Recent Developments in the Synthesis of Fatty Acid Derivatives

Recent Developments in the Synthesis of Fatty Acid Derivatives

Editors

Gerhard Knothe

National Center for Agricultural Utilization Research Agricultural Research Service U.S. Department of Agriculture Peoria, Illinois, U.S.A.

Johannes T.P. Derksen

Agrotechnological Research Institute (ATO-DLO) Wageningen, The Netherlands

Champaign, Illinois

Copyright © 1999 by AOCS Press

AOCS Mission Statement

To be a forum for the exchange of ideas, information, and experience among those with a professional interest in the science and technology of fats, oils, and related substances in ways that promote personal excellence and provide high standards of quality.

AOCS Books and Special Publications Committee

E. Perkins, chairperson, University of Illinois, Urbana, Illinois J. Derksen, Agrotechnological Research Institute, Wageningen, the Netherlands N.A.M. Eskin, University of Manitoba, Winnipeg, Manitoba J. Endres, Fort Wayne, Indiana T. Foglia, USDA—ERRC, Wyndmoor, Pennsylvania L. Johnson, Iowa State University, Ames, Iowa S. Koseoglu, Texas A&M University, College Station, Texas H. Knapp, University of Iowa, Iowa City, Iowa J. Lynn, Congers, New York M. Mathias, USDA-CSREES, Washington, D.C. M. Mossoba, Food and Drug Administration, Washington, D.C. G. Nelson, Western Regional Research Center, San Francisco, California F. Orthoefer, Monsanto Co., St. Louis, Missouri J. Rattray, University of Guelph, Guelph, Ontario A. Sinclair, Royal Melbourne Institute of Technology, Melbourne, Australia G. Szajer, Akzo Chemicals, Dobbs Ferry, New York B. Szuhaj, Central Soya Co., Inc., Fort Wayne, Indiana E. Whittle, University of Georgia, Athens, Georgia L. Witting, State College, Pennsylvania

Copyright © 1999 by AOCS Press. All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means without written permission of the publisher.

The paper used in this book is acid-free and falls within the guidelines established to ensure permanence and durability.

Library of Congress Cataloging-in-Publication Data

Recent developments in the synthesis of fatty acid derivatives/editors, Gerhard Knothe and Johannes T.P. Derksen. p. cm. Includes bibliographical references and index. ISBN 1–893997–00–6 1. Fatty acids—Synthesis. 2. Fatty acids—Derivatives—Synthesis. I. Knothe, Gerhard. II. Derksen, Johannes T.P. QD305.A2R29 1999 457′.770459—dc2l Printed in the United States of America with vegetable oil-based inks.

03 02 01 00 99 5 4 3 2 1

Copyright © 1999 by AOCS Press

99-28239 CIP

Preface This monograph grew out of a symposium “Synthesis of Novel Fatty Acid Derivatives” held at the 89th AOCS Annual Meeting in Chicago, Illinois, May 1998. An overview of recent synthetic work in the field of fatty acid chemistry is developed, and directions for future work are pointed out. Diverse derivatives and reactions are considered, with a strong representation of epoxidation as a result of its salient position in the field. This book is the first to result from symposia held by the new Industrial Oil Products Division of AOCS. The IOP Division will continue to foster dissemination of new results on synthetic work in fatty acid chemistry by regularly holding symposia in the future. This book would not have been possible without the authors who graciously wrote the various chapters. We thank all of them for their outstanding contributions. We also thank AOCS Press staff for cooperation in bringing everything into print. And we hope that readers, be they students or experienced researchers, will find interesting and exciting aspects in the book and that it will help stimulate further developments in the field. Gerhard Knothe Johannes T.P. Derksen

iii Copyright © 1999 by AOCS Press

Contents Chapter 1

Chapter 2

Chapter 3

Chapter 4 Chapter 5

Chapter 6 Chapter 7

Chapter 8 Chapter 9 Chapter 10 Chapter 11

Chapter 12 Chapter 13

Chapter 14

Preface...........................................................................................iii Novel Long-Chain Compounds Produced Through Neighboring Group Participation.........................................................................1 F.D. Gunstone Synthesis of Special Fatty Acids: A Review................................20 Marcel S.F. Lie Ken Jie and Sunny W.H. Cheung Derivatives of New Crop Oils44 Terry A. Isbell Palladium (0)-Catalyzed Reactions of Nucleophiles with Allyl Carbonates of Unsaturated Fatty Acids.......................................59 Hans J. Schäfer and Michael Zobel Synthesis of New Oleochemicals: Products of Friedel-Crafts Reactions of Unsaturated Fatty Compounds...............................80 Ursula Biermann and Jürgen O. Metzger Synthesis of New Oleochemicals: Functionalization of Fatty Compounds Using Radical Reactions.........................................90 Jürgen O. Metzger, Ralf Mahler, Gerald Francke, and Ahlke Hayen Organic Chemical Synthesis and Derivatization of Fatty Acids and Methyl Esters Obtained from Ricinus and Dimorphotheca Species. Some Industrially Feasible Preparations of New Fatty Acid–Based Products................................................................100 Piet. M.P. Bogaert, Theodoor M. Slaghek, Herman Feil, and Patrick. S.G. Tassignon Synthesis of Epoxidized Novel Fatty Acids for Use in Paint Applications..............................................................................128 G.J.H. Buisman, A. Overeem, and F.P. Cuperus Synthesis of New Derivatives from Vegetable Oil Methyl Esters via Epoxidation and Oxirane Opening.......................................141 Xavier Pagès-Xatart-Parès, Carine Bonnet, and Odile Morin New Oxidation Methods for Unsaturated Fatty Acids, Esters, and Triglycerides..............................................................................157 Mark Rüsch gen. Klaas and Siegfried Warwel Some Recent Advances in Epoxide Synthesis..........................182 George J. Piazza Cyclic Fatty Acids in Heated Vegetable Oils: Structures and Mechanisms..............................................................................196 Gary Dobson Production of Hydroxy Fatty Acids by Biocatalysis.................213 Ching T. Hou, Tsung M. Kuo, and Alan C. Lanser Some Derivatives of Fatty Compounds for Mass Spectral Structure Determination............................................................227 Gerhard Knothe and William W. Christie Index........................................................................................239 v

Copyright © 1999 by AOCS Press

Chapter 1

Novel Long-Chain Compounds Produced Through Neighboring Group Participation F.D. Gunstone

Scottish Crop Research Institute, Dundee, Scotland

Introduction

Natural fatty acids usually contain a carboxyl group and one or more double bonds with cis configuration. Occasionally there is additional functionality such as hydroxy or epoxy groups. For the most part, these functional groups react independently of each other; however, when two reactive sites are close enough, reaction at one may be influenced by the presence of the other. Sometimes this influence is merely regiospecific, i.e., instead of two products being formed in equal amounts, one is dominant. At other times, both functional groups are involved in the formation of the final product. Neighboring group participation (NGP) generally leads to novel products; these may be of interest and of potential value or they may be tiresome byproducts. The new products frequently contain carbocyclic or heterocyclic (O, N, S) systems and their yields are often solvent dependent. Thus, by appropriate selection of conditions, their formation can be maximized or minimized. This topic was reviewed some years ago (1,2).

Regioselective Effects

Regiospecific effects are illustrated in the results summarized in Tables 1 and 2. Methoxymercuration (discussed later) of methyl oleate gives methyl 9- and 10methoxystearates in high yield (98%) and in equal proportions. When this reaction is applied to ricinoleate (12-hydroxy-9c-18:1, a β-hydroxy alkene) in the form of its acetoxy or methoxy derivative, the two ethers are formed in very different proportions with the 9-methoxy derivative in excess. With the corresponding derivatives of

1 Copyright © 1999 by AOCS Press

2

F.D. Gunstone

isoricinoleate (9-hydroxy-12c-18:1, a γ-hydroxy alkene), the 13-methoxy derivative is formed in excess (3) (Table 1). Methyl 9,10-epoxystearate reacts with boron trifluoride to give the 9- and 10-oxostearates in equal amounts; this is not so, however, with methyl vernolate (12,13epoxyoleate). In addition to some cyclopropanes, which will be discussed later, the 12- and 13-oxo-oleates are formed in differing proportions. The 12-oxo ester predominates when the reaction is conducted in dioxan solution but was not detected when benzene was used as solvent (4) (Table 2).

Reactions Leading to Cyclopropanes

Cyclopropanes may be formed in reactions proceeding through a homoallylic carbocation as intermediate. Such species are easily produced from methyl ricinoleate (12hydroxyoleate) and methyl vernolate (12,13-epoxyoleate); see Scheme 1. Cyclopropanes have been obtained from ricinoleate in reactions of its mesyloxy derivative. MsO is a better leaving group than OH and reaction is more likely to proceed through a carbocation produced in an SN1 reaction. Cyclopropanes have been observed in the reaction of the mesylate with NaOMe/MeOH (56% methoxy cyclopropane), NaOAc/AcOH (30% acetoxy cyclopropane), and CaCO3/aqueous CH3CN (63% hydroxy cyclopropane). Acyclic compounds are also formed, and the cyclopropanes appear only when the reaction solutions are suitably buffered. Cyclic com-

SCHEME 1. Reactions leading to cyclopropanes.

Copyright © 1999 by AOCS Press

Novel Compounds Trough Neighboring Group Participation

3

SCHEME 2. Cyclopropanes from 12-mesyloxyoleate. *Indicates most stable intermediate—product formed under kinetic control; **indicates most stable product—formed under thermodynamic control.

pounds are not apparent in reactions conducted under acidic conditions. The cyclopropane carbocation is more stable than its acyclic forms, and cyclopropanes result when the reaction is under kinetic control. However the acyclic trans derivatives are the most stable compounds, and these dominate in reactions proceeding under thermodynamic control (5). Similar products were obtained from the tosylate of methyl ricinelaidate (6). Cyclic products were not observed in the reaction of 12-mesyloxyoleate with LiCl, MgBr2, KI, NaHS, or KSAc, nor with similar reactions of 9-MsO 12c-18:1. This last ester cannot form a homoallylic carbocation (5); see Scheme 2. A homoallylic carbocation can also be made from methyl vernolate (12,13epoxyoleate) and again this acts as a precursor of cyclopropane compounds. As already indicated, methyl epoxystearate and BF3 etherate interact to give 9- and 10-oxostearates. When this reaction was applied to methyl vernolate in boiling dioxan, the expected oxo esters were accompanied by an oxo-cyclopropane ester, which was shown to have the structure indicated in Scheme 3. The yield of cyclopropane was increased when the reaction was conducted in benzene solution (Table 2). The keto esters were mixtures of cis and trans isomers (4). The virtual absence of 12-oxo-oleate from the reaction conducted in benzene was confirmed by Ward and van Dorp (7) who recommended the reaction as a source of 13-oxo-oleate and its reduction product 13-hydroxyoleate. Much of this work was also confirmed by an Italian group (8).

Reactions Leading to Cyclopentanes

Several important metabolites of polyunsaturated fatty acids (PUFA) contain a cyclopentane ring. Compounds such as jasmonic acid, cucurbic acid, and tuberonic

Copyright © 1999 by AOCS Press

4

F.D. Gunstone

SCHEME 3. Reaction of 12,13-epoxyoleate (vernolate) with BF3-etherate.

acid (all Cl2), 12-oxophytodienic acid (C18), and PGF2α (C20) (Scheme 4) are of this type. They are produced biologically from appropriate PUFA by sequences of oxidative reactions. It was considered that cyclopentane systems might be formed from linoleate and related polyene acids either by the polar reaction sequence shown in Scheme 5 or its radical equivalent. When linoleate was reacted with acetic anhydride in a radical addition process promoted by ditertiarybutyl peroxide, 1:1 and 1:2 products were obtained. The former, after reaction with acidic methanol, were mainly unsaturated diesters such as [1], but some saturated compounds were also present and these may have been the cyclopentane derivatives ([2] and [3]) (9).

SCHEME 4. Some natural long-chain acids with a cyclopentane ring system.

Copyright © 1999 by AOCS Press

Novel Compounds Trough Neighboring Group Participation

5

SCHEME 5. Possible route to cyclopentane acids.

The conversion of appropriate dioxostearates to keto cyclopentenes and keto cyclohexenes will be discussed later.

Reactions Leading to Tetrahydrofurans and Tetrahydropyrans

There are several reactions (bromination, oxymercuration, epoxidation) of olefinic hydroxy and epoxy compounds in which the oxygen function can react intramolecularly with an intermediate formed at the double bond, leading to an oxygen-containing heterocyclic compound (tetrahydrofuran or tetrahydropyran) (Scheme 6). Interesting features of these reactions include the following:

Intramolecular processes will generally dominate over intermolecular reactions. i When the solvent is also the reagent, the reaction may be modified by replacing this with a nonreactive solvent, thus allowing intramolecular reactions only. i The reaction is dependent on the relative position of the double bond and the oxygenated function and may also give different results for cis and trans olefinic centers. i

From Linoleate by Partial Hydration

Reaction of linoleate with toluene-4-sulfonic acid could conceivably be a method of hydrating one or both of the double bonds. In fact, the major products were the 9,12and 10,13-tetrahydrofurans (44%) accompanied by the 9,13-tetrahydropyran (4%) and smaller amounts of isomeric tetrahydrofurans (8,11-, 10,13-, and 11,14-) and tetrahydropyrans (8,12- and 10,14-) (Scheme 7). These compounds were readily identified by a combination of chromatographic and spectroscopic techniques (10). Similar products were obtained with the trans isomers of linoleate (but not with the conjugated 9,11- and 10,12-dienoates) and from several oxygenated monoene esters

Copyright © 1999 by AOCS Press

6

F.D. Gunstone

SCHEME 6. Reactions leading to tetrahydrofurans and tetrahydropyrans. (A) and (B) represent intermediates in the reaction of β- and γ-hydroxy alkenes, respectively. X+ represents Br (bromination), HgOAc (oxymercuration), or O (pronated epoxide).

including methyl 12-hydroxyoleate (ricinoleate), 12-hydroxyelaidate, 12-methoxyoleate, and 9-hydroxy-12-octadecenoate (isoricinoleate). When methyl linoleate is reacted with 50% BF3-MOH at 0−5°C, the product contains four monomethoxy 18:1 esters as expected but only two dimethoxy 18:0 esters (9,12 and 10,13). This restricted range of dimethoxy 18:0 esters is explained in terms of NGP in the second stage of the reaction (11). From Polyhydroxy and Hydroxy Epoxy Compounds

Appropriate di- and trihydroxy acids are dehydrated by refluxing with methanolic sulfuric acid. The 1,4- and 1,5-diols give tetrahydrofurans and tetrahydropyrans, respectively. Thus 9,12-dihydroxystearic acid gives the 9,12-tetrahydrofuran as a mixture of cis and trans isomers. (Under similar conditions discussed below, the

SCHEME 7. Cyclic ethers from linoleate. Conditions: reagents, TsOH, MeOH, or dioxan; 100°C; 18 h.

Copyright © 1999 by AOCS Press

Novel Compounds Trough Neighboring Group Participation

7

9,12-di-oxo acid gives the corresponding furan.) Ricinoleic acid can be hydroxylated to 9,10,12-trihydroxystearic acid in four enantiomeric forms. Dehydration of each gives a 10-hydroxy-9,12-tetrahydrofuran. Similarly, isoricinoleic acid (9-OH 1218:1) gives four enantiomeric 9,12,13-trihydroxy acids, which are dehydrated to stereoisomeric 13-hydroxy-9,12-tetrahydrofurans with no evidence of a tetrahydropyran among the products (Scheme 8). These reactions involve many interesting problems of stereochemistry that are elaborated in the full paper (12). Appropriate hydroxy epoxides are also converted to hydroxytetrahydrofurans and, indeed, in some cases, the hydroxy epoxide could not be isolated but passed directly to the hydroxytetrahydrofuran. Methyl ricinoleate, epoxidized with 3chloroperoxybenzoic acid, gave the expected cis-epoxide, which was converted by BF3, to a 10-hydroxy-9,12-tetrahydrofuran. Similar reaction with methyl isoricinoleate followed a different course, i.e., the expected 12,13-epoxide could not be isolated but was transformed directly to a mixture of diastereoisomeric 13-hydroxy9,12-tetrahydrofurans. Similarly 12,13-dihydroxyoleate was converted directly to the 9,12-dihydroxy-10,13-tetrahydrofuran. These results illustrate the difference between β- and γ-hydroxy alkenes. The former (e.g., methyl ricinoleate) give epoxides that can be converted to tetrahydrofurans, whereas with the latter (e.g., methyl isoricinoleate), it is not possible to isolate the epoxide that passes directly to a tetrahydrofuran. The 12,13-dihydroxyoleate is both a β- and a γ-hydroxy alkene, and the reaction is controlled by the more reactive γ-hydroxy group (13); see Scheme 9. In a more recent application of this type of reaction, Capon et al. (14) converted 1,10,13-nonadecatriene to a mixture of stereoisomeric bisepoxides that formed acetoxyhydroxytetrahydrofurans when treated with acetic acid at 100°C. These were hydrolyzed to dihydroxrytetrahydrofurans, some stereoisomers of which had previously been recognized as metabolites of the marine brown alga Notheia anomala. In essence, this series of reactions converts a 1,4-diene to a dihydroxytetrahydrofuran via the bisepoxides in a high-yield process; see Scheme 10.

SCHEME 8. Dehydration of polyhydroxy acids. Conditions:MeOH-H2SO4 (1.5 M, 6 h, reflux).

Copyright © 1999 by AOCS Press

8

F.D. Gunstone

SCHEME 9. Rearrangement of hydroxy epoxides. Epoxidation: MCPBA or HCO3H; rearrangement: BF3 etherate/dioxan.

Halogenation of Unsaturated Hydroxy Esters

Halogenation of ricinoleate and of isoricinoleate proceeds throug chloronium, bromoniun, or iodinium intermediates such as those formulated in Scheme 6. These will have some carbocation character, and this will change in the order iodine Ɑ bromine Ɑ chlorine. Consequently, one might expect tetrahydrofurans to be formed. Cyclic ethers were not observed during chlorination but were present among the products of bromination and iodochlorination. The results, summarized in Table 3, confirm that there is more cyclic ether in the iodination process than in bromination and that the cyclized product is greater with the γ-hydroxy alkene (isoricinoleate) than with the β-hydroxy alkene (ricinoleate). There is also some indication that the trans alkene gives more cyclized product than its cis isomer (15). These tetrahydrofuran derivatives have the structures shown in Scheme 6. Siouffi and

SCHEME 10. Biometric synthesis of marine epoxy lipids. *Each structure represents two isomeric forms. R = CH3(CH2)4; R⬘ = (CH2)7CH=CH2. (i) m-CPBA; (ii) AcOH, 100°C, 16 h, R″= OAc; (iii) MeOH-NH4OH, R″ = OH.

Copyright © 1999 by AOCS Press

Novel Compounds Trough Neighboring Group Participation

9

Naudet (16) also obtained 9,12-tetrahydrofurans from methyl ricinoleate and its acetate by reaction with tert-butyl hypobromite in methanol. Oxymercuration of Olefinic Oxygenated Esters

The most common oxymercuration reaction involves alkene, an alcohol such as methanol, and a mercury salt such as the acetate. The intermediate (alkoxy mercuriacetate) is easily reduced by alkaline NaBH4 to the alkoxy derivative (Scheme 11). When applied to methyl oleate, this reaction gives a mixture of methyl 9- and 10methoxystearates in high yield. Methoxymercuration (with methanol) of a range of hydroxy alkenes is summarized in Table 4 (17) from which the following generalizations are apparent:

With β-hydroxy alkenes, there is a difference between the cis and trans isomers. The cis isomers give little or no tetrahydrofuran; with the trans isomers, however, the cyclic ether is virtually the only product, even in the presence of a participating solvent such as methanol. i γ-Hydroxy alkenes, both cis and trans isomers, give high yields of tetrahydrofurans; in molecules containing both β- and γ-hydroxy groups, the latter control the process. i δ-Hydroxy alkenes, which cannot form tetrahydrofurans, furnish tetrahydropyrans. This occurs with cis and trans octadec-5-enol. i Interesting results were reported with the alcohols prepared from arachidonic acid (AA) (∆5, a δ-hydroxy alkene) and from docosahexaenoic acid (DHA) (∆4, a γ-hydroxy alkene). Oxymercuration with methanol would give a complex mixture of products because of the large number of double bonds that would react. Simpler products result with nonparticipating solvents such as N,N-dimethylformamide (DMF); the reaction is then confined to the intramolecular process, which gives a cyclic ether with a polyunsaturated side chain ([4] and [5]). When the reaction is carried out on mixtures containing only modest levels of AA i

SCHEME 11. Oxymercuration. (i) Hg(OAc)2; (ii) NaOH, NaBH4.

Copyright © 1999 by AOCS Press

10

F.D. Gunstone

and/or DHA, the cyclic ethers produced from these materials are easily separated and distinguished from unreacted alcohols. This reaction has some resemblance to iodolactonization, which will be discussed below.

Oxymercuration of several 18:2 esters also gave cyclic ethers when carried out in water rather than methanol. This is not difficult to understand because the first (intermolecular) stage in the reaction should produce hydroxy monoenes some of which react further in an intramolecular fashion. Methyl linoleate (9c 12c), for example, yielded a mixture of 9,12- and 10,13-tetrahydrofurans (46%) along with several dihydroxystearates (52%). Some of the latter were subsequently cyclized by acid, raising the yield of cyclic ethers to ~70%. The isomeric 8c 12c diene gave 9,12tetrahydrofuran (23%), 8,12- and 9,13-tetrahydropyrans (36%), and dihydroxystearates (14%).

Copyright © 1999 by AOCS Press

Novel Compounds Trough Neighboring Group Participation

11

Oxidative Cyclization of Hydroxy Compounds

Saturated (and some unsaturated) long-chain hydroxy compounds can be converted to tetrahydrofurans by oxidation with reagents that convert the alcohol (ROH) to its alkoxide radical (RO•). These include lead tetra-acetate, silver oxide/bromine, mercuric oxide/iodine, sodium hypochlorite (domestic bleach), and N-iodosuccinimide (Scheme 12). Methyl 12-hydroxystearate gives a mixture of 9,12- and 12,15-tetrahydrofurans in 30–60% yield, depending on the reagent used; 9-hydroxy ester gives the mixed 6,9- and 9,12-tetrahydrofurans in similar yields (13). We also prepared a series of 2alkyltetrahydrofurans from the C10 to C18 alkanols (18). More recently, Isbell and Plattner (19) examined a similar reaction with the ∆5 acids in meadowfoam oil, giving the corresponding lactones. When the lead tetra-acetate reaction was applied to the olefinic hydroxy esters (ricinoleate and isoricinoleate), there was an interesting difference in the results. Isoricinoleate, reacting via a γ-olefinic alkoxy radical, gave the 13-acetoxy-9,12-tetrahydrofuran (28%) and the 12-acetoxy-9,13-tetrahydropyran (32%). Ricinoleate, reacting via a β-olefinic alkoxy radical, underwent cleavage to C7 (heptanal) and C11 products (methyl 11-acetoxy-trans-9-undecenoate and methyl 9-acetoxy-10-undecenoate in a 4:1 ratio) (13). We found our 2-alkyltetrahydrofurans to be flexible intermediates for the production of a range of interesting compounds. These are summarized in Table 5 (18).

Reactions Leading to Furans

Furanoid acids/esters such as [6] can be produced from conjugated dienes and from a range of saturated or unsaturated oxygenated acids (hydroxy, epoxy, oxo). These include the following reactions:

SCHEME 12. Oxidative cyclization of hydroxy compounds.

Copyright © 1999 by AOCS Press

12

i

i i i i i

F.D. Gunstone

Reaction of trans, trans-conjugated dienes with singlet oxygen and subsequent dehydration of cyclic peroxide to cyclic ether with iron (II) sulfate. This has been used to convert the 8,10- and 9,11-dienes to the corresponding furanoid acids (20,21). Pd(II)Cl2 and Cu(II)Cl oxidation of 12-hydroxy- and 11,12-epoxystearolic acid (9a-18:1) to give the 9,12-furan and of 12,13-dihydroxy- and 12,13-epoxyoleic acid to give the 10,13-furan (20). Oxidation by alkaline permanganate converts 12-oxostearolate to the 9,12-furan. Oxymercuration of 12-oxo-oleate gives the 9,12-furan (22,23). Reaction of 9,10-epoxy-12c-18:1, 9,10-epoxy-12-oxo-18:0, or 9,10; 12,13diepoxy-18:0 with PrI, NaI, in dimethylsulfoxide (DMSO) as solvent (23). Acidic dehydration of 1,4-diketo acids, as in the conversion of 9,12-dioxostearic acid to the 9,12-furanoid C18 acid, effected with methanolic BF3 (10). 1,4-Diketones, made from ricinoleate, linoleate, or diynoic esters by Lie Ken Jie and his colleagues (22–25), were converted to furanoid acids under acidic conditions (BF3 or toluenesulfonic acid). With 1,4-, 1,5-, and 1,6-diketones, ketocyclopentenes and ketocyclohexenes were also obtained. Typical products ([7]– [9]) are formulated:

Reactions Leading to Lactones

Lactones are easily made from hydroxy acids and from unsaturated acids. As expected, the lactones with five- (γ-lactone) and six- (δ-lactone) membered rings are

Copyright © 1999 by AOCS Press

Novel Compounds Trough Neighboring Group Participation

13

best known and most easily made, although lactones with larger rings occur naturally and can be made synthetically. Oleic acid (∆9) forms γ-stearolactone when treated with acid. The double bond is protonated reversibly and moves until the carbocation on C4 is captured by the carboxyl group. With ∆5 acids (such as the C20 acids in meadowfoam oil), the δ lactone is formed in high yield when the acid is refluxed for 3 h with perchloric acid in dichloromethane. The proportion of δ to γ lactone depends on the reaction solvent. In hexane, the ratio is ~6:1 and in dichloromethane almost 40:1. The δ-lactone is much more reactive than its γ isomer. It can be converted to hydroxy amides (amines), hydroxy acids (alkali), and alkoxy esters (alcohol and acid) with the reagents indicated in parentheses. With sodium hypochlorite (household bleach), the olefinic acid gives a chlorolactone or a range of acyclic chlorine-containing acids (19,26,27).

Iodolactonization

Appropriate olefinic acids react with KHCO3, I2, and KI to give iodolactones. Acids with a ∆4 double bond such as DHA form an iodo γ-lactone; acids with a ∆5 double bond such as AA or EPA form an iodo δ-lactone (Scheme 13). There are several interesting features about these reactions:

SCHEME 13. Iodolactonization. (i) I2, KI, KHCO3; EtOH at 25°C or tetrahydrofuran (THF) at −2 to +6°C; (ii)TMSCl and NaI in CH3CN.

Copyright © 1999 by AOCS Press

14

F.D. Gunstone

The iodolactones are neutral molecules and are therefore easily separated from unreacted acids. This provides the basis of a method for separating and isolating ∆4 and ∆5 acids from acids that do not have unsaturation at these positions. i The original unsaturated acids can be recovered from their iodolactones by reaction with trimethylsilyl iodide (or Me3SiCl and NaI). i In general, the γ-lactones are more easily formed than the δ-lactones, but the latter are more reactive. i The iodolactonization reaction is dependent on solvent, reaction temperature, and on the ratio of iodine to iodide. For example, with ethanol as solvent and a reaction temperature of 25°C, DHA gives a maximum yield of γ-iodolactone in 10 min, but EPA requires 90 min to produce the δ-iodolactone. i

Following earlier reports, the reaction was examined by a Russian group (27,28). More recently, the same group have used iodolactonization to identify new ∆4 and ∆5 acids in the freshwater sponge Baicalospongia baciliferai (30).

Reactions Leading to Sulfur Heterocycles

Long-chain hydroxy compounds may be converted to thiols in two ways, based on reactions of the corresponding mesylate. Direct reaction with sodium hydrogen sulfide converts ROMs to RSH but the thiols are very readily oxidized to disulphides (RSSR); for the successful preparation of thiols, it is necessary to carry out the reaction in a reducing atmosphere. The disulfide results from the dimerization of the thiol radical (RS). Alternatively, the mesylate can be reacted with potassium thioacetate to give the acetylthio derivative (RSAc), which gives the thiol on acidic (MeOHH2SO4) or basic (NaOMe-MeOH) hydrolysis (31,42) (Scheme 14). When these reactions are applied to a β-hydroxy alkene such as methyl ricinoleate, then the thio and acetylthio derivatives are obtained. With a γ-hydroxy alkene (methyl isoricinoleate), the acetylthio derivative is formed as expected, but the thiol is easily oxidized to give the 9,12-thioepoxide (70–75%) as major product. Octadec-4-enol (another γ-hydroxy alkene) gives both thiol and 1,4-cyclic sulfide, whereas the δhydroxy alkene (octadec-5-enol) gives the 5-thiol, which can be oxidized by iodine to the 1,5-cyclic sulfide (Scheme 15). Interesting results were obtained when these reactions were applied to the 9,10-, 9,12-, and 10,12-dihydroxystearates and to 12,13-dihydroxyoleate. The products

SCHEME 14. Formation of thiols and disulfides. (i) NaHS, RT; (ii) KSAc, 100°C; (iii) MeOH, H2SO4.

Copyright © 1999 by AOCS Press

Novel Compounds Trough Neighboring Group Participation

15

SCHEME 15. Formation of cyclic sulfide from mercapto alkene.

included dithiols, 1,4-thioepoxides, and 1,4- and 1,5-dithioepoxides as shown in Scheme 16. Lie Ken Jie and Zheng (32) converted ricinoleate to the 12-oxo-9,10-epoxide and reacted this with dimethylthioformamide and trifluoroacetic acid. The product was a mixture of oxo thioepoxide, the 9,12-furanoid acid, and the corresponding thiophene acid. The same group converted a series of conjugated octadecadiynoates (6,8-, 7,9-, 8,10-, 9,11-, and 10,12-) to heterocyclic compounds having selenium or tellurium as the hetero atom through reaction with silver acetate, methanol or ethanol, and NaHSe or Na2Te [10] (33,34).

Reactions Leading to Nitrogen Heterocycles

A range of N-heterocycles have been prepared by Lie Ken Jie and his Hong Kong group from dioxostearates or from ricinoleate or isoricinoleate. 10,12-Dioxostearate

SCHEME 16. Formation of cyclic sulfur compounds from the diols indicated. The last two structures pictured are formed from the 9,12-diol via the SH/OMs and bis-SAc derivatives, respectively.

Copyright © 1999 by AOCS Press

16

F.D. Gunstone

SCHEME 17. Pyrazoles and pyridazines from 10:12- and 9:12-dioxostearates.

reacts with hydrazine to give a fatty acid containing a pyrazole unit (35), whereas the 9,12 isomer gives a pyridazine derivative (36). Hydrazine may be replaced by substituted hydrazines (such as PhNHNH2) to give N-substituted derivatives. These reactions occur in water under moderate conditions and in high yield when conducted with ultrasound (Scheme 17). The 9,12-dioxostearate gives a pyrrole derivative when reacted with ammonium acetate in acetic acid solution (37). Ricinoleate and isoricinoleate can be converted to N-heterocyclic compounds via their azides. For example, methyl isoricinoleate, as its tosyloxy or (better) mesyloxy ester, reacts with sodium azide to give the corresponding azide at 50°C. At a reaction temperature of 80°C, the azide loses nitrogen to yield a dihydropyrrole that can be reduced with NaBH4 to the E/Z tetrahydropyrroles (38) (Scheme 18). Methyl ricinoleate was converted to 12-azido-9-oxostearate by two routes. This gave the dihydropyrrole (1-pyrrolidine) when reacted with Ph3P in tetrahydrofuran (THF) with ultrasound for 30 min (39) (Scheme 19). Methyl isoricinoleate also furnished N-heterocycles via its azido epoxide through reaction with MeI, NaI, and DMF. After 20 h, the product was a mixture of pyrrole and pyridines formed via the two possible keto intermediates. After only 15 min reaction time, the keto azides were isolated and subsequently converted to a dihydropyrrole and a tetrahydropyridine through reaction with PhP3 (40) (Scheme 20).

SCHEME 18. N-Heterocycles from methyl isoricinoleate produced via the azide. (i) MsCI, NEt3; (ii) NaN3, N,N-dimethylformamide (DMF), 50°C; (iii) 80°C; (iv) NaBH4, MeOH.

Copyright © 1999 by AOCS Press

Novel Compounds Trough Neighboring Group Participation

17

SCHEME 19. N-Heterocycles from methyl ricinoleate produced via the azide. (i)Ph3P, tetrahydrofuran (THF), ultrasound.

SCHEME 20. N-Heterocycles from methyl isoricinoleate produced via the azide. (i) MeI, NaI, N,N-dimethylformamide (DMF), 15 min; (ii) PhP3; (iii) as for (i), but for 20 h.

References

1. Gunstone, F.D., New Reactions of Long-Chain Unsaturated Acids Especially Those Containing an Oxygen Function, Riv. Ital. Sostanze Grasse 48: 510–519 (1971). 2. Gunstone, F.D., Some New Chemical Properties of Unsaturated C18 Acids, Acc. Chem. Res. 9: 34–40(1976). 3. Gunstone, F.D., and R.P. Inglis, Applications of the Oxymercuration-Demercuration Reaction to Long-Chain Unsaturated Esters, Chem. Phys. Lipids 10: 73–88 (1973). 4. Conacher, H.B.S., and F.D. Gunstone, The Rearrangement of Methyl 12,13-Epoxyoleate by Boron Trifluoride with Formation of Cyclopropane Esters, Chem. Phys. Lipids 3: 203– 220 (1969). 5. Gunstone, F.D., and A. I. Said, Methyl 12-Mesyloxyoleate as a Source of Cyclopropane Esters and of Conjugated Octadecadienoates, Chem. Phys. Lipids 7: 121–134 (1971). 6. Ucciani, E., A.Vantillard, and M. Naudet, Cyclisation du Ricinelaidate de Methyle par Solvolyse de son Tosylate, Chem. Phys. Lipids 4, 225–237 (1970). 7. Ward, J. P., and D.A. van Dorp, Synthesis of Methyl DL-13-Hydroxy-9-cisOctadecenoate and Methyl DL-19-Hydroxy-all cis-8,11,14-Eicosatrienoate, Recl. Trav. Chim. Pays-Bas 88: 1345–1357 (1969). 8. Canonica, L., J. Ferrari, J.M. Pagnoni, S. Maroni, and T. Salvatori, Tetrahedron 25: 1 (1969). 9. Gunstone, F.D., and R.G. Powell, An Addition Reaction of Methyl Linoleate Accompanied by Cyclisation, Chem. Phys. Lipids 2: 203–212 (1968).

Copyright © 1999 by AOCS Press

18

F.D. Gunstone

10. Abbot, G.G., F.D. Gunstone, and S.D. Hoyes, The Formation of 1,4-Epoxides from Methyl Linoleate and Related Esters by Reaction with Toluene-p-sulphonic Acid and an Appropriate Solvent, Chem. Phys. Lipids 4: 351–366 (1970). 11. Carballeira, N.M., M.V. Gonzalez, and M. Pagan, Neighboring Methoxyl Participation in the Acid Catalyzed Methoxylation of Methylene-Interrupted Fatty Acids, Chem. Phys. Lipids 89: 91–96 (1997). 12. Abbot, G.G., and F.D. Gunstone, The Formation of 1,4-Epoxides from Two Series of Trihydroxystearic Acids by Acid-Catalyzed Cyclisation, Chem. Phys. Lipids 7: 279–289 (1971). 13. Abbot, G.G., and F.D. Gunstone, The Formation of Some Substituted Vic-Epoxyoctadecanoates and Their Conversion to 1,4-Epoxides and Other Compounds, Chem. Phys. Lipids 7: 290–302 (1971). 14. Capon, R.J., R.A. Barrow, C. Skene, and S. Rochfort, The Biomimetic Synthesis of Marine Epoxylipids: Bisepoxides to Tetrahydrofurans, Tetrahedron Lett. 38: 7609–7612 (1997). 15. Gunstone, F.D., and B.S. Perera, The Halogenation of Some Long-Chain Hydroxy Alkenes with Special Reference to the Possibility of Neighbouring Group Participation Leading to Cyclic Ethers and Lactones, Chem. Phys. Lipids 11: 43–65 (1973). 16. Siouffi, A.M., and M. Naudet, Alcoxyhalogenation d’Acides Gras Ethyleniques III. Cas de l’Hydroxy-12-octadecene-9-oate de Methyle et de Son Acetate, Chem. Phys. Lipids 11: 103–113(1973). 17. Gunstone, F.D., and R. P. Inglis, Oxymercuration-Demercuration of Unsaturated Alcohols. A New Procedure for the Identification, Isolation, and Estimation of ∆3t, ∆4, and ∆5 Alkenols, Chem. Phys. Lipids 10: 105–113 (1973). 18. Dawes, M., Ph.D. Thesis, University of St. Andrews, Scotland, 1995. 19. Isbell, T.A., and B.A. Plattner, A Highly Regioselective Synthesis of δ-Lactones from Meadowfoam Fatty Acids, J. Am. Oil Chem. Soc. 74: 153–158 (1997). 20. Gunstone, F.D., and R.C. Wijesundera, Some Reactions of Long-Chain Oxygenated Acids with Special Reference to Those Furnishing Furanoid Acids, Chem. Phys. Lipids 24: 193–208 (1979). 21. Bascetta, E., F.D., Gunstone, and C.M. Scrimgeour, Synthesis, Characterisation, and Transformations of a Lipid Cyclic Peroxide, J. Chem. Soc. Perkin Trans. I, 2199–2205 (1984). 22. Lie Ken Jie, M.S.F., and C.H. Lam, Synthesis of Furanoid Esters from Naturally Occurring Unsaturated Fatty Esters, Chem. Phys. Lipids 20: 1–12 (1977). 23. Lie Ken Jie, M. S. F. and C. H. Lam, The Synthesis Of Furanoid, αβ-Unsaturated Oxocyclopentenyl and Oxocyclohexenyl Esters from Methyl Octadecadiynoates. Chem. Phys. Lipids 19: 275–287 (1977). 24. Lie Ken Jie, M.S.F., and F. Ahmad, Partial Synthesis of C18 Mono and Dimethyl Furanoid Fatty Acids, J. Am. Oil Chem. Soc. 60: 1783–1785 (1983). 25. Lie Ken Jie, M.S.F., and K.P. Wong, Synthesis of Phenyl Substituted C18 Furanoid Fatty Esters, Lipids 28: 43–46 (1993). 26. Isbell, T.A., Lipid Technol. 9: 140–144 (1997). 27. Isbell, T.A., and B.A. Steiner, The Rate of Ring Opening of γ- and δ-Lactones Derived from Meadowfoam Fatty Acids, J. Am. Oil Chem. Soc. 75: 63–66 (1998). 28. Imbs, A.B., D.V. Kuklev, A.D. Vereshchlagin, and N.A. Latyshev, Application of an Analytical Modification of the Iodolactonization Reaction to Selective Detection of ∆5

Copyright © 1999 by AOCS Press

Novel Compounds Trough Neighboring Group Participation

19

(∆4) Unsaturated Fatty Acids, Chem. Phys. Lipids 60: 71–76 (1991). 29. Gaiday, N.V., A.B. Imbs, D.V. Kuklev, and N.A. Latyshev, Separation of Natural Polyunsaturated Fatty Acids by Means of Iodolactonisation, J. Am. Oil Chem. Soc. 68: 230– 233(1991). 30. Imbs, A.B., and N.A. Latyshev, New ∆4 and ∆5 Unsaturated Medium- and Long-Chain Fatty Acids in the Freshwater Sponge Baicalospogia bacilifera, Chem. Phys. Lipids, 92: 117–125 (1998). 31. Gunstone, F.D., M.G. Hussain, and D.M. Smith, The Preparation and Properties of Methyl Monomercaptostearates, Some Related Thiols, and Some Methyl Epithiostearates, Chem. Phys. Lipids 13: 71–91 (1974). 32. Lie Ken Jie, M.S.F., and Y.F. Zheng, A Convenient Route to a C18 Thiophene Fatty Acid Derivative via a Keto Epithio Intermediate, Synthesis 467–468 (1988). 33. Lie Ken Jie, M.S.F., and Y.K. Cheung, Synthesis and Nuclear Magnetic Resonance Studies of a Series of Synthetic Long-Chain Selenophene Fatty Esters, J. Chem. Res. (S), 392 and (M), 2745–2778 (1993). 34. Lie Ken Jie, M.S.F., and S.H. Chau, Synthesis and Nuclear Magnetic Resonance Studies of a Series of Synthetic Long-Chain Tellurophene Fatty Esters, J. Chem. Res. (S) 428 and (M) 2642–2657 (1995). 35. Lie Ken Jie, M.S.F., and P. Kalluri, Synthesis of Pyrazole Fatty Ester Derivatives: A Sonochemical Approach, J. Chem. Soc. Perkin Trans. I, 1205–1206 (1995). 36. Lie Ken Jie, M.S.F., and P. Kalluri, Synthesis of Pyridazine Fatty Ester Derivatives: A Sonochemical Approach, J. Chem. Soc. Perkin Trans. I, 3485–3486 (1997). 37. Lie Ken Jie, M.S.F., and K.P. Wong, Synthesis of Trisubstituted C18 Pyrrole Fatty Ester Derivatives, Lipids 28: 161–162 (1993). 38. Lie Ken Jie, M.S.F., and M.S .K. Syed-Rahmatullah, Synthesis and Properties of a Novel 1-Pyrroline Fatty Acid Derivative from Methyl Isoricinoleate, J. Chem. Soc. Perkin Trans. I, 421–424 (1991). 39. Lie Ken Jie, M.S.F., M.S.K. Syed-Rahmatullah, C.K. Lam, and P. Kalluri, Ultrasound in Fatty Acid Chemistry: Synthesis of a 1-Pyrroline Fatty Acid Ester Isomer from Methyl Ricinoleate, Lipids 29: 889–892 (1994). 40. Lie Ken Jie, M.S.F., and M.K. Pasha, Synthesis of Novel Piperidine and Pyridine Containing Long-Chain Fatty Ester Derivatives from Methyl Isoricinoleate, J. Chem. Soc. Perkin Trans. I, 1331–1332 (1996). 41. Gunstone, F.D., and R.P. Inglis, Neighbouring Group Participation in the Oxymercuration-Demercuration Reaction, Chem. Phys. Lipids 10: 89–104 (1973). 42. Gunstone, F.D., M.G. Hussain, and D.M. Smith, The Preparation and Properties of Some Methyl Dimercaptostearates and Some Related Methyl Epithio- and Epidithiostearates, Chem. Phys. Lipids 13, 92–102 (1974).

Copyright © 1999 by AOCS Press

Chapter 2

Synthesis of Special Fatty Acids: A Review Marcel S.F. Lie Ken Jie and Sunny W.H. Cheung

Department of Chemistry, The University of Hong Kong, Hong Kong, China

Introduction

This review covers work published during the past six years (1993–1998) on the syntheses of an array of special fatty acids and their derivatives. It can be regarded as an extension to the review on the synthesis of rare and unusual fatty acids (1). We have also published two recent reviews on fatty acids, fatty acid analogs, and their derivatives (2,3). This review does not cover the synthesis of prostaglandins, prostacyclins, thromboxanes, and leukotrienes (lipoxins); however, the syntheses of some polyunsaturated hydroxy fatty acids are discussed.

Synthesis of Methyl-Branched Fatty Acids

Methyl-branched fatty acids are found in lipid fractions of many plants and are very common in bacterial extracts. Sixteen branched fatty acids were produced by different synthetic methods, including alkylation and hydrolysis of oxazolines to obtain 2alkyl fatty acids. This was achieved through desulfurization of alkyl-substituted thiophenecarboxylic acids for 4- and 6-alkyl fatty acids and the application of the Kolbe reaction of dioic acids to give alkyl branches at different positions of the chain (4). The synthesis of racemic 2,4-dimethyl-14:0 acid was described starting from 3,5-dimethyl-6-triphenylmethyloxyhexanal, which was chain extended¹ via the Wittig olefination reaction. Trityl deprotection and hydrogenation of the product gave 2,4-dimethyltetradecanol, which was oxidized to the requisite carboxylic acid (5). The synthesis of 15-methyl-16:1(11Z)¹ was achieved via the Wittig reaction of 4methyl-1-pentyltriphenylphosphonium bromide and 10-bromodecanal (6). (4R)Methyl-17:2(7Z,11Z) was synthesized from enantiomerically pure (S)-citronellol as the starting material. The synthesis was achieved by alkylation of 4-pentyn-1-ol with (S)-citronellyl bromide, followed by depyranylation, semihydrogenation over P-2 Ni in ethanol to the (Z)-alcohol and oxidation to the aldehyde. Wittig olefination of the last-mentioned with hexylidenephosphorane gave 2,6-dimethyl-2,9Z, 13Z-nonadecatriene, which was regioselectively epoxidized at the 2-position. Cleavage of the ¹The chain length of a fatty acid is given by the first number followed by a number after the colon, which indicates the number of unsaturation centers. Parenthetical numbers denote the positions of the unsaturated centers; the letters Z, E and A represent cis, trans olefinic, or an acetylenic bond, respectively. R or S, where present, denotes the configuration of the chiral center.

20 Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

21

SCHEME 1. Synthesis of (R)-isomer of ethyl (E)-6-methyl-24:1(4E). Reagents: (i) H2SO4, dioxane, H2O; (ii) oxalyl chloride; (iii) N,O-dimethylhydroxylamine, 4-dimethylaminopyridine, CH2CI2; (iv) LiAIH4, tetrahydrofuran (THF); (v) vinyl magnesium bromide, THF; (vi) triethyl orthoacetate, propionic acid.

epoxide with HIO4 followed by oxidation of the aldehyde intermediate gave the requisite (4R)-methyl-17:2(7Z,11Z) fatty acid (7). The racemic mixture of tuberculostearic acid, 10-methyl-18:0, was obtained by coupling of 1-iodo-2-methyldecane with 1-hydroxyoct-7-yne to give a racemic mixture of 10-methyloctadec-7-yn-1-ol, which was hydrogenated and oxidized to 10methyl-18:0 acid (8). The stereospecific syntheses of 24:1(5Z) and the (R)- and (S)-isomers of ethyl 6-methyl-24:1(4E), as presented in Scheme 1, were accomplished to test their potential role in the biosynthesis of mycolic acids. However, this study showed that only 24:1(5Z) was found to act as a mycolic acid precursor (9). Racemic 2,4-dimethyl-22:0, a major acyl component of 2,3-di-O-acyl-a,αtrehalose glycolipid antigens isolated from Mycobacterium tuberculosis, was synthesized starting from 3,5-dimethylcyclohexan-1-ols (10). Racemic 3-hydroxy-2,4,6-trimethyl-24:0 and racemic 2,4,6-trimethyl-24:1(2E) were synthesized from racemic 2,4-dimethyldocosanol. Oxidation of the latter compound gave 2,4-dimethyldocosanal, which on reaction with methyl propionate afforded racemic methyl 3-hydroxy-2,4,6-trimethyl-24:0. After hydrolysis, the Wittig reaction of 2,4dimethyldocosanal with 1-carboethoxyethylidenetriphenylphosphorane gave the required racemic 2,4,6-trimethyl-24:1(2E) acid (11).

Synthesis of Unsaturated Fatty Acids

The synthesis of long-chain 3(E)-alkenoic acids by the Knoevenagel condensation of aliphatic aldehydes with malonic acid has been described (12). Linolenic acid and the all-Z-isomers of 20:5(5,8,11,14,17), 22:6(4,7,10,13,16,19), and 22:6(4,7,10,13,16,19) were synthesized via the Wittig olefination reaction (13–15). The syntheses of 11-

Copyright © 1999 by AOCS Press

22

M.S.F. Lie Ken Jie and S.W.H. Cheung

substituted linoleic acids ([1]–[3]) starting from an acetylenic C11 acid were reported (16). The synthesis of compound [3] is presented in Scheme 2.

The syntheses of geometric isomers of a number of C20 polyunsaturated fatty acids, viz., 20:5(5Z,8Z,E,14Z,17Z), 20:5(5Z,8Z,11E,14Z,17E), 20:5(5Z,8Z,11Z, 14Z,17E) (Scheme 3), and 22:5(4Z,7Z,10Z,13Z,16Z,19E) were also achieved by chain extension (Wittig olefination) reactions (17,18). Syntheses of very long-chain fatty acid methyl esters involving the treatment of ω-iodo esters with the complexes formed from reactions of alkylcopper(I) and Grignard reagents furnished very long-chain methyl alkanoates, all-(Z) alkenoates, methylene-interrupted alkadienoates, and alkatrienoates (19). A facile access to symmetrical Z-olefins was described from phosphonium salts via an autoxidation process in a salt-free condition (modified Wittig reaction). For example, (8Z)-hexadecenedioic acid and dimethyl (10Z)-eicosenedioate were synthesized from 8-bromooctanoic acid and methyl 11-bromoundecanoate, respectively,

SCHEME 2. Synthesis of methyl 11-substituted linoleoate [3]. Reagents: (i) MeOH, thionyl chloride; (ii) periodinane, CH2CI2; (iii) 1-heptyne, 3M MeMgCI, tetrahydrofuran (THF); (iv) Pd/Pb on CaCO3, H2, pyridine, MeOH; (v) Mn2O, CH2CI2; (VI) Ph3P, MeBr, n-BuLi, hexane.

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

23

SCHEME 3. Synthesis of methyl 20:5(5Z,8Z,11Z,14Z,17E). Reagents: (i) Ph3P, I2, imidazole; (ii) Ph3P, CH3CN, CaCO3; (iii) n-BuLi, E-3-hexenal; (iv) Ph3P, Br2, CH2CI2; (v) 1-bromo-4,4-diethoxybut-1-yne, Mg, CuBr, tetrahydrofuran (THF); (vi) H2, P-2Ni; (vii) trifluoroacetic acid, CH2Cl2; (viii) Ph3P, 5-bromopentanoic acid; (ix) nBuLi, THF, hexa-methylphosphoramide (HMPA); (x) CH2N2.

by reaction of alkyltriphenylphosphonium bromide with oxygen (20). A similar approach for the preparation of symmetric cis-skipped polyenes by combining the controlled classical Wittig reaction or the Oxidative dimerization of phosphoranes has been described (21). The following radiolabeled fatty acids, (9Z,12E)-, (9E,12Z)-[114 C]-linoleic acid, (9Z,12Z,15E)-, (9E,12Z,15Z)-[1-14C]-linolenic acid, and (5Z,8Z,11Z,14E)-[1-14C]-arachidonic acid, were obtained by total synthesis by elongation steps and creation of Z-double bonds via highly stereospecific Wittig reactions. The radioactive carbon atom was introduced from [14C]-potassium cyanide, where the resulting [14C]-nitrile intermediate was hydrolyzed to the requisite radiolabeled carboxylic acid (22). Stereoselective and reductive elimination of 1,8-dibenzoate-2,4,6-trienes with sodium amalgam gave the all-E-tetraenes. This approach was applied to the synthesis of β-parinaric acid, 18:4(9E, 11E, 13E, 15E) (23). Santalbic acid or xymenynic acid, 18:2(9A, 11E), was prepared from ricinoleic acid, 12hydroxy-18:1(9Z). The first step involved the bromination and ultrasound- assisted dehydrobromination of ricinoleic acid, which gave the corresponding hydroxyacetylenic intermediate. The latter compound was mesylated, and demesylation furnished the requisite santalbic acid and its Z-isomer. Santalbic acid was isolated by urea fractionation (24). A new stereocontrolled route to conjugated dienynes has been developed, utilizing organometallic addition to pyrylium salts followed by Wittig hotnologation and dehydrohalogenation. This method was applied to the synthesis of a long-chain polyunsaturated fatty acid, 23:4(4A,6Z,8E,19A), known as Carduusyne

Copyright © 1999 by AOCS Press

24

M.S.F. Lie Ken Jie and S.W.H. Cheung

A (25). Four new C18 ethyl esters, viz., 18:3(5Z,11Z,13E), 18:3(5Z,10Z,12E), 18:3(5Z,9E,11Z), and 18:3(5Z,10Z,12Z), were synthesized; on heating, they underwent cyclization via an intramolecular Diels-Alder reaction (26).

Synthesis of Hydroxy C18, C20 Fatty Acids

Hydroxylation of oleic acid, 18:1(9Z), in the allylic positions with selenium dioxide/tert-butylhydroperoxide has been reported (27,28). 9,18- and 10,18-Dihydroxy-18:0 were prepared by condensation of two α,ω-difunctional synthons derived from suitable α,ω-alkanediols (29). The synthesis of 7,10-dihydroxy-18:1(8E) was achieved by Wittig olefination reaction of appropriate synthons (30). A chemo-enzymatic enantiospecific synthesis of (S)-coriolic acid, (13S)-hydroxy-18:2(9Z,11E), mediated via immobilized alcohol dehydrogenase of baker’s yeast has been described (31). 15,16-Didehydrocoriolic acid, 13-hydroxy18:3(9Z,11E,15Z), was stereoselectively synthesized starting from pent-2-en-4-yn-1ol (32). Four stereoisomers of 9,10,13-trihydroxy-18:1(11E) were derived from methyl 9,10-epoxy-12-octadecenoate. The latter was obtained by partial epoxidation of methyl linoleate. These trihydroxy C18 fatty acids are potential antirice blast fungal substances (33). (11S)-Hydroxy-(12S,13S)-epoxy-18:2(9Z,15Z) was synthesized from D-mannose (34). The relative and absolute configuration of 3S,12S-dihydroxypalmitic acid, a constituent of the Ipomea operculata M. resin, was confirmed by synthesis starting with dimethyl L-malate (35). An efficient synthesis of (5S)-hydroxy-20:4(6E,8Z,11Z,14Z) was accomplished by the coupling of two readily accessible synthons, methyl (5S)hydroxy-7-iodo-heptanoate and 4Z,7Z- tridecadien-1-yne (36). A highly stereoselective synthesis of β-dimorphecolic acid, (9S)-hydroxy-18:2(10E,12E), has been reported; the synthesis features a diastereoselective reduction of a keto intermediate [4] in which the tricarbonyliron lactone tether induces a 1,5-transfer of chirality followed by a stereoselective decarboxylation to create all of the stereochemical elements of β-dimorphecolic acid (37).

The total synthesis of 5(S)-hydroxy-12-oxo-20:3(6Z,8E,14Z) involving Wittig coupling reactions of tailored intermediates was described (38). 10(S)-Hydroxy20:4(5Z,8Z,11Z,14Z) was synthesized in eight steps starting from enantiomerically pure (R)-glyceraldehyde acetonide (39). Chiral adducts from Grignard or allylsilane

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

25

additions to 1,3-dioxan/1,3-dioxolan-4-ones were exploited for the total synthesis of the (R)- and (S)-isomers of 16-, 17- and 18-hydroxy-20:4(5Z,8Z, 11Z,14Z) (40). The chemoenzymatic synthesis of (13S)-hydroxy-18:2(9Z,11E) was achieved in nine steps starting from (2E)-octenal. Of importance was the enzymatic conversion of (2E)-octenal to the (S)-cyanohydrin [5] with (S)-hydroxynitrile lyase cloned from Hevea brasiliensis (41). 13- Hydroxy-10-oxo-18:1(11E) was synthesized via a Knoevenagel-type reaction of Isopropyl 11- phenylsulfinyl-10-oxoundecanoate with heptanal to form γ-hydroxyenone functionality together with carbon chain elongation (42). The regiospecific oxidation of a number of substituted unsaturated fatty esters with p-benzoquinone in the presence of palladium(II) chloride under concomitant ultrasonic irradiation was reported. For example, methyl 9-hydroxy-18:1 (12Z) furnished methyl 9-hydroxy-12-keto-18:0 exclusively (43).

A stereoselective synthesis of 5(S),6(R),15(S)-trihydroxy-20:4(7E,9E,11Z,13E) from D-xylose using zinc-mediated deoxygenation of the 4-hydroxy-2-butenoic acid moiety and base- induced double elimination of 4,5-epoxy allyl chloride as key steps was reported (44). The enantiomers (R)- and (S)-3-hydroxy-20:4(5Z,8Z,11Z,14Z) were synthesized from coupling of a chiral aldehyde intermediate with a Wittig salt, which were derived from 2-deoxy-D-ribose and arachidonic acid, respectively (45).

Synthesis of Oxo, Epoxy and Oxa Fatty Acids

Oxidation of methyl dimorphecolate, (+)-(9S)-hydroxy-18:2(10E,12E), using Oppenauer conditions furnished the corresponding 9-oxo derivative (46). Ostopanic acid [6] was synthesized by using a β-chloro vinyl ketone as a key intermediate obtained from diethyl pimeliate (47). Reaction of methyl santalbate, 18:2(9A,11E), with mchloroperoxybenzoic acid furnished a mixture of methyl trans-11,12-epoxy-9-oxo18:0 and methyl 10-oxo-18:1(11E) (48). Ultrasonically stimulated oxidation reactions of 2,5-disubstituted C18 furanoid fatty ester with magnesium monoperoxyphthalate or m-chloroperoxybenzoic acid gave methyl 9,12-dioxo-18:1(10Z) and 10,11-epoxy9,12-dioxo-18:0, respectively (49). 4,5-Epoxy-14:1(2E) was synthesized starting with two sequential additions of acetate units to 10:0 followed by epoxidation of the resulting 14:2(2E,4E) with oxone (potassium peroxomonosulphate) (50). Oxidation of olefinic fatty esters to the corresponding keto derivatives was accomplished when the substrates were treated with palladium(II) chloride and p-benzoquinone in aqueous tetrahydrofuran under ultrasonic irradiation. For example, methyl oleate furnished a mixture of methyl 9(10)-oxostearate, whereas methyl 12-hydroxy-18:1(9Z) furnished methyl 12-hydroxy-9-oxostearate exclusively (43). The synthesis of ketoand hydroxy-dienoic acids from linoleic acid in a four-step procedure has been de-

Copyright © 1999 by AOCS Press

26

M.S.F. Lie Ken Jie and S.W.H. Cheung

scribed (51,52). Oxidation reactions of acetylenic fatty esters with selenium dioxide/tert-butyl hydroperoxide gave mixtures of hydroxy- and oxo-acetylenic derivatives, but in the case of 18:2(9A,11E), the major product isolated was 8-hydroxy18:2(9A,11E) (70%) (53). Lipoxygenase-catalyzed asymmetric oxygenation of linoleic acid followed by cobalt porphyrin-catalyzed reduction-oxygenation as the key step furnished 10-oxo-11(E)-octadecen-13-olide and 13-hydroxy-10-oxo18:1(11Z) (54).

Epoxidation of oleic and linoleic acid was readily achieved by treatment with the acetonitrile complex of hypofluorous acid (55). Phase-transfer–catalyzed biphasic epoxidation of unsaturated triglycerides was accomplished with ethylmethyldioxirane in 2-butanone (56). The enantioselective formation of an α,β-epoxy alcohol by reaction of methyl 13(S)-hydroperoxy-18:2(9Z,11E) with titanium isopropoxide has been reported (57). An immobilized form of Candida antartica on acrylic resin (Novozyme 435) was used to catalyze the perhydrolysis and the interesterification of esters. Unsaturated alcohols were converted with an ester in the presence of hydrogen peroxide to esters of epoxidized alcohols (e.g., epoxystearylbutyrate) directly (58). Homoallyl ethers were obtained from olefinic fatty esters by the ethylaluminium-induced reactions with dimethyl acetals of formaldehyde, acetaldehyde, isobutyraldehyde, and pivaldehyde (59). Reaction of 18:2(9Z,12Z) with 50% BF3-methanol gave monomethoxy and dimethoxy derivatives (60). A bulky phosphite-modified rhodium catalyst was developed for the hydroformylation of methyl 18:1 (9Z) and 18:1(9E), which furnished mixtures of formylstearate and diformylstearate (61).

Synthesis of Cyclic Fatty Acids

The synthesis of a novel cyclopropyl analog of arachidonic acid [7] via a convergent synthesis that employed methyl (1R,2S)-2-formylcyclopropane-carboxylate in conjunction with the ylide from (3Z,6Z)-pentadeca-3,6-dienyl(triphenyl)-phosphonium iodide was reported (62). A new approach to cyclopropene fatty acids has been developed for the synthesis of methyl sterculate [8] and methyl 2-hydroxy-sterculate; this involves the 1,2-deiodination of 1,2-diiodocyclopropanes with butyllithium at low temperature (63). The synthesis of deuterated cyclopropene fatty esters structurally related to palmitic and myristic acids has been reported (64). Three olefinic fatty acids containing a cyclopropenyl system [9]–[11] were synthesized starting from 2-bromoalk-1-ene by dibromocyclopropanation to give the corresponding 1,1,2-tribromo-2-alkylcyclopropane. Treatment of the latter interme-

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

27

diate with butyl lithium gave the lithiopropene, which was further elaborated to the requisite olefinic cyclopropenyl fatty acids (65). The total synthesis of the methyl ester of chromomoric acids [12] was achieved via coupling of enone intermediates and retro Diels-Alder reactions as detailed in Scheme 4 (66,67). A 2,5-disubstituted C18 tetrahydrofuran fatty ester [13] was obtained from methyl ricinoleate by addition of bromine to the isomerized substrate, followed by hydrogenation over palladium on charcoal (68). Free radical and Lewis acid–induced reactions involving the double bonds of unsaturated fatty esters have been conducted by Metzger et al. (69–74); these have resulted in the production of a large number of functionalized, cyclic, and branched fatty ester derivatives (e.g., [14], [15]). The synthesis of methyl rac-2-dodecyl-cyclopentane carboxylate from methyl 2-iodo18:1(6Z) is presented in Scheme 5.

Copyright © 1999 by AOCS Press

28

M.S.F. Lie Ken Jie and S.W.H. Cheung

SCHEME 4. Synthesis of chromomoric acid B methyl ester [12]. Reagents: (i) lithium (8-t-butyldimethylsilyloxyoctyl), Cul, Et2O; (ii) 1-iodo-2-pentyne, hexamethylphosphoramide (HMPA); (iii) 5% LiOH, MeOH, tetrahydrofuran (THF); (iv) AcOH, THF, H2O; (v) pyridium dichromate, N,N-dimethylformamide (DMF), molecular sieves; (vi) CH2N2, Et2O; (vii) 270°C; (viii) Pd/CaCO3, H2, toluene.

SCHEME 5. Synthesis of methyl-rac-2-dodecyl-cyclopentane carboxylate (cis-, trans). Reagents: (i) 2,2′-azobisisobutyronitrile (AIBN), tributylin hydride (TBTH), benzene.

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

29

An efficient enantiomeric synthesis of (−)-methyl jasmonate [16] and (+)methyl epijasmonate [17] has been reported. The procedure makes use of a chiral cyclopentanoid building block that can easily be prepared from tartaric acid by phosphorus ylid chemistry (75). Jasmonic acid was also prepared by mixed Kolbe electrolysis (76). Methyl 3-methyljasmonate was synthesized from methyl jasmonate via methyl 3,7-dehydrojasmonate (77).

The Friedel-Crafts adducts of methyl oleate, olefinic C6 hydrocarbons and esters with benzene and toluene have been studied. The products are a mixture of monomer, dimer, and trimer esters together with adducts containing one, two, or three molecules of ester per mole of aromatic compound (78,79). The synthesis of polyunsaturated constituents of phenolic lipids, such as 2-hydroxy-6-(pentadeca-8Z,11Zdienyl)benzoic acid, has been described (80). The syntheses of lurlenic acid [18] and some analogs starting from D-xylose, 2,3-dimethyl-p-hydroquinone and geranylgeraniol or farnesol have been described (81). Another six analogs of lurlenic acid with a modified sugar part were synthesized (82).

Monocyclic fatty acids formed from oleic acid in heated sunflower oils have been identified as 9-(2′-butylcyclopentyl)nonanoate and 9-(2′propylcyclohexyl)nonanoate by mass spectral analysis (83). The identification of traces of cyclobutane fatty acid from fatty acids in food by γ-irradation has been reported (84). The synthesis of furan fatty esters containing a phenyl substituent at the 3- or 4position of the furan ring was reported; these involved methyl 9,10-epoxy-12oxostearate as a key intermediate (85). The mono-, di-, and triacylglycerols of a C18 furan fatty acid were prepared by chemical and enzymatic means (86). Furanoacetylene phytoalexins (wyerone [19] and dihydro-wyerone [20]) were prepared in multigram quantities from furfural (87). The synthesis route of compound [19] is presented in Scheme 6.

Copyright © 1999 by AOCS Press

30

M.S.F. Lie Ken Jie and S.W.H. Cheung

SCHEME 6. Synthesis of wyerone [19]. Reagents: (i) n-BuLi, tetrahydrofuran (THF); (ii) 2M H2SO4, Et2O; (iii) trimethyl phosphonoacetate, LiOH, THF.

The total synthesis of a naturally occurring furan fatty acid [21] has been described recently in which mercury (II)-catalyzed isomerization of 2-(1,2oxiranylcyclododecyl)-3-nonyn-2-ol was used as presented in Scheme 7 (88).

Nitrogen-Containing Fatty Acid Derivatives

12-Aminododecanoic acid was obtained by hydrogenation of vernolic acid, 12,13epoxy-18:1(9Z), to give 12,13-epoxystearic acid. The latter compound was oxidized by HIO4 to yield 12-oxododecanoic acid, which was treated with hydroxylamine to form the oxime. The oxime was then catalytically reduced to yield the requisite 12aminododecanoic acid. 11-Aminoundecanoic acid was prepared from the 12-oxodo-

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

31

SCHEME 7. Synthesis of furan fatty acid [21]. Reagents: (i) lithium acetylide, tetrahydrofuran (THF); (ii) H2O, HCOOH; (iii) heptynyllithium; (iv) aqueous t-butyl hydroperoxide, catalytic VO(acac)2; (v) HgCl2, H2SO4; (vi) pyridinium dichromate.

decanoic acid oxime via Beckmann rearrangement and Hofmann degradation, followed by hydrolysis (89). An isomer of methyl 1-pyrroline fatty esters [22] was derived from methyl ricinoleate by cyclization of methyl 12-azido-18:1(9Z) (90). The physical properties of a large number of N-substituted pyrrolinium and pyrrolidine derivatives obtained from methyl 1-pyrroline fatty ester [22] were studied (91). The synthesis of a trisubstituted C18 pyrrole fatty ester [23], containing a methyl group at the 3-position of the ring, was reported (92). Similar long-chain pyrrole fatty esters have been obtained by the reaction of lipoxygenase with methyl 9,10-epoxy-13-oxo18:1(11E) and 12,13-epoxy-9- oxo-18:1(10E) in the presence of butylamine and lysine (93).

An efficient method has been developed for the preparation of pyrazole fatty ester derivatives [24] by the reaction of 10,12-dioxostearate with hydrazine and substituted hydrazines in water under ultrasonic irradiation as presented in Scheme 8 (94).

Copyright © 1999 by AOCS Press

32

M.S.F. Lie Ken Jie and S.W.H. Cheung

SCHEME 8. Synthesis of pyrazole fatty ester derivatives [24]. Reagents: (i) Br2, Et2O; (ii) KOH, ethanol, ultrasound; (iii) BF3-MeOH; (iv) chromic acid, H2O, ultrasound; (v) R4NHN2, H2O, ultrasound.

Reactions induced by ultrasonic irradiation of methyl 9,12-dioxostearate with hydrazines in water in the presence of acidic alumina gave high yields of pyridazine fatty esters directly (95). The synthesis of novel piperidine- [25] and pyridine- [26] containing long-chain fatty ester derivatives has been reported starting from methyl iso-ricinoleate, 9-hydroxy-18:1(12Z) (96). Allylic amination of methyl oleate with bis(N-ρ-toluenesulfonyl)sulfodiimide gave a mixture of methyl 11-amino-(N-ρtoluenesulfonyl)-18:1(9E) and methyl 8-amino-(N-ρ-toluenesulfonyl)-18:1(9E) (97).

Reactions of fatty acids with diethylene triamine gave 1,3-diamides without solvent, but when carried out in a solvent (xylene), 1,2-diamide was produced. 1,3-Diamides gave imidazolines under vacuum, whereas 1,2-diamides furnished imidazolines by heating at 145°C (98).

Fatty Acids Containing Sulfur, Selenium, or Tellurium

A rapid and high-yielding two-step synthesis of fatty thioacids has been devised; it involves the reaction of an acid chloride with thioacetic acid followed by deacetylation with propylamine or butylamine (99). The synthesis of positional isomers of unsaturated and hydroxy alkylseleno fatty acids [27] (100), 2,5-disubstituted

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

33

selenophene [28], and tellurophene [29] fatty esters has been reported (101,102). The synthesis route to 2,5-disubstituted selenophene fatty esters is presented in Scheme 9.

The synthesis and the study of nuclear magnetic resonance spectroscopic properties have been reported for acetylenic tellura fatty esters in which the number of methylene groups between the acetylenic bond and the tellurium atom varied from 0 to 4 (103). Five positional isomers of thiaoleic acid, in which the number of methylene groups between the olefinic bond and the sulfur atom ranged from 1 to 5, have been prepared as models for studies on ∆12-desaturation (104). A novel analog of

SCHEME 9. Synthesis of 2,5-disubstituted selenophene fatty esters [28]. Reagents: (i) NH2OH•HCl, CuCl, EtNH2, MeOH, NaOH; (ii) BF3-MeOH; (iii) NaHSe, AgOAc, ethanol; (iv) CrO3, pyridine, CH2Cl2; (v) potassium t-butoxide, tetrahydrofuran (THF), MeCH2PPh3+Br−; (vi) Pd/C, H2

Copyright © 1999 by AOCS Press

34

M.S.F. Lie Ken Jie and S.W.H. Cheung

hepoxilin A3 [30] containing a thiirano group at the C-11/C-12 position was accomplished through allylic rearrangement of unnatural (11/R,12/R)-hepoxilin B3 under Mitsunobu conditions, followed by conversion of the epoxy group of the intermediate to the requisite thiirano ester analog (105).

Fluorinated Analogs of Fatty Acids

A series of fluorine-containing unsaturated fatty acids [31], which are potential fungicides, have been prepared by consecutive bromofluorination and dehydrobromination of ω-alkenoic acids (106). Treatment of 18:1(6Z) with perfluorooctyl iodide in the presence of lead powder and a catalytic amount of Cu(II) acetate gave a mixture of positional isomers of iodo-perfluorooctyl-18:0 derivatives (107). The fluoro analog of coriolic acid, 13-fluoro-18:2(9Z,11E), was prepared by Wittig coupling reaction of methyl 9-triphenylphosphono-bromononanoate and 4-fluoro-(2E)-nonenal (108). The synthesis of 5,6-difluoroarachidonic acid starting from α-phenylthio-β-ketoester [32] has been reported (109).

Isotope-Labeled Fatty Acids

Labeled 16,16,16-trideuterio-16:1(11A) and 15,15,16,16,16-pentadeuterio-16:2 (11Z,13Z) were synthesized; the deuterium atoms were introduced by reaction of iodoalkynes with (CD3)2CuLi followed by alkylation of a terminal diyne with

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

35

CD3 CD2I, respectively (110). Multigram quantities of methyl 12,13-dideuteriolinoleate were obtained by deuteration of methyl 18:2(9Z,12A) (isolated from Crepis alpina) in the presence of Lindlar catalyst (111). The synthesis of 14,14,15,15,17,18-hexadeuterio-linoleic acid by successive Wittig reactions starting from 2,2,3,3,5,6-hexadeuterohexanal has been described (112). Using similar strategies, Rakoff (113,114) has synthesized a number of polyunsaturated fatty acids that include 3,3,4,4,8,8,9,9-octadeuterio-20:3(11,14,17) and 8,8,9,9tetradeuterio-20:4(5,11,14,17). Luthria and Sprecher (115) have used acetylenic coupling reactions to obtain 19,19,20,20-tetradeuterio-20:4(5,8,11,14) and 19,19,20,20-tetradeuterio-20:3(8,11,14). The syntheses of (9Z,12E)- and (9E,12Z)-[1-14C]-linoleic acid, (5Z,8Z,11Z,14E)14 [1- C]- arachidonic acid (116), and (9Z,12Z,15E)- and (9E,12Z,15Z)-[1-14C]octadecatrienoic acid have also been described (117). A recent review on methods of synthesis of deuterium-labeled hydrophobic and hydrophilic synthons of lipids molecules and their analogs has been published (118).

Fullerenoid Fatty Acid Derivatives

The preparation of a number of interesting adducts of [60]-fullerene-containing ester functions have been reported. Reaction of [60]-fullerene with ethyl bromoacetate and zinc followed by quenching with trifluoroacetic acid gave 1-ethoxycarbonylmethyl-1,2-dihydro-[60]fullerene (119). Direct treatment of [60]-fullerene with ethyl or octadecyl malonate in the presence of CBr4 and diazabicyclo[5.4.0]undec-7-ene (DBU) furnished the corresponding methanofullerenes [33a] (120). Lie Ken Jie and Cheung have recently described the synthesis of similar methanofullerenes containing acetylenic, olefinic, and polyunsaturated centers in the acyl chains (121).The following fullerene adducts have been synthesized: fullerene ester–containing cholesterol [33b] (122), D2h-symmetrical tetrakis(methano)fullerenes [34] (123), tethered trisadduct [35] (124), trans-enediyne [36] (125), or triester function [37] (126).

Copyright © 1999 by AOCS Press

36

References

M.S.F. Lie Ken Jie and S.W.H. Cheung

1. Lie Ken Jie, M.S.F., The Synthesis of Rare and Unusual Fatty Acids, Prog. Lipid Res. 32: 151–194 (1993). 2. Lie Ken Jie, M.S.F., M.K. Pasha, and M.S.K. Syed-Rahmatullah, Fatty Acids, Fatty Acid Analogues and Their Derivatives, Nat. Prod. Rep. 14: 163–189 (1997). 3. Lie Ken Jie, M.S.F., and M.K. Pasha, Fatty Acids, Fatty Acid Analogues and Their Derivatives, Nat. Prod. Rep. 15: 607–630 (1998). 4. Gronowitz, S., T. Klingstedt, L. Svensson, and U. Hansson, On the Syntheses of Branched Saturated Fatty Acids, Lipids 28: 889–897 (1993). 5. Wallace, P.A., and D.E. Minnikin, Synthesis of Racemic 2,4-Dimethyltetradecanoic Acid from Mycobacterium kansasii, Chem. Phys. Lipids 82: 141–146 (1996). 6. Reyes, E.D., and N.M. Carballeira, A Short Synthesis of (Z)-15-Methylhexadec-11-enoic Acid, Synthesis, 693–694 (1996). 7. Sharma, A., and S. Chattopadhyay, A Convenient Synthesis of (4R,7Z,11Z)-(–)-4Methylheptadeca-7,11-dienal and (4R,7Z,11Z)-4-Methylheptadeca-7,11-dienoic Acid, Liebigs Ann., 529–531 (1996). 8. Wallace, P.A., D.E. Minnikin, K. McCrudden, and A. Pizzarello, Synthesis of (R,S)-10Methyloctadecanoic Acid (Tuberculostearic Acid) and Key Chiral 2-Methyl Branched Intermediates, Chem. Phys. Lipids 71: 145–162 (1994). 9. Besra, G.S., D.E. Minnikin, P.R. Wheeler, and C. Ratledge, Synthesis of Methyl (Z)Tetracos-5-enoate and Both Enantiomers of Ethyl (E)-6-Methyltetracos-4-enoate: Possible Intermediates in the Biosynthesis of Mycolic Acids in Mycobacteria, Chem. Phys. Lipids 66: 23–34 (1993). 10. Wallace, P.A., and D.E. Minnikin, Synthesis of (2R; 4R, S)- and (2S; 4R, S)-Dimethyldocosanoic Acids, Chem. Phys. Lipids 72: 87–101 (1994).

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

37

11. Wallace, P.A., and D.E. Minnikin, Synthesis of a Derivative of Racemic 3-Hydroxy2,4,6-trimethyltetracosanoic (Mycolipanolic) Acid and Racemic 2,4,6-Trimethyltetraco2-enoic (Mycolipenic) Acid, Chem. Phys. Lipids 83: 1–8 (1996). 12. Rao, B.V.S.K., P. Vijayalakshmi, and R. Subbarao, Synthesis of Long Chain (E)-3Alkenoic Acids by the Knoevenagel Condensation of Aliphatic Aldehydes with Malonic Acid, J. Am. Oil Chem. Soc. 70: 297–299 (1993). 13. Sandri, J., and J. Viala, Direct Preparation of (Z,Z) 1,4-Dienic Units with a New C6 Homologating Agent: Synthesis of α-Linolenic Acid, Synthesis, 271–275 (1995). 14. Sandri, J., and J. Viala, Syntheses of all-(Z)-5,8,11,14,17-Eicosapentaenoic Acid and all-(Z)-4,7,10,13,16,19-Docosahexaenoic Acid from (Z)-1,1,6,6-Tetraisopropoxy-2hexene, J. Org. Chem. 60: 6627–6630 (1995). 15. Taber, D.F., and K. You, New Synthon for the Convergent Construction of Skipped Conjugation Polyenes: Synthesis of Ethyl Docosa-4,7,10,13,16,19-hexaenoate, J. Org. Chem. 60: 139–142 (1995). 16. Muehldorf, A.V., Preparation of 11-Substituted Linoleic Acids, Tetrahedron Lett. 35: 6851–6852 (1994). 17. Vatele, J.M., H.D. Doan, B. Fenet, J.M. Chardigny, and J.L. Sebedio, Synthesis of Methyl (5Z, 8Z, 11E, 14Z, 7Z)- and (5Z, 8Z, 11E, 14Z, 17E)-eicosapentaenoate (EPA ∆11t and ∆11t, 17t), Chem. Phys. Lipids 78: 65–70 (1995). 18. Vatele, J.M., H.D. Doan, J.M. Chardigny, J.L. Sebedio, and A. Grandgirard, Synthesis of Methyl (5Z, 8Z, 11Z, 14Z, 17E )-eicosapentaenoate and Methyl (4Z, 7Z, 10Z, 13Z, 16Z, 19E)-docosahexaenoate, Chem. Phys. Lipids 74: 185–193 (1994). 19. Kling, M.R., C.J. Easton, and A. Poulos, Synthesis of Very Long Chain Fatty Acid Methyl Esters, J. Chem. Soc. Perkin Trans. I, 1183–1189 (1993). 20. Poulain, S., N. Noiret, and H. Patin, An Easy Access to Symmetrical Z-Olefins from Phosphorus Ylides, Tetrahedron Lett. 37: 7703–7706 (1996). 21. Sandri, J., T. Soto, J.L. Gras, and J. Viala, A Convergent Strategy Toward Linear cisSkipped Polyenes, Tetrahedron Lett. 38: 6611–6612 (1997). 22. Eynard, T., J.M. Vatele, O. Berdeaux, D. Poullain, and J.P. Noel, Synthesis of (9Z,12E), (9E, 12Z)-[1-14C]-Linoleic Acid, (9Z,12Z,15E)- (9E,12Z,15Z)-[1-14C]-Linolenic Acid and (5Z,8Z,11Z,14E)-[1-14C]-Arachidonic Acid, Grasas Aceites 47: 81–85 (1996). 23. Solladie, G., C. Kalai, M. Adamy, and F. Colobert, Stereoselective Synthesis of Conjugated all-trans-Tetraenes. Application to the Synthesis of β-Parinaric Acid Methyl Ester, Tetrahedron Lett. 38: 6917–6920 (1997). 24. Lie Ken Jie, M.S.F., M.K. Pasha, and F. Ahmad, Ultrasound-Assisted Synthesis of Santalbic Acid and a Study of Triacylglycerol Species in Santalum album (Linn.) Seed Oil,, Lipids 31: 1083–1089 (1996). 25. Charoenying, P., D.H. Davies, D. McKerrecher, and R.J.K. Taylor, A New Stereocontrolled, Pyrylium-Based Route to Conjugated Dienynes: The First Synthesis of Carduusyne A, Tetrahedron Lett. 37: 1913–1916 (1996). 26. Matikainen, J., S. Kaltia, M. Hamalainen, and T. Hase, 1,5-Hydrogen Shift and Other Isomerization Reactions of Certain Ethyl Octadecatrienoates, Tetrahedron 53: 4531– 4538 (1997). 27. Knothe, G., D. Weisleder, M.O. Bagby, and R.E. Peterson, Hydroxy Fatty Acids Through Hydroxylation of Oleic Acid with Selenium Dioxide/tert-Butylhydroperoxide, J. Am. Oil Chem. Soc. 70: 401–404 (1993). 28. Knothe, G., M.O. Bagby, D. Weisleder, and R.E. Peterson, Allylic Hydroxy Fatty Compounds with ∆5-, ∆7-, ∆8-, and ∆10-Unsaturation, J. Am. Oil Chem. Soc. 72: 703–706

Copyright © 1999 by AOCS Press

38

M.S.F. Lie Ken Jie and S.W.H. Cheung

(1995). 29. Sonnet, P.E., S.F. Osman, H.C. Gerard, and R.L. Dudley, Synthesis and Configuration Analysis of 9,18- and 10,18-Dihydroxystearic Acid Methyl Esters, Chem. Phys. Lipids 69: 121–128 (1994). 30. Mhaskar, S.Y., and R. Subbarao, Synthesis of 7,10-Dihydroxy-8(E)-Octadecenoic Acid, J. Am. Oil Chem. Soc. 70: 519–522 (1993). 31. Bhalerao, U.T., L. Dasaradhi, C. Muralikrishna, and N.W. Fadnavis, A Novel Chemoenzymatic Enantiospecific Synthesis of (S)-Coriolic Acid Mediated Via Immobilized Alcohol Dehydrogenase of Baker’s Yeast, Tetrahedron Lett. 34: 2359–2360 (1993). 32. Reddy, K.L.N., S.P. Reddy, and G.V.M. Sharma, Synthesis of 13-Hydroxy-9Z,11E,15Zoctadecatrienoic Acid, Synth. Commun. 24: 1395–1401 (1994). 33. Kato, T., T. Hirukawa, and M. Yano, Synthesis of (E)-9,10,13-Trihydroxy-11octadecenoic Acids, Bull. Chem. Soc. Jpn. 67: 839–842 (1994). 34. Wu, W.L., and Y.L. Wu, Stereoselective Synthesis of Methyl (11S,12S,13S)-(9Z,15Z)11-Hydroxy-12,13-Epoxy Octadecadienoate from D-Mannose, Tetrahedron 49: 4665– 4670 (1993). 35. Jakob, B., and H. Gerlach, Relative and Absolute Configuration of the 3,12- Dihydroxypalmitic Acids, Liebigs Ann. 2123–2129 (1996). 36. Gueugnot, S., M. Alami, G. Linstrumelle, L. Mambu, Y. Petit, and M. Larcheveque, An Efficient Total Synthesis of 5-(S)-HETE, Tetrahedron 52: 6635–6646 (1996). 37. Ley, S.V., and G. Meek, Synthesis of β-Dimorphecolic Acid Exploiting Highly Stereoselective Reduction of a Side-Chain Carbonyl Group in a π-Allyltricarbonyl Iron Lactone Complex, J. Chem. Soc. Perkin Trans. I, 1125–1130 (1997). 38. Khanapure, S.P., S.S. Wang, W.S. Powell, and J. Rokach, Synthesis of 10,11-Dihydro12-oxo-LTB4, a Key Biochemical Intermediate, J. Org. Chem. 62: 325–330 (1997). 39. Yeola, S.N., S.A. Saleh, A.R. Brash, C. Prakash, D.F. Taber, and LA. Blair, Synthesis of 10(S)-Hydroxyeicosatetraenoic Acid: A Novel Cytochrome P-450 Metabolite of Arachidonic Acid, J. Org. Chem. 61: 838–841 (1996). 40. Heckmann, B., C. Mioskowski, S. Lumin, J.R. Falck, S. Wei, and J.H. Capdevila, Chiral Acetals: Stereocontrolled Syntheses of 16-,17-, and 18-Hydroxyeicosatetraenoic Acids, Cytochrome P-450 Arachidonate Metabolites, Tetrahedron Lett. 37: 1425–1428 (1996). 41. Johnson, D.V., and H. Griengl, The Chemoenzymatic Synthesis of (S)-13Hydroxyoctadeca-(9Z, 11E)-dienoic Acid Using the Hydroxynitrile Lyase from Hevea brasiliensis, Tetrahedron 53: 617–624 (1997). 42. Nokami, J., M. Osafune, K. Shiraishi, S.I. Sumida, and N. Imai, Convenient Synthesis of Cytotoxic (11E)-13-Hydroxy-10-oxooctadec-11-enoic Acid from Undec-10-enoic Acid, J. Chem. Soc. Perkin Trans. I, 2947–2950 (1997). 43. Lie Ken Jie, M.S.F., and C.K. Lam, Regiospecific Oxidation of Unsaturated Fatty Esters with Palladium(II) Chloride/p-Benzoquinone: A Sonochemical Approach, Chem. Phys. Lipids 81: 55–61 (1996). 44. Yadav, J.S., D.K. Barma, and D. Duta, Stereoselective Total Synthesis of (5S), 6(R), 15(S)-Trihydroxy-7(E),9(E),11(Z),13(E)-Eicosatetraenoic Acid (Lipoxin A), Tetrahedron Lett. 39: 143–146 (1998). 45. Bhatt, R.K., J.R. Falck, and S. Nigam, Enantiospecific Total Synthesis of a Novel Arachidonic Acid Metabolite 3-Hydroxyeicosatetraenoic Acid, Tetrahedron Lett. 39: 249–252 (1998). 46. Tassignon, P., T. de Rijk, D. de Wit, and L. de Buyck, Synthesis of the 9-Oxo-diene Derivative of Methyl Dimorphecolate by an Efficient Oppenauer Oxidation Procedure,

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

39

Chem. Phys. Lipids 74: 39–42 (1994). 47. Bhalerao, U.T., S. Devalla, L. Dasaradhi, and B.V. Rao, A New Stereoselective Synthesis of Ostopanic Acid Via β-Chloro Vinyl Ketone, Synth. Commun. 23: 2213–2217 (1993). 48. Pasha, M.K., and F. Ahmad, Synthesis of Oxygenated Fatty Acid Esters from Santalbic Acid Ester, Lipids 28: 1027–1031 (1993). 49. Lie Ken Jie, M.S.F., M.K. Pasha, and C.K. Lam, Ultrasonically Stimulated Oxidation Reactions of 2,5-Disubstituted C18 Furanoid Fatty Ester, Chem. Phys. Lipids 85: 101– 106 (1997). 50. Cow, C., D. Valentini, and P. Harrison, Synthesis of the Fatty Acid of Pramanicin, Can. J. Chem. 75: 884–889 (1997). 51. Kuklev, D.V., W.W. Christie, T. Durand, J.C. Rossi, J.P. Vidal, S.P. Kasyanov, V.N. Akulin, and V.V. Bezuglov, Synthesis of Keto- and Hydroxydienoic Compounds from Linoleic Acid, Chem. Phys. Lipids 85: 125–134 (1997). 52. Kuklev, D.V., W.W. Christie, T. Durand, J.C. Rossi, J.P. Girard, S.P. Kasyanov, V.N. Akulin, and V.V. Bezuglov, Synthesis of Ketodienoic Compounds from Natural Unsaturated Fatty Acids. 1. Linoleic Acid, Russ. J. Bioorg. Chem. 22: 622–627 (1996). 53. Lie Ken Jie, M.S.F., M.K. Pasha, and M.S. Alam, Oxidation Reactions of Acetylenic Fatty Esters with Selenium Dioxide/tert-Butyl Hydroperoxide, Lipids 32: 1119–1123 (1997). 54. Matsushita, Y.I., K. Sugamoto, T. Nakama, T. Matsui, Y. Hayashi, and K. Uenakai, Enantioselective Syntheses of 10-Oxo-11(E)-octadecen-13-olide and Related Fatty Acid, Tetrahedron Lett. 38: 6055–6058 (1997). 55. Rozen, S., Y. Bareket, and S. Dayan, Direct Epoxidation of Unprotected Olefinic Carboxylic Acids Using HOF-CH3CN, Tetrahedron Lett. 37: 531–534 (1996). 56. Sonnet, P.E., and T.A. Foglia, Epoxidation of Natural Triglycerides with Ethylmethyldioxirane, J. Am. Oil Chem. Soc. 73: 461–464 (1996). 57. Piazza, G.J., T.A. Foglia, and A. Nunez, Enantioselective Formation of an α,β-Epoxy Alcohol by Reaction of Methyl 13(S)-Hydroperoxy-9(Z),11(E)-octadecadienoate with Titanium Isopropoxide, J. Am. Oil Chem. Soc. 74: 1385–1390 (1997). 58. Klaas, M.R., and S. Warwel, A Three-Step-One-Pot Chemo-Enzymatic Synthesis of Epoxyalkanolacylates, Synth. Commun. 28: 251–260 (1998). 59. Metzger, J.O., and U. Biermann, Ethylaluminium Dichloride Induced Reactions of Acetals with Unsaturated Carboxylic Esters: Synthesis of Homoallyl Ethers, Liebigs Ann. 1851–1854 (1996). 60. Carballeira, N.M., M.V. Gonzalez, and M. Pagan, Neighboring Methoxyl Participation in the Acid Catalyzed Methoxylation of Methylene-Interrupted Fatty Acids, Chem. Phys. Lipids 89: 91–96 (1997). 61. Muilwijk, K.F., P.C.J. Kamer, and P.W.N.M. van Leeuwen, A Bulky Phosphite-Modified Rhodium Catalyst for the Hydroformylation of Unsaturated Fatty Acid Esters, J. Am. Oil Chem. Soc. 74: 223–228 (1997). 62. Butler, P.I., T. Clarke, C. Dell, and J. Mann, Synthesis of Methyl (1R,2S)-2[(1′Z,4′Z,7′Z)-Hexadeca-1′,4′,7′-trienyl]cyclopropanecarboxylate: A Potential Inhibitor of the Enzyme 5-Lipoxygenase, J. Chem. Soc. Perkin Trans. I, 1503–1508 (1994). 63. Baird, M.S., and B. Grehan, A New Approach to Cyclopropene Fatty Acids Involving 1.2-Deiodination, J. Chem. Soc. Perkin Trans. I, 1547–1548 (1993). 64. Gosalbo, L., M. Barrot, G. Fabrias, G. Arsequell, and F. Camps, Synthesis of Deuterated Cyclopropene Fatty Esters Structurally Related to Palmitic and Myristic Acids, Lipids 28: 1125–1130 (1993).

Copyright © 1999 by AOCS Press

40

M.S.F. Lie Ken Jie and S.W.H. Cheung

65. Dulayymi, J.R.A, M.S. Baird, C.M. Dale, B.J. Grehan, and M.F. Shortt, Synthesis of Putative ∆6-,∆12-, and ∆15-Desaturase Inhibitors, Tetrahedron 53: 1099–1110 (1997). 66. Chu, X.J., H. Dong, and Z.Y. Liu, Total Synthesis of Chromomoric Acid B and F Methyl Esters, Tetrahedron 51: 173–180 (1995). 67. Liu, Z.Y., H. Dong, and X.J. Chu, Total Synthesis and Structural Revision of Chromomoric Acid C-I and C-II Methyl Esters, Tetrahedron 50: 12337–12348 (1994). 68. Mhaskar S.Y., and V.V.S. Mani, Synthesis of Methyl 9,12-Epoxyoctadecanoate from Castor Oil, J. Am. Oil Chem. Soc: 71: 543–544, (1994). 69. Metzger, J.O., and R. Mahler, Free-Radical Cyclization of Petroselinic Acid, Liebigs Ann. Chem. 203–205 (1993). 70. Metzger, J.O., and U. Biermann, Produkte der Thermischen En-Reaktion von Ungesattigten Fettstoffen and Maleinsaureanhydrid, Fat Sci. Technol. 96: 321–323 (1994). 71. Metzger, J.O., and F. Bangert, Ane Additions to Unsaturated Fatty Compounds: Thermally Initiated Additions of Alkanes to Methyl 10-Undecenoate, Fat Sci. Technol. 97: 7–9 (1995). 72. Metzger, J.O., and R. Mahler, Radical Additions of Activated Haloalkanes to Alkenes Initiated by Electron Transfer from Copper in Solvent-Free Systems, Angew. Chem. Int. Ed. Engl. 34: 902–904 (1995). 73. Metzger, J.O., and U. Biermann, Alkylaluminium Dichloride Induced Friedel-Crafts Acylation of Unsaturated Carboxylic Acids and Alcohols, Liebigs Ann. Chem. 645–650 (1993). 74. Biermann, U., and J.O. Metzger, Lewis Acid Induced Additions to Unsaturated Fatty Compounds, Fat Sci. Technol. 95: 326–328 (1993). 75. Roth, G.J., S. Kirschbaum, and H.J. Bestmann, Enantioselective Synthesis of (−)-Methyl Jasmonate and (+)-Methyl Epijasmonate, Synlett, 618–620 (1997). 76. Schierle, K., J. Hopke, M.L. Niedt, W. Boland, and E. Steckhan, Homologues of Dihydro-12-oxo-phytodienoic Acid and Jasmonic Acid by Mixed Kolbe Electrolysis, Tetrahedron Lett. 37: 8715–8718 (1996). 77. Ward, J.L., P. Gaskin, M.H. Beale, R. Sessions, Y. Koda, and C. Wasternack, Molecular Modelling, Synthesis and Biological Activity of Methyl 3-Methyljasmonate and Related Derivatives, Tetrahedron 53: 8181–8194 (1997). 78. Black, K.D., and F.D. Gunstone, The Friedel-Crafts Adducts of Methyl Oleate with Benzene and Toluene, Chem. Phys. Lipids 79: 87–94 (1996). 79. Black, K.D., and F.D. Gunstone, Friedel-Crafts Alkylation of Benzene and Toluene with Olefinic C6 Hydrocarbons and Esters, Chem. Phys. Lipids 79: 79–86 (1996). 80. Tyman, J.H.P., and N. Visani, Synthesis of Polyunsaturated Constituents of Phenolic Lipids, Chem. Phys. Lipids 85: 157–174 (1997). 81. Takanashi, S.I., and K. Mori, Synthesis of Lurlenic Acid and Lurlenol, the Sex Pheromones of the Green Flagellate Chlamydomonas, Liebigs Ann./Recueil, 825–838 (1997). 82. Takanashi, S.I., and K. Mori, Synthesis of the Analogues of Lurlenic Acid with a Modified Sugar Part: Chlamydomonas Responds Only to the D-Xyloside, Liebigs Ann./Recueil, 1081–1084 (1997). 83. Dobson, G., W.W. Christie, and J.L. Sebedio, Monocyclic Saturated Fatty Acids Formed from Oleic Acid in Heated Sunflower Oils, Chem. Phys. Lipids 82: 101–110 (1996). 84. Hamilton, L., M.H. Stevenson, D.R. Boyd, I.N. Brannigan, A.B. Treacy, J.T.G. Hamilton, W.C. McRoberts, and C.T. Elliott, Detection of 2-Substituted Cyclobutanones as Irradiation Products of Lipid-Containing Foods: Synthesis and Applications of cis- and

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

41

trans-2-(tetradec-5′- enyl)cyclobutanones and 11-(2′-oxocyclobutyl)-undecanoic Acid, J. Chem. Soc. Perkin Trans. I, 139–146 (1996). 85. Lie Ken Jie, M.S.F., and K.P. Wong, Synthesis of Phenyl Substituted C18 Furanoid Fatty Esters, Lipids 28: 43–46 (1993). 86. Lie Ken Jie, M.S.F., and M.S.K. Syed-Rahmatullah, Chemical and Enzymatic Preparation of Acylglycerols Containing C18 Furanoid Fatty Acids, Lipids 30: 79–84 (1995). 87. Delamarche, I., and P. Mosset, New Efficient Synthesis of Furanoacetylene Phytoalexins Wyerone and Dihydrowyerone, Tetrahedron Lett. 34: 2465–2468 (1993). 88. Marson, C.M., and S. Harper, Total Synthesis of the Naturally Occurring Furanoid Fatty Acid, F5, Tetrahedron Lett. 39: 333–334 (1998). 89. Ayorinde, F.O., E.Y. Nana, P.D. Nicely, A.S. Woods, E.O. Price, and C.P. Nwaonicha, Syntheses of 12-Aminododecanoic and 11-Aminoundecanoic Acids from Vernolic Acid, J. Am. Oil Chem. Soc. 74: 531–538 (1997). 90. Lie Ken Jie, M.S.F., M.S.K. Syed-Rahmatullah, C.K. Lam, and P. Kalluri, Ultrasound in Fatty Acid Chemistry: Synthesis of a 1-Pyrroline Fatty Acid Ester Isomer from Methyl Ricinoleate, Lipids 29: 889–892 (1994). 91. Lie Ken Jie, M.S.F., and M.S.K. Syed-Rahmatullah, Further Chemical Reactions of 1Pyrroline Fatty Ester: N-Substituted Pyrrolinium and Pyrrolidine Derivatives, Chem. Phys. Lipids 77: 179–186 (1995). 92. Lie Ken Jie, M.S.F., and K.P. Wong, Synthesis of Trisubstituted C18 Pyrrole Fatty Ester Derivatives, Lipids 28: 161–162 (1993). 93. Hidalgo, F.J., and R. Zamora, In Vitro Production of Long Chain Pyrrole Fatty Esters from Carbonyl-Amine Reactions, J. Lipid Res. 36: 725–735 (1995). 94. Lie Ken Jie, M.S.F., and P. Kalluri, Synthesis of Pyrazole Fatty Ester Derivatives in Water: A Sonochemical Approach, J. Chem. Soc. Perkin Trans. I, 1205–1206 (1995). 95. Lie Ken Jie, M.S.F., and P. Kalluri, Synthesis of Pyridazine Fatty Ester Derivatives in Water: A Sonochemical Approach, J. Chem. Soc. Perkin Trans. I, 3485–3486 (1997). 96. Lie Ken Jie, M.S.F., and M.K. Pasha, Synthesis of Novel Piperideine- and PyridineContaining Long-Chain Fatty Ester Derivatives from Methyl Iso-Ricinoleate, J. Chem. Soc. Perkin Trans. I, 1331–1332 (1996). 97. Herron, B.F., M.O. Bagby, T.A. Isbell, Wm. C. Byrdwell, R. Plattner, and D. Weisleder, Preparation and Characterization of Methyl 11-Amino-(N-p-toluenesulfonyl)-9-Eoctadecenoate and Methyl 8-Amino-(N-p-toluenesulfonyl)-9-E-octadecenoate, J. Am. Oil Chem. Soc. 74: 229–234 (1997). 98. Wu, Y., and P.R. Herrington, Thermal Reactions of Fatty Acids with Diethylene Triamine, J. Am. Oil Chem. Soc. 74: 61–64 (1997). 99. Shin, H.C., and D.M. Quinn, A Simple and Efficient Synthesis of Fatty Thioacids, Lipids 28: 73–74 (1993). 100. Lie Ken Jie, M.S.F., and Y.K. Cheung, Synthesis and Physical Properties of Some Unsaturated and Hydroxy Alkyseleno Fatty Acid Derivatives, Chem. Phys. Lipids 75: 71– 80 (1995). 101. Lie Ken Jie, M.S.F., and Y.K. Cheung, Synthesis of and Nuclear Magnetic Resonance Studies on a Series of Synthetic Long-Chain Selenophene Fatty Esters, J. Chem. Res. (S), 392 (1993); (M) 2745–2778 (1993). 102. Lie Ken Jie, M.S.F., and S.H. Chau, Synthesis of and Nuclear Magnetic Resonance Studies on a Series of Synthetic Long-Chain Tellurophene Fatty Esters, J. Chem. Res., (S) 428 (1995); (M) 2642–2657 (1995). 103. Lie Ken Jie, M.S.F., and S.H. Chau, Synthesis and Nuclear Magnetic Resonance Spec-

Copyright © 1999 by AOCS Press

42

M.S.F. Lie Ken Jie and S.W.H. Cheung

troscopic Properties of Some Acetylenic Tellura Fatty Acid Esters, Chem. Phys. Lipids 87: 55–63 (1997). 104. Poulain, S., N. Noiret, C. Nugier-Chauvin, and H. Patin, Preparation and Use of Thiaoleic Acids, Liebigs Ann./Recueil, 35–40 (1997). 105. Demin, P.M., D. Reynaud, and C.R. Pace-Asciak, Chemical Synthesis and Actions of 11,12-Thiiranohepoxilin A3, J. Lipid Mediat. Cell Signal, 13: 63–72 (1996). 106. Michel, D., and M. Schlosser, (ω-1)-Fluoroalk-(ω-1)-enoic Acids: Potential Fungicides, Synthesis, 1007–1011 (1996). 107. Metzger, J.O., R. Mahler, and A. Schmidt, Electron Transfer Initiated Free Radical Additions of Perfluoroalkyl Iodides and Diiodides to Alkenes, Liebigs Ann., 693–696 (1996). 108. Gree, D., R. Gree, A. Boukerb, and M. Laabassi, Synthesis of Fluorinated Analogs of Polyunsaturated Fatty Acid Metabolites (4 HNE, 13 HODE), Tetrahedron Lett. 38: 6209–6212 (1997). 109. Bildstein, S., J.B. Ducep, D. Jacobi, and P. Zimmermann, Synthesis of 5,6-Difluoroarachidonic Acid, a Potential Inhibitor of 5-Lipoxygenase, Tetrahedron Lett. 37: 4941–4944 (1996). 110. Barrot, M., G. Fabrias, and F. Camps, Synthesis of [16,16,16-²H3] 11-Hexadecynoic Acid and [15,15,16,16,16-²H5] (Z,Z)-11,13-Hexadecadienoic Acid and Their Use as Tracers in a Key Step of the Sex Pheromone Biosynthesis of the Processionary Moth, Tetrahedron 50: 9789–9796 (1994). 111. Adlof, R.O., and E.A. Emken, Large-Scale Preparation of Linoleic Acid-d2-Enriched Triglycerides from Crepis alpina Seed Oil, J. Am. Oil Chem. Soc. 70: 817–819 (1993). 112. Viala, J., and R. Labaudiniere, Synthesis of a Regioselectively Hexadeuterated Linoleic Acid, J. Org. Chem. 58: 1280–1283 (1993). 113. Rakoff. H., Synthesis of Methyl 11,14,17-Eicosatrienoate-3,3,4,4,8,8,9,9-d8, Lipids 28: 231–234 (1993). 114. 'Rakoff, H., Preparation of Methyl 5,11,14,17-Eicosatetraenoate-8,8,9,9-d4, Lipids 28: 47–50 (1993). 115. Luthria, D.L., and H. Sprecher, Synthesis of Ethyl Arachidonate-19,19,20,20-d4 and Ethyl Dihomo-γ-linolenate-19,19,20,20-d4, Lipids 28: 853–856 (1993). 116. Berdeaux, O., J.M. Vatele, T. Eynard, M. Nour, D. Poullain, J.P. Noel, and J.L. Sebedio, Synthesis of (9Z,12E)- and (9E,12Z)-[1-14C]Linoleic Acid and (5Z,8Z,11Z,14E)-[114 C]Arachidonic Acid, Chem. Phys. Lipids 78: 71–80 (1995). 117. Eynard, T., J.M. Vatele, D. Poullain, J.P. Noel, J.M. Chardigny, and J.L. Sebedio, Synthesis of (9Z,12Z,15E)- and (9E, 12Z, 15Z)-Octadecatrienoic Acids and Their [114 C]Radiolabelled Analogs, Chem. Phys. Lipids 74: 175–184 (1994). 118. Bragina, N.A., and V.V. Chupin, Methods of Synthesis of Deuterium-Labelled Lipids, Russ. Chem. Rev. 66: 975–986 (1997). 119. Wang, G.W., Y. Murata, K. Komatsu, and T.S.M. Wan, The Solid-Phase Reaction of [60] Fullerene: Novel Addition of Organozinc Reagents, J. Chem. Soc. Chem. Commun. 2059–2060 (1996). 120. Camps, X., and A. Hirsch, Efficient Cyclopropanation of C60 Starting from Malonates, J. Chem. Soc. Perkin Trans. I, 1595–1596 (1997). 121. Lie Ken Jie, M.S.F., and S.W.H. Cheung, Fullerene Lipids: Synthesis of Dialkyl 1,2[6,6]-Methano-[60]-Fullerene Dicarboxylate Derivatives, Lipids 33: 729–73 (1998) 122. Chuard, T., and R. Deschenaux, First Fullerene[60]-Containing Thermotropic Liquid Crystal, Helv. Chim. Acta 79: 736–741 (1996).

Copyright © 1999 by AOCS Press

Synthesis of Special Fatty Acids

43

123. Cardullo, F., P. Seiler, L. Isaacs, J.F. Nierengarten, R.F. Haldimann, F. Diederich, T. Mordasini-Denti, W. Thiel, C. Boudon, J.P. Gisselbrecht, and M. Gross, Bis- Through Tetrakis-Adducts of C60 by Reversible Tether-Directed Remote Functionalization and Systematic Investigation of the Changes in Fullerene Properties as a Function of Degree, Pattern, and Nature of Functionalization, Helv. Chim. Acta 80: 343–371 (1997). 124. Isaacs, L., F. Diederich, and R.F. Haldimann, Multiple Adducts of C60 by TetherDirected Remote Functionalization and Synthesis of Soluble Derivatives of New Carbon Allotropes Cn(60+5), Helv. Chim. Acta 80: 317–342 (1997). 125. Nierengarten, J.F., A. Herrmann, R.R. Tykwinski, M. Ruttimann, F. Diederich, C. Boudon, J.P. Gisselbrecht, and M. Gross, Methanofullerene Molecular Scaffolding: Towards C60- Substituted Poly(triacetylenes) and Expanded Radialenes, Preparation of a C60–C70 Hybrid Derivative, and a Novel Macrocyclization Reaction, Helv. Chim. Acta, 80: 293–316 (1997). 126. Murakami, H., Y. Watanabe, and N. Nakashima, Fullerene Lipid Chemistry: Self- Organized Multibilayer Films of a C60-Bearing Lipid with Main and Subphase Transitions, J. Am. Chem. Soc. 118: 4484–4485 (1996).

Copyright © 1999 by AOCS Press

Chapter 3

Derivatives of New Crop Oils Terry A. Isbell

USDA, Agricultural Research Service, National Center for Agricultural Utilization Research, New Crops Research, Peoria, IL 61604

Introduction

The development of new agricultural crops provides a future for farm diversity and a wealth of novel compounds for the chemist and industry. Each new crop has a unique package of natural compounds found not only in the seed oil but throughout the entire plant. The industrial chemist must seek out these new constituents and exploit the unusual chemistry that these compounds can offer. The plant kingdom is estimated to offer Ɑ300,000 species available to man (1) of which only a few hundred are in organized agriculture. In 1957, the USDA initiated a program to collect Ɑ8000 different plant species, many of which were analyzed for potential sources of starch, protein, oil, fiber, medicinal components, as well as any other unusual materials. As a result of this effort, Ɑ100 new oils were discovered. Of these new oilseed plants, three have progressed to the point of commercial production, crambe, jojoba, and meadowfoam. In addition, lesquerella is almost sufficiently developed to begin production. The unique chemical structures of these four seed oils and how they affect the chemistry of the oil will be the basis of this chapter.

Crambe and HEAR

Crambe (Crambe abyssinica) and high-erucic acid rapeseed (Brassica napus) are oilseeds that contain large quantities of erucic acid 22:1 (∆ 13) as the main fatty acid component of the triglyceride. Crambe and high-erucic acid rapeseed (HEAR) contain 59.5 and 42% erucic acid, respectively (2). HEAR has more oil in the seed (42% compared with 35% for crambe). Both oilseeds are in commercial production with acreage in the tens of thousands and are grown mainly in the northern plains of the U.S. and Canada as well as eastern Europe (3). The main derivative of erucic acid is erucamide (2,4). Erucamide is manufactured under standard amination procedures by reacting pure erucic acid (~95%) at elevated temperature and pressure with ammonia (Scheme 1). The N,N-disubstituted amides are used as slip and antiblocking agents in polyethylene, polypropylene, and other plastics (2,5). The length of the alkyl chain appears to be the key feature for its properties, allowing low concentrations (1000 ppm) in the resin, whereas shorter 44 Copyright © 1999 by AOCS Press

Derivatives of New Crop Oils

45

SCHEME 1.

chain lengths require orders of magnitude more amide to achieve properties similar to erucamide (3). Behenic acid, its esters, and other derivatives that are derived from catalytic hydrogenation of erucic acid are useful in a wide range of applications including cosmetics, fabric softeners, photothermography, mold release in rubber and plastics, and antifriction coatings in textiles (2). Brassylic acid, a 13-carbon saturated dibasic acid, can be derived from the oxidative cleavage of erucic acid and used as a feedstock for the production of nylon (Scheme 1). Brassylic acid is made by ozonolytic cleavage of erucic acid in acetic acid followed by oxidation of the resultant aldehyde by oxygen at elevated temperatures (100°C) to give the diacid. Crystallization from toluene gives a polymer-grade brassylic acid (6). Pilot-scale production of nylon-1313 (7) as well as nylon-613 was found to have exceptionally low sensitivity to moisture, excellent dimensional stability. and dielectric properties. Long-chain nylons of this type have found niche markets in automotive parts.

Jojoba

Jojoba (Simmondsia chinensis) is a shrub that is native to the Sonoran desert; its seed contains 50% of a liquid wax ester. The linear wax ester has a main fatty acid frag-

Copyright © 1999 by AOCS Press

46

T.A. Isbell

ment of C20:1 (∆11) 71%, coupled to a C20:1 (∆11) 45% or C22:1 (∆13) 44% alcohol, respectively. Consequently, the wax is composed of mainly C40:2 and C42:2 esters that remain liquid at room temperature due to its unsaturation. Jojoba is recognized as one of the more oxidatively stable oils with an oxidative stability index (OSI) time of 55.9 h at 110°C for crude jojoba oil (8). Aside from the use of the wax in cosmetics and specialty lubricant applications, the cost of the oil has limited its scope of incorporation into the industrial market. However, efforts by Shani et al. (9-12) have provided some degree of functionalization of jojoba, particularly at the olefin positions. These functionalizations fall into the classes of halogenation, allylic bromination-dehydrohalogenation, sulfurization, phosphorylation, ozonolysis, amination, oxidation, and pyrolysis. These derivatives are intended to boost the polarity of the wax or provide functionalization for binding to polymers. Bromination of jojoba oil in carbon tetrachloride yielded tetrabromojojoba derivatives at 20°C (9). When treated with excess base, these bromides yielded the corresponding acetylenes from Z,Z olefins of jojoba or allenes from the E,E isomerized jojoba (Scheme 2) with the expected hydrolysis to acid and alcohol. Allylic bromination (nonregiospecific) with N-bromosuccsinimide (NBS) followed by dehydrohalogenation yielded polyunsaturated oils with degrees of unsaturation up to the hexaenoic jojoba derivative (10). These highly unsaturated materials were envisioned

SCHEME 2.

Copyright © 1999 by AOCS Press

Derivatives of New Crop Oils

47

to be useful for drying oils. Jojobatetraene prepared from 2 equivalents of NBS yielded a diene system that provided Diels-Alder adducts when reacted with various dienophiles, including singlet oxygen (11). These adducts provided some unique functionalization to the jojoba backbone as shown in Scheme 3. Oxidation of jojoba by t-butyl hypochlorite and condensation with 6-chlorohexanol gave the near quantitative addition of chlorine and alcohol (nonregiospecifically) across the double bonds according to Scheme 4 (13). If this reaction is performed in the presence of hydrogen peroxide (no alcohol present), the reaction will yield the tetrachloro derivative substituted at the allylic positions. Treatment of jojoba with potassium permanganate yielded pelargonic acid and α,ω-dicarboxylic acids resulting from overoxidation of the olefin in conjunction with hydrolysis of the ester. Diols were not obtained in the permanganate reaction under all of the conditions examined. However, tetraols of jojoba were obtained when the oil was treated with a hydrogen peroxide and formic acid epoxidation technique. When m-chloroperbenzoic acid was used as the epoxidizing reagent, diepoxides were isolated in 80% yield (12). The diepoxide could be converted to the tetraol by refluxing in 20% aqueous hydrochloric acid. Jojoba has also been converted to its diozonide (14), which was used as an intermediate to endojojobadialdehydes, endojojobadiacids, pelargonic acid, and pelargonaldehyde (Scheme 5). The process involves the production of diozonide in a neat solution of jojoba wax at 40-50°C. Oxidative decomposition in a stream of oxygen at 90-95°C gave diacids, whereas reduction with triphenyl phosphine or steam distillation gave rise to dialdehydes. Violent decomposition of the ozonides was noted at temperatures Ɑ110°C. The ene reaction of olefins was employed to hydroxymethylate jojoba wax using Lewis acids as catalyst (15). Ethylaluminum dichloride and formaldehyde gave both

SCHEME 3.

Copyright © 1999 by AOCS Press

48

T.A. Isbell

SCHEME 4.

regio ene adducts of jojoba with yields nearing 52% for dihydroxymethyl and 32% of mono hydroxymethylated product (Scheme 6). Several papers have reported coupling of jojoba wax to polymer matrices to form a substrate capable of chelating to a metal ion. The site of chelation was developed through derivatization of the olefin region of the polymer-bound jojoba wax. Allylic brominated derivatives of jojoba (reported above) were bound to polymer matrices by a C-C bond-forming reaction using brominated jojoba and lithiated polystyrene (16). A C-N linkage to the polymer was also used to bind brominated jojoba to aminated polymers (17). Functionalization of the residual double bonds of jojoba by

SCHEME 5.

Copyright © 1999 by AOCS Press

Derivatives of New Crop Oils

49

SCHEME 6.

sulfur-chlorination or phosphonation (18) gave functional derivatives capable of forming metal chelates.

Lesquerella

Plant species of the genus Lesquerella are being developed as an alternative crop for the southwestern region of the U.S. The seed oil of Lesquerella contains 55-60% of 14-hydroxy-cis-11-eicosenoic acid (lesquerolic acid), a homologue of ricinoleic acid obtained from castor oil (19). Chemical modifications of lesquerolic acid, thus far, have nearly duplicated derivatizations of ricinoleic acid. One of the more historic reactions of hydroxy fatty acids is the alkaline cleavage of lesquerolic and ricinoleic acid to 2-octanone and 12- or 10-hydroxydecanoic acid, respectively, with one mole equivalent of alkali and temperatures of 180-200°C (Scheme 7). Reactions at higher temperatures (250-275°C) with two mole equivalents of alkali will predominately yield 2-octanol (capryl alcohol) and 1,12-dodecanedioic acid from lesquerolic acid and 1,10-decandioic acid (sebacic acid) from ricinoleic acid (20,21). The hydroxy fatty acids are dehydrogenated under the alkaline conditions to the β,γ-unsaturated ketone. Under the reaction conditions, the unsaturated ketone readily isomerizes to an α,β-unsaturated ketone, which undergoes a retro-aldol to give 2-octanone and ω-aldehydo acid. Under the reducing conditions of the reaction, an equilibrium exists between the ω-aldehydo acid and the hydrogenation of the aldehyde to 12-hydroxydodecanoic acid. This equilibrium can be shifted to the ω-hydroxy acid by the addition of excess secondary alcohol such as 2-octanol. In the presence of excess alkali and at higher temperatures (250-275°C), an irreversible conversion of the ω-aldehydo acid to the 1,12-dodecanedioic acid occurs. Manufacturing yields were reported as high as 80% for sebacic acid from castor oil and a 48% yield of the dibasic acid from lesquerolic acid (20). 1,12-Dodecanedioic acid from petroleum-based butadiene is used as the major ingredient in the synthesis of

Copyright © 1999 by AOCS Press

50

T.A. Isbell

SCHEME 7.

nylon-1212, nylon-6,12, and other molded plastics. 2-Octanol is used in perfumes, soaps, and antifoaming agents. Hydroxy fatty acids such as lesquerella can be readily converted to estolides (22,23) either as triglycerides in the presence of free fatty acid or from homopolymerization of the split fatty acids. Scheme 8 depicts the formation of estolide from triglyceride with a free carboxyl functionality. Lesquerella estolides have been synthesized using clays (24) and enzymes (22) as catalysts. Castor oil estolides demonstrate that this reaction can be run in the absence of catalyst at high temperature under vacuum or a blanket of carbon dioxide (25). Estolide triglycerides are useful as viscosity improvers in vegetable-based lubricants (26) and as basestock for lubricants (27).

Copyright © 1999 by AOCS Press

Derivatives of New Crop Oils

51

SCHEME 8.

Lesquerella oil was first dehydrated by Goldblatt and Knowles (28) using NaHSO4 as a catalyst. This method was rediscovered by Thames et al. (29). Reactions require 1-2% catalyst such as NaHSO4, KHSO4, or CuSO4 at 230-240°C. Conversion rates are good (70-95%) with reaction times ranging from 5 min to 6 h. The conjugated to nonconjugated ratio is 1.5:1 for lesquerella oil; drying times are 2.5-5 h, which is equivalent to commercially available dehydrated castor oil. The estolide of lesquerella oil could also serve as an intermediate to dehydrated lesquerella oil as has been demonstrated in castor oil (23). Lesquerella oil has been epoxidized by traditional methods (30) on a laboratory scale with m-chloroperoxybenzoic acid (MCPBA) as outlined in Scheme 9. Reactions were performed below 25°C with a 10% excess of MCPBA to provide an 86% theoretical yield of epoxidized lesquerella oil, which has an oxirane content of 4.95%.

SCHEME 9.

Copyright © 1999 by AOCS Press

52

Meadowfoam

T.A. Isbell

Meadowfoam (Limnanthes alba) is a developing new crop grown on a commercial scale (~8000 acres in 1997) in the Willamette Valley of Oregon in the northwestern United States. Other small test plots are grown elsewhere in the U.S. and abroad. Meadowfoam is composed of unique long-chain fatty acids with 5-eicosenoic acid (62%) as the major fatty acid. The other main components are 5,13-docosadienoic acid (19%), 5-docosenoic acid (3%) and 13-docosenoic acid (10%). The ∆5 unsaturation provides an excellent chemical moiety for synthetic modifications to the oil or fatty acids by providing a site at which a carboxylate stabilized carbocation can form. In addition, the ∆5 unsaturation has enhanced resistance to oxidative degradation (8) as evidenced by the high OSI of the oil (246.9 h at 110°C) or its methyl ester (69.4 h at 90°C). Estolides can be synthesized from meadowfoam fatty acids by the method depicted in Scheme 10. Estolides are oligomeric esters of fatty acids resulting from a cationic dimerization of two fatty groups in the presence of an acid catalyst. Various Brönsted acids can catalyze this oligomerization (31), but in the case of meadowfoam fatty acids (32), concentrated (70%) perchloric acid is the most effective catalyst. Yields of 60-75% are obtained for the estolide when 0.05 mol of perchloric acid are used per mole of starting fatty acid. A temperature of 65°C and a reaction time of 24 h provided the highest observable yields of estolide. One of the most important aspects of estolides is their demonstrated biodegradability (33). In addition, the oleic estolides and estolide esters show a great reduction in their pour points (pour points of −31 to −45°C) in comparison to their parent fatty acids or triglycerides. In contrast, meadowfoam estolides and the parent triglyceride

SCHEME 10.

Copyright © 1999 by AOCS Press

Derivatives of New Crop Oils

53

have nearly identical pour points, 6 and 1°C, respectively. However, meadowfoam estolides have been shown to have special properties that are useful in the cosmetic industry. δ-Lactones could be synthesized from meadowfoam oil or its fatty acids (Scheme 11) when reacted with mineral acid catalysis (34). Yields for δ-lactone were 90-100% of theoretical, based on the ∆5 unsaturation present in the starting oil. The δ-lactone ring structure is the preferred kinetic product of the reaction, although the thermodynamics favor the more stable γ-lactone ring. Conditions that favor kinetic control such as dilution and use of a nonparticipating polar solvent, which aid in cation stabilization, gave good selectivity for δ-lactone. When 500 wt% methylene chloride and perchloric acid were used, the lactonization reaction gave a 38.5:1 δ:γ ratio. δ-Lactones, due to their ring structure, are easily opened with nucleophiles (35) to yield the corresponding 5-hydroxy substituted derivatives (Scheme 12). Table 1 lists the relative rates of derivitization of meadowfoam fatty acids, γ-lactones, and δlactones in reactions with alcohols and amines. All of the relative rates were compared with the rate of esterification for meadowfoam fatty acids (FA). These data clearly demonstrate the enhanced reactivity of the δ-lactone structure with respect to fatty acids and even its analog, γ-lactone. Another aspect of the δ-lactone chemistry is the interconversion between δ-lactone and 5-hydroxy fatty acid. Under basic aqueous conditions, the equilibrium favors the formation of the 5-hydroxy fatty acids. In contrast, catalytic amounts of acid will cause rapid cyclization/dehydration of the 5-hydroxy fatty acid into the δ-lactone. By utilizing the equilibrium between these two species, 5-hydroxy fatty acids were synthesized directly from meadowfoam oil (Scheme 13). δ-Lactones and 5-hydroxy fatty acids are capable of undergoing an etherification reaction to yield a variety of secondary ethers (36). Yields of secondary ethers were 84-94% with the reaction performed as outlined in Scheme 14. Primary alcohols readily undergo etherification with either δ-lactone, γ- lactone, or 5-hydroxy fatty acids. ∆5-Unsaturated fatty acids, however, do not undergo etherification under these reaction conditions, demonstrating the necessity of an oxy-

SCHEME 11.

Copyright © 1999 by AOCS Press

54

SCHEME 12.

T.A. Isbell

genated functionality at the C-4 or C-5 position. Alcohols that form ethers with δ-lactones range in chain length from methyl through oleyl and also include the βbranched alcohols such as 2-ethylhexyl. Isopropyl alcohol failed to provide ethers, possibly due to the greatly reduced nucleophilicity of this alcohol. Etherification catalysts include mineral acids such as sulfuric and perchloric acid, but these tend to provide double the amounts of dehydration products compared with other catalysts. Lewis acids such as boron trifluoride and iron (III)-exchanged montmorillonite clays are also effective at catalyzing these reactions. Boron trifluoride provides the most consistent yields; increases in catalyst concentration generally increase ether yields and/or reduce reaction times. Iron (III) clays provided high

Copyright © 1999 by AOCS Press

Derivatives of New Crop Oils

SCHEME 13.

55

yields of ethers in small-scale reactions but failed to yield significant quantities of ether when reactions were performed on a 100-g scale. Other catalysts that provided modest yields of ether are strongly acidic ion-exchange resins and p-toluenesulfonic acid. 5-Unsaturated fatty acids, when converted to their epoxides, cyclize readily in the presence of acid to 6-hydroxy-δ-lactones (37,38). 5,6-Epoxy fatty acids could be subsequently converted to the corresponding 6-hydroxy-δ-lactone using an acid catalyst. Both the traditional peracetic/acetic acid method (37) and the modern lipase and hydrogen peroxide method (38) gave epoxides from meadowfoam fatty acids with yields of 50 and 98%, respectively. Subsequent treatment of the epoxide with sulfuric acid in toluene or dihydroxyeicosanoic acid in xylene gave the 6-hydroxy-δ-lactone in high yield (77% from the starting fatty acids). An interesting side product to the

SCHEME 14.

Copyright © 1999 by AOCS Press

56

T.A. Isbell

cyclization reaction is the formation of the hydroxy estolide, in significant quantities (41%) under some reaction conditions. Meadowfoam oil has been cross-linked with sulfur (39) to form factice (vulcanized oil). The Shore hardness of meadowfoam factice was superior to factices prepared from other vegetable oils, giving materials that had little to no stickiness. Other reported derivatives of meadowfoam include dimer acids and amides. Dimer acids were synthesized (40) from meadowfoam fatty acids at 250°C using montmorillonite clay as a catalyst with a yield of 44%. The physical properties (viscosity and lubricity) of the dimer acids were similar to those of other dimer acids prepared from highly monounsaturated fatty acids. Simple amides were also made by reaction of ammonia with fatty acids in refluxing xylenes (41). References

1. Kleiman, R., and L.H. Princen, New Industrial Oilseed Crops, in Proceedings of World Conference on Oleochemicals: Into the 21st Century, edited by T.H. Applewhite, Kuala Lumpur, Malaysia, 1990, pp. 127-131. 2. Carlson, K.D., and D. L. Van Dyne, Industrial Uses for High Erucic Acid Oils from Crambe and Rapeseed, MP-Univ.-Mo-Ext.-Div., MP678: 32 (Oct. 1992). 3. Leonard C.E., High Erucic Vegetable Oils, Ind. Crop and Prod. 1: 119-123 (1993). 4. Molnar, N.M., Eurcamide, J. Am. Oil Chem. Soc. 51: 84-87 (1974). 5. Mod, R.R., F.C. Magne, E.L. Skau, H.J. Nieschlag, W.H. Talent, and I.A. Wolff, Preparation and Plasticizing Characteristics of Some N,N-Disubstituted Amides of Erucic and Crambe Acids, I. & B.C. Prod. Res. Dev. 8: 176-182 (1969). 6. Carlson, K.D., and V.E. Sohns, Brassylic Acid: Chemical Intermediate from High-Erucic Oils, Ind. Eng. Chem. Prod. Res. Dev. 16: 95-101 (1977). 7. Nieschlag, H.J., J.A. Rothfus, V.E. Sohns, and R. Beltron Perkins, Jr., Nylon-1313 from Brassylic Acid, Ind. Eng. Chem. Prod. Dev. 16: 101-107 (1977). 8. Isbell, T.A., T.P. Abbott, and K.D. Carlson, Oxidative Stability Index of Vegetable Oils in Binary Mixtures with Meadowfoam Oil, Ind. Crop, and Prod. 9: 115-123 (1999). 9. Shani, A., Functionalization at the Double Bond Region of Jojoba Oil: I. Bromine Derivatives, J. Am. Oil Chem. Soc. 58: 845-850 (1981). 10. Shani, A., Heavily Brominated and Highly Unsaturated Derivatives of Jojoba Oil, J. Am. Oil Chem. Soc. 65: 1318-1323 (1988). 11. Shani, A., Functionalization at the Double Bond Region of Jojoba Oil: II. Diels-Alder Adducts of Jojobatetraene, J. Am. Oil Chem. Soc. 59: 228-230 (1982). 12. Shani, A., Functionalization at the Double Bond Region of Jojoba Oil. 3. Hydroxylic Derivatives, Ind. Eng. Chem. Prod. Res. Dev. 22: 121-123 (1983). 13. Galun, A.B., S. Grinberg, A. Kampf, and E. Shaubi, Oxidation and Halogenation of Jojoba Wax, J. Am. Oil Chem. Soc. 61: 1088-1089 (1984). 14. Zabicky, J., and M. Mhasalkar, Diozonide of Jojoba Wax as an Intermediate for Synthesis, J. Am. Oil Chem. Soc. 63: 1547-1550 (1986). 15. McLellan, J.F., R.M. Mortier, S.T. Orszulik, and R.M. Paton, Hydroxymethylation of Jojoba Oil by Lewis Acid-Catalyzed Ene Reaction with Formaldehyde, J. Am. Oil Chem. Soc. 71: 231-232 (1994). 16. Binman, S., S. Belfar, and A. Shani, Functionalization at the Double-Bond Region of Jo-

Copyright © 1999 by AOCS Press

Derivatives of New Crop Oils

57

joba Oil: 7. Chemical Binding of Jojoba Liquid Wax to Polystyrene Resins, J. Am. Oil Chem. Soc. 73: 1075-1081 (1996). 17. Binman, S., S. Belfar, and A. Shani, Functionalization at the Double-Bond Region of Jojoba Oil: 8. Chemical Binding of Jojoba Liquid Wax to a Polymer Matrix via an Amine “Spacer,” J. Am. Oil Chem. Soc. 73: 1083-1095 (1996). 18. Binman, S., S. Vega, S. Belfar, and A. Shani, Functionalization at the Double-Bond Region of Jojoba Oil. 9. Solid-State Nuclear Magnetic Resonance Characterization of Substituted Jojoba Wax Chemically Bonded to a Polystyrene Matrix, J. Am. Oil Chem. Soc. 75: 521-525 (1998). 19. Roetheli, J.C., K.D. Carlson, A.E. Thompson, D.A. Dierig, L.K. Glaser, M.G. Blase, and J. Goodell, Lesquerella as a Source of Hydroxy Fatty Acids for Industrial Products, United States Department of Agriculture Cooperative Research Service: Growing Industrial Materials Series (Oct. 1991). 20. Diamond, M.J., R.G. Binder, and T.H. Applewhite, Alkaline Cleavage of Hydroxy Unsaturated Fatty Acids I. Ricinoleic Acid and Lesquerolic Acid, J. Am. Oil Chem. Soc. 42: 882-884 (1965). 21. Naughton, F.C., Production, Chemistry, and Commercial Applications of Various Chemicals from Castor Oil, J. Am. Oil Chem. Soc. 51: 65-71 (1974). 22. Hayes, D.G., and R. Kleiman, Lipase-Catalyzed Synthesis and Properties of Estolides and Their Esters, J. Am. Oil Chem. Soc. 72: 1309-1316 (1995). 23. Penoyer, C.E., W. von Fischer, and E.G. Bobalek, Synthesis of Drying Oils by Thermal Splitting of Secondary Fatty Acid Esters of Castor Oil, J. Am. Oil Chem. Soc. 31: 366370 (1954). 24. Burg, D.A., R. Kleiman, and S.M. Erhan, Production of Hydroxy Fatty Acids and Estolide Intermediates, U.S. Patent 5,380,894 (1995). 25. Achaya, K.T., Chemical Derivatives of Castor Oil, J. Am. Oil Chem. Soc. 48: 758-763 (1971). 26. Lawate, S.S., Triglyceride Oils Thickened with Estolides of Hydroxy-Containing Triglycerides, U.S. Patent 5,427,704 (1995). 27. Lawate, S.S., Lubricating Oil Compositions Containing Estolides of Hydroxy-Containing Triglycerides and a Performance Additive, U.S. Patent Appl. 188,263 (1994). 28. Goldblatt, L.A., and R.E. Knowles, Dehydration of Lesquerolates, U.S. Patent 3,291,816 (1963). 29. Thames, S.F., Y. Haibin, M. Wang, and T.P. Schuman, Dehydration of Lesquerella Oil, J. Appl. Polym. Sci. 58: 943-950 (1995). 30. Carlson, K.D., R. Kleiman, and M.O. Bagby, Epoxidation of Lesquerella and Limnanthes (Meadowfoam) Oils, J. Am. Oil Chem. Soc. 71: 175-182 (1994). 31. Isbell, T.A., R. Kleiman, and B.A. Planner, Acid-Catalyzed Condensation of Oleic Acid into Estolides and Polyestolides, J. Am. Oil Chem. Soc. 71: 169-174 (1994). 32. Isbell, T.A., and R. Kleiman, Mineral Acid-Catalyzed Condensation of Meadowfoam Fatty Acids into Estolides, J. Am. Oil Chem. Soc. 73: 1097-1107 (1996). 33. Erhan, S.M., and R. Kleiman, Biodegradation of Estolides, J. Am. Oil Chem. Soc. 74: 605-607 (1997). 34. Isbell, T.A., and B.A. Plattner, A Highly Regioselective Synthesis of δ-Lactones from Meadowfoam Fatty Acids, J. Am. Oil Chem. Soc. 74: 153-158 (1997). 35. Isbell, T.A., and B.A. Steiner, The Rate of Ring Opening of γ- and δ-Lactones Derived from Meadowfoam Fatty Acids, J. Am. Oil Chem. Soc. 75: 63-66 (1998). 36. Isbell, T.A., and M.S. Mund, Synthesis of Secondary Ethers Derived from Meadowfoam

Copyright © 1999 by AOCS Press

58

T.A. Isbell

Oil, J. Am. Oil Chem. Soc. 75: 1021-1029 (1998). 37. Fore, S.P., and G. Sumrell, Some Derivatives of 5-Eicosenoic Acid, J. Am. Oil Chem. Soc. 43: 581-584 (1966). 38. Frykman, H. B. and T.A. Isbell, Synthesis of 6-Hydroxy δ-Lactones and 5,6-Dihydroxy Eicosanoic/Docosanoic Acids from Meadowfoam Fatty Acids via Lipase-Mediated SelfEpoxidation, J. Am. Oil Chem. Soc. 74: 719-722 (1997). 39. Erhan, S.M., and R. Kleiman, Meadowfoam Oil Factice and Its Performance in Natural Rubber Mixes, Rubber World 203: 33-36 (1990). 40. Burg, D.A., and R. Kleiman, Meadowfoam Fatty Amides: Preparation, Purification and Use in Enrichment of 5,13-Docosadienoic Acid and 5-Eicosenoic Acid, J. Am. Oil Chem. Soc. 68: 190-192 (1991). 41. Burg, D.A., and R. Kleiman, Preparation of Meadowfoam Dimer Acids and Dimer Esters, and Their Use as Lubricants, J. Am. Oil Chem. Soc. 68: 600-603 (1991).

Copyright © 1999 by AOCS Press

Chapter 4

Palladium(0)-Catalyzed Reactions of Nucleophiles with Allyl Carbonates of Unsaturated Fatty Acids Hans J. Schäfer and Michael Zobel

Institute of Organic Chemistry, University of Münster, 48149 Münster, Germany

Introduction

Until now, conversions in the alkyl chain of unsaturated fatty acids have been developed preferentially at the double bond in electrophilic (1–4) or radical additions (5), in cycloadditions (6,7), and in metathesis (8). If the double bond is conjugated to a carbonyl group, the vinylogous addition of nucleophiles to the double bond becomes possible (Michael-addition); this leads to a rich synthetic chemistry (9,10). Allylic acetates or carbonates can undergo nucleophilic substitutions via palladium(0)-catalysis (11). In this paper, we report on the extension of this reaction to unsaturated fatty acids by the preparation of allyl carbonates and acetates of oleic, linoleic, and 10-undecenoic acid and their substitution with carbon- and heteroatomnucleophiles by palladium(0)-catalysis. In this way, different substituents can be introduced into the alkyl chain of fatty acids. This leads to fatty acid derivatives in which the properties of biologically active compounds may possibly be combined with the amphiphilic property of the fatty acid.

Palladium(0)-Catalyzed Reactions of Nucleophiles with Allyl Carbonates

In the last decade, palladium(0)-catalyzed reactions have received great interest in the synthesis of highly functionalized and structurally complex compounds (11). Although palladium (II) compounds are electrophiles and thus react preferentially with electron-rich organic compounds, palladium(0)-complexes show a nucleophilic behavior and are reacted preferentially with halides, acetates, and triflates (12). Substrates suitable for palladium(0)-catalyzed allylic substitution are allylhalides, allylesters, and allylalcohols (13), allylphosphates (14) and allylamines (15). Here the following sequence of reactivity with regard to the leaving group holds: Cl > OCO2R > OAc >> OH. In allylacetates, the acetoxy group can be substituted by way of palladium(0)catalysis. A base must be added to generate a nucleophilic anion and to neutralize the acetic acid. Allyl carbonates A (Scheme 1) are better substrates for the substitution by palladium(0)-catalysis than actetates because they require no base and thus tolerate 59 Copyright © 1999 by AOCS Press

60

H.J. Schäfer and M. Zobel

SCHEME 1. Catalytic cycle for the palladium(0)-catalyzed addition of nucleophiles to allyl carbonates (16).

base-sensitive substrates. The first step in these reactions is the oxidative addition of palladium(0) to the allylic system A to form a σ-allylpalladium-complex, which quickly decomposes into the more stable π-allylpalladium complex B, which then undergoes decarboxylation to C (Scheme 1). The alkoxide is sufficiently basic to deprotonate a number of nucleophiles; in the presence of phosphines, it reacts with complex D. Amines or stabilized carbanions are used primarily, whereas oxygen nucleophiles have been applied in only a few cases up to now. To extend this palladium(0)-catalyzed substitution to unsaturated fatty acids, their allyl carbonates must be prepared.

Preparation of Allyl Carbonates of Unsaturated Fatty Acids Selenium Dioxide Oxidation

For the allylic oxidation of alkenes, a large variety of methods have been reported in the literature (17–20). However, after a fair number of those oxidants were applied to methyl 10-undecenoate [1], only the substoichiometric oxidation with selenium dioxide (17) was found to work satisfactorily. Methyl 10-undecenoate [1] in dichloromethane was stirred with 0.5 equivalents of selenium dioxide and t-butylhydroperoxide for 48 h at room temperature to yield 55% allylalcohol [2] and 7%

Copyright © 1999 by AOCS Press

Palladium(0)-Catalyzed Reactions of Nucleophiles

61

vinylketone [3] (Eq. 1). Selenium dioxide can be precipitated after the reaction by petroleum ether and recycled. With a small decrease in yield, the amount of selenium dioxide can also be reduced from 0.5 to 0.1 equivalents. By atom absorption analysis the amount of selenium was determined in [2] to be <1 ppm.

Similarly, methyl oleate [4] was converted into the isomeric allylic alcohols [5a,b] (Eq. 2).

Singlet Oxygen Oxidation

An economically (cheaper reagents, higher yields) and ecologically (avoidance of selenium dioxide) more favorable route for oxidation of vic-disubstitued alkenes is photo-oxygenation (21,22). This reaction has been applied to [4] and yields [5c,d] in nearly quantitative yield (Eq. 3).

Copyright © 1999 by AOCS Press

62

H.J. Schäfer and M. Zobel

Singlet oxygen is generated by irradiation of oxygen with two 1000-W sodium-vapor lamps using tetraphenylporphine as photosensitizer. After 5 h, the reaction is complete and triphenylphosphine is added until the peroxide test is negative. Then [5c,d] can be isolated by flash chromatography. Allylic Oxygenation by Cobalt (III)-Acetate

A further possibility to oxidize an unsaturated fatty acid in the allylic position is the acetoxylation of [1] and [4] with Co(OAc)3 and sodium bromide (23). When this method was applied to [1] and [4], the acetates [6a,b] and [7a,b] were obtained in moderate yields of 30% (Eq. 4).

Anodic Acetoxylation of Methyl Konjuenate

Another method to prepare compounds that are acetoxylated in the allylic position is the anodic addition of methyl konjuenate [8] (24,25), which affords the isomeric diacetates [9] in 80% yield (Eq. 5). In this reaction, [8] is oxidized at the anode to the radical cation, which reacts by acetolysis, oxidation of the formed allyl radical to an allyl cation, which in turn undergoes acetolysis to the isomeric diacetates [9].

Copyright © 1999 by AOCS Press

Palladium(0)-Catalyzed Reactions of Nucleophiles

63

Preparation of the Allyl Carbonates from the Allylic Alcohols

The allyl carbonates were obtained from the corresponding alcohols by reaction with methyl chloroformate in the presence of pyridine in dichloromethane (16). With [2], the allyl carbonate [10a,b] was obtained in 94% yield; correspondingly, [11a,b] was prepared from [5a,b] or [5c,d] in 73% yield (Eq. 6).

Addition of Carbon-Nucleophiles to the Allyl Carbonates [10a,b] and [11a,b] of Methyl 10-Undecenoate and Methyl Oleate

When allyl carbonates are used in palladium(0)-catalyzed substitutions, the reaction proceeds under neutral conditions, which tolerate base-sensitive groups in the molecule. This method, developed by Tsuji (16), was applied to [10a,b]. As carbon nucleophiles, dimethylmalonate, nitromethane, and methyl acetylacetate were used. In general, the reactions were conducted as follows: under argon, the allyl carbonate [10a,b] and the nucleophile were dissolved in absolute tetrahydrofuran. Pd2(dba)3CHCl3 (1.0–1.5 mol%) was converted in a separate flask with triphenylphosphine under argon in absolute tetrahydrofuran to the catalyst Pd(PPh3)4. The formation of the catalyst is indicated by a color change from dark red to orange. Subsequently, the catalyst is transferred to the solution with allyl carbonate and nucleophile, which is stirred at room temperature. In this way, the substituted methyl undecenoates [12], [13], and [14] were obtained in 82, 72, and 75% yield, respectively (Eq. 7). During the substitution of dimethylmalonate and methyl acetylacetate, the isomer, into which the nucleophile was introduced at C-11, is formed nearly exclusively. In the addition of nitromethane, however, a 1:1-mixture of the isomers methyl (E)12-nitro-9-dodecenoate and methyl 10-nitro-9-vinyldecanoate [13a,b] was obtained. Corresponding to [10a,b], the allyl carbonates [11a,b] were reacted via palladium(0)-catalysis with carbon nucleophiles. The substitution products [15] and [16]

Copyright © 1999 by AOCS Press

64

H.J. Schäfer and M. Zobel

were obtained in good yields as a mixture of regioisomers (Eq. 8). The product [15] was hydrolyzed to the triacid and the interfacial properties of its sodium salt were determined (see below). The surface tension in this case does not decrease to such an extent as that of other anionic surfactants, which is possibly due to the electrostatic repulsion of the cumulated carboxylates.

Palladium(0)-Catalyzed Introduction of Oxygen Nucleophiles into the Allyl Carbonates of Methyl 10-Undecenoate and Methyl Oleate

Corresponding to the reaction with carbon nucleophiles, [11a,b] was reacted with different oxygen nucleophiles such as phenols, carbohydrates, steroids, and α-tocopherol. In addition to 3-methoxyphenol, 2,4-dichlorophenol and 8-quinolinol were used, leading to the substitution products [17]–[19] as mixtures of regioisomers (Eq. 9). The substitution product [17] exhibits interesting interfacial properties as an anionic surfactant after saponification (see below). It decreases the surface tension to the same extent as commercial anionic surfactants. In addition, it possesses a good emulsifying property and a bacteriostatic effect against Escherichia coli at a concentration of 500 µg/mL; [17] is also biodegradable. 2,4-Dichlorophenoxyacetic acid

Copyright © 1999 by AOCS Press

Palladium(0)-Catalyzed Reactions of Nucleophiles

Copyright © 1999 by AOCS Press

65

66

H.J. Schäfer and M. Zobel

(2,4 D) is a herbicide that has been used frequently in the past (26,27). The substitution product [18] is a fatty acid conjugate with this herbicide. 8-Quinolinol is a classical fungicide (28,29). The reaction with [11a,b] leads to the 8-quinolinoxyfatty acid [19]. Compounds [18] and [19] are being tested for their fungicidal or herbicidal activity. Similarly, the allyl carbonate [10a,b] was treated with methanol and 3methoxyphenol to afford [20] and [21] in satisfactory to excellent yield as mixtures of regioisomers (Eq. 10).

In the regioisomers [20], the ratio of substitution at C-9:C-11 is 1:2, whereas in [21], the ratio of regioisomers at C-9:C-11 is 1.5:1; [21] belongs to the group of juvenile hormone analogs. It has been tested for suitability as an insecticide but has only a weak effect. The carboxylate of [21] shows a moderate reduction of the surface tension of water. Subsequently, an attempt was made to extend the reaction to hydroquinone as a nucleophile to obtain fatty acid derivatives with phenols having a free hydroxyl group. These would be potential antioxidants. Furthermore, this reaction could lead to dimer fatty acids, whose two carbon chains are linked together by way of the hydroquinone. Dimer fatty acids are interesting comonomers for polyesters and are of technical interest as lubricants (30). For that purpose, dimer fatty acids with a uniform structure, which are expected here, could have advantages. The allyl carbonates [10a,b] and [11a,b] were reacted by palladium(0)-catalysis with different equivalents of hydroquinone to direct the product formation either to the monomer or to the dimer. This is indeed possible as indicated by Eqs. 11 and 12 and Tables 1 and 2.

Copyright © 1999 by AOCS Press

Palladium(0)-Catalyzed Reactions of Nucleophiles

Copyright © 1999 by AOCS Press

67

68

H.J. Schäfer and M. Zobel

For both allyl carbonates, the reactions proceed with high yields in monomer and dimer depending on the ratio of allyl carbonate to [22]. With 0.5 equivalents of hydroquinone, one obtains in both cases the dimer fatty acids [24] and [26] in yields of about 80%. With 5 equivalents of hydroquinone, the monomers [23] and [25] are obtained in about 75% yield. Another possibility for substituting the hydroxyl group by a nucleophile is the Mitsunobu reaction (31,32). In this case, the alcohol is linked to the nucleophile using diethyl azodicarboxylate (DEAD) and triphenylphosphine. This reaction was applied to the allylic alcohols [5a,b] (from methyl oleate) (Eq. 13, Table 3) and [2] (from methyl 10-undecenoate) (Eq. 14, Table 4). In the Mitsunobu reaction with the allylic alcohols [2] and [5a,b] as with the palladium(0)-catalyzed reaction, good yields are obtained for the monomers [23] and [27] with an excess of nucleophile [22]. The dimers [24] and [28], however, can be prepared only in moderate yields. Because the reaction does not proceed via a π-allyl palladium complex with two reaction sites, additional isomers are not produced in this reaction. The conversion of [2] therefore yields only [27] and [28] as product. The Mitsunobu reaction saves one reaction step compared with the palladium(0)catalysis, i.e., the preparation of the allyl carbonate. It also works with nucleophiles that can deactivate the palladium(0)-complex by complexation, which can lead to a failure of palladium(0)-catalysis (33). These features can sometimes be of advantage in synthesis. Both [23] and [25] are moderate antioxidants, whereas [24] shows no antioxidant activity.

Copyright © 1999 by AOCS Press

Palladium(0)-Catalyzed Reactions of Nucleophiles

Copyright © 1999 by AOCS Press

69

70

H.J. Schäfer and M. Zobel

Palladium(0)-Catalyzed Introduction of Carbohydrates into the Allyl Carbonates of Methyl 10-Undecenoate and Oleate

Carbohydrates are gaining increasing interest in the field of surfactants. In the classical nonionic surfactants, oligoethyleneglycols are often used as the polar group; today, glucose is used increasingly as an alternative hydrophilic group for surfactants. The combination of the renewable resources, fats and oils and carbohydrates, received a new impulse by the growing public sensitivity to save fossil feedstocks. In the meantime, alkylpolyglycosides are technically produced in a one-step process using a Fischer synthesis from glucose and fatty alcohol in quantities of >100,000 t/y. They are biodegradable and exhibit synergistic interactions with other surfactants. In addition, they are more foam active and skin friendly than other nonionic surfactants (34). Because methanol could be introduced into the allyl carbonate [10] by palladium(0)- catalysis, it appeared worthwhile to try the reaction of [10] and [11a,b] with different carbohydrates. This should lead to a new type of surfactant in which the polar group is attached to the middle of the fatty acid alkyl chain. They should be acid resistant contrary to the glycosides obtained in the Fischer synthesis. In the reaction with mannofuranose [29], the allyl carbonate [11a,b] from methyl oleate leads to the fatty acid-carbohydrate conjugate [30] as a mixture of regioisomers (Eq. 15). Furthermore, allyl carbonate [10] was reacted with 1,2,3,4-tetra-O-acetyl-α-Dglucopyranose and 2,3,4,6-tetra-O-benzyl-α-D-glucopyranose. By palladium(0)catalysis, the products [31] and [32] are obtained as mixtures of regioisomers. In both reactions, predominantly the isomer bound to C-11 of the carbohydrate is formed (Eq. 16). The ratio of diastereomers could not be determined exactly because

Copyright © 1999 by AOCS Press

Palladium(0)-Catalyzed Reactions of Nucleophiles

Copyright © 1999 by AOCS Press

71

72

H.J. Schäfer and M. Zobel

the relevant signals in the ¹H NMR were superimposed. In the 13C NMR (DEPTspectrum), however, weak signals for the secondary carbon atoms are observed in the olefinic region, thus supporting the connection of the carbohydrates to C-9. The determination of the interfacial properties of [31] and [32] shows that these compounds are suited as surfactants (see below).

Palladium(0)-Catalyzed Introduction of α-Tocopherol (Vitamin E) and Ostradiol into the Allyl Carbonates of Methyl 10-Undecenoate and Methyl Oleate

Earlier results in this paper showed that phenolic oxygen nucleophiles are especially suited to be added to fatty acid–derived allyl carbonates by palladium(0)-catalysis. Therefore it seemed worthwhile to investigate whether the natural products, α-tocopherol (vitamin E) [33] and ostradiol [34], could be introduced correspondingly because they possess a phenolic structural element. Indeed, with [11a,b], the fatty acid conjugates [35] and [36] can be obtained as mixtures of regioisomers in very high yield (Eq. 17). Correspondingly, the allyl carbonate [10] reacts with α-tocopherol and ostradiol in very good yields to form fatty acid derivatives (Eq. 18). The compounds [37] and [38] are obtained as mixtures of regioisomers; [37a,b] in a ratio of 1.5:1 in favor of the C-9 isomer and [38a:b] in a ratio of 3.9:1 in favor

Copyright © 1999 by AOCS Press

Palladium(0)-Catalyzed Reactions of Nucleophiles

73

of the substitution at C-11. Because of the importance of estradiol (35) as estrogen, the new fatty acid conjugate [38] is being tested for its biological activity.

Palladium(0)-Catalyzed Introduction of Nitrogen Nucleophiles into the Allyl Carbonates of Methyl 10-Undecenoate and Methyl Oleate

Because nitrogen nucleophiles can complex with palladium(0) and thereby deactivate the catalyst, it appeared necessary to study the palladium(0)-catalyzed reaction of allyl carbonates with amines. Without deactivation of the catalyst, the reaction of [11a,b] with piperidine led to the piperidine conjugate [39] in a yield of 73% as a mixture of regioisomers (Eq. 19). After saponification, [39] proved to be a good anionic surfactant that strongly decreased the surface tension of water and formed micelles at a low concentration. However, it exhibited no bactericidal effect against Bacillus subtilis. In the same way, [10] was reacted with the primary amine, picolylamine, and the secondary amine, piperidine. In this case, the conjugates [40] and [41] were also obtained in moderate to good yields as single regioisomers (Eq. 20). Nitrogen-containing compounds are frequently used as plant-protecting chemicals. Appropriate tests with [39]–[41] in this direction, however, were negative.

Copyright © 1999 by AOCS Press

74

H.J. Schäfer and M. Zobel

Palladium(0)-Catalyzed Reaction of Carbon Nucleophiles with an Allylic Diacetate Obtained by Anodic Addition from Methyl Konjuenate

The allyl diacetate [9] can be obtained in 80% yield and in a very simple reaction by electrolysis of methyl konjuenate in an undivided cell in acetic acid as solvent (see above). It was therefore of great interest to determine whether this readily available acetate also reacts with nucleophiles under palladium(0)-catalysis. Indeed, [9] can be joined to carbon nucleophiles in moderate to good yield to form the conjugates [42] and [43] (Eq. 21). A disubstitution could not be enforced, however, despite the use of a fivefold excess of the nucleophile.

Copyright © 1999 by AOCS Press

Palladium(0)-Catalyzed Reactions of Nucleophiles

75

Investigation of Some Selected Properties of the New Fatty Acid Conjugates Tensiometric Measurements

For several of the new fatty acid derivatives, the surface tension and the cmc-values were determined. Some selected examples are given below.

Conjugated fatty acids derived from 10-undecenoate Sodium 9-hydroxy-10-undecenoate [44] σ = 44.2 mN/m CMC = 7.2 × 10−2 g/L (3.3 × 10−2 mol/L)

Mixture of sodium (E)-11-(3-methoxyresorcin-1-O-yl)-9-undecenoate [45a] and sodium (E)-9-(3-methoxyresorcin-1-O-yl)-10-undecenoate [45b] σ = 45.3 mN/m CMC = 0.13 g/L (4.2 × 10−2 mol/L)

O-1,4-[di-(E)-1-Sodium carbonyloxy-8-decen-10yl]-hydroxybenzene [46] σ = 39.5 mN/m CMC = 7.6 × 10−2g/l (0.14 mol/l)

Copyright © 1999 by AOCS Press

76

H.J. Schäfer and M. Zobel

Mixture of (E)-2,3,4,6-tetra-O-benzyl-1-O-(11sodium-carbonyloxy-2-decen-1-yl)-α-Dglucopyranose [47a] and 2,3,4,6-tetra-O-benzyl-1-O-(11-sodium carbonyloxy-1-decen-3-yl)-α-D-glucopyranose [47b] σ = 39.4 mN/m CMC = 2.9 × 10−3 g/L (4.1 × 10−3 mol/L) Conjugated fatty acids derived from oleic acid

Mixture of sodium (E)-11-hydroxy-9-octadecenoate [48a] and sodium (E)-8-hydroxy-9-octadecenoate [48b] σ = 31.1 mN/m CMC = 0.16 g/L (3.3 × 10−3 mol/L) Mixture of sodium (E)-8-(3-methoxyresorcin -l-Oyl)-9-octadecenoate and isomers [49] σ = 31.2 mN/m CMC = 0.12 g/L (0.27 × 10−3 mol/L)

Mixture of O-1,4-[di-(E)-1-sodium carbonyloxy-8-octadecen-7-yl]hydroxybenzene and isomers [50] σ = 32.8 mN/m CMC = 6.1 × 10−2 g/L (8.6 × 10−5 mol/L)

(E)-2,3,5,6-bis-O-Isopropyliden-1-O-(1-sodium carbonyloxy-8-heptadecen-7-yl)-α-Dmannofuranose and isomers [51] σ = 29.9 mN/m CMC = 2.5 × 10−2 g/L (4.1 × 10−6 mol/L)

Sodium (E)-8-piperidin-N-yl-9-octadecenoate and isomers [52] σ = 32.9 mN/m CMC = 1.4 × 10−2 g/L (3.5 × 10−5 mol/L)

Copyright © 1999 by AOCS Press

The data show that the derivatives of 10-undecenoic acid exhibit a surface activity; however, this is not as pronounced as that of the conjugates derived from oleic acid. Although the derivatives of 10-undecenoic acid exhibit surface tensions of 39– 47 mN/m at the critical micelle concentration, the values of those from oleic acid are at 29–34 mN/m. The cmc-values are lower for the oleates than for the undecenoates. These findings agree with the rule of Traube (36), which predicted a decrease of the surface tension with an increase of the alkyl chain. The introduction of double bonds or hydrophilic substituents into the alkyl chain decreases its hydrophobicity (37), which increased the solubility of the surfactant and with that, the cmc. The shift of hydrophilic groups from the end to the middle of the alkyl chain also increased (38). Bacteriostatic Activity of the Fatty Acid Derivatives

Many of the prepared compounds were examined in plate diffusion tests (PDT) or in turbidity measurements according to Klett concerning their bacteriostatic activity against gram-positive (B. subtilis) and gram-negative (E. coli) bacteria. In the plate diffusion test, only sodium (E)-8-(3-methoxyresorcin-1-O-yl)-9-octadecenoate [49] showed activity against E. coli at a concentration of 500 µg/mL.

Emulsifying Properties Methyl (E)-8-(3-methoxyresorcin-1-O-yl)-9-octadecenoate [17] displayed good emulsifying properties, which were examined by ultraviolet (UV)-turbidity tests. The hydrophilic-lipophilic balance (HLB)-value for [17] was calculated to be 1.31, indicating that [17] would be a stabilizer for water-in-oil emulsions (W/O-emulsions) because the HLB-value is <7 (39). When [17] was used as an emulsifier for a watermineral oil mixture, 50% of the turbidity of the initial emulsion remained after 24 h.

Outlook

By way of allylic oxidation with selenium dioxide, with photogenerated singlet oxygen and palladium(0)-catalyzed substitution, a route for the introduction of carbon, oxygen, and nitrogen nucleophiles into oleic and 10-undecenoic acid was developed. This allows the preparation of fatty acid–nucleophile conjugates. First investigations of these compounds showed that they share some interesting properties with surfactants. The nucleophiles can now be selected to perform the following: i modify the amphiphilic properties (cmc, interfacial tension) and the solubility of the fatty acid i possibly transfer to the derivative of bactericidal, fungicidal, herbicidal, or pharmacological properties of the nucleophile that may enhance synergistically the properties of the active compound i possibly transfer an antioxidative activity (e.g., of phenols) to the derivative, thus leading to fatty acid–bound antioxidants Copyright © 1999 by AOCS Press

78 i

H.J. Schäfer and M. Zobel

introduce carboxylic groups, hydroxy- and aminoalkyl groups into the middle of the fatty acid alkyl chain, thus affording building blocks for polyesters and polyamides.

Acknowledgments

We are grateful to the Minister of Agriculture and Forestry for financial support and the companies BASF, Ludwigshafen, Bayer, Monheim, and Henkel, Düsseldorf for the testing of compounds. References

1. Rüsch, M. gen Klaas, and S. Warwel, J. Am. Oil Chem. Soc. 73: 1453–1457 (1996). 2. Baumann, H., M. Bühler, H. Fochem, F. Hirsinger, H. Zoebelein, and J. Falbe, Angew. Chem. 100: 42–64 (1988), Angew. Chem. Int. Ed. Engl. 27: 41 (1988). 3. Metzger, J.O., and U. Biermann, Liebigs Ann. Chem. 645–650 (1993). 4. Lucas, T., and H.J. Schäfer, Fat Sci. Technol. 93: 90–96 (1991). 5. Metzger, J. O., R. Mahler, and A. Schmidt, Liebigs Ann. 693–696 (1996). 6. Sastry, G.S.R., B.G.K. Murphy, and J.S. Aggarwal, J. Am. Oil Chem. Soc. 49: 90–96 (1991). 7. Kahmen aus dem, M., and H.J. Schäfer, Fett Lipid 100: 227–235 (1998). 8. Warwel, S., Nachr. Chem. Tech. Labor 40: 314–320 (1992). 9. Bergmann, E. D., D. Ginsburg, and R. Pappo, Org. React. 10: 179–555 (1959). 10. Clark, J. H., Chem. Rev. 80: 429 (1980). 11. Godleski, S. A., in Comprehensive Organic Synthesis, edited by B. M. Trost, and I. Fleming, Pergamon Press, Oxford, 1991, vol. 4, pp. 585–661. 12. Hegedus, L.S., in Organometallics in Synthesis, edited by M. Schlosser, J. Wiley & Sons, Chichester, 1994, pp. 383–459. 13. Tsuji, J., Tetrahedron 42: 4361–4368 (1986). 14. Yanigama, Y., K. Nishimura, A. Kawasaki, and S. Murahashi, Tetrahedron Lett. 23: 5549–5553 (1982). 15. Atkins, K.E., W.E. Walker, and R.M. Manyik, Tetrahedron Lett. 3821–3829 (1970). 16. Tsuji, J., I. Shimizu, I. Minami, Y. Ohashi, T. Sugima, and K. Takahashi, J. Org. Chem. 50 :1523–1529 (1985). 17. Umbreit, M.A., and K.B. Sharpless, J. Am. Chem. Soc. 99: 5526–5528 (1977). 18. Rao, A.V.R., E.R. Reddy, A.V. Purandare, and C.V.N.S. Varaprasad, Tetrahedron 43: 4385–4394 (1987). 19. Rao, A.V.R., S.V. Mysorekar, and J.S. Yadar, Synth. Commun.17: 1339–1347 (1987). 20. Larock, R.C., Comprehensive Organic Transformations, VCH Publications, New York, 1989, pp. 116, 486. 21. Masamune, S., J. Am. Chem. Soc. 86: 290–292 (1964). 22. Schenck, G.O., and K. Ziegler, Naturwissenschaften 32: 157–166 (1944). 23. Morimoto, T., T. Machida, M. Hirano, and X. Zhung, J. Chem. Soc. Perkin Trans. 2, 909–914 (1988). 24. Schäfer, H.J., M. aus dem Kahmen, L. Hinkamp, and R. Maletz, in 3. Symposium Nachwachsende Rohstoffe, Landwirtschaftsverlag, 1994, pp. 217–234. 25. Plate, M., and H.J. Schäfer, in 5. Symposium Nachwachsende Rohstoffe, Köllen Druck Verlag GmbH, 1997, pp. 195–198.

Copyright © 1999 by AOCS Press

Palladium(0)-Catalyzed Reactions of Nucleophiles

79

26. König, K.H., Chem. Unserer Zeit 5: 217–226 (1990). 27. König, K.H., Chem. Unserer Zeit 6: 292–302 (1990). 28. Naumann, K., B. Fugmann, F. Lieb, H. Moeschler, and U. Wachendorff, Chem. Unserer Zeit 6: 317–330 (1991). 29 Naumann, K., B. Fugmann, F. Lieb, H. Moeschler, and U. Wachendorff, Chem. Unserer Zeit 1 7:35–41 (1992). 30. Haase, K.D., A.J.Heynen, and N.L.M. Laane, Fat Sci. Technol. 91: 350–353 (1989). 31. Mitsunobu, O., M. Yamada, and T. Mukaiyama, Bull. Chem. Soc. Jpn. 40: 935–939 (1967). 32. Hughes, D.L., Org. React. 42: 335–656 (1992). 33. Feldmann, G., Master’s Thesis, Universität Münster, Münster, 1998. 34. Schmid, K., Perspektiven Nachwachsender Rohstoffe in der Chemie, VCH, Weinheim, 1996, pp. 41–60. 35. Römpp, C.D., Chemie Lexikon-Version 1.0, Georg Thieme Verlag, Stuttgart, 1995. 36. Traube, I., Ann. 265: 27 (1891). 37. Klevens, H.B., J. Am. Oil Chem. Soc. 30: 74–79 (1953). 38. Evans, H.C., J. Chem. Soc. 579–586 (1956). 39. Sowada, R., and J.C. McGowan, Tenside Surf. Det. 29: 109–113 (1992).

Copyright © 1999 by AOCS Press

Chapter 5

Synthesis of New Oleochemicals: Products of FriedelCrafts Reactions of Unsaturated Fatty Compounds Ursula Biermann and Jürgen O. Metzger

Department of Chemistry, University of Oldenburg, D-26111 Oldenburg, Germany

Introduction

Unsaturated fatty compounds are of interest as renewable raw materials (1). These compounds can be functionalized at the C,C-double bond by electrophilic addition reactions to give new oleochemicals with potentially new and interesting properties. The alkylaluminum chloride–induced Friedel-Crafts acylation of unsaturated fatty compounds (Fig. 1), such as oleic acid [1a], 10-undecenoic acid [2a], petroselinic acid [3a], and erucic acid [4a], and the respective esters and alcohols yield the corresponding β,γ-unsaturated ketones (2,3). In this paper, we give some examples of ethylaluminum dichloride (EtAlCl2)induced acylations with different acylating agents such as saturated acyl chlorides, dicarboxylic acid dichlorides, cyclic anhydrides, unsaturated acyl chlorides, and aromatic and heteroaromatic acyl chlorides.

Fig. 1. Unsaturated fatty compounds: oleic acid [1a], 10-undecenoic acid [2a], petroselinic acid [3a], erucic acid [4a], and the respective esters [1b]–[4b] and alcohols [1c]–[4c] 80 Copyright © 1999 by AOCS Press

Results

Products of Friedel-Crafts Reactions

Acylations with Saturated Acyl Chlorides

81

The reaction of 10-undecenoic acid [2a], for example, with an acyl chloride such as acetyl chloride, heptanoyl chloride, hexadecanoyl chloride, and EtAlCl2 in a ratio of 1:1:2 in dichloromethane gave the corresponding β,γ-unsaturated keto carboxylic acid [5a]–[5c] (Scheme 1) after a reaction time of 2 h at room temperature with high regioselectivity. The products were obtained as a mixture of (E)/(Z)-stereoisomers ([(E)]:[(Z)] = 3:1) in isolated yields of 50–67%. Catalytic hydrogenation of the unsaturated ketocarboxylic acids [5a]–[5c] gave the saturated products in quantitative yields (2). The Lewis acid–induced acylation of unsaturated fatty compounds combined with the following reactions allows the synthesis of natural products derived from fats. Acylation of 10-undecenoic acid [2a] with heptanoyl chloride gave 12-oxo-9octadecenoic acid; reduction with NaBH4 afforded the racemate of ricinelaidic acid [6], a natural product (Fig. 2). The acylation of oleic acid [1a] with cyclopropanecarboxylic acid chloride gave in the presence of EtAlCl2 the corresponding branched β,γ-unsaturated ketocarboxylic acid [7] (Scheme 2). Product [7] was obtained as a mixture of 9- and 10-regioisomers in a ratio of 1:1 as pure (E)-adduct. The corresponding reaction of 10-undecenoic acid [2a] afforded the cyclopropyl allyl ketone [8] (Fig. 3). The solid product was obtained as a mixture of (E)/(Z)stereoisomers in a ratio of 2.3:1 and in an isolated yield of 71%.

SCHEME 1. Ethylaluminum dichloride–induced Friedel-Crafts acylation of 10- undecenoic acid [2a] with acyl chlorides.

Copyright © 1999 by AOCS Press

82

U. Biermann and J.O. Metzger

Fig. 2. Racemate of ricinelaidic acid [6] obtained by acylation of 10-undecenoic acid [2a] with heptanoyl chloride followed by reduction of the carbonyl group.

Diacylation products should be obtained by acylation reactions with dicarboxylic acid dichlorides. We carried out the reaction of methyl 10-undecenoate [2b] with adipic acid dichloride and EtAlCl2 in a ratio of 2:1:4 (Scheme 3). After a reaction time of 2 h at room temperature, the product was crystallized from petroleum ether/ether (9:1) and identified as the methyl diketocarboxylate [9]. The acylation took place at only one acyl chloride functionality of the dicarboxylic acid. The second acyl chloride functionality reacted with one equivalent of EtAlCl2in a Grignard analogous reaction to give the ethyl ketone. Acylations with Unsaturated Acyl Chlorides

The double unsaturated ketocarboxylic acid [10] was obtained by acylation of oleic acid [1a] with crotonic acid chloride (Scheme 4). Product [10], an allyl vinyl ketone, was formed as a mixture of the 9- and 10-regioisomers (3).

SCHEME 2. Ethylaluminum dichloride–induced acylation of oleic acid [1a] with cyclopropanecarboxylic acid chloride.

Copyright © 1999 by AOCS Press

Products of Friedel-Crafts Reactions

83

Fig. 3. Acylation product [8] from the reaction of 10undecenoic acid [2a] and cyclopropanecarboxylic acid chloride.

The acylation of 10-undecenoic acid [2a] with acrylic acid chloride, carried out under the same reaction conditions used for the acylation of [1a] with crotonic acid chloride, gave the corresponding allyl vinyl ketone [11] as a mixture of the (E)- and (Z)-stereoisomers ([(E)]:[(Z)] = 3:1, Fig. 4). Because of their allyl vinyl keto functionality, [10] and [11] should be suitable substrates for the Nazarov reaction. The Nazarov cyclization of acylation product [10] was carried out in two different ways. The first method we applied was the typical procedure used for Nazarov reactions (Scheme 5). The regioisomers mixture of [10] was heated for 3 h in a mixture of phosphoric acid and formic acid, yielding the expected cyclopentenone deriv-

SCHEME 3. Ethylaluminum dichloride–induced acylation of methyl 10-undecenoate [2b] with adipic acid dichloride.

Copyright © 1999 by AOCS Press

84

U. Biermann and J.O. Metzger

SCHEME 4. Ethylaluminum dichloride–induced acylation of oleic acid [1a] with crotonic acid chloride.

Fig. 4. Acylation product [11] from the reaction of 10-undecenoic acid [2a] and acrylic acid chloride.

SCHEME 5. Nazarov cyclization of acylation product [10] obtained from the reaction of oleic acid [1a] and crotonic acid chloride

Copyright © 1999 by AOCS Press

Products of Friedel-Crafts Reactions

85

SCHEME 6. Ethylaluminum dichloride–induced acylation of 10-undecenoic acid [2a] with benzoyl chloride.

atives [12]. By the second method, the allyl vinyl ketone was heated for 4 h in chloroform in the presence of montmorillonite K10. Both methods gave the cyclopentenone [12] as a regioisomeric mixture. Catalytic hydrogenation afforded the corresponding cyclopentanone derivatives [13]. Acylations with Aromatic and Heteroaromatic Acyl Chlorides

Further acylations of unsaturated fatty compounds were carried out with aromatic and heteroaromatic carboxylic acid chlorides such as benzoyl chloride and thiophene2-carboxylic acid chloride. The benzoylation of 10-undecenoic acid [2a], induced by EtAlCl2, was already complete after a reaction time of 30 min (Scheme 6). Product [14], a phenyl allyl ketone, was obtained as a mixture of (E)/(Z)-stereoisomers in an isolated yield of 49%. The reaction occurred regioselectively at C-11 of the molecule chain. The benzoylation of oleic acid [1a] yielded the corresponding branched (E)-configured phenyl allyl ketone [15] as a mixture of the 9- and 10-regioisomers in a ratio of nearly 1:1 (Fig. 5).

Fig. 5. Acylation product [15] from the reaction of oleic acid [1a] and benzoyl chloride.

Copyright © 1999 by AOCS Press

86

U. Biermann and J.O. Metzger

It should be of interest to cyclize compounds [14] and [15] to give indanone derivatives. However, the treatment of [14] and [15] with a mixture of phosphoric acid and formic acid under reaction conditions that we used for Nazarov reactions gave no cyclization products. After a reaction time of 30 min, the acylation of methyl oleate [1b] with thiophene-2-carboxylic acid chloride afforded the corresponding 2-thienyl allyl ketone [16] in an isolated yield of 64% (Scheme 7). Product [16], a 1:1 mixture of regioisomers, was obtained as pure (E)-adduct. The corresponding acylation of methyl 10-undecenoate [2b] gave the 2-thienyl allyl ketone [17] in a 59% yield as a mixture of the (E)- and (Z)-stereoisomers ([(E)17]:[(Z)-17] = 1.8:1, Fig. 6). The properties of these new oleochemicals must be examined because thiophene derivatives are of importance in areas as diverse as Pharmaceuticals, veterinary drugs, agrochemicals, and polymers. Acylations with Cyclic Anhydrides

Cyclic anhydrides are known to be useful acylating agents (2). On reaction of 10-undecenol [2c] with glutaric anhydride, the ω-hydroxy carboxylic acid [18] with an additional β,γ-unsaturated keto functionality could be synthesized (Scheme 8). After recrystallization, the product was obtained as pure (E)-isomer. The EtAlCl2-induced acylation of oleic acid [1a] with succinic anhydride afforded the corresponding β,γ-unsaturated keto dicarboxylic acid [19] as a 1:1 mixture of regioisomers with an (E)-configured double bond (Fig. 7).

SCHEME 7. Ethylaluminum dichloride–induced acylation of methyl oleate [1b] with thiophene-2-carboxylic acid chloride.

Copyright © 1999 by AOCS Press

Products of Friedel-Crafts Reactions

87

Fig. 6. Acylation product [17] from the reaction of methyl 10-undecenoate [2b] and thiophene-2carboxylic acid chloride.

Intramolecular Acylations

Intramolecular acylations are also possible with unsaturated fatty compounds (2). The intramolecular reaction of petroselinic acid chloride and EtAlCl2 gave the cyclic product [20] with an exocyclic double bond (Scheme 9). The ring closure took place regioselectively at C-6. The pure (E)-adduct was isolated in a yield of 58%. The unsaturated ketone [20] could be easily hydrogenated to give 2-dodecylcyclohexanone [21].

Conclusion

The alkylaluminum halide–induced Friedel-Crafts acylation is a very general and synthetically useful reaction that allows the functionalization of unsaturated fatty compounds. Acylations were carried out with different acylating agents such as acyl chlorides, dicarboxylic acid dichlorides, cyclic anhydrides, unsaturated acyl chlorides, and aromatic and heteroaromatic carboxylic acid chlorides, yielding a large

SCHEME 8. Ethylaluminum dichloride–induced acylation of 10undecenol [2c] with glutaric anhydride.

Copyright © 1999 by AOCS Press

88

U. Biermann and J.O. Metzger

Fig. 7. Acylation product [19] from the reaction of oleic acid [1a] and succinic anhydride.

variety of new and highly functionalized oleochemicals containing an allylketo functionality.

Experimental

Experimental instructions are given for the synthesis of products [7], [8], [9], [14], [15], [16], and [17]. The synthesis of all other products reported in this paper is described in the literature (2,3). Oleic acid (new sunflower, 82% oleic acid, 3.5 % palmitic acid, 0.6% stearic acid, 12% C18:2), and methyl oleate (new sunflower, 82.8% methyl oleate, 3.6% methyl stearate, 3.5% methyl palmitate, 8.4% C18:2) were obtained from Henkel KGaA (Düsseldorf, Germany). The amounts of the starting olefins used in the reactions were calculated on the basis of 100% purity. 10-Undecenoic acid was obtained from Atochem (Paris La Defense, France). Cyclopropanecarboxylic acid chloride, benzoyl chloride, and thiophene-2-carboxylic acid chloride were obtained from Aldrich (Deisenhofen, Germany) and used without further purification. EtAlCl2 was from Witco GmbH (Bergkamen, Germany). All acylation reactions were run under N2.

SCHEME 9. Intramolecular EtAlCl2-induced acylation of petroselinic acid chloride.

Copyright © 1999 by AOCS Press

Products of Friedel-Crafts Reactions

89

The general procedure for the synthesis of [7], [8], [14], [15], [16], and [17] was as follows: a mixture of the alkene (1 equivalent) and the acylating agent (1 equivalent) in CH2Cl2 (10 mL) was stirred magnetically under nitrogen (1 bar) for 5 min at −15°C. After dropwise addition of EtAlCl2 (1 mol/L in hexane, 2 equivalents), the sample was stirred for an additional 45 min at room temperature. The reaction was quenched by the addition of Et2O (100 mL) and H2O (40 mL); 10% HCl was added until the precipitated aluminum salts had dissolved. The organic layer was separated and washed with H2O (3 × 30 mL). The organic layer was dried (Na2SO4) and the solvent evaporated. In the case of the synthesis of [9], methyl 10-undecenoate [2b], adipic acid dichloride, and EtAlCl2 were used in a ratio of 2:1:4. The reaction procedure was carried out as described above. For characterization products, [7], [8], [14], and [15] were purified by column chromatography [28 cm × 2 cm, silica gel 60 (Merck), 70–230 mesh] with the eluent petroleum ether/EtOAc (9:1, 100 mL and 7:3, 250 mL). Fractions containing the acylation product were collected, the solvent was evaporated, and the residue dried at 20°C/0.01 mbar. The acylation products [9], [16], and [17] were purified by “Kugelrohr” distillation (5 × 10−³ bar, 250°C for [9], 170°C for [16] and 200°C for [17]). All products were characterized unambiguously by their ¹H- and 13C-NMR data and by mass spectrometry. Acknowledgments

The authors thank the Bundesministerium für Ernährung, Landwirtschaft und Forsten for financial support of this work (Förderkennzeichen 97 NR 174). Furthermore, we thank Henkel KGaA, Witco GmbH, and Atochem for providing chemicals. We thank Rosemarie Raphael for typing the manuscript. References

1. Baumann, H., M. Bühler, H. Fochem, T. Hirsinger, H. Zoebelein, and J. Falbe, Natürliche Fette und Öle—Nachwachsende Rohstoffe für die Chemische Industrie, Angew. Chem. 100: 42–64 (1988); Angew. Chem. Int. Ed. Engl. 27: 41–62 (1988). 2. Metzger, J.O., and U. Biermann, Alkylaluminium Dichloride Induced Friedel-Crafts Acylations of Unsaturated Carboxylic Acids and Alcohols, Liebigs Ann. Chem., 645–650 (1993). 3. Metzger, J.O., and U. Biermann, Lewis Acid Induced Addition to Unsaturated Fatty Compounds IV: Synthesis of Cyclopentenones from Friedel-Crafts Acylation Products of Unsaturated Fatty Compounds with α,β-Unsaturated Chlorides, Fett/Lipid 100: 2–6 (1998).

Copyright © 1999 by AOCS Press

Chapter 6

Synthesis of New Oleochemicals: Functionalization of Fatty Compounds Using Radical Reactions Jürgen O. Metzger, Ralf Mahler, Gerald Francke, and Ahlke Hayen

Department of Chemistry, University of Oldenburg, D-26111 Oldenburg, Germany

Introduction

Unsaturated fatty compounds such as oleic acid [1a], 10-undecenoic acid [2a], petroselinic acid [3a], erucic acid [4a], and the respective esters, alcohols, and native oils (Fig. 1) are alkenes and contain an electron-rich double bond that can be functionalized in many different ways by reactions with electrophilic reagents. It is therefore remarkable that >90% of oleochemical reactions have been focused on the carboxylic acid functionality and <10% have been reactions of the alkyl chain and the C,C-double bond (1). A review on radical additions to unsaturated fatty compounds that appeared in 1989 (2) quoted only very few C,C-bond–forming reactions giving branched and chain-elongated fatty compounds. Since then, modern preparative radical chemistry has been developing and has been applied also to fat chemistry (3–5). We report here on radical additions of activated haloalkanes such as alkyl 2-haloalkanoates and 2-haloalkanenitriles to unsaturated fatty compounds [1]–[4] initiated by electron transfer from copper in solvent-free systems. These additions were also car-

Fig. 1. Unsaturated fatty compounds: oleic acid [1a], 10-undecenoic acid [2a], petroselinic acid [3a], erucic acid [4a], and the respective esters [1b]–[4b] and alcohols [1c]–[4c] 90 Copyright © 1999 by AOCS Press

Functionalization Using Radical Reactions

91

ried out with fatty compounds that are halogenated in the 2-position to alkenes such as allyl alcohol, vinyl esters, and trimethylsilyl enol ethers.

Radical Additions of Alkyl 2-Haloalkanoates to Unsaturated Fatty Compounds

We reported recently on the radical additions of alkyl 2-iodoalkanoates [5] to alkenes initiated by copper metal in the absence of solvent to give γ-lactones (6,7) (Scheme 1). The reaction procedure is very simple: the alkene, iodo compound, and commercial copper powder are mixed without any pretreatment in a ratio of 1:1.3:1.3 and heated to 100–130°C in an inert atmosphere for some hours. After a simple work-up procedure, the products are obtained analytically pure and in satisfactory yields. Primary (Table 1, entry 1), secondary (entries 2–5), and tertiary (entry 6) alkyl 2iodoalkanoates [5] have been added to methyl 10-undecenoate [2b]. The alkyl 2-iodoalkanoates [5] can be formed in situ from the corresponding bromo compounds, which are more readily available, by the addition of a stoichiometric amount of sodium iodide to the reaction mixture. The products are obtained in high yields, which are similar to those obtained by direct use of the iodo compound (Table 1, entries 2–6). The products can be synthesized without problems on a multigram scale. All of the alkyl 2-haloalkanoates [5] used in this reaction gave the expected radical addition reaction. Interestingly, the reaction procedure also worked well with methyl oleate [1b], methyl petroselinate [3b], and methyl erucate [4b]. Addition of ethyl 2-iodopropionate [5b] gave the regioisomeric γ-lactones [7] and [8] in good yields and a ratio of

SCHEME 1. Additions of alkyl 2-haloalkanoates [5] to methyl 10-undecenoate [2b]. Source: Ref. 7.

Copyright © 1999 by AOCS Press

92

J.O. Metzger et al.

~1:1 (Scheme 2). It is remarkable that in these reactions, improved yields of the addition products were obtained by using ethyl 2-bromopropionate [5b] with stoichiometric amounts of sodium iodide added (Table 1, entries 6–8). We assume that the reaction is initiated by electron transfer from copper to the activated iodoalkane (Fig. 2). The electrophilic radical formed after cleavage of the halide adds to the electron-rich double bond of the alkene, and subsequent iodo abstraction yields a methyl 4-iodoalkanoate that cyclizes to a γ-lactone with elimination of iodo alkane.

SCHEME 2. Additions of ethyl 2-halopropanoate [5b] to methyl oleate [1b], methyl petroselinate [3b], and methyl erucate [4b]. Source:Ref. 7.

Copyright © 1999 by AOCS Press

Functionalization Using Radical Reactions

Fig. 2. Mechanism for the copper-initiated addition of activated iodoalkanes to alkenes (R′- 1 = alkyl 2iodoalkanoate [5], 2-iodoalkanenitrile [18]; R¹ = alkyl, (CH2)8COOMe).

93

Dicarboxylic acid derivatives were reacted as well. Additions of dimethyl 2alkyl-3-bromosuccinates [9] to methyl 10-undecenoate [2b] gave new derivatives of tetradecandioic acid with an additional γ-lactone moiety [10] (Scheme 3). Dialkyl bromomalonates, which are strongly activated, were added to alkene [2b] initiated by copper in good yields (7). An interesting example is the addition of

SCHEME 3. Additions of dimethyl 1-alkyl-2-bromo-succinates [10] to methyl 10- undecenoate [2b] in the presence of stoichiometric amounts of sodium iodide.

Copyright © 1999 by AOCS Press

94

J.O. Metzger et al.

diethyl dibromomalonate [11a] to two equivalents of methyl 10-undecenoate [2b] to give the spiro-di-γ-lactone [12] that was previously obtained also by manganese(III)acetate-initiated addition of malonic acid [11b] to fatty compound [2b] (8,9) (Scheme 4). Of particular interest is the example of diethyl 2,5-diiodoadipate [13], which was added to two equivalents of methyl 10-undecenoate [2b] to form the di-γ-lactone [14] of a tetracarboxylic acid containing 26 C-atoms in the molecule chain in a yield of 60% (Scheme 5). A remarkable rearrangement was observed in the copper-initiated reaction of 10undecenyl 2-bromohexanoate [15] (Scheme 6). The formation of the 1-iodohexadecane derivative [16] containing a γ-lactone functionality between C-10 and C-12 (yield 70%) can be explained straightforwardly by intermolecular radical addition as usual followed by lactonization. The γ-lactones obtained by this very general reaction could easily be transformed to interesting products. Scheme 7 gives two examples of the transformation of lactones [6c] and [6e] to diamide [17a] and dianilide [17b], respectively.

SCHEME 4. Additions of diethyl dibromomalonate [11a] initiated by copper powder and of malonic acid [11b] initiated by manganese(III)acetate to methyl 10-undecenoate [2b] to give spiro-di-γ-lactone [12]. Source: Ref. 7.

Copyright © 1999 by AOCS Press

Functionalization Using Radical Reactions

95

SCHEME 5. Addition of diethyl α,α′-diiodoadipate [13] to two equivalents of methyl 10-undecenoate [2b]. Source: Ref. 7.

SCHEME 6. Copper mediated rearrangement of 10-undecenyl 2-bromohexanoate [15] with stoichiometric amounts of sodium iodide added.

Copper-Initiated Additions of 2-Haloalkanenitriles to Unsaturated Fatty Compounds

2-Haloalkanenitriles add to unsaturated fatty compounds such as [2b] and [4b] in an analogous manner (6). In these cases, the iodo functionality is retained in the addition products, offering interesting possibilities for further transformations. Thus, iodoacetonitrile [18a] was added to methyl 10-undecenoate [2b] and to methyl erucate [4b]

Copyright © 1999 by AOCS Press

96

J.O. Metzger et al.

SCHEME 7. Transformation of γ-lactones [6c] and [6e] to diamide [17a] and dianilide [17b].

to give the addition products [19a] and [19b] in yields of 66 and 44%, respectively. 2-Bromohexanenitrile [18b] was added in the presence of stoichiometric amounts of sodium iodide to alkene [2b] and afforded the iodinated addition product [19c] in an isolated yield of 55% (Scheme 8).

SCHEME 8. Additions of 2-haloalkanenitriles to methyl 10-undecenoate [2b] and methyl erucate [4b]. Source: Ref. 7.

Copyright © 1999 by AOCS Press

Functionalization Using Radical Reactions

Additions of 2-Halo Fatty Compounds to Alkenes

97

The copper-initiated radical addition reaction of 2-halo fatty compounds to alkenes seems to be a well-suited method to alkylate fatty acids in the 2-position of the alkyl chain via alkyl 2- bromo- and 2-iodoalkanoates, which are readily available. One example is given in Table 1, entry 5. Methyl 2-iodopalmitate [5d] was added to methyl 10-undecenoate [2b] to give addition product [6d] in a yield of 79%. Addition of methyl 2-bromopalmitate [5d] to allyl alcohol [20] in the presence of stoichiometric amounts of sodium iodide yielded the interesting γ-lactone of an octadecan-1,2- diol [21] (Scheme 9). Additions of methyl 2-iodoalkanoates such as [5d] to vinyl acetate [22] afforded cyclic acylals such as [23] (Scheme 10). A large variety of acylals can be synthesized; these may be of interest as analogs of acetomycin, a natural compound with antibiotic and antitumor activity (10). Additions of methyl 2-bromopalmitate [5d] in the presence of sodium iodide to trimethylsilyl enol ethers [24] yielded methyl 2-alkyl-4-oxoalkanoates [25] (Scheme 11).

Conclusions

New fatty compounds have been synthesized in high yields using radical addition reactions. Alkyl 2-haloalkanoates have been added to the double bond of unsaturated fatty compounds to give γ-lactones. 2-Haloalkanenitriles have been added as well to give 4- haloalkanenitriles. 2-Halo fatty compounds, e.g., methyl 2-bromopalmitate, have been added to alkenes, allyl alcohol, vinyl esters, and trimethylsilyl enol ethers to give interesting branched and functionalized compounds. Key features of the re-

SCHEME 9. Addition of methyl 2-iodopalmitate [5d] to allyl alcohol [20].

SCHEME 10. Addition of methyl 2-iodopalmitate [5d] to vinyl acetate [22].

Copyright © 1999 by AOCS Press

98

J.O. Metzger et al.

SCHEME 11. Additions of methyl 2-bromopalmitate [5d] to trimethylsilyl enol ethers [24] (R = tBu, Ph).

action described here are as follows: no solvent is used, the work-up procedure is simple in comparison to other methods because the copper powder initiator is unproblematic for both work-up and separation, the reagents are used in essentially stoichiometric quantities, the yields are comparably high, even with 1,2-dialkylethenes, the reaction can be performed without problems on a multigram scale, and the initiator can be reused (after total oxidation of the copper, it is readily recovered by reduction). Acknowledgments

The financial support by the Bundesministerium für Ernährung, Landwirtschaft und Forsten (Förderkennzeichen 97 NR 174) is gratefully acknowledged. We thank Henkel KGaA and Elf Atochem for providing chemicals and Mrs. Rosemarie Raphael for typing the manuscript. References

1. Baumann, H., M. Bühler, H. Fochem, T. Hirsinger, H. Zoebelein, and J. Falbe, Natürliche Fette und Öle—Nachwachsende Rohstoffe für die Chemische Industrie, Angew. Chem. 100: 42–64 (1988); Angew. Chem. Int. Ed. Engl. 27: 41–62 (1988). 2. Metzger, J.O., and U. Riedner, Radikalische Additionen an ungesättigte Fettstoffe, Fat Sci. Technol. 91: 18–23 (1989). 3. Metzger, J.O., and U. Linker, New Results on Free Radical Additions to Unsaturated Fatty Compounds, Fat Sci. Technol. 93: 244–249 (1991). 4. Metzger, J.O., and U. Linker, Synthesis of Linear and Branched Perfluoroalkylated Carboxylic Acids by Radical Addition of Perfluoroalkyl Iodides to Unsaturated Fatty Acids, Liebigs Ann. Chem., 209–216 (1992). 5. Metzger, J.O., and R. Mahler, Free-Radical Cyclization of Petroselinic Acid, Liebigs Ann. Chem., 203–205 (1995). 6. Metzger, J.O., and R. Mahler, Radical Additions of Activated Haloalkanes to Alkenes Initiated by Electron Transfer from Copper in Solvent-Free Systems, Angew. Chem. 107: 1012–1015 (1995); Angew. Chem. Int. Ed. Engl. 34: 902–904 (1995). 7. Metzger, J.O., R. Mahler, and G. Francke, Radical Additions of Alkyl 2-Haloalkanoates and 2-Haloalkanenitriles to Alkenes Initiated by Electron Transfer from Copper in Solvent-Free Systems, Liebigs Ann./Recueil, 2303–2313 (1997).

Copyright © 1999 by AOCS Press

Functionalization Using Radical Reactions

99

8. Linker, U., Funktionalisierung ungesättigter Fettsäuren durch radikalische C,C-Verknüpfungsreaktionen, Ph. D. Thesis, University of Oldenburg, Oldenburg, 1991. 9. Ahmed, M.A., J. Mustafa, and S.M. Osman, Manganese(III)Acetate-Mediated One-Pot Synthesis of Some Novel Macrolides from Long-Chain Fatty Acids, J. Chem. Res. (S), 48–49 (1991). 10. Sprules, T.J., and J.-F. Lavallée, Unexpected Contrasteric Alkylation Leading to a Model for Five-Membered Ring Enolate Alkylation: Short Stereoselective Synthesis of (±)Acetomycin, J. Org. Chem. 60: 5041–5047 (1995).

Copyright © 1999 by AOCS Press

Chapter 7

Organic Chemical Synthesis and Derivatization of Fatty Acids and Methyl Esters Obtained from Ricinus and Dimorphotheca Species: Some Industrially Feasible Preparations of New Fatty Acid-Based Products

Piet M.P. Bogaert, Theodoor M. Slaghek, Herman Feil, and Patrick S.G. Tassignon Agrotechnological Research Institute (ATO-DLO), 6700 AA Wageningen, The Netherlands

Introduction

Ricinoleic Acid

Castor oil is currently the most important natural source of hydroxy fatty acids. The oil is derived from the seed of Ricinus communis, which grows as a perennial or annual in tropical and subtropical areas. Castor oil is the main source and contains ∼90% ricinoleic acid (12-hydroxy-9-Z-octadecenoic acid) [1]. This is a C18 fatty acid with a Z-double bond between C-9 and C-10 and a hydroxyl group on C-12 (Fig. 1 (1–3). Castor oil was first used for medicinal purposes, but over time, a wide

Fig. 1. Naturally occurring hydroxyl fatty acids. 100 Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

101

variety of industrial applications have developed. For example, it is commonly used as a lubricant, which is related to its long-chain triglyceride structure with up to three hydroxyl groups. These hydroxyl groups provide polarity and excellent wetting properties. The oil and esters are also used as plasticizers. Esterification of the hydroxyl group with sulfuric acid gives “turkey red oil,” which has excellent wetting and emulsifying characteristics. Sulfated oils are used in the textile industry to assist fiber wetting and as a drying agent. They also participate in the rheological control of coatings, inks, and adhesives. Dehydrated castor oil and castor fatty acids are used in coatings and provide the following characteristics: adhesion; flexibility; water, chemical, and abrasion resistance; light colors; and color retention. Important usage continues in the basic alkyl, epoxy, and acrylic film–forming polymer systems. Hydrogenated castor oil retains the hydroxyl group. Its lithium and other alkali metal soaps are outstanding thickeners, providing a gel structure important to the multipurpose greases that are used for automotive, truck, aviation, railroad, and industrial applications. Insolubility of the hydroxyl stearic soap is the key to the formation of a stable and durable gel structure. Castor oil is also often used in cosmetics as simple esters, mono- and diglycerides, glycol esters, and hydroxyl stearates. Dimorphecolic Acid

Dimorphecolic acid (Fig. 1) is an example of a hydroxydiene fatty acid, which occurs in vegetable oils. The acid contains 18 carbon atoms and a conjugated diene system localized between C-10 and C-13. The diene system has an E,E-configuration and on the C-9 of the alkyl chain, a hydroxyl group is present [9(S)-hydroxy-10,12-E,Eoctadecadienoic acid] [2] (4–9). Dimorphecolic acid was isolated first from the seed oil of Dimorphotheca sinuata. It represents 65–67% of the oil of this species and is also the predominant fatty acid in oils of other Dimorphotheca species and certain species of the related genera, Osteospermum and Castalis (10). Comparative research between Dimorphotheca pluvialis, Dimorphotheca sinuata, and Osteospermum ecklonis was performed within the framework of the National Oil Seeds Research Programme (NOP). This showed that the collection of D. pluvialis contains accessions with the highest yields. In addition, D. sinuata and Osteospermum have more agronomically disadvantageous properties than D. pluvialis (for The Netherlands). D. pluvialis is a 30–70 cm high composite with white flowers originating from South Africa and is known in Europe as an ornamental. It is a flowery plant with poor seed retention. Dimorphotheca forms two types of achenes: winged and unwinged. The plant is an annual; its life cycle in The Netherlands lasts from April to mid-August (11). D. pluvialis yields 0.6–1.7 tons of seed per hectare (11,12). The seeds have an average oil content of 20%. Agronomic bottlenecks associated with Dimorphotheca include its low yield and asynchronous Saturation in combination with seed shattering. The growth of the leaf area after emergence is slow. The plant starts flowering while total leaf surface is limited. The result is that

Copyright © 1999 by AOCS Press

102

P.M.P. Bogaert et al.

maximum photosynthesis occurs only for a short period. The genetic variation with regard to the potential seed yield is high, i.e., plant types with clearly different synchronicity of maturation have been found (8,9,11,13). The best harvesting result was obtained when Dimorphotheca was chemically desiccated, at an early stage of maturation, before harvesting (14,15). Isolation and Purification of Dimorphecolic Acid [2]

Dimorphecolic acid was first isolated as the methyl ester from D. sinuta seeds (16). Smith et al. (6) and Diamond et al. (17) reported different transesterification methods. From the mixture of methyl esters, the methyl ester [3] of [2] was isolated by countercurrent separation (18) with an acetonitrile/hexane solvent system (6). Other oxygenated fatty acids were isolated by partition between petroleum ether and methanol (19,20). Low-temperature crystallization from methanol and pentane gave [3] in good yield. Chromatographic fractionation (21) on silica gel was also used to purify [3]. This method yields pure samples but is suitable only on the scale of milligrams to grams. Distillation is an unsuitable method for purification of [3] because of its tendency to decompose upon heating. The free fatty acids (FFA) have been prepared by saponification of the oil and subsequent acidification. Dissolving the acids in acetonitrile and extracting the solution with hexane (17) could enrich [2]. Isolation of [2] or [3] on a large scale with acetonitrile as cosolvent is rather cumbersome. Therefore, investigations have been carried out by Tassignon et al. (22) towards upscaleable and economically feasible isolation methods. In the method described by Tassignon (22), D. pluvialis seed oil was hydrolyzed and transesterified, resulting in FFA and fatty acid methyl esters (FAME), respectively. Gas chromatography (GC) analysis (area percentage of silylated methyl esters) of these mixtures showed 58% dimorphecolic acid, 32% unsaturated octadecanoic acids (oleic, linolenic, and linoleic acid), 2% stearic acid, and 2% palmitic acid as well as minor compounds. In the following discussion, CFA refers to the fatty acids and CFM to the methyl esters of the CFA. Isolation of [2]

FFA were prepared by hydrolysis of the oil with a Candida cylindracea lipase. This method was preferred over saponification of the oil and acidification of the salts to minimize the formation of trienes from [2]. Subsequently, [2] was isolated by selective low-temperature crystallization from hexane containing a small amount of ethyl acetate. An overall yield of 54% with a purity of 97% was obtained. Ethyl acetate as a cosolvent decreased the required volume of solvents and distinctly improved the crystallinity. Isolation of [3]

For isolation on a larger scale, it is useful to obtain the methyl ester by a base-catalyzed transesterification of the glycerides and separation of [3] from its contaminants

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

103

by a method of countercurrent distribution (CCD). Values for the distribution coefficients of the compounds of interest make possible a straightforward development of an efficient separating system. For this reason, the fractions of [3] and CFM have been analyzed in the different phases of 1:1 (vol/vol) aqueous methanol/hexane and acetonitrile/hexane systems after one equilibration. These data are presented in Table 1. The distribution coefficients (K) of these compounds have been determined and are given in Table 2, together with the corresponding separation factors (α). The a factors for the aqueous methanol/hexane systems show an improved separation of [3] from CFM with increasing water content. From these results, it can be concluded that, in contrast with earlier studies (17), aqueous methanol/hexane yields a better separation than the acetonitrile/hexane system. The α factors for systems with 90 and 92% aqueous methanol/hexane predict successful isolation of [3] with a minimum of vessels. The distribution of the different compounds between the two phases can be described by a binomial equation (23) as follows:

where p is the fraction of total solute of a compound that is in phase S of any vessel k, q is the fraction in phase M of the same vessel and n is the number of transfers. In combination with the values, the number of transfers (n) and the minimal number of

Copyright © 1999 by AOCS Press

104

P.M.P. Bogaert et al.

vessels required for the separation of [3] from the CFM with a yield >99% can be calculated. Table 3 presents these parameters for the 90 and 92% aqueous methanol/hexane system and shows entry 5 as the most efficient and economical procedure in terms of minimum quantity of vessels and number of transfers. Figure 2 shows clearly that after four vessels, [3] is separated from CFM (area under each curve is equal to one). The large α factors predict the CCD system with aqueous methanol as mobile phase to be an interesting, economically feasible process especially for kilogram-scale production of [3] from Dimorphotheca oil. Experimental conditions were evaluated to obtain optimized procedures for the synthesis of methyl 9(R)-hydroxyoctadecanoate [3] and methyl 12-hydroxyoctadecanoate from Dimorphotheca and Ricinus seed oil. The preparations are very suitable for valorization of oils that are of inferior quality because of aging. The effectiveness of the simple CCD method was demonstrated for the isolation of [2], The hydroxyl group on the C-9 or C-12 position of the hydroxy fatty acid esters is an interesting tool for further chemical modifications of compounds [1]–[3]. In this section, a few chemical modifications and addition reaction of the hydrogenated methyl esters of the compounds [1],[2] are described. These methyl esters [4],[5] can be obtained from dimorphecolic and ricinolic oil after hydrogenation and transesterification of the compounds [1],[2]. The methyl esters [4],[5] were also pu-

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

105

Fig. 2. Calculated normalized separation of common fatty acid methyl esters (—, CFM), dimorphecolic acid methyl ester (– –) and methyl 9-oxo-10,12-E,E-octadecadienoate (– – –) after 14 transfers using a 90% aqueous methanol/hexane system and Vs/Vm = 2.

rified using the CCD method. Chemical reactions and modification of the nonhydrogenated compounds [1],[2] were also done, but these modifications will be reported in the future. In the following paragraphs a few interesting modifications and applications of the compounds [4],[5] are explained. Selective Oxidation of [4] and [5]

Reduction of [2] affords (yield >97%) 1,9(R)-octadecanediol [6] as a primary-secondary diol. A challenge is the cheap and selective oxidization of the secondary hydroxyl group of [6]. Such oxidation gives the 1-hydroxy-9-octadecanone [7] as a very interesting building block for further synthesis. For instance, with a Grignard reaction at the ketone, branched-chain molecules can be synthesized. Branched-chain fatty acids, in general, have found many applications in industry related to their unusual physical properties (24). The same selective conversion can be performed with the commercially available 1,12-octadecanediol [8] from castor oil. Oxidation of alcohols is often accomplished conveniently by or via hypochlorites; in such treatment of primary-secondary diols, a selectivity of 1:7–20 is reached for the secondary hydroxyl group. In general, sodium hypochlorite in acetic acid solution is used for this purpose (25). We obtained very good results, based on chlorine or trichloroisocyanuric acid (TCIA) in methanol in acid-buffered medium. The hydroxyketone [7] can be formed reliably with a selectivity of 1:50–65. The improved selectivity is based on the use of the lowest convenient temperature and on the Promotion of the exchange of “positive chlorine” between primary and secondary alcohol positions (including from relatively stable methyl hypochlorite).

Copyright © 1999 by AOCS Press

106

P.M.P. Bogaert et al.

Reduction of methyl 9(R)-hydroxyoctadecanoate [2] with lithium aluminum hydride in tetrahydrofuran afforded the l,9(R)-octadecanediol [6] (yield >93%). An optimized Oppenauer oxidation (24) of [6] with benzoquinone (3 mol/mol [6]) and aluminum t-butoxide (0.2 mol/mol [6]) gave 1-hydroxy-9-octadecanone [7] with a selectivity of 7:1 for the secondary alcohol function (96–99% conversion after 6–16 h at 55°C with 20% of [6] in tetrahydrofuran). Another method (26) based on TCIA (1.36 equivalents/mol [6]) and pyridine in acetone (4–5% of [6]) also gave a 7:1 proportion (99.6% conversion in 20 min at 20°C). A “positive chlorine” method, such as the method based on TCIA, looks the most attractive as a starting point because it is simple and allows further fine-tuning of the selectivity, e.g., by lowering the reaction temperature. However, the most direct way to bring “positive chlorine” in contact with alcohols is the introduction of chlorine gas into the solution. Methanol, as the simplest alcohol, can even be used as a solvent. In general, the principal reaction of Scheme 1 can be proposed. The alkyl hypochlorites are relatively unstable (27) toward decomposition to a carbonyl compound (oxidation) in the following general sequence of stability: CH3OCI > RCH2OCl>> R2CHOCl. The transfer of “positive chlorine” from methyl hypochlorite, generated with chlorine in cold methanol in the presence of some excess of sodium bicarbonate, was studied first. The formation of hydrogen chloride tends to hold the equilibrium to the left side. The presence of MeOCl can be detected by ¹H and 13C nuclear magnetic resonance (NMR) spectroscopy: ¹H δ = 3.87 ppm, 13 C δ = 69.7 ppm. At room temperature, those signals disappear slowly; after several hours, only the signals of methyl formate and methylal remain. When 2-octanol, as a model, is added to a fresh methyl hypochlorite solution (at 5–15°C), a reaction is occasionally detectable by its exothermicity. With ¹H and 13C NMR, oxidation to 2octanone is measured. This oxidation starts only when 2-octanol is added before complete decolorization of the yellow color of chlorine. If the reaction does not start, apparently all of the chlorine is converted to the relatively stable methyl hypochlorite (see Scheme 1), by which transfer to 2-octanol is prohibited. The reaction can be initiated by the addition of a few drops of concentrated hydrogen chloride. Thus it is obvious that for an effective transfer of “positive chlorine,” the methyl hypochlorite has to be activated by protonation of the hypochlorite and/or chlorine itself (Scheme 2). The decomposition of a relatively unstable alkyl hypochlorite (e.g., to a ketone) produces hydrogen chloride, promoting further reaction. However, an excess of hydrogen chloride is not permitted because the carbonyl derivatives then become enolized and further chlorinated, resulting in the formation of chloroketones. Therefore

SCHEME 1. Reversible formation of alkyl hypochlorites.

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

107

SCHEME 2. Transfer of “positive chlorine” from alcohols.

conditions must be created in which the amount of hydrogen chloride is kept at a constant level, by which protonation of hypochlorite derivatives is promoted. These protonated derivatives are consequently effective as transmitters of “positive chlorine.” In a mixture of alcohols, an equilibration of hypochlorite derivatives can be expected. The least stable hypochlorite derivative will decompose selectively to oxidized products without further chlorination (Scheme 3). In practice, we can maintain a suitable level of acidity in the following two ways: (i) in general, the most convenient way is to use a suspension of sodium bicarbonate that dissolves at a rate adjustable by its particle size and by the presence of some acetic acid; (ii) in some cases, the acidity is better controlled with a dissolved buffer based on dichloroacetic or trichloroacetic acid and their salts. Simple oxidation of 2-octanol with a reagent based on methyl hypochlorite (freshly formed from 80–85% excess of chlorine, methanol, sodium bicarbonate, and sodium acetate, 20–30 min 7
25°C) gave 99% conversion to the ketone. Oxidation at a lower temperature (bath -14°C, 40% excess of chlorine) of an equimolecular mixture of 2-octanol and 1-hexanol resulted overnight in 99% 2-octanone and a selectivity of 28:1 with respect to oxidation of 1-hexanol to methyl hexanoate (3.5%) and a trace of the dimethyl acetal. These procedures require adaptation in cases of rather poorly soluble substrates such as the alcohols and diols derived from long-chain fatty acids that are the focus of our interest. In the standard experiments, the production of hydrogen chloride and its delayed neutralization with undissolved sodium bicarbonate maintain a relatively high level of acidity. However, this acidity is not sufficient to promote unwanted chlorination of the ketone if some soluble sodium acetate and acetic acid are present. Sodium acetate instead of bicarbonate produces slow (overnight) oxidation. In the case of [6] and methyl 12-hydroxyoctadecanoate [5], the low solubility of the substrates requires an acidity of the reaction medium that is more fixed in the correct range. Thus, the use of acid buffers based on dichloro- or trichloroacetic acid is re-

Copyright © 1999 by AOCS Press

108

P.M.P. Bogaert et al.

SCHEME 3. Decomposition of alkyl hypochlorites.

quired for the diol [6], whereas the oxidation of the hydroxy ester [6] proceeds very well with a combination of dichloromethane and acetic acid as promoters of solubility and acidity, respectively (see Table 4). The solvent systems employed have the advantage of being useful at quite low temperatures. In this way, the highest selectivity can be reached. The differences in stability of alkyl hypochlorites are exploited, under equilibrating conditions, to effect preponderant oxidation of the alcohol that produces the least stable hypochlorite. Excess of chlorinating agent can suffer autodestruction or is easily quenched (bisulfite, 2-propanol). For the selective oxidation of the secondary function in primary-secondary diols, the competition with an added primary alcohol (e.g., ethanol) makes a very careful adjustment of the oxidant unnecessary. TCIA can be a rather inexpensive alternative for chlorine. A fresh solution of TCIA in methanol showed by NMR analysis that about one third of the available chlorine atoms are converted to methyl hypochlorite. Addition of 2-octanol does not result in its oxidation. When pyridine or a mixture of sodium acetate and acetic acid is subsequently added, the TCIA precipitates almost completely (anion of DCIA recovers “positive chlorine” from hypochlorite). These results show a special mechanism for the mobility of “positive chlorine” for TCIA (see Scheme 3). The problem of precipitation can be avoided by using a relatively strong acid buffer (trichloroacetic acid) and a high proportion of cosolvents (acetone, dichloromethane) that are more effective in dissolving the TCIA. Experiment 6 of Table 4 shows optimized results when chlorine was replaced with TCIA, showing that their use can be interchanged when appropriate. On a 10g scale of the oxidation of [6] with chlorine (selectivity 52:1), l-hydroxy-9octadecanone [7] was isolated in 74% yield in two crops of 99.5 and 98% purity, respectively (from ethanol). The oxidation of 1,12-octadecanediol [8] with TCIA on a 5- to 10-g scale was frequently incomplete, until acetone (20%) was used as a cosolvent to improve the solubilities.

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

109

Copyright © 1999 by AOCS Press

110

P.M.P. Bogaert et al.

In summary, it can be stated that “positive chlorine” is not easily transferred from a neutral alkyl hypochlorite to a neutral alcohol oxygen. Either the alkyl hypochlorite must be activated by protonation before it can transfer “positive chlorine” to an alcohol, or the electron density of the alcohol oxygen must be increased by (partial or complete) deprotonation by a base. The latter reaction path is suggested to proceed during the selective oxidation of an 1,2 diol by t-BuOCl in the presence of pyridine7 (unusual higher acidity of the 1,2-diol system). In principle, transfer of “positive chlorine” can also be achieved using the equilibrium between partially dechlorinated derivatives of TCIA and alkyl hypochlorites in the presence of weak bases. This unexploited possibility might be useful if acid is to be avoided. The transfer of “positive chlorine” of alkyl hypochlorites is promoted by acid. Methods are described that allow excellent oxidation of secondary alcohols such as 2-octanol and methyl 12-hydroxyoctadecanoate [5] and very selective oxidation of 1,9- and 1,12-octadecanediol with methyl hypochlorite in the presence of a suitable buffer. It is shown that TCIA can replace chlorine as the source of “positive chlorine.”

Amination of Methyl 9-Oxo-Octadecanoate and Derivatives

In a few patents, pharmaceutical companies have been showing an increased interest in methyl 12-aminooctadecanoate [9] and its C-9 isomer (methyl 9-aminooctadecanoate [10]) (Fig. 3) (28). Pharmaceutical formulations containing these fatty acid esters are useful in effecting modifications of the lipid structure of all membranes (28) and in modifying the degree of expression of cell receptors and antigen determinants, e.g., in the treatment of a variety of clinical and veterinary disorders associated with a lowering of the saturation index of cell membranes, including malignancies (28– 30). In addition, methyl 12-aminooctadecanoate can be used in the relief of pain and in bioengineering processes (28–30). The amino derivatives are not only valuable for the preparation of pharmaceutical compounds but they also have potential as monomers (31), as surface-active agents (32), and as chemical intermediates for other industrial products (32). In the past, a few methods have been reported for preparing 12-aminooctade-

Fig. 3. Chemical structure of methyl 9 [10] or methyl 12-aminooctadecanoate [9].

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

111

canoic acid [10] and its methyl ester [9] starting from 12-oxooctadecanoic acid [12]. One method was based on a reduction of the oxime of 12-oxooctadecanoic acid [12] and its methyl ester [13] with sodium in methanol (31). An economically more attractive process, however, used the reductive amination of 12- oxooctadecanoic acid [12] to its corresponding amino derivative (32–34) in the presence of an excess of ammonia. In spite of the industrial feasibility, some disadvantages could not be overcome, i.e., the use of large amounts of catalyst (17–34%), the high reaction temperature (100–150°C), and the high hydrogen pressures (1793– 13,790 MPa). For the preparation of methyl 9-aminooctadecanoate [10] also, a few methods, including the two methods mentioned above (31–34), have been reported. Reactions such as a SnCl4- induced Ritter reaction on oleic acid [14] followed by methanolysis (35), or the addition of sodium azide to methyl oleate [15] with manganese(III)-acetate, prepared in situ, and reduction of the addition products to the corresponding methyl aminooctadecanoate gave a mixture of methyl 9(10)-aminooctadecanoate [10] and [15] (10-amino) with a low yield (40–55%) (36). Even lower yields (35– 45%) were obtained when methyl oleate [15] was treated with colorless 70% nitric acid followed by a hydrogenation over Raney nickel catalyst (37). All of the abovementioned methods, however, led to a mixture of methyl-9(10)-aminooctadecanoate [10]. One single regioisomer could be obtained if methyl 9-bromooctadecanoate [16] was treated with phthalimide followed by a reduction with sodium borohydride in 2propanol (28). For the preparation of the aminoesters [9]–[10] and the aminoalcohols (12amino- octadecanol [17] and 9-aminooctadecanol [18], the keto esters [13],[19] or keto [20]–[21] were reacted with hydroxylamine. Two different reaction sequences have been followed. In the first sequence (Scheme 4, method 1), methyl 12-oxooctadecanoate [13], methyl 9- oxooctadecanoate [19], 1-hydroxy-12-octadecanone [20] or 1-hydroxy-9-octadecanone [21] was dissolved in methanol and refluxed in the presence of hydroxylamine hydrochloride whereby triethylamine was used as a hydrochloric acid acceptor. When the reaction was complete, the triethylaminehydrochloride salts were removed with water; the oximes were extracted with hexane from the reaction mixture and isolated in a pure form as determined by 13C NMR. One of the major problems in this reaction sequence is the stochiometric formation of triethylamine-hydrochloride as a side product. Experimentally, it became clear that in a one-pot reaction without separation of the triethylamine hydrochloride, the catalytic reduction of the oximes to their corresponding amines [9]–[10],[17]–[18] was inhibited. However, this problem could be solved if NH2OH could be used jnstead of the HCl-salt. Unfortunately, the commercially available NH2OH solutions in water (25% w/w) could not be used because a large amount of water prevents the oxime formation. This problem was solved by generating NH2OH in a solvent mixture of CH2Cl2 and diethylether (1:1) from hydroxylammonium chloride and triethylamine, removing the precipitated triethylammonium chlorides (Scheme 4, method 2). Industrially, the preparation of NH2OH can be achieved via a titanium silicate-1 zeolite (TS- l)-

Copyright © 1999 by AOCS Press

112

P.M.P. Bogaert et al.

SCHEME 4. Amination of functionalized keto-octadecanoate derivatives. Conditions: A: (method 1) 1.05 equivalents NH2OH•HCI, 1.05 equivalents Et3N/MeOH, reflux 2 h; B: (method 2) 1.05 equivalents NH2OH, MeOH, reflux 2 h; C: H2 (40 psi), Pd/C, room temperature.

catalyzed reaction of ammonia with hydrogen peroxide in an alcohol. TS-1 catalyzes the formation of hydroxylamine from ammonia with good selectivity (38,39). The corresponding oximes [22]–[25] were formed by the reaction of the hydroxylamine solution with the keto derivatives [13],[19]–[21] dissolved in methanol under reflux. This second reaction sequence is more favorable because the resulting oxime [22]– [25] solutions were immediately ready for the hydrogenation procedure, causing no inhibition problem as mentioned earlier. The corresponding yields of the compounds [9]–[10],[20]–[21] using method 2 are shown in Table 5. For the reduction of the oxime derivatives, several reducing agents and conditions were examined. Table 6 shows modifications of the reduction procedure designed to achieve milder reaction conditions while improving the yield and purity of the products. All given data are for the reduction of methyl 12-oximeoctadecanoate [22] to methyl 12-aminooctadecanoate [9]. For the hydrogenation of the other compounds [23]–[25], similar conditions were used as described in entry 9 of Table 6. Entries 1–4 (Table 6) represent attempts to reduce the oxime derivatives with hydrides. No methyl 12-aminooctadecanoate [20] was formed. However, many side products were detected by gas-liquid chromatographic (GLC) and thin-layer chromatographic (TLC) analyses of the crude mixtures. It has been proposed in the reduction of imines that secondary amines are formed by addition of the primary amine to the imine followed by hydrogenation of the C=N double bond and subsequent loss

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

113

of ammonia (32). No improvement was noted when Raney nickel was used as a catalyst at a hydrogen pressure between 207 and 276 MPa. The predominant formation of the oxo compound [13] is surprising. In the literature (32), it has been mentioned also that the oximation reaction could be reversible. As shown in entries 1–6 (Table 6) no methyl 12-aminooctadecanoate [9] could be obtained if NaBH4, LiAIH4, or Raney-Ni was used as reducing agent. However, with the use of palladium on activated coal, the formation of compound [9] was accomplished. As shown in entries 7 and 8 (Table 6), the combination of Pd on activated coal and acetic acid (as secondary catalyst) was not very successful. If Pd on activated coal in methanol at room temperature and a hydrogen pressure

Copyright © 1999 by AOCS Press

114

P.M.P. Bogaert et al.

of 275 MPa was used, a high yield and purity were found for methyl 12-aminooctadecanoate [9]. Extending the hydrogenation time from 24 to 48 h, an excellent yield (98.3%) was obtained at room temperature and at low hydrogenation pressure (275 MPa). The final entry was conducted at a higher temperature (50°C) and side products were identified (40). The conversion of long-chain oxo derivatives to amines has been accomplished by oximation (free NH2OH) and catalytic hydrogenation in an efficient and simplified procedure. In this procedure, the hydrogenation was accomplished at low hydrogen pressure (275 MPa) and room temperature conditions.

Synthesis of Branched Alkyl Polyglycoside

Up to now, many effective methods have been developed for the coupling reaction of primary alcohols with saccharides (e.g., the Koenigs-Knorr reaction or the trichloroacetimidate method) (41). The coupling of long-chain, branched, secondary alcohols with glycosides has offered more problems; subsequently, fewer approaches toward the development of these products have been made (42). Therefore, we started to explore the oleochemistry and the potential use of two secondary hydroxy fatty acid esters, namely, methyl 9- [4] and methyl 12- hydroxyoctadecanoate [5]. Coupling of these fatty acid derivatives to carbohydrates will result in compounds with surface-active properties. Esters of alkyl polyglycoside (APG) are already known as a new generation of nonionic surfactants with interesting technological and economical properties (43–46); they have excellent detergent and lathering properties, are easily rinsed, are not irritating, and lessen the irritation of the skin and the mucous membrane of the eye caused by other surfactants used in combination with them (47,48). In addition, they have low environmental effect as a result of their biodegradability. To accomplish the binding of the methyl hydroxy fatty acid esters [4]–[5] to acetylated sugar bromides [26a–d], two methods have been elaborated. The first is based on the Koenigs-Knorr method. In the original method, dating from 1901, and the subsequently developed, more efficient variants, the activation is achieved through the formation of glycosyl halides (bromides, chlorides); their reaction is preferentially performed in the presence of silver or mercury salts (Scheme 5; method 1) (41). For the synthesis of the compounds [27a–h], glycosyl bromides [26a–d] and mercury salts were used (Helferich method) (Scheme 5). This coupling reaction is usually promoted by addition of catalytic amounts of KI to the reaction mixture. As shown in Table 7, the coupling of the fatty acid ester to the glycosyl bromides [26a– d] was performed either at 0°C or at room temperature. In the latter case, a significantly lower yield was observed for the coupling reaction of the peracetylglycosylbromides [26a–d]. If [27a] was synthesized at room temperature, the overall yield was 15%; in contrast, if compound [27a] was synthesized at 0°C, then the overall yield was 35%. The same differences were found for the compounds [27b,e–f]. This observation may be explained by the formation of fewer side products at

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

115

SCHEME 5. Reaction of methyl 9- or methyl 12-hydroxyoctadecanoate and amino derivatives with mono- and disaccharides.

0°C compared with reactions at room temperature. The amount of side products was gravimetrically analyzed after column-chromatography. The bromine atom is more reactive at elevated temperature and can induce side reactions on the glycosyl group. A few side products were detected as methyl 9-acetoxy [28] (C-9 13C-NMR δ : 74.34 ppm) or methyl 12-acetoxy octadecanoate [29] (C-12 13C-NMR δ : 74.02 ppm), and product with a hydroxy group on the C-2⬘ position instead of an acetoxy group [30a– h]. These compounds could be formed via the following reaction pathway. The coupling reaction started, after dehalogenation of the glycosylbromide with HgBr2, with the formation of an acetoxonium ion [31], which was attacked by a secondary alcohol

Copyright © 1999 by AOCS Press

116

P.M.P. Bogaert et al.

[4],[5], resulting in the formation of an orthoacetate [32]. The side products could be formed if the orthoacetate [32] was protonated at the oxygen attached to C-2. This leads to acetate esters [28]–[29] and the 2-hydroxyglycosyl ion [33], which can be trapped by the alcohols [4] or [5] to give 2-hydroxyglycosides [30a–h] (49). During this reaction, ∼6–10% of the methyl hydroxyoctadecanoate ester [4,5] was acetylated and therefore excluded from the coupling reaction. The products [28] and [29] could be regenerated to [4,5] with Na2CO3 in methanol. There was only a 15–40% conversion. At room temperature, the mixture still contained a small amount of the starting materials. Gravimetric analysis revealed that 10–15% of the methyl esters [4,5] and 5% of the compounds [26a–d] were present. However, prolongation of the reaction time did not result in a better conversion and yield. The fact that the glycosyl acceptor was a secondary alcohol, which is less reactive than a primary alcohol, provides an explanation for that observation. The glycosidation reaction via acetylated glucopyranosyl bromide as glycosyldonor and mercury bromide as catalyst, at low temperature, led to α/β ratios of 5:95– 17:83. “Neighboring group participation” on the acetyl function on C-2 explains the major formation of the β-anomer. In spite of the development of efficient variants, severe, partly inherent disadvantages of the Koenigs-Knorr method could not be overcome. A major disadvantage includes the use of expensive and toxic heavy-metal salts. Mercury salts represent a major hazard, particularly in largescale reactions. Many attempts have been made to develop competitive methods that do not require the use of heavy-metal salts. The newly developed coupling methods such as the trichloroacetimidate, thioglycosides, and phosphate-induced method are not suitable for large-scale production (41). Therefore we have studied a new approach toward the development of a new coupling method for the synthesis of compounds [27a–b,e–f]. This method is based on the use of Et3N as a HBr acceptor instead of mercury salts and KI (Scheme 5; method 2). The catalytic activity of Et3N is based

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

117

on its ability to deprotonate the hydroxyl groups of compounds [4] and [5] and as bromine acceptor. During the reaction, no side products could be detected. However, when the reaction was executed at room temperature, a low conversion efficiency was obtained (8–10%, gravimetric analysis). It was necessary to increase the temperature to ∼81 °C, the boiling point of acetonitrile, to obtain a higher yield. The conversion efficiency for short-chain secondary alcohols at boiling temperature was very high (i-propanol, 97%; t-butanol, 91%). In spite of the fact that the catalytic activity of Et3N is lower than that of mercury salts, this method gives rise to at least the same conversion and yield as that of the Koenigs-Knorr reaction. This may be explained from the fact that Et3N induces fewer side reactions on the glycosyl group. The reaction conditions and yields are shown in Table 8 (41–51%). For the sake of completeness, it should be mentioned that a similar method could not be used for the preparation of compounds [27c–d,g–h]. The mixtures were purified by flash-chromatography using hexane and ethyl acetate as solvents. A 35:65 volumetric ratio of ethyl acetate to hexane was selected (Rf values; Table 7). The purified compounds were analyzed by NMR. With the aid of proton-proton NMR (COSY-NMR) spectra and proton-detected multiple bond coherence NMR (HMBC-NMR) spectra, the structures of the coupled compounds [27a– h] were determined. The HMBC-NMR assignments of compound [27c] (lactosyl derivative) are shown as an example in Figure 4. Previously, it was reported that the chemical shift of the proton atoms of the C12 of the fatty acid methyl ester can be found at 3.503 ppm (50). In the HMBC-NMR spectrum, there is a cross peak between this proton of C-12 and the carbon atom with a chemical shift of 95.23 ppm. This is a tertiary carbon atom. The discrimination between quaternary and tertiary 13C signals has been performed by using an attached proton test (APT) pulse sequence, one each second. The cross peak refers to the threebond coupling of the H-12 protons with the anomeric center C-1⬘ of lactose. The ¹Hdata of C-1⬘ were found with the one-bond coupling constant, which is 150 MHz. With a COSY-NMR spectrum and its corresponding HMBC spectrum, the ¹H and 13C NMR values of C-2⬘, C-3⬘, .^.^. could be determined. For the deacetylation of compounds [27a–h], sodium in methanol was used. The

Copyright © 1999 by AOCS Press

118

P.M.P. Bogaert et al.

Fig. 4. [1H,13C] Multiple-bond coherence (HMBC) nuclear magnetic resonance (NMR) spectrum of compound [27c]. The numbers in the proton and carbon spectra refer to the corresponding proton and carbon atoms of the assigned structures. Cross-peaks resulting from single- and multiple-bond coherences are indicated with two numbers, the first referring to the proton and the second to the carbon atoms involved in the coherence. For example, 12/1⬘ indicates a coherence between H-12 and C-1⬘ of structure [27c]. Cross-peaks of single bond coherences, indicated by identical proton and carbon numbers, are split by the large single bond proton-carbon couplings.

products were dissolved in a solvent mixture of methanol and dichloromethane (4:1, vol/vol) because the glycosyl part of [27a–h] is not soluble in methanol. The mixture was allowed to reflux for 15 h during which small crystals of [35a–h] were formed. The determination of the degree of deacetylation of the isolated crystalline products could be done by Fourier transform infrared spectroscopy (FTIR). The spectra of the compounds [35a–h] showed a large absorption-range band between 3600 and 3100 cm-1, corresponding to the hydroxyl groups of the carbohydrate part. The specific signals for the C=O of the acetyl groups (1751–1740 cm-1) disappeared, which proved the deacetylation to be complete. Moreover, shifts in the CH3-area (2952–2700 cm1 ) were observed. The FTIR signals and yields of compounds [35a–h] are shown in Table 9. From the experiments, it can be concluded that two methods can be used for the coupling reaction of methyl 9 [4] or methyl 12-hydroxyoctadecanoate [5] with monoand disaccharides. The first is the well-known Koenigs-Knorr reaction; the second is a newly developed more environmentally friendly method using triethylamine instead of heavy-metal salts. The combination of the compounds [4,5] and carbohydrates leads to the formation of a new class of alkyl polyglycosides. The next paragraphs concern the development of a new class of N-alkylamino-

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

119

glycosides by the coupling of methyl 9-aminooctadecanoate [10] and its C-12 [9] derivative with glucose, galactose, lactose, and maltose. To accomplish the glycosidic linkage between the methyl amino fatty acid esters [9–10] and carbohydrates, two methods have been elaborated. We employed the method described by Helferich (51,52) in which acetylated sugar bromides react with the amino fatty acid esters [9– 10] in the presence of a mercury catalyst. For the synthesis of methyl 9- [36c,d,g,h] or methyl 12-glycosyl aminooctadecanoate [36a,b,e,f], the Helferich method, which proved useful for the coupling reaction of carbohydrates to carbohydrates and for the coupling of mono- and disaccharides to methyl hydroxyoctadecanoate (51), can be followed. The reaction, whose reaction pathway is shown in Scheme 5 (method 3), starts with the peracetylation of glucose [34a], galactose [34b], lactose [34c], and maltose [34d] with the use of acetic anhydride and sodium acetate. To accomplish the binding of the methyl amino fatty acid esters [9–10] to the acetylated sugar bromides [26a–d], both compounds were dissolved in anhydrous dichloromethane in the presence of molecular sieves (4A) and mercury bromide, and small amounts of KI as catalyst (Scheme 5, method 3). The corresponding yields of the coupling products [37a–h] are mentioned in Table 10. Utilization of Et3N as a catalyst, instead of mercury bromine, (method 2) did not give any formation of the desired compounds [36a–h]. The fact that the catalyst had almost the same structure as the reacting functionality of the fatty acid esters [9–10] offers a possible explanation. For the purification of the reaction mixtures, flash-chromatography was used. The solvent was a 35:65 (vol/vol) hexane/ethyl acetate system, and the Rf values are mentioned in Table 10. The chemical structure of the compounds was identified with the aid of NMR spectroscopy. For the deacetylation of compounds [36a–h], sodium in methanol was used. The products were dissolved in a solvent mixture of methanol and dichloromethane (6:1, vol/vol) because the protected glycosyl part of [36a–h] is not soluble in methanol. The FTIR data and yields of compounds [37a–h] are shown in Table 11.

Copyright © 1999 by AOCS Press

120

P.M.P. Bogaert et al.

Physicochemical Properties

Surfactant Solubility and the Krafft Temperature

A primary driving force for the industrial development of synthetic surfactants was the problem of the precipitation of the fatty acid soaps in the presence of multivalent cations such as calcium and magnesium. Although most common surfactants have a substantial solubility in water, that characteristic can change significantly with changes in the length of the hydrophobic tail and the nature of the head group. For many ionic and nonionic materials, it was found that the overall solubility of the material in water increased as the temperature increased. That effect is the result of physical characteristics of the solid phase, such as the crystal energy and heat of hydration of the material being dissolved. It is often observed that the solubility of the material undergoes a sharp, discontinuous increase at a characteristic temperature, the so-called Krafft temperature. Below that temperature, the solubility of the surfactant

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

121

is determined by the crystal structure and energy. Above that temperature, the solubility is higher than the critical micelle concentration (cmc) (53). The Krafft temperatures of the compounds [35a–h], [37a–h] are mentioned in Table 12. Some general remarks about the influence of the different chemical functionalities on the TKrafft can be deduced as follows: i i i

i

the mean TKrafft of glucose derivatives is higher than the mean TKrafft of galactose the mean TKrafft of monosaccharides is lower than the mean TKrafft of the disaccharides great differences in Krafft temperature can be found between the aminooctadecanoate monoglycosyl derivatives (Table 12) and the aminooctadecanoate diglycosyl derivatives (Table 12); the mean TKrafft is 8 and 14.5°C, respectively, for these derivatives the lowest mean TKrafft are found for the C-9 hydroxy- or C-9 aminooctadecanoate galactosyl derivatives [35f] and [37f]

Apparently, the type of carbohydrate is a far more determining factor of the TKrafft than the nature (hydroxy- or aminooctadecanoate) and position (C-9 or C-12) of the group that links the head group and the C-18 alkyl chain. However, the TKrafft is determined mainly by the effective alkyl chain length. In general, the TKrafft increases with an increasing amount of methylene groups of the alkyl chain. For example, the TKrafft of C-12 APG to C-16 APG increases from 38 to 57°C,

Copyright © 1999 by AOCS Press

122

P.M.P. Bogaert et al.

respectively. On the basis of this information, a higher TKrafft could be expected, but the introduction of branched alkyl chains is a classic method for lowering the effective chain length and the corresponding TKrafft (54–57).

Surface Tension, Critical Micelle Concentration, Maximum Surface Excess and Molecular Area

Table 12 lists the cmc and surface tension at the cmc for the compounds [35a–h] and [37a–h]. Differences in the surface-active properties of the individual compounds can be deduced as follows:

1. The mean cmc and the γ at cmc for the monosaccharides are significantly lower than the mean cmc and the γ value of the disaccharides. The mean cmc and γ values for the monosaccharides are 1.8 × 10-3 mol/L and 34.1 mN/m, respectively, whereas for the disaccharides, the mean values of cmc and γ are 9.8 × 10-3 mol/L and 38.2 mN/m. 2. The mean cmc value for the compounds that have lactose as the common carbohydrate part is 2.5 times higher than the mean cmc of the compounds that have maltose as the common carbohydrate group. The differences between the two cmc values are much lower if the cmc value of compound [37g] is not taken into account for the determination of the mean cmc of lactose group (mean cmc is then 6.6 × 10-3 mol/L) However, only a small difference of 0.9 mN/m (2%) is found for the γ values.

The surface tension of some alkyl (poly)glucosides was investigated by Shinoda (58), Lange et al. (59), and Rosen et al. (61) as a function of the alkyl chain and the degree of polymerization (DP) by using samples differing in composition. In addition to pure surfactants for characterizing the basic dependencies, large amounts of data on surfactants of technical quality are available because of the interest in using alkyl polyglycosides on an industrial scale. A few selected values are given in Table 13. The cmc values of the pure alkyl polyglycosides and the technical alkyl polyglycosides are comparable to those of typical nonionic surfactants and decrease distinctly with increasing alkyl chain length. As can be seen in Table 13, the cmc values of linear APG varied between 2.5 × 10-3 mol/L for the β-D-C8G1 and 1.9 × 10-4 mol/L for the-β-D-C12G1 APG. As a general approximation, the cmc is lowered by a factor of 10 when the chain length is increased by two methylene units. Van Doren (55) has determined the cmc values of a homologous series of N-octyl, N-decyl, and N-dodecyl glucosamines and found the same line for the cmc as for the APG (Table 13). Although cmc values decrease with increasing alkyl chain length, the cmc values of the compounds [35a–h] and [37a–h] are slightly higher then the cmc values of the linear APG (Table 12). Normally, lower cmc values would be expected, but according to Varadaraj et al. (57) and Rosen et al. (61), these higher cmc values can be explained by the fact that branching of linear alkyl chains reduces the effective chain

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

123

length. This means that the compounds [35a–h] and [37a–h] do not have the effective chain length of a C-18 chain but that of a lower one. Comparison of the cmc values of the compounds [9a,e] with the values of the APG β-D-C8G1 and β-D-C10G1 shows that the compounds [1], [5] have an effective chain length of ∼8–10 carbon atoms (both have 1 glucose as glycosyl group). For the compounds [37a,e], an effective chain length of 11-12 carbon atoms is observed. Only small differences of the cmc can be found if methyl 12-hydroxyoctadecanoate or the C-9 derivative is excluded. The same conclusion can be made for the amino derivatives, if compound [32] is excluded. The hydrophilicity of the disaccharides can explain the higher cmc values of the disaccharide compounds in comparison with the values of the monosaccharides. The maximum difference in cmc observed for the compounds with the same alkyl chain length was well within 1.5 order of magnitude. Apparently, the alkyl chain length is far more the determining factor for the cmc value than the type of carbohydrate and the nature of the group that links the head group and the alkyl chain. The surface tension–reducing power of the compounds [35a–h], [37a–h] is lower than that of the corresponding linear compounds (Tables 12 and 13). As reported in the literature (56,57), branching in the hydrophobic group will result in a reduction in the hydrophobicity of a surfactant chain relative to that of a related straight-chain material with the same total carbon content. For example, carbon atoms located on branch sites will contribute approximately two thirds as much to the character of a surfactant molecule as one located in the main chain (56). The above-mentioned surfactants [35a–h], [37a–h] follow this general trend. In addition, it has generally been found that the presence of alkyl groups attached to the nitrogen seems to have little effect on the surface tension reduction of a surfactant. This general trend also holds for the γ values of the compounds [35a– h], [37a–h].

Copyright © 1999 by AOCS Press

124

P.M.P. Bogaert et al.

When one discusses the effectiveness of adsorption, as defined by the maximum lowering of surface tension regardless of surfactant concentration, the value of γmin is determined only by the system itself and represents a more firmly fixed point of reference. The value of γmin for a given surfactant will be determined by one of two factors, i.e., the Krafft temperature or the cmc. The surfactants used below TKrafft are unable to achieve their maximum degree of adsorption at the solution/vapor interface, and therefore the only controlling factor for the determination of the effectiveness of a surfactant will be the cmc (56).

General Conclusions

For several reasons, a growing number of industries feel very strongly about sustainable product development. It is unavoidable that industries will be confronted with problems regarding the supply of petrochemical-based materials. Another aspect is the growing criticism from society and political circles of the effects of using such raw materials. The number of problems involved with the treatment of waste originating from by-products and main products at the end of their service life is increasing. Consequently, waste treatment will tend to become more expensive. For agro-raw materials to compete with synthetic raw materials, it is a prerequiste that the level of knowledge of agro-materials be comparable to that of synthetic raw materials. This requires that lost ground will have to be made up because the chemical industry, the supplier of the synthetic raw materials, had been conducting extensive research and development for several decades. For synthetic raw materials, the basic knowledge, the applications, and the commercial options have been largely developed. Most of the vegetable oils processed today are converted to a limited number of base compounds, i.e., fatty acids, fatty methyl esters, and fatty alcohols. In a succeeding step, these intermediates may then be further processed to a large array of high-value-added final products. Saturated and unsaturated fatty acids have found a wide application in cosmetic, pharmaceutical, resin, and polymer industries. To further exploit these markets, there is a growing interest in the development of novel oil seeds and oil seed processing. In this paper, two new promising oil seeds were reported. To extend their market, the chemical modification of new seed oils by organic synthesis serves as a promising route to the preparation of functionalized fatty-acid derivatives. The oxidized fatty acid and esters, which were described earlier, can find an application in the pheromones industry; the alkyl polyglycosides can be used in the detergent, cosmetic, and paper industries. The amino derivatives are base-building blocks for the pharmaceutical and polymer industries. Therefore, continuous research is required to develop new products from and new markets for multifunctionalized fatty acids and esters. References

1. Achaya, C.R., Chemical Derivatives of Castor Oil, J. Am. Oil Chem. Soc. 48: 758–763 (1971).

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

125

2. Noughton, F.C., Production, Chemistry and Commercial Applications of Various Chemicals from Castor Oil, J. Am. Oil Chem. Soc. 51: 65–70 (1974). 3. Vingnolo, R., and F. Naughton, Castor: A New Sense of Direction, INFORM 2: 692–699 (1991). 4. Dytham, R.A., and B.C.L. Weedon, Fission of Ethylenic Acids (Varrentrapp Reaction), Tetrahedron 8: 231–238 (1960). 5. Dytham, R.A., and B.C.L. Weedon, Fission of Oxo and Hydroxy Acids, Tetrahedron 9: 246–260 (1961). 6. Smith, C.R., I.L. Wilson, E.H. Melvin, and I.A. Wolff, Dimorphecolic Acid—A Unique Hydroxydienoid Fatty Acid, J. Am. Chem. Soc. 82: 1417–1421 (1960). 7. Badami, R.C., and L.J. Morris, The Oxygenated Fatty Acid of Calendula Seed Oil, J. Am. Oil Chem. Soc. 42: 1119–1121 (1965). 8. Applewhite, T.H., R.G. Binder, and W. Gaffield, Absolute Configuration of Dimorphecolic, Lesquerolic and Densipolic Acids, J. Chem. Soc. Chem. Commun., 255–256 (1965). 9. Applewhite, T.H., R.G. Binder, and W. Gaffield, Optical Rotatory Dispersion and Absolute Configuration of Some Long-Chain Hydroxy Acids, J. Org. Chem. 32: 1173–1178 (1967). 10. Earle, F.R., K.L. Mikolajczak, I.A. Wolff, and A.S. Barclay, Seed Oils of the Calendulae, J. Am. Oil Chem. Soc. 41: 345–347 (1964). 11. Van Soest, L.J.M., Spil no. 95–96/97–98, 75 (1991). 12. Mulder, F., End Report 1990–1994 National Oil Program, Platform NOP CPRO-DLO, The Netherlands (1995). 13. Derksen, J.T.P., B.G. Muuse, and F.P. Cuperus, Designer Oil Crops, edited by D.J. Murphy, VCH Publishers, Inc., Weinheim (1994). 14. Barclay, A.S., and F.R. Earle, South African Calenduleae as a Source of Oil Seeds, Econ. Bot. 19: 33–43 (1965). 15. Breemhaar, H.G., and A. Bouman, Mechanical Harvesting and Cleaning of Calendula officinalis and Dimorphotheca pluvialis, Ind. Crops Prod. 3: 281–284 (1995). 16. Tassignon, P.S.G., Synthetic and Chemical Aspects in the Development of Dimorphotheca pluvialis as Industrial Crop, Ph.D. thesis, University of Ghent, Ghent, 1995. 17. Diamond, M.J., R.E. Knowles, R.G. Binder, and L.A. Goldblatt, Hydroxy Unsaturated Oils: II Preparation and Characterization of Methyl Dimorphecolate and Methyl Lesquerolate from Dimorphotheca and Lesquerella Oils, J. Am. Oil. Chem. Soc. 41: 430– 433 (1964). 18. Craig, L.C., and D. Craig, Technique of Organic Chemistry, edited by A. Weissberger, Interscience, New York, 1950. 19. Zilch, K.T., and H.J. Dutton, Analysis of Fat Acid Oxidation Product, Anal. Chem. 23: 775–784 (1951). 20. Bharucha, K.E., and F.D. Gunstone, Vegetable Oils (IV) Detergents, the Component Acids of Oils Containing Epoxy and (Or) Hydroxy Acids, J. Sci. Food Agric. 6: 373– 380 (1955). 21. Frankel, E.N., D.G. McConnel, and CD. Evans, Analysis of Lipids and Oxidation Products by Partition Chromatography: Hydroxy Fatty Acids and Esters, J. Am. Oil. Chem. Soc. 39: 297–301 (1962). 22. Tassignon, P.S.G., P. de Waard, T. de Rijk, H. Tournois, D. de Wit, and L. de Buyck, An Efficient Countercurrent Distribution Method for the Large-Scale Isolation of Dimorphecolic Acid Methyl Ester, Chem. Phys. Lipids 71: 187–196 (1994).

Copyright © 1999 by AOCS Press

126

P.M.P. Bogaert et al.

23. Pecsok, R.L., L.D. Shields, T. Crains, and I.G. McWilliam, Modern Methods of Chemical Analysis, Wiley, New York, 1976. 24. Kinsman, D.V., Fatty Acids in Chemistry, edited by R.W. Johnson, and E. Fritz, Marcel Dekker, Inc., New York, 1989. 25. Skarzewski, J. and R. Siedlecka, Org. Prep. Proc. Int. 24: 623 (1992). 26. Hiegel, G.A., and M. Nalbandy, The Oxidation of Secondary Alcohols to Ketones with Trichloroisocyanuric Acid, Synth. Commun. 22: 1589–1595 (1992). 27. De Buyck, L., High-Yield Preparation of 2,3,3-Trichlorinated Tetrahydropyran with an Excess of Refluxing Sulfuryl Chloride in the Dark, Bull. Soc. Chim. Belg. 101: 303–305 (1992). 28. Medeva, P, GB Patent 2237018 (1989). 29. Wood, C.B., N.A. Habib, K. Apostolov, W.R. Barker, A. Thompson, M. Hershman, and L.H. Blumgart, Eur. J. Surg. Oncol. 597: 258 (1980). 30. Apostolov, K., W. Barker, D. Catovsky, and J. Goldman, Cyclic Effects of Interferon and Its Antagonist on the Saturation of 18-Carbon Fatty Acids, Ann. Virol. 135: 245–256 (1984). 31. Colonge, J., and P. Guyot, Sur l’Acide Amino-12 Stéarique, Bull. Soc. Chim. France, 339–342 (1954). 32. Freedman, B., and G. Fuller, Reductive Amination of 12-Ketostearic Acid, J. Am. Oil Chem. Soc. 47: 311–312 (1970). 33. Floyd, D., U.S. Patent 2,610,212 (1952). 34. Handford, W, U.S. Patent 2,312,967 (1943). 35. Biermann, U., and J. Metzger, Lewis Acid Induced Additions to Unsaturated Fatty Compounds, Fat Sci. Technol. 9: 326–328 (1993). 36. Metzger, J. and U. Linker, New Results on Free Radical Addition to Unsaturated Fatty Compounds, Fat Sci. Technol. 7: 244–249 (1991). 37. Malins, D., and C. Houle, Nitration of Methyloleate with Acetylnitrate: Synthesis of Methylaminostearate, J. Am. Oil Chem. Soc. 40: 43–45 (1963). 38. Petrini, G., M. Padovan, F. Genomi, G. Leofanti, P. Poffia, A. Cesona, Eur. Patent Appl. EP 384,390 (1990). 39. Zecchina, A., G. Spoto, S. Bordiga, S. F. Geobaldo, G. Petrini, G. Leofanti, M. Padovan, M. Mantegazza, and P. Roffia, Stud. Surf. Sci. Catal. 75: 719 (1972). 40. Bogaert, P.M.P., G.G.M. van den Bosch , P.S.G. Tassignon, D. de Wit, T.M. Slaghek, and P. Van der Meeren, Low Temperature and Pressure Hydrogenation of Methyl 9-(Hydroxyimino)-octadecanoate and Derivatives to Their Corresponding Amino Derivatives, Fett/Lipid 99: 282–286 (1997). 41. Schmidt, R.R., New Methods for the Synthesis of Glycosides and Oligosaccharides— Are There Alternatives to the Koenigs-Knorr Method? Angew. Chem. Int. Ed. Engl. 25: 212–235 (1986). 42. Zhou, Q.-H., and N. Kosaric, Utilization of Canola Oil and Lactose to Produce Biosurfactants with Candida bombicola, J. Am. Oil Chem. Soc. 72: 67–71 (1995). 43. Hovelmann, P., Changes in Detergents and Personal Care Products—Chances for Surfactants, Abstract, 21st World Congress and Exhibition of the International Society for Fat Research, Den Haag, 1 (1995). 44. Garlisi, S., L. Turchini, A. Albanini, and D. Fornara, EP N. Patent 25881481 (1987). 45. Garlisi, S., D. Fornara, and P. Bernardi, EP N. Patent 92106794.2 (1992). 46. Bernardi, P., D. Fornara, and S. Garlsis, EP N. Patent 92106753.4 (1992). 47. Muizushima, H., A. Yamamuro, and Y. Yokato, EP N. Patent 91303192.8 (1991).

Copyright © 1999 by AOCS Press

Fatty Acids from Ricinus and Dimorphotheca Species

127

48. Facino, R.M., M. Carini, G. Depta, P. Bernardi, and B. Casetta, Atmospheric-Pressure Ionization Mass-Spectrometric Analysis of New Anionic Surfactants: The Alkylpolyglycoside Esters, J. Am. Oil Chem. Soc. 72: 1–9 (1995). 49. Weuthen, M., R. Kawa, K. Hilland, and A. Ansmann, Long Chain Alkyl Polyglycosides. A New Generation of Emulsifiers, Fat Sci. Technol. 97: 209–211 (1995). 50. Tassignon, P.S.G., D. de Wit, T.C. de Rijk, and L.F. De Buyck, Selective Oxidation of Primary-Secondary Diols with Methyl Hypochlorite in Acid Buffered Medium, Tetrahedron 51: 11863–11872 (1995). 51. Bogaert, P.M.P., F.K.G. Bakker, D. de Wit, T.M. Slaghek, and P. Van der Meeren, New Coupling Method Without Heavy Metals for the Synthesis of a New Class of Fatty Acid Methyl Ester Oligoglycoside Ethers, J. Surfact. Deterg. 1: 65–72 (1998). 52. Erickson, J.G., Reactions of Long Chain Amines. V: Reactions with Sugars, J. Am. Chem. Soc. 77: 2839–2843 (1955). 53. Crook, E.H., D.B. Fordyce, and G.F. Trebbi, Molecular Weight Distribution of Nonionic Surfactants. I: Surface and Interfacial Tension of Normal Distribution and Homogeneous p,t-Octylphenoxyethoxyethanols, J. Phys. Chem. 67: 1987–1994 (1963). 54. Nickel, D., T. Förster, and W. Von Rybinski, Alkyl Polyglucosides: Technology, Properties and Applications, VCH Publishers, Inc., Weinheim, 1997. 55. Van Doren, H.A., Carbohydrates as Organic Raw Materials, edited by H.V. Bekkum, H. Roper, and T. Voragen, VCH Publishers, Inc., New York, 1996. 56. Myers, D., Surfactant Science and Technology, VCH Publishers, Inc., New York, 1988 and references. 57. Varadaraj, R., J. Bock, P. Valint, S. Zushma, and R. Thomas, Fundamental Interfacial Properties of Alkyl-Branched Sulfate and Ethoxy Sulfate Surfactants Derived from Guerbet Alcohols. 1. Surface and Instantaneous Interfacial Tensions, J. Phys. Chem. 95: 1671–1679 (1991). 58. Shinoda, K., The Effect of Alcohols on the Critical Micelle Concentrations of Fatty Acid Soaps and the Critical Micelle Concentrations of Soap Mixtures, J. Phys. Chem. 58: 1136–1141 (1954). 59. Lange, E., Über Empirische und Willkürliche Grundlagen van Ma-en und Ma-systemen, Z. Phys. Chem. 204: 245–260 (1955). 60. Nickel, D., T. Förster, and W. Von Rybinski, Alkyl Polyglucosides: Technology, Properties and Applications, VCH Publishers, Inc., Weinheim, 1997. 61. Rosen, M.J., Z.H. Zhu, B. Gu, and D.S. Murphy, Relationship of Structure to Properties of Surfactants. 14. Some N-AIkyl-2-Pyrrolidones at Various Interfaces, Langmuir 4: 1273–1277 (1988).

Copyright © 1999 by AOCS Press

Chapter 8

Synthesis of Epoxidized Novel Fatty Acids for Use in Paint Applications G.J.H. Buisman, A. Overeem, and F.P. Cuperus

ATO-DLO, Bornsesteeg 59, 6700 AA, Wageningen, The Netherlands

Introduction

The effect on modern industry of an embargo on fossil oil has been known since the early 1970s. One might conclude, therefore, that today’s interest in vegetable oils originates from that time. However, in ancient times, the Egyptians already used vegetable oils to facilitate the workload during transport of heavy blocks to build their pyramids. Today, vegetable oils and animal fats find major applications in the food industry as frying oils (e.g., soybean oil and rapeseed oil) and spreads; however, these applications will not be discussed here (1). As mentioned, the first non-food applications of vegetable oil and animal fats were as lubricants. Many centuries later, it was recognized that the “thickness” of vegetable oils increased on exposure to air. This phenomenon was exploited to prepare simple varnishes and paints, which were used to make decorative drawings and paintings. At the end of the Middle Ages, it was a common art to make paints and varnishes based on drying oils. The lubricating properties of the oils were used to grind the color substances (i.e., pigments) into pastes. By grinding the pigments into finely dispersed particles, good paint characteristics (wide color pallettes, bright and nonfading colors with high glosses) could be obtained. With the invention of modern letter press techniques in the early 19th century, resins based on polymerized vegetable oils were further improved to be applicable for printing ink applications as well (2,3). In the last two centuries, an enormous impetus for industrial growth was stimulated by the chemical industry. Fossil raw materials were exploited extensively to satisfy the needs for both industrial and automotive fuels. Additionally, huge quantities of fossil fuels are used as raw materials for the synthesis of bulk chemicals. Impressive distilling plants all over the world are producing large volumes of both aromatic and aliphatic hydrocarbons; these have found widespread applications in asphalts, fuels, solvents, and all kinds of chemical intermediates (e.g., monomer feed streams). The awareness that fossil raw material supplies are finite has stimulated research to find renewable alternatives. Another driving force to search for substitutes is based on the higher biodegradability that is often reported for products based on renewable resources (e.g., vegetable oils, starch, or proteins) (4). Currently, a major drawback of these feedstocks is that the prices are not competitive with bulk chem128 Copyright © 1999 by AOCS Press

Epoxidized Fatty Acids for Use in Paint Applications

129

icals synthesized by the petrochemical industry. However, with the aid of many sponsoring programs and agricultural lobbies, several commodities (e.g., epoxidized vegetable oils or bioplastics) based on renewable materials are available today (5,6). For specialty products, market shares are small, but customer prices are often much higher and make products based on renewables competitive. Today’s research is focused on the possibilities for genetic modifications and industrial enzymatic and microbial conversions to synthesize new products (7). It has been demonstrated by many examples now that these techniques can be used efficiently to synthesize highvalue-added products in moderate-to-good yields. In this monograph, the use of epoxidized novel fatty acid derivatives in coating systems will be discussed.

Vegetable Oils in Paint Systems

Unsaturated vegetable oils have found widespread application as drying oils in alkyd paints (8). By far, the most famous coating systems are solvent-borne alkyds. Traditionally, they contain pigments, fillers, synthetic binders, polyfunctional alcohols, drying oils, organic solvents, and additives (dryers or siccatives). The formulated paints show very good paint performance (high gloss, strong films, good durability). However, a major drawback of the formulations is the large amount (often >50%) of organic solvents (i.e., white spirit) required to dissolve the binders in these paints. Since the early 1970s, considerable research has been devoted to a reduction in the amount of volatile organic solvents in paints. With traditional alkyd systems, this was not an easy task. Therefore, polymerization chemistry in water was used to develop paint systems that contained polymerized monomers (resins). Normally, these systems do not contain vegetable oils or organic solvents. The fact that water-borne acrylics cannot compete with the coating performance of alkyds on gloss, or in terms of film flexibility and repaintability, has led to some recent developments (9). Highsolid alkyds were introduced in the late 1980s. Here the resin molecules were made easily soluble in organic solvents by developing low-molecular-weight resins. As a result, only small amounts (between 10 and 30%) of organic solvents are required to dissolve the resins of high-solid alkyds. Recently, water-borne alkyds have also been successfully developed. In these paints, the alkyd resins have been chemically modified so that they are emulsifiable in water (10). Upon drying, water evaporates and the alkyd particles coagulate to form an alkyd film with almost the same characteristics as those of traditional solvent-borne alkyds. Currently under investigation are paint systems in which acrylic monomers and unsaturated vegetable oils are polymerized in one system. It is expected that the films will exhibit the advantages of both classical acrylics and alkyd paints, i.e., high gloss, flexible films, and good preservation of the substrate (11). Traditionally, in alkyd paint preparation, vegetable oils are transesterified to polyols (e.g., pentaerythritol or trimethylolpropane) at high temperature (270°C) to become part of the binder system. Upon (oxidative) drying, the double bonds in the alkyd chains are cross-linked to form the final coating network. Therefore, agricul-

Copyright © 1999 by AOCS Press

130

G.J.H. Buisman et al.

tural crops with oilseeds rich in unsaturated oils (e.g., linseed or soybean) are traditionally used in alkyd paints. In addition, castor oil and dehydrated castor oil (DCO) are widely used in paint systems because alkyds based on these oils show much reduced yellowing upon aging. Another interesting finding is the use of modified unsaturated vegetable oils as reactive diluents in high-solid paints. As an example, the process in which vegetable oils from Calendula officinalis are transesterified to methyl esters for use in paint formulations is patented by DSM resins BV (12). By following this route, the last 15–30% of organic solvents present in high solids can be replaced by methyl esters with conjugated double bonds, which become part of the polymer network upon exposure to air.

Epoxidized Vegetable Oils in Powder Coatings

Powder coatings have been widely accepted in the industrial coating market because they are completely solvent free and show excellent mechanical and physical properties (13). Usually they contain a synthetic binder, a reactive cross-linker, pigments, and several additives to improve flow and other film characteristics. Various application techniques have been developed in recent years. Powder coatings are normally sprayed electrostatically on grounded substrates and are heat-cured thereafter. Because these systems do not contain organic solvents, no emission of organic solvents takes place. However, the need for an oven to melt the resin and start the chemical reactions makes this environmentally friendly process unsuitable for do-it-yourself markets. In general, acid, hydroxy, isocyanide, or epoxy functionalized resins, with defined molecular weights, are used in powder coatings in combination with a chemical cross-link reagent. Triglycidyl isocyanurate (TGIC, Fig. 1), is widely applied as an epoxy cross-linker in powder coating formulations; however, environmental factors and health risks associated with TGIC indicate the need for a replacement of this compound (14). Cross-linkers that are based on epoxidized fatty acid derivatives seem to offer a good alternative for TGIC. In a recent patent, DSM Resins BV has claimed the use of aliphatic oxiranes based on vegetable oils as cross-linkers in commercial powder

Fig. 1. Triglycidyl isocyanurate (TGIC).

Copyright © 1999 by AOCS Press

Epoxidized Fatty Acids for Use in Paint Applications

131

coatings (15). For this purpose, unsaturated vegetable oils are epoxidized, and the highly functional epoxides are used in acid-epoxy cross-link reactions. Aliphatic oxiranes show good performance albeit with some technical drawbacks if compared with TGIC cross-linkers. Normally, aliphatic oxiranes have melting points below or close to room temperature, and clots can be formed if these liquids are added to the powders without care. Currently, special techniques are used to formulate stable powders, free from clots, that can be transported or stored at ambient temperatures. In addition, vegetable oils and epoxidized oils are known to act as plasticizers (16,17). For technical reasons, it is very important that only small amounts of liquid cross-linkers be added to guarantee good storage stability of the powder coating. In particular, when aliphatic oxiranes synthesized from triglycerides with specific fatty acid compositions are used as cross-linkers, the degree of yellowing of the coating is relatively high (18,19). It was expected that the above-mentioned drawback of aliphatic oxiranes could be reduced by using unsaturated vegetable oils with high iodine values, preferably above 150. The higher percentage of oxirane oxygen obtainable after epoxidation with specific oilseed crops justifies the addition of less of this substance. As a direct result, less yellowing and a smaller decrease in glass transition temperature (Tg) is expected. Of course, the minimal amount of aliphatic oxirane that could be added strongly depends on the amount of functional groups present in the resin. Examples of commercially available vegetable oils with high iodine values (I.V.) are linseed oil (I.V. = 170–180) and soybean oil (I.V. = 145–155). Linseed oil is a well-known source of linolenic acid (~55% of C18:3) and is used for its drying properties in oil-modified alkyd resins and alkyd emulsions. In addition, epoxidized linseed oil and soybean oil are readily available as commodities because of their current applications as weakeners in plastics. The oilseeds of Lallemantia iberica (or Iberian dragon head, Fig. 2) also show a very high content of linolenic acid (67–74%) (20). Originally, Lallemantia iberica was found in Asia (Syria, Israel, Iran, Iraq) and the Caucasus. Today, the crop also appears in Central and Southern Europe. In Germany, Austria, and Canada, its introduction was not successful probably due to the relatively wet climate conditions in these areas. The projected oilseed yields varied between 1300 and 1800 kg/ha. The oil percentage of the seeds varied between 34 and 38%, which is similar to the oil content of linseed (21). Table 1 compares the fatty

Fig. 2. Typical fatty acids of Lallemantia iberica oil.

Copyright © 1999 by AOCS Press

132

G.J.H. Buisman et al.

acid composition of Lallemantia iberica with that of commercial linseed oil. The high content of linolenic acid (67–74%) exceeds that of linseed oil (52–62%), making this highly unsaturated oil an interesting crop for industrial applications (e.g., airdrying coatings or epoxy-curing reagents). Another interesting oilseed crop that has been researched in our laboratories for several years is Euphorbia lagascae (22) (Fig. 3). This oilseed crop contains vernolic acid (12, 13-epoxy-cis-9-octadecenoic acid), with a natural epoxide content of ~65% (see Table 2). Vernolic acid also occurs to an even greater extent (72–78%) in Vernonia galamensis and other Vernonia species that are cultivated in the United States and Africa (23–25). Recently, it was found that seeds of Bernardia pulchella (Euphorbiaceae) contain the highest natural epoxy acid content yet described for a seed oil (26). This oilseed crop, grown in Brazil, contains >90% vernolic acid in the oil (25% oil content); for this reason, it is interesting to consider this wild plant for breeding experiments in the near future. An interesting application of such low-viscous epoxy-containing oils is as reactive diluents in paint formulation. Here, the oil functions as a solvent, making dispersion or solubilization of the paint formula constituents with volatile organic solvents superfluous. It has been calculated that a 10 wt% addition of vernonia oil per gallon of paint would reduce volatiles by as much as 160 million pounds per year across the United States (27). On the other hand, the oil can react with other components (i.e., resins) to form an integral part of the dried coating (28). Other advantages of using epoxy functionalized triglycerides are that both the mechanical and physical characteristics of the final polymer coatings or elastomers can be tailor-made. It is obvious

Fig. 3. Typical fatty acids of Euphorbia lagascae oil.

Copyright © 1999 by AOCS Press

Epoxidized Fatty Acids for Use in Paint Applications

133

that when epoxy functional triglycerides are used as binders, the nonfunctional fatty acids always present in minor amounts act as plasticizers in the final three-dimensional network. Thus by varying the epoxide concentration, it is possible to monitor both chemical and physical properties (e.g., Tg) of the cross-linked polymer network (29).

Chemical and Enzymatic Epoxidations

Epoxidized triglycerides of Lallemantia iberica and Euphorbia lagascae are highly functional (that is, there is a high degree of oxirane oxygen substitution); in principle, they are suitable cross-linkers for the formation of three-dimensional networks. Chemical epoxidation reactions were carried out with hydrogen peroxide (35%) and different organic acids/peroxides [performic and peracetic acid (30), m-CPBA (31), and MMPP (32)] at different temperatures. Enzymatic epoxidations were conducted with hydrogen peroxide (35%) and Candida antarctica at 30°C in the absence of organic solvents (33–35) (Scheme 1). Fewer colored epoxidation products were obtained when the epoxidation reactions were carried out following the enzymatic route, which is indicative of the very

SCHEME 1. Enzymatic epoxidation of unsaturated double bonds.

Copyright © 1999 by AOCS Press

134

G.J.H. Buisman et al.

mild enzymatic process applied. Typical examples of epoxidation products based on Lallemantia iberica and Euphorbia lagascae are given in Figures 4 and 5. In a recent patent, DSM Resins BV has claimed the use of aliphatic oxiranes based on vegetable oils as cross-linker in a commercial powder coating formulation that clearly demonstrates the applicability of these epoxidized oils in solvent-free coating formulations (15). Although the aliphatic oxiranes are liquids, stable powder formulations can be formed by making a careful selection of different modified oils. A major advantage of using these vegetable oil—based cross-linkers is that they are neither toxic nor mutagenic, in contrast with the still widely applied TGIC-based curing systems. The resulting coatings meet all standard requirements for modern powder coatings; further developments in these fields can be expected in the near future. Typical powder coating formulations are presented in Table 3; these are based on epoxidized Lallemantia iberica, Euphorbia lagascae, and linseed oil (18). It has been clearly demonstrated that epoxidized vegetable oils act as weakeners in powder coating formulations. The epoxide functionality of ~1.7 per chain in Euphorbia lagascae triglycerides requires ~17.6 g of cross-linker to react with the acid groups in Uralac. In comparison, epoxidized linseed oil and epoxidized Lallemantia iberica, with epoxide functionalities of 2.1 and 2.5 per chain, respectively, require ~11.0 and 9.8 g of epoxidized oil. As a direct result, much higher glass transition temperatures and powder stabilities have been observed with these latter two epoxides. Surprisingly, strong differences in overbake stability (yellowing) tests were found; these were attributed to differences in epoxidized fatty acid compositions of the cross-linkers. For example, it is known that diepoxystearates undergo cyclization into furan-like products at ambient temperatures (36). Possibly, di- and triepoxides synthesized from linoleic and linolenic acids, respectively, give rise during heat-cure to the formation of by-products that then contribute to coating yellowing.

Model Epoxy Cross-Linkers

To investigate the effect of different epoxidized fatty acids on the final powder coating characteristics, model compounds have been studied. If desirable, lower cross-

Fig. 4. Epoxidized triglyceride of Lallemantia iberica oil.

Copyright © 1999 by AOCS Press

Epoxidized Fatty Acids for Use in Paint Applications

135

Fig. 5. Epoxidized triglyceride of Euphorbia lagascae oil.

link densities can be obtained by using epoxidized alkyl esters of oleic, linoleic, and linolenic acid. These aliphatic oxiranes are easily synthesized in high purities (>95%) from high oleic sunflower oil, Euphorbia lagascae, and Lallemantia iberica oilseeds, respectively. Methyl 9,10-monoepoxystearate, methyl 9,10:12,13-diepoxystearate (17), and methyl 9,10:12,13:15,16-triepoxystearate were synthesized according to literature procedures in high purities starting from oilseed crops rich in unsaturated fatty acids. Methyl 9,10-monoepoxystearate was prepared from refined trisun oil as shown in Scheme 2. Trisun oil, rich in oleic fatty acids (>80%), was converted to fatty acid methyl esters (FAME) and epoxidized following standard routes (37–39). Methyl 9,10-monoepoxystearate was easily obtained in pure form via vacuum distillation. For the synthesis of methyl 9,10:12,13-diepoxystearate (27,30), Euphorbia lagascae, which is rich in vernolic acid, was used as the raw material. Following a one-step transesterification with MeOH/NaOH, pure methyl vernolate was obtained in high chemical yields by vacuum distillation. Methyl 9,10:12,13-diepoxystearate

Copyright © 1999 by AOCS Press

136

G.J.H. Buisman et al.

SCHEME 2.

was easily obtained after epoxidation of methyl vernolate with peracetic acid (Scheme 3).

The same chemical procedure was followed to synthesize methyl 12-hydroxy 9, 10-monoepoxystearate (Fig. 6) from castor oil [80% ricinoleic acid, 12-OH 18:1 (9c)] (40). Methyl ricinoleate contains a polar hydroxy group, making it easily separable from the unsaturated methyl esters present in minor amounts in the FAME mixtures. Methyl 9,10:12,13:15,16-triepoxystearate (Scheme 4) was obtained by using column-chromatography on silica gel. Relatively polar triepoxides have much longer eluting times than mono- and diepoxides, making large-scale purification rather time consuming. Other purification routes that have been successfully reported for epoxystearates are crystallizations from apolar solvents (Skellysolve B) (30). Epoxidized methyl esters used as cross-linkers in powder coating formulations show that the amounts of cross-linker (varying between 9.8 and 23.6 g) have an enormous effect on the glass transition temperature (Tg) and the powder stability of the powder coating during storage (Table 4). For example, the relatively low epoxy functionalities for mono epoxy stearate (ƒ = 1) and mono epoxy ricinoleate (ƒ = 1) make it necessary to add ~23–24 g of cross-linker to only 119 g of resin. As an immediate plasticizing effect, the glass transition temperatures varied between 21 and 24°C. Another interesting observation is that, except for mono epoxy ricinoleate, yellowing constants between 7.5 and 8.7 were found.

SCHEME 3.

Copyright © 1999 by AOCS Press

Epoxidized Fatty Acids for Use in Paint Applications

137

Fig. 6. Methyl 12-hydroxy 9,10 monoepoxystearate.

It seems, then, that epoxidized methyl esters of oleic, linoleic, and linolenic acid contribute equally to yellowing, taking into consideration the added amount of crosslinker. The observed yellowing with epoxidized methyl ricinoleate is rather poor, and the reason for that is not yet fully understood. For all of the model compounds under investigation, it can be concluded that the films show poor coating characteristics after heat-curing. The films are too brittle and do not resist the standard reversed-impact tests. It is likely that epoxide functionalities between 1 and 3 per alkyl chain are too low to form flexible three-dimensional networks.

Conclusions

Epoxidized vegetable oils and their derivatives have recently found industrial application as cross-linkers in environmentally friendly solvent-free powder coatings. In these systems, toxic and mutagenic TGIC can be replaced efficiently. However, these powder coating formulations still contain large amounts of synthetic binders to obtain the desired coating characteristics. In this paper, it was shown that functionalized vegetable oils can serve as building blocks for the preparation of (plastic) binders based on renewables exhibiting good drying properties. The mechanical and physical performance of these binders (flexibility, adhesion, drying mechanisms) can be monitored easily by a proper choice of the vegetable oils used to prepare tailormade coatings for all kinds of substrates. The creation of networks from fully or partially polymerized vegetable oils or their derivatives is a concept applicable to many coat-

SCHEME 4.

Copyright © 1999 by AOCS Press

138

G.J.H. Buisman et al.

ing, ink, or constructing systems and has opened the possibility of creating novel compositions free from organic solvents. Recently, epoxidized vegetable oils have also found application in cationic polymerization reactions. Upon exposure to UV radiation, epoxidized vegetable oils can be cured within a few seconds to form complete three-dimensional interpenetrating networks (IPN) or binders with strong mechanical properties (41,42). In addition, the plants themselves may be modified genetically to produce even further structural variations within the triglycerides and their derivatives. References

1. Introduction to Fats and Oils Technology, edited by P.J. Wan, American Oil Chemists’ Society, Champaign, IL, 1991. 2. Podhajny, R., The History of Printing, Ink World, pp. 32–36, March (1998). 3. Erhan, S.Z., M.O. Bagby, and H.W. Cunningham, Vegetable Oil-Based Printing Inks, J. Am. Oil Chem. Soc. 69: 251–256 (1992). 4. Derksen, J.T.P, F.P. Cuperus, and P. Kolster, Paints and Coatings from Renewable Resources, Ind. Crops Prod. 3: 225–236 (1995). 5. Kugler, D.E., Policy Directions for Agricultural Industrial Products in the United States, Ind. Crops and Prod. 6: 391–396 (1997). 6. Kerckow, B., C. Mangan, and L. Breslin, Industrial Crops and Products and European Union Research Policy, Ind. Crops Prod. 6: 325–331 (1997). 7. Hasan, Z.A.A., Biotransformation in the Oleochemical Industry, Palm Oil Dev. 26: 13– 19 (1997). 8. Wicks, Z.W., Jr., F.N. Jones, and P.S. Pappas, Organic Coatings: Science and Technology, Vol. 1: Film Formation, Components, and Appearance, John Wiley & Sons, Inc., 1992, pp. 133–142.

Copyright © 1999 by AOCS Press

Epoxidized Fatty Acids for Use in Paint Applications

139

9. Hurley, R., and F. Buona, J. Coating Technol. 54: 55 (1982). 10. Kumanotani, J., H. Hironori, and H. Masuda, Org. Coatings Sci. Technol. 6: 35 (1984). 11. Hamersveld, E.M.S., F.P. Cuperus, and J.T.P. Derksen, Modified and Novel Vegetable Oils in a New Generation of Emulsion Paints. Agro-Food-Industry High-Tech, pp. 23– 25, July/August (1996). 12. Zonjee, T., Resin Composition Containing an Alkyd Resin and a Reactive Diluent, European Patent Application, EP 0 685 543 A2 (1995). 13. Bodnar, E., Powder Coatings on the Road to the 21st Century. Eur. Coating J. 1–2: 44– 49 (1997). 14. Charlesworth, T., Taking Care with TGIC: Health & Safety Rules, Measurements of Exposure Levels, New Formulations, Surface World 5: 14–18 (1988). 15. DSM BV, European Patent EP 600 546 (1994). 16. Meffert, A., Technical Uses of Fatty Acid Esters, J. Am. Oil Chem. Soc. 61: 255–258 (1984). 17. Rusling, J.F., G.R. Riser, M.E. Snook, and W.E. Scott, Epoxidation of Alkyl Ester of 12, 13-Epoxyoleic Acids and Evaluation of Diepoxides as Plasticizers for Poly(Vinyl Chloride), J. Am. Oil Chem. Soc. 45: 760–763 (1968). 18. Buisman, G.J.H., 17th International Conference on Coatings, Inks and Adhesives for Plastic and Elastomers, October 1997, Milan, Italy, pp. 27–28. 19. Buisman, G.J.H., A. Overeem, and F.P. Cuperus, Biodegradable Binders and Crosslinkers from Renewable Resources, P.P.C.J., April, 14–16 (1998). 20. Hondelmann, W., and M. Dambroth, Identification and Evaluation of Oilseed-Bearing Wild Species of Forbs as Potential Crops for the Extraction of Industrial Raw Materials, Plant Res. Dev. 31: 38–49 (1990). 21. van Soest, L.J.M., M. Doorgeest, and E. Ensink, Introduction Demonstration of New Potential Crops, 1987 (In Dutch). Center for Genetic Resources, CPRO-DLO, Wageningen, The Netherlands, 1987, pp. 29–31. 22. Muuse, B.G., F.P. Cuperus, and J.T.P. Derksen, Composition and Physical Properties of Oils from New Oilseed Crops, Ind. Crops Prod. 1: 57–65 (1992). 23. Grinberg, S., V. Kolot, and D. Mills, New Chemical Derivatives Based on Vernonia galamensis Oil, Ind. Crops Prod. 3: 113–119 (1994). 24. Ayorinde, F.O., G. Osman, R.L. Shepard, and F.T. Powers, Synthesis of Azelaic Acid and Suberic Acid from Vernonia galamensis Oil, J. Am. Oil Chem. Soc. 65: 1774–1777 (1988). 25. Thompson, A.E., D.A. Dierig, and R. Kleiman, Variation in Vernonia galamensis Flowering Characteristics, Seed Oil and Vernolic Acid Contents, Ind. Crops Prod. 3: 175–183 (1994). 26. Spitzer, V., K. Aitzetmüller, and K. Vosmann, The Seed Oil of Bernardia pulchella(Euphorbiaceae)—A Rich Source of Vernolic Acid, J. Am. Oil Chem. Soc. 73: 1733–1734 (1996). 27. Carlson, K.D., and S.P. Chang, Chemical Epoxidation of a Natural Unsaturated Epoxy Seed Oil from Vernonia galamensis and a Look at Epoxy Oil Markets, J. Am. Oil Chem. Soc. 62: 934–939 (1985). 28. Tipton, M.B., J.D. Elmore, W.J. Degooyer, and J.H. Kaiser, Vernonia Oil Modification of Epoxy Resins, European Patent Specification, EP 0 555 543 (1993). 29. Barret, L.W., L.H. Sperling, and C.J. Murphy, Naturally Functionalized Triglyceride Oils in Interpenetrating Polymer Networks, J. Am. Oil Chem. Soc. 70: 523–534 (1993).

Copyright © 1999 by AOCS Press

140

G.J.H. Buisman et al.

30. Maerker, G., E.T. Haeberer, and S.F. Herb, Epoxidation of Methyl Linoleate. II. The Two Isomers of Methyl 9, 10:12,13-Diepoxystearate, J. Am. Chem. Soc, 43: 505–508 (1966). 31. Brougham, P., M.S. Cooper, D.A. Cummerson, H. Heaney, and N. Thompson, Synthesis, 1015–1016 (1987). 32. Heaney, H., Oxidation Reactions Using Magnesium Monoperphthalate and Urea Hydrogen Peroxide, Aldrichim. Acta 26: 35–45 (1993). 33. Rüsch gen. Klaas, M., and S. Warwel, Chemoenzymatic Epoxidation of Unsaturated Fatty Acid Esters and Plant Oils, J. Am. Oil Chem. Soc. 73: 1453–1457 (1996). 34. Rüsch gen. Klaas, M., and S. Warwel, New Oxidation Methods for Unsaturated Fatty Acids, presented during the 1998 AOCS Annual Meeting, Chicago, IL. 35. Kramer, G.F.H., S.Th. Bouwer, R.W. van Gemert, J.T.P. Derksen, and F.P. Cuperus, Enzymatic Peroxycarboxylic Acid Formation in a Hollow-Fibre Membrane Reactor: Kinetics and Mass Transfer, Catalysis Today 22: 537–547 (1994). 36. Debal, A., G. Rafaralahitsimba, A. Bonfand, and E. Ucciani, Catalytic Epoxidation of Methyl Linoleate—Cyclisation Products of the Epoxidized Esters, Fat Sci. Technol. 97: 269–273 (1995). 37. Debal, A., G. Rafaralahitsimba, and E. Ucciani, Epoxidation of Fatty Acid Methyl Esters with Organic Hydroperoxides and Molybdenum Oxide, Fat Sci. Technol. 95: 236–239 (1993). 38. Agarwal, R., M.H. Ansari, M.W.Y. Khan, M. Ahmad, and K.D. Sharma, Synthesis and Antimicrobial Activity of Fatty 2-Morpholinones Prepared from Epoxy Fatty Acid Methyl Esters, J. Am. Oil Chem. Soc. 66: 825–827 (1989). 39. Pages, X., C. Bonnet, and J. Laur, Synthesis of New Derivatives from High-Oleic Sunflower Methyl Esters via Epoxidation and Oxirane Opening, Presented during the 1998 AOCS Annual Meeting, Chicago, IL. 40. Berdeaux, O., W.W. Christie, F.D. Gunstone, and J.-L. Sebedio, Large-Scale Synthesis of Methyl cis-9, trans-11-Octadecadienoate from Methyl Ricinoleate, J. Am. Oil Chem. Soc. 74: 1011–1015 (1997). 41. Crivello, J.V., and R. Narayan, Epoxidized Triglycerides as Renewable Monomers in Photoinitiated Cationic Polymerization, Chem. Mater. 4: 692–699 (1992). 42. Crivello, J.V., and R. Ghoshal, UV-Curable Coatings, U.S. Patent 5,318,808 (1994).

Copyright © 1999 by AOCS Press

Chapter 9

Synthesis of New Derivatives from Vegetable Oil Methyl Esters via Epoxidation and Oxirane Opening Xavier Pagès-Xatart-Parès, Carine Bonnet, and Odile Morin ITERG, French Institute for Fats and Oils, F 33600 Pessac, France

Introduction

The preparation and characterization of epoxides were described, as early as 1861, by the chemists Berthelot, Wurtz, and Reboul (1). More recently, epoxides have received increased attention because they are of interest both as end-products and as chemical intermediates; epoxidized oils, mainly soybean oil, and their ester derivatives have thus found important applications as plasticizers and additives for polyvinyl chloride (PVC) (2). Because of the high reactivity of the oxirane group (Fig. 1), it undergoes a wide variety of ring-opening reactions with a broad range of electrophile and nucleophile agents (3). Therefore, new routes toward an interesting range of functional groups are provided. For example, cleavage of the epoxy group by reaction with carboxylic acids leads to ester alcohols, with water to diols, and with alcohols to ether alcohols. In our current research effort to find new industrial uses for vegetables oils, we are studying the cleavage of the oxirane group to develop new oleochemicals for applications in the fields of lubricants and detergents, and as new chemical intermediates.

Fig. 1. Reactivity of the oxirane group. 141 Copyright © 1999 by AOCS Press

142

Pagès-Xatart-Parès et al.

Fatty acid methyl esters are now the main intermediates in oleochemistry. Epoxidation can be considered as a transformation currently applied to triglycerides that is easy to perform on an industrial scale, compared with the production process for fatty alcohol. Therefore, why should epoxidized fatty acid methyl esters not become one of the commodities of the future? The commercial development of these compounds requires easy, environmentally friendly (e.g., avoiding catalyst use) routes of low production cost as well as identified industrial outputs. Such considerations were taken into account in this study: both rapeseed methyl esters (RME) and higholeic sunflower methyl esters (HOSME) were used as starting materials. The epoxidation of unsaturated fatty material has been studied by numerous investigators (4). Epoxidation of unsaturated oils or fatty derivatives such as methyl esters can be performed using either pre- or in situ-formed peracids; in situ epoxidation can be accomplished by using hydrogen peroxide with formic or acetic acids (5). Formic acid is preferred to acetic acid because of its high reactivity. The key points are to avoid hazards resulting from the use of hydrogen peroxide, to achieve a high conversion rate, and to decrease the transformation cost that arises from the excess of hydrogen peroxide. All these matters have been particularly studied by different authors (6,7) and companies (8–12). It is also interesting to report that new original ways of epoxidation such as perhydrolysis catalyzed by lipase (13) are now being investigated. In fact, on an industrial scale (14), in situ epoxidation is often preferred because the separate preparation and handling of peracids can be time-consuming and not without hazards because of the high oxidation potential of the hydrogen peroxide and the peracids involved. This last effect can be avoided when the peracids are consumed by the reaction immediately after their formation. The main disadvantage of the in situ epoxidation is a lower final yield because the obtained epoxides can be converted to diols, for example, by the action of the organic acid on the oxirane group. In this work, epoxidation of starting esters was made as performed on triglycerides in industry with H2O2 and formic acid, and the epoxidized products were further processed without purification. Cleavage of epoxy esters leads to a number of trifunctional compounds, depending on the nature of the reactants. Applications for these interesting products have not yet been fully investigated (3). Recently, different results have been communicated by research teams and private companies. Zaher (15), in particular, has studied the kinetics of cleavage with acetic and formic acids to elucidate the role of this acid on the epoxidation yield. To date, different fields of application have been tested. These include polyurethane foams (by opening the oxirane ring with polyols) (16–17) detergents (18–20), or products for other industrial sectors (21). The possibility of further functionalization of these ring-opened derivatives from epoxidized esters leads to a large array of possible investigations in the future.

Copyright © 1999 by AOCS Press

Methyl Esters: Epoxidation and Oxirane Opening

143

In this work, cleavages of the oxirane group of epoxidized RME and HOSME were studied according to the following routes: i acid opening by reaction with carboxylic acids (acetic acid, heptanoic acid) and fatty acids from high-oleic sunflower oil i alcoholysis by reaction with ethanol, heptanol, and octanol (acid and base catalysis) i reaction with butylamine

Experimental Procedures

Preparation of Epoxidized Esters

Materials. Commercial rapeseed methyl esters (RME) and high-oleic sunflower methyl esters (HOSME) were from industrial sources and provided by NOVANCE Cy (Compiègne, France). The main characteristics of these raw materials are reported in Tables 1 and 2. The oleic acid content of the sunflower methyl esters is ~80%, which is common for high-oleic varieties. The chemical values of these methyl esters are also classical. Epoxidation was conducted with analytical grade formic acid (99%, Merck, France) and hydrogen peroxide (50%, Prolabo, France). Method of Epoxidation. In this work, in situ epoxidation with formic acid was performed following the usual process conditions described in Figure 2. The reactions were carried out in a heatable glass reactor (capacity of 1–4 L), fitted with a reflux condenser and equipped with a stirrer and a temperature controller; the reactor was immersed in a silicone bath regulated at the desired temperature.

Copyright © 1999 by AOCS Press

144

Pagès-Xatart-Parès et al.

Methyl esters were mixed with formic acid (0.3 mol/mol of esters) and heated to 40°C. Hydrogen peroxide (3 mol/mol of esters) was then added slowly (e.g., dropwise), and the stirring rate was controlled to obtain a complete dispersion of the oil in the mixture. The temperature was maintained below 70°C. After 3 h, the reaction mixture was cooled to room temperature. After decantation, the aqueous layer was

Fig. 2. Epoxidation process of fatty acid methyl esters.

Copyright © 1999 by AOCS Press

Methyl Esters: Epoxidation and Oxirane Opening

145

drawn off and the ester layer was washed with successive sodium carbonate-saturated solutions and water. The remaining traces of water were removed at 80°C under reduced pressure. The samples were analyzed for the iodine, hydroxyl, acid, saponification, and oxirane values. Gas chromatographic (GC) analysis of epoxidized esters was also achieved.

Cleavage of Epoxidized Esters

Materials. The carboxylic acids used were analytical grade: acetic acid (99.9%, Carlo Erba, France), sulfuric acid (98%, Prolabo, France), and heptanoic acid (98%, Fluka, France). Fatty acids were produced from HOSME on a pilot scale (1 kg) by saponification and acid splitting. Fatty acid composition is reported in Table 1. The alcohols were analytical grade: ethanol (99.8%, Carlo Erba, France), heptanol (99%, Fluka, France), octanol (99%, Fluka, France). Para-toluene sulfonic acid (PTSA), zinc chloride, and methanol were also analytical grade. Butylamine (98%) was provided by Fluka.

Methods of Epoxidized Esters Opening. For cleavage of the oxirane group at atmospheric pressure, epoxidized esters were mixed with reactant [generally in mass ratio of 1:2 (esters/reactant) in a glass reactor (100 mL to 2 L)]. Catalyst (generally 1% by weight), when applicable, was added. The mixture was heated and maintained at the desired temperature during the reaction time. The reaction product was then neutralized with a sodium carbonate solution, when applicable, and washed with water. The remaining traces of water were removed at 80°C under reduced pressure. For cleavage of the oxirane group under pressure, the previous procedure was used. In this case, an autoclave was also used. Nitrogen was injected after introduction of reactants, then the reactor was closed and heated. Different ester/reactant ratios were studied. Analytical Methods

The analytical methods for the characterization of raw materials and reaction products are given in Table 3. They are mainly standardized methods and most of them are classical. Determinations of polymer content and hydroxyl values were performed to control side reactions such as dimerization and diol formation. Other methods to determine specific parameters were specially developed in ITERG’s laboratory. For the fatty acid composition of the epoxidized esters by GC, a low polar stationary phase and an alkaline transesterification with trimethylammonium hydroxide were successfully used to improve the separation of all of the components and particularly the polar epoxy esters (see below). For nitrogen derivatives obtained by reaction with butylamine, special conditions of GC were decided such as a low polar stationary phase and the silylation of the samples. For these same reaction products, classical hydroxyl and oxirane determinations are not suitable because of analytical interferences; the use of infrared spectroscopy is therefore highly advisable for analyzing the side products.

Copyright © 1999 by AOCS Press

146

Pagès-Xatart-Parès et al.

Results and Discussion Epoxidation

The epoxidized esters produced by in situ epoxidation using formic acid, under the reaction conditions described above show high oxirane values and low iodine values. In addition, they present satisfactory aspect and color. The absence of a hydroxyl group (<0.1) indicates a negligible oxirane cleavage, in other words, a very limited diol formation. The chemical characteristics of the epoxidized materials are shown in Table 4. On a pilot scale, epoxidation of both rapeseed methyl esters (RME) and higholeic sunflower methyl esters (HOSME) generally yields 85–90%. Epoxidized RME have oxirane values ranging from 4.5 to 5.2 and iodine values ranging from 5 to 1.7. Epoxidized HOSME have oxirane values ranging from 4.5 to 5 and iodine values ranging from of 1.7 to 0. The fatty acid compositions of expoxidized methyl esters determined by GC are given in Table 5.

Copyright © 1999 by AOCS Press

Methyl Esters: Epoxidation and Oxirane Opening

Cleavage of Epoxidized Esters

147

Reactions of Epoxidized Esters with Carboxylic Acids and High-Oleic Sunflower Fatty Acids (HOSFA). Reactions of epoxidized RME were carried out with acetic acid and heptanoic acid at atmospheric pressure. Operating conditions (temperature, acid quantity, and reaction time) were optimized. The best results were obtained without catalyst, with a molar ratio of esters/acid of 1:2, at 80°C for 12 h. The characteristics of the reaction products are reported in Table 6. The observed low oxirane values (0.3 with acetic acid as reactant and 0.0 with heptanoic acid) show that the oxirane group cleavage is completely achieved; the increase of the saponification values indicates ester formation. GC analysis of the final products obtained with acetic acid is reported in Table 7. The products contain 70% of a diester, the methyl acetate of hydroxystearate, and 6% of a triester, the methyl

Copyright © 1999 by AOCS Press

148

Pagès-Xatart-Parès et al.

acetate of hydroxyoleate, resulting from the cleavage of epoxidized oleic and linoleic acid methyl esters, respectively. As shown by GC analysis, cleavage with heptanoic acid leads to the formation of the diester, methyl heptanoate of hydroxystearate, representing 50% of the reaction products. GC analysis also indicates the formation of secondary products; among them, the occurrence of dehydration products is indicated by the increasing iodine value. The reaction of epoxidized HOSME with HOSFA can be performed through an easy, environmentally friendly route. We found that no catalyst or solvent was needed. The reactions were carried out under atmospheric pressure. The reaction produced mainly estolides but also small quantities of dimers and polymers as side products. High-performance liquid chromatographic (HPLC) analysis of the reaction mixture, before and after saponification followed by acid splitting, allowed determination of the estolide content because dimer and polymer fractions are not saponifiable. The analysis was performed according to the IUPAC method 2508 [Column, PLgel (Polymer Laboratories, UK) column length, 300 mm; diameter, 7 mm; particle diameter, 5µm; pore size, 100 Å; detector, differential refractor; solvent, tetrahydrofuran (THF)]. The results of a pilot trial are reported in Table 8. The best results were obtained at 175°C with an equimolar ratio of reactants, for 17 h. Under these conditions, the conversion rate (cleavage of the oxirane group) reached nearly 91%. After neutralization of the oleic acid excess with caustic soda, the total estolide content reached 100%. The high hydroxyl value (174) is due to the ring opening, resulting also in the formation of alcohol functions. The main estolide (80% of the reaction products) was a C36 diester, namely, the oleate of methylhydroxystearate (Fig. 3). It is important to note that a triester (C54 estolide) was also formed by transesterification and, as previously mentioned, that some dimers and polymers were obtained from estolides in small quantity.

Copyright © 1999 by AOCS Press

Methyl Esters: Epoxidation and Oxirane Opening

149

Reaction of Epoxidized Esters with Alcohols. Alcoholysis study of epoxidized RME with ethanol and heptanol was conducted to determine the best cleavage conditions of the oxirane group; acid catalysts (zinc chloride, PTSA) and a basic catalyst (sodium methylate) were tested. The results are reported in Tables 9 and 10. As expected, no cleavage of the oxirane group occurred when no catalyst or basic catalyst was used at temperatures <100°C. With ethanol, the best results were obtained with PTSA. The oxirane ring cleavage was complete at 90°C, after 12 h, under 1 bar. Ether formation was confirmed by different determinations, particularly the saponification value and GC analysis of the reaction products. The main component of the reaction mixture was the methyl ethoxyhydroxystearate (>50%, by GC). Secondary products were also formed during the reaction including the following: (i) fatty derivatives by dehydration, catalyzed by acid catalyst (PTSA), as shown by the increasing iodine value; (ii) products resulting from the transesterification with ethanol such as the ethyl ethoxyhydroxystearate; and (iii) dimers. With heptanol as reactant, the use of PTSA (100°C, 12 h, 1 bar) led to fatty ethers with an oxirane value of 0.0 and a saponification value of 130. The main component was the methyl heptyloxyhydroxystearate formed out of the epoxidized oleic methyl

Fig. 3. Synthesis of estolides via acidification of higholeic sunflower methyl esters with high-oleic sunflower fatty acids: main reaction product.

Copyright © 1999 by AOCS Press

150

Pagès-Xatart-Parès et al.

esters. Transesterification with heptanol also occurred and led to secondary products such as heptyl heptyloxyhydroxystearate; dimers were also probably formed in small quantity. As mentioned for reactions with ethanol and heptanol, the cleavage of the oxirane group by alcohols requires an acid catalyst. The best results were obtained with PTSA (0.5% w/w) under the following conditions: 100°C for 17 h, with a molar esters/alcohol ratio of 1:3 for a conversion rate close to 100%. The characteristics of the reaction products are reported in Table 11. As indicated by the oxirane value after reaction (0), the ring opening is complete under these conditions. Other chemical values reported, as well as GC analysis of the

Copyright © 1999 by AOCS Press

Methyl Esters: Epoxidation and Oxirane Opening

151

reaction products confirm the formation of etheralcohols and side products. The main final compound obtained (>60% of the final mixture) was the methyl octoxy-hydroxystearate (Fig. 4). Secondary products were also formed during the reaction as follows: (i) fatty derivatives obtained by dehydration (catalyzed by PTSA) as shown by the increased iodine value; (ii) products resulting from the transesterification with octanol; and (iii) dimers. The physical characteristics of the final products are reported in Table 12. For comparison, we also indicate the corresponding values for the methyl esters of high-oleic sunflower oil. It appears that the viscosity of octylic ethers is higher; the cold performances are similar, but the oxidation stability is better due to the very low unsaturation. Our industrial partner is investigating the solvent and lubricant properties of this product. Other ether alcohols of HOSME are also being tested. Cleavage with Butylamine

Oxirane ring opening of epoxidized HOSME with amines and particularly with butylamine is more difficult than with the former reactants. The difficult point is that the opening of the oxirane group competes with the transamidation reaction. Figure 5 describes the competition between these reactions. We actually observed that transamidation always joins amination. The first trials, r eaction with different catalysts under “soft conditions,” for example, at 80°C (the boiling point of butylamine), were not successful. The final optimized operating conditions must work under pressure (3 bar) at 140/200°C without catalyst, with a molar ratio of 1:36 (epoxidized HOSME/butylamine) for 3 h. The re-

Fig. 4. Synthesis of octylic ethers via alcoholysis of high-oleic sunflower methyl esters: main reaction product.

Copyright © 1999 by AOCS Press

152

Pagès-Xatart-Parès et al.

action temperature appears to be a key parameter (see below). The reaction products were then washed with distilled water and dried under reduced pressure. The compounds from routes A (1-butylamido epoxidized oleic methyl esters) and B (1-butylamido 10-hydroxy-9 butylamino oleic methyl esters) mentioned in Figure 5 are found in the final reaction products at levels depending on the operating parameters. Table 13 reports the different concentrations obtained for A and B, as a function of the reaction temperature. The characterization of the final reaction products is a delicate issue. As indicated, classical hydroxyl and oxirane determinations are not suitable because of analytical interferences. Different chromatographic methods have been developed to determine the content of nitrogen derivatives and side reaction products. From the analytical results, complete transamidation and amination can be achieved only with high temperature (180/200°C). Infrared spectroscopy clearly detected that, under the reported operating conditions, imine and polycondensation products, obtained by dehydration, are also formed during the reaction (Fig. 6). The

Fig. 5. Cleavage of epoxidized high-oleic sunflower methyl esters with butylamine: main reactions.

Copyright © 1999 by AOCS Press

Methyl Esters: Epoxidation and Oxirane Opening

153

presence of dimers formed by dimerization of nonepoxidized transaminated methyl esters is also characterized, probably in very small quantity (Fig. 7).

Conclusions

Epoxidation of both rapeseed methyl esters (RME) and high-oleic sunflower methyl esters (HOSME) was successfully performed following the classical in situ formic acid process with a high conversion rate and satisfactory characteristics. The conver-

Fig. 6.Cleavage of epoxidized high-oleic sunflower methyl esters with butylamine: side reactions from B compound.

Copyright © 1999 by AOCS Press

154

Pagès-Xatart-Parès et al.

Fig. 7. Cleavage of epoxidized high-oleic sunflower methyl esters with butylamine: side reaction from nonepoxidized transamidated esters.

sion yield was ~85–90%, and the low hydroxyl values indicated that negligible oxirane cleavage occurred, allowing the conclusion that the diol side formation is minimized. Cleavage trials of these epoxidized methyl esters with different reactants, both on a laboratory and pilot scale (up to 5 kg), demonstrate their potential interest as starting materials to produce the following under “soft” conditions: i i i

esters alcohols by reaction with carboxylic acids hydroxy esters or estolides by cleavage with fatty acids from high-oleic sunflower oil with a conversion rate of 90% ether alcohols with monoalcohols as reactants

For these last reactions, acid catalysis gives a better conversion rate than basic catalysis, which promotes transesterification and the formation of side products. The reaction of epoxidized rapeseed methyl esters with heptanol and PTSA (100°C, 12 h, 1 bar) led to fatty ethers with an oxirane value of 0.0 and a saponification number of 130; methyl heptyloxy-hydroxystearate was the main reaction product. With epoxidized high-oleic sunflower methyl esters and octanol as reactant (100°C, 17 h), the conversion rate was close to 100%, producing mainly (60%) the methyl-octoxy-hydroxystearate. The final products obtained on a pilot scale also contained side products and are currently being tested to determine their potential applications in the field of lubrication and as chemical intermediates for detergency. Using more drastic conditions (pressure and temperature) with butylamine as reactant, it is possible to obtain hydroxy—butylamine derivatives. Under the reported conditions, transamidation is the inevitable co-reaction of amination, and side products such as imine and polycondensation products have been identified. Further work is still ongoing to optimize both operating conditions and results. Currently, these final products obtained on a pilot scale are also being tested.

Copyright © 1999 by AOCS Press

Methyl Esters: Epoxidation and Oxirane Opening

155

Acknowledgments

This work was supported by the European Community and the French organization of oilseeds producers (SOFIPROTEOL). References

1. Ashland Oil & Refining Company, Epoxidation Process Using Peracetic Acid Produced in situ, U.S. Patent 3,360,531 (1967). 2. Gan, N.L.H., S.H. Goh, and K.S. Ooi, Kinetic Studies of Epoxidation and Oxirane Clevage of Palm Olein Methyl Esters, J. Am. Oil Chem. Soc. 69: 347–351 (1992). 3. Zoebelein, H., Renewable Resources for the Chemical Industry, INFORM 3: 721–725 (1992). 4. Findley, T.W., D. Swern, and J.T. Scanlan, Epoxidation of Unsaturated Fatty Materials with Peracetic Acid in Glacial Acetic Acid Solution, J. Am. Chem. Soc. 67: 412–414 (1945). 5. Gunstone, F. D., Epoxidized Oils, in Lipid Technologies and Applications, edited by F.D. Gunstone and F. B. Padley, Marcel Dekker Inc., New York, 1997, pp. 759–769. 6. Schmitz, W.R., and J.G. Wallace, Epoxidation of Methyl Oleate with Hydrogen Peroxide, J. Am. Oil Chem. Soc. 31: 363–365 (1954). 7. Chadwick, A.F., D.O. Barlow, A.A. D’addieco, and J.G. Wallace, Theory and Practice of Resin-Catalyzed Epoxidation, J. Am. Oil Chem. Soc. 35: 355 (1958). 8. Solvay and Cie, Cyclic Epoxidation Process for Unsaturated Fatty Acid, DE Patent 1,257,767 (1962). 9. Argus Chemical Corporation, U.S. Patent 1,137,762 (1966). 10. Degussa and Henkel, Verfahrung zur epoxydierung olefinisch ungesattigter Verbindungen, DE Patent 0,032,990 (1981). 11. Henkel Kgaa, Continuous Epoxidation, Especially of Soya Oil, with Performic Acid Production in situ, Using Cascade of Reactors Operated with Cross Countercurrent Flow, DE Patent 3,320,219 (1984). 12. Henkel Kgaa, Preparation of Epoxidized Fatty Alcohol from Unsaturated Alcohol by Reaction with Aqueous Solutions of Hydrogen Peroxide and Formic Acid in Presence of Buffer, Without Catalyst, EP Patent 286937 (1988). 13. Rûsch gen. Klaas, M., and S. Warwel, Chemo-Enzymatic Epoxidation of Unsaturated Fatty Acids Esters and Plant Oils, J. Am. Oil Chem. Soc. 73: 1453–1457 (1996). 14. Dieckelmann, G., and H.J. Heinz, The Basics of Industrial Oleochemistry, edited by Peter Pomp Gmbh, Essen, 1988, pp. 133–144. 15. Zaher, F.A., and S.M. El-Shami, Oxirane Ring Opening by Formic Acid, Grasas Aceites 41: 361–365 (1990). 16. Nor, H.M., M.H. Mahmood, H. Kifli, and M.A. Rahman, The Use of Epoxidized Palm Oil Products for the Synthesis of Radiation Curable Resins, in Proceedings of World Conference on Oleochemicals: Into the 21st Century, edited by T.H. Applewhite, American Oil Chemists’ Society, Champaign, IL, 1991, pp. 311–314. 17. Ahmad, S., P. Siwayanan, O.T. Lye, R. Ali, A. Rafaei, Z. Zainuddin, D. Wiese, C.M. Choo, Characteristics of Polyurethane Foams Derived from Palm Oil Products, 21st ISF Congress, The Hague, 1-5 October, 1995. 18. BASF Ag, DE Patent 195 40 091 Al (1995).

Copyright © 1999 by AOCS Press

156

Pagès-Xatart-Parès et al.

19. BASF Ag, WO Patent 97/16408 (1997). 20. Henkel Kgaa, Fatty Alcohol Derivatives with Vicinal Hydroxy and Alkoxy Groups. Production by High Pressure Hydrolysis of Ester Prepared by Acylation, Epoxidation and Ring Opening, DE Patent 3,829,735 (1990). 21. Dahlke, B., S. Hellbardt, M. Paetow, and W.H. Zech, J. Am. Oil Chem. Soc. 72: 349–353 (1995).

Copyright © 1999 by AOCS Press

Chapter 10

New Oxidation Methods for Unsaturated Fatty Acids, Esters, and Triglycerides Mark Rüsch gen. Klaas and Siegfried Warwel

Institute for Biochemistry and Technology of Lipids, H.P. Kaufmann-lnstitute, Federal Centre for Cereal, Potato and Lipid Research, Piusallee 68, D-48147 Münster, Germany

Introduction

C=C-Oxidations of unsaturated fatty acids and their derivatives are important methods in oleochemistry. Two processes, the epoxidation of unsaturated plant oils by short-chain peracids and the Oxidative cleavage of oleic acid by ozone, are conducted on a large industrial scale. Both processes, principally unchanged since their introduction in the 1950s and early 1960s, employ potentially dangerous oxidants of high oxidizing power without catalysts. Therefore, we want to show catalytic reactions with more convenient oxidants for the following: i i i

epoxidation direct synthesis of epoxide derivatives (diols, hydroxy ethers) oxidative cleavage

for unsaturated fatty acids, esters, and alcohols

Epoxidation

Prileshajev-Epoxidation by Peroxy Acids

In industry, plant oils are epoxidized by the Prileshajev reaction with peroxy acids (Fig. 1). A short-chain peroxy acid, preferably peracetic acid, is prepared from hydrogen peroxide and the corresponding acid either in a separate step or in situ (1). To avoid the handling of peroxy acids, the in situ method is generally preferred for largescale epoxidation. The price that has to be paid for the gain in safety is the presence of a strong mineral acid that is essential for peroxy acid formation. The strong mineral acid is considered to be the main cause for side reactions, leading viaoxiranering opening to diols, hydroxyesters, estolides, and other dimers (2); thus the selectivity of industrial plant oil epoxidation rarely exceeds 80%. Furthermore, the strong acid in an Oxidative environment causes corrosion problems. Instead of peroxy acid in situ, performic acid in situ may also be used for epoxidation (3). In this case, no mineral acid is necessary to catalyze peroxy acid formation because formic acid itself is sufficiently acidic. However, it is also acidic enough to catalyze the epoxide ring opening (see p. 167). 157

Copyright © 1999 by AOCS Press

158

M.Rüsch gen. Klass and S. Warwel

Fig. 1. Epoxidation of plant oils by peracetic acid in situ.

The major application for epoxidized oils made by these means is their use as polyvinyl chloride (PVC)-plasticizers and stabilizers because of their ability to catch free HCl and slow degradation. Additionally, they can be used as reactive diluents for paints and as intermediates for polyurethane-polyol production. The most important product today is epoxidized soybean oil. The total worldwide production amounts to ~200,000 t/y (4). Although preformed short-chain peroxy acids are quite unstable, those with a higher molecular weight, i.e., a lower active oxygen content, are generally less demanding. Hence, m-chloroperbenzoic acid (MCPBA) remains the workhorse for the epoxidation of oleochemicals in laboratory-scale synthesis [for an experimental procedure, see e.g., Carlson et al. (5)]. In the case of plant oils and esters, the resulting MCPBA can be removed easily by alkaline washing, but this is not possible when free fatty acids are epoxidized; in that case, other work-up procedures, e.g., chromatography, must be used. In spite of its wide use, MCPBA, even in its diluted commercial form, is potentially explosive (6).

Copyright © 1999 by AOCS Press

New Oxidation Methods

159

Fig. 2. Epoxidation of oleic acid by autoxidation of benzaldehyde to perbenzoic acid in situ. Source: Ref. 7.

To escape the dilemma between using pure but risky preformed peroxy acids and in situ peroxy acid formation accompanied by a strong acid that may decrease selectivity, peroxy acids have been generated from aldehydes and molecular oxygen catalyzed by transition metal compounds. This technique has also been applied to the epoxidation of oleic acid (7). Benzaldehyde was used as the peroxy acid source and Co3+ as the catalyst (Fig. 2); the yield was 83% at 100% conversion after 1.5 h. In a similar way, 18O-marked epoxy fatty acid esters have been synthesized (8). Chemoenzymatic Epoxidation

A new way to prepare peroxy acids was discovered by Novo Nordisk, DK (9–11). They showed that some lipases catalyze the conversion of fatty acids with hydrogen peroxide (preferably 60%) to peroxy fatty acids; Novozym 435, an immobilized lipase from Candida antarctica on polyacrylic Lewatit, is the most active and stable biocatalyst for this purpose (Scheme 1). Recently we found that Novozym 435 is also capable of catalyzing perhydrolysis (12), i.e., the reaction of carboxylic acid esters with hydrogen peroxide to percarboxylic acids (Scheme 2). If the ester is applied both as solvent and as reactant, it is possible to use commercial 30–35% hydrogen peroxide without loss of reactivity. Additionally, perhydrolysis has a much broader substrate range, i.e., not only peroxy fatty acids but also branched and chiral peracids as well as peracetic and peracrylic acid can be prepared in situ. Both methods of biocatalytic peracid formation are extremely useful for epoxi laation in oleochemistry because both the necessary carboxyl group and the C=C-

Copyright © 1999 by AOCS Press

160

M.Rüsch gen. Klass and S. Warwel

SCHEME 2.

bonds are conveniently situated in one starting material. On the basis of the conversion of free fatty acids to peroxy fatty acids, we developed a convenient method for the chemoenzymatic “self-“ epoxidation of unsaturated fatty acids (13) (Fig. 3). Yields of 72–91% for natural fatty acids with internal C=C-bonds (including ricinoleic acid) and 61–68% for terminal unsaturated fatty acids were achieved; selectivity was ≥98% in all cases. The method has been employed for the epoxidation of meadowfoam fatty acids by Frykman and Isbell (14). Because there are no other organic components in the solvent than the fatty acid, the work-up procedure 3 is extremely simple. The immobilized lipase is removed by filtration and multiple use is possible; therefore a scale-up for commercial applications seems feasible. A similar method for the “self-“ epoxidation of plant oils, one based on perhydrolysis, yields 88–96% epoxides with a selectivity of ≥92% (15) (Fig. 4). The method is characterized by the use of 35% hydrogen peroxide and the addition of a small amount of free fatty acids, which is necessary not for peroxy acid formation but to prevent the formation of mono- and diglycerols by (per-)hydrolysis.

Fig. 3. Chemoenzymatic “self-“ epoxidation of unsaturated fatty acids: reaction principle.

Copyright © 1999 by AOCS Press

New Oxidation Methods

161

Fig. 4. Chemoenzymatic “self-“ epoxidation of plant oils.

Unsaturated fatty alcohols can also be epoxidized by lipase-catalyzed perhydrolysis (16). Interestingly, the outcome of the reaction depends on the ester applied for peroxy acid generation (Fig. 5). Fatty acid esters such as butyric acid ethylester react to epoxyalkanolacylates in a three-step one-pot reaction. Carbonic acid esters such as diethyl carbonate also form peroxy acids (percarbonic acid derivatives) and epoxidize the unsaturated alcohol; however, in a water-containing environment, they are obviously not stable enough to esterify the hydroxyl group. Thus, the end product is the epoxy alcohol. Epoxidation by Dioxiranes

Although dioxiranes have been known for some time, the breakthrough for their synthetic use was the discovery that they can be prepared in a simple manner from monopersulfates and ketones. Since then, they have been applied for a wide range of oxidations (17). There are two ways in which to utilize dioxiranes for epoxidation, i e., preparation of dioxiranes as distilled, storable ketone solutions (Fig. 6) and preparation of dioxiranes in situ (Fig. 7). Both methods have been successfully applied to unsaturated fatty acid methyl esters (18) and triglycerides (19). Preformed (dimethyl-)dioxirane in acetone is an excellent epoxidizing agent; as Prepared by the method of Adam et al. (20), it is routinely used in our group to pre-

Copyright © 1999 by AOCS Press

162

M.Rüsch gen. Klass and S. Warwel

Fig. 5. Chemoenzymatic epoxidation of unsaturated fatty alcohols with ethylbutyrate vs. diethylcarbonate as the peroxy acid source.

pare analytic samples of any epoxide. The unsaturated compound is solved in the acetone solution of dioxirane and stirred at room temperature for some hours. Then, only acetone must be removed to isolate the epoxide. Unfortunately, this convenient procedure is useless except on the smallest scale. Dioxirane solutions in acetone are usually 0.08 mol/L, and the utilization of monopersulfate in the dioxiran synthesis is ~5%. Taking into account the high molecular weight of monopersulfate (Oxone: 2 KHSO5 × KHSO4 × K2SO4; active oxygen ~4.7%), this means that ~4 kg of salt would be required for the epoxidation of 30 g of linolenic acid.

Fig. 6. Preformed distilled dioxiranes for C=C-epoxidation.

Copyright © 1999 by AOCS Press

New Oxidation Methods

163

Fig. 7. Epoxidation of triglycerides with ethylmethyldioxirane in situ. Source: Ref. 19.

The in situ procedure as proposed by Sonnet et al. (18) is much more attractive for synthetic applications. With the use of only a moderate excess of monopersulfate (C=C: KHSO5 = 1:2–2.4), they achieved an 80% yield for the epoxidation of oleic acid methyl ester and 81–96% for the epoxidation of various plant oils. It is a twophase reaction with a crown-ether as phase-transfer catalyst; yet a considerable amount of inorganic waste (six times the weight of the product) is produced. In a recent work (21), the phase-transfer catalyst was replaced by acetonitrile as a polar solvent. In summary, epoxidation by dioxiranes is a promising new method for oleochemistry, especially because it also works in combination with metal catalysts to influence diastereoselectivity (22); an enantioselective epoxidation with “sugardioxiranes” has also been reported (23). Transition Metal–Catalyzed Epoxidation

Countless transition metal-catalyzed epoxidation systems using a variety of oxidants(e.g., hydroperoxides, NaOCl, NaIO4, K3Fe(CN)6, PhIO, or amine oxides) have been proposed (24). The direct use of oxygen for C=C-epoxidation (excepting the silver-catalyzed production of ethylenoxide) has so far been elusive; thus for

Copyright © 1999 by AOCS Press

164

M.Rüsch gen. Klass and S. Warwel

economy and ecology, any process striving for industrial realization must use hydrogen peroxide or alkyl hydroperoxides (25).

Transition Metal–Catalyzed Epoxidation with Hydrogen Peroxide. There are few transition metal–catalyzed epoxidations of unsaturated fatty acids with hydrogen peroxide. By using a procedure developed by Ishii et al. (26), Bavaj (27) in our group epoxidized various fatty acid esters in a two-phase system with a combination of a tungsten heteropoly acid and a phase-transfer catalyst, using 30% H2O2 as the oxidant. (Table 1). Methyltrioxorhenium (CH3ReO3) is an especially versatile catalyst (see p. 168). Although the compound itself has been known for some time (28), Herrmann et al. (29) discovered both a simple synthesis (Fig. 8) and its ability to catalyze, among other reactions, C=C-oxidation (30). After only a single oleochemical example was conducted by Hermann et al. (31), Gerle (32) examined the epoxidation of oleic acid methyl ester, linseed oil fatty acid methyl esters, high-oleic sunflower oil, and linseed oil (Table 2). The reactions were carried out either at 0°C or in the presence of 2,2bypyridine to reduce the Lewis-acidity of CH3ReO3 because these conditions ensure that no epoxide ring opening by water occurs. The CH3ReO3-catalyzed epoxidation proceeds under mild conditions, and the selectivity is high. Commercial 30% hydrogen peroxide is used as the oxidant, but water must be removed by treatment of the H2O2 with MgSO4 in an organic solvent because the catalytic cycle (not CH3ReO3 itself) is sensitive to water (30). Gerle (32) also showed that the epoxide can be hydrogenated to obtain products that closely resemble hydrogenated castor oil, thus indicating a new potential market for epoxy oils. The most serious disadvantage of CH3ReO3-catalyzed oxidations is the problem of catalyst recovery. As can be easily seen by the color of the reaction mixture or by infrared (IR)-spectra, CH3ReO3 decomposes during the reaction. Although the rhenium content may be recovered, total turnover would have to be much higher to ren-

Copyright © 1999 by AOCS Press

New Oxidation Methods

Copyright © 1999 by AOCS Press

165

166

M.Rüsch gen. Klass and S. Warwel

der this oxidation system competitive. However, the method has been refined recently by Sharpless et al. (33); the catalyst has been supported on silica and was active for four cycles (34). Both methods, tungsten- and rhenium-catalyzed, illustrate the principal difficulties caused by hydrogen peroxide as the oxidant for transition metal—catalyzed epoxidations. On the one hand, hydrogen peroxide is not very soluble in many organic solvents, leading to the need for phase-transfer catalysts. On the other hand, when working in polar solvents, the unavoidable water bears the danger of epoxide ring hydrolysis to diols.

Transition Metal–Catalyzed Epoxidation with Alkyl Hydroperoxides. Alkyl hydroperoxides are attractive oxidants on a technical scale because they can be produced by autoxidation of branched alkanes with oxygen. This concept has been realized on the largest scale in the so-called Halcon process, i.e., the transition metalcatalyzed epoxidation of propylene to propylene oxide (35) (Fig. 9). Homogeneous and heterogeneous titanium, vanadium, and molybdenum catalysts are capable of catalyzing the C=C-epoxidation by alkyl hydroperoxide (for a review see Ref. 36). The reaction was also applied successfully in oleochemistry, and the epoxidation of unsaturated fatty acid esters and triglycerides has been thoroughly studied by Mar-

Fig. 9. Epoxidation of propylene by alkyl hydroperoxides (Halcon process).

Copyright © 1999 by AOCS Press

New Oxidation Methods

167

Fig. 10. Molybdenum-catalyzed epoxidation of unsaturated fatty acid esters by alkyl hydroperoxides. Source: Ref. 38.

tinez de la Cuesta et al. from 1978 to 1985, with results published in more than a dozen articles. They used homogeneous catalysts such MoO2(acac)2 or Mo(CO)6 and cumol hydroperoxide as oxidant. By epoxidizing soybean oil at 128–144°C, for example, they achieved a yield of ~60% epoxide after < 1 h reaction time (37). Ucciani et al. (38) used a commercial MoO3-Al2O3 contact to epoxidize various fatty acid methyl esters with yields and selectivities >95% in 3–9 h at 115°C (Fig. 10). Taking into account the facts that the catalyst is heterogeneous and can be reused (or even used continuously in a fixed bed reactor) and that oxygen is the primary oxidant, this process must be considered the most attractive epoxidation method. Why has it not yet replaced the nasty peroxy acid epoxidation in oleochemistry? If it is economic for propylene oxide, it should be even more so for oleochemicals for which the product/byproduct ratio is much more favorable. One reason may be that a huge petrochemical company has better outlets for the by-product (tert-butanol, styrene) than the oleochemicals industry. Another reason surely is plant capacity, because a typical halcon plant produces ~100,000 t/y propylene oxide. To sum up, this process may be superior to anything else, but only a real “global player” with a large market will consider it.

Direct Syntheses of Epoxide Derivatives

As mentioned, epoxidized fatty acids, esters, and alcohols are used for polyurethanepolyols after ring opening with water or alcohols (4) (Fig. 11). However, the ring opening of internal epoxides, in contrast to those of ethylene or propylene oxide, is not trivial, and quite harsh reaction conditions (160–200°C, sulfuric acid) have to be used (39). Therefore, the direct synthesis of these epoxide derivatives is a worthy target. Vicinal Hydroxylation by Performic Acid

The standard procedure for the vicinal hydroxylation of oleic acid in the laboratory is the use of peroxy formic acid in situ with subsequent alkaline hydrolysis of the hy-

Copyright © 1999 by AOCS Press

168

M.Rüsch gen. Klass and S. Warwel

Fig. 11. Vicinal diols and hydroxyethers as epoxide derivatives.

droxy monoformiates (Fig. 12) (40). Because the reaction sequence includes an ester hydrolysis, it is unsuitable for fatty acid esters and triglycerides. Additionally, peroxy formic acid is extremely hazardous, and its use in the laboratory should be restricted to a small scale. Transition Metal–Catalyzed Vicinal Hydroxylation with Hydrogen Peroxide

Transition metal–catalyzed vicinal hydroxylation, with the notable exception of the cis- hydroxylation by OsO4 and KMnO4, is mechanistically closely related to epoxidation (Fig. 13), i.e., epoxides are formed as intermediates and afterwards are hydrolyzed by water. Not surprisingly, the same catalysts are often used for transition metal—catalyzed epoxidation and vicinal hydroxylation, and the reaction conditions are adjusted accordingly. As early as 1967, Swern et al., (41) examined the direct hydroxylation of oleic acid and its derivatives with H2WO4 and 70% H2O2. They achieved 70–80% diols. The catalysts were preconditioned by partial solution in the oxidant and the reaction was carried out in emulsion without solvent. A very similar reaction catalyzed by MoO3 or MoO2(acac)2 has been disclosed by Degussa AG (42). CH3ReO3 is also capable of catalyzing vicinal hydroxylation; oleic acid methyl ester (31) and high-oleic sunflower oil (32) were hydroxylated with

Copyright © 1999 by AOCS Press

New Oxidation Methods

Fig. 12. Vicinal hydroxylation of oleic acid by peroxy formic acid in situ. Source: Ref. 40.

169

yields of 85–89% diols at 20–50°C. Rhenium(VII)oxide, the precursor of methyltrioxorhenium synthesis, is itself an active catalyst for C=C-hydroxylation (43), i.e., Re2O7-catalyzed vicinal hydroxylation of oleic acid methyl ester with 60% H2O2 in 1,4-dioxane yields after 16 h at 40°C 80% 9,10-dihydroxystearic acid methyl ester. Comparing these four methods (Fig. 14), we found the H2WO4-catalyzed method by Swern the most convenient because the catalyst is cheap and recoverable, the reaction conditions are moderate, no solvent is needed, and the space-time yield is high with good selectivity. We scaled the reaction up to kilogram scale, and it now appears ripe for commercial exploitation. Polyunsaturated fatty acid methyl esters are generally epoxidized similarly to monounsaturates, albeit sometimes with lower selectivity, whereas polyunsaturated

Fig. 13. Mechanism of Re-, W-, and Mo-catalyzed vicinal hydroxylations

Copyright © 1999 by AOCS Press

170

M.Rüsch gen. Klass and S. Warwel

Fig. 14. Four catalytic systems for the vicinal hydroxylation of oleic acid methyl ester by H2O2.

fatty acid esters and triglycerides cannot be hydroxylated with satisfying results. The amount of hydroxyl groups produced is always lower than predicted. After the first epoxide ring in a 9,10-12,13-diepoxy functionality has been hydrolyzed, for example, the resulting hydroxyl groups are in an excellent position to act as nucleophiles on the second epoxide ring (Fig. 15), thus leading to dihydroxytetrahydrofurane rings (44). Transition Metal–Catalyzed Alkoxyhydroxylation

The availability of various transition metal–catalyzed methods for direct vicinal hydroxylation prompted us to look for similar methods for the direct synthesis of hydroxyethers (Fig. 11). Our starting point was the direct vicinal hydroxylation of C=C-bonds with hydrogen peroxide catalyzed by Re2O7. Using the oxidation of 7tetradecen as a test reaction (in contrast to fatty acids, symmetric olefins do not lead to regioisomeric hydroxyethers), we compared vicinal hydroxylation and methoxyhydroxylation under very similar conditions (Fig. 16). In both cases, the total yields are comparable (71 vs. 67%). Hydroxylation gives only the diol, and methoxyhydroxylation gives predominantly 8-methoxy-7-tetradecanol (58%). Obviously. Re2O7 is acidic enough to catalyze the epoxide ring opening by alcohols nearly as

Copyright © 1999 by AOCS Press

New Oxidation Methods

171

Fig. 15. Formation of tetrahydrofurane rings by hydroxylation of polyunsaturates.

well as the opening by water. However, the hydroxyether is accompanied by 9% diol, which is of course formed by water, brought into the reaction system with the H2O2. These products are difficult to separate, and this has also been the key problem in those rare attempts at direct alkoxyhydroxylation that have been reported in the literature (45,46). We therefore decided that a selective method of alkoxyhydroxylation must avoid the use of H2O2 as the oxidant. We succeeded in modifying the molybdenum-catalyzed epoxidation by alkyl hydroperoxide as described earlier. First, the reaction was carried out in the alcohol (the nucleophile) as solvent. In addition to slowing down the reaction rate, the use of alcohols as the solvent is not known to have any effect on the transition metal-catalyzed C=C-epoxidation by alkylhydroperoxides (47). Furthermore, because the molybdenum catalyst was not active for epoxide ring opening at all, a second catalyst was added. After testing various homogeneous and heterogeneous Lewis and Bronstedt acids as well as some bases, we found boron trifluoride most suitable. BF3 has been used as a catalyst for fatty epoxide ring opening (48); its use as an industrial catalyst for the production of ethylbenzene is well known, and it can be conveniently applied in the laboratory in the form of its etherate or alcoholate. With this system (Fig. 17), oleic acid methyl ester was converted to 65% 9(10)-methoxy-10(9)hydroxystearic acid methyl easter. The selectivity of the reaction was 92%; the only

Copyright © 1999 by AOCS Press

172

M.Rüsch gen. Klass and S. Warwel

Fig. 16. Vicinal hydroxylation vs. methoxyhydroxylation by Re2O7/H2O2.

Fig. 17. Molybdenum-catalyzed methoxyhydroxylation of oleic acid methyl ester by tert-butyl hydroperoxide.

Copyright © 1999 by AOCS Press

New Oxidation Methods

173

by-product was the epoxide, which could be recycled together with the unreacted oleic acid methyl ester. In addition to their use for polyurethanes, hydroxyethers from unsaturated fatty acids may also be intermediates for detergents (49).

Oxidative C=C-Cleavage Ozonolysis

In addition to epoxidation, ozonolysis is the only C=C-oxidation in oleochemistry carried out on an industrial scale. By cleavage of fatty acid mixtures enriched in oleic acid, azelaic (nonanedioic) acid and pelargonic (nonanoic) acid are produced (Fig. 18). Although the mechanism of oleic acid ozonolysis has been discussed in detail (50), published information about the technical process and its problems is scarce (51,52). Ozonolysis itself is a highly selective reaction, but because the feedstock contains only 70% oleic acid, a costly downstream processing is unavoidable; even then the azealic acid is only ~90% (“polymer grade”) (53). Ozone is extremely toxic and aggressive. Furthermore, ozone itself and the intermediate ozonides are explosive (54). Considerable efforts have been concentrated on the economics of ozone generation (55) and on the safe decomposition of the ozonides (56). The result of these problems connected with the use of large amounts of ozone is that only a single plant of the Henkel Corporation (Cincinnati, OH) with a capacity of a few 10,000 t/y is operating, although there seems to be a robust market for azelaic acid in the fields of polymers, lubricants, and plasticizers (53).

Fig. 18. Azelaic and pelargonic acid by ozonolysis of oleic acid.

Copyright © 1999 by AOCS Press

174

M.Rüsch gen. Klass and S. Warwel

Transition Metal–Catalyzed C=C-Cleavage

C=C-bonds in unsaturated fatty acid esters can be cleaved without ozone in the laboratory, e.g., by KMnO4/NaIO4 (57) or OsO4/Cr(VI) (58). As an industrial alternative to ozonolysis, these (and similar) methods are far too expensive. Zaidman et al. (59) showed that only H2O2 and NaOCl can theoretically cleave unsaturated fatty acids at a cost comparable to that of ozone; unfortunately, they did not include alkyl hydroperoxides in their calculation. A ruthenium-catalyzed procedure for the Oxidative C=C-cleavage of oleic acid by NaOCl has been published (60). According to our own experiments in this field, the need to adjust the pH-value twice (alkaline for oxidation and acidic for extraction) and especially the formation of a small (~2%) but rather suspicious amount of chlorinated by-products render this process not feasible for industrial application. We have been studying the transition metal–catalyzed cleavage of unsaturated fatty acids for some years and reviewed our results twice (61,62). Fig. 19 shows some of the examined pathways. The direct cleavage of internal C=C-bonds in native fatty acids by peracetic acid and Ru-catalyst or by hydrogen peroxide with homogeneous and heterogeneous Re-, W-, and Mo-catalysts results in yields of only 50–60%. On the other hand, yields ≥80% are achieved for the cleavage of terminal unsaturated fatty acid ester by either Ru-catalysts/CH3CO3H (63) or by Re2O7/H2O2 (64). The two-step-process via metathesis of oleic acid methyl ester with ethylene, yielding 9-decenoic acid methyl ester and 1-decene (65,66), has the advantage that the production of the two products of direct cleavage can be decoupled. 1-Decene will be submitted to Oxidative cleavage only if there is a market for pelargonic acid. Be-

Fig. 19. Pathways for the Oxidative cleavage of unsaturated fatty acids.

Copyright © 1999 by AOCS Press

New Oxidation Methods

175

cause this decoupling is independent of the oxidation step, it could also be used in combination with ozonolysis. Another two-step procedure uses the diol as the intermediate. Two-step Oxidative cleavage via diols or epoxides was proposed several years ago (67). However, Hinze was obviously not aware of direct processes for vicinal hydroxylation and therefore concentrated on epoxides as the intermediate. In light of the excellent methods for direct vicinal hydroxylation (see p. 168), diol cleavage now deserves renewed attention. Transition Metal–Catalyzed Diol Cleavage

Transition Metal–Catalyzed Diol Cleavage by Oxygen. Since 1954, various processes for diol cleavage by oxygen have been described (68–71). They all used Co2+/Co3+ catalysts, but in other aspects were often contradictory. According to our own experience, selectivities of >60% are not reproducible. A new promising catalyst for Oxidative diol-cleavage by oxygen is Pb/BixRu8-xO13 (72,73); to our knowledge, it has not yet been applied to dihydroxy fatty acids. In addition to these transition metal-catalyzed oxidation systems, dihydroxy fatty acids may also be cleaved similarly to ricinoleic acid by “caustic” oxidation with alkali at ~250°C (74,75). Transition Metal–Catalyzed Diol Cleavage by Hydrogen Peroxide. Modifying our systems for the Oxidative C=C-cleavage to fit diol cleavage, we found that 9,10dihydroxystaeric acid methyl ester is converted with nearly 80% yield when Re2O7 is used as the catalyst and water is withdrawn by azeotropic distillation during the reaction with a particular solvent system (Fig. 20). Unfortunately, the reaction requires 85% H2O2, which may cause safety problems. As mentioned earlier, ruthenium cat-

Fig. 20. Re2O7-catalyzed Oxidative cleavage of 9,10-dihydroxystearic acid methyl ester by H2O2. Conditions: Re2O7:H2O2:diol = 1:100:360 (molar); 16 h at 90°C in acetic acid anhydride/1,4-dioxane/toluene or acetic acid anhydride/triethylphosphate/toluene.

Copyright © 1999 by AOCS Press

176

M.Rüsch gen. Klass and S. Warwel

Fig. 21. Ruthenium-catalyzed Oxidative cleavage of 9,10-dihydroxystearic acid methyl ester by H2O2 in a two-phase system. Conditions: RuCI3:DDAB:diol:H3O2 = 1:20:500:4000 (molar); 2 h at 80°C in 1,2-dicholethane. DDAB, didodecyldimethyl ammonium bromide.

alysts are generally the first choice for any carbon-carbon-bond cleavage. However, these compounds do not work together with hydrogen peroxide but decompose it very quickly, usually accompanied by the precipitation of RuO2. That problem has been overcome by a two-phase system for the oxidation of styrene (76), and the system also works for the oxidation of 9,10-dihydroxystearic acid methyl esters (Fig. 21). We obtained 82–86% yield of cleavage products; some chain shortening is typical for ruthenium-catalyzed oxidations. To avoid the use of an expensive phase-transfer catalyst and a chlorinated solvent, we modified the system. Using Ru(PPh3)3Cl2 as the catalyst, only moderate H2O2-decomposition occurred (probably because no RuO2 precipitated); with 35% hydrogen peroxide, 9,10-dihydroxystearic acid methyl ester was cleaved with a yield of 78–82% (Fig. 22). In summary, Oxidative cleavage of diols by Ru(PPh3)3Cl2/H2O2 is a further but not final step in the development of a catalytic alternative to ozonolysis. As in the case of epoxidation, the use of alkyl hydroperoxides instead of hydrogen peroxide may bring additional benefits; however, this field has scarcely been explored (77).

Copyright © 1999 by AOCS Press

New Oxidation Methods

177

Fig. 22. Ru(PPh3)3Cl2-catalyzed Oxidative cleavage of 9,10dihydroxystearic acid methyl ester by H2O2 in a one-phase system. Conditions: Ru(PPh3)3CI2:diol:H2O2 = 1:500:4000 (molar); 2 h at 80°C in tert-butanol.

References

1. Formo, M.W., Miscellaneous Oil and Fat Products, in Bailey’s Industrial Oil and Fat Products, 4th edn., edited by D. Swern, New York, 1982, pp. 366–371. 2. Rangarajan, B., A. Havey, E.A. Grulke, and P.D. Culnan, Kinetic Parameters of a TwoPhase Model for in situ Epoxidation of Soybean Oil, J. Am. Oil Chem. Soc. 72: 1161–1169 (1995). 3. Weigert, W., W. Merck, H. Offermans, G. Prescher, G. Schreyer, and O. Weiberg, Wasserstoffperoxid und -Derivate als Oxidationsmittel in der Organischen Chemie, in Wasserstoffperoxid und seine Derivate, edited by W. Weigert, Huthig, Heidelberg, 1978, pp. 80–81. 4. Eierdanz, H., Industrielle Fettchemie - Rohstoffe, Verfahren, Produkte, in Nachwachsende Rohstoffe - Perspektiven fur die Chemie, edited by M. Eggersdorfer, S. Warwel, and G. Wulff, VCH, Weinheim, 1993, pp. 23–32. 5. Carlson, K.D., R. Kleiman, and M.O. Bagby, Epoxidation of Lesquerella and Limnathes (Meadowfoam) Oils, J. Am. Oil Chem. Soc. 71: 175–182 (1994). 6. Heaney, H., Novel Organic Peroxygen Reagents for Use in Organic Synthesis. Top. Curr. Chem. 164: 2–19 (1993). 7. Kuo, M.-C, and T.-C. Chou, Kinetics and Mechanism of the Catalyzed Epoxidation of Oleic Acid with Oxygen in the Presence of Benzaldehyde, Ind. Eng. Chem. Res. 26: 277– 284 (1987). 8. Schottler, M., and W. Boland, Efficient Synthesis of Isotopically [18O]-Epoxides Using Molecular Oxygen, Synlett, 91–92 (1997). 9. Björkling, F., S.E. Godtfredsen, and O. Kirk, Lipase-Mediated Formation of Peroxycarboxylic Acids Used in Catalytic Epoxidation of Alkenes, J. Chem. Soc. Chem. Commun., 1301–1303 (1990).

Copyright © 1999 by AOCS Press

178

M.Rüsch gen. Klass and S. Warwel

10. Björkling, F., H. Frykman, S.E. Godtfredsen, and O. Kirk, Lipase Catalyzed Synthesis of Peroxycarboxylic Acids and Lipase Mediated Oxidations, Tetrahedron 48: 4587– 4592 (1992). 11. Kirk, O., M.W. Christensen, T. Damhus, and S.E. Godtfredsen, Enzyme Catalyzed Degradation and Formation of Peroxycarboxylic Acids, Biocatalysis 11: 65–77 (1994). 12. Rüsch gen. Klaas, M., and S. Warwel, Lipase-Catalyzed Preparation of Peroxy Acids and Their Use for Epoxidation, J. Mol. Catal. A 117: 311–319 (1997). 13. Warwel, S., and M. Rüsch gen. Klaas, Chemo-Enzymatic Epoxidation of Unsaturated Carboxylic Acids. J. Mol. Catal. B 1: 29–35 (1995). 14. Frykman, H.B., and T.A. Isbell, Synthesis of 6-Hydroxy δ-Lacotones and 5,6-Dihydroxy Eicosanoic/Docosanoic Acids from Meadowfoam Fatty Acids via Lipase-Mediated SelfEpoxidation, J. Am. Oil Chem. Soc. 74: 719–722 (1997). 15. Rüsch gen. Klaas, M., and S. Warwel, Chemo-Enzymatic Epoxidation of Unsaturated Fatty Acids Esters and Plant Oils, J. Am. Oil Chem. Soc. 73: 1453–1457 (1996). 16. Rüsch gen. Klaas, M., and S. Warwel, A Three-Step-One-Pot Chemo-Enzymatic Synthesis of Epoxyalkanolacylates, Synth. Commun. 28: 251–260 (1998). 17. Murray, R.W., Dioxiranes, Chem. Rev. 89: 1187–1201 (1989). 18. Sonnet, P.E., M.E. Lankin, and G.P. McNeill, Reactions of Dioxiranes with Selected Oleochemicals, J. Am. Oil. Chem. Soc. 72: 199–204 (1995). 19. Sonnet, P.E., and T.A. Foglia, Epoxidation of Natural Triglycerides with Ethylmethyldioxirane, J. Am. Oil Chem. Soc. 73: 461–464 (1996). 20. Adam, W., J. Bialas, and L. Hadjiarapoglou, A Convenient Preparation of Acetone Solution of Dimethyldioxirane, Chem. Ber. 124: 2377 (1991). 21. Lie Ken Jie, M.S.F., and M.K. Pasha, Epoxidation Reactions of Unsaturated Fatty Esters with Potassium Peroxomonosulfate, Lipids 33: 633–637 (1998). 22. Foglia, T.A., P.E. Sonnet, A. Nunez, and R.L. Dudley, Selective Oxidations of Methyl Ricinoleate: Diastereoselective Epoxidation with Titanium IV Catalysts, J. Am. Oil. Chem. Soc. 75: 601–607 (1998). 23. Wang, Z.-X., Y. Tu, M. Frohn, J.-R. Zhang, and Y. Shi, An Efficient Catalytic Epoxidation Method, J. Am. Chem. Soc. 119: 11224–11235 (1997). 24. Jorgensen, K.A., Transition-Metal-Catalyzed Epoxidations, Chem. Rev. 89: 431–458 (1989). 25. Sheldon, R.A., Fine Chemicals by Catalytic Oxidation, Chemtech, 566–576 (1991). 26. Ishii, Y., K. Yamawaki, T. Ura, H. Yamada, T. Yoshida, and M. Ogawa, Hydrogen Peroxide Oxidation Catalyzed by Heteropoly Acids Combined with Cetylpyridinium Chloride: Epoxidation of Olefins and Allylic Alcohols, Ketonization of Alcohols and Diols, and Oxidative Cleavage of 1,2-Diols and Olefins, J. Org. Chem. 53: 3587–3593 (1988). 27. Bavaj, P., Zu molybdän- und wolframkatalysierten Oxidationen von C=C-Doppelbindungen petro- und fettchemischer Verbindungen mit Wasserstoffperoxid, Ph.D. Thesis, Rheinisch-Westfälische Technische Hochschule, Aachen, 1995. 28. Beattie, I.R., and P.J. Jones, Methyltrioxorhenium. An Air-Stable Compound Containing a Carbon-Rhenium Bond, J. Am. Chem. Soc. 97: 2318 (1979). 29. Herrmann, W.A., F.E. Kühn, R.W. Fischer, W.R. Thiel, and C.C. Romao, Simple and Efficient Synthesis of MeReO3. A General Method, Inorg. Chem. 31: 4431–4432 (1992). 30. Herrmann, W.A., R.W. Fischer, M.U. Rauch, and W. Scherer, Alkylrhenium Oxides as Homogeneous Epoxidation Catalysts: Activity, Selectivity, Stability and Deactivation, J. Mol. Catal. 86: 245–266 (1994).

Copyright © 1999 by AOCS Press

New Oxidation Methods

179

31. Herrmann, W.A, D. Marz, W. Wagner, J.G. Kuchler, G. Weichselbaumer, and R. Fischer, Deutsche Offenlegungsschrift 3902357 Al (1989). 32. Gerle, M., Zur Metathese und Oxidation ausgewahlter Fettsäuremethyl ester und Triglyceride an Rheniumkatalysatoren. Beiträge zur Nutzung nachwachsender Rohstoffe, Ph.D. Thesis, Rheinisch-Westfälische Technische Hochschule, Aachen, 1997. 33. Rudolph, J., K.L. Reddy, J.P. Chiang, and K.B. Sharpless, Highly Efficient Epoxidation of Olefins Using Aqueous H2O2 and Catalytic Methyltrioxorhenium/Pyridine: PyridineMediated Ligand Acceleration, J. Am. Chem. Soc. 119: 6189–6190 (1997). 34. Neumann, R., and T.-J. Wang, Methyltrioxorhenium Supported on Silica Tethered with Polyethers as Catalyst for the Epoxidation of Alkenes with Hydrogen Peroxide, Chem. Commun., 1915–1916 (1997). 35. Landau, R., G.A. Sullivan, and D. Brown, Propylene Oxide by Co-Product Processes. Chemtech, 602–607 (1979). 36. Sharpless, K.B., and T.R. Verhoeven, Metal-Catalyzed Highly Selective Oxygenations of Olefins and Acetylenes with t-BuOOH. Practical Considerations and Mechanisms, Aldrichchim. Acta 12: 63–82 (1979). 37. Martinez de la Cuesta, P.J., E. Rus Martinez, and L.M. Cotoluero Minguez, Epoxidacion de Aceite de Soja Refinado Mediante Hidroperoxido de Isopropilbenceno, Grasas Aceites 36: 181–185 (1985). 38. Debal, A., G. Rafaralahitsimba, and E. Ucciani, Epoxidation of Fatty Acid Methyl Esters with Organic Hydroperoxides and Molybdenum Oxide, Fat Sci. Technol. 95: 236–239 (1993). 39. Dahlke, B., S. Hellbardt, M. Paetow, and W.H. Zech, Polyhydroxy Fatty Acids and Their Derivatives from Plant Oils, J. Am. Oil Chem. Soc. 72: 349–353 (1995). 40. Swern, D., J.T. Scanlan, and G.B. Dickel, 9,10-Dihydroxystearic Acid, Org. Synth. Coll. IV, 317–320 (1963). 41. Luong, T.M., H. Schriftmann, and D. Swern, Direct Hydroxylation of Fats and Derivatives with a Hydrogen Peroxide Tungstic Acid System, J. Am. Oil Chem. Soc. 44: 316– 320 (1967). 42. Dankowski, M., G. Goor, and G. Prescher, Stets geforscht... Chemieforschung im DegussaForschungszentrum Wolfgang, Publication of Degussa AG, Frankfurt, 1988, vol. 2, p. 63. 43. Warwel, S., M. Rüsch gen. Klaas, and M. Sojka, Formation of Vicinal Diols by Re2O7 Catalyzed Hydroxylation of Alkenes with Hydrogen Peroxide, Chem. Commun., 1578– 1579 (1991). 44. Weber, N., K. Vosmann, E. Fehling, K.D. Mukherjee, and D. Bergenthal, Analysis of Hydroxylated Fatty Acids from Plant Oils, J. Am. Oil Chem. Soc. 72: 361–368 (1995). 45. Smith, C.W., D.G. Norton, and G.B. Paqyne, Production of Hydroxy Ethers, U.S. Patent 2808442, 1953. 46. Gulliver, D.J., and S.J. Kitchen, Alkoxy-Alcohols from Olefins, UK Patent application 2252556, 1991. 47. Sheldon, R.A., and J.A. van Doom, Metal Catalyzed Epoxidation of Olefins with Organic Hydroperoxides, J. Catal. 31: 427–435 (1973). 48. Bischoff, M., U. Zeidler, and H. Baumann, Ätheralkohole und Esteralkohole - neue Tensidrohstoffe auf Basis von Olefinoxiden, Henkel-Referate 13: 44–49 (1977). 49. Bischoff, M., H. Baumann, H. Andree, and E. Sung, l-Methoxyalkyl-sulfate-2, ihre Herstellung und Verwendung in Wasch- und Reinigungsmitteln, Deutsche Offenlegungschrift 2651925, 1976.

Copyright © 1999 by AOCS Press

180

M.Rüsch gen. Klass and S. Warwel

50. Rebrovic, L., The Peroxidic Species Generated by Ozonolysis of Oleic Acid or Methyl Oleate in a Carboxylic Acid Medium, J. Am. Oil Chem. Soc. 69: 159–165 (1992). 51. Rebrovic, L., and F.D. Gunstone, Oxidative Cleavage of Unsaturated Fatty Acids, Lipid Technol. 78: 135–137 (1996). 52. Fayter, R.G., Technical Reactions for the Production of Oleochemical Monomers, in Perspektiven nachwachsender Rohstoffe in der Chemie, edited by H. Eierdanz, VCH, Weinheim, 1996, pp. 107–117. 53. Janakiefski, N.G., The Unique Chemistry of Azelaic Acid, NLGI Spokesman 61: 14–23 (1997). 54. Schober, B.D., Ozonolysis and Reduction in Fine Chemicals Industry, Chim. Oggi., 21– 26 (1995). 55. Klein, H.-P., Large-Scale Production and Application of Highly Concentrated Ozone, J. Am. Oil Chem. Soc. 61: 306–312 (1984). 56. Rebrovic, L., Catalyzed Process for Oxidation of Ozonides of Unsaturates to Carboxylic Acids, U.S. Patent 5399749 (1995). 57. von Rudloff, E., Periodate-Permanganate Oxidations. Oxidation of Lipids in Media Containing Organic Solvents, Can. J. Chem. 34: 1413–1418 (1956). 58. Henry, J.R., and S.M. Weinreb, A Convenient, Mild Method for Oxidative Cleavage of Alkenes with Jones Reagent/OsmiumTetraoxide. J. Org. Chem. 58: 4745 (1993). 59. Zaidmann, B., A. Kisilev, Y. Sasson, and N. Garti, Double Bond Oxidation of Unsaturated Fatty Acids, J. Am. Oil Chem. Soc. 65: 611–615 (1988). 60. Foglia, T.A., P.A. Barr, and A.J. Malloy, Oxidation of Alkenes with Use of Phase Transfer Catalysts, J. Am. Oil Chem. Soc. 54: 858A–861A (1977). 61. Rüsch gen. Klaas, M., P. Bavaj, and S. Warwel, Transition-Metal Catalyzed Oxidative Cleavage of Unsaturated Fatty Acids, Fat Sci. Technol. 97: 359–367 (1995). 62. Warwel, S., and M. Rüsch gen. Klaas, Oxidative Cleavage of Unsaturated Fatty Acids Without Ozone, Lipid. Technol. 79: 10–14 (1997). 63. Warwel, S., M. Sojka, and M. Rüsch gen. Klaas, Synthesis of Dicarboxylic Acids by Transition-Metal Catalyzed Oxidative Cleavage of Terminal Unsaturated Fatty Acids, Top. Curr. Chem. 164: 79–89 (1993). 64. Warwel, S., and M. Rüsch gen. Klaas, Production of Carboxylic Acids, U.S. Patent 5321158 (1994). 65. Warwel, S., H.-G. Jägers, and S. Thomas, Metathese ungesättigter Fettsäurester - ein einfacher Zugang zu langkettigen Dicarbonsäuren, Fat Sci. Technol. 94: 323–328 (1992). 66. Warwel, S., Bavaj. P., Rüsch gen. Klaas, M. and Wolff, B., Polymerbausteine aus Pflanzenölen durch katalytische Reaktionen, in Perspektiven nachwachsender Rohstoffe in der Chemie, edited by H. Eierdanz, VCH, Weinheim 1996, pp. 119–135. 67. Hinze, A.G., Oxidatives Verfahren zur Herstellung kurzkettiger Carbonäsuren aus ungesättigten Fettsäuren, Fat Sci. Technol. 89: 339–342 (1987). 68. Mackenzie, J.S., and C.S. Morgan, Oxidation of Fatty Acids, U.S. Patent 2820046 (1954). 69. de Vries, G., and A. Schors, Oxygenation of Organic Compounds. A Novel Oxidative Cleavage of 1,2-Glycols. Tetrahedron Lett., 5689–5690 (1968). 70. Schwarze, W., and W. Weigert, Verfahren zur Oxidation vicinaler Diolgruppen- Deutsche Offenlegungsschrift 2035558 (1970). 71. Zeidler, U., Verfahren zur Herstellung von Ketonen, Deutsche Offenlegungsschrift 2144117 (1971).

Copyright © 1999 by AOCS Press

New Oxidation Methods

181

72. Felthouse, T.R., Catalytic Oxidative Cleavage of Vicinal Diols and Related Oxidations by Ruthenium Pyrochlore Oxides: New Catalysts for Low-Temperature Oxidations with Molecular Oxygen, J. Am. Chem. Soc. 109: 7566–7568 (1987). 73. Arts, S.J.H.F., F. van Rantwijk, and R.A. Sheldon, Oxidation Studies of Carbohydrates Using Oxygen and a Bismuth-Ruthenium Oxide Catalyst, J. Carbohydr. Chem. 15: 317– 329 (1996). 74. Dytham, R.A., and B.C.L. Weedon, Organic Reactions in Strong Alkalis. Fission of Keto- and Hydroxy-Acids, Tetrahedron 8: 246–260 (1960). 75. Offermanns, H., W. Schwarze, and W. Weigert, Verfahren zur Herstellung von Carbonsauren, Deutsche Offenlegungsschrift 2034014 (1970). 76. Barak, G., and Y. Sasson, Dual-Function Phase-Transfer Catalysis in the Metal-Assisted Oxidation by Hydrogen Peroxide of Styrene to Benzaldehyde and Acetophenon, Chem. Commun., 1266–1267 (1987). 77. Kaneda, K., K. Morimoto, and T. Imanaka, Molybdenylacetyl-Acetonate Complex as a Reagent for Oxidative Cleavage of vic-Diols, Chem. Lett., 1295–1296 (1988).

Copyright © 1999 by AOCS Press

Chapter 11

Some Recent Advances in Epoxide Synthesis George J. Piazza

Eastern Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture, Wyndmoor, PA 19038

Introduction

Commercial processes for preparing epoxidized oils and fatty acid epoxy esters from unsaturated oils and esters were developed in the 1940s (1). These processes use either preformed peracid or peracid prepared in situ by the addition of H2O2 to acetic acid or formic acid in the presence of sulfuric acid. The resulting epoxidized fatty materials are used commercially as plasticizers. The epoxide functionality is a versatile intermediate for the production of many different chemical products. Shown in Scheme 1 are several examples of common reactions that epoxides undergo (2,3). Research on epoxide synthesis has continued for several reasons, including the following: (i) the acid present or generated by the peracid procedure catalyzes opening of the oxirane ring, causing epoxide yields to be less than quantitative; (ii) the peracid procedure shows only modest selectivity; and (iii) the spent peracid must be removed from the epoxide product and disposed of or recycled. When considering the manufacture of commodity chemicals such as those primarily produced from fats and oils, the cost of by-product disposal is often a major component of overall cost. Excellent reviews of epoxidation syntheses have been published (4–8). This chapter will review some recent advances in this area with emphasis on work using fatty substances.

SCHEME 1. Epoxide as facile intermediate for synthesis. 182 Copyright © 1999 by AOCS Press

Epoxide Synthesis

Epoxidation with Dioxiranes

183

A dioxirane is a three-membered ring containing two oxygen atoms. The most studied dioxirane is dimethyl dioxirane (DMDO), which is conveniently prepared by the action of potassium peroxymonosulfate with acetone in the presence of bicarbonate (Scheme 2). DMDO can be isolated by vacuum distillation, or it can be generated in situ in a biphasic system with a phase transfer agent (9). The main by-product formed when DMDO is used for epoxidation is acetone. In addition, a small amount of methyl acetate is formed from auto-decomposition of DMDO (10). Because DMDO epoxidations are conducted at alkaline pH, acid-sensitive epoxides can be prepared. Even when an in situ biphasic reaction system is used, oxirane ring opening is minimal. DMDO, like peracids, exhibits high stereoselectivity in its reactions with cis and trans alkenes. In addition, DMDO reacts with cis-alkenes 7–9 times faster than with trans alkenes (9). Allylic alcohols have the potential to promote diastereoselective epoxidation with DMDO provided that a nonpolar solvent is used; this allows hydrogen bonding between DMDO and the alcohol (11). DMDO also has the ability to oxidize methylene groups to ketones. However, its reactivity toward methylene groups is much lower than its epoxidation activity on double bonds (12). Dioxirane has been used to epoxidize fatty materials. The epoxidation of methyl oleate took place cleanly and completely (12). The action of DMDO/acetone on methyl ricinoleate did not give satisfactory yields of epoxy alcohol. However, when a biphasic reaction system was used with butanone to give ethylmethyldioxirane (EMDO), high yields of epoxide were obtained. Oils from several sources were subjected to the action of EMDO, prepared in situ in a biphasic medium (13). Generation of EMDO required two additions of peroxymonosulfate to achieve high levels of epoxide formation with oils containing higher levels of unsaturation. Yields were poor only with tung oil, which contains esters of the conjugated polyunsatured fatty acid α-eleostearic. When methyl ricinoleate was treated with an excess of EMDO, the

SCHEME 2. Synthesis of dimethyl dioxirane (DMDO) and its use for epoxidation.

Copyright © 1999 by AOCS Press

184

G.J. Piazza

ketoepoxide, methyl (Z)-9,10-oxido-12-oxo-octadecanoate, was produced (14). The ketoepoxide was acid labile and rearranged to methyl 8-(5-hexylfuran-2-yl)octanoate. Dioxiranes were generated in the presence of trifluoroacetone or methyl pyruvate (15). A diepoxide could be obtained from a fatty methyl ester containing two conjugated double bonds with dioxirane, whereas with m-chloroper-oxybenzoic acid, only the monoepoxide could be generated. Epoxy-hydroxystearate was generated by dioxirane from iso-ricinoleic acid, but m-chloroperoxybenzoic acid gave only the tetrahydrofuran derivative.

Aldehyde and Oxygen Epoxidations

A wide variety of metal catalysts have been examined for their ability to promote epoxide formation (4–8). In most cases, these catalysts require higher energy oxidants such as hydrogen peroxide, organic peroxides, or iodosylbenzene. With some metal catalysts, molecular oxygen can be used as the oxidant, provided that an aldehyde is added to the reaction mixture (16,17). Epoxide formation is accompanied by co-oxidation of the aldehyde to a carboxylic acid (Scheme 3). The kinetics of oleic acid epoxidation by Co+3 in the presence of benzaldehyde were determined (18). The kinetics could be described by a mechanism in which the benzaldehyde donates an electron to Co+3, and the resulting radical then reacts with O2 to give the peroxybenzoate radical. Reaction of the peroxybenzoate radical with oleic acid was the main source of epoxy oxygen (16). Cobalt acetate was bound to a membrane by ion exchange (19). This membrane was used as a catalyst for the epoxidation of oleic acid with benzaldehyde and oxygen. The membrane-bound cobalt catalyst could be recycled. Synthetic metalloporphyrins are analogs of the prosthetic group of heme-containing enzymes that selectively catalyze various reactions, i.e., oxygenation, oxidation, oxidative chlorination, and dismutation (6). In recent years, the number of studies of metalloporphyrins and other organic metal complexes has increased because the organic ligands can modulate the activity and selectivity of the metal (8,16,20,21). With respect to fatty epoxide formation by benzaldehyde/O2, there has been only one study. The epoxide of oleic acid was formed using cobalt (II) tetraphenylporphyrin as a catalyst (22). It was found that the use of 4-methoxybenzaldehyde increased the rate of epoxidation sevenfold, whereas 4-chlorobenzaldehyde decreased the rate of epoxidation compared with benzaldehyde.

SCHEME 3. Epoxidation with aldehyde and O2 with a metal catalyst.

Copyright © 1999 by AOCS Press

Epoxide Synthesis

Epoxidation with Hydrogen Peroxide

185

Methyltrioxorhenium as Catalyst

Methyltrioxorhenium (methylrhenium trioxide), CH3ReO3 (MTO), catalyzes the epoxidation of olefins with hydrogen peroxide (23). The catalytically active species is formed when 1 or 2 mol of H2O2 react with MTO to give organorhenium peroxo complexes with a ligand of water (Scheme 4) (24). MTO also catalyzes the hydrolysis of the oxirane ring, leading to trans-configured 1,2-diols. Oxirane ring opening can be suppressed by conducting the epoxidation at temperatures as low as −25°C or by adding an amine as an acid scavenger. Epoxidation of cis-β-methylstyrene and transβ-methylstyrene led to the cis and trans epoxides, respectively (25). This observation appears to rule out either a radical mechanism or one involving nucleophilic attack of one of the olefinic carbon atoms on the peroxo oxygen. These studies were extended to substituted alkenes (26). Epoxidation rates increased with increasing alkyl substitution of the double bond and decreased when electron-attracting groups were present. Rates of epoxidation were about an order of magnitude greater in CH3CN/H2O than in methanol. It was concluded that the data were consistent with a mechanism in which the double bond of the alkene attacks a peroxidic oxygen in a concerted manner. In a successful attempt to reduce oxirane ring opening, MTO was used with a urea/H2O2 adduct, thus eliminating the presence of H2O (27). Diastereoselectivity was examined with a series of allylic alcohols and found to be approximately the same as that given by m-chloroperbenzoic acid, with the threo epoxy alcohol predominant in most cases. It also was found that the addition of pyridine to MTO substantially increased the degree of epoxidation and decreased ring opening of the epoxide even when aqueous H2O2 was used as the oxidant (28). Although the mechanism by which this acceleration occurs is not known, the observation that 2-picoline cannot provide reaction acceleration eliminates a simple base effect as the origin of the rate enhancement. Epoxidation of selected cyclic dienes resulted in high diastereoselectivity in diepoxide formation. In another study of epox-

SCHEME 4. Epoxidation with methyltrioxorhenium/H2O2 with subsequent oxirane ring opening in the presence of water. The second line shows the structures of the epoxidizing agents.

Copyright © 1999 by AOCS Press

186

G.J. Piazza

idation of conjugated dienes with MTO/aqueous H2O2, diols were the major product in many cases (29). When urea/H2O2 complex was used as the oxidant, monoepoxide was the predominant product. Tungstate as Catalyst

This procedure is uniquely suitable for epoxide formation on terminal olefins although internal olefins also are reactive. The original procedure for 1-octene epoxidation used H2O2 and a Na2WO4-H3PO4-quaternary ammonium chloride combined catalyst in a 1,2-dichloroethane-water biphasic system. Later work was performed with tungsten peroxo complexes (30). Both procedures were unsatisfactory in that excess olefinic substrate was used, limiting yields of epoxide to ≤60%. Another research group reported that tungstate ion alone could not catalyze the epoxidation of 1-octene, but that the heteropoly acid, H3PW12O40, would give the epoxide in 96% yield when the reaction was conducted with cetylpyridinium chloride, H2O2, in CHCl3/H2O (Scheme 5) (31). A variety of allylic alcohols such as trans-2-buten-1ol were also epoxidized in good yield with this system. In a further study of epoxidation by H3PW12O40 (termed Ishii-Venturello chemistry), the active catalytic epoxidation species was confirmed to be {PO4[WO(O2)2]4}3- (32). Monitoring of gaseous products showed that H2O2 disproportionation was the dominant side reaction. Catalysis stopped after 500 turnovers as a result of catalyst inactivation by product epoxide. Ishii-Venturello chemistry has been applied to the epoxidation of undecylenic acid (10-undecenoic acid) and its methyl ester (33). Undecylenic acid is manufactured by pyrolysis of the sodium or calcium salts of ricinoleic acid from castor oil. In terms of yield and selectivity, results were superior to those obtained with preformed and in situ prepared peroxyacetic acid. In the biphasic system 1,2-dichloroethane/H2O in the presence of methyltricaprylammmonium chloride, [WZnMn2(ZnW9O34)2]12− was shown to be a very active catalyst for epoxidation of cyclohexene, cyclooctene, and several aliphatic olefins (34). The disubstituted manganese polyoxometalate was also active in toluene/H2O, but not active under monophasic conditions in solvents such as acetonitrile or tert-butyl alcohol. In a subsequent study, it was shown that selectivity for epoxidation over allylic oxidation could be improved by lowering the reaction temperature to 2°C (35). The catalytically active species was not fully characterized, but the available data suggested that it was best described as a tungsten peroxo compound. Another variation of this procedure used Na2WO4, (aminomethyl) phosphonic acid (NH2CH2PO3H2), and methyltri-n-octylammonium hydrogensulfate {[CH3

SCHEME 5. Epoxidation with H3PW12O40 with a phase-transfer catalyst (PTC).

Copyright © 1999 by AOCS Press

Epoxide Synthesis

187

(n-C8H17)3N]HSO4} in a 2:1:1 molar ratio with H2O2 as the oxidant (36). The oxidation can be performed without solvent. This procedure gave high yields except with very acid-sensitive epoxides such as that from styrene in which the yield of epoxide was <25%. Titanosilicates as Catalysts

Zeolites are crystalline aluminosilicate minerals that are used commercially as molecular sieves, drying agents, and catalysts (37). In the early 1980s, it was demonstrated that a zeolitic material containing titanium, known as titanosilicate (TS-1), was a good selective oxidation catalyst capable of converting propylene to propylene oxide and phenol to hydroquinone in the presence of hydrogen peroxide (Scheme 6) (38). The second coordination shell of the titanium atoms in TS-1 is silicon; clusters in the form of TiO2 are not present (39). In titanium-substituted zeolites such as TS1, titanium is uniformly distributed in the crystalline framework by isomorphous substitution of a part of SiIV with TiIV (40). The second line of Scheme 6 shows an example of how this substitution might occur in the case of a boron-containing zeolite; it is speculated that the titanium is inserted into the vacancy created by the loss of boron (39). The complexed titanium reacts either with H2O2 or organic hydroperoxide to produce a reactive complex as shown on the third line of Scheme 6, although an alternative structure for the reactive species has been proposed (41). Because the pore size in TS-1 is small, steric restrictions do not allow penetration by the substrate

SCHEME 6. Epoxidation with titanosilicates. The second line shows the substitution of titanium in a borosilicate. The third line shows the formation of the epoxizing agent.

Copyright © 1999 by AOCS Press

188

G.J. Piazza

when an alkyl hydroperoxide is used. Accordingly TS-1 is an active epoxidation catalyst only with H2O2 and relatively small alkenes. Another titanosilicate catalyst is a titanium-doped zeolite (Ti-β), which is prepared by mixing TiO2, SiO2, tetraethylammonium hydroxide, and hydrogen peroxide, followed by heating (42–44). Ti-β has a three-dimensional pore system containing 12-membered ring apertures. This larger pore size makes Ti-β suitable for branched and cyclic alkenes. Catalytic activity also is observed with organic hydroperoxide oxidants as well as H2O2. A third type of titanosilicate that will be discussed here is based upon an ordered microporous molecular (MCM) sieve (45). These were initially prepared by the calcination of aluminosilicate gels in the presence of surfactants. MCM contain uniform channels, and their dimensions can be varied through the choice of surfactant. MCM containing titanium (Ti-MCM) can be prepared in several ways. One preparation, designated as Ti→MCM41, accommodates the titanium ions within the silica walls (46,47). In another preparation, designated as Ti↑MCM41, tetracoordinated titanium ion was grafted onto the inner surface of the MCM channels (48). Ti-β and Ti→MCM41 have been used as catalysts for the epoxidation of methyl oleate with the use of H2O2 or tert-butyl hydroperoxide (49). Generally Ti-β gave better yields of epoxide than did Ti→MCM41. Acetonitile was a better solvent than methanol, although when Al-free Ti-β was washed with an aqueous solution of sodium cations, good yields of epoxide were obtained even in CH3OH. Higher yields of epoxide could be obtained with Ti→MCM41 when tert-butyl hydroperoxide was used as the oxidant. Research on titanosilicates as well as other forms of immobilized titanium catalysts is rapidly advancing, and novel forms of titanosilicates are still being discovered (50,51). Two recently described advances that may prove useful are the following: (i) tetraneopentyltitanium was reacted with surface hydroxyl groups on several types of silicas. With the use of this catalyst, epoxidation occurred with H2O2 and tert-butyl hydroperoxide (52,53); (ii) Silica was reacted with TiF4; dropwise addition of H2O2 to the defluorinated catalyst led to epoxidation of cyclohexene (54). Lipase-Catalyzed Peracid Formation

Hydrogen peroxide can react with carboxylic acid to form peracid when lipase enzyme is used as a catalyst instead of inorganic acid. The peracid will epoxidize alkenes, which results in the regeneration of the carboxylic acid (Scheme 7). The best results were obtained in the nonpolar solvents toluene and hexane or when mixtures of alkene and carboxylic acid were used without solvent (55). Lipases were immobilized on cellulose, polysulfone membranes, or polypropylene beads (56). The lipases were used to convert caprylic acid to its peracid. The peracid was used to convert oleic acid to its epoxide. The best yield of epoxy stearic acid (81%) was obtained with the lipase from Candida antartica. Using a commercial preparation of immobilized C. antartica lipase, various unsaturated carboxylic acids were treated with

Copyright © 1999 by AOCS Press

Epoxide Synthesis

189

SCHEME 7. Lipase generation of a peracid and its use for epoxidation.

H2O2; yields of epoxide varied from 0 to 91% (57). Formation of dihydroxy acids or estolides did not occur. The lipase was still active after 15 uses. Forty carboxylic acids were used as substrates for the lipase-catalyzed formation of peracids (58). Good agreement between the amount of peracid formed and epoxidation of 1-octene was obtained. Neither α-substituted carboxylic acids nor aromatic acids was converted to peracids. The perhydrolysis of carboxylic acid esters has also been studied. In this reaction, H2O2 reacts with an ester to produce a peracid and an alcohol. Using perhydrolysis, a number of α-substituted peracids could be prepared. Formation of epoxides by running the perhydrolysis procedure in isobutyric acid methyl ester was found to be less expensive than buying peracids. A one-pot synthesis of epoxyalkanolacylates using lipase and H2O2 has been devised (59). The procedure used an unsaturated alcohol and an ethyl ester. Peracid is produced from the ester. The peracid is used for epoxidation, and the lipase also functions as a catalyst for interesterification to convert the alcohol to its corresponding ester. A notable example was the reaction of oleyl alcohol with the ethyl ester of acrylic acid to produce 9,10-epoxystearyl-acrylate, a compound that is not easily produced by chemical methods because of its tendency to polymerize.

Fatty Acid Hydroperoxide as a Source of Epoxy Alcohols Sources of Fatty Acid Hydroperoxide

Polyunsaturated fatty acids react with O2 to give fatty acid hydroperoxides. This reaction is catalyzed by a variety of catalysts. However, a mixture of regio- and stereoisomers is obtained. From a synthetic perspective, the best catalyst is the enzyme lipoxygenase (LOX) because under appropriate conditions, the reaction exhibits an extremely high degree of enantio- and regioselectivity and gives a product with the hydroperoxide moiety in a single position (60). Because LOX requires the presence of a (Z),(Z)-pentadiene moiety, for industrial purposes the use of LOX would be limited to linoleic acid because this fatty acid can be obtained in large amounts from several seed oils. LOX-1 from soybean catalyzes the addition of oxygen to linoleic acid to give the 13-hydroperoxide [13(S)-hydroperoxy-9(Z),11(E)octadecadienoic acid] (61). Another class of LOX found in potato tubers and maize

Copyright © 1999 by AOCS Press

190

G.J. Piazza

kernels converts linoleic acid into the 9-hydroperoxide [9(S)-hydroperoxy10(E),12(Z)-octadecadienoic acid]. Although most LOX enzymes contain ferric iron in their active state, another type of manganese-containing LOX is found in some species (62). From linoleic acid, Mn-LOX from Gaumannomyces graminis produced two hydroperoxides. One product was the 13-hydroperoxide with the R configuration [13(R)-hydroperoxy-9(Z),11(E)-octadecadienoic acid]. The other product resulted from bis-allylic oxygenation of linoleic acid to give the 11-hydroperoxide [11(S)hydroperoxy-9(Z),12(Z)- octadecadienoic acid]. Metal Ion Catalysis of Epoxidation from Hydroperoxide

The catalysis of epoxide formation from organic hydroperoxides by transition metals has been studied extensively (7). Asymmetric epoxidation is conveniently achieved by the Sharpless epoxidation using titanium (63). The hydroperoxide of methyl linoleate [methyl 13(S)-hydroperoxy-9(Z),11(E)-octadecadienoate, Me-HPODE] was prepared using LOX, and Me-HPODE was subjected to the action of vanadyl acetylacetonate [VO(acac)2] (Scheme 8); only the double bond adjacent to the hydroperoxide was epoxidized (64). Treatment with titanium(IV) isopropoxide [Ti(O-i-Pr)4] gave predominantly the threo isomer, i.e., methyl 11(R), 12(R)-epoxy-13(S)hydroxy-9(Z)-octadecenoate (65). Treatment of MeHPODE with niobium(V)ethoxide [Nb(OC2H5)5] gave the erythro isomer, i.e., methyl 11(S), 12(S)-epoxy-13(S)-hydroxy-9(Z)-octadecenoate (66). Multiple structure epoxy alcohols can be generated from fatty hydroperoxides by Fe 2+(67), hematin (68), hemoglobin (69), and myoglobin (70).

SCHEME 8. Formation of a fatty acid hydroperoxide with O2 and lipoxygenase (LOX). The second line shows the conversion of the hydroperoxide to an epoxy alcohol with metal ion catalysts and the enzyme epoxygenase.

Copyright © 1999 by AOCS Press

Epoxide Synthesis

191

Enzymatic Catalysis of Epoxidation from Hydroperoxide

Epoxy alcohol derivatives of arachidonic acids, termed hepoxilins, can have biological activity in mammalian systems. Although hepoxilins can be prepared with catalysis by hemin and hemoglobin, there is an enzymatic route to these compounds as well that has not yet been characterized (71,72). Epoxy alcohols also are obtained by the treatment of fatty acid hydroperoxide with a cyctochrome P450-dependent activity (68), hydroperoxide isomerase (67,73), and epoxygenase (peroxygenase) (74). It is notable that epoxygenase acts only on cis-double bonds (Scheme 8).

Conclusion

As noted in the introduction, epoxidation by the peracid procedure suffers from a number of drawbacks associated with the use of a volatile acid catalyst. Fortunately, as outlined here, research activity in this area has been vibrant, and the processor has been offered a wide variety of new alternatives for fatty epoxide production. References

1. Lutz, J.T., Epoxidation, in Encyclopedia of Chemical Technology, Vol. 9, edited by M. Grayson, John Wiley & Sons, New York, 1980, pp. 251–266. 2. Grundt, E., and G. Stoll, Ketone Production by Epoxide Rearrangement Using Catalyst Comprising Quatranary Ammonium or Phosphonium Ions and Iodide Ions, U.S. Patent 5,245,062 to Henkel KGAA (1993). 3. Hamberg, M., C.A. Herman, and R.P, Herman, Novel Biological Transformation of 15Ls-Hydroperoxy-5,8,11,13-eicosatetraenoic Acid, Biochim. Biophys. Acta 877: 447–457 (1986). 4. Rao, A.S., Addition Reactions with Formation of Carbon-Oxygen Bonds: (i) General Methods of Epoxidation, in Comprehensive Organic Synthesis, Vol. 7, edited by B.M. Trost and I. Fleming, Pergamon Press, Oxford, 1991, pp. 357–387. 5. Mimoun, H., Metal Complexes in Oxidation, in Comprehensive Coordination Chemistry, Vol. 6, Pergamon Press, Oxford, 1987, pp. 317–419. 6. Meunier, B., Metalloporphyrins as Versatile Catalysts for Oxidation Reactions and Oxidative DNA Cleavage, Chem. Rev. 92: 1411–1456 (1992). 7. Jørgensen, K.A., Transition-Metal-Catalyzed Epoxidations, Chem. Rev. 89: 431–458 (1989). 8. Katsuki, T., Mn-Salen Catalyst, Competitor of Enzymes, for Asymmetric Epoxidation, J. Mol. Cat. A: Chem. 113: 87–107 (1996). 9. Murray, R.W., Dioxiranes, Chem. Rev. 89: 1187–1201 (1989). 10. Singh, M., and R.W. Murray, Chemistry of Dioxiranes. 21. Thermal Reactions of Dioxiranes, J. Org. Chem. 57: 4263–4270 (1992). 11. Adam, W., and A.K. Smerz, Solvent Effects in the Regio- and Diastereoselective Epoxidations of Acyclic Allylic Alcohols by Dimethyldioxirane: Hydrogen Bonding as Evidence for a Dipolar Transition State, J. Org. Chem. 61: 3506–3510 (1996). 12. Sonnet, P.E., M.E. Lankin, and G.P. McNeill, Reaction of Dioxiranes with Selected Oleochemicals, J. Am. Oil Chem. Soc. 72: 199–204 (1995).

Copyright © 1999 by AOCS Press

192

G.J. Piazza

13. Sonnet, P.E., and T.A. Foglia, Epoxidation of Natural Triglycerides with Ethylmethyldioxirane, J. Am. Oil Chem. Soc. 73: 461–464 (1996). 14. Foglia, T.A., P.E. Sonnet, A. Nuñez, and R.L. Dudley, Selective Oxidations of Methyl Ricinoleate: Diastereoselective Epoxidation with TitaniumIV Catalysts, J. Am. Oil Chem. Soc. 75: 601–607 (1998). 15. Lie Ken Jie, M.S.F., and M.K. Pasha, Epoxidation Reactions of Unsaturated Fatty Esters with Potassium Peroxomonosulfate, Lipids 33: 633–637 (1998). 16. Nam, W., H.J. Kim, S.H. Kim, R.Y.N. Ho, and J. Selverstone Valentine, Metal ComplexCatalyzed Epoxidation of Olefins by Dioxygen with Co-Oxidation of Aldehydes. A Mechanistic Study, Inorg. Chem. 35: 1045–1049 (1996). 17. Hamamoto, M., K. Nakayama, Y. Nishiyama, and Y. Ishii, Oxidation of Organic Substrates by Molecular Oxygen/Aldehyde/Heteropolyoxometalate System, J. Org. Chem. 58: 6423–6425 (1993). 18. Kuo, M.-C, and T.-C. Chou, Kinetics and Mechanism of the Catalyzed Epoxidation of Oleic Acid with Oxygen in the Presence of Benzaldehyde, Ind. Eng. Chem. Res. 26: 277–284 (1987). 19. Kuo, M.-C, and T.-C. Chou, Epoxidation of Oleic Acid with Oxygen in the Presence of Benzaldehyde Using Heterogenized Homogeneous Co-Type Ion-Exchange Membrane as Catalyst, Can, J. Chem. Eng. 68: 831–838 (1990). 20. Pietikäinen, P., Convenient Asymmetric (Salen)Mn(III)-Catalyzed Epoxidation of Unfunctionalized Alkenes with Hydrogen Peroxide Using Carboxylate Salt Cocatalysts, Tetrahedron 54: 4319–4326 (1998). 21. Campbell, L.A., and T. Kodadek, Enantioselective Catalysis of Epoxidation by Metalloporphyrins: Application to Enantioselective Synthesis, J. Mol. Cat. A: Chem. 113: 293–310 (1996). 22. Chou, T.-C., and S.-V. Lee, Epoxidation of Oleic Acid in the Presence of Benzaldehyde Using Cobalt (II) Tetraphenylporphyrin as Catalyst, Ind. Eng. Chem. Res. 36: 1485– 1490 (1997). 23. Herrmann, W.A., R.W. Fischer, and D.W. Marz, Methyltrioxorhenium as Catalyst for Olefin Oxidation, Angew. Chem. Int. Ed. Engl. 30: 1638–1641 1991). 24. Herrmann, W.A., R.W. Fischer, W. Scherer, and M.U. Rauch, Methyltrioxorhenium (VII) as Catalyst for Epoxidations: Structure of the Active Species and Mechanism of Catalysis, Angew. Chem. Int. Ed. Eng. 32: 1157–1160 (1993). 25. Al-Ajlouni, A.M., and J.H. Espenson, Expoxidation of Styrenes by Hydrogen Peroxide as Catalyzed by Methylrhenium Trioxide, J. Am. Chem. Soc. 117: 9243–9250 (1995). 26. Al-Ajlouni, A.M., and J.H. Espenson, Kinetics and Mechanism of the Epoxidation of Alkyl-Substituted Alkenes by Hydrogen Peroxide, Catalyzed by Methylrhenium Trioxide, J. Org. Chem. 61: 3969–3976 (1996). 27. Adam, W., and CM. Mitchell, Methyltrioxorhenium (VII)-Catalyzed Epoxidation of Alkenes with the Urea/Hydrogen Peroxide, Angew. Chem. Int. Ed. Engl. 35: 533–535 (1996). 28. Rudolph, J., K.L. Reddy, J.P. Chiang, and K.B. Sharpless, Highly Efficient Epoxidation of Olefins Using Aqueous H2O2 and Catalytic Methyltrioxorhenium/Pyridine: Pyridine-Mediated Ligand Acceleration, J. Am. Chem. Soc. 119: 6189–6190 (1997). 29. Tan, H., and J.H. Espenson, Kinetics and Mechanism of the Dihydroxylation and Epoxidation of Conjugated Dienes with Hydrogen Peroxide Catalyzed by Methylrhenium Trioxide, Inorg. Chem. 37: 467–472 (1998).

Copyright © 1999 by AOCS Press

Epoxide Synthesis

193

30. Venturello, C., and R. D’Aloisio, Quaternary Ammonium Tetrakis (diperoxotungsto) phosphates (3-) as a New Class of Catalysts for Efficient Alkene Epoxidation with Hydrogen Peroxide, J. Org. Chem. 53: 1553–1557 (1988). 31. Ishii, V., K. Yamawaki, T. Ura, H. Yamada, T. Yoshida, and M. Ogawa, Hydrogen Peroxide Oxidation Catalyzed by Heteropoly Acids Combined with Cetylpyridinium Chloride: Epoxidation of Olefins and Allylic Alcohols, Ketonization of Alcohols and Diols, and Oxidative Cleavage of 1,2-Diols and Olefins, J. Org. Chem. 53: 3587–3593 (1988). 32. Duncan, D.C., R.C. Chambers, E. Hecht, and C.L. Hill, Mechanism and Dynamics in the H3[PW12O40]-Catalyzed Selective Epoxidation of Terminal Olefins by H2O2. Formation, Reactivity, and Stability of {PO4[WO(O2)2]4}3-, J. Am. Chem. Soc. 117: 681–691 (1995). 33. Yadav, G.D., and D.V. Satoskar, Kinetics of Epoxidation of Alkyl Esters of Undecylenic Acid: Comparison of Traditional Routes vs. Ishii-Venturello Chemistry, J. Am. Oil Chem. Soc. 74: 397–407 (1997). 34. Neumann, R., and G. Mohammad, Highly Active Manganese-Containing Polyoxometalate as Catalyst for Epoxidation of Alkenes with Hydrogen Peroxide, J. Am. Chem. Soc. 116: 5509–5510 (1994). 35. Neumann, R., and G. Mohammad, The Manganese-Containing Polyoxometalate, [WznMn2(ZnW9O34)2]−12, as a Remarkably Effective Catalyst for Hydrogen Peroxide Mediated Oxidations, J. Am. Chem. Soc. 117: 5066–5074 (1995). 36. Sato, K., A. Masao, M. Ogawa, T. Hashimoto, and R. Noyori, A Practical Method for Epoxidation of Terminal Olefins with 30% Hydrogen Peroxide Under Halide-Free Conditions, J. Org. Chem. 61: 8310–8311 (1996). 37. Freemantle, M., Layered Titanolsilicate Catalyst Has Five-Coordinate Titanium Atoms, C&E News 74: 47–49 (1996). 38. Serrano, D.P., H.-X. Li, and M.E, Davis, Synthesis of Titanium-Containing ZSM-48, J. Chem. Soc. Chem. Commun.: 745−747 (1992). 39. Dartt, C.B., and M.E. Davis, Characterization and Catalytic Activity of Titanium Containing SSZ-33 and Aluminum-Free Zeolite Beta, Appl. Catal. A: General 143: 53−73 (1996). 40. Hütter, R., T. Mallat, and A, Baiker, Titania-Silica Mixed Oxides, J. Catal. 153: 177–189 (1995). 41. Adam, W., A. Corma, A. Martínez, C.M. Mitchell, T.I. Reddy, M. Renz, and A.K. Smerz, Diastereoselective Epoxidation of Allylic Alcohols with Hydrogen Peroxide Catalyzed by Ttitanium-Containing Zeolites or Methyltrioxorhenium Versus Stoichiometric Oxidation with Dimethyldioxirane: Clues on the Active Species in the Zeolite Lattice, J. Mol. Catal. A: Chem. 117: 357–366 (1997). 42. Murugavel, R., and H.W. Roesky, Titanosilicates: Recent Developments in Synthesis and Use as Oxidation Catalysts, Angew. Chem. Int. Ed. Engl. 36: 477–419 (1997). 43. Camblor, M.A., M. Costantini, A. Corma, L. Gilbert, P. Esteve, A. Martínez, and S. Valencia, Synthesis and Catalytic Activity of Aluminium-Free Zeolite Ti-Oxidation Catalysts, J. Chem. Soc. Chem. Commun., 1339–1340 (1996). 44. Blasco, T., M.A. Camblor, A. Corma, P. Esteve, J.M. Guil, A. Martínez, J.A. Perdigón Melón, and S. Valencia, Direct Synthesis and Characterization of Hydrophobic Aluminum-Free Ti-Beta Zeolite, J. Phys. Chem. B 102: 75–88 (1998). 45. Kresge, C.T., M.E. Leonowicz, W.J. Roth, J.C. Vartuli, and J. S. Beck, Ordered Mesoporous Molecular Sieves Synthesized by a Liquid-Crystal Template Mechanism, Nature 359: 710–712 (1992).

Copyright © 1999 by AOCS Press

194

G.J. Piazza

46. Corma, A., M.T. Navarro, and J. P. Pariente, Synthesis of an Ultralarge Pore Titanium Silicate Isomophous to MCM-41 and Its Application as a Catalyst for Selective Oxidation of Hydrocarbons, J. Chem. Soc, Chem. Commun., 147–148 (1994). 47. Oldroyd, R.D., J.M. Thomas, T. Maschmeyer, P.A. MacFaul, D.W. Snelgrove, K.U. Ingold, and D.D.M. Wayner, The Titanium(IV)-Catalyzed Epoxidation of Alkenes by tertAlkyl Hydroperoxides, Angew. Chem. Int. Ed. Engl. 35: 2787–2790 (1996). 48. Maschmeyer, T., F. Rey, G. Sankar, and J.M. Thomas, Heterogeneous Catalysts Obtained by Grafting Metallocene Complexes onto Mesoporous Silica, Nature 378: 159–162 (1995). 49. Camblor, M.A., A. Corma, P. Esteve, A. Martínez, and S. Valencia, Epoxidation of Unsaturated Fatty Esters Over Large-Pore Ti-Containing Molecular Sieves as Catalysts: Important Role of the Hydrophobic-Hydrophilic Properties of the Molecular Sieve, J. Chem. Soc. Chem. Commun., 795–796 (1997). 50. Karjalainen, J.K., O.E.O. Hormi, and D.C. Sherrington, Highly Efficient Heterogeneous Polymer-Supported Sharpless Alkene Epoxidation Catalysts, Tetrahedron Asymmetry 9: 1563–1575 (1998). 51. Karjalainen, J.K., O.E.O. Hormi, and D.C. Sherrington, Heterogeneous Asymmetric Epoxidation of cis-Allylic Alcohols: Use of Polymer-Supported Ti(IV)-Catalyst, Tetrahedron Asymmetry 9: 2019–2022 (1998). 52. Holmes, S.A., F. Quignard, A. Choplin, R. Teissier, and J. Kervennal, Tetraneopentyltitanium Derived Supported Catalysts. Part 1. Synthesis and Catalytic Properties for the Epoxidation of Cyclohexene with Aqueous Hydrogen Peroxide, J. Catal. 176: 173–181 (1998). 53. Holmes, S.A., F. Quignard, A. Choplin, R. Teissier, and J. Kervennal, Tetraneopentyltitanium Derived Supported Catalysts. Part 2. Influence of the Hydrophobic Character of Silica on the Catalytic Properties, J. Catal. 176: 182–191 (1998). 54. Jorda, E., A. Tuel, R. Teissier, and J. Kervennal, Synthesis, Characterization, and Activity in the Epoxidation of Cyclohexene with Aqueous H2O2 of Catalysts Prepared by Reaction of TiF4 with Silica, J. Catal. 175: 93–107 (1998). 55. Björkling, F., S.E. Gadtfredsen, and O. Kirk, Lipase-Mediated Formation of Peroxycarboxylic Acid Used in Catalytic Epoxidation of Alkenes, J. Chem. Soc. Chem. Commun., 1301–1303 (1990). 56. Cuperus, F.P., S.T. Bouwer, G.F.H. Kramer, and J.T.P. Derksen, Lipases Used for the Production of Peroxycarboxylic Acids, Biocatalysis 9: 89–96 (1994). 57. Warwel, S., and M. Rüsch gen. Klaas, Chemo-Enzymatic Epoxidation of Unsaturated Carboxylic Acids, J. Mol. Cat. B: Enz. I: 29–35 (1995). 58. Rüsch gen. Klaas, M., and S. Warwel, Lipase-Catalyzed Preparation of Peroxy Acids and Their Use for Epoxidation, J. Mot. Cat. A: Chem. 117: 311–319 (1997). 59. Rüsch gen. Klaas, M., and S. Warwel, A Three-Step-One-Pot Chemo-Enzymatic Synthesis of Epoxyalkanolacylates, Syn. Comm. 28: 251–260 (1998). 60. Solomon, E.I., J. Zhou, F. Neese, and E.G. Pavel, New Insights from Spectroscopy into the Structure/Function Relationships of Lipoxygenases, Chem. Biol. (Land.) 4: 795– 808 (1997). 61. Gardner, H.W., Recent Investigations into the Lipoxygenase Pathway of Plants, Biochim. Biophys. Acta 1084: 221–239 (1991). 62. Hamberg, M., C. Su, and E. Oliw, Manganese Lipoxygenase. Discovery of a bis-Allylic Hydroperoxide as Product and Intermediate in a Lipoxygenase Reaction, J. Biol. Chem. 273: 13080–13088 (1998).

Copyright © 1999 by AOCS Press

Epoxide Synthesis

195

63. Johnson, R.A., and K.B. Sharpless, Addition Reactions with Formation of Carbonoxygen Bonds: (ii) Asymmetric Methods of Epoxidation, in Comprehensive Organic Synthesis, Vol. 7, edited by B.M. Trost and I. Fleming, Pergamon Press, Oxford, 1991, pp. 389–435. 64. Hamberg, M., Vanadium-Catalyzed Transformation of 13(S)-Hydroperoxy-9(Z),11(E)Octadecadienoic Acid: Structural Studies on Epoxy Alcohols and Trihydroxy Acids, Chem. Phys. Lipids 43: 55–67 (1987). 65. Piazza, G.J., T.A. Foglia, and A. Nuñez, Enantioselective Formation of an α,β-Epoxy Alcohol by Reaction of Methyl 13(S)-Hydroperoxy-9(Z), 11(E)-octadecadienoate with Titanium Isopropoxide, J. Am. Oil Chem. Soc. 74: 1385–1390 (1997). 66. Piazza, G.J., T.A. Foglia, and A. Nuñez, Enantioselective Conversion of Linoleate Hydroperoxide to an α,β-Epoxy Alcohol by Niobium Ethoxide, J. Am. Oil Chem. Soc. 75: 939–943 (1998). 67. Gardner, H.W., Lipoxygenase as a Versatile Biocatalyst, J. Am. Oil Chem. Soc. 73: 1347– 1357 (1996). 68. Chang, M.S., W.E. Boeglin, F.P. Guengerich, and A.R. Brash, Cytochrome P450-Dependent Transformations of 15R- and 15S-Hydroperoxyeicosatetraenoic Acids: Stereoselective Formation of Epoxy Alcohol Products, Biochemistry 35: 464–471 (1996). 69. Hamberg, M., Decomposition of Unsaturated Fatty Acid Hydroperoxides by Hemoglobin: Structures of Major Products of 13L-Hydroperoxy-9,11-octadecadienoic Acid, Lipids 10: 87–92 (1975). 70. Hamberg, M., Myoglobin-Catalyzed bis-Allylic Hydroxylation and Epoxidation of Linoleic Acid, Arch. Biochem. Biophys. 344: 194–199 (1997). 71. Pace-Asciak, C.R., D. Reynaud, and P. Demin, Mechanistic Aspects of Hepoxilin Biosynthesis, J. Lipid Mediat. Cell Signal. 12: 307–311 (1995). 72. Pace-Asciak, C.R., D. Reynaud, and P. Demin, Hepoxilins: A Review on Their Enzymatic Formation, Metabolism and Chemical Synthesis, Lipids 30: 107–114 (1995). 73. Hamberg, M., Stereochemical Aspects of Fatty Acid Oxidation: Hydroperoxide Isomerases, Acta Chem. Scand. 50: 219–224 (1996). 74. Blée, E., Phytooxylipins: The Peroxygenase Pathway, in Lipoxygenase and Lipoxygenase Pathway Enzymes, edited by G.J. Piazza, AOCS Press, Champaign, IL, 1996, pp. 138–161.

Copyright © 1999 by AOCS Press

Chapter 12

Cyclic Fatty Acids in Heated Vegetable Oils: Structures and Mechanisms Gary Dobson

Scottish Crop Research Institute, Invergowrie, Dundee DD2 5DA

Introduction

During heating of vegetable oils, such as during frying, triacylglycerols undergo a variety of chemical changes including oxidation and, in the absence of oxygen, formation of new carbon-carbon bonds (1). New bonds between two different fatty acids can result in dimeric acids, either within or between triacylglycerol molecules, and further cross-linking produces polymers. Cyclic fatty acids are the products of new bond formation within a fatty acid molecule. The nature of cyclic fatty acid monomers (CFAM) has been the subject of study for over 40 years (2–5). Interest in CFAM was driven by the finding that feeding CFAM to pregnant rats led to increased mortality in the offspring (3). CFAM are formed in low amounts (0.01–0.7%) during deep frying (3) and inevitably constitute part of the human diet. They have been detected in a human milk sample (6), but the effect on humans is unknown. A general review on cyclic fatty acids by Sébédio and Grandgirard (3) included substantial sections on the structures, occurrence, quantification, and biological effects of CFAM in frying oils, and synthesis of model compounds. Subsequently, Le Quéré and Sébédio (4) updated the structural, quantitative, and model compound aspects, and a review by Dobson (5) focused particularly on the considerable advances that have been made recently in structural characterization of CFAM, as well as quantitative analysis. These reviews should be consulted for a more comprehensive treatment of the subject. This article will concern itself with an overview of the most recent literature on structural aspects and possible mechanisms for the formation of CFAM.

Structural Studies on Cyclic Fatty Acids

CFAM in heated vegetable oils occur as complex mixtures of acids, each with 18 carbons and either a five- or six-membered ring. The position of the ring, the geometry of the side chains about the ring, and the number, position (including within the ring or in the side chains) and geometry of double bonds are all variables that give a large number of potential structures (Fig. 1). The degree of unsaturation of the acid reflects the origin. Saturated (7), monoenoic (8,9), and dienoic (9–13) acids are derived from oleic, linoleic, and α-linolenic acids, respectively, with one double bond, 196 Copyright © 1999 by AOCS Press

Cyclic Fatty Acids in Heated Vegetable Oils

197

Fig. 1. Monomeric cyclic fatty acids derived from heated vegetable oils: I–IV, cyclic dienes from α-linolenate; V–XIV, cyclic monoenes from linoleate; XV–XVIII, monocyclic saturated acids from oleate; XIX–XXII, bicyclic saturated acids from linoleate (the dotted lines indicate that the positions of the bonds within the rings were not determined); XXIII–XXVI, cyclic dienes from γ-linolenate.

lost upon cyclization. It follows that the type of vegetable oil affects the nature of CFAM. For example, oleic acid is the major acid in olive oil and high-oleic sunflower oil, linoleic is the predominant acid in normal sunflower and soybean oils, and αlinolenic acid is a significant component of soybean and rapeseed oils. Using linseed oil and high-oleic and normal sunflower oils as models, only recently have the complete CFAM structures from the three unsaturated precursors

Copyright © 1999 by AOCS Press

198

G. Dobson

been elucidated. Typically, oils were heated at 275°C for 12 h under nitrogen; a CFAM fraction was isolated by preparing methyl esters from the heated oil, collecting a nonpolar fraction by column chromatography, and after two urea treatments, collecting the nonadducts (14). Further purification of CFAM from heated sunflower oil, using reversed-phase high-performance liquid chromatography (RP-HPLC), was necessary. About 10 times the amount of CFAM was recovered from linseed oil (11.1% CFAM of total fatty acids) compared with sunflower oil (1.0% CFAM), reflecting the relative ease of cyclization of α-linolenic acid compared with linoleic acid (14). The CFAM fractions were complex mixtures. The degree of unsaturation of the acids was determined by gas chromatography-mass spectrometry (GC-MS) of the methyl esters. Although some structures were proposed by comparison to the mass spectra of synthetic standards (9), it was difficult to obtain structural information from the mass spectra of the methyl esters. The mixtures could be simplified by hydrogenation, which eliminated positional and geometrical double bond isomers. The skeletal structures were then elucidated by GC-MS of the methyl esters; some were confirmed by comparison to synthetic standards (15). A number of additional skeletal CFAM structures from heated linseed and sunflower oils were proposed with the use of a novel fast-atom bombardment (FAB) GC-MS/MS approach (16). Gasphase carboxylate ions were generated from electron-capture ionization of hydrogenated CFAM pentafluorobenzyl esters, and collisionally activated dissociation (CAD) mass-analyzed ion kinetic energy (MIKE) spectra of the carboxylate ion were characterized by charge-remote fragmentation to the ring. An on-line hydrogenation method in conjunction with GC-MS was developed to correlate the mass spectra of the methyl esters of intact CFAM and their hydrogenated analogs (17). Ten CFAM from heated linseed oil were examined, and some tentative dienoic structures, with either cyclopentenyl or cyclohexenyl rings and a double bond of unknown location in the side chain, were suggested. The geometrical configurations of the ring isomers of the major hydrogenated CFAM from sunflower and linseed oils (15) were determined by comparison of GC retention times to those of synthetic standards (18,19). In all cases, the trans isomer eluted before the cis isomer. The geometrical configurations of the double bonds of a number of CFAM from heated sunflower and linseed oils were first determined by Sébédio and colleagues by GC-Fourier transform infrared (FTIR) spectroscopy of the methyl esters (9). More recently, the technique was used in conjunction with other techniques (see below) to obtain full structural information on CFAM from sunflower oil (8) and linseed oil (10,12,20,21). Mossoba and colleagues have examined methyl esters (12,20) and DMOX derivatives (21) by GC-matrix isolation (MI)-FTIR spectroscopy to obtain greater sensitivity. Specific stretch and deformation bands for Z double bonds in cyclopentenyl and cyclohexenyl rings (E double bonds cannot occur in rings) were evident, and a strong deformation band around 970 cm−1 was highly characteristic for E double bonds in the chain. A complete structural characterization of the CFAM from heated oils was not

Copyright © 1999 by AOCS Press

Cyclic Fatty Acids in Heated Vegetable Oils

199

possible because of the complexities of the mixtures and the absence of a method that would give full structural information, especially with respect to double bond positions. Two advances have now made this possible. First, a CFAM mixture (obtained by urea adduction) can be separated into simpler fractions, containing only a few acids, by silver-ion HPLC. Second, the positions of the double bonds, and the types and positions of the rings, can be determined by GC-MS analysis of the picolinyl esters or DMOX derivatives of the fractions (7,8,10). In conjunction with GC-FTIR (to determine the geometrical configurations of the double bonds) and GC of the hydrogenated acid methyl esters (to elucidate the geometrical configurations of the side chains about the rings), complete structural elucidation is now possible. The CFAM were separated by silver-ion HPLC as phenacyl esters (because of improved separation compared with methyl esters), according to the degree of unsaturation and geometrical configurations of the double bonds (E isomers eluted before Z isomers) (8,10) and also sometimes on the basis of sizes of the rings and configurations of the side chains about the rings (10). In this way, several fractions, each containing only a limited number of usually well resolved GC components, some of which were unresolved from other peaks in the total mixture, were obtained. Each fraction was then analyzed by GC-MS of either the picolinyl ester or DMOX derivative. In general, the mass spectra of the two types of derivatives were analogous and could be easily interpreted without the need for standards by using a few basic principles. The skeletal structures were confirmed by GC-MS of the derivatives of the hydrogenated fractions, and GC-MS analyses of the derivatized deuterated fractions were particularly useful for verifying positions of double bonds in the chains. The above approach was used to study the CFAM from linseed oil heated to 275°C; these were shown to be a complex mixture by GC (Fig. 2) (10,22). They were separated into nine fractions by silver-ion HPLC, and the cyclic dienes occurred in fractions 3–8 (Fig. 3). Each fraction contained up to five acids; by applying the various analytical techniques outlined above and correlating all of the information, 16 cyclic dienoic acids were characterized. There were about equal amounts of eight cyclopentenyl (a–h, Fig. 2) and eight cyclohexenyl (i–p) acids. There were two basic cyclopentenyl structures, one with a ring from C-10 to C-14 of the parent α-linolenic acid and a double bond at C-15 (Structure I, Fig. 1) and the other with a ring from C-11 to C-15 and a double bond at C-9 (II). Therefore the double bond in the straight chain remained at the original position, that in the ring presumably remained at the C-12 position (and must have Z configuration), and one double bond was lost upon ring formation. There were two basic cyclohexenyl structures, both with rings between C-10 and C-15, one with a double bond at C-8 (III) and the other at C-16 (IV), having shifted from the original positions at C-9 and C-15, respectively. Again the double bond in the ring remained at C-12 and one double bond was lost upon cyclization. The carbon atoms involved in cyclization are summarized in Figure 4A. There were four acids of each basic structure, representing all possible combinations of geometrical configurations of the double bond in the straight chain and of the straight chains about the ring. There were similar amounts of cyclopentenyl acids with Z and

Copyright © 1999 by AOCS Press

200

G. Dobson

Fig. 2. Partial gas chromatogram of total cyclic fatty acid methyl esters derived from α-linolenic acid. A Cp-Wax 52CB (25 m × 0.25 mm i.d., 0.20 µm film thickness) capillary column was used. There was an initial temperature of 160°C for 5 min followed by a program at 0.5°C/min to 180°C. Peaks a–h were cyclopentenyl acids and i–p were cyclohexenyl acids. Source: Ref. 22.

E double bonds, and a predominance of trans ring isomers. There was a predominance of cyclohexenyl acids with E double bonds, presumably as a result of the double bond shifting position, and similar amounts of both ring isomers. The mass spectrum of a picolinyl ester of a cyclopentenyl acid with a ring from C-11 to C-15 and a cis double bond at C-9 [10-(2′-propyl-cyclopentenyl) dec-cis-9enoate] is given as an example (Fig. 5A). The molecular ion was at m/z 369; regular gaps of 14 amu from m/z 164 to 234 indicated a saturated straight chain up to C-8; a gap of 66 amu between m/z 260 and 326 located the cyclopentenyl ring between C11 and C-15. A gap of 26 amu between m/z 234 and 260 suggested a double bond at C-9. There were no diagnostic ions for locating the position of the double bond in the ring, but it was assumed that it remained at the C-12 position. In the mass spectrum of the hydrogenated ester, the molecular ion was 4 amu higher at m/z 373, and a gap of 68 amu between m/z 262 and 330 verified the position of the ring. In the spectrum of the deuterated ester, the molecular ion was a further 4 amu higher at m/z 377; a gap of 70 amu between m/z 264 and 334 verified two deuterium atoms in the ring, and gaps of 15 amu from m/z 234 to 249 and from m/z 249 to 264 verified deuterium

Copyright © 1999 by AOCS Press

Cyclic Fatty Acids in Heated Vegetable Oils

201

Fig. 3. Silver-ion high-performance liquid chromatogram of phenacyl esters of cyclic fatty acids derived from heated linseed oil. A column of Nucleosil 5SA (250 × 4.6 mm i.d.; 5 µm particle size) was used in the silver ion form. The mobile phase was composed of dichloromethane-dichloroethane (50:50 vol/vol; Solvent A) and dichloromethane-dichloroethane-acetonitrile (49:49:2, by vol; Solvent B); the flow rate was 1 mL/min. There was a linear gradient from 100% A to 75% A/25% B over 50 min, then to 100% B over a further 5 min. An evaporative light-scattering detector was employed. Source: Ref. 10.

atoms at C-9 and C-10. The mass spectra of the derivatives of the other cyclopentenyl acid (Structure I, Fig. 1) and the cyclohexenyl acid with a double bond at C-8 (III) were also readily interpretable in a similar way. One additional feature in all cyclohexenyl spectra was the presence of an ion (m/z 315 for picolinyl esters and m/z 277 for DMOX derivatives) due to a retro Diels-Alder fragmentation that could arise only if the double bond in the ring was located at C-12. Interpretation of the mass spectra of derivatives of the cyclohexenyl acids with a double bond at C-16 [e.g., 9-(2′-prop-trans-1-enyl-cyclohex-cis-4-enyl)nonanoate; Fig. 5B] was more difficult. A gap of 78 amu between m/z 248 and 326 suggested a cyclohexadiene structure, yet the presence of a retro Diels-Alder fragment (m/z 315) indicated only one double bond in the ring. The mass spectrum of the deuterated ester had gaps of 84 amu from m/z 248 to 332 and of 30 amu from m/z 332 to 362, clearly showing that there was only one double bond in the ring with the other at C-

Copyright © 1999 by AOCS Press

202

G. Dobson

Fig. 4. Carbon atoms involved in cyclization of (A) α-linolenic acid, (B) linoleic acid, and (C) oleic acid. Source: Ref. 28.

16. This example clearly illustrates the necessity for examining the deuterated as well as the native acid derivatives. Mossoba and colleagues (11,12), studying a similar CFAM sample from linseed oil but analyzing the DMOX derivatives of the total mixture by GC-MS, characterized 13 out of the 16 acids. There was agreement with the previous study (10) for the majority of acids. However, two acids were claimed to have a cyclohexadiene structure, but we believe that the true structures had cyclohexenyl rings between C-10 and C-15 and Z double bonds at C-16 (10). Two C18 bicyclic acids with a six-membered ring fused to a five-membered ring, containing one double and with either a methyl or an ethyl substituent on the ring, were also characterized (12,13). The cyclic monoenoic acids from linoleate in sunflower oil heated to 275°C were analyzed by the silver-ion HPLC/GC-MS approach outlined above (8). Several acids were characterized; the carbon atoms involved in cyclization are illustrated in Figure 4B. The major CFAM were two cyclopentenyl acids, one with the ring from C-10 to C-14 (V, Fig. 1) and the other from C-8 to C-12 (VI), present in similar amounts. The positions of the double bonds in the rings could not be confirmed from the mass spectra but were presumably in the original C-12 and C-9 positions, respectively. Only one geometrical ring isomer of each acid was detected; on the basis of a study on the hydrogenated acids of a similar mixture (15), it can be assumed that the configuration was trans. There were minor cyclopentenoic acids with rings between C9 and C-13 (VII). Two acids (VIII, IX) with structures similar to the major ones were detected, but because each gave two products (presumably both ring isomers)

Copyright © 1999 by AOCS Press

Cyclic Fatty Acids in Heated Vegetable Oils

203

Fig. 5. Mass spectra (70 eV) of picolinyl ester derivatives of cyclic dienoic acids derived from α-linolenic acid in heated linseed oil: (A) 10-(2′-propyl-cyclopentenyl)dec-cis-9-enoate and (B) 9-(2′-prop-trans-1-enyl-cyclohex-cis-4enyl)nonanoate. Source: Ref. 10.

upon hydrogenation, it was postulated that the double bonds must be at one of the two substituted carbons. These are the only examples in which the double bond in the ring has shifted from the original position in the parent straight-chain acid. Interestingly, CFAM with cyclohexenyl rings were not detected. Of the remaining acids, there were several with a cis double bond in the straight chain at either C-9 or C-12 and either cyclopentyl [from C-13 to C-17 (X) or from C-5 to C-9 (XI), respectively] or cyclohexyl [from C-12 to C-17 (XII)/ C-13 to C-18 (XIII) or from C-5 to C-10 (XIV), respectively] rings. Analogous CFAM with a trans double bond occurred only as cyclohexyl (from C-5 to C-10 and C-12 to C-17) acids.

Copyright © 1999 by AOCS Press

204

G. Dobson

The saturated CFAM, isolated by silver-ion HPLC from high-oleate and normal sunflower oils used in a small-scale frying operation of potato strips (7,23), were shown by GC to be composed of two different groups of CFAM (Fig. 6). The first group (q-x) was presumably derived from oleate and constituted a greater proportion of the total CFAM in high-oleate (~67%) compared with normal (~25%) sunflower

Fig. 6. Partial gas chromatograms of saturated cyclic fatty acid methyl esters derived from (A) high-oleate sunflower oil and (B) normal sunflower oil. A Cp-Wax 52CB (25 m × 0.25 mm i.d., 0.20 µm film thickness) capillary column was used. There was an initial temperature of 170°C for 5 min followed by a program at 2°C/min to 210°C. Peaks q–x were monocyclic acids, and later eluting peaks had bicyclic structures. Source: Ref. 7.

Copyright © 1999 by AOCS Press

Cyclic Fatty Acids in Heated Vegetable Oils

205

oil (7). Eight monocyclic acids were comprised of four basic structures. There were two cyclopentyl acids with a ring either from C-5 to C-9 (XV, Fig. 1) or from C-10 to C-14 (XVI), and two cyclohexyl acids with a ring either from C4 to C-9 (XVII) or from C-10 to C-15 (XVIII), each represented by two ring isomers. The carbon atoms involved in cyclization are summarized in Figure 4C. The cyclopentyl acids were about twice as abundant as the cyclohexyl acids. The second group of peaks (Fig. 6) comprised a greater proportion of the total CFAM in normal (~14%) than in high-oleate (~4%) sunflower oil and was probably derived from linoleate (23). There were several peaks with a molecular weight 2 amu less than that of saturated monocyclic acids; they were unchanged after hydrogenation, confirming their saturated nature. The mass spectra of the DMOX derivatives and picolinyl esters suggested the presence of five- and six-membered rings with internal ring bonds to give bicyclic structures; the positions of the bonds could not be determined from the mass spectra. Again there were four basic structures, i.e., two five-membered ring [from C-8 to C-12 (XIX) and C-10 to C-14 (XX)] acids and two six-membered ring [from C-7 to C-12 (XXI) and C-10 to C-15 (XXII)] acids. The acids could occur in a number of different forms, possibly stereoisomers and/or isomers with different positions of the internal carbon-carbon bond in the ring. The cyclohexyl acids were more abundant. The cyclopentyl acids had structures analogous to the major cyclopentenyl CFAM (V and VI) derived from linoleate in laboratoryheated sunflower oil (8), and a common origin was suggested (23). Several CFAM were characterized from a lightly partially hydrogenated soybean oil, both before and after heating, by GC-MS of the methyl esters of nonurea adduct mixtures after hydrogenation to simplify the mixtures (24). The mixtures would be expected to be particularly complicated because oleate (45%), linoleate (37%), and α-linoleate (2%) were all present in the starting oil. The profile may have been complicated even more by CFAM from geometrical and positional double bond isomers formed by partial hydrogenation. Some of the expected products (cyclohexyl acids with rings from C-10 to C-15 and cyclopentyl acids with rings from C-10 to C-14 and C-11 to C-15) were detected, but others, for example, cyclopentyl acids with rings from C-8 to C-12 (a major product from linoleate), were not found. Conversely, some cyclohexyl acids (with C-8 to C-13, C-9 to C-14, and C-11 to C-16 rings) were identified, but their nonhydrogenated counterparts have not been encountered. There are few data on the types of CFAM from oils used under real frying conditions. The studies, cited above, of saturated monocyclic (7) and bicyclic acids (23) formed in sunflower oils are examples. In another study (25), the CFAM from peanut and soybean oils, used in the frying of frozen prefried french fries, were analyzed after hydrogenation. The types of CFAM were essentially similar between the two oils, with differences only in the relative proportions of components. This observation may seem surprising because the fatty acid compositions of the two oils differ (only soybean oil contains significant amounts of α-linolenic acid); this may be explained in part by the derivation of the same hydrogenated CFAM from different intact CFAM. However, the detection of an acid with a cyclopentyl ring between C-11 and

Copyright © 1999 by AOCS Press

206

G. Dobson

C-15 in peanut oil was anomalous because it would appear that this acid could be derived only from α-linolenic acid. Further studies on oils under real frying conditions using the analytical methodology outlined above for model heating experiments would be beneficial. There is also limited information on the effect of heating conditions on CFAM composition. In one study, the same types of CFAM seemed to be present when sunflower and linseed oils were heated at 200°C under air or at 275°C under nitrogen. Only the relative proportions and concentrations of CFAM varied (9). A number of model cyclic fatty acids have been synthesized as an aid to assigning the structures of a few of the natural acids. The synthetic routes have been reviewed in detail by Sébédio and Grandgirard (3) and more recently by Le Quéré and Sébédio (4). The skeletal structures of trans and cis ring isomers of cyclohexyl acids with rings from C-10 to C-15 and C-5 to C-10, from hydrogenated CFAM derived from heated linseed and sunflower oils, respectively, were verified by comparison to the mass spectra and retention times of the methyl esters of the synthetic compounds (15). Similarly, the isomers of the major cyclopentyl acids, with rings from C-8 to C12, were identified in the hydrogenated CFAM from heated sunflower oil (9). The mass spectrum of the methyl ester of a synthetic cyclohexenyl acid with a ring between C-10 and C-15 and a double bond at C-8 (Structure III, Fig. 1) aided identification of the cyclohexenyl dienoic acids from heated linseed oil, but positions of double bonds could not be confirmed (9). The structures, including double bond positions, of cyclopentenyl acids (I, II) in heated linseed oil were confirmed by comparison of the mass spectra of the picolinyl esters and DMOX derivatives with those of the authentic compounds (10). In an attempt to understand the mechanisms of cyclic fatty acid formation more completely, the cyclic dienoic acids from γ-linolenic acid in heated evening primrose oil have been studied recently (26). The products were compared with those formed from α-linolenic acid in heated linseed oil (10) and were totally analogous in the sense that the triene unit reacted in a similar way irrespective of its position along the chain. Therefore, there were 16 acids comprising two basic cyclopentenyl structures (with rings from C-8 to C-12 and C-7 to C-11 and double bonds at C-6 and C-12, respectively; XXIII and XXIV, Fig. 1) and two basic cyclohexenyl structures (with rings from C-7 to C-12 and double bonds at either C-5 or C-13; XXV and XXVI). All geometrical isomers of the double bonds in the straight chains and of the straight chains about the rings were observed. The proportions of acids with different structural features (e.g., Z compared with E double bonds, and C6 relative to C5 rings) were similar for both types of oils.

The Mechanisms of Cyclic Fatty Acid Formation

It has been more than 25 years since the first attempt at proposing a mechanism for formation of cyclic fatty acids in heated vegetable oils was made (27). It was proposed that the diene system in linoleic acid underwent a thermal intramolecular re-

Copyright © 1999 by AOCS Press

Cyclic Fatty Acids in Heated Vegetable Oils

207

arrangement to give a monoenoic acid with a cyclopentenyl ring. However, the resultant ring would be from C-9 and C-13 of the parent molecule, and this was subsequently found not to be one of the observed products (8). Any mechanism must explain the full range of observed products, including the restricted range of positional and geometrical double bond and ring isomers. It has been proposed that formation via radical intermediates is the most likely mechanism (7,8,10,23,28). A double bond can give rise to an allylic radical that may or may not undergo a hydrogen shift before reacting with an olefinic carbon to form a ring. 1,6- and 1,5-Hydrogen shifts result in five- (Fig. 7A) and six- (Fig. 7B) membered rings, respectively. Such mechanisms can account for all observed cyclic acids from oleate (7,28), linoleate (8,23,28), and α-linolenate (10,28).

Fig. 7. Possible mechanisms for the formation of (A) 9-(2′-butylcyclopentyl)nonanoic acid and (B) 9-(2′-propylcyclohexyl]nonanoic acid from oleic acid via free radical intermediates. Source: Ref. 28.

Copyright © 1999 by AOCS Press

208

G. Dobson

The simplest case is the formation of saturated monocyclic fatty acids from oleate that may produce four allylic radicals at C-8, C-9, C-10, and C-11. There must then be a 1,5- or 1,6-hydrogen shift before the radical can cyclize with an olefinic carbon, so that there are two possible products from each radical (not including ring stereoisomers), giving a total of eight possible structures (Table 1). A study of the saturated cyclic acids from high-oleate sunflower oil used in frying revealed only four (eight including ring isomers) basic structures (7), i.e., two cyclopentyl (Formulae XV and XVI) and two cyclohexyl acids (Formulae XVII and XVIII). Each radical produced one product, but it was unclear why the formation of the alternative products was not favored. In linoleate, allylic radicals may be formed at every carbon from C-8 to C-14. The 9,12-pentadiene unit will give the C-11 radical most easily, but the more stable radicals at C-9 and C-13 will be more favored in the five-carbon delocalized system. In addition, the double bonds can be involved individually to produce radicals at C8 and C-14, and delocalization will give radicals at C-10 and C-12, respectively. Cyclization between a radical and an olefinic carbon may occur with or without a preceding 1,5- or 1,6-hydrogen shift to give monocyclic monoenoic fatty acids. In total, there are sixteen possible products (not including double bond and ring geometrical isomers) (Table 2). Free radical mechanisms can account for eight (Formulae VVII, X-XIV) of the 10 observed products; the remaining two (Formulae VIII and IX) may be formed by thermal rearrangement of the ring double bond of two of these products (Formulae V and VI). Considering that the radical at C-11 is relatively unstable, it is not surprising that the four possible products from this radical were not observed (8). Presumably, the remaining expected products were not detected because competing mechanisms favored the formation of the observed products. Indeed, as observed for oleate, all radicals produced at least one product. All products with a double bond in the chain could occur as Z isomers. Only in cyclohexyl acids could the double bond in the straight chain take the E configuration, suggesting that stereomutation of the double bond took place before cyclization, and subsequent formation of a five-membered ring was not favored (8).

Copyright © 1999 by AOCS Press

Cyclic Fatty Acids in Heated Vegetable Oils

209

Bicyclic acids, derived from linoleate, have been characterized from sunflower oil used in frying (23). It has been suggested that the five-membered ring bicyclic acids (XIX, XX) could be alternative products to the cyclopentenyl ring acids (V, VI); both types of acids may be formed via 1,6-hydrogen shifts (but not by direct cyclizations) of the same allylic radicals. These mechanisms would fix the internal carbon-carbon bonds either between C-9 and C-11 (XIX) or C-11 and C-13 (XX). However, other positions would also have to occur to account for all of the isomers detected, indicating that alternative or additional mechanisms may occur. It was not possible to suggest mechanisms for the formation of six-membered ring bicyclic acids. Mechanisms involving 1,5-hydrogen shifts for the formation of cyclohexenyl acids give rise to free radicals outside the ring that could not be the precursors to bicyclic structures. α-Linolenate can produce 10 allylic radicals, and free radical mechanisms predict up to 20 CFAM (not counting stereoisomers; Ref. 28). However, only four basic structures (Formulae I–IV) were observed. A similar number of analogous products (XXIII–XXVI) were characterized from γ-linolenate (26). These observations stress the importance of the triene unit in determining the limited number of CFAM formed. As for oleate and linoleate, the restricted number of products can be explained partly by competing mechanisms from any one allylic radical. However, in contrast to oleate and linoleate, obviously not all possible radicals give rise to a cyclic product. Although the outer double bonds were involved in cyclization, products involving loss of the middle double bond were not detected (Fig. 4A). Also, products involving cyclizations between olefinic carbons of the outer double bonds and radicals even nearer the molecule extremities were not detected. The types of products formed could depend on the relative ease of formation and stability of the free radicals and their reactivity with the olefinic carbons. The additional presence of the double bonds at C-15 and C-6 in α- and γ-linolenate, respectively, compared with linoleate, must in-

Copyright © 1999 by AOCS Press

210

G. Dobson

fluence one or more of these factors to such an extent that the majority of expected products is excluded; how this can occur is unclear. The restricted range of stereoisomers from linoleate, in contrast to the presence of all possible double bond and ring isomers from linolenate, is also an enigma. There is much opportunity for further studies on the mechanisms of CFAM formation.

Summary

In recent years, there has been considerable progress in the structural characterization of cyclic fatty acids formed in heated vegetable oils. The structures of CFAM from linolenic, linoleic, and oleic acids have now been almost completely elucidated. This has been possible by advances in analytical methodology with the use of silverion HPLC for simplifying the mixtures, GC-MS of nitrogen-containing derivatives for structural characterization, and GC-FTIR for determining double bond geometry. There are still some areas requiring further study. For example, the detection of other possible minor components, suggested by studies of hydrogenated mixtures, and further characterization of complex mixtures of bicyclic acids. Also, apart from some information on heating at different temperatures in the presence or absence of air (9), little is known about the effects of frying conditions on the types and quantities of CFAM. It has been proposed that CFAM are formed by mechanisms involving allylic radicals (7,8,10,28). Although these mechanisms can explain the formation of most observed CFAM, further work is required to understand why other possible products are not favored. Examining the CFAM formed from individual fatty acids, covering a range of double bond positional isomers, should help to further the understanding of the mechanisms of CFAM formation. Acknowledgment

This paper is published as part of a program funded by the Scottish Office Agriculture, Environment and Fisheries Department. References

1. White, P.J., Methods for Measuring Changes in Deep-Fat Frying Oils, Food Technol. 45: 75–80 (1991). 2. Wells, A.F., and Common, R.H., Chemical Aspects of Thermal Damage to the Nutritive Value of Vegetable Oils. II. The Possible Formation of Cyclized or Branched Monomeric Acyl Radicals, J. Sci. Food Agric. 4: 233–237 (1953). 3. Sébédio, J.-L., and A. Grandgirard, Cyclic Fatty Acids: Natural Sources, Formation During Heat Treatment, Synthesis and Biological Properties, Prog. Lipid Res. 28: 303–336 (1989). 4. Le Quéré, J.-L., and J.-L. Sébédio, Cyclic Monomers of Fatty Acids, in Deep Frying: Chemistry, Nutrition and Practical Applications, edited by E.G. Perkins and M.D. Erickson, AOCS Press, Champaign, IL, 1996, pp. 49–88. 5. Dobson, G., Cyclic Fatty Acids: Qualitative and Quantitative Analysis, in Lipid Analysis in Oils and Fats, edited by. R.J. Hamilton, Blackie A & P, London, 1998, pp.136–180.

Copyright © 1999 by AOCS Press

Cyclic Fatty Acids in Heated Vegetable Oils

211

6. Chardigny, J.-M., R.L. Wolff, E. Mager, J.-L. Sébédio, L. Martine, and P. Juanéda, Trans Mono- and Polyunsaturated Fatty Acids in Human Milk, Eur. J. Clin. Nutr. 49: 523–531 (1995). 7. Dobson, G., W.W. Christie, and J.-L. Sébédio, Monocyctic Saturated Fatty Acids Formed from Oleic Acid in Heated Sunflower Oils, Chem. Phys. Lipids 82: 101–110 (1996). 8. Christie, W.W., E.Y. Brechany, J.-L. Sébédio, and J.-L. Le Quéré, Silver Ion Chromatography and Gas Chromatography-Mass Spectrometry in the Structural Analysis of Cyclic Monoenoic Acids Formed in Frying Oils, Chem. Phys. Lipids 66: 143–153 (1993). 9. Sébédio, J.-L., J.-L. Le Quéré, E. Sémon, O. Morin, J. Prévost, and A. Grandgirard, Heat Treatment of Vegetable Oils. II. GC-MS and GC-FTIR Spectra of Some Isolated Cyclic Fatty Acid Monomers, J. Am. Oil Chem. Soc. 64: 1324–1333 (1987). 10. Dobson, G., W.W. Christie, E.Y. Brechany, J.-L. Sébédio, and J.-L. Le Quéré, Silver Ion Chromatography and Gas Chromatography-Mass Spectrometry in the Structural Analysis of Cyclic Dienoic Acids Formed in Frying Oils, Chem. Phys. Lipids 75: 171–182 (1995). 11. Mossoba, M.M., M.P.Yuracwecz, J.A.G. Roach, H.S. Lin, R.E. McDonald, B.D. Flickinger, and E.G. Perkins, Rapid Determination of Double Bond Configuration and Position along the Hydrocarbon Chain in Cyclic Fatty Acid Monomers, Lipids 29: 893– 896 (1994). 12. Mossoba, M.M., M.P. Yuracwecz, J.A.G. Roach, H.S. Lin, R.E. McDonald, B.D. Flickinger, and E.G. Perkins, Elucidation of Cyclic Fatty Acid Monomer Structures. Cyclic and Bicyclic Ring Sizes and Double Bond Position and Configuration, J. Am. Oil Chem. Soc. 72: 721–727 (1995). 13. Mossoba, M.M., M.P. Yuracwecz, J.A.G. Roach, R.E. McDonald, and E.G. Perkins, Confirmatory Mass-Spectral Data for Cyclic Fatty Acid Monomers, J. Am. Oil Chem. Soc. 73: 1317–1321 (1996). 14. Sébédio, J.-L., J. Prévost, and A. Grandgirard, Heat Treatment of Vegetable Oils I. Isolation of the Cyclic Fatty Acid Monomers from Heated Sunflower and Linseed Oils, J. Am. Oil Chem. Soc. 64: 1026–1032 (1987). 15. Sébédio, J.-L., J.-L. Le Quéré, O. Morin, J.M. Vatèle, and A. Grandgirard, Heat Treatment of Vegetable Oils. III. GC-MS Characterization of Cyclic Fatty Acid Monomers in Heated Sunflower and Linseed Oils after Total Hydrogenation, J. Am. Oil Chem. Soc. 66: 704–709 (1989). 16. Le Quéré, J.-L., J.-L. Sébédio, R. Henry, F. Couderc, N. Demont, and J.-C. Promé, Gas Chromatography-Mass Spectrometry and Gas Chromatography-Tandem Mass Spectrometry of Cyclic Fatty Acid Monomers Isolated from Heated Fats, J. Chromatogr. 562: 659–672 (1991). 17. Le Quéré, J.-L., E. Sémon, B. Lanher, and J.-L. Sébédio, On-Line Hydrogenation in GC-MS Analysis of Cyclic Fatty Acid Monomers Isolated from Heated Linseed Oil, Lipids 24: 347–350 (1989). 18. Vatèle, J.M., J.-L. Sébédio, and J.-L. Le Quéré, Cyclic Fatty Acid Monomers: Synthesis and Characterization of Methyl ω-(2-Alkylcyclopentyl) Alkenoates and Alkanoates, Chem. Phys. Lipids 48: 119–128 (1988). 19. Rojo, J.A., and E.G. Perkins, Chemical Synthesis and Spectroscopic Characteristics of C18 1,2-Disubstituted Cyclopentyl Fatty Acid Methyl Esters, Lipids 24: 467–476 (1989). 20. Mossoba, M.M., M.P. Yuracwecz, H.S. Lin, R.E. McDonald, B.D. Flickinger, and E.G. Perkins, Application of GC-MI-FTIR Spectroscopy to the Structural Elucidation of

Copyright © 1999 by AOCS Press

212

G. Dobson

Cyclic Fatty Acid Monomers, Am. Lab. (Shelton) 27: 16K–O (1995). 21. Mossoba, M.M., M.P. Yuracwecz, J.A.G. Roach, R.E. McDonald, B.D. Flickinger, and E.G. Perkins, Analysis of Cyclic Fatty Acid Monomer 2-Alkenyl-4,4-dimethyloxazoline Derivatives by Gas Chromatography-Matrix Isolation-Fourier Transform Infrared Spectroscopy, J. Agric. Food Chem. 44: 3193–3196 (1996). 22. Dobson, G., W.W. Christie, and J.-L. Sébédio, Gas Chromatographic Properties of Cyclic Dienoic Acids Formed in Heated Linseed Oil, J. Chromatogr. A 723: 349–354 (1996). 23. Dobson, G., W.W. Christie, and J.-L. Sébédio, Saturated Bicyclic Fatty Acids Formed in Heated Sunflower Oils, Chem. Phys. Lipids 87: 137–147 (1997). 24. Rojo, J.A., and E.G. Perkins, Cyclic Fatty Acid Monomer Formation in Frying Fats. 1. Determination and Structural Study, J. Am. Oil Chem. Soc. 64: 414–421 (1987). 25. Sébédio, J.-L., M. Catte, M.A. Boudier, J. Prévost, and A. Grandgirard, Formation of Fatty Acid Geometrical Isomers and of Cyclic Fatty Acid Monomers During the Finish Frying of Frozen Prefried Potatoes, Food Res. Int. 29: 109–116 (1996). 26. Dobson, G., and J.-L. Sébédio, Monocyclic Dienoic Fatty Acids Formed from γLinolenic Acid in Heated Evening Primrose Oil, Chem. Phys. Lipids 97: 105–118 (1999). 27. Gast, L.E., Schneider, W.J., Forest, C.A., and J.C. Cowan, Composition of Methyl Esters from Heat-Bodied Linseed Oils, J. Am. Oil Chem. Soc. 40: 287–289 (1963). 28. Dobson, G., W.W. Christie, and J.-L. Sébédio, The Nature of Cyclic Fatty Acids Formed in Heated Vegetable Oils, Grasas Aceites 4: 26–33 (1996).

Copyright © 1999 by AOCS Press

Chapter 13

Production of Hydroxy Fatty Acids by Biocatalysis Ching T. Hou, Tsung M. Kuo, and Alan C. Lanser

Oil Chemical Research, NCAUR, ARS, USDA, Peoria, IL 61604

Introduction

Plant systems produce hydroxy fatty acids, which are important industrial materials. The hydroxy group gives a fatty acid special properties, such as higher viscosity and reactivity compared with other fatty acids. At present, imported castor oil and its derivatives are the only commercial source of these industrial hydroxy fatty acids. Because of their special chemical attributes, hydroxy fatty acids are used in a wide range of products, including resins, waxes, nylons, plastics, corrosion inhibitors, cosmetics, and coatings. Furthermore, they are used in grease formulations for highperformance military and industrial equipment. Ricinoleic and sebacic acids, two of the castor oil derivatives, are classified by the Department of Defense as strategic and critical materials. 12-Hydroxystearates (esters with C-10 to C-12 alcohols) are used in leather coatings requiring oil resistance and water imperviousness and in roll leaf foils because of their alcohol solubility and excellent wetting and adhesion to metallic particles (1). Because of fluctuating supplies and prices for castor oil, some companies have sought alternative raw materials, primarily petroleum-based feedstocks. Like ricinoleic acid, the hydroxy fatty acids of lesquerella also have double bonds and a carboxyl group that provide sites at which chemical reactions can occur (Fig. 1). We have been investigating the production of value-added products from soybean oil. A Japanese patent application by Soda et al. (2) claimed the production of ricinoleic acid from oleic acid by Bacillus pumilus. Our initial goal was to produce ricinoleic acid from oleic acid by biocatalysis and hence to reduce the dependency on imported castor oil. Although we could not demonstrate the production of ricinoleic acid from oleic acid as did other investigators, including Soda’s own group (2), our efforts led to discoveries of many new hydroxy fatty acids. These new products have potential industrial applications. Microbial oxidation of unsaturated fatty acids was reviewed recently (3).

Monohydroxy Fatty Acid

Production of 10-Hydroxystearic Acid

Microbial hydration of unsaturated fatty acid was first reported by Wallen et al. (4) from our laboratories. They found that a Pseudomonad isolated from fatty material 213 Copyright © 1999 by AOCS Press

214

C.T. Hou et al.

Fig. 1. Hydroxy fatty acids from castor oil and lesquerella oil.

hydrated oleic acid at the cis-9-double bond produced 10-hydroxystearic acid (10HSA) with a 14% yield. The 10-HSA is optically active (5,6) and has the d-configuration (6). Incubation of this organism with oleic acid in a medium enriched with deuterium oxide yields 10-HSA containing one deuterium atom (6). Moreover, this deuterium was shown to be on carbon atom 9 and in the L-configuration. By using soluble cell-free extracts of a Pseudomonad, Niehaus and Schroepfer (7) were able to demonstrate the conversion of oleate to 10-HSA under anaerobic conditions and the reversibility of the reaction. These findings, coupled with the observed stereospecific uptake of one atom of solvent hydrogen into 10-HSA and the lack of conversion of either the cis- or trans- 9,10-epoxystearic acid to 10-HAS, are compatible with a mechanism involving hydration of the double bond of oleic acid and rule out an epoxide intermediate (8). The same enzyme preparation was later found to catalyze both the hydration of cis- and trans-9,10-epoxystearic acids to yield threo- and erythro9,10-dihydroxystearic acid, respectively (9). Niehaus et al. (10) demonstrated the interconversion of oleic acid and 10-HSA by a soluble (105,000 × g supernatant) enzyme preparation from a Pseudomonad. This further ruled out the possible intermediate role of an epoxide in the overall conversion for two reasons. First, the enzymatic conversion of oleate to 10-HSA was observed to proceed under anaerobic conditions, a feature not characteristic of enzymatic epoxidations of olefins. Second, neither the DL-cis-9,10-epoxystearate nor the DL-trans-9,10-epoxystearate served as precursors of either oleate or 10-hydroxystearate under the conditions studied. By using the squalene screening method (11), Seo et al. (12) isolated a culture, Corynebacterium sp. S-401, from soil, that hydrates the squalene molecule to form tertiary alcohols. They found that resting cells of strain S-401 also stereospecifically hydrated oleic acid to 10-ketostearic (10-KSA) and (-)-10R-hydroxystearic acids with 22.4 and 9.1% yield, respectively. Strain S-401 failed to catalyze hydration of

Copyright © 1999 by AOCS Press

Industrial Hydroxy Fatty Acids by Biocatalysis

215

oleoamide, oleonitrile, oleyl alcohol, oleyl aldehyde, or cis-9-octadecene. Accordingly, the carboxy group of oleic acid seems to be essential in this reaction. Cells of Rhodococcus rhodochrous also hydrated oleic acid to 10-HSA and 10KSA at 55 and 12% yields, respectively (13). Hydration of oleic acid to 10-HSA was also demonstrated in resting cell suspensions of seven Nocardia species under anaerobic conditions (14). Nocardia cholesterolicum NRRL 5769 gave a yield >90% with optimum conditions at pH 6.5 and 40°C. A minor amount of 10-KSA was detected. The reaction proceeds via hydration of the double bond as shown by labeling experiments with deuterium oxide and 18O-labeled water. The system was specific for fatty acids with cis unsaturation at the 9 position. Anaerobiosis favors bioconversion to 10HSA (15) and higher pH favors bioconversion to 10-KSA (14). Thus far, the microbial hydration of oleic acid was found in Pseudomonas (4), Nocardia (Rhodococcus) (13,14), Corynebacterium (12), Sphingobacterium (16), and Micrococcus (17). Works of El-Sharkawy et al. (18) considerably extended the genera of microorganisms known to hydrate oleic acid to include a range of eucaryotic organisms. Strains from several other genera including Absida, Aspergillus, Candida, Mycobacterium, and Schizosaccharomyces were also found capable of catalyzing the hydration of oleic acid. Resting cells of Saccharomyces cerevisae (baker’s yeast, type II: Sigma, St. Louis, MO) converted oleic acid to 10-HSA with a 45% yield (18). Three other cultures, Nocardia aurantia ATCC 12674, Nocardia sp. NRRL 5646, and Mycobacterium fortuitum UI 53378, all converted oleic acid to 10-KSA with 65, 55, and 80% yields, respectively. Small amounts of 10-HSA were also produced by these cultures except strain NRRL 5646. The stereospecificity of microbial hydrations of oleic acid to 10-HS A was investigated by Yang et al. (19) on the basis of the ¹H nuclear magnetic resonance (NMR) spectral analysis of diastereomeric S-(+)-O-acetylmandelate esters of hydroxystearates (18). They found that although R. rhodochrous ATCC 12674-mediated hydration of oleic acid gave mixtures of enantiomers 10(R)-hydroxystearic acid and 10(S)-hydroxystearic acid, Pseudomonas sp. NRRL B-3266 produced optically pure 10(R)-hydroxystearic acid. The remaining microorganisms investigated (16) stereoselectively hydrated oleic acid to 10(R)-hydroxystearic acid containing 2 and 18%, respectively, of the contaminating 10(S)-hydroxystearic acid. Although hydration of oleic acid to 10-HSA was investigated at the cell-free enzyme level (7–10), attempts to purify hydratase were not successful. Very little was known about the physical and chemical properties of oleate hydratase. Purification and characterization of oleate hydratase from Nocardia cholesterolicum NRRL 5767 were investigated by Huang et al. (20). The cell-free extracts, which were obtained after French Press disintegration of the cells and centrifugation, were fractionated by ammonium sulfate. The enzyme activity was found in the fraction of 60–75% ammonium sulfate saturation. The enzyme fraction was further purified through a MonoQ ion exchange and Superose (Pharmacia, Piscataway, NJ) gel filtration column chromatography. The purified enzyme fraction showed a single Protein band on acrylamide gel electrophoresis. The hydration proceeded linearly for 6 h. The optimum

Copyright © 1999 by AOCS Press

216

C.T. Hou et al.

pH for the enzyme reaction is between 6.5 and 7. The Km value for the hydratase reaction at 30°C is 2.82 × 10−4 mol/L. The molecular weight estimated from Supherose HR 10/30 gel filtration column is ~120,000 Da and from SDS-PAGE is ~32,000 Da (21). Therefore, oleate hydratase is a tetramer, composed of four identical subunits. Lanser (22) reported the conversion of oleic acid to 10-ketostearic acid by a microorganism from other genera, Staphylococcus sp. The yield was >90% with <5% byproduct, 10-hydroxystearic acid. Hou reported (23) that Flavobacterium sp. DS5 converted oleic acid to 10-KSA in 85% yield. Optimum time, pH, and temperature for the production of 10-KSA are as follows: 36 h, 7.5, and 30°C. A small amount of 10-HSA (~10% of the main product 10-KSA) is also produced during the bioconversion. 10-KSA is not further metabolized by strain DS5 and accumulates in the medium. In contrast to growing cells, a resting cell suspension of strain DS5 produces 10-HSA and 10-KSA in a ratio of 1:3. The cell-free crude extract obtained from ultrasonic disruption of the cells converts oleic acid mainly to 10-HSA (10-HSA: 10-KSA = 97:3). This result strongly suggested that oleic acid is converted to 10-KSA via 10-HSA. Stereochemistry of 10-HSA from strain DS5, determined by ¹H NMR of the mandelate esters, showed 66% enantiomeric excess in 10(R) form. The Flavobacterium DS5 enzyme system also catalyzes the conversion of linoleic acid. In contrast to oleic acid substrate, which yields mainly the keto product, linoleic acid substrate yields mainly 10-hydroxy-12(Z)-octadecenoic acid (10-HOA) with 55% yield (24). The optimum conditions for the production of 10-HOA were pH 7.5, temperature 20–35°C, and 36 h of incubation. Two minor products produced were 10-methoxy-12-octadecenoic acid and 10-keto-12-octadecenoic acid (10KOA). Strain DS5 oxidized unsaturated but not saturated fatty acids. The relative activities were in the following order: oleic > palmitoleic > arachidonic > linoleic > linolenic > γ-linolenic > myristoleic acids. With the resting cells suspension, the ratio of products, 10-HOA:10-KOA was 97:3. Less 10-KOA was produced in comparison with that of growing cells. The cells were disrupted with ultrasonic oscillation and centrifuged to obtain cell-free crude extract. The linoleic acid conversion enzyme(s) resided in the cell-free crude extract, and only 10-HOA was produced from linoleic acid. Positional Specificity of Strain DS5 Hydratase

From substrate specificity studies (23,24), it seems that DS5 hydratase hydrates a specific carbon position of the unsaturated fatty acid substrates. To clarify this point and the effect of substrate carbon chain length on the strain DS5 hydratase activity, we studied the hydration of mono-, di-, and triunsaturated C-18 fatty acids as well as other carbon chain length monounsaturated fatty acids. Strain DS5 converted α-linolenic acid to 10-hydroxy-12,15-octadecadienoic acid and a minor product 10-keto-12,15-octadecadienoic acid (24). Strain DS5 also con-

Copyright © 1999 by AOCS Press

Industrial Hydroxy Fatty Acids by Biocatalysis

217

verted γ-linolenic acid to 10-hydroxy-6(Z),12(Z)-octadecadienoic acid. The enzyme hydrated 9-unsaturation but did not alter the original 6,12-unsaturations. Strain DS5 converted myristoleic acid to two products, 10-keto myristic and 10-hydroxymyristic acids. Palmitoleic acid also gave two bioconversion products, 10-ketopalmitic and 10-hydroxypalmitic acids. Previously, the strain DS5 bioconversion products from oleic and linoleic acids were identified as 10-ketostearic (23) and 10-hydroxy-12(Z)-octadecenoic acid (24), respectively. It is interesting to find that all unsaturated fatty acids tested are hydrated at the 9,10 positions with the oxygen functionality at C-10 despite their varying degree of unsaturations. DS5 hydratase was not active on saturated fatty acids and other non-9(Z)-unsaturated fatty acids such as elaidic [9(E)-octadecenoic], arachidonic [5(E),8(E),1 l(E),14(E)-eicosatetraenoic], and erucic [13(E)-docosenoic] acids (25). From all of the data gathered, it is concluded that DS5 hydratase is indeed a C-10 positional-specific enzyme. The fact that elaidic acid was not hydrated indicates that the unsaturation must be in the cis configuration for DS5 hydratase activity. The strain DS5 system produced more keto product from palmitoleic and oleic acids and more hydroxy product from myristoleic, linoleic, and α- and γ-linolenic acids. The reason for this preference is not clear. Among the 18-carbon unsaturated fatty acids, an additional double bond in either side of the C-10 position lowers the enzyme hydration activity. A literature search revealed that all microbial hydratases hydrate oleic and linoleic acids at the C-10 position (Fig. 2). Therefore, the positional specificity of microbial hydratases might be universal. Hydration of Other Fatty Acids

Hydrations of unsaturated fatty acids other than oleic acid were also reported. Wallen et al. (26) prepared three new unsaturated 10-hydroxy fatty acids, all optically active,

Fig. 2. Bioconversion products from unsaturated fatty acids by strain DS5 hydratase.

Copyright © 1999 by AOCS Press

218

C.T. Hou et al.

by the anaerobic microbial hydration of a cis-9-double bond. Substrates that formed these new hydroxy fatty acids are linoleic, linolenic, and ricinoleic acids. The yields were as follows: linoleic acid to 10-hydroxy-12(Z)-octadecenoic acid, 20 mole %; linolenic acid to 10-hydroxy-12(Z),15(Z)-octadecadienoic acid, 21 mol %; and ricinoleic acid to 10, 12-dihydroxystearic acid, 41 mol %. Giesel-Buhler et al. (27) reported the production of 10-hydroxy-12-octadecenoic acid from linoleic acid by resting cells of Acetobacterium woodii through hydration. In a patent disclosure, Litchfield and Pierce (13) claimed that cells of Rhodococcus rhodochrous catalyzed the hydration of linoleic acid to 10-hydroxy-12octadecenoic acid at 22% yield with 10-keto-12-octadecenoic acid as a co-product. The hydration enzyme is inducible by the presence of oleic acid at the early stage of cell growth. More recently, Koritala and Bagby (28), using washed resting cells suspension of Nocardia cholesterolicum under anaerobic conditions, reported the hydration of linoleic and linolenic acids to 10-hydroxy-12(Z)-octadecenoic (yield 71%) and 10hydroxy-12(Z),15(Z)-octadecadienoic acids (yield 77%), respectively. The production of 10-hydroxy fatty acids by hydratase from various microbes is summarized in Table 1. Other than hydration, the hydroxylation of oleic acid was also reported. Lanser et al. (29) found that two strains of Bacillus pumilus (NRRL BD-174 and BD-226) produced 15-, 16-, and 17-hydroxy-9-cis-octadecenoic acids.

Dihydroxy Unsaturated Fatty Acid

In our continuing screening program for new industrial chemicals from vegetable oils and their component fatty acids, we isolated a new bacterial strain, PR3, which converted oleic acid to a new compound, 7,10-dihydroxy-8(E)-octadecenoic acid

Copyright © 1999 by AOCS Press

Industrial Hydroxy Fatty Acids by Biocatalysis

219

(DOD), involving both hydration and possibly hydroxylation (30,31). Strain PR3, isolated from a water sample at a pig farm in Morton, IL, formed a smooth, round, white colony on agar plate. The microorganisms were motile, short, rod-shaped bacteria. Flagella stain showed multiple polar flagellae. Strain PR3 grew aerobically and could not grow anaerobically. The oxidase activity of the cells was positive. Based on these observations, strain PR3 belongs to the genus Pseudomonas (30). Strain PR3 produced fluorescein on King’s medium B as well as pyrocyanin on King’s medium A, suggesting that the organism was a strain of P. aeruginosa. Further identification was conducted with DNA reassociation measurements (32). Strain PR3 converted oleic acid to a new compound, 7,10-dihydroxy-8(E)octadecenoic acid. The absolute configuration of DOD was determined with the aid of circular dichroism to be R,R (33). The production of DOD from oleic acid reached a maximum after 48 h of incubation with a yield of 63%. Further incubation reduced DOD content in the medium, thus strain PR3 metabolizes DOD. The production of DOD from oleic acid is unique in that it involves an addition of two hydroxy groups at two positions and a rearrangement of the double bond of the substrate molecule. The reaction at the ∆9,10 position resembles hydration, and the reaction at the C-7 position seems like a hydroxylation. Subsequent investigation of reactions catalyzed by PR3 led to the isolation of another new compound, 10-hydroxy-8-octadecenoic acid (HOD) (34). From the structure similarity between HOD and DOD, it is likely that HOD is an intermediate in the formation of DOD from oleic acid by strain PR3. Kinetic studies (34) showed that the conversion from HOD to DOD is not a rate-limiting step. The bioconversion pathway for the production of DOD from oleic acid is postulated as follows (Fig. 3): a hydratase in strain PR3 attacks oleic acid at the C-10 position, introduces a hydroxy group, and at the same time shifts the double bond from C-9 to C-8. The resulting product (HOD) is then oxidized by a hydroxylase at the C-7 position to produce DOD. Recently, we were able to produce DOD with a cell-free enzyme preparation at a higher yield. The yield of DOD production by strain PR3 was improved to >80%.

Fig. 3. 7,10-Dihydroxy-8(E)-octadecenoic acid produced from oleic acid by Pseudomonas aeruginosa PR3.

Copyright © 1999 by AOCS Press

220

C.T. Hou et al.

It was also found that strain PR3 converted ricinoleic acid to a new compound, 7,10,12-trihydroxy-8(E)-octadecenoic acid at 35% yield (35). The reaction mechanism is the same as that for the conversion of oleic acid to DOD. Physiologic activity tests of DOD revealed that DOD has some activity against Bacillus subtilis and a common pathogen, Candida albican. A similar type of compound, dihydroxyoctadecenoic acid, was produced by Pseudomonas 42A2 (36). However, the positions for the double bond and hydroxy groups in that report were determined later and were shown to correspond to 7,10dihydroxy-8(E)-octadecenoic acid (37,38).

Trihydroxy Unsaturated Fatty Acid

Recently, we discovered the production of a new compound, 12,13,17-trihydroxy9(Z)-octadecenoic acid from linoleic acid by a new microbial isolate. Production of trihydroxy unsaturated fatty acids in nature is rare. The only compounds reported were all produced in trace amounts. 8,9,13-Trihydroxy docosaenoic acid was produced by yeast as an extracellular lipid (39). 9,10,13-Trihydroxy-11(E)- and 9,12,13trihydroxy-10(E)-octadecenoic acids were detected in beer (40). It has been suggested that these trihydroxy fatty acids are formed from linoleic acid during the processes of malting and mashing of barley (41). Gardner et al. (42) reported the production of diastereomeric (Z)-11,12,13-trihydroxy-9-octadecenoic acids and four isomers of (E)-9,12,13(9,10,13)-trihydroxy-10(11)-octadecenoic acids by acidcatalyzed transformation of 13(S)-hydroperoxylinoleic acid. Kato et al.(43,44) reported that hydroxy and epoxy unsaturated fatty acids present in some rice cultivars acted as antifungal substances and were active against rice blast fungus. It was postulated that these fatty acids were derivatives of linoleic and linolenic acid hydroperoxides. Recently, mixed hydroxy fatty acids, isolated from the Sasanishiki variety of rice plant suffering from the rice blast disease, were shown to be active against the fungus (45). Their structures were identified as 9S,12S,13S-trihydroxy-10-octadecenoic acid and 9S,12S,13Strihydroxy-10,15-octadecadienoic acid (46,47). 9,12,13-Trihydroxy-10(E)-octadecenoic acid was also isolated from Colocasia antiquorum inoculated with Cemtocystis fimbriata and showed activity against black rot fungus (48). Other than extraction from plant materials, our discovery is the first report on production of trihydroxy unsaturated fatty acids by microbial transformation. The microorganism that performs this unique reaction was isolated from a dry soil sample collected from McCalla, AL. Strain ALA2 is a gram-positive, nonmotile rod (0.5 µm × 2 µm). The strain was identified as Clavibacter sp. ALA2 (49). The chemical structure of the new compound was determined by mass spectrometry (MS), Fourier transform infrared Spectroscopy (FTIR) and NMR. The chemical ionization mass spectrum of the methyl ester prepared with diazomethane gave a molecular ion of mlz 345. Fragments of 327 (M-18), and 309 (M-2 × 18) were also seen. The electron impact spectrum of the methylated product provided more frag-

Copyright © 1999 by AOCS Press

Industrial Hydroxy Fatty Acids by Biocatalysis

221

ments for structural analysis. Large fragments corresponding to α-cleavage with ions mlz 227 (25%) and 129 (100%) place two hydroxy groups at the C-12 and C-13 positions and the third hydroxy group at a position higher than carbon 13. Proton and 13 C NMR analyses further confirmed the structure. Resonance signals (ppm) and corresponding molecular assignments, given in Table 2, located three hydroxy groups at C-12, C-13, and C-17 and further confirmed the identity of the bioconversion product as 12,13,17-trihydroxy-9(Z)-octadecenoic acid. The coupling constant of 10.7 Hz at C-9,10 confirmed our infrared data that the unsaturation is in cis configuration (49) (Fig. 4). The structure of THOA resembles those of plant self-defense substances. Other Reaction Products

Typical reaction products produced from linoleic acid by strain ALA2 analyzed by gas chromatography (GC) are shown in Figure 5. In addition to the main reaction product at retention time (Rt) 24 min, there were small amounts of products at Rt 17 and 10 min. Mass spectral analysis of fragments indicated that these were 12-[5ethyl-2-tetrahydrofuranyl]-7,12-dihydroxy-9Z-dodecenoic for Rt 17 and 12-[5-ethyl2-tetrahydrofuranyl]-12-hydroxy-9Z-dodecenoic acid for Rt 10. The yield of the main product (THOA) was 35% and the relative amounts of these products produced were THOA:Rt 17:Rt 10 = 9:1.3:1 (50).

Copyright © 1999 by AOCS Press

222

C.T. Hou et al.

Fig. 4. 12,13,17-Trihydroxy-9(Z)-octadecenoic acid produced from linoleic acid by Clavibacter sp. ALA2.

Fig. 5. A typical gas chromatography of strain ALA2 reaction products. 1: Internal standard, palmitic acid; 2: substrate, linoleic acid; 3–7: unknown; 8: product, 12,13,17trihydroxy-9(Z)-octadecenoic acid.

Copyright © 1999 by AOCS Press

Industrial Hydroxy Fatty Acids by Biocatalysis

223

The optimum conditions for the bioconversion of linoleic acid to THOA were pH 7.0 and temperature 30°C (51); the maximum production of THOA was found after 5–6 d of reaction. Further incubation did not reduce THOA content in the medium, indicating that strain ALA2 did not metabolize THOA. The biological activity of THOA at 200 ppm concentration was tested against many plant pathogenic fungi (52,53). The results, expressed in percentage of growth inhibition, are as follows: Erisyphe graminis (common disease name, wheat powdery mildew) 77%; Puccinia recondita (wheat leaf rust) 86%; Pseudocercosporella herpotrichoides (wheat foot rot) 0%; Septoria nodorum (wheat glume blotch) 0%; Pyricularia grisea (rice blast) 0%; Rhizoctonia solani (rice sheath blight) 0%; Phytophthora infestans (potato late blight) 56%; and Botrytis cinerea (cucumber botrytis) 63%. References

1. Naughton, F.C., Production, Chemistry, and Commercial Applications of Various Chemicals from Castor Oil, J. Am. Oil Chem. Soc. 51: 65–71 (1974). 2. Soda, K., and T. Kido, Manufacture of Hydroxy Unsaturated Fatty Acids with Bacillus pumilus, Japanese Kokai Tokkyo Koho JP 01051092 A2 890227 Heisei, Application JP 87–207994. 3. Hou, C.T., Microbial Oxidation of Unsaturated Fatty Acids, in Advances in Applied Microbiology, edited by A.I. Laskin, Academic Press, Orlando, 1995, vol. 41, pp. 1–23. 4. Wallen, L.L., R.G. Benedict, and R.W. Jackson, The Microbial Production of 10-Hydroxystearic Acid, Arch. Biochem. Biophys. 99: 249–253 (1962). 5. Schroepfer, G.J., Jr., and K.J. Block, Enzymatic Stereospecificity in the Dehydrogenation of Stearic Acid to Oleic Acid, J. Am. Chem. Soc. 85: 3310–3315 (1963). 6. Schroepfer, G.J., Jr., and K.J. Block, Enzymatic Stereospecificity in the Conversion of Oleic Acid to 10-Hydroxystearic Acid, J. Biol. Chem. 240: 54–65 (1965). 7. Niehaus, W.G., and G.J. Schroepfer, Jr., The Reversible Hydration of Oleic Acid to 10DHydroxystearic Acid, Biochem. Biophys. Res. Commun. 21: 271–275 (1965). 8. Schroepfer, G.J., Jr., Stereospecific Conversion of Oleic Acid to 10-Hydroxystearic Acid, J. Biol. Chem. 241: 5441–5447 (1966). 9. Niehaus, W.G., and G.J. Schroepfer, Jr., Enzymatic Stereospecificity in the Hydration of Epoxy Fatty Acids, J, Am. Chem. Soc. 89: 4227–4228 (1967). 10. Niehaus, W.G., A. Kisic, A. Torkelson, D.J. Bednarczyk, and G.J. Schroepfer, Jr., Stereospecific Hydration of the ∆9-Double Bond of Oleic Acid, J. Biol. Chem. 245: 3790– 3797 (1970). 11. Yamada, Y., H. Motoi, S. Kinoshita, N. Takada, and H. Okada, Oxidative Degradation of Squalene by Arthrobacter Species, Appl. Microbiol. 29: 400–404 (1975). 12. Seo, C.W., Y. Yamada, N. Takada, and H. Okada, Hydration of Squalene and Oleic Acid by Corynebacterium sp. S-401, Agric. Biol. Chem, 45: 2025–2030 (1981). 13. Litchfield, J.H., and G.E. Pierce, Microbiological Synthesis of Hydroxy-Fatty Acids and Keto-Fatty Acids, U.S. Patent 4,582,804 (1986). 14. Koritala, S., L. Hosie, C.T. Hou, C.W. Hesseltine, and M.O. Bagby, Microbial Conversion of Oleic Acid to 10-Hydroxystearic Acid, Appl. Microbiol. Biotechnol. 32: 299–304 (1989).

Copyright © 1999 by AOCS Press

224

C.T. Hou et al.

15. Davis, E.N., L.L. Wallen, J.C. Goodin, W.K. Rohwedder, and R.A. Rhodes, Microbial Hydration of cis-9-Alkenoic Acids, Lipids 4: 356–362 (1969). 16. Kaneshiro, T., J.-K. Huang, D. Weisleder, and M.O. Bagby, 10R-Hydroxystearic Acid Production by a Novel Microbe, NRRL B-14797, Isolated from Compost, J. Ind. Microbiol. 13: 351–355 (1994). 17. Blank, W., H. Takayanagi, T. Kido, F. Meussdoerffer, N. Esaki, and K. Soda, Transformation of Oleic Acid and Its Esters by Sarcina lutea, Agric. Biol. Chem. 55: 2651–2652 (1991). 18. El-Sharkawy, S.H., W. Yang, L. Dostal, and J.P.N. Rosazza, Microbial Oxidation of Oleic Acid, Appl. Environ. Microbiol 58: 2116–2122 (1992). 19. Yang, W., L. Dostal, and J.P.N. Rosazza, Stereospecificity of Microbial Hydration of Oleic Acid to 10-Hydroxystearic Acid, Appl. Environ. Microbiol. 59: 281–284 (1993). 20. Huang, J.-K., C.T. Hou, and M.O. Bagby, Isolation and Characterization of Oleate Hydratase from Nocardia cholesterolicum NRRL 5767, Annual Meeting, Society of Industrial Microbiology, Philadelphia, PA, Abstract p. 29 (1991). 21. Huang, J.-K., C.T. Hou, and M.O. Bagby, Purification and Characterization of Oleate Hydratase from Nocardia cholesterolicum NRRL 5767: Physical and Chemical Properties, 34th West Central States Biochemistry Conference, Ames, IA, Abstract p. 51 (1991). 22. Lanser, A.C., Conversion of Oleic Acid to 10-Ketostearic Acid by Staphylococcus sp., J. Am. Oil Chem. Soc. 70: 543–545 (1993). 23. Hou, C.T., Production of 10-Ketostearic Acid from Oleic Acid by a New Microbial Isolate, Flavobacterium sp. (NRRL B-14859), Appl. Environ. Microbiol. 60: 3760–3763 (1994). 24. Hou, C.T., Conversion of Linoleic Acid to 10-Hydroxy-12(Z)-Octadecenoic Acid by Flavobacterium sp. DS5, J. Am. Oil Chem, Soc. 71: 975–978 (1994). 25. Hou, C.T., Is Strain DS5 Hydratase a C-10 Positional Specific Enzyme? Identification of Bioconversion Products from α- and γ-Linolenic Acids by Flavobacterium sp. DS5, J. Ind. Microbiol 14: 31–34 (1995). 26. Wallen, L.L., E.N. Davis, Y.V. Wu, and W.K. Rohwedder, Stereospecific Hydration of Unsaturated Fatty Acids by Bacteria, Lipids 6: 745–750 (1971). 27. Giesel-Buhler, H., O. Bartsch, H. Hneifel, H. Sahm, and R. Schmid, Proceedings International Symposium on Biocatalysis in Organic Media held in Wageningen, edited by Laane, Tramper and Lilly, Elsevier Press, Amsterdam, 1987, p. 241. 28. Koritala, S., and M.O. Bagby, Microbial Conversion of Linoleic and Linolenic Acids to Unsaturated Hydroxy Fatty Acids, J. Am. Oil. Chem. Soc. 69: 575–578 (1992). 29. Lanser, A.C, R.D. Plattner, and M.O. Bagby, Production of 15-, 16- and 17-Hydroxy9-octadecenoic Acids by Bioconversion of Oleic Acid with Bacillus pumilus, J. Am. Oil Chem. Soc. 69: 363–366 (1992). 30. Hou, C.T., and M.O. Bagby, Production of a New Compound 7,10-Dihydroxy-8(E)octadecenoic Acid from Oleic Acid by Pseudomonas sp. PR3, J. Ind. Microbiol. 7: 123– 130 (1991). 31. Hou, C.T., M.O. Bagby, R.D. Platner, and S. Koritala, A Novel Compound, 7,10- Dihydroxy-8(E)-octadecenoic Acid from Oleic Acid by Bioconversion, J. Am. Oil Chem. Soc. 68: 99–101 (1991). 32. Hou, C.T., L.K. Nakamura, D. Weisleder, R.E. Peterson, and M.O. Bagby, Identification of NRRL Strain B-18602 (PR3) as Pseudomonas aeruginosa and Effect of Phenazine1-carboxylic Acid Formation on 7,10-Dihydroxy-8(E)-octadecenoic Acid Accumulation, World J. Microbiol. Biotechnol. 9: 570–573 (1993).

Copyright © 1999 by AOCS Press

Industrial Hydroxy Fatty Acids by Biocatalysis

225

33. Knothe, G., M.O. Bagby, R.E. Peterson, and C.T. Hou, 7,10-Hydroxy-8(E)-octadecenoic Acid: Stereochemistry and a Novel Derivative, 7,10-Dihydroxyoctadecanoic Acid, J. Am. Oil Chem. Soc. 69: 367–371 (1992). 34. Hou, C.T., and M.O. Bagby, 10-Hydroxy-8(Z)-Octadecenoic Acid, an Intermediate in the Formation of 7,10-Dihydroxy-8(E)-octadecenoic Acid from Oleic Acid by Pseudomonas sp. PR3, J, Indust. Micwbiol. 9: 103–107 (1992). 35. Kuo, T.M., L.K. Manthey, and C.T. Hou, Fatty Acid Bioconversion by Pseudomonas aeruginosa PR3, J. Am. Oil Chem. Soc. 75: 875–879 (1998). 36. Mercade, E., M. Robert, M.J. Espuny, M.P. Bosch, M.A. Manreesa, J.L. Parra, and J. Guinea, New Surfactant Isolated from Pseudomonas sp. 42A2, J. Am. Oil Chem. Soc. 65: 1915–1916 (1988). 37. de Andres, C, E. Mercade, J. Guinea, and A. Manresa, 7,10-Dihydroxy-8E-octadecenoic Acid Produced by Pseudomonas sp.42A2: Evaluation of Different Cultural Parameters of the Fermentation, World J. Microbiol. Biotechnol. 10: 106–109 (1994). 38. Guerrero, A., I. Casals, M. Busquets, Y. Leon, and A. Manresa, Oxidation of Oleic Acid to (E)-10-Hydroperoxy-8-octadecenoic and (E)-10-Hydroxy-8-octadecenoic Acids by Pseudomonas sp. 42A2, Biochim. Biophys. Acta 1347: 75–81 (1997). 39. Stodola, F.H., R.F. Vesonder, and L.J. Wickerham, 8,9,13-Trihydroxydocosanoic Acid, an Extracellular Lipid Produced by a Yeast, Biochemistry 4: 1390–1394 (1965). 40. Graveland, A., Enzymatic Oxidation of Linoleic Acid and Glycerol-1-Monolinoleate in Doughs and Flour-Water Suspensions, J. Am. Oil Chem. Soc. 47: 352–361 (1970). 41. Baur, C., and W. Grosch, Study on the Taste of Di-, Tri- and Tetrahydroxy Fatty Acids, Z. Lebensm. Unters. Forsch. 165: 82–84 (1977). 42. Gardner, H.W., E.C. Nelson, L.W. Tjarks, and R.E. England, Acid-Catalyzed Transformation of 3(S)-Hydroperoxy-Linoleic Acid into Epoxyhydroxyoctadecenoic Acid and Tri-Hydroxyoctadecenoic Acids, Chem. Phys. Lipids 35: 87–101 (1984). 43. Kato, T., Y. Yamaguchi, N. Abe, T. Uyeharaa, T. Nakai, S. Yamanaka, and N. Harada, Unsaturated Hydroxy Fatty Acids, the Self-Defensive Substances in Rice Plant Against Rice Blast Disease, Chem. Lett. 25: 409–412 (1984). 44. Kato, T., Y. Yamaguchi, T. Uyehara, T. Yokoyama, T. Namai, and S. Yamanaka, Self-Defensive Substances in Rice Plant Against Rice Blast Disease, Tetrahedron Lett. 24: 4715– 4718 (1983). 45. Kato, T., Y. Yamaguchi, N. Abe, T. Uyehara, T. Namai, M. Kodama, and Y. Shiobara, Structure and Synthesis of Unsaturated Trihydroxy C-18 Fatty Acids in Rice Plant Suffering from Rice Blast Disease, Tetrahedron Lett. 26: 2357–2360 (1985). 46. Kato, T., Y. Yamaguchi, S. Ohnuma, T. Uyehara, T. Namai, M. Kodama, and Y. Shiobara, Structure and Synthesis of 11,12,13-Trihydroxy-9(Z),15(Z)-octadecadienoic Acids from Rice Plant Suffering from Rice Blast Disease, Chem. Lett. 27: 577–580 (1986). 47. Suemune, H., T. Harabe, and K. Sakai, Synthesis of Unsaturated Trihydroxy C-18 Fatty Acids Isolated from Rice Plants Suffering from Rice Blast Disease, Chem. Pharm. Bull. 36: 3632–3637 (1988). 48. Masui, H., T. Kondo, and M. Kojima, An Antifungal Compound, 9,12,13-Trihydroxy(E)-10-octadecenoic Acid, from Colocasia antiquorum Inoculated with Ceratocystis fimbriata, Phytochemistry 28: 2613–2615 (1989). 49. Hou, C.T., A Novel Compound, 12,13,17-Trihydroxy-9(Z)-octadecenoic Acid, from Linoleic Acid by a New Microbial Isolate Clavibacter sp. ALA2, J. Am. Oil Chem. Soc. 73: 1359–1362 (1996).

Copyright © 1999 by AOCS Press

226

C.T. Hou et al.

50. Hou, C.T., H. Gardner, and W.K. Brown, Production of Polyhydroxy Fatty Acids from Linoleic Acid by Clavibacter sp. ALA2, J, Am. Oil Chem. Soc. 75: 1483–1487 (1998). 51. Hou, C.T., W. Brown, D.P. Labeda, T.P. Abbott, and D. Weisleder, Microbial Production of a Novel Trihydroxy Unsaturated Fatty Acid from Linoleic Acid, J. Ind. Micmbiol. Biotechnol. 19: 34–38 (1997). 52. Hou, C.T., Antimicrobial Activity of Hydroxy Fatty Acids, presented at SIMB Annual Meeting, Denver, CO, p. 2, 1998. 53. Hou, C.T., 12,13,17-Trihydroxy-9 (Z)-Octadecenoic Acid and Derivatives and Microbial Isolate for Production of the Acid, U.S. Patent #5,852,196, Dec. 22, 1998.

Copyright © 1999 by AOCS Press

Chapter 14

Some Derivatives of Fatty Compounds for Mass Spectral Structure Determination Gerhard Knothea and William W. Christieb

a National Center for Agricultural Utilization Research, Agricultural Research Service U.S. Department of Agriculture, Peoria, IL 61604 and bScottish Crop Research Institute, Invergowrie, Dundee DD2 5DA, Scotland, UK

Introduction

Mass spectrometry (MS), often applied in conjunction with gas chromatography (GC), is one of the most powerful methods for the structural analysis of fatty compounds. Accordingly, in recent years, several reviews (1–5) in addition to numerous research papers have been published on this subject. The mass spectra of underivatized fatty compounds derived from GC-MS are often nondistinctive. For example, compounds containing double bonds are subject to double bond migration. Thus, the position of the functional group in the chain often cannot be determined. For this reason, several methods have been developed for enabling more facile interpretation of mass spectra. The development of methods for sample derivatization has facilitated structural investigations of fatty compounds. The synthesis of some newer types of derivatives is discussed here. All of the methods discussed are suitable for determining double bond positions in fatty acid chains, although some can be used for determining the positions of other functional groups as well.

Discussion

Generally, the structure of fatty compounds offers two possibilities for performing derivatization reactions with the objective of improved fragmentation patterns in mass spectrometry. The first possibility is derivatization at the carboxylic acid moiety (C1; “remote-site derivatization”); the second possibility is derivatization of functional groups in the main fatty acid chain. Often both kinds of derivatization are carried out on one sample to facilitate analysis. Because this article is concerned mainly with synthetic aspects of the derivatives, mass spectral details will be discussed only briefly; instead, the reader is referred to the literature (1–5 and references therein) for that information.

227 Copyright © 1999 by AOCS Press

228

G. Knothe and W.W. Christie

Derivatization at the Carboxyl Group

Two major methods for derivatizing fatty acids at the carboxyl group (besides formation of methyl esters, of course, which will not be dealt with here) that have been developed are finding current practical applications (1–5). These are 4,4-dimethyloxazoline (DMOX) derivatives and picolinyl (3-hydroxymethyl pyridinyl) esters formed by reacting the carboxyl moiety with 2-amino-2-methyl-l-propanol (AMP) or 3-(hydroxymethyl) pyridine (HMP), respectively. In both cases, a nitrogen-containing heterocycle is introduced at the carboxyl moiety. Other heterocyclic derivatives described in the literature include pyrrolidides (6–8), nicotinates (9), alkenylbenzoxazoles (10), as well as piperidyl and morpholinyl esters (11). The introduction of the nitrogen-containing heterocycle causes the nitrogen atom to carry the charge upon ionization of the sample in the mass spectrometer, thus minimizing double bond ionization and migration (1,2). The nitrogen atom causes the molecular ion to be odd numbered, whereas most other distinctive fragments are even numbered. Cleavage in the chain occurs in most cases in a regular fashion, each bond in the chain providing a potential cleavage point. Thus, in the saturated parts of the chain, regular differences of 14 amu are observed. This regularity is interrupted when a double bond in the chain is encountered, leading to differences of 26 amu for picolinyl esters (with other characteristic ions further down the chain) and 12 amu for DMOX derivatives. Each derivative has advantages and disadvantages; thus they should be seen as complementary (1,3). DMOX derivatives are only slightly less volatile than methyl esters (1,2), whereas picolinyl esters require temperatures ~50°C higher for GC. On the other hand, for some alkyl-branched fatty compounds, the position of the branch is readily identifiable with the picolinyl esters in contrast to the DMOX derivatives. Generally, DMOX derivatives have yielded more easily interpretable spectra (1), as long as there is one fatty acid in the peak. Picolinyl esters were found to be preferable in the case of less well-resolved peaks. DMOX Derivatives

DMOX derivatives of fatty compounds for the purpose of mass spectral investigation were introduced in 1988 by Zhang et al. (12,13) for determining double bonds in olefinic fatty acids. While studying the location of double bonds in polyunsaturated fatty acids, Fay and Richli (14) proposed an improved preparation (incomplete reaction in Refs. 12,13) of the DMOX derivatives with fatty acid methyl esters as starting materials (formation of DMOX derivatives from the methyl esters and AMP by heating at 180°C overnight), in a slight modification from a method described previously (15). Other authors (16) used the method in a similar fashion. Similarly, a direct preparation of DMOX derivatives from oil or total lipid extracts involving heating at 180°C for 18 h under nitrogen was reported (17). 3(E)-Hexadecenoic acid (18) isomerized to the 2(Z)-isomer during the DMOX derivatization reaction, probably as a result of the high reaction temperature. A milder method for DMOX derivatization

Copyright © 1999 by AOCS Press

Fatty Acids Derivatives for Mass Spectrometry

229

using the acid chloride as intermediate was described recently by Christie (19) for ene-yne acids (crepenynic acid = octadeca-9-en-12-ynoic acid and octadeca-14-en17-ynoic acid). It has advantages in terms of preventing isomerization artifacts. The two methods presented here are the “usual method” and that based on the acid chloride using oxalyl chloride as acid chloride-forming reagent and trifluoroacetic anhydride as dehydration-cyclization reagent. The second method via the acid chloride is depicted in Scheme 1. As an example, the mass spectrum of the DMOX derivative of oleic acid is given in Figure 1. DMOX derivatives give prominent peaks characteristic of the DMOX moiety at m/z 113 (McLafferty rearrangement ion) and 126. The gap of 12 amu between m/z 196 and 208 locates the double bond in the example in Figure 1.

Experimental procedures. The first method is based on that in Ref. 16 (microprocedure for the preparation of fatty acid DMOX derivatives from free fatty acids, fatty acid methyl esters, or intact lipids). The reaction does not always go to completion and some N-acyl intermediate may remain. In this case, the products can be purified by Florisil chromatography, although the yield may not be good. Up to 2 mg lipid sample is added to AMP (0.25 g). The reaction tube is flushed with nitrogen, stoppered, and heated at 180°C overnight. The reaction tube is allowed to cool to room temperature. Dichloromethane (1 mL) is added, followed by isohexane (5 mL) and water (3 mL), and the mixture thoroughly shaken. Addition of some potassium or sodium chloride may be necessary to break up any emulsions. The organic layer is transferred to a fresh test tube, distilled water (3 mL) added, and the mixture shaken. The solvent layer is passed through a short (3 cm) column of anhydrous sodium sulfate to dry. The sample is washed through the column with isohexane (2 mL). The sample is dried under a gentle stream of nitrogen on a heating block at 30°C and is then ready for GC-MS analysis. When further purification of the derivatives is required, the sample is washed through a Florisil column (prewashed with isohexane) with isohexane/acetone (94:6, vol/vol, 4 mL) and evaporated to dryness as above; however, recoveries may be poor.

New acid chloride preparation of DMOX derivatives. (Procedure from Ref. 19.) For preparation of the acid chloride, lipid samples (up to 5 mg) must first be hydrolyzed to the free fatty acids (KOH in aqueous ethanol). For formation of the acid

SCHEME 1. Synthesis of 4,4-dimethyloxazoline (DMOX) derivatives.

Copyright © 1999 by AOCS Press

230

G. Knothe and W.W. Christie

Fig. 1. Mass spectrum of the 4,4-dimethyloxazoline (DMOX) derivative of oleic acid.

chloride, oxalyl chloride is added to the fatty acids and this mixture is stirred at room temperature overnight. The oxalyl chloride is removed in a stream of nitrogen to yield the acid chloride. This is used immediately for the preparation of the required derivatives. Reaction of acid chloride with AMP. A solution of AMP in dichloromethane (10 mg/mL) is prepared and stored over anhydrous sodium sulfate. The solution is stable in a refrigerator for ~1 mo. The stock solution of AMP in CH2Cl2 (0.5 mL) is added to the freshly prepared acid chloride, cooled in an ice bath, and the mixture is then left to warm up to room temperature for 1 h. The solvent is removed in a stream of nitrogen and then trifluoroacetic anhydride (0.5 mL) is added. After 1 h on a heating block at 50°C or 3 h at room temperature, the reagent is removed in a stream of nitrogen. Isohexane (5 mL) is added, followed by water (2 mL). The mixture is shaken thoroughly and then the solvent layer is removed by Pasteur pipette. It is dried over sodium sulfate via a small column and concentrated under nitrogen. If necessary, the product can be purified by applying to a short column of Florisil as described earlier. Compounds apparently closely related to the DMOX derivatives but eluting a little later on GC can be 1–2% of the total; currently, there is no way to eliminate these. Picolinyl Esters

Picolinyl (3-hydroxymethyl pyridinyl) esters as derivatives of fatty compounds for the purpose of mass spectral structure determination were originally introduced by Harvey (20). Picolinyl esters have replaced the originally developed pyrrolidides (6– 9) because of easier preparation and ions of greater abundance (20). For both pyridinecarboxylates and hydroxymethyl pyridinyl esters, it was shown that the nitrogen atom must be in the 3-position in order to obtain spectra with peaks of suffi-

Copyright © 1999 by AOCS Press

Fatty Acids Derivatives for Mass Spectrometry

231

cient abundance (9,20). Picolinyl esters have been used in the structure determination of numerous fatty compounds with different functional groups in the chain (1,3–5). In their studies of the characterization of the picolinyl esters of epoxides of polyunsaturated fatty acids, Balazy and Nies (21) prepared the picolinyl esters via unstable imidazolide intermediates, which were generated by reaction of the epoxy acids with 1,1′-carbonyldiimidazole. For the epoxides, a method using acid chlorides as intermediates (22) and thionyl choride as reagent was not recommended because of the instability of the epoxides in acidic solution. However, using oxalyl chloride as reagent for formation of the acid chloride (Scheme 1) is a viable route and was used in the formation of derivatives of ene-yne acids (19). Reaction with HMP yielded the desired picolinyl ester (Scheme 2). The method of Balazy and Nies (21) as well as the oxalyl chloride—based acid chloride method of Christie (19) are given. As an example, the mass spectrum of the picolinyl ester of oleic acid is given in Figure 2. Picolinyl esters give prominent peaks characteristic of the picolinyl moiety at m/z 92, 108, 151 (McLafferty rearrangement ion), and 164. The gap of 26 amu between m/z 234 and 260 locates the double bond in the example in Figure 2.

Experimental procedures. (Based on Ref. 21.) The lipid sample should be present in the form of free fatty acids. The reaction is moisture sensitive; thus all reagents should be dried before use. A fresh diimidazolide solution should be used each time samples are prepared. The picolinyl-derivatizing reagent is prepared by dissolving HMP (900 µL) in dichloromethane (9 mL) and triethylamine (9 mL) and adding a small amount of anhydrous sodium sulfate (2–3 mm in tube). This reagent can be stored in a refrigerator for 1 mo if properly sealed. The free fatty acid sample (up to 1 mg; in a centrifuge tube) is dissolved in dichloromethane (100 µL) and a freshly prepared solution of 1,1′-dicarbonydiimidazole (100 µL, 100 mg/mL) is added. This solution is left for 1 min at room temperature (the reaction time is critical; longer reaction times give reduced recoveries). Then 200 µL picolinyl reagent is added; the mixture is agitated and left for 10 min at 37°C before acetic acid (10 µL) is added. It is then dried at 30°C with a gentle stream of nitrogen. Then 5 mL isohexane and 2 mL distilled water are added and the mixture is shaken and centrifuged. The upper isohexane layer is eluted through a short (4 cm) column of anhydrous sodium sulfate prewashed with isohexane. The aqueous layer is reextracted with fresh isohexane (2 mL) and passed through the column. If necessary, some crystalline potassium chloride can be added to break up any emulsions. Finally, it is washed through the column

SCHEME 2. Synthesis of picolinyl esters.

Copyright © 1999 by AOCS Press

232

G. Knothe and W.W. Christie

Fig. 2. Mass spectrum of the picolinyl ester of oleic acid.

with isohexane (1 mL) and the combined isohexane eluents evaporated as described previously (heating block). An optional procedure to ensure removal of excess reagents or by-products is as follows: The sample is dissolved in isohexane (1 mL) and passed through a short Florisil column (4 cm) prewashed with isohexane (3 mL). The column is eluted with isohexane/acetone (98:2, vol/vol, 5 mL) and the eluant discarded. Then the column is eluted with isohexane/acetone (80:20, vol/vol, 8 mL) and the sample evaporated to dryness as done previously.

New acid chloride-based preparation of picolinyl esters. A solution of HMP in CH2Cl2 (20 mg/mL) is prepared and stored over anhydrous sodium sulfate. The solution is stable in a refrigerator for ~1 mo. The stock solution of HMP in CH2Cl2 (0.5 mL) is added to the freshly prepared acid chloride (synthesis of the fatty acid chloride, see procedure for DMOX derivatives), cooled in an ice bath, and the mixture is then left to warm up to room temperature for 1 h (a white precipitate may form). The solvent is added in a stream of nitrogen; isohexane (5 mL) is added, followed by 0.5% aqueous sodium bicarbonate. The mixture is shaken thoroughly and then the solvent layer is removed by Pasteur pipette. It is dried via a small sodium sulfate column and concentrated under nitrogen. If necessary, the product can be purified by Florisil chromatography as described above. Derivatization in the Chain

Numerous well-known methods exist for analyzing functional groups in a fatty acid chain. These methods include deuteration, hydrogenation, epoxidation, hydroxylation, and silylation. Ozonolysis, although strictly speaking not a derivatization because the fatty acid chain is cleaved and the cleavage products analyzed, can also

Copyright © 1999 by AOCS Press

Fatty Acids Derivatives for Mass Spectrometry

233

be included here. Although derivatizations at the carboxyl group often have wider applicability to a variety of functional groups in a chain, a chain derivatization is usually targeted at analyzing one specific kind of functional group. Some chain derivatizations can be carried out extremely easily, for example silylation (where often just adding the silylation reagent to the sample in the GC vial suffices), whereas others, such as ozonolysis, are considerably more involved. Two newer chain derivatizations will be considered here. They are Diels-Alder adducts of conjugated double bonds with 4-methyl-l,2,4-triazoline-3,5-dione (MTAD) and addition of dimethyl disulfide (DMDS) to double bonds. DMDS [bis(methylthio)] Derivatives

The alkylthiolation of double bonds for the determination of their position in fatty compounds (23; see Scheme 3) and linear alkenes (24) was originally reported in 1981; the procedure has recently been reviewed (3). The seemingly straightforward addition reaction, however, possesses some more intriguing aspects. The addition of DMDS is Stereospecific, leading to the formation of erythro- and threo-diastereomers from E and Z double bond configurations, respectively, when assuming trans addition of the methylthio groups (25). The threo diastereomers eluted at slightly lower temperature in the GC-MS than their erythro congeners (25,26) and elution temperatures increased with increasing distance of the original double bond from C-l (25). DMDS derivatization was extended to diunsaturated fatty compounds by Vincenti et al. (27). Only when the two double bonds in dienes are separated by at least four methylene groups do the expected di-adducts arise (27). When the two double

SCHEME 3. a) Synthesis of dimethyl disulfide (DMDS) derivatives of isolated double bonds (monounsaturated compounds or double bonds separated by four or more CH2 groups); b) DMDS adducts from dienes with double bonds separated by 1, 2, or 3 CH2 groups; c) DMDS adducts from conjugated double bonds.

Copyright © 1999 by AOCS Press

234

G. Knothe and W.W. Christie

bonds are separated by one, two, or three CH2 groups, four-, five-, and six-membered heterocyclic thioethers (containing one S atom; see Scheme 3) and two methylthio substituents on the chain carbons to the ring arise (27; see Scheme 3), with only carbons from positions between the double bonds contributed by the fatty compound. Conjugated double bonds form five-membered cyclic thioethers (all carbons from double bond positions) with two methylthio substituents attached to heterocyclic carbons. Due to the resulting fragmentation pattern, the double bonds may easily be recognized as conjugated, but their position in the chain may be uncertain (27). In derivatives of monounsaturated fatty compounds, the most intense ions are caused by the cleavage of the bond between the two carbons carrying the methylthio groups, i.e., the site of the original double bond. These fragments allow for easy location of the double bond position. The methylthio adducts exhibit a characteristic peak at m/z 61 (23), which is identified as CH3S+=CH2. The methyl ester-containing ion also gives a strong peak due to the loss of methanol (−32). Besides the molecular ion, the spectra show M-31 and M-47 (loss of methoxyl and methylthienyl, respectively). As an example, the mass spectrum of the DMDS derivative of methyl oleate is given in Figure 3. In this instance, the ions at m/z 173 and 217 represent cleavage at the carbons originally constituting the double bond. Experimental procedures. (According to Ref. 2.) Alkenoate (0.01 mmol) is added to dimethyl disulfide (0.06 mmol) containing 0.0005 mmol iodine. The resulting mixture is purged with nitrogen and stirred for 24 h at room temperature. The mixture is then analyzed directly by GC-MS. On a nanogram scale, according to Ref. 26, a heptane solution (1–20 µL) containing 50–500 ng monounsaturated fatty compound is treated with DMDS (50 µL) and 0.06% iodine in diethyl ether (5 µL). The mixture

Fig. 3. Mass spectrum of the dimethyl disulfide (DMDS) derivative of methyl oleate.

Copyright © 1999 by AOCS Press

Fatty Acids Derivatives for Mass Spectrometry

235

is reacted overnight at 40°C. Then heptane (10–50 µL) and aqueous 5% sodium thiosulfate (25 µL) are added to quench the reaction. The heptane phase is concentrated under a stream of nitrogen. For reaction with 10–50 ng starting material, only 10 µL DMDS, 2 µL iodine solution, and 10 µL sodium thiosulfate solution are used. MTAD Adducts

Due to the considerable interest in the physiologic effects of conjugated linoleic acid (CLA), the analysis thereof is also highly significant. Conjugated double bonds undergo the Diels-Alder reaction. Therefore, it is reasonable to assume that a DielsAlder adduct amenable to GC-MS investigation could be developed. Although maleic anhydride adducts have been known for some time (28), they are suitable for derivatizing only dienes with E,E configuration. Young et al. (29) reported on locating the position of conjugated dienes in an aliphatic chain by means of mass spectrometry (direct insertion) of 4-phenyl-l,2,4triazoline-3,5-dione (PTAD) adducts synthesized according to a previously described procedure (30). As a means for increasing volatility of the adducts, 4-methyl-1,2,4triazoline-3,5-dione (MTAD) derivatives were introduced (31). Preparation of the MTAD dienophile proceeded according to the PTAD method with the exception of the use of N2O4 as oxidizing agent instead of tert-butyl hypochlorite for the urazole (4-methyl-1 ,2,4- triazolidine-3,5- dione) starting material. The method of Young et al. (31) was modified (28) and used in the analysis of CLA. This is the procedure given here. The MTAD derivatives are suitable for dienes with different double bond geometries; however, some reactivity differences that depend on the diene configuration have been reported (28 and references therein). Christie (19) reported the use of MTAD adducts for proving (by isomerization) the formation of 7-(5-pentylcyclohexadienyl)-heptadecanoate during the DMOX derivatization of crepenynic acid; they have also been used to characterize long-chain metabolites of CLA (32). A depiction of the MTAD adduct formation is given in Scheme 4. The MTAD moiety gives characteristic peaks at m/z 165/166 [M - R′- R″ + H]+ and distinct ions representing cleavage on either side of the six-membered ring.

Experimental procedures. (According to Ref. 28, modified from Ref. 31.) Solutions containing known amounts of CLA sample and MTAD (which is commercially available) in dichloromethane are mixed at 0°C with <10 s agitation. CLA methyl ester (1.15 mmol/L) and MTAD (5.79 mmol/L; molar ratio CLA:MTAD 1:5) in CH2Cl2 (650 µL) were used. The reaction is quenched immediately by adding a twofold molar excess (relative to MTAD) of 1,3-hexadiene, followed by a few seconds of agitation. The mixture is dried under a nitrogen stream at 30°C and dissolved in CH2Cl2 to give ~0.1% (wt/vol) solutions of the MTAD adducts. The sample can then be analyzed directly by GC-MS. Young et al. (31) used the method employed for PTAD adducts (30). Reaction of PTAD with a diene was carried out by dissolving both in CH2Cl2 and mixing them at room temperature.

Copyright © 1999 by AOCS Press

236

Acknowledgment

G. Knothe and W.W. Christie

SCHEME 4. Synthesis of 4methyl-1,2,4-triazoline-3,5dione (MTAD) derivatives.

This work was funded in part by the Scottish Office of Agriculture, Environment and Fisheries Department. References

1. Christie, W.W., Gas-Chromatography-Mass Spectrometry Methods for Structural Analysis of Fatty Acids, Lipids 33: 343–353 (1998). 2. Spitzer, V., Structure Analysis of Fatty Acids by Gas Chromatography-Low Resolution Electron Impact Mass Spectrometry of Their 4,4-Dimethyloxazoline Derivatives—A Review, Prog. Lipid Res. 35: 387–408 (1997). 3. Christie, W.W., Structural Analysis of Fatty Acids, in Advances in Lipid Methodology— Four, edited by W.W. Christie, The Oily Press, Dundee, 1997, pp. 119–169. 4. Murphy, R.C., Mass Spectrometry of Lipids, Handbook of Lipid Research 7, edited by F. Snyder, Plenum Press, New York, 1993. 5. Harvey, D.J., Mass Spectrometry of Picolinyl and Other Nitrogen-Containing Derivatives

Copyright © 1999 by AOCS Press

Fatty Acids Derivatives for Mass Spectrometry

237

of Lipids, in Advances in Lipid Methodology—One, edited by W.W. Christie, The Oily Press, Dundee, 1992, pp, 19–80. 6. Vetter, W., W. Walther, and M. Vecchi, Pyrrolidide als Derivate für die Strukturaufklärung aliphatischer und alicyclischer Carbonsäuren mittels Massenspektrometrie, Helv. Chim. Acta 54: 1599–1605 (1971). 7. Andersson, B.Å., and R.T. Holman, Pyrrolidides for Mass Spectrometry Determination of the Position of the Double Bond in Monounsaturated Fatty Acids, Lipids 9: 185–190 (1974). 8. Andersson, B.Å., W.-H. Heimermann, and R.T. Holman, Comparison of Pyrrolidides with Other Amides for Mass Spectral Determination of Structure of Unsaturated Fatty Acids, Lipids 9: 443–449 (1974). 9. Vetter, W., and W. Meister, Nicotinates as Derivatives for the Mass Spectrometric Investigation of Long-Chain Alcohols, Org. Mass Spectrom. 16: 118–122 (1981). 10. Yu, Q.-T., J.-Y. Zhang, and Z.-H. Huang, A Novel Approach to Double Bond Location in Long-Chain Olefinic Acids: The Mass Spectra of 2-Alkenylbenzoxazoles, Biomed. Environ. Mass Spectrom. 13: 211–216 (1986). 11. Jutila, M., and J. Jalonen, Piperidyl and Morpholinyl Esters as Derivatives for the Structural Determination of Long-Chain Fatty Acids by Mass Spectrometry, Biomed. Environ. Mass Spectrom. 17: 433–436 (1988). 12. Zhang, J.Y., Q.T. Yu, B.N. Liu, and Z.H. Huang, Chemical Modification in Mass Spectrometry IV—2-Alkenyl-4,4-Dimethyloxazolines as Derivatives for the Double Bond Location of Long-Chain Fatty Acids, Biomed. Environ. Mass Spectrom. 15: 33–44 (1988). 13. Yu, Q.T., B.N. Lu, J.Y. Zhang, and Z.H. Huang, Location of Methyl Branchings in Fatty Acids: Fatty Acids in Uropygial Secretion of Shanghai Duck by GC-MS of 4,4-Dimethyloxazoline Derivatives, Lipids 23: 804–810 (1988). 14. Fay, L., and U. Richli, Location of Double Bonds in Polyunsaturated Fatty Acids by Gas-Chromatography-Mass Spectrometry after 4,4-Dimethyloxazoline Derivatization, J. Chromatogr. 541: 89–98 (1991). 15. Lepage, G., and C.C. Roy, Direct Transesterification of All Classes of Lipids in a OneStep Reaction, J. Lipid Res. 27: 114–120 (1986). 16. Luthria, D.L., and H. Sprecher, 2-Alkenyl-4,4-dimethyloxazolines as Derivatives for the Structural Elucidation of Isomeric Unsaturated Fats, Lipids 28: 561–564 (1993). 17. Garrido, J,L., and I. Medina, One-Step Conversion of Fatty Acids into Their 2-Alkenyl4,4-dimethyloxazoline Derivatives Directly from Total Lipids, J. Chromatogr. A 673: 101–105 (1994). 18. Lamberto, M., and R.G. Ackman, Positional Isomerization of trans-3-Hexadecenoic Acid Employing 2-Amino-2-methyl-propanol as a Derivatizing Agent for Ethylenic Bond Location by Gas Chromatography/Mass Spectrometry, Anal. Biochem. 230: 224– 228 (1995). 19. Christie, W.W., Mass Spectrometry of Fatty Acids with Methylene-Interrupted Ene-Yne Systems, Chem. Phys. Lipids 94: 35–41 (1998). 20. Harvey, D.J., Picolinyl Esters as Derivatives for the Structural Determination of Long Chain Branched and Unsaturated Fatty Acids, Biomed. Mass Spectrom. 9: 33–38 (1982). 21. Balazy, M., and A.S. Nies, Characterization of Epoxides of Polyunsaturated Fatty Acids by Mass Spectrometry via 3-Pyridinylmethyl Esters, Biomed. Environ. Mass Spectrom. 18: 328–336 (1989).

Copyright © 1999 by AOCS Press

238

G. Knothe and W.W. Christie

22. Harvey, D.J., Picolinyl Derivatives for the Structural Determination of Fatty Acids by Mass Spectrometry: Applications to Polyenoic Acids, Hydroxy Acids, Di-Acids and Related Compounds, Biomed. Mass Spectrom. 11: 340–347 (1984). 23. Francis, G.W., Alkylthiolation for the Determination of Double-Bond Position in Unsaturated Fatty Acid Esters, Chem. Phys. Lipids 29: 369–374 (1981). 24. Francis, G.W., and K. Veland, Alkylthiolation for the Determination of Double Bond Positions in Linear Alkenes, J. Chromatogr, 219: 379–384 (1981). 25. Buser, H.-R., H. Arn, P. Guerin, and S. Rauscher, Determination of Double Bond Position in Monounsaturated Acetates by Mass Spectrometry of Dimethyl Disulfide Adducts, Anal. Chem. 55: 818–822 (1983). 26. Leonhardt, B.A., and E.D. DeVilbiss, Separation and Double-Bond Determination on Nanogram Quantities of Aliphatic Monounsaturated Alcohols, Aldehydes and Carboxylic Acid Methyl Esters, J. Chromatogr. 322: 484–490 (1985). 27. Vincenti, M., G. Guglielmetti, G. Cassani, and C. Tonini, Determination of Double Bond Position in Diunsaturated Compounds by Mass Spectrometry of Dimethyl Disulfide Derivatives, Anal. Chem. 59: 694–699 (1987). 28. Dobson, G., Identification of Conjugated Fatty Acids by Gas Chromatography-Mass Spectrometry of 4-Methyl-l,2,4-triazoline-3,5-dione Adducts, J. Am. Oil Chem. Soc. 75: 137–142 (1998). 29. Young, D.C., P. Vouros, B. Decosta, and M.F. Holick, Location of Conjugated Diene Position in an Aliphatic Chain by Mass Spectrometry of the 4-Phenyl-l,2,4-triazoline3,5-dione Adduct, Anal. Chem. 59: 1954–1957 (1987). 30. Cookson, R.C., S.S.H. Gilani, and I.D.R. Stevens, Diels-Alder Reactions of 4-Phenyll,2,4-triazoline-3,5-dione, J. Chem. Soc. C: 1905–1909 (1967). 31. Young, D.C., P. Vouros, and M.F. Holick, Gas Chromatography-Mass Spectrometry of Conjugated Dienes by Derivatization with 4-Methyl-l,2,4-triazoline-3,5-dione, J. Chromatogr. 522: 295–302 (1990). 32. Sébédio, J.L., P. Juanéda, G. Dobson, I. Ramilison, J.C. Martin, J.M. Chardigny, and W.W. Christie, Metabolites of Conjugated Isomers of Linoleic Acid (CLA) in the Rat, Biochim. Biophys. Acta 1345: 5–10 (1997).

Copyright © 1999 by AOCS Press