Corporation) U. S. Patent 5,087,770 (1992). 14. W. C. Frank, (Union Camp Corporation) U. S. Patent 5,095,152 (1992); W. C. Frank, (Union Camp Corporation) U. S. Patent 5,321,173 (1994). 15. W. C. Frank, (Union Camp Corporation) U. S. Patent 5,204,322 (1993). 16. W. C. Frank, (Union Camp Corporation) U. S. Patent 5,206,217 (1993). 17. W. C. Frank, (Union Camp Corporation) U. S. Patent 5,292,719 (1 994); W. C. Frank, (Union Camp Corporation) U. S. Patent 5,292,720 (1994); W. C. Frank, (Union Camp Corporation) U. S. Patent 5,401,720 (1995); W. C. Frank, (Union Camp Corporation) U. S. Patent 5,403,823 (1995). 18. H. E. Baumgarten, Ed. in Chief, ‘Organic Synthesis’, Collective Volume 5, John Wiley and Sons, New York (1973) 49. 19. E. J. Corey and J. 0. Link, Tetrahedron Letters, 1992, 33, 3431, and related references; S. M. Roberts, ‘Comprehensive Organic Chemistry: The Synthesis and reactions of Organic Compounds’, I. 0. Sutherland, Ed., Pergammon Press Ltd., London, 1979, Chapter 9.4,739, and references cited therein. 20. W. C. Frank, R. L. Veazey, J. J. Mahurter, M. J. Jenkins, and N. R. Fairfax, (Union Camp Corporation) U. S. Patent 5,457,239 (1995).

Synthesis and Application of Thiocarbonyl Compounds Shuichi Hayashi', S. Hashimotol, H. Karneoka2 and K. Sugimoto2

'

NAGAOKA PERFUMERY CO. LTD, R&D CENTRE, 1-3-30, ITSUKAICHI, IBARAKI CITY, PSAKA 567, JAPAN DEPMTMENT OF APPLIED CHWIISTRY, KINKI UNIVERSITY, 3-4-1, KOWAKAE, HIGASHIOSAKA CITY, OSAKA 577, JAPAN

1 INTRODUCTION Every flavourist and perfumer within the flavour and fragrance industry is always seeking good aroma chemicals. A good aroma chemical should have a unique odour, low threshold, low cost, be safe to use, and have additional functionality. Having a unique odour is especially of high importance, thus researching such materials is both interesting and valuable. It is well known that chemicalswhich have sulphur atoms in their molecular structure often have a unique odour'.Such organosulphur compounds have been identified in various materials. Some examples are, from mango, guava, grapefruit, durian, and tomato, the molecules dimethyldisulphide,2-methyl-1-propanethiol, 1-propanethiol,and 1para-menthene-8-thiol have been identifiedzd.Roasted coffee is known to contain methyl2-methyl-3-furyl sulphide,difiufuyl sulphide, 2-methylthiazole and 3-methylthiophene etc.'. Not only are organosulphur chemicals found in fruitand vegetables, they can also be found in cooked meat'. The organosulphur compounds in cooked meat are often key flavour components. The organosulphur molecules described before can be categorized by their functional groups, i.e. sulfide, di- or higher sulphides, thiol, thioacid and heterocyclic compounds. The organosulphur compounds which contain a thiocarbonyl group have not been extensively investigated with regard to the flavour and fragrance industry. In this report the authors describe the synthesis and the application of thiocarbonyl compunds as aroma chemicals. In addition the bioactivity of these molecules is also discussed. Finally, to obtain more unique aroma chemicals, the authors reduced the thiocarbonyl compounds to the correspondingthiols.

References 1. H.Kameoka and K.Furukawa, 'Kaori to kurashi', Shokabo, Tokyo, 1994, p. 135. 2. A.J. MacLeod and N.M.Nirmala, Phyochem., 1984,23,2361.

3.O.Nishimura, K.Yamaguchi, S.Mihara and T.Shibamoto, JAgric. Food Chem., 1989, 37, 1, 139. 4. E.Demole, P.Enggist, and GOhloff. Hefv. Chim. A m . , 1982,65,6, 1785.

210

Flavours and Fragrances

5. J.Baldry, J.Dougan, and G.E.Howard, Phytochem., 1972, 11,2,2081. 6 . R.G.Butery, G.Takeoka, R.Teranishi, L.C.Ling, J. Agric. Food Chem., 1990,38, 1 1 , 2050. 7. R. Tressl, and R.Silwar, J. Agric. Food Chem., 1981,29, 5 , 1078. 8. G.J.Hartman, Q.Z.Jin, G.J.Collins, K.N.Lee, C.T.Ho, and S.Stephen,J. Agric. Food Chem., 1983,31,5, 1030.

2 SYNTHESIS OF THIOCARBONYLS The synthesis of the thiocarbonyl group has been reported, but the methods described do have some drawbacks such as low conversion ratios, and the use of dangerous reagents like hydrogen sulphide". The simplest way of obtaining the thiocarbonyl group is to substitute the oxygen atom of the carbonyl group directly with a sulphur atom. Lawesson's reagent has been reported as a convenient reagent to cany out this substitution smoothlyJ-' (Figure 1). The organoleptic properties of the products have not been reported. The Authors synthesized a range of thiocarbonyl compounds using Lawesson's reagent, and then assessed their odour.

I

I

S

Lawesson's Reagent

R

R\

I

R 2 = O R-C-OR

II

->

S

R-C-OR

II

S

0 R-C-SR'

'c= RJ

-I

R-C-SR'

II

S

Figure 1 Lawesson's Reagent and Typical Thiation ofCarbonyls

21 1

Synthesis and Application of Thiocarbonyl Compounds

2.1 Synthesis of Thiolactones The delta and gamma lactones have been identified in various h i t s , milk, and butter, and the lactones are often key components of the flavour. In this chapter the Authors describe the synthesis of thiolactones. A range of gamma lactones from butyro to dodecalactonewere converted to their corresponding thiolactones. The gamma lactones were mixed with half a mole equivalent of Lawesson's reagent in toluene and then refluxed for three hours with stimng. After the reaction was complete the Lawesson's reagent was removed by filtration and the thiolactones were isolated by distillation and column chromatography(silica gel, mixed solvent of hexane/ether). The corresponding delta lactones were refluxed in xylene for five hours, and then purified as for the gamma lactones. In result, all lactones were converted to the correspondingthiolactones (Figure 2). The yield of the delta lactones was lower than that for the corresponding gamma lactones. The reason behind this is currently being researched. As the electric environment around the carbonyl group is similar, the Authors assumed that the conformation of the side chain is directly connected to the reactivity. The most stable conformation of the delta lactone ring is the chair type. When a side chain is in an axial position, the carbonyl and side chain would be in close proximity to each other, thus this environment may disturb the reaction.

Lawesson's Reagent in Toluene I Reflux I 3 h

H

:a92%

C4H9 : e 91%

.

C5H11 f 75% C6Hi3 : g 8 1 % OH15 : h 79% C8Hi7 : i 88%

Lawesson's Reagent in Xylene I Reflux I 5 h

C4H9 : a47% C5Hii : b 63% c6H13 : C 5 5 % C7Hi5 : d 6 9 %

Figure 2 Thiation of y and &lactones by Lawesson's reagent

212

Flavours and Fragrances

The products of the reaction were identified by instrumental analysis. The Infra red spectrum shows the absorption of a typical lactone carbonyl to be 1772 cm”. In the thiolactone products the absorption has shifted to 1238 cm”. The mass spectrum showed that the molecular ion peaks had increased by 16 after the reaction. These results were obtained for each delta and gamma lactone reaction product. (Table 1)

Table 1 IR and MS Analysis Data of Thiolactones

IR Spectral Data (vmax cm-I) 1772 (C=O)

1238 (C=S)

MS Spectral Data ~~~~

R

Lactone

H CH3 C2H5 C3H7 C4H9 CSHl 1 C6H 13 C7H 15 C8H17

m/z 86(M+) m/z 100(M+)

m/z 114(M+)

m/z 128(M+) m/z m/z m/z m/z m/z

142(M+) 156(M+) 170(M+) 184(M+) 198(M+)

Thiolactone m/z 102(M+) m/z 116(M+) m/z 130(M+) m/z 144(M+) m/z 158(M+) m/z 172(M+) m/z 186(M+) m/z 200(M+) m/z 2 14(M+)

IR Spectral Data (vmax cm-I) 1772 (C=O)

1238 (C=S)

MS Spectral Data R C4H9 C5H11 C6H 13 C7H15

Lactone m/z 156(M+) m/z 170(M+) m/z 184(M+) m/z 198(M+)

Thiolactone m/z 172(M+) m/z 186(M+) m/z 200(M+) m/z214(M+)

Synthesis and Applicarion of Thiocarbonyl Compounds

213

The 13C-NMR data is shown in Table 2. The chemical sift of the carbon at No.2 position has obviously changed from 172 ppm to 222 ppm during the reaction. Every delta and gamma lactone commonly showed these results. Table 2 IH and 13C NMR Data of Thiobctones I I

I

R= C6H13 (y-Decathiolactone)

1H N M R 6 (PPm) 4.87 (H) 3.04 (2H) 1.79 1.98 (2H) I .61 1.75(2H) 1.21 1.58 (8H) 0.90 (3H)

---

I

13C N M R

6 @pm) 222.48 C2 O172.32 (C=O) 44.85 C3 34.05 C4 91.05 CS 31.53 29.56 28.90 25.21 22.40 13.95

R=CSH 1 1 (GDecathiolactone) IHNMR

13C N M R

6 @pm)

6 (PP4

4.35 (H) 2.92 (2H) 1.92 2.00 (2H) 1.56 1.72 (2H) 1.24- 1.38(8H) (3H) 0.90

223.60 40.65 35.19 26.89 83.58 31.55 24.73 22.46 18.09 13.97

--

c2 *174.50 (C=O) C3 C4 C5 C6

2.2 Application of thiolactones to flavour and fragrance. The thiolactones obtained from the reactions showed unique and strong odours. The odour ofgamma butyrothiolactone was strong, green, oily, and fatty. The thiocarbonyl group

214

Flavours and Fragrances

seemed to especially impart the odour character of green. Gamma valero and gamma hexathiolactones had similar green odours. Gamma heptathiolactone showed a mellow, green, and fatty odour. From gamma octa to gamma undecathiolactone, the odour was also strong, wonderfully mellow, green, and fatty. The corresponding delta thiolactones had also changed their odour clearly from their lactone analogues. Delta nonathiolactone had a mild sulphury green and ether like note. Delta decathiolactone had a mellow, green, and fatty odour, delta undecathiolactone was sulphur. Both the gamma and delta thiolactones show unique strong green notes, and because of this they can be used for the creation of flavours and fragrances. The Authors applied the thiolactones to flavour and fiagrances and examined the results. (Table 3). The butter flavour with thiolactones included had a mellow, natural, and long lasting odour and taste. The white peach flavour with thiolactones included showed a mellow, natural, and gave the impression of freshly squeezed juice from a well ripened peach. The lily of the valley fragrance including gamma undecathiolactone had a mellow and long lasting scent. These results suggest that the thiolactones give volume, richness, and natural freshness to fruits, daily products, cooking flavours, and that they give body and naturality for fragrances.

Table 3 Application of thiolactones to Flavour and Fragrance

Butter Flavour Ingredient Diacetyl Butyl Acetate Acetoin Butyric Acid Hexanoic Acid Octanoic Acid 2-Heptanone 2-Nonanone GNonalactone GNonathiolactone &Decalactone GDecathiolactone GUndecalactone BUndecathiolactone GDodecalactone BDodecathiolactone Propylene glycol Ethanol

Formula A

Formula B

2.0 2.0 4.0 6.0 8.0 4.0 I .o 2.0 5.0

2.0

40.0 1 .o

25.0

200.0 700.0

2.0 4.0 6.0 8.0

4.0 I .o 2.0 5.0 0.1 40.0 0.1 I .o 0.1 25.0 0.1

200.0 700.0

Synthesis and Application of Thiocarbonyl Compounds

215

Table 3 Continued. White Peach Flavour Ingredient Ethyl maltol Vanillin Ethyl Acetate Ethyl Myristate Benzaldehyde Benzyl alcohol Linalool 2-Methyl butyric acid y-Octalactone y-Oetathiolactone y-Decalactone y-Decathiolactone GDecalactone &Decathiolactone y-Undecalactone y-Undecathiolactone GDodecalactone &Dodecathiolactone NPC Peach Base Propylene glycol Ethanol

Lily of the Valley fragrance Ingredient Hydroxycitronellal Linalool Ylang ylang oil Jasrnin absolute Rhodiol Terpineol Benzaldehyde Heliotropine Caldamon oil Benzyl propionate Benzyl salicylate y-Undecalactone (1 0%) y-Undecathiolactone (10%) Vanillin Anisyl formate Cyclamen aldehyde Undecenal Civetone (3%) Citronellyl acetate

Formula A

Formula B

0.1 0. I 5.0 8.0 1.o 30.0 7.0 3.0 1.o

0.1 0.1

3.0 5.0

20.0 3.8 10.0

200.0 700.0

5.0 8.0 1 .o

30.0 7.0 3.0 1.o 0.1 3.0 0.1 5.0 0.1 20.0 0.1 3.8 0.1 10.0 200.0 700.0

Formula A

Formula B

300 150 20 30

300 150

20 30

150

I50

I50

150

1 25

I 25 I

I 15

15

60

60 0.9

1

2

30 3 2 20 40

0.1 2 30 3 2 20 40

216

Flavours and Fragrances

References 1. J.W.Scheeren, P.H.J.Ooms and R.J.Nivard, Synthesis, 1973, 149. 2. H.Hoffmann and G.Schumacher, Tetrahedron Lett., 1967,31,2963. 3. S.Oae, A.Nakanishi and N.Tsujimoto, Chem. and Ind., 1972,575. 4. P.Salama, M.Poirier and M.Caissie, Heterocycles, 1995,41, 1 1,2481. 5. B.S.Pedersen, SScheibye, N.H.Nilsson and S.O.Lawesson, Bull. SOC.Belg. 1978,87,3, 223. 6. BSPedersen, S.Scheibye, K.Clausen and S.O.Lawesson, Bull. SOC.Belg. 1978,87,4, 293. 7. S.Scheibye, J.Knstensen and S.O.Lawesson, Tetrahedron, 1979,35, 1339.

3 OTHER FUNCTIONALITY OF THIOLACTONES

It is well known in Japan that some plants that contain organosulphur compounds have special bioactivity. For example wasabi (Japanese horseradish) has antimicrobial activity, and garlic is known to activate the human body"'. In this chapter the Authors describe the antimicrobial activity and Superoxide dismutase like activity of the thiolactones. 3.1 Antimicrobial Activity The antimicrobial activity was anaylzed by the paperdisk diffusion method. The procedure is as follows, firstly the organisms diluted in a suitable solvent were coated onto the culture medium. Secondly the paper disk is set onto the medium, and finally the solution being tested is added to the disk, and everything is then incubated for 24 or 48 hours. If the sample has antimicrobial activity an inhibitory circle will form around the paper disk. The diameter of this inhibitory circle is measured as the intensity of antimicrobial activity. The organisms tested are shown in table 4, and the results of the tests are shown in table 5. Activity was shown against only E. coli and S.aureus. The highest activity was shown by the low molecular weight gamma thiolactones. The high molecular weight derivatives showed a much weaker effect. The delta thiolactones showed the same trend. Gamma butyro, gamma valero, and gamma hexathiolactones may be used as antimicrobial additives. Table 4 Examined Organisms

Staphylococcus aureus Bacillus subtilis Escherichia coli Saccharomyces cerevisiae Penicillium chrysogenum Aspergillus niger

I F 0 3060 IF0 1213 IF0 12734 sw-33 I F 0 6352 IF0 6342

217

Synthesis and Application of Thiocarbonyl Compounds

Table 5 Intensity of Antimicrobial Activity

Sample

dose

Intensity (mm) E. coli S. aureus

13 7 11 6

y-Heptathiolactone y-Octathiolactone y-Nonathiolactone y-Decathiolactone y-Undecathiolactone y-Dodecathiolactone

1000 1000 1000 1000 1000 1000

8 6 4 3

11 5 10 4 5 3 3 2 2 2

1 0

0

6-Nonathiolactone 6-Decathiolactone GUndecathiolactone bDodecathiolactone

1000 1000 1000 1000

3 1 1 0

2 2 1 0

y-Butyrothiolactone

1000

500 y-Valerothiolactone

1000

y-Hexathiolactone

1000

500 500

10 4

1

not active against other organisms

3.2 Superoxide Dismutase Like Activity

As the second functional property superoxide dismutase like activity was examined. The superoxide radical is well known and has been studied greatly'". These radicals are generated inside our bodies by autoxidation of haemoglobin or other systems. The radicals damage DNA or denature proteins, thus can initiate cancers or cause ageing. Our bodies have a natural inhibitor of superoxide anion radical, it is the enzyme called super oxide dismutase (SOD). The SOD reduces the superoxide anion to hydrogen peroxide, which is then converted to water by the catalase enzyme. It has been reported that some extracts of herbs reduce the superoxide anion radical just like SOD. Some reports describe some organosulphur compounds as having SOD like activity'. The Authors examined the activity by the NBT method'". The outline of the NBT method is shown in figure 2. This method uses xanthine and xanthine oxidase as the generator of superoxide anion radicals. Initially xanthine is oxidized to uric acid by xanthineoxidase, it is during this oxidation that the superoxide radicals are generated. The sample and nitroblue tetrazolium (NBT) are then mixed into this system. When a thiolactone has SOD like activity, a proportion of the superoxide radicals are reduced to hydrogen peroxide, the

218

Flavours and Fragrances

rest react with NBT causing a change to the formazan form. At this time the colour of NBT changes from yellow to blue. The intensity of activity is found by measuring the absorbance of the formazan form. If a sample has a very strong SOD like activity, the superoxide radical anions will be consumed, thus leaving too few to cause a colour change. The intensity of SOD like activity was assessed by the following expression.

EXPRESSION (B-(A-C))IBxlOO (Yo) A; is absorbance of 560 nm of sample with XOD system. B; is absorbance of 560 nm of only XOD system. Without sample system. C; is absorbance of sample without XOD. When sample is coloured, the absorbance should be modified.

Nltroblue tetrazolium (yellow)

/“IJric acid

f Formazan (blue)

Diformazan (blue)

Figure 3 Outline of NBT Method (Xanthine-xanthinoxidase, Nitroblue tetrazolium method) The results of the examination of SOD like activity are shown in table 6 . Only gamma thiolactones were examined, and the strongest activity was seen from the higher molecular weight derivatives. Gamma deca, and gamma undecathiolactones showed especially high inhibitory ratios. These results show the opposite in terms of molecular weight from that found for antimicrobial activity. Gamma deca and gamma undecathiolactones can be used as functional flavour or fragrance ingredients.

Synthesis and Application of Thiocarbonyl Compounds

219

Table 6 Intensity of SOD like Activity Inhibitory ratio Sample y-Butyrothiolactone y-Valerothiolactone y-Hexathiolactone y-Heptathiolactone y-Octathiolactone y-Nonathiolactone y-Decathiolactone y-Undecathiolactone y-Dodecathiolactone

(YO) 7 15

6 16 29 50 75 92 47

References 1. H.Kameoka and K. Furukawa, ‘Kaori to kurashi’, Shokabo, Tokyo, 1994, p.95. 2. H.Kameoka and K. Furukawa, ‘Kaori to kurashi’, Shokabo, Tokyo, 1994, p.99. 3. H.Kameoka and K. Furukawa, ‘Kaori to kurashi’, Shokabo, Tokyo, 1994, p.199. 4. A.Battistoni, S.Folcarelli, R.Gabbianelli, C.Capo and G.Rotilio, Bi0chem.J.,1996,320, 713. 5 . G.Cao, E.Sofic and R.L.Prior, J.Agric. Food Chem.,1996,44,3426. 6. A.Rotman, Mol. Pharmacol, 1976,12,887. 7. T.Kuramoto, Shokuhin to kaihatsu, 1992, 27,3,22. 8. T.Miyazawa, Food Chem., 1995,119,27.

4 REDUCTION OF THE THIOCARBONYLS TO THIOLS To obtain more aroma chemicals pocessing a unique odour, the reduction of the thiocarbonyl group to the corresponding thiol was examined. Ethyl cinnamate, cis-jasmon, menthyl acetate, camphor, and menthone were converted to their corresponding thiocarbonyl compounds using Lawesson‘s reagent (Figure 4). The thiocamphor and thiomethone have unique green, and bumt sugar like odours. Both of these two compounds were then reduced with lithium aluminium hydride to their corresponding thiols (thiobomeol, and thiomenthol)(Figure 5). The thiobomeol had a cool, green, and tropical fruits like odour, whilst the thiomenthol had a strong green, mint like odour

220

Flavours and Fragrances

Ethyl cinnamate

P

0

S

sweet. fnrity, balsamic

I

long lasting. honey like and masted

fruity and jasmin like

caramalic, honey like, burnt. masted

Menthyl acetate

-

-

----›

fruity. bergamot and lavendar like

Q.-Isweet, cooked vegetables

Figure 4 Thiation of Other Aroma Chemicals

Lawesson's Reagent Camphor

Thiocamphor

Thioborneol

camphoric

nutty. burned sugar

cool, tropical fruits

I

Menthone Menthol like

Thiomenthone Caramalic. burned sugar

Thiomenibol green, tropical fruits, grapefruit

~~

Figure 5 Reduction of Thiocarbonyl Compounds

Synthesis and Application of Thiocarbonyl Compounds

22 1

5 CONCLUSION Various unique flavour and fragrance chemicals were obtained by the conversion of carbonyl compounds to thiocarbonyl compounds. Especially the thiolactones have a high potential to be key compounds for use in flavours and fragrances. Some thiolactones showed bioactivity thus they may be used in the future as functional additives. The development of new flavour chemicals is expected from the reduction of thiocarbonyls to thiols.

Heteropoly Acids as Catalysts for Fine Chemicals Synthesis Ivan V. Kozhevnikov LEVERHULME CENTRE FOR INNOVATIVE CATALYSIS, DEPARTMENT OF CHEMISTRY, THE UNIVERSITY OF LIVERPOOL, LIVERPOOL L69 3BX, UK

1 INTRODUCTION

Catalysis by heteropoly acids (HPAs) and related polyoxometalate compounds is a field of growing importance.'-" Many new and exciting developments are taking place in the field in both research and technology. This paper concentrates mainly on the use of HPAs as acid and redox catalysts for low-temperature, liquid-phase organic reactions relevant to the synthesis of fine chemicals. A series of reviews, discussing properties of HPAs and various aspects of HPA catalysis have been published.'-'' 1.1 Structure

HPAs are complex proton acids incorporating polyoxometalate anions (heteropolyanions) having metal - oxygen octahedra as the basic structural units.'' The first characterized and the best known is the Keggin heteropolyanion XMIz04,"-* where X is the central atom (Si4+, Pst, etc.), x is its oxidation state, and M is the metal ion (Mo6+,W6+, V5+,etc.). This anion is composed of a central tetrahedron XOa surrounded by 12 edge- and comer-sharing metal - oxygen octahedra M 0 6 (Figure 1).l2

Figure 1 The Keggin structure of a-XM,204t-s Among a wide variety of HPAs, the Keggins are the most stable and more easily available; these are the most important for catalysis. In this paper HPAs are understood as the Keggin acids, unless otherwise stated. Generally, their formulas will be

Heteropolyacids and Related Compounds as Catalysts for Fine Chemicals Synthesis

223

abbreviated to XM, e.g., PW and SiW for H3PW12040 and H4SiWIZ040, respectively. 1.2 Properties

Why are HPAs interesting for catalysis? HPAs have several advantages as catalysts which make them environmentally and economically a t t r a c t i ~ e . ~They . ~ . ~have: (1) Discrete and mobile ionic structure. The structure is composed of discrete heteropolyanions loosely linked by countercations, unlike the network structure of e.g. zeolites and metal oxides. The structural mobility of solid HPAs is important for their use in heterogeneous catalysi~.~.~ (2) Very strong Bronsted acidity. HPAs are stronger than conventional mineral acids and solid acids such as H,SO, Si0,-A120,, zeolites, etc. Some solid HPAs, e.g. PW and its acid Cs salt CS,,~H,,PW, become superacids after thermal re treatment.^ (3) Redox properties. Particularly important is the ability of some polyoxometalates to reversible multielectron redox reactions under mild conditions as well as the ability to activate such common oxidants as 0, and H,O,.””” Moreover, acidbase and redox properties can be varied by changing the chemical composition of HPAs. (4) Very high solubility in water and oxygenated organic solvents such as lower alcohols, ethers, etc.; but HPAs are insoluble in hydrocarbons. ( 5 ) Fairly high thermal stability in the solid state which allows to use HPAs in heterogeneous catalysis at moderately high temperatures. (6) Such a unique property as “pseudoliquid phase”,’.’ etc. Not surprisingly, HPAs make efficient acid, redox and bifunctional catalysts both in homogeneous and heterogeneous systems. HPAs are used as model systems for fundamental research, providing unique opportunities for mechanistic studies at the molecular level. At the same time, they become increasingly important for applied catalysis. 1.3 Industrial application

In the last two decades, the broad utility of HPA catalysis has been demonstrated in a wide variety of synthetically useful selective transformations of organic substances.’-’ Several new industrial processes based on HPA catalysis have been developed and commercialized, for example oxidation of methacrolein, hydration of olefins (propene, n-butene and isobutene) and polymerization of tetrahydrofuran.’ There are also a few small-scale industrial processes relevant to tine chemicals synthesis to be discussed later.

’

2 ACID CATALYSIS 2.1 Advantages

Generally, HPAs have the following advantages over conventional acid ( I ) HPAs offer broad operational choice - they can be used in homogeneous, heterogeneous or biphasic systems. (2) Being stronger acids, and, therefore, more efficient proton donors, HPAs are generally more active catalysts than mineral acids and conventional solid-acid catalysts.

224

Flavours and Fragrances

In particular in organic media, the molar catalytic activity of HPAs is 100 to 1000 times higher than that of H,S0,.‘6 (3) Efficient operating under milder conditions which is prerequisite of higher selectivity. (4)Lack of side reactions like sulfonation, chlorination, etc. due to the inertness of heteropolyanions towards organic molecules. ( 5 ) Heteropolyanions can play important role in reaction promoting by stabilizing organic intermediates, e.g. carbenium ions.’ 2.2 Organic reactions

Below are given some selected examples of liquid-phase acid-catalyzed reactions, demonstrating that HPAs in many cases offer strong options for more efficient and cleaner processing. 2.2.1 Homogeneously catalyzed reactions. The HPA-catalyzed hydration of C3-C4 olefins (Eq. ( I ) ) is an industrially important reaction, the hydration of propene being the first commercial process based on HPA catalysis.’ RCH=CH2 + H2O + RCH(OH)CH3

(1)

The hydration of isobutene is used for the separation of isobutene from the C4 hydrocarbon stream produced by cracking. As the catalyst, a concentrated aqueous solution of HPA is used. Compared to mineral acids, such as H2S04, HNO,, and HC104, HPA is 2-4 times more active per equal H’ concentration and shows a higher selectivity, minimizing side reactions such as isobutene ~ligomerization.””~ Mechanistic studies” showed that heteropolyanions play important role in promoting the reaction by stabilizing intermediate carbenium ions. HPAs are reportedly more efficient catalysts than H2S04 and HClO4 in the hydration of phenylacetylene’ (Eq. (2)). PhCECH + H2O + PhCOCH3

(2)

Cycloalkenes are hydrated to cycloalkanols with 99 YOselectivity in the presence of a catalyst consisting of a concentrated aqueous solution of an arylsulfonic acid and tungsten HPA.” PW and SiW are efficient catalysts for the homogeneous hydration of camphene to isobomeol, which is an intermediate in the synthesis of camphor.16 Synthesis of glycosides catalyzed by HPA is of industrial importance.’ Glycosides are used as new effective and biodegradable surfactants. HPA is several times more active than the conventional catalysts such as toluene sulfonic acid and ZnC12. Thus acetylated monosaccharides interact readily with alcohols in a homogeneous phase in the presence of 2% of HPA with respect to the sugar derivative at 7O-13O0C, yielding 7090% of glycosides” (Eq. (3)).

Heteropolyacids and Related Compounds as Catal.ystsfor Fine Chemicals Synthesis

AcO S Ac

O

A OAc

c

+

RoH

-

225

OAc P Ac.

c

O

OAc ~ O +R AcOH

The synthesis of vitamin E with HPA proceeds via the condensation of trimethylhydroquinone with isophytol (Eq. (4)). HPA provides the same yield and quality of vitamin E as ZnCl,, which is the best industrial catalyst. But in contrast to ZnCI,, the HPA catalyst is used in much smaller amounts and can be recycled.'*

2.2.2 Biphasic reactions. Separation of products and recovery and recycling of a catalyst often becomes much easier if a homogeneously catalyzed reaction can be performed in a biphasic system. HPAs due to their special solubility properties, i.e., high solubility in a variety of polar solvents and insolubility in nonpolar solvents, are suitable catalysts for operating under phase-transfer conditions. Polymerization of THF is used for the preparation of polyoxytetramethyleneglycol (PTMG), which is employed for manufacturing Spandex fibers and polyurethanes. PTMG is commercially produced by a two-step process, including ring-opening polymerization of THF with acetic anhydride catalyzed by HC104, followed by hydrolysis of the terminal acetate groups in the prepolymer. Aoshima et have developed a one-step process for the THF polymerization to directly yield PTMG (Eq. (5)) in the presence of PW as catalyst. nTHF + HzO + HO-[-(CHZ)~-O-],-H

(5)

The reaction proceeds in a two-phase system. The upper (product) phase is a PTMG solution in THF. The lower (catalyst) phase is a concentrated solution of HPA in THF. The polymer is formed in the HPA phase and transferred to the product phase. The process is performed continuously. The PTMG with a molecular weight of 500-2000 and a narrow molecular weight distribution is obtained from the product phase. PMo, PW, and SiW catalyze the biphasic cyclotrimerization of aldehydes, such as ethanal, propanal, butanal, 2-methylpropanal, etc., to produce 2,4,6,-trialkyl-1,3,5trioxanes in high yields (Eq. (6)), as reported by Sat0 et a/.''

R

Catalyst turnover number is more than 10000 for the propanal cyclotrimerization. At

226

Flavours a n d Fragrances

high conversions of aldehyde, the reaction mixture spontaneously separates into two phases, a product phase and a catalyst phase, which, depending on aldehyde can be solid or liquid. For the propanal cyclotrimerization, the reaction mixture separates into two liquid phases, and the recovered catalyst phase can be easily reused. Similarly, the formation of acetals between alcohols or diols and a range of aldehydes or ketones is catalyzed by HPAs in two-phase system.2’ HPAs catalyze acetoxylation and hydration of dihydromyrcene to yield dihydromyrcenol and dihydromyrcenyl acetate (Eq. (7)) which are useful as perfume ingredients.” These reactions occur with a 90% selectivity at room temperature in a twophase system. The heterogeneously catalyzed conversion with Si0,-supported HPA is also feasible.

OAc

OH

The alkylation of hydroquinone with isobutene to yield 2-t-butylhydroquinone and 2,5-di-t-butylhydroquinone catalyzed by H3PW12040,H6PzW18062and H6PzWz1071 under phase-transfer conditions in a biphasic system has been reported.’’ 2.2.3 Heterogeneottsly catalyzed reactions. The advantage of heterogeneous systems over homogeneous is easy separation of catalyst from reaction products. HPAs are very efficient and versatile catalysts for alkylation, dealkylation and transalkylation of These reactions are widely used for the preparation of antioxidants, bioactive substances, positional protection, etc. Bulk and 90,-supported PW and SiW are much more active than H,SO,, ion-exchange resins, aluminosilicates, etc. Even Nafion’, a polymeric perfluororesinsulfonic acid comparable in its strength to 100% H,SO,, is less active than PW.’4 Alkylation of p-cresol by isobutene with the use of HPA has been commercialized in Russia. It is a step in the synthesis of antioxidants. The use of HPA instead of H2SO4 provides the gain in selectivity of 7- 10% and almost completely eliminates toxic water pollution.’ HPAs and their salts are promising solid-acid catalysts for the Friedel-Crafts reactions, to replace the conventional homogeneous catalysts such as AICI, and H,SO,, which bring about serious environmental and operational problems.’ Insoluble acidic salt Cs, ’H,,PW shows high efficiency in acylation of activated arenes, such as p-xylene, anisole, mesitylene, etc., by acetic and benzoic anhydrides and acyl chlorides. This catalyst provides higher yields of acylated arenes than the parent PW, the latter being partly soluble in the reaction mixture.2b Esterification of dipicolinic acid with butanol is a step in the synthesis of pharmaceuticals (Eq. (8)). PW as a homogeneous catalyst is almost as efficient as sulfuric acid, yielding 100% of diester. An acidic salt, Ce,,,H,,PW, practically insoluble in butanol, was found to be fairly active as a heterogeneous catalyst. Although less

221

Heteropolyacids and Related Compounds as Catalystsfor Fine Chemicals Synthesis

active than homogeneous PW, it can be easily recovered and reused.” + 2BuOH

H02C

+ 2H20

(8)

BuO$ f i 0 2 B u

The insoluble salt C S ~ , ~ H ~ .and ~ P even W PW itself, which is highly soluble in water, can be included in the silica matrix by means of a sol-gel technique to be waterinsoluble and easily separable microporous solid-acid catalysts. The catalysts thus obtained have large surface areas (400-800 m2 g-’), strong acidity and are thermally more stable than Amberlyst- 15. They effectively catalyze the hydrolysis of ethyl acetate in aqueous phase, showing higher activities than Amberlyst-15 and H-ZSM-5. Remarkably, the immobilization of PW into the silica matrix effectively suppresses the HPA leaching to as low as 0.3% during the hydrolysis reaction (6OoC, 3 h).’* Silica-supported PW has been found to be an active and recyclable catalyst for the Diels-Alder reaction (Eq. (9)) in toluene medium, providing a 70-80% yield.29The Ce3’ salt, Ce0.87H0.4PW,showed a fairly high activity too, while C S ~ , ~ H ~ was . ~ P practically W inactive. Bulk PW exhibited a very low activity probably due to the small surface area and site blocking by diene polymers. 0

0

2.2.4 Shape-selective catalysis. Supported HPAs are important for applications because bulk HPAs are nonporous materials with low surface area. As supports, especially interesting, are materials with regular pore structure, e.g. zeolites, providing a shape selectivity. However, conventional zeolites are not suitable for this because their pores are too small to adsorb large HPA (12 A) molecules. Recently HPA (PW) supported on a novel mesoporous pure-silica molecular sieve MCM-4 1, having uniform pores 32 A in size, was prepared and chara~terized.’~~’’ The PW/MCM-41 compositions with PW loadings from 10 to 50 wt% have -30 8, uniformly sized mesopores. As shown by transmission electron microscopy, the PW species are mainly located inside the MCM-41 pores rather than on the outer surface.” PW/MCM-41 exhibits a higher catalytic activity than H,SO, or bulk PW and shows a shape selectivity in liquid-phase phenol alkylation.” This catalyst may be promising for shape-selective reactions of large organic molecules, particularly for the synthesis of fine chemicals.

3 LIQUID-PHASE OXIDATION The liquid-phase oxidation of organic substances catalyzed by polyoxometalates (POM) proceeds in homogeneous or biphasic systems, oxygen, hydrogen peroxide,

228

Flavours and Fragrances

alkylperoxides, etc., being applied as oxidants. In contrast to acid catalysis, where easily available classical Keggin-type HPAs dominate, in the liquid-phase oxidation, a wide variety of transition-metal-substituted heteropolyanions are used. Multicomponent POM redox catalysts are often considered as robust inorganic metalloporphyrin 9.1 1.32 analogues. In this part, we will discuss oxidations with the use of the most important and environmentally benign oxidants - dioxygen and hydrogen peroxide. Several reviews on catalysis by POM in the liquid-phase oxidation have been pub~ished.2~3~5~9~l 1.33 3.1 Oxidation with dioxygen This section mainly reviews recent applications of Keggin-type mixed-addenda heteropolyanions PMo,~-,,V,,O~~(~+")~ (HPA-n) as catalysts for aerobic liquid-phase oxidation. Here HPA-n can mean either anion or acid as well as a partially protonated anion. The HPA-n catalytic system, discovered by Matveev et is the most efficient and versatile one in the POM series for oxidation by 02.3935*36 3.1.1 General principles. In liquid-phase oxidation, HPA-n with n = 2-6 are used. These compounds are remarkable because they are the reversibly acting oxidants under mild condition^.^"'^^ Generally reactions catalyzed by HPA-n proceed via a stepwise redox mechanism represented by Eqs. (10) and (1 1): HPA-n + Red + m H' + H,(HPA-n) + Ox H,(HPA-n) + m/4 O2 + HPA-n + m/2 HzO This mechanism includes stoichiometric oxidation of the substrate (Red) by HPA-n followed by reoxidation of the reduced form of the oxidant, H,(HPA-n), with dioxygen, where H,(HPA-n) = H , ( P M O ~ ~ - , , ~ + V , , - ~ + Vm~<) (n.~ + Actually, ~ ) - , it is the V5+t)v"' transformation that is responsible for the redox properties of HPA-n. The reduction of HPA-n in solution is accompanied by its protonation to maintain the charge of the polyanion. Reactions (10) and (1 1) can be carried out in the same reactor (one-stage process) or separately as a combination of two stoichiometric reactions (two-stage process)."' Thus HPA-n combines the advantages of stoichiometric and catalytic oxidations. As the catalysts, either one-component, two, andor multicomponent systems are used. The most important two-component system includes HPA-n and Pd(I1). Other twocomponent systems are also known, e.g., those including complexes of Pt, Ir, Ru, etc.'" The thermodynamic condition for the occurrence of reactions (10) and (11) is: E(Red) 5 E(HPA-n) I E(02) = 1.23 V.3 The redox potential E(HPA-n) is 0.7 V vs. normal hydrogen electrode for HPA-n with n = 1-4 at pH 1.37 Hence the above condition fulfills for a wide variety of organic substrates as well as for many inorganic redox systems to be used as co-catalysts with HPA-n. 3.1.2 One-component HPA-n system. A range of hydrocarbons, including alkanes, can be homogeneously oxidized with HPA-n as catalyst under mild c o n d i t i o n ~ . ~Many ."~~~

''

Heteropolyacids and Related Compounds as Catalysts for Fine Chemicals Synthesis

229

synthetically attractive oxidations of hydrocarbons have been developed to date. For example, highly efficient dehydrogenation reactions with HPA-2 in CH2C1CH2Cl/tetraglymesolution have been reported by Neumann et al.38 Thus reaction (12) yields 98% anthracene (70°C, 16 h).

+

11202

-

\

\

\

+2Hg

Oxidation of 2-methylnaphthalene to quinone (Eq. (13)) is a step of the vitamin K3 synthesis. In industry it is carried out with highly toxic Cr03 as a stoichiometnc oxidant with a low yield (ca. 50%). With HPA-n, we have succeeded to carry out the catalytic oxidation in AcOH-H20 solution at 12O-14O0C with 82% selectivity at 78% conversion.39

0

Gorodetskaya et aLa reported oxidative bromination of a range of aromatics catalyzed by HPA-n with HBr as a brominating agent and O2 as a terminal oxidant. More recently Neumann et aL4' reported the oxidative bromination of phenols in the para-position (Eq. (14)) in a homogeneous HPA-2/CH2C1CH2Cl/tetraglyme system with a remarkably high yield of 99% (2OoC, 5 h).

Ishii et ~ 1 . ~have ' studied a number of interesting oxidations with O2 in the presence of an aldehyde, e.g. 2-methylpropanal, as a sacrificial reductant and (NH&H4PMo6V6040as a catalyst in dichloroethane solution. Thus the epoxidation of olefins proceeds at 25'C with a very high selectivity of 81-100% (Eq. (15)). A carboxylic acid as a cooxidation product is also formed. In addition, the Baeyer-Villiger oxidation of cyclic ketones to lactones was accomplished with 78-87% selectivity by using benzaldehyde as the aldehyde (25'C, 20 h, 1 atm 0,). These reactions may be synthetically useful if the carboxylic acid is a useful product, provided a less harmful solvent instead of CH2ClCH2Clis employed.

230

alkene + O2 (f RCHO) + epoxide (+ RCOOH)

Flavours and Fragrances

(15)

A variety of oxygenates, e.g., aldehydes, ketones, phenols, etc., can be oxidized with HPA-n under mild conditions with very good yields. Bregeault et u I . reported ~~ the oxidative cleavage of cyclic ketones. For example, the cleavage of 2methylcyclohexanone is catalyzed by HPA-4 in aqueous solution to yield 97% of ketoacid (60°C, 8 h) (Eq. (16)).

Ishii et d4'studied the oxidation of aldehydes to carboxylic acids with the use of the NH4+salt of HPA-6 as a catalyst. The reaction proceeds under very mild conditions (2S°C, 1 h) in dichloroethane with 96-100% selectivity at 60-78% aldehyde conversion. We44*45 and also Neumann et ~ 1 reported . ~ synthetically ~ attractive oxidations of alkylphenols. The oxidation of 2,3,5-trimethylphenoI (TMP) to trimethylbenzoquinone (TMBQ) (Eq. (1 7)) is of practical interest as a step to synthesize vitamin E. This reaction (HPA-4, AcOH-H20, 5OoC) yields TMBQ with 86% selectivity at 100% TMP conversion. 45

The oxidative coupling of 2,6-dialkylphenols to diphenoquinones (Eq. (1 8)) catalyzed by copper amine complexes is used for the synthesis of antioxidants. In this reaction, HPA-n shows a higher activity than copper complexes. The oxidative coupling proceeds in H20, AcOH-H20, or MeOH-H20 as solvents at 25-5OoC under 0 2 pressure of 1-5 atm. Diphenoquinones, insoluble under these conditions, can be readily isolated from the reaction mixture by crystallization. Thus, 2,2',6,6'-tetramethyI- and 2,2'6,6'tetra-t-butyl-diphenoquinoneshave been obtained in a nearly 100% yield. The catalyst can be recycled without loss of its activity.44

Heteropolyacids and Related Compounds as Catalysts for Fine Chemicals Synthesis

23 1

3.1.3 Two-component system HPA-n + Pd(I4. The two-component redox system HPA-n + Pd(II), discovered by M a t ~ e e v ”has ~ ~attracted considerable attention. A wide variety of stoichiometric oxidations of organic compounds (SH2), e.g., olefins, alcohols, arenes, etc., by Pd(I1) can be carried out catalytically with the use of HPA-n as a coX5.35.36 catalyst and O2as an oxidant (Eq. (1 9)).

Generally, such reactions proceed via a stepwise redox mechanism (Eq. (20)).’*3It is analogous to the Wacker-type oxidation with CuCI2-PdCl2 as the catalyst. But the HPA-n + Pd(I1) catalyst can work in the absence of or at a very low concentration of C1’ ions. Consequently, it has the advantage of having a higher selectivity and is less 1 corrosive.

One of the most promising reactions is the Wacker-type oxidation of olefins (Eq. (2 1)):

M a t ~ e e v ”first ~ ~ proposed the use of HPA-n as chloride-free oxidant in the Wacker reaction. The oxidation of ethylene can be represented by Eqs. (22)-(24). C2H4+ Pd(I1) + H 2 0 + CH3CH0 + Pd(0) + 2 Hi Pd(0) + HPA-n + 2 H’ + Pd(I1) + H2(HPA-n) H2(HPA-n) + 1/2 O2 + HPA-n + H 2 0 Catalytica3’ has recently developed new technology for the Wacker oxidation of ethylene and higher olefins with the use of the HPA-n + Pd(I1) system. In this process, over 99% of the production of chlorinated by products is eliminated, decreasing to less than 0.01% yield on ethylene. 3.1.4 Biphasic oxidation. The advantage of biphasic systems is that they combine a catalytic reaction and product separation in one unit. This is especially promising for fine and specialty chemical synthesis to avoid difficult separation problems. Matveev et ~ 1 . ~ reported ’ two-stage oxidation of 2-methyl-1-naphthol to

232

Flavours and Fragrances

menadione in a high yield with space-separated steps (25) and (1 1), step (25) being camed out in an aqueous-hydrocarbon two-phase system at 5OoC with 89% yield of menadione at a 100% naphthol conversion. This is a promising alternative to the conventional synthesis of vitamin K3 by oxidizing 2-methylnaphthalene with Cr03.

0

We48 have developed the efficient one-stage oxidation of TMP to yield 84% TMBQ at a 100% TMP conversion in a two-phase system at 5OoC and 1 atm O2pressure (Eq. (17)). The catalyst (lower) phase is a HPA-4 solution in AcOH-H20. The upper phase is a TMP + TMBQ solution in a hydrocarbon solvent, e.g., heptane, benzene, etc. Monflier et al.49reported the biphasic Wacker oxidation of 1-decene to 2-decanone at 80°C with excellent yield (98%) and practically no double bond migration. In this case, the catalyst (lower) phase is a three-component system, containing Pd(II), HPA-6 and Cu(I1) in an aqueous solution, where Cu(I1) is suggested to promote the Pd reoxidation. P-Cyclodextrin as a phase-transfer catalyst is used; it transfers 1 -decene to the aqueous phase by a host-guest interaction. 3.2 Oxidation by H 2 0 2 Tungsten and molybdenum POMs catalyze various oxidations of organic substances by hydrogen peroxide such as epoxidation of olefins, oxidation of alcohols, glycols, phenols, etc., in homogeneous or two-phase systems.' Peroxo polyoxometalates have been shown to be the active intermediates in these reactions." Two highly effective and mechanistically closely related catalyst systems relevant to POM mediated hydroperoxide oxidation are of particular interest, namely, those developed by groups of Venturello and Ishii. According to Venturello et al.,50,5' the selective epoxidation of alkenes is performed in a biphasic system, e.g., CHC13-H20, at 60-70°C with the use of diluted H202 (2-15%) and tungstate and phosphate as the catalyst in the presence of a phasetransfer catalyst - quaternary ammonium cation (Q') with C6-C18alkyl groups. In this system, a peroxo POM, { P04[WO(02)2]4}3-was isolated and characterized.'' This peroxo complex was found to be remarkably effective in stoichiometric as well as in catalytic biphasic oxidation of alkenes by H20z and was postulated to be the active oxygen transfer agent in this reaction. The anion { P04[WO(Oz)2]4}3~ has the Cz symmetry and consists of the central PO4 tetrahedron linked through its oxygen atoms to two pairs of edge-sharing distorted pentagonal bipyramids W(O2)2O3 (Figure 2). Each tungsten atom is linked to two peroxo

Hereropolyacids and Related Compounds as Catalystsfor Fine Chemicals Synthesis

233

groups - one non-bridging (q2) and the other bridging (q2,q’)- located in the equatorial plane of the pentagonal bipyramid.”

Ishii et ~l.’~demonstratedthat oxidations of a wide variety of organic substrates with commercially available 35% H20z can be effectively performed in homogeneous phase or more often in a biphasic system with the use of a catalyst comprising heteropoly acids PW or PMo and cetylpyridinium chloride, the latter providing phasetransfer function. The PW catalyst is usually superior to PMo. Tris-cetylpyridinium salt of PW, [x-C5HSN’(CH2)&H3]3(PW12040)3(CPW), is prepared by interacting PW with 3 equivalents of cetylpyridinium chloride. This salt catalyzes the epoxidation of alkenes and allylic alcohols with H202 in a biphasic system using CHC13 as the solvent, while a,&unsaturated carboxylic acids are readily epoxidized in water at pH 6 to 7. secAlcohols and diols are dehydrogenated by the CPW-H202 system in t-butyl alcohol to give the corresponding ketones in good yields. Primary alcohols are much less reactive under such conditions. However, a,w-diols give lactones in fair yields. The CPW-H202 system is also efficient for oxidative cleavage of carbon-carbon bonds of vic-glycols and alkenes to yield carboxylic acids. Other reactions, which can be readily camed out with this system include the conversion of internal alkynes to the corresponding a$-epoxy ketones, allenes to a-alkoxy or a-hydroxy ketones, amines to oximes and nitrones, anilines to azoxy-nitroso- and nitrobenzenes and sulfides to sulfoxides and sulfones. According to a schematic mechanism proposed by Ishii et bl. the two-phase epoxidation of alkenes by hydrogen peroxide in the presence of PWI2O403-as the catalyst precursor proceeds as follows. An active peroxo POM is formed in the aqueous phase by interacting the Keggin heteropolyanion with hydrogen peroxide. With the phase-transfer catalyst, Q’, the peroxo POM is almost filly transferred into the organic phase because its Q’ salt is readily soluble in organic solvents. The reaction takes place preferentially in the organic phase via the oxygen atom transfer from the peroxo POM to the substrate. The peroxo POM is regenerated at the interface by the interaction with H202. In the recent years, the Venturello-Ishii chemistry has been studied by several groups. Thus in the studies by Brtgeault et ~ 1 . and ~ Griffith ~ 9 ~ et~al.,” a wide variety of new mono-, di-, tri-, and tetranuclear tungsten and molybdenum peroxo POMs with q2,ql-peroxo moieties and various central groups - PO4, PhP03, Ph2P02, SO4, As04 HAs04, CH~ASOJ,etc. have been isolated and characterized by X-ray and tested in oxidations with H 2 0 2 . Ballistreri et a1.’6and Hill et u1.” have shown that the Venturello-Ishii epoxidation is first-order in both {P04[WO(02)2J4}3-and alkene. Hill et al.” obtained kinetic and spectroscopic evidence that two processes dominate during the reaction: a slow

-

234

Flavours and Fragrances

epoxidation (Eq. (26)) {P04[W0(02)2]4}3-+ alkene -+ PW4, PW3, and PW2 + epoxide

(26)

followed by a rapid regeneration of { P04[WO(02)2]4}3with H202. Catalyst deactivation as well as the use of chlorinated hydrocarbons as solvents have been mentioned as the major drawback to the Venturello-Ishii process.” 4 CONCLUSION The selected examples reviewed show the broad scope of potentially promising applications of HPAs as acid and redox catalysts in low-temperature liquid-phase organic reactions. Due to their unique physicochemical properties, HPAs can be profitably used in homogeneous, biphasic or heterogeneous systems, often providing higher activities and selectivities and allowing for cleaner processing compared to conventional catalysts. The high effectiveness of HPAs as acid catalysts is primarily due to their strong Bronsted acidity, greatly exceeding that of ordinary mineral acids and solid-acid catalysts. Such important HPA properties as the ability of polyanions to stabilize organic intermediates, “pseudoliquid phase” and lack of side reactions like sulfonation, etc., also contribute to enhancing the catalytic activity and reaction selectivity. In liquid-phase oxidation, polyoxometalates can be applied either as stoichiometric oxidants or as catalysts in conjunction with such environmentally friendly oxidants as O2 and H20z.The use of a variety of multicomponent redox systems based on polyoxometalates greatly extends the scope of possible oxidations. However, for these potentially promising reactions to be used in practice, complete recovery and recycling of HPA catalysts will be required.

5 ACKNOWLEDGMENTS The capable assistance of co-workers referred to in the citations and the financial support from The Netherlands Science Foundation (NWO), Unichema International and Quest International are gratefully acknowledged.

References 1. K.I.Matveev, Kinet. Katal., 1977, 18, 862. 2. Y.Ono, in ‘Perspectives in Catalysis’, J.M.Thomas and K.I.Zamaraev, Eds., Blackwell, London, 1992, p. 43 1. 3. 1.V.Kozhevnikov and K.I.Matveev,Appl. Catal. , 1983,S, 135. 4. I.V.Kozhevnikov, Russ. Chem. Rev., 1987,56, 8 1 1. 5. I.V.Kozhevnikov, Catal. Rev. Sci. Eng., 1995,37, 3 11. 6. M.Misono, Catal. Rev. Sci. Eng., 1987, 29, 269. 7. M.Misono and N.Nojiri, Appl. Catal.. 1990,64, 1. 8. Y.Izumi, K.Urabe, and M.Onaka, ‘Zeolite, Clay and Heteropoly Acid in Organic Reactions’, KodanshaNCH, Tokyo, 1992, p. 99.

Heteropolyacids and Related Compounds as Catalystsfor Fine Chemicals Sytithesis

235

9. T.Okuhara, N.Mizuno, and M.Misono, Adv. Catal., 1996,41,113. 10. R.J.J.Jansen, H.M. van Veldhuizen, M.A.Schwegler, and H. van Bekkum, Recl. Trav. Chim. Pays-Bas, 1994,113,115. 11. C.L.Hill and C.M.Prosser-McCartha, Coord. Chem. Rev., 1995,143,407. 12.M.T.Pope, ‘Heteropoly and Isopoly Oxometalates’, Springer, Berlin, 1983. 13.K.Urabe, K.Futsita, and Y.Izumi, Shohbai, 1980,22,223. 14. A.Aoshima, S.Yamamatsu and T.Yamaguchi, Nippon K a g a h Kaishi, 1987, 976; 1987,984;1990,233. 15. 0.Mitsuo and Y.Fukuoka, U.S. Patent 4,528,407,1985. 16.M.A.Schwegler and H. van Bekkum, Bull. SOC.Chim. Belg., 1990,99,113. 17. K.Kiromura, S.Kitazawa, Y.Takata, and TSakakibara, U S . Patent 4,874,852, 1988. 1 8. I.V.Kozhevnikov, S.M.Kulikov, N.G.Chukaeva, A.T.Kirsanov, A.B.Letunova, V.I.Blinova, React. Kinet. Catal. Lett., 1992,47,59. STonomura, and S.Yamamatsu, Polymers for Advanced 19. A.Aoshima, Technologies, 1990,2,127. 20. SSato, H.Furuta, TSodesawa, and F.Nozaki, J. Chem. SOC. Perkin Trans. 2, 1993, 385. 21.SSato, KSugara, H.Furuta, and F.Nazaki, J. Mol. Catal. A , 1996,114.209. 22. I.V.Kozhevnikov, A.SiMema, A.J.A. van der Weerdt, and H. van Bekkum, J. Mol. Catal. A, in press. 23.M.N.Timofeeva and I.V.Kozhevnikov, React. Kinet. Catal. Lett., 1995,54,413. 24. I.V.Kozhevnikov, S.M.Kulikov, M.N.Timofeeva, A.P.Krysin, and T.F.Titova, React. Kinet. Catal. Lett., 1991,45,257. 25.I.V.Kozhevnikov and M.N.Timofeeva, J. Mol. Catal., 1992,75,179. 26.Y.Izumi, M.Ogawa, W.Nohara, and K.Urabe, Chem. Lett., 1992, 1987. 27. M.N.Timofeeva, R.I.Maksimovskaya, E.A.Paukshtis, and I.V.Kozhevnikov, J. Mol. Catal. A, 1995,102,73. 28. Y.Izumi, M.Ono, M.Kitagawa, M.Yoshida, and K.Urabe, Microporous Materials, 1995,5,255;Y.Izumi, K.Urabe, and M.Onaka, Catal. Today, 1997,35,183. 29.G.J.Meuzelaar, L.Maat, I.V.Kozhevnikov, and R.A.Sheldon, J. Org. Chem., 1997, in press. 30. I.V.Kozhevnikov, K.R.Kloetstra, A.Sinnema, H.W.Zandbergen, and H. van Bekkum,J. Mol. Catal. A, 1996,114,287. 3 1. I.V.Kozhevnikov, A.SiMema, R.J.J.Jansen, K.Pamin, and H. van Bekkum, Catal. Lett., 1995,30,241. 32.R.A.Sheldon, Top. Curr. Chem., 1993,164,22. 33. N.Mizuno and M.Misono, J. Mol. Catal., 1990,64,1. 34. K.I.Matveev, E.G.Zhizhina, N.B.Shitova, and L.I.Kuznetsova, Kinet. Katal., 1977, 18,380. 35. J.H.Grate, D.R.Hamm, and S.Mahajan, in ‘Polyoxometalates: From Platonic Solids to Anti-Retroviral Activity’, M.T.Pope and A.Muller, Eds., Kluwer, Dordrecht, 1994,p. 281. 36.I.V.Kozhevnikov, J. Mol. Catal. A, 1997,117,151. 37. V.F.Odyakov, L.I.Kuznetsova, and K.I.Matveev, Zhur. Neorg. Khim., 1978, 23, 457.

236

Flavours and Fragrances

38. R.Neumann and M.Lisse1, J. Org. Chem., 1989,54,4607. 39. T.A.Gorodetskaya, I.V.Kozhevnikov, and K.I.Matveev, USSR Patent 1,121,255, 1984. 40. T.A.Gorodetskaya, I.V.Kozhevnikov, and K.I.Matveev, Kinet. Katal., 1982, 23, 992. 41. R.Neumann and I.Assael, J. Chem. SOC. Chem. Commun., 1988, 1285. 42. M.Hamamoto, K.Nakayama, Y.Nishiyama, and Y.Ishii, J. Org. Chem., 1993, 58, 642 1. 43. A.Atlamsani and J.-M.Bregeault, J. Org. Chem., 1993, 58, 5663; B.E.AIi, J.M.Brigeault, J.Mercier, J.Martin, C.Martin, and O.Convert, J. Chem. SOC. Chem. Commun.. 1989,825. 44. T.A.Gorodetskaya, I.V.Kozhevnikov, and K.I.Matveev, USSR Patent 1,155,580, 1985. 45. O.A.Kholdeeva, A.V.Golovin, R.I.Maksimovskaya, and I.V.Kozhevnikov, J. Mol. Catal., 1992, 75,235. 46. M.Lisse1, H.Jansen in de Wal, and R.Neumann, Tetrahedron Lett., 1992,33, 1795. 47. KLMatveev, E.G.Zhizhina, V.F.Odyakov, and V.N.Parmon, Rum. Chem. Bull. 1994, 3, 1 142; K.I.Matveev, V.F.Odyakov, and E.G.Zhizhina, J. Mol. Catal. A , 1996,114, 151. 48. I.V.Kozhevnikov, A.Sinnema, and H. van Bekkum, J. Chem. Res. (S). 1996,238. 49. E.Monflier, E.Blouet, Y.Barbaux, and A.Mortreux, Angew. Chem. Int. Ed. Engl., 1994,33,2 100. 50. C.Venturello, R.D‘Aloisio, J.C.J.Bart, and M.Ricci, J. Mol. Catal., 1985,32, 107. 51. C.Venturello and M.Gambaro, Synthesis, 1989,295. 52. Y.Ishii and M.Ogawa, in ‘Reviews on Heteroatom Chemistry’, Vol. 3, A.Ohno and N.Furukawa, Eds., MYU, Tokyo, 1990, p. 121. 53. L.Salles, C.Aubry, RThouvenot, F.Robert, C.Doremieux-Morin, G.Chottard, H.Ledon, Y.Jeannin, and J.-M.BrCgeault, Inorg Chem., 1994,33,871. 54. L.Salles, J.Y.Piquema1, R.Thouvenot, C.Minot, and J.-M.Brkgeault, J. Mol. Catal. A , 1996,117, 375. 55. N.M.Gresley, W.P.Griffith, A.C.Laemme1, H.I.C.Nogueira, and B.C.Parkin, J. Mol. Catal. A , 1997, 117, 185. 56. F.P.Ballistreri, A.Bazzo, G.A.Tomaselli, and R.M.Toscano, J. Org. Chem., 1992, 57,7074. 57. D.C.Duncan, R.C.Chambers, E.Hecht, and C.L.HiI1, J. Am. Chem. Soc., 1995, 117, 681.

Subject Index

Abbalide, 196,199 Absolute, 70,98 Ageing, 2 17 AIDS, 18 Aldehyde, 3,7,21,24,3 1 Aldehydic muguet, 21,23,27 Aldol condensation, 23 Aldoses, 159 Alginate, 140 Allene oxide synthase, 13 Aluminium chlorohydrate, 23 Amadori rearrangement, 153,159,166 Amarige, 40-4 1 Ambrox, 4 1-42,44,106 American brown oil extraction, 70,75 Amino acid oxidase (L), 17 Amino acid, 3,5 Antigen, 5 Antiperspirant, 23-24 Appel's intensity, 52 Arabica coffee, 155 Aroma chemicals, 12 Aroma extract dilution analysis, 154-155,166 Aromatic resins, 96- 1 14 Arthritis, 18 Aspergillus niger, 17 Atropine, 3 Aura, 36,38 Autoxidation, 24,217 BacdanoI, 40 Bacillus macerans, 57 Bacteria, 11 Balsam, 98 BBA, 196,198-199 Beef flavour, I55 Beefy meaty peptide, 14

Benzenoid musk model, 33 Benzoin, 96,111-1 13 Benzyl acetate, 40 Benzyl acetone, 38 Benzyl salicylate, 44 BINAP (Pd), 187 BINAP (Rh), 185 BINAP (Ru), 189-192 Biocatalysis, 137-15 1 Biodegradation of biocatalyst waste, 138 of rose oxide, I88 Biosynthetic pathway, 13 Biotechnology, 11 Bisabolene (alpha), 104 Bitter orange oil leaf, 70,75 flower, 70,75,82-83 BMP (see beefy meaty peptide) Boiling point, 36 Bottom note, 36-47 Bourgeonal, 21-22,24 Bovine tongue papillae, 18 Bread flavour, 154,158 Brbnsted acid, 197,205 Browning reaction, 116 Browning, 153,158 Camphene, 193 Cancer, 18,217 Canola oil, 16 Carbohydrases, 138 Carbohydrate fragmentation, 166 Carbon dioxide extract, 98 Carbon dioxide extraction supercritical fluid, 70-71,8 1 supercritical liquid, 8 1 Carboxylic acid, 23

238

Carene (delta-3), 58 Carvone (L), 15 Caryophyllene (beta), I09 Cashmeran, 40,44 Cedarwood, 73 Cedramber, 40,44 Cedrol, 44 Celestolide, 197 Charm analysis, 154,159,166 Cheese flavour, 14,16- 1 7,172- 173 Chicken flavour, 155 Chiral synthesis, 185-195 Chocolate flavour, 153-154,158 Citronellol, 16 Citrus oils, 70-7 1,91 Clove oil, 90 Cocoa flavour, 154,158 Coffee flavour, 154- 155 Concretes, 70 Conformational analysis, 196 Copaiba, 96,109-1 10 Com, 11 Correlated mutation analysis, 4-6 Cotton, 11 Coumarin, 45 Cyclamen aldehyde, 2 1-22 Cyclamen purpurascens, 1 88 Cycloamiloses (see cyclodextrin) Cyclodextrin glycosyltransferases, 57 Cyclodextrin, 57-67,93 Cyclogalbaniff, 41 -42 Cycloglucans (see cyclodextrin) Cyclomatooligoses (see cyclodextrin) Cyclopentadecanolide, 41 -42.44 Cysteine, 153,155,158-159 Damascenone (beta), 12 Damascone (alpha), 42 Damascone (beta), 42 Damping theory, 179-181 Darwinian selection, 5-6 Deamination. 17 Decalactone (delta), 12 Decalactone (gamma), 12 Dehydrogenation, 193. Dendrobium superbum, 37-38 Deoxyosones, 153-154, 163

Flai:oursand Fragrances

Determinants, 7-9 Diacetyl, 166 Diastereomeric composition, 185 Diffusivity, 36,38,42-43 Dihydrofamesol, 22,188- 189 Dihydromyrcene hydration, 226 Dipole moment as calculation parameter, 8,5 1-52 Disulphide bond, 4 DNA sequence, 13 Dupical, 2 1,26 Eclipse of sun, 36 Electronic parameters, 5 1 Elemi, 96,107- 109 Elemol, 107 Enantiodifferentiation, 185 Enantiomers, 57-67,185- 195 Enantiomeric excess, 64-65 Enzymatic conversion, 12 Enzymatic degradation, 57 Enzyme, 137-151 Escherichia coli, 17 Essential oil, 15,65,70-86,90-92 Esterases, 138 Ether hydrolase, 138 Ethyl linalool, 41-42 Ethyl vanillin, 40,44 Eugenol, 140 Expression of citrus fruits, 75 Extract, 98 Exudate, 70,96 Farnesol, 2 1 Fechners law, 178 Fermentation, 7 1 Fish tasty peptides in, 14, flavour, 154 Flavourzyme, 139 Floralozone, 4 1-42 Florosa Q, 21,24,29-3 1 FMC corporation process, 70,75 Food irradiation, 16 Formylation, 203 French fries flavour, 154-158 Friedel-Crafts, 197,202,226

Subject Index

Frontier molecular orbitals, 48 Fructose, 140 Fungus, 17 Furaneol, 12,14 Furanose sesquiterpenes, 102 Furfural, 154 Furfbryl thiol, 155 G-protein, 18 Galaxolide, 42,44,197,199 Galbanum, 96,99-101 Gardenia, 189 Gas chromatography, 58,87,93,154 enantioselective, 58 multidimensional, 64 mass spectroscopy, 64 FT-IR, 64 Smelling, 28,32-33,154-155 Gene, 13,16,18 Genetic Engineering, 11,16 Geraniol, 15-16 Ginger oil, 90 Givaudan-Roure, 42,145,197 Givenchy, 40 Givescone, 42 Glucose oxidase, 146 Glucose, 140 Glycoside synthesis, 225 Glyoxal, 166 Grob test, 58 Guaicol, 193 Guerlain, 39 Gum, 98 Headspace analysis, 173 Headspace dilution analysis, 154,166 Headspace, 38 Hedione, 40,42,44 Heteropolyacids, 222-234 Keggin structure, 222 solubility of, 223 industrial application, 223 use in catalysis, 222-234 Hexenol (cis-3), 12-14,7 1,144 Heyns rearrangement, 153,159,166 Histidine, 5-7 Hops, 78

239

Hydrodiffusion, 70-7 1,76 Hydrodistillation, 70,75 Hydrogen peroxide oxidation, 232-234 Hydrogen sulphide, 155,159,210 Hydrolases, 137-139 Hydrolysis, 7 1 Hydroperoxide lyase, 12-13 Hydroxyacetone, 166 IFF, 36,42,44,197 IFRA, 1 10 Immune system, 5 In-mouth concentration measurement of, 174 In-nose concentration measurement of, 171,174-175 Indane, 199-200,202-203 Infra-red analysis, 212 Insulin, 11 Ionone (beta), 42,45 Iso-E-super, 40-4 1 Isoeugenol, 140 Isomerases, 137,139-140 Isopulegol, 187 Italian pellatrice method, 70,75 Jasmine, 75 Jasmone (cis), 42 Ketoses, 159 Klebsiella pneumoniae, 57 Labdanum, 78,96,105-106 Lactone (delta), 192,211 Lactone (gamma), 191-192,211 Lavender(spike), 73 Lavender, 63 Lawesson’s reagent, 2 10-21 1,219 Leaf alcohol (see Hexenol (cis-3)) Ledol, 106 Lemon oil, 65 Lewis acid, 197,205 Ligases, 137,143 Lilial, 3,5,18,21-22,24 Lily of the valley (see muguet) Limonene, 15,40,44,70,75 Linalool, 38,40,44,65

240

Linalyl acetate, 40-41,44,63-64 Linolenic acid, 12,14 Lipases, 16,I38 Lipoxygenases, 12- 13,144 Living flower technology, 38 Lolitol, 42 Lyases,137,140-142 Lyral, 3,5,18,21,24 Maillard reaction, 116,153-168 Majantol, 2 1 Mammals, 9 Mass spectroscopy atmospheric pressure ionisation, 171,174 Matsutake, 191 Matsutakeol, 191 Mayol, 22 Meat flavour, 14,16,153-155 Menthol, 15,180-181,188-1 87 Menthonethiol, 155,219 Metabolic pathway, 16,137,138 Methional, 12,17 Methyl anthranilate, 12 Methyl ionone, 40,42 Methyl jasmonate, 13 Methyl octin carbonate, 42 Microbes, 11 Microbial conversion, 12 Middle note, 36-47 Milk protein, 16 Mint oil, 12,70-71 Molecular modelling, 4,29,196 Molecular weight As a calculation parameter, 36 Monoamine oxidase. 17 Mouthfeel, 154 Muguet, 3,2 1-34,188,2 15 Multiple regression analysis, 93 Musk ambrette, 197 Musk ketone, 197 Musk xylene, 196-197 Musk xylol, 40 Musk, 48-53,196-206 intensity, 48 macrocyclic, 196 nitro, 196- 197

Flavours and Fragrances

Musk (continued) polycyclic, 196-198,202 substantivity of, 200,203 Musklactone (see cyclopentadecanolide) Mutation, 5 Myrrh, 96,102-1 04 Nagarmotha oil, 83 National Geographic magazine, 36-37 Natural flavours, 12 Natural isolates, 70-86 preparation of, 70 Nerolidol, 1 10 Neural network, 87,173 Neurospora crassa, 17 Nitrile, 7 Nitroblue tetrazolium method, 2 17-2 18 Nuclear magnetic resonance Heteronuclear correlated, 1 19-133 INEPT spectra, I 19-133 Nitrogen-15, 119-133 Nonalactone (gamma), 12 Nosespace analysis, 174 NOVO, 139 Oakmoss, 77 Odour activity values, 154,159 Odour binding pocket, 4 Odour intensity, 29 Odour perception, 9,185 Odour sensing system 87-94 Odour threshold 30,36,43 Odourant, 18 Old testament, 97 Olefin hydration, 224 Oleoresin, 98 Olfactory bulb, 4 Olfactory epithelium, 174 Olfactory system, 3 Olfactory transduction, 3 Olibanum, 96,101-1 02 Oligosaccharides, 57 Opoponax, 96,104- 105 Opsins, 5 Optical purity, 185 Orchid, 37-38 Organoleptic stability, 24

Subject Index

Orris root, 71 Overexpression, 17 Oxidases, 17,144-148 Oxidation, 23-24 Baeyer-Villiger, 192 by hydrogen peroxide, 232-234 in liquid phase, 227-228 of lipids, 158 Wacker-type, 23 1 with dioxygen, 228-23 1 Oxidative cleavage of cyclic ketones, 230 Oxidoreductases, 137,143 Oxyphenylon, 38 Paper disk difksion method, 2 16 Passionfruit compound, 42 Pectin, 140 Peppermint oil, 12,65,90 Peptidases, 138-139 Peroxidases, 148 Peru resin, 96,110- 1 1 1 PFW, 197 Phantolid, 197 Phellandrene, 107 Phenol oxidases, 144 Phenylalanine, 7 Phosphatidyl lipids, 93 Pinene (alpha), 58,65 Pinene (beta), 58,65 Plant engineering, 15 Polysantol, 42 Polysioloxane, 58-59,6 1,65 Popcorn flavour, 153-154,158 Potato flavour, 1 1,153 Primary alcohol, 28 Principal component analysis, 90,9 1.93 Process flavourings, is3,iss Propranolol, 3 Protease, 15,16-17,138-139 Protein disulphide-isomerase, 140 Protein hydrolysate, 16 Psychophysical laws, 171-1 72, 177-179 Pyrazines,116-133, 155,158 Pyruvaldehyde, 166 Quartz crystal microbalance, 87-94 Quartz resonator sensor, 88

24 1

Quest International, 145,149 Raspberry odour, 38 Receptor, 9,185 adrenergic (beta), 4 G-protein coupled, 3 muscarinic acetylcholine, 5 musk, 48 olfactory, 3,18,180 taste, 18 trigeminal, 180 Recombinant DNA technology, 11,15, 16,18 Recombinant yeast cells, 13-14,18 Reductases, 144,148-149 Resin, 98 Resinoid, 98 Rhodopsin, 185 Robusta coffee, 155 Rose 2 1,26,31,33 Rose oil, 70,75,82,187 Rose oxide, 187-188 Rosemary, 73 Sacrificial olefins, 201-202 Sandalwood oil, 193 Sandalwood, 73 Santalene (alpha), 104 Santalex T, 193-194 Santalol, 44 Secondary alcohol, 28 Sensing film, 92 Sensory analysis, 171 Sensory perception, 175 Sensory time-intensity measurements, 171-182 Serine, 7 Sfumatrice method, 70,75 Shalimar, 39-40 Shape-selective catalysis, 227 Signal transduction mechanism, 18 Solid-phase micro extraction, 38 Solvent extraction, 70,77 Soybeans, 11 Starch, 57 Steam distillation, 70-7 1,75 Stecker degradation, 153

242

Stereoisomer, 185-195 Steric hinderance, 28 Stevens law, 172,178 Strawberry, 189 Structure activity relationship, 2 1,26,34, I96 Structure determination, 1 16 Structure odour relationship, 18 Styrallyl acetate, 40,189- 190 Sulphydryl oxidase, 146 Super-santalex, 193- 194 Superoxide dismutase (SOD), 21 7-2 19 Superoxide radical, 2 17-2 18 Sweetness, 171 Takasago, 193 Taste generation, 137- 151 Tasty peptides, 14 Tea flavour, 154 Terpenoids, 7 1 Terpinen-4-01,65 Terpineol (alpha), 65 Tetrahydrofuran polymerisation, 225 Tetralide, 196,200-203 Tetralin, 199,202 Thermally processed foods, 154 Thiamine, 159 Thiazolidine carboxylic acid. 163 Thioborneol, 2 19 Thiocamphor, 2 19 Thiocarbonyl, 209-22 1 Thiolactones, 2 1 1-2 19 antimicrobial action of. 21 6-2 17

Flavours and Fragrances

Thiols, 155,2 19 Thiomenthol, 2 19 Thiomenthone, 155,2 19 Tinctures, 70 Tolu, 96,113 Tomato, 1 1 Tonalid, 42,45,197-198,200 Topnote, 36-47 Transferases, 137,142-143 Transgenic field trial 1 1 Transgenic plants, 16 Treemoss, 73,77 Tricholoma matsutake, 191 Trimerization of aldehydes, 225 Tuber melanosporum, 191 Undecavertol, 42 Union Camp, 198 Uric acid, 2 17 Van der Waals radii, 52 Vanillin, 110,112,140 Vapour pressure, 36 Versalide, 197 Vitamin E, 225 Volatility, 36 Xanthan gum, 140 Xanthine oxidase, 21 7-218 Xanthine, 2 1 7 Yoghurt, 14