Advances in Physical Organic Chemistry, Volume 28

Advances in Physical Organic Chemistry Volume 28

Edited by

D. BETHELL The Robert Robinson Laboratories Department of Chemistry University of Liverpool P.O. Box 147 Liverpool L69 3BX

ACADEMIC PRESS Harcourt Brace Jovanovich, Publishers London San Diego New York Boston Sydney Tokyo Toronto

ACADEMIC PRESS LIMITED 24/28 Oval Road London NWl 7DX United States Edition published by ACADEMIC PRESS INC. San Diego, CA 92101 Copyright 0 1993 by ACADEMIC PRESS LIMITED All rights reserved

No part of this book may be reproduced in any form by photostat, microfilm, or any other means, without written permission from the publishers

A catalogue record for this book is available from the British Library ISBN 0-12-033528-X ISSN 0065-3160

FILMSET BY P&R TYPESETTERS LTD, SALISBURY, WILTSHIRE. UK AND PRINTED IN GREAT BRITAIN BY HARTNOLLS LIMITED, BODMIN, CORNWALL

Contents

vi i

Preface

viii

Contributors t o Volume 28

Electron Storage and Transfer in Organic Redox Systems w i t h Multiple Electrophores

1

MARTIN BAUMGARTEN, WALTER HUBER KLAUS MULLEN

AND

Introduction 1 Design and synthesis 5 Extended redox sequences 10 Mechanism of successive electron transfers 14 5 Structural transfer relevant to intramolecular electron transfer 17 6 Conclusion 39

1 2 3 4

Chirality and Molecular Recognition in Monolayers at the Air-Water Interface

PHILIP L. ROSE, NOEL G. HARVEY EDWARD M. ARNETT 1 2 3 4

Introduction 45 Monolayer methods 49 Chiral monolayers 71 Conclusions 133 ... 111

AND

45

CONTENTS

IV

Transit ion-St ate Theory Revisited

139

W. JOHN ALBERY 1 2 3 4 5 6 7 8 9 10 11 12

Introduction 139 The classical paradox 140 An alternative derivation 140 Where is the transition state? 143 The adiabatic case 145 Multistep reactions in solution 147 Proton transfers to cyanocarbon bases 152 Extension to include XR 153 Reactions at liquid/liquid interfaces 155 Colloidal deposition 158 Free energy profiles 163 Summary 167

Neighbouring Group Participation by Carbonyl Groups in Ester Hydrolysis

171

KEITH BOWDEN 1 2 3 4 5

Introduction 171 Catalysis by carbonyl groups 172 Intramolecular catalysis in ester hydrolysis 173 Implications for enzymic catalysis 202 Conclusions and summary 203

Electrophilic Bromination of Carbon-Carbon Double Bonds: Structure, Solvent and Mechanism

MARIE-FRANCOISE RUASSE 1 Introduction 208 2 Methods for obtaining reliable bromination rate constants 21 1 3 Bromine-olefin charge transfer complexes as essential intermediates in bromination 216 4 The ionic intermediates: bridged bromonium ions or open fi-bromocarbocations 220 5 Kinetic substituent effects 243

207

CONTENTS

6 Solvent effects and solvation in bromination 267 7 The reversible formation of bromonium ions 279 8 Concluding remarks 285 Author Index

293

Cumulative index of Authors

303

Cumulative index of Titles

305

ADVISORY BOARD W. J. Albery, FRS University of Oxford, Oxford A. L. J. Beckwith The Australian National University, Canberra R. Breslow Columbia University, New York L. Eberson Chemical Center, Lund H. Iwamura University of Tokyo G. A. Olah University of Southern California, Los Angeles Z. Rappoport The Hebrew University of Jerusalem P. von R. Schleyer Universitat Erlangen-Niirnberg G. B. Schuster University of Illinois at Urbana-Champaign

Preface

In pursuit of the objective of the series, which is to present considered reviews of areas concerned with the quantitative study of organic compounds and their behaviour - physical organic chemistry in its broadest sense - in a manner accessible to a general readership, this twenty eighth volume contains five contributions on a diversity of topics. Two of these reflect the increasing importance of physical organic studies in providing fundamental knowledge relevant to the development of new materials with novel physical properties. The others represent more traditional areas of physical organic interest, where recent research has thrown new light. As ever, the Editor and his Advisory Board invite comments, criticism (preferably, but not necessarily, constructive) and proposals from readers and potential writers. Suggestions concerning developments in physical organic chemistry where a forward-looking review might help in the development of a new research area, or of established topics in need of an up to date treatment, should be directed to any of us. D. BETHELL

vii

Contributors t o Volume 28

W. John Albery Physical Chemistry Laboratory, University of Oxford, South Parks Road, Oxford, UK Edward M . Arnett Department of Chemistry, Duke University, Durham, North Carolina 27706, USA

M. Baumgarten Max-Planck-Institut fur Polymerforschung, AckermannWeg 10, Mainz, Germany Keith Bowden Department of Chemistry and Biochemistry, University of Essex, Wivenhoe Park, Colchester, Essex C 0 4 3SQ, UK Noel G. Harvey Exploratory Plastics Research, Rohm and Haas Company, PO Box 219, Bristol, Pennsylvania 19007, USA

W. Huber Hoffmann La Roche Ltd, Grenzacher Strasse 124,4002 Basel, Switzerland K. Mullen Max-Planck-Institut fur Polymerforschung, Ackermann-Weg 10, 6500 Mainz, Germany Philip L. Rose Exploratory Chemicals Research, Rohm and Haas Company, 727 Norristown Road, Spring House, Pennsylvania 19477, USA Marie-FranGoise Ruasse Institut de Topologie et de Dynamique de Systhmes, Universite de Paris 7, associe au CNRS-URA 34, 1 rue Guy de la Brosse, 75005 Paris, France

...

Vlll

Electron Storage and Transfer in Organic Redox Systems w i t h Multiple Electrophores M. BAUMGARTEN,~ W. H U B E R and ~ K. M ~ L L E N ~ Max-Planck-lnstitut fur Polymerforschung, Ackermann- Weg 10,6500 Mainz, Germany Hoffmann La Roche Ltd, Grenzacher Strasse 124,4002 Basel, Switzerland

Introduction 1 Design and synthesis 5 Highly charged states via extended redox sequences 10 Mechanisms of successive electron transfers 14 Structural factors relevant to intramolecular electron transfer The nature of the subunit 22 The length of the bridging group 25 The conformation of the bridging group 29 Ion pairing 32 The mode of linking 36 6 Conclusion and outlook 39 References 40

1 2 3 4 5

1

17

Introduction

Unsaturated carbo- and hetero-cycles are known to constitute active electrophores that can be subjected to chemical or electrochemical redox processes forming persistent cations or anions. Depending on the number of electrons transferred, the charged products can be either diamagnetic or paramagnetic. Crucial questions in describing the redox-activity of cyclic 7c-systems concern: (i) the number of accessible redox-states (Meerholz and Heinze, 1989, 1990); (ii) the mode of charge distribution (Heilbronner and Bock, 1978; Salem, 1966; Eliasson et al., 1986, 1990); (iii) the stabilization of charge by ion-pairing and/or conjugative effects (Hogen-Esch, 1977; Miillen et al., 1990); (iv) the possibility of charge-induced configurational and conformational changes (Eliasson et al., 1986; Huber and Miillen, 1986); (v) the chemical reactivity (Hogen-Esch, 1977; Szwarc, 1968; Miillen, 1986, 1 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 28 ISBN 0-12-033528-X

Copyright 0 1993 Academic Press Limited A / / rights of reproduction in any form reserved

2

M. BAUMGARTEN ET AL.

1987). A particularly attractive group of substrates are the annulenes, since electron transfer allows an interconversion of (4n + 2)n- and (4n)x-systems, and thus a switch between x-bond delocalization and n-bond localization (Mullen, 1984). A relevant step from the viewpoint of electron-transfer activity is the linking of two or more identical electrophores to bis-, oligo- or polyelectrophoric systems. In an attempt to extrapolate the description of the single electrophore to that of the higher homologues, the above questions maintain their significance, but additional attractive aspects arise. First of all, it is appropriate to subdivide multi-electrophoric species into two classes depending on whether the linkage between the building blocks is an unsaturated or saturated chemical unit. In the former case, with extended x-conjugation, one expects successiveelectron-transfer steps to create strongly interacting redox states in which the excess charge is distributed over the whole n-system. In the latter case, with electronically “decoupled” electrophores, one expects the electron-transfer steps to occur independently, with only electrostatic interactions of the redox states; the excess charge is localized on the electrophores (Smith et al., 1976; Flanagan et al., 1978). Within the domain of materials science, the classification of multielectrophoric organic systems according to the mode of linking of the subunits has led to the definitions of conducting polymers and redox polymers (Wegener, 1981; Baughman et al., 1982; Heinze, 1990; Nowak et al., 1980; Murray, 1984). The present review is restricted to “dimeric” or (low molecular weight) “oligomeric” redox systems with identical, electronically decoupled electrophores A. Consider the biselectrophore A-1- A, where 1 represents a saturated spacer (Fig. 1A). Injection of an electron into one subunit A under formation of a radical anion raises the immediate question of whether the electron will tend to localize in the original subunit or whether it will undergo a degenerate electron transfer to the neighbouring unit. In the terminology of electron-transfer kinetics such electron hopping between identical redox groups is termed self-exchange (Cannon, 1980). Depending upon the rate of this process, one will observe an “effective delocalization” over two or more units within the timescale of the experiment, or a localization of charge on one unit. Upon further contact with a redox reagent or at higher redox potentials, additional electrons can be transferred. After a two-electron transfer, each redox unit can accept one charge with formation of a singlet or triplet dianion, or (less favourably from an electrostatic point of view) both charges can enter one redox unit. Here again, an intramolecular electron-exchange process is possible. Further relevant questions concern the nature of the highest accessible

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

A-I-A

a.

3

a. or

m a. or

m

or

Fig. 1 Charging of a biselectrophore (A) and a triselectrophore (B).

redox state and the way in which the molecule copes with the increasing electrostatic repulsion. Clearly, the charging mechanism of the substrate and the particular electron-transfer kinetics are expected to be more complicated if the redox groups are different; this can be the case in the homologous tris-electrophore (Fig. lB), since the outer and inner electrophoric units now have a different substitution pattern, or in the bis-electrophore A-1-A when one unit A is slightly perturbed, e.g. by alkyl substitution. A major concern of this review is the tailoring of the redox behaviour of organic compounds, i.e. the optimization of such systems for electron storage and electron hopping. While the emphasis is on reduction and thus on anion formation, it has been shown on many occasions that oxidative cation formation leads to analogous conclusions (Meerholz and Heinze, 1990; Lewis and Singer, 1965). The structure of this text is thus obvious. (a) One will first have to consider the design and synthesis of suitable systems in which the structural conditions relevant for the energy profiles of inter- and intra-molecular electron-transfer processes can be systematically varied. The variation of the structure comprises the nature of the redox-unit A and of the spacer 1. It will be shown that

4

M. BAUMGARTEN ET AL.

this can be brought about readily by the synthetic technique of reductive alkylation (Mullen, 1984, 1986, 1987). (b) The second question is whether such species can serve as efficient electron acceptors in successive charging reactions. This aspect, if expressed in terms of structure, will focus on electrophores A that can be charged to a high charge density and on spacer groups 1 that tend to minimize the consequent Coulombic repulsion. It is obvious that this approach is closely related to the search for organic electronstorage materials that can be used as battery electrodes ( McDiarmid, 1979; Bitthin et al., 1987; Shacklette et al., 1987). (c) A third question concerns the sequence of the successive charging processes. It is clear that, depending on the particular charging mechanism and upon the number of electrons transferred, one can arrive at para- or dia-magnetic products that differ in the prevailing charge (spin) density distribution. (d) The fourt!? and main topic is how the intramolecular electron transfer between the redox groups of the systems depends upon structural phenomena. The energy profiles of intramolecular electron-transfer processes are important for many areas of chemistry. Some representative examples are (i) the photochemically induced electron transfer and the lifetime of charge-separated states in porphyrin-quinone diads, which serve as model compounds in photosynthetic studies (Gust et al., 1986, 1988; Wasielewski et al., 1985); (ii) the possibility of a long-range electron transfer and the relation between rate and reaction enthalpy according to the Marcus theory (Marcus, 1956, 1963, 1965) in benzenoid species attached to a steroidal spacer (Closs and Miller, 1988; Closs et al., 1989; Liang et al., 1990); and (iii) the influence of chain length and chain conformation on electron transfer between metal complexes (Reimers and Hush, 1990; Gray and Malmstrom, 1989), e.g. at the ends of an oligopeptide chain (Isied et al., 1988). In contrast with these approaches, the present account is restricted to degenerate, intramolecular electron transfer occurring in the charged ground state of the bis-electrophore A-1-A. It should be emphasized that our approach is a purely empirical one: knowledge of the structural dependence of the charge-storage capacity and of intramolecular electron-hopping processes might enable us specifically to design dimers, oligomers and polymers in which the electron-transfer rate, and thus the resulting charge distribution, can be controlled precisely.

5

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

2

Design and synthesis

It is appropriate to begin with an overview of the compounds that have been specifically selected and synthesized for the present purposes. They are made up of redox-active subunits such as the benzenoid hydrocarbons naphthalene [ 11 and anthracene [2], the [4n]annulene cyclo-octatetraene [3] and the bridged (aromatic = 4n + 2) [ 14lannulene [4]. Each subunit is known to form a radical anion and dianion in chemical or electrochemical reductions (Meerholz and Heinze, 1989; Rabinovitz, 1988). Cyclo-octatetraene is an outstanding electrophore because, in the course of an electron transfer, its tub-shaped structure undergoes a flattening of the ring, while the other hydrocarbons possess rigid n-systems that do not experience extensive structural change upon undergoing a redox process (Katz, 1960; Anet et al., 1964; Heinze et al., 1974). R

2 m

3 4

0 0 -

5

When linking these redox-active building blocks to form dimers and oligomers, care has to be taken to vary the steric and electronic interaction of the subunits systematically. In [51 an orthogonal arrangement of the two of the subunits systematically. In [51 an orthogonal arrangement of the two cyclo-octatetraene units is enforced (via the spiroconjugation). Compounds [6]-[ 103 contain a flexible connection of the anthracene species, while in the arrangement. Another type of face-to-face-arrangement is found in the multi-layered annulene systems [ 141 and [ 151. Here, unlike the para-cyclophanes, in which the phenyl units are connected by “external” alkanediyl groups, the stacking of annulene layers is achieved by “internal” linkages, with the bridge located inside the n-clouds. The advantage of the latter structure is that one can

M. BAUMGARTEN ET AL.

6

&yJ C6l

a:n=2 b:n=3 c:n=4 d:n=6 e:n=ll

3

&

‘ 0

C8l 0

c71

a:n=3 b:n=4

%I&0

c91

0

0

0

0

0

0

L

Jn

a: m = 3, n = 1 b:m=3,n=2 c: m = 11, n = 2

readily vary the inter-plane distance and that the multi-layered system can be created in a true polymer forming reaction (see below). The common feature of compounds [5]-[15] is that the electrophoric units are linked by saturated spacers, thus establishing only weak electronic (through-bond or through-space) interaction of the .n-systems. In contrast, the binaphthyl [ 161, the biperylenyl [ 171 and the bianthryl [ 181 as well as the structurally related homologues [19], [20] and [21] allow for a direct .n,.n-interactionof the subunits; it will be shown, however, that for both steric and electronic reasons the inter-ring conjugation can be weak and thus lead to electronically independent redox groups in a similar fashion as in [51-[151.

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

[14]a: R = CH3

b: R = 2-dodecyl

%%

[ 15]a: R = CHs

b: R = 2-dodecyl

%p%%

&6 '0 0

T p

I/

c 171

21 a : R = R ' = hexyl 21 b: R = H,R'= lsopentyl

In

c 191

21c: R = hexyl

7

a

M. BAUMGARTEN H A L .

[22]

a:n=l b:n=2

c:n=3

Compounds [6] and [7] have been described in the literature (Mullen, 1987; Fiedler et al., 1986; Huber et al., 1983). The ortho-anthracenophane [13] has been prepared by Diels-Alder cycloaddition as part of a project devoted to ladder-type polymers (Wagner et al., 1988; Wegener and Miillen, 1991; Pollmann et al., 1990). Compound [ 171 as well as the oligomers and polymers [ 19)-[22] have been prepared recently using various methods of aryl-aryl coupling (Bohnen et al., 1990; Fahnenstich et al., 1989; Koch and Mullen, 1991; Baumgarten et al., 1992a; Schenk, 1989). The most appropriate method, however, for the versatile linking of conjugated hydrocarbons by saturated spacer groups is reductive alkylation (Mullen, 1984, 1986, 1987; Bender et al., 1988, 1989; Bender and Mullen, 1988; Krummel and Mullen, 1988).This approach is based on the formation of carbanionic hydrocarbons by reduction of conjugated n-systems or by deprotonation of dihydroprecursors and their subsequent reaction with electrophilic reagents such as haloalkanes in SN2-type reactions. For the synthesis of the bis-cyclo-octatetraene compound [51 (Krummel et al., 1987; Auchter-Krummel and Mullen, 1991), cyclo-octatetraene dianion was quenched with tetrabromoneopentane to give the bis-adduct [231, which exists in an equilibrium between valence isomers [23a) and [23bl. Hexacycle [23a] was actually isolated in about 60% yield (Fig. 2) (Krummel et al., 1987). Accordingly, in the subsequent dehydrogenation, the formation of [23a] must be avoided by working at low temperatures; in this case it was possible to deprotonate the originally formed isomer [23b], obtaining a

Fig. 2 Quenching of cyclo-octatetraene with tetrabromoneopentane.

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

9

tetra-anion that was then subjected to oxidation with cadmium chloride to obtain the target molecule [51. This concept of reductive alkylation can be extended to polymer synthesis (Bender et al., 1988, 1989). A typical starting compound is the anthracene dianion, which is treated with a bifunctional electrophile such as a 1,n-dihaloalkane to obtain the intermediate [241. The latter constitutes both a nucleophile and an electrophile and one can now attach a second carbanion such as [25] to the electrophilic end. This process, after protonation and subsequent dehydrogenation, gives rise to the dianthrylalkane compound [6]. A related procedure provides dianthryl compounds such as [7] in which the oligomethylene spacer is replaced by an oxoethylene chain (Fiedler et al., 1986). Under appropriate experimental conditions the bifunctional intermediate [241 can undergo polymerization. Not surprisingly, the polymerization also proceeds upon mixing of the bis-nucleophile [261 and the bis-electrophile [27]. The yield of the polymer is higher than 95%, the average molecular weight is about 10000, and the product possesses high structural homogeneity. Finally, the poly( dihydroanthrylene) systems can be dehydrogenated to obtain the corresponding unsaturated polyanthrylene compounds [lo] (Bender et al., 1988, 1989).

A crucial aspect of the present carbanion alkylation is its high regioselectivity. Thus, under the prevailing experimental conditions, the kinetically controlled alkylation only proceeds at the positions of the highest local n-charge of the carbanions. In the anthracene dianion or in [25] these positions are at C-9 and C-10, respectively. An important consequence of this regioselectivity is given for the example of the pyrene isomer [28]. The corresponding dianion can be regioselectively alkylated at the two inner positions since these are the positions of the highest local n-charge. When applying a haloalkane as electrophile, the bridged [ 14lannulene [4] is obtained, and by carefully controlling the stoichiometry of nucleophile and electrophile one can, for example, obtain the doubly layered system [ 141

10

M. BAUMGARTEN H A L .

and the triply layered system [ 151 with varying spacer groups. This reaction can even be extended to polymer synthesis yielding [29b1, where, again, the crucial intermediate [29a] is obtained after alkylation of the dianion [28]’with a bifunctional species such as [271 (Irmen et al., 1984; Alexander et al., 1989). According to crystal-structure analysis, the products do indeed adopt a face-to-face arrangement of the annulene layers (Irmen et al., 1984; Maresch et al., 1989). In general, the technique of carbanion alkylation is the method of choice for providing a broad series of compounds for the study of electron-transfer reactions. In this way, indeed, it is possible systematically to vary (i) the type of the n-conjugation and the rigidity of the electrophoric subunits; (ii) the length and conformational behaviour of the spacer; and (iii) the relative orientation and the degree of overlap of the n-systems. These structural factors are relevant for the redox-activity of bis- and oligo-electrophoric systems and for the reorganization energy associated with intramolecular electron hopping.

3

Highly charged states via extended redox sequences

The first question is whether the redox systems can be subjected to successive electron-transfer reactions in extended redox sequences. What one needs to know thereby are the number of charges that can be transferred and what is the Coulombic repulsion arising between the charged subunits. The experimental methods that have to be applied are obvious. Cyclic

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

11

voltammetry (Heinze, 1984) provides information on the number of charges being transferred and on the energies of the available redox states. The ionic species formed along an extended redox sequence can be characterized spectroscopically: diamagnetic products by nmr spectroscopy, and the paramagnetic products by esr and ENDOR spectroscopy (Eliasson et al., 1990; Huber and Mullen, 1986; Mullen, 1987; Klabunde et al., 1987; Gerson, 1967, 1970; Kurreck et al., 1988). Quenching reactions can also be applied since they reflect the number of charges in the reduced system, and they can prove that the framework of molecules remains intact upon charging. From the pronounced tendency of cyclo-octatetraene toward dianion formation upon chemical or electrochemical reduction, it is not surprising that the rigid bis-electrophore [51 transforms into a tetra-anion upon chemical or electrochemical reduction. The geometry of the tetra-anion does not lead to significant Coulombic repulsion. The effective compensation of electrostatic energies in oligocyclo-octatetraenyls is even more astonishing if the subunits are brought into conjugative interaction. This is the case in the linear n-chains [30] and [31] that have recently become available and

r

c:n=2

c 321

12

M. BAUMGARTEN E T A L .

in which each cyclo-octatetraene unit accepts two electrons upon reduction ( Auchter-Krummel and Mullen, 1991; Staley et al., 1985).

The situation is less obvious for other electrophores such as anthracene, which upon electrochemical reduction forms a dianion salt, although at a much more negative potential than cyclo-octatetraene. The redox activity of a corresponding “dimer” will, of course, depend upon the chemical mode of linking. The example of l,n-di(9-anthryl) alkanes [6] is revealing (Huber et al., 1983; Becker et al., 1991). Here, the Coulombic interaction between two monocharged anthracene units depends sensitively on the length of the alkanediyl spacer. This energy is reflected by the difference between the first and second reduction potentials in the cyclic voltammogram (Heinze, 1986). Thus an interaction energy can be detected for spacers shorter than butanediyl. Further charging steps up to tri- and tetra-anions are distinctly shifted to more negative potentials than for the anthracene dianion itself and, additionally, a higher repulsion between the doubly charged redox groups is reported (Mortensen et al., 1991). Accordingly, dianthryl compounds with a close proximity, such as [6] and even [ 181,are capable of accepting four electrons although an appreciable electrostatic repulsion is built up (Becker et al., 1991; Huber and Miillen, 1980).When considering the question of how a bis-electrophore accommodates the extra charge it is important to note that, for example, the tetra-anion of di(9-anthry1)ethane [61 adopts an anti-conformation with respect to the central C-C bond, thus minimizing the electrostatic repulsion (Huber et al., 1983). In this context it is noteworthy to refer to the unsaturated analogue 1,2-di(9-anthryl)ethene [ 321 (Weitzel and Mullen, 1990; Weitzel et al., 1990). Like [6] (Becker et al., 1991), compound [32] forms a stable dianion and tetra-anion upon reduction. In the cyclic voltammogram of [321, the first two electrons are transferred at nearly the same potential, pointing to an effective minimization of the Coulombic repulsion between the charged anthryl units (Bohnen et al., 1992). This situation, which again corresponds to that in [6], could imply a torsion about the central olefinic bond (Bock et al., 1989). In contrast to the conformationally mobile dianthrylethanes, the rigid cyclophane [ 111, with a face-to-face arrangement of the n-layers, is electrostatically less favourable for reduction and only gives rise to a dianion upon alkali-metal reduction (Huber et al., 1983). It is straightforward in the charging of a layered electrophore to increase the interplane distance of the n-layers and thus to “relax” the resulting Coulombic strain; this approach is discussed for the doubly layered annulenes [ 141. In this particular case, a stable tetra-anion is available which can be characterized by a highly resolved ‘H-nmr spectrum (Irmen et al., 1984).

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

13

There are two remarkable findings: (i) While the neutral compound is made up of two diatropic [14]annulenes, the tetra-anion is composed of two strongly paratropic dianionic subunits; therefore the reduction is accompanied by drastic chemical shift changes, i.e. the ring protons of the tetra-anion appear at very high field (A& ca. - 8 ppm) and the bridge protons at very low field (AdH ca. 18 ppm). (ii) Unlike the neutral compound, the tetra-anion exists in two isomers that from an analysis of the 'H-nmr chemical shifts can be identified as syn- and anti-conformers of [ 141. Surprisingly enough, these two conformational isomers of the tetra-anion do not interconvert even at about + 80°C. Another question concerns the behaviour of redox systems containing more than two separate electrophores. It is characteristic that the tetra-anthrylene [ lOc] with a long undecanediyl spacer group can be reversibly charged with two electrons per electrophore without significant electrostatic interactions between the anthracenes (Becker et al., 1991; Mortensen et al., 1991; Bohnen et al., 1992). A related redox activity is found for the trisanthrylene [ lOa] with short propanediyl spacers (Bohnen et al., 1992). It should be noted, however, that in the cyclic voltammogram there is only one wave for a three-electron transfer; in other words, each anthracene can be charged with one electron without the creation of strong Coulombic repulsion, and only further reduction will then give rise to detectable electrostatic effects. Independent evidence for the redox capacity of these systems can be obtained from quenching reactions; the structure of the hydro-derivative obtained upon protonation of the charged hydrocarbon, i.e. the number of electrophiles being incorporated, reflects the number of charges in the anion. Figure 3 indicates that one can thus chemically prove the formation of both a dianion and a tetra-anion. This example is significant because the reduction of 1,2-di-(9-anthryl)ethanehas been claimed in the literature to cause a rapid cleavage of the ethane o-bond (Gerson et al., 1976; Hammerich and Saveant, 1979). In contrast, under appropriate experimental conditions, even the tetra-anion is a stable species (Huber et al., 1983). The homologous series of oligo( 1.4-naphthylene)~[ 191(Anton et al., 1992) and oligo(9,lO-anthrylene)~ [21] (Baumgarten et al., 1992a) are examples of directly connected electroactive units. In their cyclic voltammograms the first reduction and oxidation potentials are independent of the chain length due to the steric hindrance of conjugation. The second charging step is facilitated with increasing number of electroactive units because of the smaller Coulombic repulsion between the charges residing on the outermost units.

+

14

M. BAUMGARTEN E T A L .

Fig. 3 Quenching of the di- and tetra-anions of 1,2-dianthrylethane [6a] with protons.

At the end of a charging process, nearly every n-unit is singly charged in the case of the naphthylene derivatives and doubly charged in the oligoanthrylenes. The conclusion from the above examples is that under appropriate experimental conditions these systems can be subjected to successive electron-transfer reactions forming highly charged derivatives with intact molecular frameworks.

4

Mechanism of successive electron transfers

As has been pointed out already, the description of successive reduction processes meets with a fundamental problem. If one electron has been injected into a bis-electrophoric system (see Fig. l), the second charge can enter the same unit, forming a dianionic moiety, or it can enter the second neutral unit. The former case is only conceivable if the two extra charges on one redox group give rise to conjugative or ion-pairing effects compensating for the electrostatic repulsion. In the latter, electrostatically more favourable, case the dianion could exist either as a singlet or a triplet species. The situation is even more complicated when a tris-electrophoric system is charged. The first question is which subunit will be charged initially. Then, if a dianion with the two electrons in separate electrophores has been formed, the charges can reside either on neighbouring electrophores or on those allowing the greatest possible distance between charges (see Fig. 1 ). What experimental evidence is available to clarify the resulting mode of spin-density and charge-density distribution? If one characterizes the reduction of (i) a single annulene, (ii) a doubly layered and (iii) a triply layered analogue by cyclic voltammetry, one derives a first criterion (Alexander et al., 1989; Bohnen et al., 1992; Fry et al., 1985). The potential

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

15

wave for the diannulenyl [14] shows two very close redox steps, each describing a one-electron transfer. The simulation of these waves provides the potential difference for the first and second reduction; from this, after correcting for the entropy term, one obtains an interaction energy of about 1.7 kcal mol- '. In a similar fashion, the triply layered analogue [ 151 reveals one wave indicating a three-electron transfer. The interaction energy corresponding to a two-fold charging of [ 151 is below 0.2 kcal mol-'. Accordingly, one must describe the dianion of [l5] as a species in which the first two extra electrons reside in the outer layers. Only the injection of the third electron, then, creates a significant interaction energy. It is clear that the charging sequence can be affected not only by the (inner/outer) position of the electrophore within the oligomeric chain, but also by substituents slightly affecting the redox potentials. The oligo( 1.4naphthy1ene)s [ 191carry t-butyl groups only in the terminal units, in contrast to the corresponding oligo( 1,5-naphthylene)s [203, which are substituted with (cation-stabilizing) alkyl groups in each unit. This different substitution pattern explains why the oxidation of [20] occurs with greater ease ( A E z 110 mV) than that of [19] (Bock et al., 1989; Anton et al., 1992). The tris-9,lO-anthrylenes [21a] and [2lb] (Baumgarten et al., 1992a) possess a different number of solubilizing alkyl groups in the central anthracene unit. As a result, the first oxidation of [21a] occurs at a potential 120 mV lower than for [21b] owing to the second stabilizing group. When , alkyl substituents are placed turning to the tetrameric anthrylene [2 1 ~ 1the on the terminal units, which are then preferentially oxidized to the dication, and consequently no Coulombic repulsion is seen. The reduction, on the other hand, is favoured in the central unsubstituted units and the second charging step takes place against a higher interaction energy. As we have shown previously for two different series of oligo( p-pheny1ene)s [22] and [33] (Bohnen et al., 1991; Heinze and Meerholz, 1990), the substitution of central units of a chain can influence the charging behaviour drastically. In [331 with methyl substituents, the conjugation is interrupted as soon as two substituents interfere, leading to hindered rotation around the connecting bond. Thus, independent units of biphenyl ([ 33a], [33cl and [33e]) or terphenyl ([33b] and [33d]) - depending on the hindrance between the subunits -are charged, even in the case of long phenylene chains. While cyclic voltammetric experiments provide thermodynamic and kinetic information on the charging processes (Heinze, 1986), only indirect information on the structure of the redox products is available. Fortunately, independent evidence can be obtained from spectroscopic experiments. Figure 4 depicts the esr spectra recorded for both a dianion and a trianion of [21a] in solid solution. From the zero-field splitting D, which is proportional to the inverse distance of the unpaired electrons, one can roughly

16

M. BAUMGARTEN ET AL.

estimate the location of the charge; there is no doubt that in the dianion the two electrons have entered the outer anthracene moieties. Analogous findings can be obtained for the related tetramer [21c] (Baumgarten et al., 1992a). The detection of zero-field splitting for dianions of [ 181 or [21] is very important; it reveals not only the existence of a triplet state, but it also provides information on the mode of spin density distribution. Even more

Fig. 4 Esr spectra of high-spin trianthrylene [2la] at 120K in MTHF/K: A, biradical triplet state [2la]*-' (labelled 0); B, triradical quartet state [21aI3-' (labelled *); C , biradical triplet state [21aI4-' (labelled m).

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

17

subtle structural information is available from the esr spectra, however. Thus, for the conformationally mobile dianthrylalkanes [6]-[9], the zero-field splittings observed for solid solutions of the dianions together with model considerations indicate that the oligomethylene spacers adopt an all-anti conformation, so that the redox-active subunits have the largest possible separation. A related finding appears for the dianions of the doubly layered system [ 141; their esr spectra taken in the glass provide evidence for a triplet structure, and from the D-value the inter-plane distance in the charged species can be readily determined (Irmen et al., 1984; Alexander et al., 1989). The important conclusion to be drawn from the cyclic voltammetric and esr spectroscopic data is that the charging of redox systems with electronically separate units follows a certain sequence. Thus, according to the esr spectra, the first electron will enter the inner unit of the triply layered system [ 151 while, according to the cyclic voltammogram and zero-field splitting derived from the esr spectra, the corresponding dianion has the two electrons in the outer layers. Accordingly, transition from the monoanion to the dianion will require a redistribution of charge, i.e. an electron-hopping process (Alexander et al., 1989).

5

Structural factors relevant t o intramolecular electron transfer

The occurrence of intramolecular electron-hopping processes will now be discussed in detail, considering on an empirical basis systematic variation of relevant structural factors in the substrates. First, however, a few theoretical and experimental aspects will be summarized. In a general description of intramolecular electron-transfer (ET) processes one has to differentiate between charge separation in donor/acceptor (D/A) systems via the formation of photoexcited states and a “charge-transfer” or “charge-shift” reaction that is thermally activated (Cannon, 1980; Fox and Chanon, 1988; Meyer, 1978). For systems such as A-1-A, where each subunit A can act as donor or acceptor, the latter case may also be described in some cases as resonance as in ( l ) , because the electronic configuration can be written without a

difference in energy. If we think of a charge transfer (Ulstrup, 1979) between particle A+‘/-‘ and A, however, it is possible to define a rate constant k for

18

M. BAUMGARTEN E T A L .

this process, which, according to the Arrhenius equation, possesses an exponential temperature dependence (2), where A is a proportionality factor,

E , the activation energy,k , the Boltzmann constant and T the temperature. Following Eyring’s transition-state theory (Eyring, 1935a,b), one may replace A by Z K , where Z is either the collision frequency in bimolecular reactions or the vibrational frequency in intramolecular reactions in bridged systems, and K is the transmission coefficient, i.e. the probability that activated complexes (transition states) yield the product. By using the free energy of activation AG*, Marcus (1956, 1963, 1965) described the rate constant as a function of AG* and the thermal energy k,T as in (3). At large values of r, k = Zrc(r)exp( -AG*/k,T).

(3)

is assumed to change exponentially with r. The kind of process involved (adiabatic or non-adiabatic) depends on the J may be coupling J between the wavefunctions Y(+”-’lo)and Y(ol+.’-’). expressed as a function of overlap S of both states (Heilbronner and Bock, 1978; Hoffmann, 1971) by (4), in which the parameters are defined as in (4a) K

and He, is the Hamiltonian for the system in the Born-Oppenheimer approximation. The adiabatic and non-adiabatic processes may be visualized in terms of two intersecting potential curves, one of them representing the electronic configuration ( +./-.lo) and the other one the configuration (01 +./ -.). This is demonstrated in Fig. 5A-C for the cases of very strong, strong and weak coupling, where the first two examples are generally called adiabatic and the last non-adiabatic. A very strong coupling (large value of J ) results in the formation of a single minimum surface with a symmetric chargedelocalized ground-state ion (either anion or cation, Fig. 5A). For somewhat smaller values of J the process is still adiabatic, and the resulting surface has a double-minimum potential. Interconversion between the degenerate charge-localized states takes place via the avoided crossing region on the

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

4

A

B

19

C

nuclear coordinates

Fig. 5 Potential energy hypersurfaces as a function of the reaction coordinate for adiabatic (A, single-minimum potential; B, double-minimum potential) and non-adiabatic (C) electron-transfer reactions.

lower surface, while the electronically excited state surface has a single minimum. The magnitude of the avoided crossing region is given by twice the electron coupling integral J (Fig. 5B).In the non-adiabatic region with two separate curves an energy of 1would be necessary for electron transfer between the states, if no reorganization of the solvation shell takes place. Thermal activation of the solvent to a configuration in which the free energy of the ion is unchanged, is, however, possible at lower energies. For isoenergetic parabolic potential curves the energy of this point is $2 (Fig. 5C). Comparing J with the thermal energy, one can differentiate between the two processes. For adiabatic surfaces the coupling J >> kb’T; while in the non-adiabatic case J << kb i? A quantitative description of the two types of process can be made using the Landau-Zehner parameter g (Ziilicke, 1985; Onuchic et al., 1986) defined in ( 5 ) . Here, o,is the frequency of the vibrational g=-

2rcIJ(’ ho,1

mode determining the ET. If g >> 1, the reaction is adiabatic, and for g << 1 it is non-adiabatic. If we consider ET in the ground state, we shall deal with an adiabatic case where the occurrence of charge transfer depends on the activation energy between the double-minimum potential (Meyer, 1978). When defining the activation energy AG +, one typically uses the approximations of harmonic

M . BAUMGARTEN ET AL.

20

oscillators and parabolic potential curves, yielding a quadratic dependence (6) on 1 and AGO (Marcus, 1956, 1963, 1965; Closs and Miller, 1988; AG+

= (AGO

+~ ) ~ / 4 1

(6)

Closs et al., 1989; Liang et al., 1990; Hush, 1961, 1975), where 1 is the total reorganization energy, and AGO the difference in free energy of the two minima. The reorganization energy 1 is the sum of two terms, As + ili, the inner sphere reorganization ,Ii for changes of D/A (structure, geometry, and vibrational states) and the solvent reorganization 1,.In a simple model A, can be expressed by (7) in terms of the optical (cop) and static ( E , ) dielectric

1, = e2 +al + +a2 -

(

--)(& 1

-

t)

(7)

constants of the medium, the radii of donors and acceptors ai, the electronic charge e and the centre-to-centre distance r (Closs and Miller, 1988; Closs et al., 1989; Liang et al., 1990; Wasielewski, 1989). il may be determined experimentally, for example from charge-transfer bands ( Wasielewski, 1989; Siders and Marcus, 1981) or from the self-exchange rates for identical subunits where AGO is zero and k,, depends only on 1 through (8) (Grampp et al., 1990a).

k,,

= ZKexp[

+

-

'1

4AkbT

In the quantum mechanical formulation of electron transfer (Atkins, 1984; Closs et al., 1986) as a radiationless transition, the rate of ET is described as the product of the electronic coupling term J 2 and the Frank-Condon factor FC, which is weighted with the Boltzmann population of the vibrational energy levels. But Marcus and Sutin (1985) have pointed out that, in the high-temperature limit, this treatment yields the semiclassical expression (9). k

= 4n ~

h

J2(

-)'"

1 4nlk,T

exp[ -

+

41kb T

(9)

Thus the distance dependence of ET can be divided into the distance dependence of J and that of 1.For J an exponential decrease with increasing distance is assumed (Wasielewski, 1989). A principal distinction is made between outer-sphere and inner-sphere mechanisms in ET reactions (Kochi, 1988). In the outer-sphere reactions the

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

21

donor and acceptor are viewed as separate units even in the transition state of the transfer, and the kinetics show an exponential decrease with increasing distance (Kuznetsov and Ulstrup, 198 1). In the inner-sphere mechanism a compact transition state is formed where the coordination spheres of D and A penetrate each other (overlap), and the structure of the reactants is changed. The outer-sphere reactions may be described by the Marcus theory. This approach is widely used for explaining inter- and intra-molecular chargetransfer processes including systems such as A-1-A (Kuznetsov and Ulstrop, 1981) as long as no direct overlap occurs between the orbitals of each A (Fukuzumi et al., 1980; Hay et al., 1975). A number of different techniques have been applied to test the distance and orientation dependence of ET reactions (Closs and Miller, 1988; Closs et al., 1989; Liang et al., 1990; Reimers and Hush, 1990; Fox and Chanon, 1988; Wasielewski, 1989; Paddon Row and Jordan, 1988; Joachim et al., 1990; McConnell, 1961). Our method of analysing the mode of charge distribution in charged species is esr spectroscopy, which defines the timescale of the detectable dynamic species (Gerson, 1967; Kurreck et al., 1988; Wertz and Bolton, 1972). If an electron transfer is slow relative to the esr timescale ( < l o - ’ s) the spectrum corresponds to that of monomeric model compounds with a single electrophore. If the hopping process is rapid on the esr timescale, one will detect an effective delocalization. The charged electrophores are prepared by chemical or electrochemical electron transfer (reduction or oxidation) in solution and exist as ion pairs (Hogen-Esch, 1977; Huber and Miillen, 1986; Gerson, 1967). This situation differs significantly from that in which, for example, radical anions are prepared by pulse radiolysis (Closs and Miller, 1988; Closs et al., 1989; Liang et al., 1990). In the latter only “naked” anions exist, which possess no interaction with counterions. As has been described above, intramolecular hopping of electrons between electrophoric units requires a reorganization of the ion pair and of the solvent shell. Not surprisingly, therefore, the rate constants of such processes are similar to those observed in pulse-radiolysis experiments. It should be added that under the prevailing experimental conditions (e.g. high dilution) intermolecular electron-exchange processes are slow and do not affect the appearance of the esr spectra (Kurreck et al., 1988; Becker et al., 1991). In several cases studied in the literature the rates of exchange processes are within the esr timescale and give rise to dynamic line broadening (Ward and Weissman, 1957; Freed and Fraenkel, 1964). With the assumption of a short time for the jump itself, an adequate description of the exchange process is given by the modified Bloch equation (Gutowsky and Holm, 1956). The lineshape for an equilibrium is then described by the complex transverse magnetization M . The alternating linewidth effects have been particularly

M. BAUMGARTEN E T A L .

22

useful within kinetic analyses (Freed and Fraenkel, 1964; Sullivan and Bolton, 1970). Once the rate constants have been evaluated by spectral analysis for a series of temperatures, the determination of the activation parameters is achieved. While similar examples will also be reported herein, the emphasis is on the discrimination between localization of spin density in a single redox unit and effective delocalization over the entire molecule, i.e. on a yes/no decision between two borderline cases. Thus, in the present approach, the major focus is on the question of how we can influence the “external” parameters like solvent and counterion and the “intrinsic” structural parameters within the systems A-1-A to force the electron-hopping process into the timescale of the experiment, or at least to establish clearly the borderline cases. That we are still looking at an electron-hopping process in the case of “effective charge delocalization” over the entire molecule and not at a pure resonance phenomenon may be reassured by VIS/NIR spectroscopy of the neutral and charged species; the absorption of a single chromophore should be detected unless a very fast process > lo’, Hz is taking place. THE NATURE OF THE SUBUNIT

The first structural factor that is relevant for electron hopping is the electrophoric subunit. The reduction of the spirodicyclo-octatetraenylsystem [ 5 ] again serves as a suitable introductory case. Upon prolonged metal contact in solution [ S ] can be charged to a tetra-anion, where the number of H- and 3C-nmr signals reflects the D,, symmetry (Krummel et al., 1987; Auchter-Krummel and Mullen, 1991;Krummel, 1988).The intermediate ions such as the monoanion and the dianion are more interesting since they can represent “mixed”-valence states (Reimers and Hush, 1990; McConnell, 1961). It follows from the esr spectrum of the monoanion, which closely corresponds to that of a “monomeric” cyclo-octatetraene radical anion, that the unpaired electron is localized on one ring and that, even at elevated temperatures, there is no rapid hopping between the two rings. A completely analogous finding is obtained for the corresponding dianion. The 13C-nmr spectrum of the latter can be looked at as a superposition of a “neutral” and a “dianionic” spectrum. It is thus clear that the two excess charges are both localized on one subunit; even at +90°C the spectra fail to show any changes and to reveal any hopping process. Another important point can be deduced from the splitting pattern of the methylene protons in the ‘H-nmr spectrum of the dianion salt [5I2-/2Li+. It appears from the equivalence of, for example, H-9,11 and the nonequivalence of H-9’,11’ that the charged ring is planar, while the uncharged ring is tub-shaped (Fig. 6) (Krummel et al., 1987). The latter

’

’

23

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

H - 9 a .b

H-9'a.b

T

O°C

I

11 I

8

6

.

.

~

'

~

t

'

'

4

~

'

~

~

'

T = -6OOC

.

2

.

"

'

.

'

~

~

'

.

.

0

Fig. 6 Nmr spectra ofthe dianion of [ 51 with the molecular positions and assignment of protons.

"

'

'

24

M. BAUMGARTEN E T A L .

undergoes a rapid ring inversion whose slow-exchange domain is reached at -60°C. The localization of two electrons in one ring and the failure to observe an intramolecular electron transfer in [ 51 and [512- is again related to the specific redox-behaviour of the parent system, cyclo-octatetraene (Anet et al., 1964; Krummel et al., 1987; Staley et al., 1985).It is characteristic for the latter that (i) upon injection of an electron, the ring experiences a flattening and that (ii) the electronic stabilization of the dianion overcompensates the Coulombic energy associated with the second electron transfer upon transition from the radical monoanion to the dianion (Katz, 1960; Anet et al., 1964; Eliasson and Staley, 1985). Such a flattening is also likely to occur for the second (uncharged) ring of [5]’- in the course of an intramolecular electron transfer. This flattening will require some steric strain, an energy term that contributes to the inner-sphere reorganization energy Izi of the intramolecular electron transfer, and this will increase the overall activation barrier. Accordingly, if we want to bring about a rapid intramolecular electron transfer between identical electrophores, we shall have to introduce rigid electrophores that do not experience a significant structural change upon charging. The “extra stabilization” of charged cyclo-octatetraene and the reluctance of cyclo-octatetraenes to be involved in intra- or inter-molecular electrontransfer processes is also documented by the behaviour of the cyclooctatetraenylenevinylenes [301 and [311 ( Auchter-Krummel and Mullen, 1991; Staley et al., 1985). In the dianion of e.g. [30a] and [31a] the charge is localized on one eight-membered ring, although a complete 7c-charge delocalization within the conjugated systems is not inhibited by a saturated spacer. The reason is again the steric strain associated with the flattening of the cyclo-octatetraene moiety. Characteristically enough, a charge-shift occurs in [30a12 - at temperatures above - 20°C whereas [3 la] does not exhibit dynamic behaviour even above room temperature. A localization of the spin density on one redox-group of a bis-electrophore has also been observed for the radical anion of spiro [5.51undeca-1,4,6,9tetraene-3,8-dione [341 (Gerson et al., 1972). The authors emphasize, however, that in this case the localization is mainly due to the interaction of the oxygen atoms with the counterions. In contrast to this finding, the unpaired electron is delocalized over both z-systems in the radical anion of the spirobifluorene [35] (Gerson et al., 1974). This seems to support the assumption that an effective delocalization can be brought about by introducing two rigid electrophores for which injection of an extra electron induces no conformational change; an orthogonal arrangement of the two electrophores, however, does not necessarily inhibit rapid electron transfer. In contrast to the dianion of the spirodicyclo-octatetraenyl [51, the dianion of the spirobifluorene [35] exists as a triplet biradical. This spin state of the -’

’-

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

25

dianion can be rationalized in terms of a double degeneracy of the lowest antibonding orbitals of [351.

THE LENGTH OF THE BRIDGING GROUP

1,n-Di-(9-anthry1)alkanes [61 are suitable for investigating the influence of the distance between the single redox groups A, since the length of the conformationally mobile alkane chain 1 can readily be varied. As pointed out above, injection of an electron into the bis-electrophore forming a radical anion raises two main questions: (i) is the unpaired electron localized in one subunit or does a rapid electron-hopping process occur so that within the timescale of the esr experiment one observes an effective delocalization over both units; and (ii) can one clearly distinguish between the extremes experimentally? The obvious criteria are the magnitude -of the esr/ENDOR hyperfine coupling constants and the multiplicity of the hyperfine coupling constants, which reflects the number of equivalent hydrogen atoms coupling with the unpaired electron (Kurreck et al., 1988; Wertz and Bolton, 1972). It is clear from Fig. 7 that the potassium salt of the dianthrylethane radical anion in a mixture of dimethoxyethane (DME) and hexamethylphosphoramide (HMPA) as solvent has the unpaired electron effectively delocalized over both units (Fiedler et al., 1986). This outcome, however, depends on the

M. BAUMGARTEN E T A L .

26

A

B

Fig. 7 Esr and ENDOR spectra of 1,2-dianthrylethane [6a] for different ion-pair conditions: A, DMEiHMPA, K'; B, MTHF, K + at 190 K in solution.

length of the alkanediyl spacer group. Thus, an analogous finding, i.e. effective delocalization over both anthracene units, can be made for the propanediyl-, and butanediyl bridged system, while the corresponding dianthrylhexane radical anion shows localization of the unpaired electron in one anthracene unit under all experimental conditions, both in different solvent systems and at different temperatures. In the latter case, the esr- and ENDOR spectra closely correspond to those of the 9-alkylanthracene radical anion (Fiedler et al., 1986; Huber et al., 1983). Thus an electron transfer is hampered by the activation barrier between the double-minimum potential; we can no longer differentiate between an adiabatic process with large activation and a non-adiabatic process (McConnell, 1961; Kuznetsov and Ulstrup, 1981). This occurs even for the flexible alkanediyl bridging chains, where a rapid rotation and reorientation of the spacer and electrophores is possible in solution. Furthermore, the delocalization is strongly distance-dependent, so that a through-space mechanism seems to be responsible for the exchange, while through-bond transfer is not favourable in longer alkanediyl chains. The average distance between the electrophores is too large for an effective through-space coupling; therefore, no Coulombic interaction (CV) and zero-field splitting (esr) is observed in the doubly charged species, which should be a biradical (Becker et al., 1991). The distance dependence of the electron-transfer rate can also be demonstrated for the non-benzenoid systems [ 141 and [ 151, which

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

27

incorporate [ 14lannulenes (Alexander et al., 1989; Irmen et al., 1984). Like the dianthrylalkanes, they possess flexible spacer groups; the connection of the Ir-layers, however, enforces a direct n,n-interaction (Irmen et al., 1984). The radical anion of the bridged [14]annulene [14] can again serve as reference system. Its esr hyperfine coupling constants are in accord with a simple MO model (Mullen, 1978). When one proceeds to the radical anion of the analogous doubly layered system with a propanediyl chain as a spacer [ 14a], one observes only half the magnitude of the coupling constants (Irmen et al., 1984). This is convincing evidence for an effective delocalization of the spin density over both layers, and thus for rapid electron transfer. When the inter-plane distance is increased, however, the esr and ENDOR spectra closely correspond to those of the monomer. Consequently, by setting the appropriate length of the spacer, one can control the rate of the degenerate electron transfer and switch between a spin-localized and spin-delocalized situation. Accordingly, from spectroscopic evidence we can clearly distinguish whether radical anions derived from bis-electrophores exhibit rapid or slow electron hopping on the timescale of the experiment. Furthermore, there is a certain maximum of the spacer length beyond which intramolecular electron transfer is no longer detectable. For long, flexible spacer groups and electronically “decoupled” electrophores, the contribution of spacer reorganization to the activation barrier of the exchanger process is too large. A theoretical treatment of electron hopping in bis-electrophores with flexible spacers will have to consider different conformations of the spacer. Inclusion of rigid biselectrophores (with a fixed distance) is therefore relevant. The question of “how far an electron can be transferred” in radical anions of bis-electrophores has been the subject of several articles (Harriman and Maki, 1963; Shimada et al., 1972; Shimada and Szwarc, 1974, 1975; Gerson et al. 1984); the systems studied possess both flexible and rigid spacer groups between the electrophores. The findings accumulated for a series of 1,n-dinaphthylalkanes (Shimada et al., 1972; Shimada and Szwarc, 1975) are in qualitative agreement with the above results on dianthrylalkanes. An analogous scaling for rigid spacers does not exist, but it was demonstrated for compound [36a] and higher homologues [36b] and [36c] with more than one spiro-bonded cyclobutane unit (Shimada and Szwarc, 1974; Gerson et al., 1984) that electron hopping can occur for distances up to 0.8 nm. Long-range electron transfer has also been reported in a recent esr/ENDOR study of stable radical anions of naphthalenes attached to rigid norbornane spacers of variable length (Gerson et al., 1990). No counterion or solvent effects were found for the dinaphthobicycloheptadienc [37al with a small edge-to-edge (e-e) separation [ r = 0.24 nm(e-e); r = 0.63 nm (centre-tocentre)], and complete delocalization could generally be detected. The esr/ENDOR spectra of dinaphthotetracyclododecadiene [ 37b] and di-

28

M. BAUMGARTEN ET AL.

naphthohexacyclohexadecadiene [37c] revealed that the electron transfer between the naphthalene moieties in the radical anion is slowed down, but still within the esr timescale. This occurs even though the centre-to-centre distances between naphthylene groups are r = 0.8 nm in [37b] and r = 1.03 nm in [37c]. The long-range interaction of n-systems across rigid saturated bridges has also been studied by PES in the gas phase, where the ionization potentials provide a direct measure of 25 (Paddon-Row and Jordan, 1988). These studies demonstrated that through-bond coupling is considerable and easily exceeds through-space coupling in rigidly linked olefins. Linear extrapolation of J-values for the cation radicals of [38a; n = 13 and [38b] as a function of the number of single bonds separating the alkene moieties allowed estimation of the rate constants for thermal electron transfer in the ground state cation radicals of [ 38a; n = 2, 3 and 41 by use of equation (9). In a study of long range ET between aromatic donor (biphenyl) and acceptor molecules separated by steroid spacers [391, pulse radiolysis and electron beam techniques have been used for the injection of electrons (Closs and Miller, 1988; Closs et al., 1989; Liang et al., 1990). Here, the reaction rates (observed by changes in the absorption spectra) pass through a

[37b]

a

a

c401

29

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

maximum for different acceptors, and the results confirm the inverted region predicted by Marcus (A < (AG'I) (Miller et al., 1984). For a cyclohexane and decalin bridge, the dependence of ET rates on the orientation of a biphenyl donor and naphthyl acceptor was studied. Although a considerable difference in the centre-to-centre distance results for two decalin isomers (2,7-a,a and 2,7-e,e) [40a] and [40b], the reaction rates did not change in a dramatic way. Through-bond coupling is therefore assumed to dominate over a through-space mechanism. THE CONFORMATION OF THE BRIDGING GROUP

Turning again to flexible spacers, it appears that in oligo( anthrylene) species [lo] a long, flexible (CH,),-chain improves the solubility of the compound, which is particularly important in the related polymers. The question therefore arises whether one can bring about rapid electron transfer even for long spacer groups and thus whether a fine-tuning of the electron-transfer behaviour is possible for a given spacer length. The solution of the problem comes from a conformational analysis of oligomethylene chains, which are known to adopt preferentially an all-anti conformation (Dale, 1974). A different situation is encountered in oxaethylene chains, which prefer a gauche conformation with respect to the carbon-oxygen bond, thus giving rise to a helical conformation of the chain. The crystal structure of the biselectrophore [7a] provides a convincing example (Fiedler et al., 1986). The corresponding hydrocarbon species with the same spacer length, the dianthrylundecane [6e], forms a radical anion that always exists as a spin-localized species. The esr spectrum obtained for the radical anion of the structurally related ether [7a] (see Fig. 8) is temperature-dependent, which

Ill

A. \

T=233K

Fig. 8 Esr spectra of the alkoxy-bridged dianthryl [7] at two temperatures: A, 190 K; B, 233 K.

30

M. BAUMGARTEN E T A L .

indicates that intramolecular electron transfer occurs on the timescale of the experiment. The measurement has been performed on the potassium salt in dimethoxyethane, and similar results have been obtained for the lithium salt in 2-methyltetrahydrofuran. By means of lineshape analysis, we can determine the free energy of activation of the intramolecular electron transfer as AG* = 3.5 kcal mol-’. A similar rate-increasing effect is detected with the homologous ether [7b] (Fiedler et al., 1986). When rationalizing the significant difference of the hydrocarbon- and ether-bridged radical anions, the main aspect will certainly be the conformation of the oxyethylene chain, which brings the electrophores into closer contact. An additional aspect follows from the ability of the oxygen centres along the chain to chelate the counterion and thus to fix the cation between the electrophores. It is not possible from the available experimental evidence to discriminate between the two effects. The role of ion pairing and the relative position of the counterion and carbanion will be dealt with below. There is, however, another obvious effect of conformation upon the electron-transfer rate. In the p-anthracenophane [ 111, the mode of linking enforces a face-to-face arrangement of the 7c-layers (Huber et al., 1983; Gerson et al., 1976). Correspondingly, for its radical anion effective delocalization has been detected under all experimental conditions in solution. It is not clear from the available evidence whether a hopping process occurs, thus giving rise to a double-minimum profile, or whether there is a true spin-delocalized structure representing a single-minimum energy profile. This question will be answered below. When comparing the esr hyperfine coupling constants of [ 111-. with those of the flexible dianthrylethane radical anion [61 -’, a close correspondence can be seen between flexible and rigid species (Becker et al., 1991). Even the coupling constants of the methylene protons are very similar, although they are known to be sensitively dependent upon conformation. What these data seem to suggest is that the dianthrylethane radical anion adopts a conformation in solution with an (at least partially) eclipsed position of the electrophores. Such a structure would differ from that of the triplet dianion [6a]’-., where the zero-field splitting suggests an anti-conformation of the anthryl units with respect to the central C-C bond (Becker et al., 1991). In agreement with this are the results for other “phane” structures, which may also be compared with their corresponding flexible alkanediyl bridged forms as in paracyclophane [41] (McConnell, 1961; Gerson and Martin, 1969), syn- and anti- [2,2] naphthalenophane [42] (Gerson et al., 1976; Hamacher et al., 1986; Konishi and Reddoch, 1977) and syn- and anti-[2,2]( 1,4)-anthracenophane [ 121 (Nemoto et al., 1980). For bridged naphthalenes with a flexible alkanediyl chain [43] and a more rigid cyclohexane bridge [441 it was found that the intramolecular electron transfer

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

31

is around five times faster within the open chain structure possessing the same number of bonds as for the rigid cyclohexane (Shimada and Szwarc, 1974). Aromatic spacers have also been used; thus it could be shown for a phenylene-bridged diporphyrinyl system that electron hopping between both subunits can occur (Huber et al., 1990). Here the phenylene bridge is twisted out of conjugation with the two porphyrin rings, and the effect is observed for 1,4-(para) [45a] and the 1,3-(meta) [45b] substitution. While the rate of an electron-transfer process and thus the mode of the resulting spin-density distribution can be controlled by the length and the conformation of the spacer groups, the ion pairing creates an additional factor that can serve as some “external” control even for a given choice of subunits and spacers.

Plr

Ar

[45a] Ar=CCHs,M=Zn

Ar

[45b] Ar=C6Hs,M=Zn

32

M. BAUMGARTEN €T AL.

I O N PAIRING

It has been pointed out already that formation of a radical anion by a redox process in solution produces an ion pair and that any hopping of the electrons will thus be bound to the migration of the cation, which then becomes rate-limiting (Gerson et al., 1972, 1974, 1990). The ,ion-pair structure of the radicals is mainly affected by the size of the counterions and the ion-solvating capability of the solvent (Hogen-Esch, 1977; Szwarc, 1968). The example in Fig. 7 is significant for the reorganization of the ion-pair structure in the course of an electron transfer. It has been shown already that the radical anion of dianthrylethane as its potassium salt in a DME/HMPA solution possesses an effective delocalization of the spin (Huber et al., 1983; Becker et al., 1991). If this solvent system is replaced by 2-methyltetrahydrofuran, leaving all other conditions unchanged, one can readily switch from a spin-delocalized to a spin-localized situation. It is known from various pieces of evidence (Huber et al., 1983; Szwarc, 1974) that 2-methyltetrahydrofuran has a poor cation solvating ability and strongly favours contact ion pairs, i.e. a tight interaction of counterion and carbanion, while, on the other hand, HMPA and dimethoxyethane with their strong cation coordinating ability generate solvent-separated ion pairs. If an intramolecular electron transfer requires a reorganization not only of the redox units but also of the ion-pair structure, it is clear that the existence of a tight ion pair increases the overall activation barrier of the electron-hopping process. Not surprisingly, therefore, increasing cation coordinating ability of the solvent allows one to accelerate the intramolecular electron transfer. This feature provides the opportunity of controlling the electron-transfer rate just by the experimental conditions. It thus constitutes an “external” factor in contrast to the “internal” factors such as the nature of the redox group and of the spacer. The same “control mechanism” can be put to work for the doubly layered electrophores [ 141. From esr measurements it appears that a change only of the ion pairing brings about a different hopping rate and creates a different spin-density distribution within the timescale of the experiment. The important role of ion pairing for the rates of intramolecular electron transfers has been addressed in nearly all articles concerned with this subject. The most direct pieces of evidence for the high influence of anion-cation interaction result from the investigation of radical anions of phane species such as [ l l ] , [12] and [41]-[43] (Huber and Miillen, 1980; Gerson et al., 1976; Gerson and Martin, 1969; Hamacher et al., 1986; Konishi and Reddoch, 1977; Nemoto et al., 1980). Owing to the strong interaction of the two n-systems as a result of the face-to-face arrangement, rapid electron

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

33

transfer is expected for all these derivatives. However, even for the radical anions, the intramolecular self-exchange is somewhat slowed down in solvents of low cation-solvating power. For defining the timescale of the actual electron-transfer processes, it is crucial to apply solvents of high solvating ability, since otherwise the exchange rate is tied in with the significantly lower rate of concomitant migration of the cation. Fixing the location of the counterion midway between two identical electrophores has been achieved in the radical anion of the dibenzo [ 18lcrown-6 derivative [46] (Mazur et al., 1980). When its radical anion exists as the ion pair with Na or K +,the intramolecular electron transfer becomes detectable on the esr timescale. The activation energy determined for the electron transfer (1.4 kcal mol-') clearly demonstrates that in this case a significant contribution to the activation barrier from ion pairing can be negated. In the radical anions of the norbornane-linked naphthalenes [ 371 mentioned earlier (Gerson et al., 1990) no counterion effects were detected for [37a], which has a small spatial separation, but the esr/ENDOR spectra of [37b] and [37c] -. indicate that the electron-spin transfer between the naphthalene moieties is determined by the rate of synchronous counterion migration (Gerson et al., 1990). For tight ion pairs the electron is localized, while for loose ion-pair conditions, e.g. by using solvents of high cation-solvating power, the transfer becomes fast on the hyperfine timescale (k > lo7 Hz). Esr spectroscopy has also been used to study pure solvent dynamics in electron self-exchange reactions (Grampp et al., 1990a; Grampp and Jaenicke, 1984a,b). When the systems are not linked by a spacer (i.e. TCNQ-'/TCNQ (TCNQ = tetracyanoquinodimethane), the homogeneous bimolecular rate constants khom are given by ( 1O), with kA the association constant and kET +

-'

the actual electron-transfer rate, comparable to our intramolecular exchange rate. The solvent dependence of the rate constant is explained in terms of Marcus theory by using an outer-sphere reorganization energy. This leads to an experimental reaction distance between the two molecules in the transition state for this adiabatic process (Grampp et al., 1990a). As an extension of the intermolecular self-exchange described above, the solvent-induced intramolecular electron exchange kinetics in radical anions of 1,3-dinitrobenzene [47] and benzene 1,3-dicarbaldehyde [48] have been studied by several authors (Freed and Fraenkel, 1964; Grampp et al., 1989, 1990b; Shohoji et al., 1987).The advantage of [47] and [48] is their structural simplicity and their high stability, which allows measurements even in protic

M.BAUMGARTEN ET AL

34

NC

ray

CN

0

solvents; in particular, the fixed distance and orientation of the redox centres open the possibility of a straightforward application of theoretical models. From alternate line-broadening effects in the esr spectra, the solventdependent intramolecular exchange rates can be derived, which in turn allows the calculation of values for the outer-sphere reorganization energy. If, on the other hand, the electron transfer in solution is determined by some rearrangement within the ion-pair structure, it is crucial to investigate the feasibility of electron transfer for an immobilized ion-pair structure in the solid state. For three substrates, the doubly layercd annulene [ 14a], the dianthrylethane [6a] and the anthracenophane [ 111, the radical anions were prepared under experimental conditions bringing about a spin-delocalized structure in solution. The solutions were then frozen into a glass by cooling, and the solid-state ENDOR spectra measured using the Davis pulse method (Davis, 1974; Grupp and Mehring, 1990; Rautter et al., 1992). It appears from the interpretation of the experimental spectra that, independently of the temperature in the glass, there is only a spin-localized structure. This occurs in the case of a doubly layered system [14a] and for the dianthryl [6a] (Fig. 9) (Rautter et al., 1992). What is even more surprising is that it also holds for the anthracenophane [ 111 with a short interplane distance of about 3 A. In solution the slow-exchange domain of the hopping process could never be achieved. Accordingly, if in a rigid system with strong n,n-overlap of the electrophoric group the position of the cation is fixed, the localization of the electron on one side of a bis-electrophoric molecule is enforced. Computer simulation of the experimental spectra in the frozen solution points to a spin-localized structure independent of the temperature. This interpretation is unambiguous because, in the solid state, the dipolar contributions of the hyperfine interaction A( dip) for all three principal

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

35

Fig. 9 Pulse ENDOR spectra of 1,3-dianthrylpropane[6b] at different temperatures.

directions must sum to zero, leaving the isotropic components a(iso), which are directly proportional to the spin densities at the carbon centres [see ( 1 la,b)]. Furthermore, by warming the samples until the solutions liquify A

= a(iso)

+ A(dip)

( 1l a )

(ca. 200 K), a drastic decrease in hyperfine splittings and thus spin densities

to one-half of the original values can be observed within the same experiment. For the examples of [11]-'/K+ and [14]-'/K+, it could also be proved that the transition between localized and delocalized states takes place at the phase transition between solid and liquid state (Rautter et al., 1992). In order to obtain kinetic parameters for the electron transfer of [ 1l ] - ' / K + , the dephasing time t, of the electron-spin echo near the phase-transition temperature T, was measured. These"experiments gave a correlation time t , of 100 ns for the electron transfer at T, = 170 K. From the assumption of an exponential decrease oft, in solution, a value of 100 ps was estimated for t , at room temperature (Rautter, 1989; Rautter et al., 1992).

36

M. BAUMGARTEN ET AL.

It has been shown so far that “internal” and “external” factors can be combined in the control of the electron-transfer rate. Although in most cases a simple theoretical treatment, e.g. by the Marcus approach, is prevented by the coincidence of these factors, it is clear that the observed features for the isoenergetic self-exchange differ by the electronic coupling and the free energy of activation. Then it is also difficult to separate the inner- and outer-sphere reorganization energies.

THE MODE OF LINKING

It will be shown that even more subtle structural changes can bring about drastic consequences in the electron-transfer kinetics. At first glance, the “dianthryl” compounds [61, [81 and [9] are closely related since they possess the same electrophoric subunit and the same ethanediyl bridge (Becker et al., 1991).Yet it appears that the different mode of linking profoundly affects the redox behaviour. The first evidence comes from cyclic voltammetric data. The question is whether the presence of the second redox group could facilitate the reduction oT the first one. This aspect could in principle be evaluated by comparison with the first redox potentials of suitably substituted monomeric analogues. More important in the present context is the interaction energy between the two electrophores, each carrying one extra electron. The experimental measure of this interaction is clearly the difference AE 1 , 2 between the first and second redox potentials (Bohnen et al., 1992). Considering first the anthracenophanes [ l l ] and [12], it is quite reasonable that the strong interaction of the n-systems as a result of the face-to-face arrangement should give rise to a significant potential difference upon dianion formation. More interestingly, a significant potential difference AE1,2can be detected for the di(9-anthry1)ethane [6] but not for the isomeric species [8] and [9] in which the bridge is attached to the centres C-1 and C-2 respectively. Even in the ortho-cyclophane [ 131 with two linkages there is no detectable interaction energy (Becker et al., 1991). It is thus possible to classify the closely related dianthryl systems according to the degree of interaction of the subunits; the question that immediately follows is whether such a clustering also develops from esr spectroscopic study of the corresponding radical anions. In the radical anions of the di(2-anthryl) system [9] and of the ortho-cyclophane [ 131 a spin-localized situation prevails independently of temperature and ion pairing. On the other hand, for the corresponding di(9-anthryl) species and even for the di( 1-anthryl) species a rapid hopping process can be detected if suitable ion pairing is established (Becker et al., 1991). Thus, according to both cyclic voltammetric and esr spectroscopic

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

37

evidence, the closely related dianthrylethane systems can be classified into three groups: (i) compounds [ 111 and [ 123 with strong interaction energy whose radical anions under all experimental conditions undergo rapid electron hopping in solution;

(ii) compounds [6a] and [8] whose radical anions show rapid or slow hopping, that is spin localization or effective delocalization, depending on the experimental conditions; and (iii) compounds [9a] and [ 131 with weak interaction, whose radical anions fail to exhibit electron transfer on the timescale of the experiment. When rationalizing the different behaviour of closely related biselectrophores, one must be aware that the intramolecular electron-transfer process requires some mixing of product and educt wave functions. The mode of linking seems to allow a fine tuning of the coupling integral J. The degree of interaction between the starting and the final situation will, of course, depend upon the spin density at the bridgehead position because these positions are closest in space. This may also be expressed by the distances between the average centre of spin density of the single electrophore. The lowest unoccupied molecular orbital of anthracene (see Fig. lo), i.e. the orbital that actually accepts the excess electron, has a high AO-coefficient at C-9 and a relatively small AO-coefficient at C-2 (Heilbronner and Bock, 1978). What is found experimentally is that rapid hopping is detectable in all those cases in which the linking is at C-9 or at C-1, i.e. at centres with a large or moderate AO-coefficient. This may be expressed in terms of through-space and through-bond mechanisms, which are both favoured in the case above, while a through-bond mechanism becomes much less probable as soon as the spin density at the bridgehead position is small as for [9] and [13]. Even though [9] is much more flexible than [13], which has two bridges, the situation encountered is comparable. It seems that a second bridge at a position of low spin density does not improve the through-bond exchange and thereby diminish the activation barrier. Thus it appears that

Fig. 10 Atomic orbital coefficients for the LUMO of anthracene.

38

M. BAUMGARTEN ET AL.

even with the same electrophore and with the same linking group very subtle structural changes, i.e. the position of attachment of the spacer to the electrophore, can influence the degree of interaction and the rate of electron hopping. Biaryl compounds, although formally possessing direct n-conjugation between the aryl moieties, can give rise to electronically independent subunits as outlined in Section 3. From the viewpoint of redox chemistry, there are two basic criteria by which to classify this situation: for "decoupled" electrophores (i) injection of one electron will give rise to spin localization and (ii) injection of a second electron may produce a triplet dianion. Clearly such a situation does not prevail in biphenyl and its analogue the 1,l'-binaphthyl derivative [ 161. In the radical anion of the latter, one detects a set of coupling constants [ a H= 0.39 mT ( n = 2), 0.225 mT ( n = 2), etc.], which indicates a simultaneous distribution of the spin density over both entities. From the asymmetry of spin density at positions 4 and 5, a twisting angle of 50" has been determined (Koch, 1991; Baumgarten et al., 1992b). Another extreme is 9,9'-bianthryl [ 181; as a result of the near orthogonality of the anthryl rings, the dianion exists as a triplet, and, in agreement with this finding, the esr spectrum of the biradical dianion shows a large zero-field splitting D = 15.5 mT and a half-field signal for the Ams = 2 transition (Baumgarten et al., 1992a). When considering the prevailing non-bonded interaction in the binaphthyl [ 161 and the corresponding biperylenyl [ 171,

MA h

T=240K

Fig. 11 Temperature-dependent esr spectra of the monoanion of biperylene [ 171 in MTHF.

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

39

a similar inter-ring torsional angle would be expected. It is characteristic, however, that the esr spectrum of [ 171 reveals also the presence of localized forms for strong counterion conditions in solution, while lowering the temperature to the frozen state (ca. 160K) yields nearly exclusively the localized structure (Fig. 11). If the inter-ring torsional angle is comparable in both cases, the differences in the exchange rates may be traced back to the A 0 coefficients and spin densities at the bridgehead position ( p = 0.2 vs. 0.13) and the large differences in the centre-to-centre distance (0.49 vs. 0.87 nm).

6

Conclusion and outlook

The rate of degenerate intramolecular electron-transfer processes in biselectrophoric redox systems, and the observed spin-density distribution over one or two units depend upon the overall reorganization energy and thus upon

(i) the nature of the subunit; (ii) the spacer between the subunits; (iii) the ion pairing; and (iv) the way in which the subunits are linked. In essence, these empirical findings allow control of the rate of electrontransfer processes by creating the appropriate structural conditions. It is, of course, straightforward to extend such a correlation of structure and electron-transfer kinetics to higher homologues. The esr spectra of the radical anion obtained for the triply layered annulene system [lS] points to a situation with one electron being effectively delocalized over three layers (Alexander et al., 1989). The same holds true for the higher homologues. A closely related finding can be made for the anthrylene structure [lo], i.e. a redox system with three anthracene electrophores. It is obvious from the esr data for the radical anion that, depending on the ion pairing, the unpaired electron resides on the inner anthracene unit or is effectively delocalized over the whole chain (Fiedler et al., 1986). Accordingly, when one proceeds to the related oligomers and polymers, by a proper choice of the subunit, the bridging group and the ion pairing, one can control the energy profiles of intramolecular charge-transport processes in an analogous fashion. Systems such as [6], [lo], [I81 or [21] have also been the subject of photophysical studies (Rettig, 1988; Mataga et al., 1989; Yao et al., 1989)

40

M. BAUMGARTEN € T A L .

in which energy transfer between the separate anthracene chromophores is considered as a function of the molecular geometry. Although a direct comparison is not straightforward, for example, in view of the role of the ion pairing, it is a tempting approach to compare the processes of energy and charge transport (Heine et al., 1990). Another aspect is that intramolecular electron transfer is a fundamental process in intramolecular electronics in which a single molecule is used to process electrical signals (Joachim et al., 1990; Joachim and Launay, 1990). In this context, the electron transfer through a spacer group has been related to its “conductance” if placed between two conducting wires. Thus, accepting the definition of conduction of a single molecule or spacer group, the control of an electron transfer between redox centres closely corresponds to the control of the current passing through an electrical circuit. Upon electrochemical oxidation at a platinum electrode, many polycyclic compounds such as naphthalene, fluoranthene and pyrene form deeply coloured, crystalline radical cation salts (Maresch et al., 1989; Enkelmaan et al., 1985; Endres et al., 1985; Krohnke et al., 1980). According to crystal structure analysis, these salts adopt a stack-type structure with a face-to-face arrangement of the n-layers and an alternating array of neutral and monocationic units. The complexes have a high electrical conductivity along the stacking axis, and one way of explaining this conductivity is based on the assumption of a charge-hopping process between the n-layers. Such a charge-transport mechanism is thus closely related to the electron transfer in multi-layered systems such as [lo], [l5] and [21].

References Alexander, J., Ehrenfreund, M., Fiedler, J., Huber, W., Rader, H.-J. and Miillen, K. (1989). Angew. Chem. 101, 1530; Int. Ed. Engl. 28, 1531 Anet, A. L., Bourn, A. J. R. and Lin, Y. S. (1964). J . Am. Chem. SOC. 86, 3576 Anton, U., Bohnen, A., Koch, K.-H., Naarmann, H., Rader, H.-J. and Miillen, K. (1992). Adu. Mat. 4,91-93 Atkins, P. W. ( 1984).Molecular Quantum Mechanics. Oxford University Press, Oxford Auchter-Krummel, P. and Miillen, K. (1991). Angew. Chem. 103, 996; Znt. Ed. Engl. 30, 1003 Baughman, R. H., Bredas, J. L., Chance, R. R., Elsenbaumer, R. L. and Shacklette, L. W. (1982). Chem. Rev. 82, 209 Baumgarten, M., Miiller, U., Bohnen, A. and Miillen, K. (1992a). Angew. Chem. 104, 482-485 Baumgarten, M., Koch, K.-H. and Miillen, K. (1992b). Accepted for publication Becker, B., Bohnen, A., Ehrenfreund, M., Wohlfarth, W., Sakata, Y., Huber, W. and Miillen, K. (1991). J. Am. Chem. SOC. 113, 1121 Bender, D. and Miillen, K. (1988). Chem. Ber. 121, 1187

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

41

Bender, D., Herbst, H., Schade, P. and Miillen, K. (1988). Chem. Ber. 121, 1177 Bender, D., Przybylski, M. and Miillen, K. (1989). Makromol. Chem. 190, 2071 Bitthin, R., Ely, G., Wolffler, F., Munstedt, H., Naarmann, H. and Naegele, D. (1987). Makromol. Chem. Macromol. Symp., 8, 51 Bock, H., Ruppert, K. and FenskelD: (1989). Angew. Chem. 101, 1717; Int. Ed. Engl. 28. 1685 Bohnen, A., Koch, K.-H., Liittke, W. and Miillen, K. (1990). Angew. Chem. 102, 548; Int. Ed. Engl. 29, 525 Bohnen, A., Heitz, W., Miillen, K., Rader, H.-J. and Schenk, R. (1991). Makromol. Chem. 192, 1679 Bohnen, A., Rader, H.-J. and Miillen, K. (1992) Synth. Met., 47, 37-63. Cannon, R. D. ( 1980). Electron-Transfer Reactions. Butterworth, London Closs, G. L. and Miller, J. R. (1988). Science 240, 440 Closs, G. L., Calcaterra, L. T., Green, N. J., Penfield, K. W. and Miller, J. R. (1986). J . Phys. Chem. 90, 3673 Closs, G. L., Johnson, M. D., Miller, J. R. and Piotrowiak, P. (1989). J . Am. Chem. SOC. 111, 3751 Dale, J. (1974). Tetrahedron 30,1683 Davies, E. R. (1974). Phys. Lett. 47A, 1 Eliasson, B. and Staley, S. W. (1985). Prepr. Am. Chem. SOC.Diu. Pet. Chem. 30, 620; Chem. Abstr. 104, 50494a Eliasson, B., Edlund, U. and Miillen, K. (1986). J . Chem. SOC.Perkin Trans. 11, 937 Eliasson, B., Johnels, D., Sethson, I., Edlund, U. and Miillen, K. (1990). J . Chem. SOC. Perkin Trans. 11, 897 Endres, H., Keller, H. J., Miiller, B. and Schweitzer, D. (1985).Acta Cryst. C41,607 Enkelmann, V., Gockelmann, K., Wieners, G. and Monkenbusch, M. (1985). Mol. Cryst. Liq. Cryst. 120, 195 Eyring, H. (1935a). J . Chem. Phys. 3, 107 Eyring, H. (1935b). Chem. Rev. 17, 65 Fahnenstich, U., Koch, K.-H. and Miillen, K. (1989). Makromol. Chem., Rapid Commun. 10, 563 Fiedler, J., Huber, W. and Miillen, K. (1986). Angew. Chem. 98, 444; Int. Ed. Engl. 25,443 Flanagan, J. B., Margel, S., Bard, A. J. and Anson, F. C. ( 1978). J . Phys. Chem. 82,1101 Fox, M. A. and Chanon, M. (1988). Photoinduced Electron-Transfer. Elsevier, Amsterdam Freed, J. H. and Fraenkel, G. K. (1964). J . Phys. Chem. 41, 699 Fry, A. J., Powers, T. A., Miillen, K. and Irmen, W. ( 1985). Tetrahedron Lett. 26,4879 Fukuzumi, S., Wong, C. L. and Kochi, J. K. (1980). J . Am. Chem. SOC. 102,4790 Gerson, F. ( 1967). HochauJlosende ESR-Spektroskopie. Verlag-Chemie, Weinheim; ( 1970). High Resolution ESR Spectroscopy. Verlag-Chemie, Weinheim Gerson, F. and Martin, W. A. Jr (1969). J . Am. Chem. SOC. 91, 1883 Gerson, F., Gleiter, R., Moshuk, G. and Dreiding, A. S. (1972). J . Am. Chem. SOC. 94, 2919 Gerson, F., Kowert, B. and Peake, P. M. (1974). J . Am. Chem. SOC. 96, 118 Gerson, F., Martin, W. B. Jr and Wydler, C. (1976). J . Am. Chem. SOC. 98, 1318 Gerson, F., Huber, W., Martin, W. B. Jr, Caluwe, P., Pepper, T. and Szwarc, M. (1984). Helu. Chim. Acta 67, 416 Gerson, F., Wellauer, T., Oliver, A. M. and Paddon-Row, M. N. (1990). Helu. Chim. Acta. 73, 1586

42

M. BAUMGARTEN ET AL.

Grampp, G. and Jaenicke, W. (1984a). Ber. Bunsenges. Phys. Chem. 88, 325 Grampp, G. and Jaenicke, W. (1984b). Ber. Bunsenges. Phys. Chem. 88, 335 Grampp, G., Sohoji, M. C. B. L. and Herold, B. J. (1989). Ber. Bunsenges. Phys. Chem. 93, 580 Grampp, G., Harrer, W. and Hetz, G. (1990a).Ber. Bunsenges. Phys. Chem. 94,1343 Grampp, G., Sohoji, M. C. B. L., Herold, B. J. and Steenken, S. (1990b).Ber. Bunsenges. Phys. Chem. 94, 1507 Gray, H. B. and Malmstrom, B. G. (1989). Biochemistry 28, 7449 Grupp, A. and Mehring, M. (1990). In Modern Pulsed and Continuous Wave Electron Spin Resonance, p. 195. Wiley, New York Gust, D., Moore, T. A,, Moore, A. L., Barrett, D. and Harding, L. 0. (1986). J . Am. Chem. SOC.108, 8028 Gust, D., Moore, T. A., Moore, A. L., Barrett, D., Harding, L. O., Makings, L. R., Liddell, P. A., De Schryver, F. C., Van der Auweraer, M., Bensasson, R. V. and Rougee, M. (1988). J . Am. Chem. SOC.110, 321 Gutowsky, H. S. and Holm, C. H. (1956). J . Chem. Phys. 25, 1228 Hamacher, V., Plato, M. and Mobius, K. (1986). Chem. Phys. Lett. 125, 69 Hammerich, 0. and Saveant, J.-H. (1979). J . Chem. SOC. Chem. Commun., 938 Harriman, J. E. and Maki, A. W. (1963). J . Chem. Phys. 39, 778 Hay, P. F., Thibeault, J. C. and Hoffmann, R. (1975). J . Am. Chem. SOC.97, 4884 Heilbronner, E. and Bock, H. (1978). Das HMO Model1 und seine Andwendungen. Verlag-Chemie, Weinheim/New York Heine, B., Sigmund, E., Maier, S., Port, H., Wolf, H. C., Effenberger, F. and Schlosser, H. (1990). J . Mol. Electron. 6, 51 Heinze, J. (1984). Angew. Chem. 96, 823; Int. Ed. Eng. 23, 831 Heinze, J. (1990). Top. Curr. Chem. 152, 1 Heinze, J. and Meerholz, K. (1990). Angew. Chem. 102, 695; Int. Ed. Engl. 29, 692 Heinze, J., Serafimov, 0.and Zimmermann, H. E. (1974). Ber. Bunsenges. Phys. Chem. 78, 652 Hoffmann, R. (1971). Acc. Chem. Res. 4, 1 Hogen-Esch, T. E. (1977). Adu. Phys. Prog. Chem. 15, 153 Huber, M., Kurreck, H., Maltzan, B. von, Plato, M. and Mobius, K. (1990). J . Chem. SOC.Faraday Trans. 86, 1087 Huber, W. and Miillen, K. (1980). J . Chem. SOC. Chem. Commun., 698 Huber, W. and Miillen, K. (1986). Acc. Chem. Res. 19, 300 Huber, W., Unterberg, H. and Miillen, K. (1983). Angew. Chem. 95, 239; Int. Ed. Engl. 22, 242 Hush, N. S. (1961). Trans. Faraday SOC.,57, 557 Hush, N. S. (1975). Chem. Phys. 10, 361 Irmen, W., Huber, W., Lex, J. and Miillen, K. (1984). Angew. Chem. 96, 800; Int. Ed. Engl. 23, 818 Isied, S. S., Vassilian, A., Wishart, J. F., Creutz, C., Schwarz, H. A. and Sutin, N. (1988). J . Am. Chem. SOC. 110, 635 Joachim, C. and Launay, J. P. (1990). J . Mol. Electron. 6, 37 Joachim, C., Launay, J. P. and Woitellier, S. (1990). Chem. Phys. 147, 131 Katz, T. J. (1960). J . Am. Chem. SOC.82, 3784 Klabunde, K.-U., Huber, W., Lex, J. and Miillen, K. (1987). Chem. Ber. 120, 1491 Koch, K.-H. (1991). Dissertation, Mainz Koch, K.-H. and Miillen, K. (1991). Chem. Ber. 124, 2091. Kochi, J. K. (1988). Angew. Chem. 100, 1331; Int. Ed. Engl. 27, 1227

ORGANIC REDOX SYSTEMS WITH MULTIPLE ELECTROPHORES

43

Konishi, S. and Reddoch, A. H. (1977). J . Magn. Res. 28, 41 Krohnke, C., Enkelmann, V. and Wegner, G. (1980). Angew. Chem. 92,941; Int. Ed. Engl. 19, 912 Krummel, G., Huber, W. and Miillen, K. (1987). Angew. Chem. 99, 1305; Int. Ed. Engl. 26, 1290 Krummel, G. (1988). Ph.D. dissertation, Mainz Krummel, G. and Miillen, K. (1988). Tetrahedron Lett. 29, 2809 Kurreck, H., Kirste, B. and Lubitz, W. (1988). Electron nuclear double resonance spectroscopy of radicals in solution. In Methods in Stereochemical Analysis (ed. A. P. Marchand). VCH Publishers, Weinheim/New York Kuznetsov, A. M. and Ulstrup, J. (1981). J . Chem. Phys. 75, 2047 Lewis, I. C. and Singer, L. S. (1965). J . Chem. Phys. 43, 2712 Liang, N., Miller, J. R. and Closs, G. L. (1990). J . Am. Chem. SOC. 112, 5353 Marcus, R. A. (1956). J . Chem. Phys. 24, 979 Marcus, R. A. (1963). J . Phys. Chem. 67, 853, 2889 Marcus, R. A. (1965). J . Chem. Phys. 43, 679 Marcus, R. A. and Sutin, N. (1985). Biochim. Biophys. Acta 811, 265 Maresch, G. G., Mehring, M., Schiitz, J. U. von, Werner, H.-P., Gokelmann, K., Enkelmann, V., Miillen, K. and Klabunde, K.-U. ( 1989).J . Chem. Phys. 91,4543 Mataga, N., Yao, H., Okada, T. and Rettig, W. (1989). J . Phys. Chem. 93, 3383 Mazur, S., Dixit, V. M. and Gerson, F. (1980). J . Am. Chem. SOC. 102, 5343 McConnell, H. M. (1961). J . Chem. Phys. 35, 508 McDiarmid, A. D. (1979). J . Chem. SOC.Chem. Commun., 594 Meerholz, K. and Heinze, J. (1989). J . Am. Chem. SOC. 111, 2325 Meerholz, K. and Heinze, J. (1990). J . Am. Chem. SOC. 112, 5142 Meyer, T. J. (1978). Acc. Chem. Res. 11, 94 Miller, J. R., Calcaterra, L. T. and Closs, G. L. (1984). J . Am. Chem. SOC. 106, 3047 Mortensen, J., Heinze, J., Herbst, H. and Miillen, K. (1992). J . Electround. Chem., accepted for publication Miillen, K. (1978). Helu. Chim. Acta 61, 2307 Miillen, K. (1984). Chem. Rev. 84, 603 Miillen, K. (1986). Pure Appl. Chem. 58, 177 Miillen, K. (1987). Angew. Chem. 99, 192; Int. Ed. Engl. 26, 204 Miillen, K., Heinz, W., Klarner, F.-G., Roth, W. R., Kindermann, I., Adamczak, O., Wette, M. and Lex, J. (1990). Chem. Ber. 123, 2349 Murray, R. W. (1984).In Electroanalytical Chemistry, Vol. 13 (ed. A. J. Bard), p. 191. Dekker, New York Nemoto, F., Ishizu, K., Toyoda, T., Sakata, Y. and Misumi, S. (1980).J . Am. Chem. SOC. 102, 654 Nowak, R. J., Schultz, F. A,, Umana, M., Lam, R. and Murray, R. W. (1980). Anal. Chem. 52, 315 Onuchic, J. N., Beratan, D. A. and Hopfield, J. J. (1986).J . Phys. Chem. 90, 3707 Paddon Row, M. N. and Jordan, K. D. (1988). In Modern Models of Bonding and Delocalization (ed. J. F. Liebman and A. Greenberg), Chap. 3. VCH Publishers, New York Pollmann, M., Wohlfarth, W., Lex, J. and Miillen, K. ( 1990).Tetrahedron Lett. 31,2701 Rabinovitz, M. (1988). Top. Curr. Chem. 146, 99 Rautter, J. (1989). Diploma thesis, Stuttgart Rautter, J., Grupp, A., Mehring, M., Alexander, J., Miillen, K. and Huber, W. (1992). Mol. Phys. accepted for publication

44

M. BAUMGARTEN ET AL.

Reimers, J. R. and Hush, N. S. (1990). Inorg. Chem. 29, 3886 Rettig, W. (1988). Mol. Struct. Energetics 6, 229 Salem, L. (1966). The Molecular Orbital Theory of Conjugated Systems. Benjamin, New York Schenk, R. (1989). Ph.D. thesis, University of Mainz Shacklette, L. W., Maxfield, M., Gould, S., Wolf, J. F., Jow, T. R. and Baughman, R. H. (1987). Synth. Met. 18, 611 Shimada, K., Moshuk, G., Connor, H. D., Caluwe, P. and Szwarc, M. (1972). Chem. Phys. Lett. 14, 396 Shimada, K. and Szwarc, M. (1974). Chem. Phys. Lett. 28, 540 Shohoji, M. C. B. L., Herold, B. J. and Novais, H. M. (1987). J . Chem. SOC. Perkin Trans. 11, 1465 Siders, P. and Marcus, R. A. (1981). J . Am. Chem. SOC.103, 748 Smith, T. W., Kuder, J. E. and Wychick, D. (1976). J. Polym. Sci. 14, 243 Staley, S. W., Dustman, C. K., Facchine, K. L. and Linkowsky, G. E. (1985). J . Am. Chem. SOC. 107,4003 Sullivan, P. D. and Bolton, J. R. (1970). Adu. Magn. Reson. 4, 39 Szwarc, M. ( 1968). Carbanions, Living Polymers and Electron-Transfer Processes. Wiley, New York Szwarc, M. ( 1974) Ions and Ion-Pairs in Organic Reactions. Wiley-Interscience, New York Ulstrup, J. ( 1979). Charge- Transfer Processes in Condensed Media. Springer-Verlag, Berlin Wagner, M., Wohlfarth, W. and Miillen, K. (1988). Chimia 42, 377 Ward, R. L. and Weissman, S. I. (1957). J . Am. Chem. SOC.79, 2086 Wasielewski, M. R., Niemczyk, M. P., Svec, W. A. and Pewitt, E. B. (1985). J. Am. Chem. SOC. 107, 5562 Wasielewski, M. R. (1989). In Photoinduced Electron-Transfer (ed. M. A. Fox and M. Chanon), Chap. 1.4. Elsevier, Amsterdam Wegner, G. (1981). Angew. Chem. 93, 352; Int. Ed. Engl. 20, 361 Wegener, S. and Mullen, K. (1991). Chem. Ber. 124, 2101 Weitzel, H.-P. and Miillen, K. (1990). Makromol. Chem. 191, 2837 Weitzel, H.-P., Bohnen, A. and Miillen, K. (1990). Makromol. Chem. 191, 2837 Wertz, J. E. and Bolton, J. R. (1972) Electron Spin Resonance. McGraw-Hill, New York; reprinted by Chapman and Hall, London, 1986 Yao, H., Okada, T. and Mataga, N. (1989). J. Phys. Chem. 93, 7377 Ziilicke, L. (1985). Quantenchemie, Bd.2 p 49ff. VEB Deutscher Verlag der Wissenschaften, Berlin

Chirality and Molecular Recognition in Monolayers a t the Air-Water Interface PHILIP L. ROSE; NOELG. H A R V Eand Y ~ EDWARD M. ARNETT' Exploratory Chemicals Research, Rohm and Haas Company, 727 Norristown Road, Spring House, Pennsylvania 19477, U S A bExploratory Plastics Research, Rohm and Haas Company, PO Box 219, Bristol, Pennsylvania 19007, U S A Department of Chemistry, Duke University, Durham, North Carolina 27706, U S A a

1 Introduction 45 2 Monolayer methods 49 Film balance techniques 49 Equilibrium thermodynamic properties 51 n / A curves and phase transitions 54 Dynamic methods 57 Mixing criteria as applied to the elucidation of intermolecular interactions at surfaces 63 Techniques for visualizing films and aggregates 68 3 Chiral monolayers 71 Enantiomeric systems 71 Diastereomeric systems 102 4 Conclusions 133 Symbols 134 References 135

1

Introduction

The purpose of this chapter is to provide a review of research conducted in the senior author's laboratory over the past 15 years on the stereochemistry of intra- and inter-molecular interactions in a special kind of highly oriented system-monolayers. On two previous occasions we have reviewed the behavior of chiral surfactants at the air-water interface; first as an introduction to the 45 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 28 ISBN 0-12-033528-X

Copyright 0 1993 Academlc Press Limited All rights of reproduction in any form reserved

46

P. L. ROSE ET AL.

opportunities for research in the field (Stewart and Arnett, 1982) and, more recently, as a brief outline, by the present authors, of work accomplished during the first decade (Arnett et aZ., 1989). In contrast, the present report presents an extensive summary of what has been done, and what was learned, from the perspective of a decade and a half. Since the majority of the research on chiral monolayers has been done in our laboratories, this chapter can be regarded as a nearly complete review of research accomplished in the field. From the perspective of physical organic chemistry, the study of chiral monolayers is a classic example of how the purposeful interaction of two fields that were separated by their traditional identifications with organic and physical chemistry can help to illuminate both areas. Stereochemistry is one of the most powerful tools of organic chemistry for demonstrating, and exploiting, the influences of molecular shape and symmetry on reactivity. The extensive role of chirality for providing selectivity in biochemical processes has stimulated enormous activity by organic chemists towards achieving asymmetric syntheses of ever greater stereochemical control and breadth of scope. During the 15 years of this project, organic chemists have also become increasingly aware of the importance of intermolecular attractive forces for the development of the detailed molecular architecture of complexes, intermediates and transition states. This awareness has been manifested by the introduction of such terms as molecular recognition, enzyme mimicry, and host-guest chemistry. Although organic chemists have always been sensitive to the specific effects of repulsive interactions through the popular term “steric hindrance”, the skillful manipulation of such attractive forces as London-van der Waals, dipole-dipole, n-n stacking, and especially hydrogen bonding, for the generation of multimolecular aggregated systems of closely specified geometry has now engaged some of the best minds in organic chemistry. Parallel developments in the physical chemistry of surfaces have also proceeded rapidly during the same period. An extensive battery of new spectroscopic and microscopic techniques have brought analysis and even observation down to the molecular and atomic ideal of “seeing” and manipulating these ultimate units of chemistry. Much of the driving force for these advances has come from the microelectronics industry, where the ability towards mass production of microstructures approaching nanometer dimensions is proceeding with remarkable speed and success. Again, we are reminded that Nature provides the ultimate model for emulation in the use of cooperative interactions of an enormous number of small structural components through many weak, reversible attractions and repulsions to produce such complex microstructures as proteins, enzymes, viruses, and cells with virtually perfect fidelity (Whitesides, 1991). One important strategy for producing ultra-thin films of promise for microelectronics

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

47

employs the well-established Langmuir-Blodgett technology for handling monolayers of surfactants at the air-water surface and is related directly to the subject of this review. Although the notion of monomolecular surface layers is of fundamental importance to all phases of surface science, surfactant monolayers at the aqueous surface are so unique as virtually to constitute a special state of matter. For the many types of amphipathic molecules that meet the simple requirements for monolayer formation it is possible, using quite simple but elegant techniques over a century old, to obtain quantitative information on intermolecular forces and, furthermore, to manipulate them at will. The special driving force for self-assembly of surfactant molecules as monolayers, micelles, vesicles, or cell membranes (Fendler, 1982) when brought into contact with water is the hydrophobic eflect. The immiscibility of oil and water has been noted since antiquity and been put to use for calming troubled seas. Benjamin Franklin, benefitting from the rapidly developing atomic-molecular theory of discontinuous matter, noted that a teaspoon of oil spread over the half acre of Clapham pond not only stilled all surface agitation but produced a film so thin as to give “prismatic colors”.’ Self-organization at interfaces with water can be achieved by compounds whose molecules combine an appropriate polar functionality (e.g., carboxylate, hydroxyl, or alkylammonium) that is strongly attracted to water, and an extensive hydrocarbon framework or chain that is rejected by it. If molecules of such amphipathic compounds (i.e., soaps, surfactants, and surface-active agents) are introduced to a water surface, the polar headgroup is attracted to the subphase and the fatty tail “floats” on the surface. The actual orientation of the surfactant molecules is determined by their surface concentration in terms of area per molecule and the attractive-repulsive interactions between nonpolar portions of adjacent molecules. Thus, the molecules are preorganized at the surface through the balance of hydrophilic attraction to the headgroup and hydrophobic repulsion of the fatty portion. However, most importantly, the rest of the organization process can be varied at will by changing the surface area and measuring the balance of repulsive and attractive forces that resist compression of the monolayer in a direction parallel to the surface. This point will be re-emphasized below. The beautifully simple and elegant tool for studying the response of monolayers on a water surface to compression was invented in 1882 by Fraulein Agnes Pockels when she was 20 years old from observations made at her kitchen sink. As physical chemistry developed, an appreciation for the

’

For extensive and authoritative monographs on the nature of the hydrophobic effect and the history of its use, see Tanford, 1973, 1989.

48

P. L. ROSE ET AL.

significance of the monolayer state as a means for studying intermolecular forces grew with it. In the hands of Harkins at the University of Chicago, McBain, N. K. Adam, and, most notably, Irving Langmuir and his colleague Kathleen Blodgett at General Electric Co., the basic techniques for handling monolayer films and the correct physico-chemical interpretations of their behavior were developed. Langmuir’s development of Fraulein Pockel’s trough for handling monolayers bears his name and is described in detail below. In 1932 Langmuir received the Nobel Prize for this research. Since its first important flowering in the hands of purely physical chemists, interest in monolayers at the air-water interface has waxed and waned with a frequency of roughly 25 years. The first resurrection of interest came from biochemistry, primarily during the 1955-65 decade as phospholipid monolayers were studied as models for the cell membrane (see, for example, Chapman, 1968).This is still a very productive field of biophysical research. The second recrudescence has been driven by thin film physics and technology as noted above. During their work at General Electric, Langmuir and Blodgett learned how to transfer monolayers from the surface of water to various solid surfaces such as glass. A variety of clever routines were developed for building up multilayers containing up to 200-300 layers, most commonly of alkaline earth stearates (Gaines, 1966). A commercial step-gauge of exact thickness based on the known thickness of stearate monolayers was actually marketed by General Electric. Polymerizable functionalities, such as acetylene linkages, may be used to bind the chains of surfactant molecules, properly oriented at the surface, to each other and thus manufacture polymeric microfilms of known thickness as an ultimate step towards etch-resistant films. An important recent development is the self-organization of long chain thiols on gold surfaces (for a brief review, see Bain and Whitesides, 1988). Against this background, the opportunity of using the techniques of monolayer chemistry to manipulate the approach of preoriented chiral molecules to each other at an aqueous surface in attractive. To the organic chemist, it provides a means for determining how the interplay of molecular structure, conformation and proximity of packing can affect the way chiral molecules “read” each other’s shapes and finally pack in the lowest energy state of surface aggregation. To the physical chemist, the introduction of stereochemistry provides greater subtlety in considering intermolecular interactions in condensed matter. For our part, the study of chiral molecules has provided an enjoyable and often surprising excursion into truly interdisciplinary physical organic chemistry.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

2

49

Monolayer methods

FILM BALANCE TECHNIQUES

The elegant tool by which monolayers may be studied and also manipulated at the air-water interface was developed by Langmuir and Blodgett during the 1920s and 1930s. The film balance described here was developed from that designed by Langmuir (Langmuir and Schaefer, 1937);a picture is presented in Fig. 1. For the purposes of detecting stereochemical effects, a much more sensitive trough is necessary; a description of the apparatus used in the authors' laboratory follows. The trough was milled from solid Teflon (80 cm x 14 cm x 0.3 cm) and mounted on a magnesium base, which rests on four leveling screw legs. A serpentine glass coil connected to a constant-temperature circulating bath may be placed in the trough base in order to hold the subphase temperature in the desired range. In addition to the water circulation coil, the surface temperature is controlled further by placing the entire apparatus in a Puffer-Hubbard Unitherm temperature control cabinet. In the 15-40°C temperature range, these controls are capable of regulating surface temperatures to k0.5"C. After being filled with subphase water, a floating hydrophobic Teflon barrier (12cm x 1 cm x 1 mm) is connected to both sides of the trough by means of two flexible Teflon Mylar ribbons so as to divide the surface of the trough into two sections. The floating barrier is arranged so that it floats freely but remains in contact with the torsion arm lever of the torsion strip spring system. The torsion strip is connected to a transducer coil, which in turn is connected to a strip-chart recorder and an torsion strip + & , hanging weight + holder

ever arm transducer

barrier

Fig. 1 Schematic representation of the Langmuir film balance used for the measurement of pressure-area monolayer film properties. Reprinted with permission from Arnett et al., 1989. Copyright 1989 American Chemical Society.

50

P. L. ROSE €T AL.

analog-to-digital converter system (Johnson, 1985). The film balance is therefore divided into two compartments, with the monolayer film being spread on the side with the greater surface. The area of the compartment containing the monolayer may be varied by means of a motor-driven Teflon barrier. As the area per molecule of the film is reduced, the surface tension of the film-covered side of the trough is lowered with respect to the other side, and a surface pressure is registered against the floating barrier. This surface pressure is related to surface tension at any given film area by expression ( l ) , where ll is the surface pressure, y o is the surface tension of

the film-free surface, and y is the surface tension of the film-covered surface. The result is a continuous surface pressure versus area ( I I I A ) isotherm. Because of the extreme sensitivity of monolayers to contamination and the resulting likelihood of erroneous results, every precaution must be taken to ensure that all materials and instruments are clean. A great bonus for the stereochemical studies reported here is that comparison of the properties of enantiomers provides a rigorous check for internal consistency and purity that is not available to studies of achiral systems. Details have been given elsewhere and will not be included here (Thompson, 1981). Appropriate molecules for study as monolayers are natural or synthetic surfactants. The surfactants have divergent functional groups that give rise to their amphiphilic nature: a highly polar functionality (carboxylate, amino, or hydroxyl) as a head group and a fatty chain of 10 carbons or more as a tail group. If a very dilute solution of surfactant molecules, in a suitable spreading solvent, is deposited on the surface of pure water, and the solvent evaporates or dissolves, the surface-active molecules are left at the interface with their polar head groups bound to the aqueous subphase and their fatty tails assuming various orientations relative to the surface plane. The orientation of the hydrocarbon tail will be dependent upon the available surface area per molecule. Because surfactant molecules have hydrocarbon chains that are insoluble in the aqueous subphase, they are constrained to stay at the interface, or aggregate to form microcrystals. If the area per molecule is large (over 1000 A2/molecule), a “gaseous” monolayer is usually formed in which the molecules float freely and independently on the surface. The Langmuir film balance allows for the manipulation of surfactant molecules from molecules floating freely with little or no interaction to highly condensed phases where the hydrocarbon chains are interacting significantly. The manipulation of the available area per molecule is all that is needed to effect such a phase transition.

-

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

51

The significance of the above discussion is to point out that the orientation of the hydrocarbon chains with respect to the surface and to each other can be monitored, and controlled, as a function of the surface pressure (which in turn is directly proportional to the surface tension if the monolayer is insoluble in the subphase). Another means of measuring the properties of insoluble films at the air-water interface is through the use of surface potentials. Surface potential (A V) measures the charge separation created by the vector component of the surfactant’s molecular dipole that is perpendicular to the air-water interface. Thus, the surface potential yields information about the orientation of the surfactant molecules. Surface potential values are often expressed alternatively as surface dipole moments p L I according to (2), where n is the 1 AV pcLI= --

4x n number of molecules per cm2 in the monolayer. If the area A is expressed in A2/molecule and A V in millivolts then (2) reduces to ( 3 ) where p l is A AV PLI =-

12x

(3)

obtained in millidebye units (mD). A common method for measuring the surface potential of a monolayer film uses an ionizing electrode. The ionizing electrode utilizes a radioactive source (e.g. lOmCi, 224Cm)to ionize the air gap above the film so that it becomes conducting. The potential difference between the electrode above the film and the one in the aqueous subphase can then be measured directly. A typical experimental set-up is shown in Fig. 2. Surface potentials are measured either with the monolayer film at a constant area per molecule or while the film is undergoing compression. Static measurements are performed by delivering a known amount of surfactant in a suitable spreading solvent to the subphase surface in a container of known dimensions. Dynamic surface potential measurements are made concurrently with compression and expansion cycles on the Langmuir trough.

EQUILIBRIUM THERMODYNAMIC PROPERTIES

The Langmuir film balance experiment as described above is generally considered to be thermodynamically determined by the relation between the independent variables temperature, surface pressure, and molecular area.

P. L. ROSE E T A L .

52

glass electrode reference electrode

ionizing reference electrode

glass electrode

subphase

I

Fig. 2 Monolayer surface potential measurement apparatus.

However, the spreading of a surfactant monolayer from a volatile solvent leaves behind a film that may not be in thermodynamic equilibrium with its bulk crystalline form or the aqueous subphase. It has been proposed that this is a result of the relatively high energy barriers to film collapse or dissolution into the subphase as compared with lowered interfacial free energy when a stable, insoluble surfactant monolayer is formed (Gershfeld, 1976). The rate at which a whole system approaches true equilibrium in such a system is often so slow that the monolayer film can be treated for most purposes as though it were at equilibrium with the subphase. The question may then be raised as to whether insoluble monolayers may really be treated in terms of equilibrium thermodynamics. In general, this problem has been approached by considering: (i) the equilibrium spreading pressure of the monolayer in the presence of the bulk crystalline surfactant, and (ii) the stability of the monolayer film as spread from solution. These quantities are obtained experimentally and are necessary in any consideration of film thermodynamic properties. In both cases, time is clearly a practical variable.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

53

7he equilibrium spreading pressure ( E S P ) The equilibrium spreading pressure (ESP) is defined as the surface pressure of the monolayer when it is in equilibrium with the stable crystalline phase (Fig. 3) (Gaines, 1966). This surface pressure is determined experimentally by depositing surfactant crystals on the water surface and allowing the development of a constant surface pressure at constant temperature over the course of several hours. This operation may be performed on a Langmuir film balance of the type shown in Fig. 1. This approach allows the experimentalist to observe the evolution of surface pressure with time. Alternatively, the ESP may be determined by surface tensiometry (duNuoy, 1929). Spreading will occur from the crystal under these conditions only if a reduction in the free energy of the interfacial system results. In principle, three-way equilibrium is set up between the crystal, the monolayer, and the surfactant monomer that has dissolved in the subphase (Gershfeld, 1982). It is generally assumed that the amount of dissolved surfactant is negligible, and that the monolayer-forming material is nonvolatile. These assumptions allow the monolayer-crystal equilibrium to be treated thermodynamically as a two-phase equilibrium analogous to the sublimation of a solid, and allows the calculation of surface excess free energy of spreading, AGs, by (4)

AG, = -IIeAe

(4)

Air Phase

Water Phase

Fig. 3 The model of equilibrium spreading from a surfactant crystal at the air/water interface. The ESP is attained when crystal, monolayer, and subphase solution are at equilibrium.

P. L. ROSE ET AL.

54

(Harkins et al., 1940), where II" is the ESP, and A" is the average area per molecule at the ESP as obtained from the I I / A isotherm of the spread film. The temperature dependence of the ESP may then be used to calculate the excess surface entropies from (5) and enthalpies of spreading from (6).

ASs = A

drI" " dT

AHs = AGs + T AS = TA'-

~

dll" - II'A" dT

(5)

(6)

Monolayer stability limits The monolayer stability limit is defined as the maximum pressure attainable in a film spread from solution before the monolayer collapses (Gaines, 1966). This limit may in some cases correspond directly to the ESP, suggesting that the mechanism of film collapse is a return to the bulk crystalline state, or may be at surface pressures higher than the ESP if the film is metastable with respect to the bulk phase. In either case, the monolayer stability limit must be known before such properties as work of compression, isothermal compressibility, or monolayer viscosity can be determined. During the film balance experiment, the monolayer stability limit is determined by compressing the spread film isothermally to a selected surface pressure, halting compression, and measuring the decay in pressure as a function of time. Two criteria for stability have been proposed: Cadenhead (1969) suggested that films decaying in surface pressure at a rate no more than 1.0 dyn cm-' min-' be considered stable, while Gershfeld (1982) has proposed a more stringent criterion of 0.1 dyncm-' min-'. It is important to note that, ultimately, all monolayers are probably metastable and that these criteria are simply useful rules of thumb for comparing relative stabilities, which must be time dependent.

n/A CURVES AND PHASE TRANSITIONS

The film balance may be regarded as a two-dimensional piston, and the most commonly studied property is the surface pressure (II) versus area ( A ) isotherm. The analogy to a PV isotherm is so appropriate that in the "gaseous monolayer regime" the two-dimensional analogue of the ideal gas law pertains: rIA = nRT. It is therefore reasonable to relate discontinuities in I I / A isotherms as the monolayer film is compressed in two dimensions to

55

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

Collapse to a rnuItiIayer

8

s,

30-

? t

“Solid”

20-

8

I ,

rn

“Gaseous”

0

10

20

30

40

Are&

50

60

70

80

90

100

/molecule)

Fig. 4 Idealized surface pressure ll versus area A isotherm detailing the inferred molecular orientation and aggregation states during a compression cycle. Reprinted with permission from Arnett et a[.,1989. Copyright 1989 American Chemical Society.

phase changes like those that take place in three-dimensional condensed matter. Figure 4 represents a typical surface pressure versus area isotherm, and the inferred orientation of molecules with respect to the surface as the monolayer is compressed from a gaseous to a solid-like state. Assuming that the polar head group is constrained to lie in the liquid surface throughout the compression cycle, what is the true nature of the changes in molecular orientation and packing as we pass from random arrangements in the gaseous state to the highly ordered “two-dimensional crystalline state”? The “two-dimensional” monolayer film may be taken through a variety of phase changes. At large areas per molecule, the film is presumed to be the two-dimensional analogue of a gas, with random interactions occurring between amphiphile molecules, and a high degree of hydrocarbon chain disorder. As the film is compressed, these chains interact in a manner analogous to those in a liquid hydrocarbon until the ‘‘lift-off’point is reached (defined as the point where a surface pressure greater than a few tenths of a dyn cm- is first registered). Further reduction in molecular area results in more highly ordered packing, with the polar headgroups interacting directly at the aqueous subphase surface and the tail sections neatly stacked above it. Over-compression past this point generally leads to a collapse of the monolayer to a multilayer or back to the crystalline state. The physical transformations are usually signalled by the presence of a

P. L. ROSE €T AL.

56

“plateau” (indicative of a first-order phase change characterized by a plateau region where the surface pressure is invariant to a change in the area per molecule), or a sudden discontinuity in the isotherm (indicative of a second-order phase change characterized by a discontinuity in the slope of the pressure/area curve). These phase changes, along with the finite thickness of the film, illustrate the true three-dimensional character of the monolayer by taking account of the layer of hydrocarbon chains immediately above the water surface. An interesting example of the manual manipulation of surfactant orientation on a water surface is shown by Fig. 5, which compares the isotherms for dipalmitoyl phosphatidylcholine with a similar compound bearing a hydroxyl group at the 12-position along the hydrocarbon chain. The enormous difference in area occupied by this molecule at low surface pressure is readily attributable to the fact that at large molecular areas this molecule has two polar groups interacting with the water surface, the usual choline system and the 12-hydroxyl group, which anchors part of the chain to the surface. Upon reduction in the available area per molecule (i.e., driving the piston in the above analogy), extra work must be invested to pull the hydroxyl group out of the surface to bring the chain to the standing position.

I

g

>. e

\

2

\

10-

,-

_ \ \ \

07

,

- - __ - _ _ A2hnolecule

Fig. 5 Surface pressure/area isotherm for the compression cycle of dipalmitoylphosphatidyl choline (dashed line) and 1-palmitoyl-2-(12-hydroxystearoyl)phosphatidyl choline (solid line) on a pure water subphase at 25°C. Reprinted with permission from Arnett et al., 1989. Copyright 1989 American Chemical Society.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

57

Once the chains are out of the water surface, the cross-sectional areas of the molecules are defined by the choline head groups and have similar limiting molecular areas. The area enclosed between the two isotherms is a quantitative measure of the extra free energy required to pull the molecule from the inchworm arrangement with the hydroxyl group in the surface to the vertical arrangement with the hydroxyl group out of the water surface (L. C . Gold, unpublished; see also Tachibana and Hori, 1977).

DYNAMIC METHODS

The dynamic behavior of fluid interfaces is usually described in terms of surface rheology. Monolayer-covered interfaces may display dramatically different rheological behavior from that of the clean liquid interface. These time-dependent properties vary with the extent of intermolecular association within the monolayer at a given thermodynamic state, which in turn may be related directly to molecular size, shape, and charge (Manheimer and Schechter, 1970). Two of these time-dependent rheological properties are discussed here : surface shear viscosity and dynamic surface tension. Surface shear viscosity

Viscosity, defined as the resistance of a liquid to flow under an applied stress, is not only a property of bulk liquids but of interfacial systems as well. The viscosity of an insoluble monolayer in a fluid-like state may be measured quantitatively by the viscous traction method (Manheimer and Schechter, 1970), wave-damping (Langmuir and Schaefer, 1937), dynamic light scattering (Sauer et al., 1988) or surface canal viscometry (Harkins and Kirkwood, 1938; Washburn and Wakeham, 1938). Of these, the last is the most sensitive and experimentally feasible, and allows for the determination of Newtonian versus non-Newtonian shear flow. As originally designed by Harkins (Harkins and Kirkwood, 1938) and modified for continuous viscosity measurement (Fig. 6 ) , the surface canal viscometer is a “two-dimensional” analogue of the familiar Ostwald viscometer. A monolayer spread and compressed to a fixed surface pressure in one compartment of the trough is allowed to flow isobarically through a canal of known length and width to a clean water surface in a second trough compartment. According to the theory of monolayer flow as proposed by Moore and Eyring (1938), the flowing monolayer may be-modeled as a series of “rows” of molecules within the viscometer canal (Fig. 7). The monolayer responds to an applied surface pressure at one end of the canal by flowing toward the low-pressure end, resulting in a shear force (or drag) between

P. L. ROSE ET AL

58

Cahn Electrobalsnce

Surface Shear Viscomefer

Fig. 6 Detail of the Verger film balance modified for measurement of surface shear viscosity. Reprinted with permission from Harvey et al., 1988. Copyright 1988 American Chemical Society.

Direction of Pressure Gradient

-

Canal Wall

/A

Fig. 7 The Eyring model of Newtonian monolayer flow. The white circles represent “holes” vacated by flowing molecules.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

59

each row of molecules and the canal wall. As the distance between a given row of molecules and the canal is increased, a laminar flow is induced in which those rows farthest from the canal wall experience the least drag and are the first to flow. In this manner, the rate of the monolayer flow is a direct function of the canal dimensions and may be used to calculate the monolayer’s shear viscosity using the Harkins-Kirkwood equation (7), where rs is the

monolayer viscosity coefficient, ylo is the viscosity of the pure water surface (usually taken as its bulk liquid viscosity), IT, - IT, is the surface pressure difference, Q is the film flow rate, L is the canal length, and Wis the canal width. The viscosity exhibited by a monolayer may be Newtonian, where the viscosity is independent of the rate of shear, or non-Newtonian, where viscosities vary with the rate of flow. Experimentally, this may be determined simply by varying the width of the canal and measuring the surface viscosity at different A l l (Jarvis, 1965). According to Eyring (Moore and Eyring, 1938) and Joly ( 1956), Newtonian flow in a monolayer is the result of a cohesive attraction between surfactant molecules. For every molecule that flows from higher to lower surface pressure in a motion parallel to the canal walls, there is another molecule ready to “fill the hole” vacated by the first. The mechanism for this cohesive flow is presumably attractive van der Waals interactions between hydrocarbon chains. This model assumes that the average intermolecular separation in a surface-continuous monolayer does not exceed the cross-sectional area of the molecule as defined by the average molecular area A of the film at the surface pressure l7 in the pressurized compartment of the viscometer. Non-Newtonian flow may result if the monolayer array consists of molecules that interact by specific Coulombic or dipole interactions to form “floating islands”, which in turn may interact by van der Waals forces around their peripheries (Joly, 1956). Non-Newtonian flow may also be a property of collapsed films. The resulting differences in viscosity over a range of flow rates may then reflect film-component segregation or partial monolayer collapse. The surface shear viscosity of a monolayer is a valuable tool in that it reflects the intermolecular associations within the film at a given thermodynamic state as defined by the surface pressure and average molecular area. These data may be used in conjunction with l l / A isotherms and thermodynamic analyses of equilibrium spreading to determine the phase of a monolayer at a given surface pressure. This has been demonstrated in the shear viscosities of long-chain fatty acids, esters, amides, and amines (Jarvis, 1965). In addition,

60

P. L. ROSE ET AL.

factors such as temperature, surface pressure, and subphase pH have been shown to alter monolayer flow properties dramatically (Jarvis, 1965; Joly, 1956). It is not surprising, therefore, that this dynamic technique was chosen to provide information on chiral interactions in compressed films. Given that these stereochemically differentiated systems may have dramatically different l l / A isotherm characteristics, and hence different packing arrangements, it is plausible that their flow properties are stereochemically differentiated as well. Dynamic surface tension

The dynamic surface tension of a monolayer may be defined as the response of a film in an initial state of static quasi-equilibrium to a sudden change in surface area. If the area of the film-covered interface is altered at a rapid rate, the monolayer may not readjust to its original conformation quickly enough to maintain the quasi-equilibrium surface pressure. It is for this reason that properly reported l l / A isotherms for most monolayers are repeated at several compression/expansion rates. The reasons for this lag in equilibration time are complex combinations of shear and dilational viscosities, elasticity, and isothermal compressibility (Manheimer and Schechter, 1970; Margoni, 1871;Lucassen-Reynders et al., 1974).Furthermore, consideration of dynamic surface tension in insoluble monolayers assumes that the monolayer is indeed insoluble and stable throughout the perturbation; if not, a myriad of contributions from monolayer collapse to monomer dissolution may complicate the situation further. Although theoretical models of dynamic surface tension effects have been presented, there have been very few attempts at experimental investigation of these time-dependent phenomena in spread monolayer films. The difference between the static or equilibrium and “dynamic” surface tension is often observed in the compression/expansion hysteresis present in most monolayer l l / A isotherms (Fig. 8). In such cases, the compression isotherm is not coincident with the expansion one. For an insoluble monolayer, hysteresis may result from very rapid compression, collapse of the film to a surfactant bulk phase during compression, or compression of the film through a first or second order monolayer phase transition. In addition, any combination of these effects may be responsible for the observed hysteresis. Perhaps understandably, there has been no firm quantitative model for time-dependent relaxation effects in monolayers. However, if the basic monolayer properties such as ESP, stability limit, and composition are known, a qualitative description of the dynamic surface tension, or hysteresis, may be obtained.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

61

“Clockwise”Hysteresis Rotation

c

1 w

Compression

“Counterclockwise”Hysteresis Rotation

ArealMolecule

Fig. 8 Representative compression and expansion l l / A isotherms showing “clockwise” and “counterclockwise” hysteresis.

The simplest model of hysteresis in monolayers is based upon a loose analogy with the closed Carnot cycle (Dragcevic et al., 1986), in which the work of compression is compared to the work of expansion for a stable film in order to obtain the net work done during the cycle. The work of compression from an initial average molecular area A , to the final average molecular area A , at which the monolayer is stable is the change of Helmholtz free energy of compression F, (this symbol is used in surface chemistry because A is universally used for area) given by (8). Conversely, the work

AF,= -j;IIdA of expansion from A , to A , is described by (9). The net. work of the cycle,

AFcycle,may then be expressed as ( 10). The simple difference between the AFcycle= AF,

- AF,

(10)

62

P. L. ROSE E T A L .

work of compression and the work of expansion allows for a simple treatment of the hysteresis “loop” exhibited in the n / A isotherm. As shown in Fig. 8, the compression/expansion hysteresis may be ‘‘clockwise’’ or “counterclockwise” in nature. If the hysteresis is clockwise, the expansion regime occurs at a higher surface pressure than the compression throughout the cycle. The net work done by the surroundings on the monolayer during the closed cycle, as calculated from the simple relations above, is therefore negative, and the monolayer system is considered to have done work on the surroundings. In essence, this implies that the film is elastic, or springy, in its dynamic response to compression. If the sense of hysteresis is counterclockwise, the net work done by the surroundings on the monolayer system is positive, and energy that was expended upon compression has been lost to some other process (monolayer component reorganization, film collapse, dissolution, heat transfer, etc.) upon expansion back to the original average molecular area. The sense of hysteresis has been shown to depend significantly on the rate of film compression and expansion (Munden and Swarbrick, 1973). This treatment of monolayer dynamic surface tension is simplistic in that it does not take into account the fact that most monolayers are metastable, and therefore may not be used for quantitative thermodynamic analysis of irreversible systems. However, it does provide a qualitative comparison of the dynamic properties of differing film systems, and the direction of energy transfer for the cycle may be used in determining the nature of molecular interactions during dynamic processes. The few examples of deliberate investigation of dynamic processes as reflected by compression/expansion hysteresis have involved monolayers of fatty acids (Munden and Swarbrick, 1973; Munden et al., 1969), lecithins (Bienkowski and Skolnick, 1974; Cook and Webb, 1966), polymer films (Townsend and Buck, 1988) and monolayers of fatty acids and their sodium sulfate salts on aqueous subphases of alkanolamines (Rosano et al., 1971). A few of these studies determined the amount of hysteresis as a function of the rate of compression and expansion. However, no quantitative analysis of the results was attempted. Historically, dynamic surface tension has been used to study the dynamic response of lung phosphatidylcholine surfactant monolayers to a sinusoidal compression/expansion rate in order to mimic the mechanical contraction and expansion of the lungs. Until very recently, there has been little or no experimental protocol for obtaining quantitative dynamic surface tension data on monolayer films. In most cases, the experimental set-up has consisted of a simple Langmuir film balance equipped with a variable-speed motor to drive the moving barrier. Hysteresis data were then obtained at a number of compression/expansion rates and compared qualitatively. This experimental set-up was improved considerably by Johnson (Arnett et al., 1988a), who modified a special

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

63

To Electmbalance

Drive Cable Detector Bamer System

Trough

Fig. 9 Schematic of the modified dynamic surface tension apparatus.

Langmuir trough (Cahn Instruments Surface Tension Accessory No. 099-002910-30)in which the film could be rapidly compressed and expanded at a linear rate by means of moving barriers at either end of the balance (Fig. 9). This equipment was redesigned to reduce leaks and to follow a linear rather than sinusoidal cycle. Unlike the Langmuir film balance experiment, the dynamic surface tension (DST) apparatus compresses and expands the film from both ends of the small Teflon trough (11.4 cm x 4.8 cm x 1.8 cm) by means of two moving Teflon barriers, which are connected by a clean Viton elastomer (12 cm x 1.0 cm x 0.035 cm). Over the period of the experiment, the barriers are driven at a constant rate by means of a spindle and pulley system powered by a Hurst CA 1/4 rpm motor. The rates of compression and expansion may be varied by a set of interchangeable gears of 3:1,2: 1, 1: 1, 1:2, and 1:3 gear ratio. Ultimately, the barrier compression/expansion cycle speeds may be varied from 0.11 to 1.0 cycle min-'. The reversal of compression and expansion modes is accomplished by means of an infrared emitter-detector diode pair that is switched by a pair of indicator flags attached to the drive cable. The concomitant changes in surface pressure can then be detected via a Wilhelmy plate, which is connected to a Cahn RG electrobalance detector.

MIXING CRITERIA AS APPLIED TO THE ELUCIDATION OF INTERMOLECULAR INTERACTIONS AT SURFACES

It is possible to study monolayers composed of more than one compound by the techniques described above. Mixed monolayers are of considerable

P. L. ROSE ET AL.

64

interest because of their relevance to natural systems, such as biological membranes. If one considers that the monolayer properties of these multicomponent films vary with the composition of the film, binary monolayers provide an opportunity to study the effects of molecular structure on intermolecular interactions between species whose only common feature is surface activity. Conversely, these films also provide an opportunity to check interactions between molecules whose main structural difference is isomerism. One may then reasonably treat monolayers cast from enantiomers as a specialized case of a mixed binary monolayer. Of course, films cast from diastereomers, structural isomers, or purely nonisomeric pairs may be treated in the same manner. The first question that must be addressed in considering interactions in binary monolayers is whether or not an intimately mixed monolayer is formed at all. Unlike three dimensional systems (such as a mixture of oil and water), the miscibility or immiscibility of the two monolayer components usually cannot be determined visually. Spreading a mixture of surfactants from a homogeneous solution of surfactants 1 and 2 does not ensure that they mix homogeneously at the air-water interface (Fig. 10). If the film components are immiscible, they will segregate into separate microdomains (Gershfeld, 1972). The small size of these domains renders their direct microscopic observation or physical detection difficult, so that thermodynamic methods are needed to provide criteria for film component miscibility. Non-ideally miscible and ideally miscible components will of course have differing degrees of randomness in their component distribution, which depend on the mole fraction of each component in (or composition of) the film. As in the case with three-dimensional systems, the associations between film components may then be inferred by the dependence of monolayer properties on composition. Two methods have been used classically to determine component

Segregated

Mixed

Fig. 10 Idealized representation of mixed and unmixed monolayers.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

65

miscibility in binary films, using data readily obtained from the l T / A isotherm: (i) the monitoring of average molecular area (or some other monolayer property) as a function of film composition at constant l7 (Gaines, 1966); and (ii) the monitoring of phase transition pressures for one of the components as a function of film composition (Crisp, 1949; Defay, 1932). When these tests indicate ideal or at least partial miscibility, quantitative methods may be used to determine the energetics of mixing (Goodrich, 1957). Combined, these techniques provide a potentially powerful tool for determining the stereospecific associations in monolayer mixtures of chiral surfactants. Determination of miscibility by additive properties such as average molecular area can give only qualitative interpretations of I I / A data. For example, if the film components are ideally miscible, the average molecular area of the binary film at a fixed surface pressure will be the sum of each of the molecular areas of the individual components 1 and 2 in their pure films, and will follow equation ( 1 l), where N is the mole fraction. Unfortunately,

exactly the same result will be obtained if the components are completely immiscible, since they will segregate and occupy the same average molecular areas in their domains as they would in their separate pure films. At best, a linear relationship between the measured molecular area and composition may indicate ideal miscibility or complete immiscibility at a given surface pressure, but cannot distinguish between them; however, any deviation from this ideal line is useful in that it indicates at least partial miscibility. This additive relationship may be applied to any other property of the system (i.e., viscosity, surface potential, etc.) under isothermal and isobaric conditions with the same result (Gaines, 1966). When a reversible transition from one monolayer phase to another can be observed in the n/Aisotherm (usually evidenced by a sharp discontinuity or plateau in the phase diagram), a two-dimensional version of the Gibbs phase rule (Gibbs, 1948) may be applied. The transition pressure for a phase change in one or both of the film components can be monitored as a function of film composition, with an ideally miscible system following the relation (12). A completely immiscible system will not follow this ideal law, but will

yield instead a flat, horizontal plot (Fig. 1 1 ) showing no deviation from transition pressure of the pure components (Gershfeld, 1972). This result can be understood in the light of Crisp and Defay’s proposal of a “two-dimensional” phase rule (Crisp, 1949; Defay, 1932). At constant

P. L. ROSE E T A L .

66

0

100

Composition

100 % # 2 0 %#1

Fig. 11 Defay-Crisp diagram for a binary monolayer: A, ideal mixing; B, non-ideal mixing; C , complete immiscibility. II, and II, are the phase transition pressures of components 1 and 2.

temperature and external pressure, the number and degrees of freedom, f, for the system is given by (13), where Cb is the number of components

equilibrated throughout the system (air and water), C, is the number of components in the surface monolayer (2), P, is the number of bulk phases equilibrated throughout the system (air and water), and Psis the number of surface phases in equilibrium with each other. The quantity 2 represents the two variables surface pressure and film composition. If components 1 and 2 are miscible, then the number of surface phases in equilibrium at the transition pressure is two, and f = 2. In this case, the surface pressure varies continuously with film composition. If the components are immiscible, the number of surface phases in equilibrium at the transition pressure will be 3, andf= 1. Variation of film composition will not alter the transition pressure. The utility of this method is limited by the fact that not all spread films show a true phase transition in their l l / A isotherms and also by the dubious application of equilibrium thermodynamics to metastable monolayers. Gershfeld has determined carefully the gas-liquid transition pressures of simple fatty acid mixtures (Pagano and Gershfeld, 1972) and the phosphatid ylcholine-cholesterol binary film system (Gershfeld and Pagano, 1972a,b,c)and applied the Defay-Crisp phase rule. However, these transitions occur at surface pressures of only a few mdyn cm- and are out of the range of detection for most film balances. This presents a problem in dealing with

',

67

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

binary films in which neither of the components undergoes a readily observed, reversible phase change in the course of compression. Furthermore, this method indicates miscibility only at surface pressures equal to the transition pressure and higher (Crisp, 1949; Defay, 1932). The phase rule has been applied more conveniently to ESP measurements taken as a function of temperature. Again, Gershfeld (1982) has shown that a plateau or discontinuity in the ESP versus temperature plot may be indicative of a three-way equilibrium between the floating crystal and the separate monolayer phases that have spread from this crystal. This treatment has been used to argue for the existence of surface bilayers of phosphatidylcholine derivatives (Gershfeld, 1986, 1988). Once it has been established that the components of a binary monolayer are to some degree miscible, the energetics of their interaction may be calculated directly from the II/A isotherms of the mixture and its individual components. As proposed by Goodrich (1957), this technique employs the differences in the work of compression of the binary film and the work required to compress each of the films of the pure components to the same surface pressure. The result is the total free energy of mixing as expressed by the sum of the excess and ideal free energies of mixing in (14), where N ,

S,n

Acmix

=

( A , , , - N , A , - N,A,)dn

+ R T ( N , In N , + N , In N , ) (14)

and N , are the mole fractions of components 1 and 2, A , and A , are the average molecular areas of components 1 and 2 in their own pure films at n, and A , , , is the average molecular area of the mixed film at a given n. The excess free energies of mixing are then easily determined by integrating the area under the I I / A isotherms from ll = 0 to the specified surface pressure n. An excess free energy versus composition diagram may be plotted if data are taken at several different compositions, affording a quantitative determination of the energy required for non-ideal mixing. The excess entropy and enthalpy of mixing may also be determined if the n/A isotherms are obtained at several temperatures. The inherent assumptions of this technique are: (i) the monolayers of the components and their mixtures are stable and behave reversibly, and the upper limit of integration is below the ESPs of each of the components; (ii) ideal free energies of mixing describe the mixing behavior of the films accurately at very large average molecular areas so that the ll = 0 lower limit of integration may be assumed valid at the film’s lift-off point; (iii) the film components do not reorganize or reaggregate during the compression process; and (iv) the monolayer is surface-continuous at all surface pressures.

68

P. L. ROSE ET AL.

These assumptions have been expanded upon (Shah and Capps, 1968; Lucassen-Reynders, 1973; Rakshit and Zografi, 1980), especially in regard to the application of the ideal mixing relationship in gaseous films (Pagano and Gershfeld, 1972). It has been pointed out that water may contribute to the energetics of film compression if the molecular structures of the surfactants are sufficiently different (Lucassen-Reynders, 1973). It must be noted that this treatment assumes that the compression process is reversible and the monolayer is truly stable thermodynamically. It must therefore be applied with considerable reservation in view of the hysteresis that is often found for II/A isotherms. Excess free energies of mixing have been used indirectly to determine the effects of surfactant structure on mixing. Excess thermodynamic properties have been reported for structurally simple molecules such as the straight chain fatty acids (Pagano and Gershfeld, 1972; Rakshit and Zografi, 1980), but have also been determined for more complex systems, such as vitamin K , and chlorophyll-a (Gaines et al., 1965), octadecanol and docosyl sulfate (Goodrich, 1957), and hexadecanol/tetradecanol mixtures, with sodium hexadecyl sulfate (Goodrich, 1957). Recently, free energies of interaction were determined for mixtures of chlorophyll-a with a-tocopherylquinone and plastoquinone-3 and -9 (Guay and LeBlanc, 1987). Comparisons of the effect of molecular structure on mixing suggested that the carbonyl moieties of the quinone were responsible for the large negative deviations from ideal mixing in mixtures with chlorophyll-a. The implications for films cast from mixtures of enantiomers is that diagrams similar to those obtained for phase changes (i.e., melting point, etc.) versus composition for the bulk surfactant may be obtained if a film property is plotted as a function of composition. In the case of enantiomeric mixtures, these monolayer properties should be symmetric about the racemic mixture, and may help to determine whether the associations in the racemic film are homochiral, heterochiral, or ideal. Monolayers cast from nonenantiomeric chiral surfactant mixtures normally will not exhibit this feature. In addition, a systematic study of binary films cast from a mixture of chiral and achiral surfactants may help to determine the limits for chiral discrimination in monolayers “doped” with an achiral diluent. However, to our knowledge, there has never been any other systematic investigation of the thermodynamic, rheological and mixing properties of chiral monolayers than those reported below from this laboratory. TECHNIQUES FOR VISUALIZING FILMS AND AGGREGATES

Although the standard l l / A isotherms of chiral monolayers can disclose much information on stereospecific packing arrangements, they do not allow

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

69

for direct visualization of packing patterns. Standard crystallographic techniques can be used on the bulk solids of the surfactants used to form the monolayer, but there is no guarantee that the packing arrangements responsible for the display of chiral discrimination within the film are identical to those in the bulk solid phase. X-ray crystallography and ellipsometry have been performed on Langmuir-Blodgett films of fatty acids and related materials transferred to solid substrates, but these techniques can only give quantitative information on the orientation of the headgroup and the tilt of the hydrocarbon chains at highly defined surface pressures. We have found that, when combined with the quasi-thermodynamic information provided by II/A isotherms, surface rheology, and equilibrium spreading pressures, direct visualization of the monolayers and film aggregates via ultramicroscopy provides useful insights into stereospecific packing. nansmission and scanning electron microscopy

Electron microscopy of Langmuir-Blodgett films is generally confined to multilayers of surfactant deposited on to a solid substrate. Scanning electron microscopy allows for direct visualization of the uppermost layer of molecules deposited on the substrate, and this may be sufficient for the distinction of packing differences in chiral films. Transmission electron microscopy can be used in the same multilayer aggregates to give information on the number of layers deposited, the average thickness of each layer, and the topographical features of the uppermost layer. These techniques are currently incapable of resolving reproducibly the features of a single-layer small molecule film. Because they operate reliably on only multilayer films, they are not adequate for describing the more subtle packing arrangements in loosely aggregated monolayers. Scanning tunneling microscopy

Scanning tunneling microscopy (STM) allows for the imaging of a single layer of non-insulating molecules transferred to a solid substrate. Under optimal conditions, scanning tunneling microscopy has the necessary resolution to provide visual evidence of stereo-dependent packing between R- and S-enantiomers versus interactions between S- and S- and R- and R-enantiomers. At lower resolutions, the technique can provide dramatic pictures of chiral packing in a broad monolayer array or in a sufficiently conductive multilayer. Unfortunately, the films of surfactant must be stationary and highly ordered for the film to be effectively imaged. This is usually accomplished by transferring the monolayer from the water surface to a solid, conducting surface (usually pyrolytic graphite). This transfer does

70

P. L. ROSE ET AL.

not allow for an accurate description of the film as it sits on the water surface, and there is no guarantee that the film is ordered on the solid substrate in the same fashion as at the air-water interface. However, the simplicity of the technique and the availaility of several commercial instruments and imaging software make it an attractive investigative tool. Epijluorescence optical microscopy

Unlike electron and scanning tunneling microscopy, the use of fluorescent dyes in monolayers at the air-water interface allows the use of contrast imaging to view the monolayer in situ during compression and expansion of the film. Under ideal circumstances, one may observe the changes in monolayer phase and the formation of specific aggregate domains as the film is compressed. This technique has been used to visualize phase changes in monolayers of chiral phospholipids (McConnell et al., 1984, 1986; Weis and McConnell, 1984; Keller et al., 1986; McConnell and Moy, 1988) and achiral fatty acids (Moore et al., 1986). The optical epifluorescence technique utilizes a small Langmuir trough (usually made of Teflon) mounted on the stage of a photomicroscope. The surface pressure of a spread film can be manipulated if the trough is equipped with a moveable barrier driven by a smoothly operating motor. The surface pressure can be monitored by use of a small Wilhelmy plate and electrobalance, and disruption of the film-covered surface by air currents can be minimized by placing a microscope coverslip or some other optically transparent glass over the trough. Visualization is accomplished by mixing dilute solutions of the surfactant of interest with dilute solutions of a fluorescent-label tagged analogue of the surfactant. For example, in work performed with chiral phospholipids, an NBD-functionalized phospholipid was used as a dye, and added to the lipid-spreading solution in two mole percent concentration. The solution is spread on the water surface, the spreading solvent allowed to evaporate, and fluorescence is then excited by an argon or helium laser. As the film is compressed, the image of the monolayer usually appears dark against a brightly fluorescing field. This allows for direct observation of monolayer aggregation during compression. Despite its attractive capabilities, the epifluorescence technique has some drawbacks. The fluorescent surfactant probe must not be miscible with the major phase of interest, and must not interact with the major phase in any way that changes the rheological flow or compression characteristics of the film. In addition, the probe itself must form a stable monolayer on the air-water interface. The area in which this work is to be performed must also be clean enough for accurate film balance work and must be free of vibration.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

3

71

Chiral monolayers

The following sections describe examples of the application of these approaches to the study of chiral molecular recognition in thin film systems, using the battery of techniques outlined above.

ENANTIOMERIC SYSTEMS

N- (a-Methylbenzy1)stearamide

Our first investigations of the stereospecific aggregation of molecules in a monolayer involved the use of a novel chiral surfactant, N-(amethylbenzyl)stearamide, spread on aqueous acid subphases ( Arnett and Thompson, 1981; Arnett et al., 1982). This surfactant was chosen for study because of the potential for strong hydrogen bonding between enantiomers, which should in theory yield closely packed aggregates in a film system. In the bulk crystalline phases, large differences exist in the properties of the racemic mixture and the pure enantiomers. X-ray powder diffraction patterns showed that the racemic mixture was a true racemate, and the melting transition points and heats of fusion of the racemate were markedly different from those of the pure enantiomers [which were identical (Arnett and Thompson, 1981)l. Enantiomeric recognition was clearly displayed in films spread from solution and films in equilibrium with their crystals, and was sharply dependent on the acidity of the subphase. Protonation of the amide group appeared to be necessary for spreading to stable monolayers. For example, the crystals of the racemate deposited on a ION H,S04 solution at 25°C spread quickly to yield a film with an ESP of 7.7 dyn cm-', while the single enantiomers spread only to a surface pressure of 3.9dyncm-' (Table 1). Similar effects are observed at 15 and 35°C. The effect of stereochemistry on equilibrium spreading is even more pronounced at lower subphase acidities. On 6~ sulfuric acid, the racemate spread to an equilibrium surface pressure of 4.9 dyn cm-', while the enantiomeric systems spread to less than 1 dyn cm-'. When spread from dilute hexane solution, acid-dependent enantiomeric discrimination was observed in the l l / A compression isotherms of the monolayer at 25°C (Fig. 12). It is interesting to note that at higher subphase acidities, both racemic and enantiomeric film systems become more highly expanded, and the surface pressures where enantiomeric discrimination commences occur at high (85-90 A2/molecule) average molecular areas. This may be taken as direct evidence of headgroup ionization effects. The surface

Table 1 Equilibrium spreading pressures ( ESPs) of racemic and optically pure N - ( cemethylbenzy1)stearamides at various temperatures and subphase acidities."*b ION H,SO, Racemate

6~ H,SO, Enantiomers

Temp./"C

II/dyn cm-'

Areac

lT/dyn cm-

15 25 35

4.8 f 0.3 7.7 f 0.3 20.9 & 0.6

76 f 3 70 f 3

0.89 f 0.1 3.9 f 0.4 10.3 & 0.5

Racemate

Enantiomers

Area'

II/dyn cm-'

Area'

lT/dyn cm-'

Area'

86 f 3 82 f 3

4.9 f 0.2

73 f 1

0.6 f 0.2

77.5 f 0.1

"Reprinted with permission from Arnett et al., 1982. Copyright 1982 American Chemical Society. obtained from normalized 11-A isotherms, is area at He at 15 and 12°C. Weight of crystals used was about 300mg. Area of trough was 565.4cm2. lTeis equilibrium spreading pressure (ESP); A' is area per molecule at He. 'Az/molecuIe.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

301 7

5

0

---------"r% H

20 -

I

....,.

I11

U h

2

73

10-

***... t \

'....

I1 '\* *.

--&+ I

Fig. 12 Force/area curves of stearamide films on 6~ H$O, at 25°C: 1, natural racemate; 11, enantiomeric stearamide; 111, mixture of solutions of enantiomers on the surface in 1: 1 ratio. Reprinted with permission from Arnett and Thompson, 1981. Copyright 1981 American Chemical Society.

pressures required to pack the racemic system are higher at every average molecular area than that required in the enantiomeric systems. In addition, the racemic monolayer is more highly expanded as a function of subphase acidity than the enantiomeric system at surface pressures less than 5 dyn cm- Above this surface pressure, the enantiomeric system becomes unstable and apparently collapses to some other surface state at subphase . observation is consonant with the ESP acid concentrations less than 6 ~This values obtained from the bulk crystals (Table 1). Several aspects of the properties of the chiral N - ( wmethylbenzy1)stearamide film systems indicate that kinetic, as well as thermodynamic, factors may contribute to enantiomeric discrimination. The instability of the spread films of the enantiomeric system and the low ESP values of films spread from enantiomeric crystals suggest that slow collapse to a surface crystalline state may be responsible for film instability. Conversely, the racemic monolayer spread from solution is stable at higher ( > 5 dyncm-') surface pressures and attains a higher surface pressure when in equilibrium with its crystal, ~ acid subphase. However, when a racemic film of especially on 1 0 sulfuric R( + )- and S ( - )-N-(wmethylbenzy1)stearamide is spread on this same subphase from solutions of the separate enantiomers (first depositing a film of one enantiomer and then depositing a film of the opposite enantiomer over the same surface), the I I / A isotherm of the resulting monolayer is very similar to that of the single enantiomer. These seemingly contradictory data suggest that, when spread from a homogeneous solution, collapse of the racemic film to separate enantiomeric crystalline domains is probably

'.

74

P. L. ROSE ET A L .

inhibited by the presence of the opposite enantiomer. Resolution of the racemic film into separate homochiral domains may therefore be kinetically inhibited. These results for spread film and equilibrium spreading suggest that films of racemic N - ( a-methylbenzy1)stearamide may be resolved by seeding the racemic film with crystals of either pure enantiomer. Indeed, when a monolayer of racemic N-( a-methylbenzy1)stearamide is compressed to 45 A'/molecule (27 dyn cm-'), deposition of a crystal of either R ( + )- or S ( - )-enantiomer results in a decay of surface pressure from the initial 28 dyn cm-' film pressure to 3.0 dyn cm-', the ESP of the enantiomeric ~ acid subphase (Table 1). When the experiment systems on a pure 1 0 sulfuric is repeated with racemic crystals, the system reaches an equilibrium surface pressure of 11 dyn cm- l, nearly the ESP of the racemic crystal on the clean acidic interface. In either case, equilibrium pressure is reached within a two hour time period. The implication of these results is that deposition of, for example, R( +)-crystals on to the racemic films provides a nucleation site for R( + )-molecules in the film, leaving behind a partially resolved film of predominantly S ( - )-molecules. Deposition of S ( - )-crystals should, alternatively, leave behind a film composed predominantly of R( + )molecules. This model is supported by the ESP data obtained on the clean acidic surface, where the free energy of enantiomer crystals appears to be lower compared with liquid-like film states than that of the racemic crystals. The combined data given above indicate that intermolecular attractions within the monolayer at any given state are stronger for the enantiomeric systems than for their racemic mixture. It is evident from the fact that the film systems are formed only on strongly acidic subphases that hydrogen bonding and ionization of the headgroups are major factors in the transmission of shape-selective information from one molecule to the next. The quasi-thermodynamic ESP data and film stabilities, combined with the reported resolution phenomena, suggest that enantiomeric recognition in these films occurs most readily in condensed film states. The question remains as to whether enantiomeric discrimination may be observed in the more highly expanded, fluid-like states that are characteristic of most biological surfactants in uiuo. Phospholipids

Phospholipids are perhaps the most ubiquitous of chiral surfactants in cell biology. It is well known that only the L-isomer is naturally occurring in the cell membrane of most living organisms, yet the question of whether or not this homochirality plays a role in the regulation of cell chemistry has barely

75

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

been addressed. Such a large proportion of biochemical reactions are stereospecific that one might reasonably conclude that membrane chirality plays a significant role in cell processes. The environment of a cell membrane is often modeled by a monolayer of phospholipid on the air-water interface. Our attempts to detect enantiomeric recognition in such films has largely involved dipalmitoylphosphatidyl choline (DPPC), which has a chiral headgroup situated at the junction of two 16-carbon unit chains. The l l / A isotherms of the racemic and enantiomeric forms of DPPC are identical within experimental error under every condition of temperature, humidity, and rate of compression that we have tested. For example, the temperature dependence of the compression/expansion curves for DPPC monolayers spread on pure water are identical for both the racemic mixture and the D- and L-isomers (Fig. 13). Furthermore, the equilibrium spreading pressures of this surfactant are independent of stereochemistry in the same broad temperature range, indicating that both enantiomeric and racemic films of DPPC are at the same energetic state when in equilibrium with their bulk crystals. The dynamic surface tension properties of the heterochiral and homochiral DPPC films are also independent of stereochemistry. Figure 14 shows the hysteresis loops of five successive compression/expansion cycles obtained on the modified Cahn DST apparatus. Although all five compression/expansion cycles of the DPPC films are not coincident, the areas and shapes of the

-'

6

30-

h

ze

2010 - _ _ _ I

50

60

70

80

90

100

110

A */ molecule Fig. 13 Force/area curves of dipalmitoylphosphatidyl choline monolayers spread on ure water at 25°C (solid line) and 45°C (dashed line). The compression rate is 7.2 '/molecule per minute. The shape of the isotherms is identical for homochiral and heterochiral films.

x

P. L. ROSE ET AL.

76

1 . a

40

2 3

I

20

70

50

30

W * / molecule Fig. 14 Compression/expansionhysteresis loops for monolayers of dipalmitoylphosphatidyl choline at 25°C on pure water subphase. Rate of compression/expansion is 12.5 A2/molecule per minute.

hysteresis loops are independent of chirality. Furthermore, this lack of stereodifferentiation is constant throughout the 15-60°C temperature range and at every compression/expansion rate we have tested. Shorter chain analogs of DPPC were also investigated in order to determine if the lack of stereo-differentiation in monolayer properties could be due to DPPC’s higher gel point or complicating steric effects. Figure 15 shows the compression/expansion isotherms of DPPC as compared with racemic and enantiomeric dimyristoylphosphatidyl choline (DMPC) and dilauroyl phosphatidyl choline (DLPC). Again no stereodifferentiation in monolayer properties was observed as reflected by l T / A isotherms or dynamic surface tension. One point of interest in the comparison of DPPC with its shorter chain analogs is the apparent first-order phase transition in the DPPC compression/expansion isotherm. The constant surface pressure plateau is indicative of a change in film state from a highly expanded surface phase to a more tightly packed, highly condensed film. In the latter intermolecular associations are dominated by short-range forces, which should yield some expression of stereo-recognition in film properties. Although we have not been able to detect any dependence of the phosphatidyl choline monolayers’ thermal or mechanical properties on chirality in this I I / A region, epifluorescence microscopy has yielded dramatic evidence of chiral aggregation in the plateau region. McConnell and others (McConnell et al., 1984, 1986; Weis and McConnell, 1984; Keller et al., 1986; McConnell and Moy, 1988) have shown that films of the ( - )-isomer form star-shaped aggregates having a swirl-like pattern to the domain symmetry opposite to that of films cast from the ( + )-isomer when compressed through the constant pressure region

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

40

77

t

I

E

U

h c

2-

Fig. 15 Force/area curves for monolayers of dipalmitoylphosphatidyl choline (solid line), dimyristoylphosphatidyl choline (dashed line), and dilauroylphosphatidyl choline (dotted line) at 25°C.

of the I I / A isotherm. The swirl symmetry is a result of stereoselective packing and consists of a rotation of the domain shape, which is mirrored in the shape of ( - )- versus ( + )-domains. Conversely, films cast from the racemic mixture have aggregates exhibiting no definite symmetry when compressed to this film state. The question remains as to why this stereo-differentiation is not as easily detected in film properties. In any event, these results highlight the extremely subtle nature of the forces which play a role in chiral molecular recognition. Stearoylamino acids and esters

Both the N-(a-methylbenzy1)stearamide and phospholipid systems as detailed above proved to be difficult systems with which to work. The inability of N - ( a-methylbenzy1)stearamideto form stable monolayers or even to spread from the crystal on anything but very acidic subphases presents a significant technical challenge despite the presence of a chiral headgroup that is unobstructed by other molecular features. On the other hand, the phospholipid surfactants that spread to form stable films both from solution and from their bulk crystals on pure water subphases at amBient temperatures displayed no discernible enantiomeric discrimination in any film property. The chiral functionality on these biomolecules is apparently shielded from intermolecular interactions with other chiral centers to the extent

78

P. L. ROSE € T A L .

that we are unable to detect any expression of stereoselectivity in monolayer properties. One may contemplate the significance of this lack of stereoselectivity in such an important living structure as the cell membrane considering the widespread employment of chiral recognition for natural selection in many other biological systems. The primary focus of our work with stearoylamino acids and esters was essentially the same as the above-mentioned efforts with the phospholipids: to test the extent of molecular recognition between chiral molecules of biological significance. While our work with the phospholipids was limited to natural materials and their antipodes, the work discussed here is based upon fatty versions of natural amino acids constructed by us in the laboratory. This work is also related to the N-(a-methylbenzu1)stearamidestudies in that the molecules as constructed are simple, single-chain surfactants with an exposed chiral headgroup well-known for intermolecular hydrogen bonding. N-Stearoylamino acids and their methyl esters were synthesized from enantiomeric and racemic forms of tyrosine, serine, alanine, and tryptophan (Fig. 16). Analogs of these molecules were investigated initially over 30 years ago by Zeelen and Havinga, who found stereochemical differentiation in the monolayer n/A isotherms of these materials (Zeelen, 1956; Zeelen and Havinga, 1958). We have extended this study using more sensitive Langmuir balances, a wider array of dynamic and equilibrium techniques, and the N-stearoyl methyl esters of the amino acids (Harvey et al., 1989; Harvey and Arnett, 1989). Figure 17 shows the l l / A isotherms of racemic and enantiomeric films of the methyl esters of N-stearoyl-serine, -alanine, -tryptophan, and -tyrosine on clean water at 25°C. Although there appears to be little difference between the racemic and enantiomeric forms of the alanine surfactants, the N-stearoyl-tyrosine, -serine, and -tryptophan surfactants show clear enantiomeric discrimination in their II/A curves. This chiral molecular recognition is first evidenced in the lift-off areas of the curves for the racemic versus enantiomeric forms of the films (Table 2). As discussed previously, the lift-off area is the average molecular area at which a surface pressure above 0.1 dyn cm- is first registered. The packing order differences in these films, and hence their stereochemical differentiation, are apparently maintained throughout the compression/expansion cycles. In addition, it should be noted that none of the compression and expansion cycles for these films are coincident. The considerable hysteresis exhibited during the compression/expansion cycle is evidenced at every compression/expansion rate investigated, and is indicative of a stereoselective kinetic process that must occur upon film compression. Table 3 gives the monolayer stability limits of the amino acid methyl ester films as defined by

79

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

OH

HO

Stearoyltyrosinemethyl ester

R-(4-

%+)-

Stearoylserine methyl ester

S(+b

Stearoylalaninemethyl ester

S(+)-

R-0-

H,C02C C,,H,,CONH

y7

R-C)-

H

3

C,,H,,CONH

c

0

o"'H

2

c

Q $0 y2 ,cH3

/ \ / \ / \ N

I

H

R-H-

w

NHCOC17H35

N

I

H

Stearoyltryptophanmethyl ester

S(+b

Fig. 16 Chiral amino acid soaps.

Table 2 Lift-off areas per molecule for different chiral surfactants studied at 25°C on a pure water subphase.

Surfactant Stearoylserine methyl ester Stearoylalanine methyl ester Stearoyltryptophan methyl ester Stearoyltyrosine methyl ester

Enantiomeric monolayer"

Racemic monolayer

42.5 f 0.9* 51.5 f 3.2 75.9 f 1.4 74.5 & 0.6

67.4 f 1.4 49.8 f 3.2 79.3 f 3.5 59.9 5.1

Compression and expansion rates of 7.1 A'/molecule/min.: Values in units of A2/molecule. bStandard deviations are at the 95% confidence level.

P.

80

L. ROSE E T A L .

D

Azhnolecule

Fig. 17 Surface pressure/area isotherms for the compression and expansion cycles of racemic (dashed line) and enantiomeric (solid line) stearoylserine (A), stearoylalanine (B), stearoyltryptophan (C), and stearoyltyrosine methyl esters (D) on a pure water subphase at 25°C carried out at a compression rate of 7.1 b;2/molecule per minute. Arrows indicate the direction of compression and expansion.

Table 3 Stability results for the stearoylamino acid esters on pure water subphases at 25°C using Cadenhead's criterion for stability. Surfactant

Enantiomeric monolayef

Racemic monolayer

Stearoylserine methyl ester Stearoylalanine methyl ester Stearoyltryptophan methyl ester Stearoyltyrosine methyl ester

Unstableb Unstable Stable to 27' Stable to 4'

Stable to 2.5' Unstable Stable to 25' Unstable

'Compression and expansion rates of 7.1 A2/molecule per minute. bGreater than 1 dyn cm-' min-' deviation in surface pressure at low surface pressure. 'Values in dyn cm-'.

the criteria given previously. None of these films is stable at pressures exceeding 5 dyn cm-'. If the collapse of the monolayer, or its dissolution into the aqueous subphase, occurs at a rate greater than or equal to the rate of compression, the hysteresis in the l l / A isotherms may be readily explained.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

81

The instability of these chiral monolayers may be a reflection of the relative stabilities of their bulk crystalline forms. When deposited on a clean water surface at 25"C, neither the racemic nor enantiomeric crystals of the tryptophan, tyrosine, or alanine methyl ester surfactants generate a detectable surface pressure, indicating that the most energetically favorable situation for the interfacial/crystal system is one in which the internal energy of the bulk crystal is lower than that of the film at the air-water interface. Only the racemic form of N-stearoylserine methyl ester has a detectable equilibrium spreading pressure (2.6 0.3 dyn cm- I). Conversely, neither of its enantiomeric forms will spread spontaneously from the crystal at this temperature. Despite the clear evidence of stereodifferentiation exhibited in the II/A isotherms of these chiral surfactants, the instabilities of the films as spread from solution at temperatures of experimental feasibility prevent a thorough description of the factors that might lead to molecular recognition in monolayers at equilibrium with their environments. Our next line of approach to this problem has been to conduct a broad investigation of the most attractive candidate from this group. N-Stearoylserine methyl ester. We chose N-stearoylserine methyl ester (SSME) as the molecule which would most probably demonstrate the highest degree of molecular recognition in monolayer properties at temperatures between 20 and 100°C. This choice was based on several factors: (i) the monolayers of the racemic form are stable up to approximately 2.5 dyn cm-' when spread from hexane solution at 25°C while the films spread from either of the enantiomers are unstable; (ii) the racemic form spreads spontaneously from the crystal at the same temperature (ESP = 2.6 If: 0.3 dyn cm-') while the enantiomeric forms do not; and (iii) the differences in the I I / A isotherms of the enantiomeric versus racemic films at 25°C are very large (Fig. 16a). These clear differences in the film properties as spread from solution and as spread under equilibrium conditions from the crystal are evidence of molecular recognition and discrimination in highly ordered film and bulk crystalline states. Our experimental strategy was therefore designed to elucidate the state or phase in which clearly stereo-dependent physical properties could be detected. We began this approach by determining some of the fundamental properties of the bulk crystals. Associations within the bulk crystalline phase. The physical property of enantiomeric solids and their mixtures which is cited most often is melting point. Figure 18 gives the melting point versus composition diagram for mixtures of S ( + )- and R( - )-SSME. The solid-liquid coexistence line of

P. L. ROSE €T AL.

82

P

P 0

al

40

8

m

100

%RC)smm

Fig. 18 Melting point versus composition diagram for stearoylserine methyl ester crystals. Reprinted with permission from Harvey et al., 1989. Copyright 1989 American Chemical Society.

Table 4 Heats and entropies of fusion for crystalline stearoylserine methyl ester.'.b Sample ( S ) - ( + ) - or ( R ) - (-)-stearoylserine methyl ester ( R , S ) - ( )-stearoylserine methyl ester

Transition temp./K

AH:/ kcal mol-'

362.1

23.0 f 1.3

-63.5

367.2

16.4 f 1.5

-45.0 f 4.1

ASP/

cal K - ' mol-' 4.0

ASf = AH,/'&. Reprinted with permission from Harvey et a[( 1989).Copyright 1989 American Chemical Society.

this diagram is highly suggestive of a racemic compound or racemate in which the smallest unit consists of a heterochiral pair of molecules (Eliel, 1962).The onset of eutectic formation at the extreme ends of the composition diagram indicates th?? the racemate is stable (Jacques et al., 1981; Petterson, 1956). The strength of the crystalline interactions is also shown in the solid-liquid transition thermodynamics. The thermal solid-liquid phase transition enthalpies and entropies for enantiomeric and racemic forms of SSME as obtained by differential scanning calorimetry indicate significant differences in the energetics of intermolecular interaction during melting (Table 4),with the racemic crystals having a lower enthalpy of fusion despite a higher melting point. In addition, the infrared spectra of these materials indicated significant differences in packing in the homochiral versus heterochiral crystals. The

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

83

crystals of the racemate display a pronounced broadening in the -CH, -CH, bending bands between enantiomeric and racemic crystals, indicating differences in packing about the chiral headgroups in both forms. Repeated attempts to grow crystals of sufficient quality for X-ray analysis failed, however. Given the information above, the question remains as to the nature of the monolayer states responsible for the stereo-differentiation of surface properties in racemic and enantiomeric films. Although associations in the crystalline phases are clearly differentiated by stereochemical packing, and therefore reflected in the thermodynamic and physical properties of the crystals, there is no indication that the same differentiations occur in a highly ordered, two-dimensional array of molecules on a water surface. However, it will be seen below (pp. 107-127) that conformational forces that are readily apparent in X-ray and molecular models for several diastereomeric surfactants provide a solid basis for interpreting their monolayer behavior. Associations within the monolayer jilm. The difficulties in comparing the properties of crystals and monolayers spread from them, as described above, result from ambiguity regarding molecular associations in a dry, tightly packed, crystalline environment in contrast to those in a more loosely packed, wet monolayer spread on an aqueous subphase. The ideal interpretation of the molecular associations responsible for chiral recognition and the monolayer phase in which they occur would have to account for the orientation of chiral functionality at the air-water interface, the disruption and depth of the water surface, the hydration state of the chiral headgroup at every I I / A state of the film, and the dependence of these associations on variables such as temperature and humidity. Until all these structural and energetic factors can be accounted for, the best line of approach is to measure the mechanical, rheological, and thermodynamic properties of the continuum as represented by the monolayer. The first line of comparison between associations within a monolayer and those in a crystal can be obtained from equilibria between crystal and film. Table 5 gives the equilibrium spreading pressures and surface free energies, enthalpies, and entropies of spreading for racemic and enantiomeric films of SSME spread from their crystals in the 20-40°C temperature range. It is clear that the ESPs of these films are dependent upon chirality. In each case, the free energy of spreading is small and negative, with spreading of the racemic form being more spontaneous than that of the enantiomers. Within the propagation of experimental error, the heats and enhopies of spreading for these films are identical. Figure 19 gives the I I / A isotherms of the racemic and enantiomeric films of SSME at several different temperatures. At the lower temperature limit

Table 5 Equilibrium spreading pressures of SSME and surface free energies, enthalpies, and entropies of spreading for the resulting film". n'/dyn cm-'

Ae/A2/molecule

AGJkcal mol- '

ASJcal K - ' mol-'

AH:/kcal mol-

'

~~

293 298 303 313

2.5 f 0.3 4.2 f 0.3 11.3 f 1.5

0.5 f 0.2 5.7 f 0.7

58 f 3 54 If: 2 49 2

64 f 3 54 f 4

-0.21 f 0.03 -0.33 f 0.03 -0.77 0.11

47 f 4 -0.05 f 0.02 -0.44 f 0.06

"Reprintedwith permission from Harvey et al., 1989. Copyright 1989 American Chemical Society.

44 f 4

14.3

+ 1.2

13.3 f 1.2

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

85

of 20°C, there appears to be little difference in the isotherms for either enantiomeric or racemic films. However, at higher temperatures the shapes of the compression/expansion isotherms are clearly stereo-differentiated, with the racemic films being more highly expanded. It should be noted that the shapes of the I'I/A isotherms indicate clearly that a phase transition, possibly first-order, occurs during the compression of both racemic and enantiomeric films at temperatures above 20°C. This transition occurs at lower surface

I "

a) 20"c

1 0 ° 10 L

'i

b, 25"c

'L 0

10

Fig. 19 Surface pressure/area isotherms for the compressioq/expansion cycle of enantiomeric (dashed line) and racemic (solid line) SSME monolayers on pure water subphase at (a) 20"C, (b) 25"C, (c) 30°C and (d)40°C. The compression rate is 29.8 A'/ molecule/min. Reprinted with permission from Harvey et al., 1989. Copyright 1989 American Chemical Society.

86

P. L. ROSE ET AL

i

c)

30°C

Fig. 19 Continued.

pressures for the enantiomeric films. There appears to be the same general trend noted in the ESPs of films spread directly from the crystalline state, for which equilibrium is reached at lower surface pressures in the homochiral systems. The general, qualitative conclusion that may be reached by comparison of these data is that upon compression, the enantiomeric films may undergo transition to a tightly packed, solid-like state at lower surface pressures than do the racemic monolayers. It is interesting to note that when varying ratios of S ( +)- and R( - )-enantiomem are spread as monolayer films, their lift-off areas show a compositional dependence with temperature from 20 to 30°C (Fig. 20).

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

a

1

a7

T I 30-40 "C

%

R-(-)-SSME

Fig. 20 "Liftoff area versus composition diagram for SSME films on pure water subphase at 20,25,30, and 40°C. Reprinted with permission from Harvey et al., 1989. Copyright 1989 American Chemical Society.

+

At 25"C, chiral differentiation occurs as a function of R ( - ) / S ( ) ratio. At this temperature the stereo-differentiation in film l l / A isotherm properties is most completely expressed. Below this temperature, the racemic and enantiomeric films are nearly identical in compression/expansion properties, while at the higher temperature extremes the degree of expansion of both homochiral and heterochiral films are identical at lower surface pressures. Although the general shape of the lift-off area versus composition diagram at 25°C is similar to the melting point diagram of the crystalline surfactant mixtures, it must not be taken as a definite indication of surface racemic compound formation. It must be remembered that the enantiomeric and racemic films at this temperature have significantly different stability limits and equilibrium spreading pressures. We believe that the compositional dependence of the lift-off area is an indication of the onset of the formation of a more highly expanded, liquid-like phase induced by the mixing of the enantiomers. The stability limits of these monolayers are given in Table 6 , and demonstrate the fact that the more highly expanded racemic film is also the more stable system over the temperature range studied. However, the general trend that can be observed in the l I / A isotherms and the monolayer stability limits is that, as the enantiomeric films become increasingly stable, their isotherms begin to take on the characteristics of the racemic system. Coupled with the observation of the temperature and compositional dependence of the

88

P. L. ROSE ET AL.

Table 6 SSME monolayer stability limits, llL. n,/dyn cm-' TI K

Racemic

Enantiomeric

293 298 303 313

Unstable at all II 2.5 15 -21

Unstable at all ll Unstable at all ll 0.5 19

Table 7

-

SSME surface shear viscosity."sb

ll = 2.5 dyn cm-'

TIK 293, 298 303 308 313

ll = 5.0dyncm-'

Racemic

Enantiomeric

Racemic

Enantiomeric

C

C

C

C

0.553 f 0.026 0.472 f 0.026 0.419 f 0.047

2.00 f 1.11 0.504 f 0.038 0.393 f 0.036

0.573 f 0.42 0.535 f 0.040 0.507 f 0.039

No flow 0.666 0.109d 0.493 f 0.020

"Reprinted with permission from Harvey et al. (1989). Copyright 1989 American Chemical Society. 'Surface viscosity in millisurface poise. Condensed films, no surface flow. Measurable non-Newtonian flow.

lift-off area, this may be taken as direct evidence of the thermodynamic dependence of the recognition phenomena, and suggests that the enantiomeric discrimination exhibited is a function of the phase (i.e., solid-like versus liquid-like) of the film. The latter point is illustrated by the surface shear viscosities of the homochiral and heterochiral films at surface pressures below the monolayer stability limits. Table 7 gives the surface shear viscosities at surface pressures of 2.5 and 5 dyn cm- in the temperature range given in Fig. 19 (20-40°C). Neither enantiomeric nor racemic films flow under these conditions at the lower temperature extreme, while at 30°C the racemic system is the more fluid, Newtonian film. However, in the 35-40°C temperature range, the racemic and enantiomeric film systems are both Newtonian in flow, and have surface shear viscosities that are independent of stereochemistry. These results are not surprising when one considers that (i) when the monolayer stability limit is below the surface pressure at which shear viscosity is measured, the film system does not flow, or flows in a non-Newtonian manner; (ii) when the monolayer stability limit is above the surface pressure

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

89

at which shear viscosity is measured, the flow is Newtonian and independent of stereochemistry; and (iii) when the shear viscosity of the film is stereochemically dependent, the stability limit of the enantiomeric system is below the surface pressure at which the shear viscosity is measured. Taken together, the equilibrium spreading pressures of films spread from the bulk surfactant, the dynamic properties of the films spread from solution, the shape of the l l / A isotherms, the monolayer stability limits, and the dependence of all these properties on temperature indicate that the primary mechanism for enantiomeric discrimination in monolayers of SSME is the onset of a highly condensed phase during compression of the films. This condensed phase transition occurs at lower surface pressures for the R( - )- or S ( + )-films than for their racemic mixture. Although the quasi-thermodynamic and rheological data show the effects of stereochemistry on film properties, ultramicroscopy of the film in situ at the water surface provides a glimpse at the effect of chiral recognition on film packing as the monolayer is compressed. Figure 21 shows the epifluorescence micrographs of racemic and enantiomeric films of SSME at surface pressures below 10 dyn cm- at 25°C. The racemic films show a high degree of ordering, forming swirling clusters akin to those formed by chiral phospholipids under similar conditions (McConnell et al., 1984, 1986; Weis and McConnell, 1984; Keller et al., 1986; McConnell and Moy, 1988). There appears to be no set packing pattern. Conversely, the enantiomeric films show only a highly condensed film in which no ordering can be observed. Along with scanning tunneling and transmission electron microscopy of these films transferred to carbon substrates (Harvey et al., 1989), the message of all the data on the SSME system is that a tightly packed crystalline or quasi-crystalline surface state is necessary for detection of enantiomeric discrimination by our methodology. N-Stearoyltyrosine. The case of N-stearoylserine methyl ester illustrates temperature-dependent enantiomeric discrimination in both monolayers spread from solution and in equilibrium with the bulk phase. Although the n / A isotherms suggested large differences in the intermolecular associations in homochiral and heterochiral films of SSME, there exist chiral systems in which enantiomeric discrimination as exhibited in .film compression properties is much more subtle. N-Stearoyltyrosine (STy) is such a system. When spread from a benzene/hexane solution on to a slightly acidic water subphase, spread films of racemic and enantiomeric STy exhibit nearly the same lT/A isotherms (Fig. 22) and surface shear viscosities (Harvey et al., 1990). The shapes of these isotherms and the apparently small differences between the compression/expansion characteristics of these fluid homochiral and heterochiral monolayers is conserved throughout the

90

P. L. ROSE ET AL.

Fig. 21 In situ epifluorescencemicrographs of (a) “fluid” racemic and ( b )“crystalline” enantiomeric SSME monolayers at the air-water interface at 25°C. Lighter domains are fluorescing probe L-NBD PC; darker domains are SSME. Total magnification is - 5 0 0 0 ~ . Reprinted with permission from Harvey et al., 1989. Copyright 1989 American Chemical Society.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

91

Area Az/molecule

Fig. 22 Surface pressure/area isotherms for the compression cycles of stearoyltyrosine on a buffered pH 6.86 subphase carried out at a compression rate of 19.24 A2/molecule per minute at 16, 19,22,25,28,31, and 34°C. Reprinted with permission from Harvey et al., 1990. Copyright 1990 American Chemical Society.

T e m p e r a w "C

+

Fig. 23 Equilibrium spreading pressures of (R,S)-(& ) - and ( R ) - ( )-stearoyltyrosine on an aqueous subphase of pH 6.86 (potassium phosphate/disodium phosphate buffer) as a function of temperature. Film type I1 is the film at. temperatures above the transition and film type I is the film at temperatures below the transition. Reprinted with permission from Arnett et al., 1990.Copyright 1990American Chemical Society.

15-35°C temperature range. However, when spread from their bulk crystalline phases, the equilibrium spreading pressures of these films are clearly dependent on stereochemistry (Fig. 23) across the same temperature range. The conclusion that can be reached from these preliminary data is

92

P. L. ROSE ET AL .

that the presence of the bulk surfactant phase is necessary for detectable expression of enantiomeric discrimination in films of N-stearoyltyrosine. If we assume that the data of Figs. 22 and 23 can be treated by equilibrium thermodynamics, the discontinuities in the ESP versus temperature phase diagram should indicate the presence of a three-way equilibrium between bulk surfactant and two different film types in both homo- and hetero-chiral systems. The surface heats of transition ( U )between the two film types in either system may be obtained by relation (15), where lIe is the equilibrium

spreading pressure, A' is the average molecular area in the monolayer at the equilibrium spreading pressure as obtained from the l l / A isotherm of the spread film, and T' is the temperature at the transition "kink" in the ESP versus temperature diagram in Fig. 23. The subscripts I and I1 refer to the film states that exist in equilibrum at the transition temperature. This simultaneous approach eliminates the internal energy of the bulk surfactant in water, solving for the surface thermodynamic properties of film transition only. Similar transformations may be used to determine the surface free energies and entropies of spreading. Table 8 gives the results of this thermodynamic analysis for the spreading of film types I and I1 from the bulk, and the direct transition from film types I and 11. It is obvious that the Helmholtz free energies, entropies, and enthalpies are differentiated stereochemically. We believe that the chiral discrimination reflected by the transition thermodynamics represents the energetics of molecular recognition operating during an equilibrium transition to the more highly ordered crystalline state. Considering the very small differencesin the l l / A isotherms of the monolayers of racemic and enantiomeric STy as spread from solution, the equilibrium spreading pressure and the thermodynamic quantities derived from them strongly suggest that enantiomeric discrimination in this film system is heavily dependent on the presence of the crystalline bulk phase of the surfactant at the surface. The point of similarity between this system and the Nstearoylserine methyl ester system is the necessity of a closely packed state for enantiomeric recognition to be reflected by monolayer physical properties. Mixed chiral monolayers: special systems diferentiated only by symmetry

Given the strong expression of enantiomeric discrimination in monolayers of N-stearoylserine methyl ester, the question arises as to whether this discrimination can be maintained in the presence of other, achiral surfactants

Table 8 Helmholtz free energy, entropy, and internal energy of spreading and of transition for N-stearoyltyrosine on an aqueous subphase of pH = 6.86 at the transition temperature for each film."

bulk -+ film Ib bulk -+ film II' film I + film I1

- 1.40 f 0.07 - 1.50 f 0.07

0.10 f 0.09

-0.22 f 0.04 -0.24 f 0.05 0.02 f 0.01

38 f 6 126 f 7 88 f 9

0 40 6 40 f 6

a Reprinted with permission from Arnett et al., 1990. Copyright 1590 American Chemical Society. bFilm I: T < '7;. 'Film 11: T > 7;.

10.10 f 1.80 36.71 f 2.20 26.26 f 2.80

-0.23 f 0.04 11.93 f 1.80 12.10 f 1.80

P. L. ROSE € T A L .

94

where stereochemical interactions in the film would be -“diluted (Harvey and Arnett, 1989). Accordingly, dilute spreading solutions of SSME were mixed with solutions of simple fatty acids such as stearic or palmitic acid. Figure 24 shows the II/A isotherms of mixed monolayers of palmitic acid and racemic and enantiomeric SSME as a function of composition. At compositions of 16.7% and 33.3% palmitic acid, there is no difference between the compression/expansion isotherms of the racemic and enantiomeric systems, and the films have the I I / A characteristics of a pure palmitic acid film. However, at compositions of 50% SSME and above, the l l / A isotherms of the homochiral and heterochiral systems are significantly different. It is also at the 50% SSME mark that the monolayer stability limits of the mixed films begin to diverge according to stereochemistry (Table 9), with the enantiomeric systems being less stable. The difference in the l l / A properties of these mixed chiral/achiral systems was also observed in the films’ dynamic properties. Figure 25 gives the surface shear viscosities of the palmitic acid/SSME systems at surface pressures of 2.5 and 5.0 dyn cm-’ at 25°C. It is clear that stereo-dependence of film flow

A ’/ molecule Fig. 24 Surface pressure/area isotherms for palmitic acid/stearoylserine methyl ester films at 25°C on a pure water subphase and compressed at 29.8 A2/molecules per minute. A, 16.7-33.3%; B, 50%; C, 66.6%; D, 83.3% SSME.

95

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

Table 9 Monolayer stability limits of palmitic acid/stearoylserine methyl ester films at 25°C. Stability limit/dyn cm Monolayer composition (palmitic acid/SSME)

( R ) - (- 1 or ( S ) - ( + 1

( R , S ) - (f 1

----

--20.0 --15.0 15.0 2.0 1.0 Unstable at all ll

20.0 15.0 15.0

100% palmitic acid 5/1 (16.7% SSME) 2/1 (33.3% SSME) l / l (50.00/, SSME) 1/2 (66.6% SSME) 1/5 (83.3% SSME) 100% SSME

~

10.0 8.5 6.0 2.5

Unstable at all ll

I

25-

I I

20-

I

15-

0

I

I

/

a

D

4

o

m

a

0

1

a

l

%ssME

0

2

0

4

0

m

m

1

a

O

%ssME

Fig. 25 Surface shear viscosity us. film composition for the palmitic acid/stearoylserine methyl ester film system at 25°C.

96

P. L. ROSE E T A L .

properties occurs at compositions greater than 50% SSME, with the enantiomeric systems being much more resistant to flow than the racemic. As in the case of the pure SSME films, the lower stabilities and greater viscosities of the enantiomeric films indicate a higher degree of aggregation than in the more stable, fluid racemic monolayers. The question remains as to whether the mechanism of enantiomeric recognition is a collapse back to a surface quasi-crystalline state, even in the presence of the achiral diluent. Enantiomeric discrimination and its relation to film component reorganization upon compression can also be observed in dynamic surface tension hysteresis loops. Figure 26 shows the l l / A isotherms generated upon five successive compression/expansion cycles (from ll = 0 to 10 dyn cm- ) of racemic and enantiomeric films containing 17 mole percent palmitic acid. The hysteresis loops, obtained on the apparatus described in Section 2 (p. 63), show that the first compression/expansion cycle of the racemic system is repeated in each successive cycle. Upon expansion of the film from the maximum surface pressure back to 0 dyn cm-', the racemic film returns to its original state without detectable reorganization of the components. However, the

A 1.09 midcycle 75.2 AZ/molecule/min

. x

W

E

30

m

al

A2tm01em1.e

B

-

10

1.09 midcycle 7 1.6 A2/molecule/min

I

i s

x n

2

0

25

Az/molecule Fig. 26 Hysteresis isotherms for the 1/5 palmitic acid/(A) racemic and ( B ) enantiomeric stearoylserine methyl ester ( 17% palmitic acid) monolayer system at 25°C. Arrows indicate direction of expansion and compression. Reprinted with permission from Arnett et al., 1989. Copyright 1989 American Chemical Society.

97

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

enantiomeric system apparently reorganizes significantly from its original expanded state, as indicated by the large hysteresis loop, to a more condensed state as exemplified by the subsequent, smaller hysteresis loops. This observation is consistent with the lower monolayer stability limits of the homochiral films, their higher surface shear viscosities, and their lower ESP values. In order to test the mechanism of recognition, equilibrium spreading pressures of both racemic and enantiomeric forms of SSME were obtained in pre-spread films of palmitic acid/SSME mixtures. The films were spread from solution and then compressed to their lift-off areas. A crystal of the racemic SSME was placed on surface film mixtures of the fatty acid with racemic SSME, and the enantiomeric crystals were placed on surface film mixtures of the fatty acid and enantiomeric SSME. The results of the equilibrations are given in Fig. 27. As on pure water substrates, the enantiomeric crystals of SSME did not spread on the enantiomeric SSME/palmitic acid monolayer-covered surfaces, while the spreading of the racemic crystals on the racemic film-covered water was actually enhanced. The palmitic acid crystals deposited on either racemic or enantiomeric film covered substrates spread to the same surface pressure, independent of stereochemistry. When treated by the modified Gibbs phase rule (Crisp, 1949; Defay, 1932), these results suggest that at equilibrium, the enantiomeric monolayer system

- - -5~,,--4- _.

-

-p- 4- - 4

b) R,SW

i

I

?

8

3

a)

R O or S(+)

,+.- .-. - .1 ,

'

I

'

.-.I

'

I

-

_.+

-Q

' . '

98

P. L. ROSE E T A L .

consists of three phases: bulk crystalline SSME enantiomer, palmitic acid monolayer domains, and enantiomeric SSME monolayer domains. This situation appears to occur at every monolayer composition, and suggests that the considerably higher monolayer stability limits of the spread enantiomeric films reflect a metastability of the multicomponent film at surface pressures above the ESP. It is therefore reasonable to assume that the reorganization of the monolayer at surface pressures above the stability limit is most probably due to a squeezing out of the less stable enantiomeric SSME component. Conversely, the racemic film system appears to be “solubilized” by the achiral fatty acid component. At compositions of 10-33% palmitic acid, the ESP of the racemic system varies linearly with film composition, indicating that the monolayer in equilibrium with the racemic crystal is a homogeneous mixture of racemic SSME and palmitic acid. At compositions of less than 33% palmitic acid, the ESP is constant, indicating that three phases consisting of palmitic acid monolayer domains, racemic SSME monolayer domains, and racemic SSME crystals exist in equilibrium at the surface. This last result is consonant with the observation that enantiomeric discrimination as reflected by the l l / A isotherms is not observed until the composition reaches greater than 33% palmitic acid. On statistical average, the 33% achiral fatty acid mark is the lowest palmitic acid composition at which each chiral headgroup in the mixed film can be surrounded completely by six neighboring achiral carboxylatc headgroups, effectively blocking the communication of shape-specific packing from one chiral center to another. It is therefore not surprising that there is a first-order transition in the ESP versus composition diagram for the racemic film system, and that this occurs at a composition of 33% palmitic acid. All of the experiments in pure and mixed SSME systems, as well as in the N-stearoyltyrosine systems, have one common feature, which seems characteristic of chiral molecular recognition in enantiomeric systems and their mixtures: enantiomeric discrimination as reflected by monolayer dynamic and equilibrium properties has only been detected when either the racemic or enantiomeric systems have reverted to a tightly packed, presumably quasi-crystalline surface state. Thus far it has not been possible to detect clear enantiomeric discrimination in any fluid or gaseous monolayer state. Following the successful examination of stearoylserine methyl ester monolayers by a variety of techniques, a detailed study of the effects of several subtle changes in the headgroup has been performed by Heath (1991; Heath and Arnett, 1992). Figure 28 compares the structural variations as the hydroxyl group of SSME is replaced by a thiol in stearoylcysteine methyl ester (SCME);two SCME chains are linked through their sulfurs to produce

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

Stearoylserine Methyl Esters (SSME)

Y C O C 17Hz

NHCOCI7H3, I

c s

C R

H,COOC* ~'CH,OH H

HOH2Cf H'COOCH,

L-SSME

D-SSME

Stearoylcysteine Methyl Esters (SCME) NHCOCI7H,,

YCOC17H35

I

C S

C R H , C O O C ~H ~'CH,SH

HSH,C~~'COOCH,

LSCME

D-SCME

Dilauroylcystine Dimethyl Esters (DLCDME)

NHCOCI~HB I YCOCliH,, C R C R H,COOC':'CH,S SH,C'~'H H COOCH,

-

~COCiiHz,

yCOCilH23

cs

c s

H'~'CH,S-SH,C'~'COOCH, EOOCHR H

L-DLCDME

D-DLCDME

YCOCiiHz,

FcoC1lH23

C R

C S

H,COOC*~'CH,S-SH,C'~'COOCH, H

H

meso-DLCDME

Stearoylthreonine Methyl Esters (STME) H3COOC,y

NHCOC 17H,

H,,C~,OCHN,H/COOCH,

c'2

C

c 3

c

I

HO'~\CH,

LSTME (2S.3R) H3COOC,!,NHCOC,,H,

I C

H,C$' OH

Lallo-STME (2S,3S)

I

H,C'~\OH D-STME (2R,3S)

H,CI,0CHN,~/COOCH3 C I C HO'~'CH~

D-allo-STME (2R,3R)

Fig. 28 Structural comparisons of chiral surfactants.

99

100

P. L. ROSE € T A L .

diastereomeric ( + )- and ( - )-dilauroylcystine dimethyl esters (DLCDME) and a methyl group is attached to the hydroxyl-bearing carbon in SSME to produce the bulkier stearoylthreonine methyl esters (STME) and the ( + )- and ( - )-do-stearoylthreonine methyl esters [( + )- and ( - ) - d o STME]. Figures 19 (SSME), 29 (SCME), 30 (STME), and 31 (DLCDME) compare the effects of these changes, all of which are significant except that the normal and allo isomers of STME are indistinguishable. The figures speak for themselves in representing the effects of these changes in headgroup structure and temperature on the energetics of compression and phase changes with kinetic effects for relaxation in hysteresis. Melting point versus composition diagrams for the four sets of compounds showed racemate formation that was echoed in plots of the monolayer phase transitions points versus enantiomeric composition. Similar analysis of the lift-off areas versus composition for mixed monolayers of many combinations of the four systems showed great sensitivity to headgroup functionality and stereochemistry. Even the differences between combinations of normal and do-STME were defined clearly. Comparison of these many results indicates that the principal locus of stereoselectivity in the packing of these films lies in the stereoche-nical arrangements around the carbon between the ester and amide functions. The importance of the hydroxyl group of SSME, presumably through its interaction with the aqueous subphase, is clearly evident.

AZ/mOleeule

Fig. 29 Surface pressure/area isotherms for compression and expansion of enantiomeric (dashed line) and racemic (solid line) stearoylcysteine methyl esters (SCME) on an aqueous subphase at various temperatures. Arrows indicate the direction of compression and expansion.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

101

Fig. 30 Surface pressure/area isotherms for compression and expansion of enantiomeric (dashed line) and racemic (solid line) stearoylthreonine methyl esters (STME) on an aqueous subphase at various temperatures. Arrows indicate the direction of compression and expansion.

Fig. 31 Surface pressure/area isotherms for compression and expansion of (A) enantiomeric (dashed line) and racemic (solid line) lauroylcysteine methyl esters and (B) enantiomeric (dashed line) and meso (solid line) dilauroylcystine esters at 25°C. Arrows indicate the direction of compression and expansion.

102

P. L. ROSE H A L .

DIASTEREOMERIC SYSTEMS

Interactions between chemically distinct chiral species that are not enantiomers may be treated by a parity relationship analogous to that for enantiomers. Interactions between two different sets of antipodes, R or S and R' or S', are related by (16). These interacting pairs may be regarded as

homochiral (S-S', R - R ' ) or heterochiral (R-S', S-R') diastereomeric complexes. The equation holds for interactions between any dissimilar chiral molecules, whether they are diastereomeric or non-isomeric [or, in Craig's terminology (Craig and Mellor, 1975), chirodiastaltic] pairs. This rule is the basis for the resolution of enantiomeric mixtures by chiral resolving agents, as was first demonstrated by Pasteur's (1848) resolution of sodium ammonium tartrate with ( - )-cinchonhe. The energetics of chiral aggregation have been approached by calorimetry. Diastereomeric discrimination in the energetics of chiral ion aggregation has been demonstrated by Arnett and Zingg (1981). Heats of neutralization for ephedrine and pseudoephedrine with mandelic acid were obtained in dimethyl sulfoxide, dioxane and water. They showed clear differences in the energetics of ephedrinium mandelate and pseudoephedrine mandelate salt formation. The structures of these salts were confirmed by high-field nmr spectroscopy and X-ray crystallography. Calorimetric techniques related the structural differences to thermochemical terms (Zingg et al., 1988). Examples of diastereomeric complex formation between chiral solutes in achiral solvents abound, and have been reviewed elsewhere (Zingg, 1981). Diastereomeric complex formation is also used as the model for the stereoselectivity of chiral chromatography ( Pirkle and Pochapsky, 1987). Much less often examined are the interactions in neat chiral media into which another chiral species is introduced. In principle, interactions of the enantiomeric forms of an asymmetric molecule with an asymmetric environment or medium must be different (Craig and Mellor, 1975). If the differences in the physicochemical properties of the pure chiral components of the system are sufficiently large then chiral discrimination may be detected more easily than is the case in enantiomeric systems. If the components do not interact ideally then the properties of the mixture may reflect stereochemically dependent deviations from ideality. When the two components do not interact homogeneously, any chiral recognition effects must arise from interfacial interactions, and may therefore be more difficult to detect. Reports of differences in I I / A properties of monolayers cast from diastereomers are few. Stallberg-Stenhagen and Stenhagen ( 1951) reported

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

103

differences in the n/A properties of the diastereomers of 7-keto-acids of ( )-2(~ ) - 9 -L( or D)-dimethyltetracosanoic acid, although they questioned the purity of their materials. However, their experiments compared monolayer properties of individual diastereomeric surfactants with each other rather than mixtures of two different optically active surfactants. Lundquist, in her quasi-racemic’ monomolecular film experiments, demonstrated differences in the force-area characteristics for mixed films of S-( )-Ztetracosanyl acetate and methyl esters of R-( - )- or S-( + )-2-methylhexacosanoic acid (Lundquist, 1965). Malcolm ( 1973, 1975) reported possible observable differences between the pressure-area curves of poly( L-alanine) mixed with poly( D-a-amino-nbutyric acid) and the corresponding mixture containing poly( D-alanine). Shafer (1974) observed differences in the film pressure between ( R ) phosphatidyl serine with poly-L-lysine and the corresponding film with poly-D-lysine injected under the film. It is evident then that good precedent exists for detection of diastereomeric discrimination in monolayer films at the air/water interface. The method of quasi-racemates has been considered as a means of examining steric factors in living systems (Fredga, 1944). Lundquist recognized its potential for testing the arrangement of asymmetric molecules in monolayers in view of the loosely packed ordering which has been proposed for them. Our interest was aroused in the further application of the method of quasi-racemates to chiral monolayers because of its potential for establishing the configuration of very small quantities of biomolecules. The methyl esters of stearoylalanine [ 11 and stearoylserine [2] were considered as quasi-racemate candidates because of their slight structural differences. No quasi-racemate behavior was observed, however, in their force-area isotherms although clear diastereomeric discrimination was seen for this combination (Verbiar, 1983). We have seen no indication of quasi-racemate behavior for any other mixed chiral monolayers.

+

+

0

H3C0,C

I

II C17H35-C-N-C-CH3 I

H R-( -)-

H

H 3C-C-N-C-C

i 1

7H35

I

H

s-(+ )-

A quasi-racemate is a molecular compound that is related to a true racemic compound by a small structural change in one of the enantiomers (Fredga, 1944).

P. L. ROSE € T A L

104

0

II TOzCH3 C1,H3,-C-N-C-CHzOH I

H,CO,C

I

HOHzC-C-N-C-C17H35

I

H

H

C2I

R-( - )-

s-(+)-

Mixtures of N- (a-methylbenzyl) stearamide and difeerent esters of stearoylamino acids

When compressed to surface pressures greater than their stability limits (see Table lo), diastereomeric mixtures of N-(a-methylbenzy1)stearamides with both stearoylalanine and stearoylserine methyl esters provided clear evidence of chiral discrimination. Force-area isotherms at 35"C'for homochiral and heterochiral pairs of N - ( a-methylbenzy1)stearamide and stearoylalanine methyl ester show differences in both their lift-off and touchdown (the area per molecule where the surface pressure returns to zero on the expansion arm of the isotherm) areas per molecule (Fig. 32). In addition, monolayers of the heterochiral pair could be compressed to lower areas per molecule than monolayers of the homochiral pair. Heterochiral mixtures of N - ( a-methylbenzyl)stearamide and stearoylserine methyl ester show a collapse of the film at 20 &/molecule to some unstable state (Fig. 33). The homochiral mixture can be compressed to a higher

-

Table 10 Stability limits for diastereomeric monolayers cast from heterochiral (R-S', S-R) and homochiral ( S - S ' , R - R ) surfactant pairs. Diastereomeric monolayers"

+ +

Stearoylalanine methyl ester N - ( a-methylbenzy1)stearamide Stearoylalanine methyl ester stearoylserine methyl ester Stearoylserine methyl ester + N - ( wmethy1benzyl)stearamide Stearoyltyrosine methyl ester + N - ( CI -methylbenzyl)stearamide Stearoyltryptophan methyl ester N - ( a-methylbenzy1)stearamide

+

Combination

Stabilityb

Heterochiral Homochiral Heterochiral Homochiral Heterochiral Homochiral Heterochiral Homochiral Hzterochiral Homochiral

Unstable' Unstable Stable to 4 Unstable Stable to 15 Stable to 9 Stable to > I 0 Stable to >10 Stable to > 11 Stable to > 11

"Compression rates were 7.1 A2/molecule per minute. bValuesin dyn cm-I. 'Greater than one dyn cm-' min-' decrease in surface pressure even at low pressure.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

105

A-; a

0

10

P

m

W

--- - B’)

1

m

A%noMe

Fig. 32 Surface pressure/area isotherms for the compression/expansion cycles of diastereomeric monolayers of ( R or S ) - N - (a-methylbenzy1)stearamidesmixed 1 : 1 with (R’ or S’)-stearoylalanine methyl esters on a pure water subphase at 35°C. Dashed lines denote heterochiral pairs ( R :S’ or R ‘ : S )and solid lines denote homochiral pairs (R:R’ or S:S’).

Fig. 33 Surface pressure/area isotherms for the compression/expansion cycles of diastereomeric monolayers of ( R or S ) - N - (a-methylbenzy1)stearamidesmixed 1 : 1 with (R’ or S’) stearoylserine methyl esters on a pure water subphase at 35°C. Dashed lines denote heterochiral pairs (R:S’or R ’ : S )and solid lines denote homochiral pairs ( R : R ‘ or S:S’).

pressure than the corresponding heterochiral pair before collapsing. It must be noted, however, that the diastereomeric discrimination detected in these systems is not due to stereochemical differences in packing arising from interactions in a stable fluid film. Rather, detectable differences are a reflection of the stereochemistry as it is expressed in the reorganization and collapse

P. L. ROSE €T AL.

106

A

4)-

aE

B 9)-

E

; my

10-

0-

1

m

30

4)

8)

8)

m

1

m

Aymoleeule

Fig. 34 Surface pressure/area isotherms for the compression/expansion cycles of diastereomeric monolayers of N - ( a-methylbenzy1)stearamides mixed 1:1 with ( A ) stearoyltyrosine methyl esters and ( B ) stearoyltryptophan methyl esters on a pure water subphase at 35°C. All isotherms are for R:R' o r S:S' and R:S' or S:R' films.

of the films as they are compressed beyond their stability limits to some metastable or unstable state. Mixtures of N - ( a-methylbenzy1)stearamides with both stearoyltyrosine and stearoyltryptophan methyl esters show no discrimination in their pressure-area relationships at 35"C, regardless of the surface pressure to which the films are compressed (Fig. 34). The n/A curves for homo- and hetero-chiral pairs are exactly coincident. No discrimination in the pressure/area characteristics was seen for diastereomeric monolayer films spread from all possible mixtures of pure racemates ( R - and S - ) and their racemic mixtures ( R - ,S - ) of stearoylalanine, stearoylserine, stearoyltyrosine, and stearoyltryptophan methyl esters. The one exception was heterochiral and homochiral mixtures of stearoylalanine methyl esters and stearoylserine methyl esters at 35°C. The force/area

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

107

Fig. 35 Surface pressure/area isotherms for the compression/expansion cycles of diastereomeric monolayers of ( R or S)-stearoylalanine methyl esters mixed 1 : 1 with (R’ or S’) stearoylserine methyl esters on a pure water subphase at 35°C. Dashed lines denote heterochiral pairs (R:S’ or R ’ : S )and solid lines denote homochiral pairs ( R : R ’ or S:S’). Reprinted with permission from Arnett et al., 1989. Copyright 1989 American Chemical Society.

isotherms are given in Fig. 35. Again, no discrimination is noted until surface pressures are reached which exceed the stability limits of the film (greater than 1 dyn cm-’ min- decrease in surface pressure once the compression is stopped). The film spread from the homochiral pair is unstable at all ll, while that spread from the heterochiral pair is stable to 4 dyn cm-’.

’

Diazene-linked diacids Meso- and ( f )-azobis [6-( 6-cyanododecanoic acid)] were synthesized by Porter et al. (1983) as an amphipathic free radical initiator that could deliver the radical center to a bilayer structure controllably for the study of free radical processes in membranes. The decomposition pathways of the diazenes are illustrated in Fig. 36. When the initiator was decomposed in a DPPC multilamellar vesicle matrix, the diazenes showed stereo-retention yielding unprecedented diastereomeric excesses, as high as 70%, in the recombination of the radicals to form meso- and ( f )-succinodinitriles (Brittain et al., 1984). When the methyl esters of the diazene surfactants were decomposed in a chlorobenzene solution, poor diastereoselectivity was observed, diastereomeric excesses of 2.6% and 7.4% for meso- and ( f )-isomers respectively, which is typical of free radical processes in isotropic media (Greene et al., 1970). Porter et a f .(1986a) found that these surfactants also exhibited significantly enhanced diastereoselectivity when photolyzed alone in a pH 10 buffer

P. L. ROSE E T A L .

108

COOH

<

COOH

>

HOOC

R=H,CH,

Fig. 36 Structure of meso- and ( f)-azobis[6-(6-~yanododecanoicacid)] and their decomposition scheme.

solution. Studying the meso/( ) product ratio for the succinodinitrile recombination products, significant retention of stereochemistry was observed; ( f)-diaxenc resulted in greater ( f)-succinodinitrile product and, likewise, meso-diazene resulted in greater meso-product. At high dilution, well below their CMCs, stereochemical retention was preferred with a ratio of about 3.5:l. Above the CMC, ( f )-diazene exhibited a remarkable 6 : l retention of stereochemistry most likely as a result of the viscous, restricted environment of the micellar matrix. Counterintuitively, meso-diazene showed a decrease in diastereoselectivity, dropping from 3.5: 1 to 2: 1, upon micellization. We were interested in examining the corresponding monolayer photolysis products of the diazenes as a function of surface pressure. This goal was not reached due to our inability to collect enough material from the film balance experiment (typically, 1OI6 molecules per run) for HPLC analysis. However, the physical data acquired by interfacial techniques has given information about the nature of the aggregates and their diastereoselectivity. The meso and ( k )-azobis [6 4 6-cyanododecanoic acid)] provided strong

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

109

Fig. 37 Surface potential versus molecular area plot for monolayers of meso- and ( f )-azobisC6-(6-cyanododecanoic acid)] spread statically on a pH 3 subphase.

evidence of chiral discrimination in every property examined. It was predictable, though, that discrimination resulting from the intramolecular interactions of chiral centers within diastereomeric compounds would be far greater than the discrimination resulting from the looser intermolecular interaction of enantiomers. All of the monolayer data reflected clear and consistent discrimination between properties of the meso- and (+)-films. The meso-films exhibited lower surface potentials, indicating less orientation of their molecular dipoles

(Fig. 37) and lower viscosity, indicating less intermolecular association; they were also more expanded, indicating that their molecules had a greater tendency to spread themselves out over the interface in comparison to the ( f )-diazene (Fig. 38). In fact, the meso-films were so expanded that significant surface pressures were measured with more than 200 A2 available per molecule. Lift-off areas of this magnitude are typical of ionized monolayers which are highly expanded because of the electrostatic repulsions exerted between head groups. The insensitivity of the diazene films to the subphase acidity ruled out the possibility that the difference observed in the degree of expansion of the films was a result of a difference in pK,-values in the meso- and ( k )-monolayer. The higher degree of association of the ( f)-surfactant was also reflected in other aggregated forms. Micellization of the ( f)-diastereomer was more exergonic or, in other words, more spontaneous. Porter et al. (1986a) found correspondingly anomalous behavior in the photolysis of these compounds as presented above. While the physical data describing the aggregate forms are most probably related to the stereochemistry of the aggregates, we have no precedent for the following interpretation.

110

P. L. ROSE ET AL.

--_._ -.

*...

i

. ---_.._

0 J)

I

m

I

9D

I

120

I

160

I

180

I

210

Fig. 38 Surface pressure/area isotherms for the compression/expansion cycles of meso- (dashed line) and ( f)-(solid line) azobis- [6-(6-cyanododecanoic acid)] on a pH 3 subphase at 22°C. Compressed at a rate of 15.5 Az/molecule per minute. Reprinted with permission from Porter et al., 1986a. Copyright 1986 American Chemical Society.

The decomposition of the diazene methyl ester derivatives in chlorobenzene solution resulted in no retention of stereochemistry in the succinodinitrile products. Since under these conditions, the diazenes were probably monomeric rather than aggregated, the inherent electronic properties of the isolated molecules have been excluded as a possible source of the stereochemical retention. When the free acid derivatives of the diazenes were decomposed in chlorobenzene solution, retention was observed. In this non-polar solvent, the diazene diacids were probably aggregated as reversed micelles with their head groups linked by intra- and inter-molecular hydrogen bonding, decreasing the radical centers' opportunity to escape from one another and thereby resulting in more direct recombination of the radicals. In water, decomposition of the free acids at concentrations below the CMCs also led to stereochemical retention in the products. Stereochemical retention under these conditions was most probably due to hydrogen bonding between the diazene head groups as before, and hydrophobic association of the aliphatic chains. These factors would again reduce the mobility of the radicals. Up to this point, we have not had to address the unexpected photolysis behavior of the meso-compound, namely, that the stereochemical retention in the products resulting from the decomposition of this stereoisomer actually decreased at concentrations above the CMC. Intuitively, one would expect that the restricted environment created by micellization of the diazenes would enhance stereochemical retention. The (_+j-diazene behaves as we would have predicted, but the meso-isomer does not. To explain the behavior of the meso- and ( f)-diazenes upon micellization, the differences in the molecular structures and the aggregates they form must both be examined since either (or both) may be responsible for the anomaly.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

mesa

b

111

‘twisted” mesa C

Fig. 39 Possible conformations of ( & )- and meso-diazenes. Reprinted with permission from Porter et al., 1986a. Copyright 1986 American Chemical Society.

By placing both acid head groups on the same end of the molecule (a requirement for packing the surfactant into a monolayer, bilayer, or a micelle), the diazene moiety of the ( f )-surfactant is constrained to the “S-planar” conformation illustrated in Fig. 39( a). Incidentally, this conformation allows for the maximum anomeric interaction of the diazene lone pair electrons and the adjacent carbon-cyano bond. However, when the meso compound is arranged with both head groups proximately, the diazene moiety adopts the “U-planar” conformation as shown in Fig. 39( b). Alternatively, if the diazene moiety of the meso form remains fixed in the “S-planar” conformation to take advantage of the available anomeric interaction, disposing the head groups to the same end of the molecule would result in the “twisted” meso-conformation depicted in Fig. 39(c). Examining the monolayer data in terms of these models yields two possible, but unresolvable, explanations of the results (Porter et al., 1986a). In one scenario, the monolayer packing differences expressed by their degree of expansion are explained in terms of the symmetry of the diazene moiety. If, in the monolayer, the meso- and ( f )-diazenes adopt the “U-planar” and “S-planar” conformations, respectively, a projection of the ( )-structure parallel to the diacid chains shows nearly perfect C,, symmetry whereas the meso isomer’s projection is dissymmetric. The CZv symmetry of the (-t-)-diazene may allow for extended packing in the monolayer similar to that found in crystalline AIBN (Jaffe et al., 1972). The tight, crystal-like packing available to the ( f )-stereoisomer could result in highly condensed, highly ordered monolayers, which is consistent with the experimental data. By the same reasoning, the dissymmetry of the meso’s “U-planar” conformation would not allow for extended packing and would result in

112

P. L. ROSE ET AL.

more expanded, more disordered monolayers relative to those of the ( f )-isomer.

If, on the other hand, the meso-isomer maintains the anomeric interaction, thereby adopting the “S-planar” conformation, the head groups would be disposed on opposite ends of the molecule. Maintaining this conformation about the diazene fragment while constraining the head groups to the interfacial plane of the monolayer would force the molecule to adopt the “twisted” conformation as the area available per molecule is reduced. This conformation would result only if the gauche interactions created in the chain were energetically less costly than the loss of whatever anomeric interaction exists. The broad, irregular shape of the “twisted” conformer would also be consistent with the highly expanded, disordered films observed. In analyzing these two models, it is also important to note that the anomeric and gauche interactions may trade-off their optimal conformations, resulting in a structure that is some compromise of the two. With either model, though, the retention of stereochemistry may be the result of aggregation forces which reduce the mobility of the radical centers after formation thereby maintaining them in an optimum position to recouple directly after the expulsion of the nitrogen molecule. In the micellar matrix, the ( f)-molecules suffer a dramatic decrease in their mobility, leading to increased stereochemical retention in the photolysis products. Due to the geometric problems involved in packing the meso-molecules into micelles described by either model above, the meso-molecules are less apt to pack tightly and therefore more easily lose their stereochemistry. By these models, the decrease in stereochemistry of the meso-isomer upon micellization reflects a decrease in the aggregation forces experienced by them. Keto-linked diacids Recently, Porter et al. (1986b, 1988) have reported the synthesis of both meso- and (+)-forms of a series of two-chain carbonyl diacids made by joining two pentadecanoic acid units by a carbonyl group at the 3,3‘, 6,6, 9,9‘ and 12,12’ positions, 3,5-didodecyl-4-oxoheptanedioic acid (C- 15:3,3’), 6,8-dinonyl-7-oxotridecanedioicacid (C-15:6,6’), 9,lldihexyl-10-oxononadecanedioic acid (C-15:9,9‘) and 12,14-dipropyl-13oxopentacosanedioic acid ((2-15: 12,127, respectively. The diacids were used to probe further the question of stereochemical preference in two-chain amphiphiles. The method used for examining the diastereomeric preference was equilibration by base-catalyzed epimerization in homogeneous, bilayer and micellar media. This method allows for stereoselection based on hydrophobic/ hydrophilic considerations rather than classic steric size effects. When the diastereomeric diacids were epimerized in aqueous base at 6WC,

113

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

at concentrations above the critical micelle concentration, all of them showed a preference for the meso-form in accordance with predictions from molecular mechanics based on the model compound 2,4-dimethyl-3-pentanone (Cram, 1984). However, epimerization in homogeneous media (e.g., benzene solution, TsOH catalyst) or in water (hydroxide ion catalyst) below the CMC gave 50: 50 mixtures of meso- and ( )-diastereomers. It was thus concluded that hydrophobic forces perturb the equilibrium between diastereomers in favor of the meso-form. We report here a study of these eight diastereomeric compounds as monolayers at the air-water interface, a special set of circumstances representing an extreme of hydrophobic control in which intermolecular and intramolecular forces may be opposed directly. Figures 40-43 compare the position of the bridging carbonyl group in each diastereomeric pair and the effects of stereochemistry on their force-area curves at 25°C.It is clear that there is a large effect of stereochemistry on the energetics of compression and expansion for a wide variety of ketodiacid surfactants; all of the ketodiacids in this study showed a dependence of the shapes of their l 7 / A isotherms on their stereochemistry. Several facts are striking. In every case there is a sharp differentiation between the behavior of films cast from meso- and (*)-isomers. The isotherms for the

*

-

40

-

30

-

c-15:3,3’

I

E

=x

2

20 -

10 -

0 - ,

,

,

,

,

I

I

,

I

I

I

1

A * / molecule Fig. 40 Surface pressure/ area isotherms for the compression and expansion cycles of ( + ) - and rneso-C-15:3,3’ ketodiacids on a pure water subphase at 25°C carried

out at a compression rate of 19.24A2/molecule per minute. Arrows indicate the direction of compression and expansion. Reprinted with permission from Harvey et al., 1988. Copyright 1988 American Chemical Society.

P. L. ROSE ET AL.

114

40

HOOC HOOC

30

C-I5:6.6‘

E

0

h 20 e

F!

e

10

0 C1

u)

40

60

100

80

120

I

I

I

I

140

160

180

200

A*/moIecuIe

Fig. 41 Surface pressure/area isotherms for the compression and expansion cycles of ( * ) - and meso-C-15:6,6 ketodiacids on a pure water subphase at 25°C carried out at a compression rate of 19.24!IZ/molecule per minute. Arrows indicate the direction of compression and expansion. Reprinted with permission from Harvey et al., 1988. Copyright 1988 American Chemical Society.

1

HOOC HOOC-

0 C-15:9,9’

30

-

:

I

x

20-

P e

10 -

0 -0

20

40

60

80

100

120

140

160

180

200

A2/moIecuIe

Fig. 42 Surface pressure/area isotherms for the compression and expansion cycles of ( + ) - and meso-C-15:9,9’ ketodiacids on a pure water subphase at 25°C carried out at a compression rate of 19.24!12/molecule per minute. Arrows indicate the direction of compression and expansion. Reprinted with permission from Harvey et al., 1988. Copyright 1988 American Chemical Society.

115

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

-

40

-

30

1-

I

r

HOOC

t

-

35

20-

e

I0I

i 7

HOOC meso

C-15: 12,12‘

10 -

020

, 40

I

60

80

100

120

140 160 180 200

I

I

220

240

A*/mo~ecu~e

Fig. 43 Surface pressure/area isotherms for the compression and expansion cycles of ( f )- and meso-C-15:12,12’ ketodiacids on a pure water subphase at 25 “C carried out at a compression rate of 19.24 A/molecule per minute. Arrows indicate direction of compression and expansion. Reprinted with permission from Harvey et al., 1988. Copyright 1988 American Chemical Society.

meso-diastereomers are virtually identical at surface pressures greater than 5 dyn cm- regardless of the position of the carbonyl bridge or the length of the hydrocarbon chains. The limiting molecular area for the meso compounds is in the neighborhood of 60 A2/molecule, and their isotherms are highly reminiscent of the condensed nature of single-chain fatty acids. In contrast, films cast for the ( f )-diacids are more highly expanded films at lower pressures and have collapse pressures that are dependent on the position of the bridging carbonyl group. These results are quite compatible with the preferred conformations of the two-chain carbonyl diacids in aqueous media mentioned above (Porter et al., 1986b, 1988). The meso-compounds’ preferred collinear conformation, which places the hydrogens at the asymmetric carbons in a nearly eclipsed position relative to each other (Fig. 44), is more stable than that of the ( & )-diastereomer by about 1.2 kcal mol- In this conformation, the two carboxylic acid groups at the ends of the chain can be attached to the water surface side-by-side. The entire molecule can then behave as a “goodamphiphile” whose structure is similar to a pair of single-chain fatty acid molecules bound side-by-side, each chain mirroring the other about the molecules’ plane of symmetry.

’,

’.

116

P. L. ROSE E T A L .

-1

compression

HOOb

bOOH

preferred conformation

preferred conformation

HOOC- C

meso

It)

subphase

subphase

O

O

7

mesn

I f )

Fig. 44 Preferred lowest energy conformations of ( f )- and meso-diastereomersand the effect of stereochemistry on the mechanism of film compression. Reprinted with permission from Arnett et al., 1989. Copyright 1989 American Chemical Society.

There is little variation in the shape of the isotherms for a series of meso structural isomers (C-15:3,3’-C-15: 12,12’)upon moving the carbonyl group up the hydrocarbon chain. There is an expansion of the film, as would be expected if the head groups were allowed to drift apart at large areas ( > 240 A’) per molecule. A comparison of the r I / A isotherms for the series of structural isomers shows that the meso-diastereomers become more expanded at low surface pressures as the carbonyl group is moved up the hydrocarbon chain. In contrast to the preferred conformation of the meso-diastereomers, the ( & )-isomers are most stable if the carboxylate head groups are maximally separated with a rather wide range of pseudotorsional angles through which the hydrogens on the carbons carrying the bridging carbonyl group can be twisted until a minimum energy is reached when they are nearly eclipsed (Fig. 44). When one of the ( k )-diacids is spread on an aqueous surface, the carboxylate acid groups are separated from each other and the orientation of the molecule can be compared to a pair of open hedge-clippers resting on the water surface. Since the head groups are separated from one another ,along the axis of the molecule, it takes up more space on the surface and produces a more expanded monolayer, typical of “bolaform” amphiphiles (Furhop and Fritsch, 1986; Furhop et al., 1986; Jeffers and Daen, 1965; Morawetz and Kandaniem, 1966). The steady increase in the area per molecule at low surface pressures for the series of structural isomers correlates with the movement of the position of the carbonyl bridge away from the head groups. The differences in the energetics of compression can be related ultimately

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

117

9)-

I

5 -?

m-

10

-

0 -

I

0

a,

I

e

I 8)

I 8 )

I m

1

I 2

o

I u

o

I m

1

I 8

o

I

m

A?/mOleCUlt?

Fig. 45 Surface pressure/area isotherms for the compression cycle of 12ketooctadecanoic acid (A) and octadecanoic acid ( B ) on a buffered subphase ( A R hydrochloric acid; pH4.0) at 30°C carried out at a compression rate of -2.0-3.0 b;’/molecule per minute.

to the stereochemically dependent intramolecular conformation of the isomers at the air-water interface. During the compression cycle for rneso-isomers, work must be done on the system to raise the hydrocarbon portion of the molecule out of the water surface until the hydrocarbon chains are normal to the water surface. Since there is an absence of a plateau region or any apparent resistance to compression, other than raising the hydrocarbon tail out of the surface, in the I I / A isotherm for all meso-diastereomers, it can be inferred that the carbonyl linkage group has only a small interaction with the water subphase. This is in contrast to the I I / A isotherm shown in Fig. 45, where a carbonyl group on a monosubstituted single-chain fatty acid adds a large energetic contribution to the work of compression necessary for raising the hydrocarbon chain out of the water surface (Nagaranjan and Shah, 1981). During the compression cycle of the ( f)-isomers, the surface area is reduced and the carboxylate groups are forced together in the plane of the water interface. The chains are gradually forced together until they are side-by-side (Fig. 44), an arrangement similar to that proposed for the rneso-diacid under similar compression. However, for the ( f)-acids, in addition to the work necessary to raise the hydrocarbon chains from the surface, additional work must be done against intramolecular conformational forces resulting in chains which are collinear and normal to the water surface.

118

P. L. ROSE €T AL.

Table 11 Work ofcompression of the C-15:3,3', C-15:6,6, C-15:9,9', and C-15: 1 2 3 ' ketodiacids on a pure water subphase at 25°C compressed at a rate of 19.24 A2/molecule per minute. Work of compression"/cal mol-' Monolayer

C-15: 3,3' C-l5:6,6' C-15:9,9' C-15: 12,12'

n / d y n cm-'

15.0 15.0 15.0 15.0

meso

78.89 f 3.41 157.83 3.90 245.17 f 7.50 460.94 f 16.48

(f) 166.95 f 5.65 333.80 f 21.00 857.64 f 21.40 1493.79 f 67.94

( f)-meso

88.06 f 6.60 175.97 f 21.36 612.47 f 22.68 1032.85 f 69.91

'Work of compression calculated from the equation W=

:j n

dA,

where A , and A, are the initial and final areas per molecule, respectively.

This additional work of compression has been calculated from the differences in the I I / A isotherms to be 15dyncm-' (see Table 11). The further apart the carboxylate groups are to begin with, the greater is this n/A work and the greater the integrated area under the curve. It must be kept in mind that the ( f )-diastereomer is a mixture of enantiomers that were not separable by reverse phase HPLC. Since the method for separation of ( f)- and meso-diastereomers (reverse phase HPLC, C- 18 column) is based on the difference in conformation between isomers (related ultimately to their differential attraction to the column), it is unlikely that the two enantiomers of the ( f)-diastereomer have different conformations at the air-water interface. It is not known, however, how the energetics of compression and expansion will differ for films cast from either R,R or S,S enantiomer from those cast from the racemic mixture. Table 12 shows the equilibrium spreading pressures of each diacid. It is immediately apparent that for three of the diastereomeric pairs there are statistically significant differences. These distinctions relate stereochemical preferences in the spontaneous spreading of ( f )- versus meso-monolayers in equilibrium with their respective crystalline phases. However, there appears to be no discernible trend in either the ( f)- or meso-ESPs as a function of carbonyl position despite clear trends seen in their monolayer properties in the absence of any bulk crystalline phase. It is readily apparent that the (f)-C-15:6,6, C-15:9,9' and C-15:12,12' compounds show a collapse to some three-dimensional state, as has been observed for most over-compressed lipid films (Handa and Nagaki, 1979; Stewart and Arnett, 1982). However, the ensuing plateau region is

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

119

Table 12 Equilibrium spreading pressures of the C-15:3,3’, C-15:6,6’, C-15:9,9’, and C-15: 12,12’ ketodiacids on a pure water subphase at 25°C.

Monolayer

meso

C-15: 3,3‘ C-15:6,6’ C- 15 9,9’ C-15:12,12

15.48 f 0.84 17.90 0.84 16.75 f 1.43 3.45 f 0.43

(f) 0.97 _+ 1.36 12.93 f 0.76 20.73 f 3.82 10.45 f 0.56

extraordinarily stable. No change in surface pressure is seen even when compression is halted for several hours. In accordance with the twodimensional phase rule of Defay (1932) and Crisp (1949), this indicates that there are two phases in equilibrium. The most likely composition is one of monolayer and three-dimensional aggregates. Compression in the plateau region below the limiting molecular area of the collinear molecule does not result in increased surface pressure. The exact nature of the collapsed phases in this plateau region is not clear.

Surface viscosities. In view of the dramatic expression of the stereochemically directed conformation on rI/A curves, it seems reasonable to ask if corresponding differences in surface viscosities may be found. Variations in surface pressure as a function of molecular area must be a result of confounding intramolecular conformation and intermolecular interactions between molecules as they lie in the surface and are gradually compressed. Rheological properties such as the surface viscosity might reasonably show considerable differences depending on the stereochemistry of the diacid and the surface pressure. Table 13 presents all the surface shear viscosity data that could be measured. Several trends may be perceived in the data. As the surface pressure is increased, the surface viscosities increase. This is not surprising owing to increased intermolecular chain-chain contact at higher surface pressures. There are differences between meso- and ( f )-isomers for all measurable pairs of diastereomers, albeit small for 6,6- and 9,9‘diastereomers and negligible for 3,3’-diastereomers. However, if the precise thermodynamic state of the film is considered, i.e., surface pressure/area relationships, then significant differences in the free energies of activation to flow are seen (Table 13). Free energies of activation to flow increase as the carbonyl group is moved up the hydrocarbon chain. This may relate to the increased ability for intercalation of the molecules as they sit at the water surface. We saw previously that movement of the carbonyl linkage up the hydrocarbon chain

120

P. L. ROSE H A L .

Table 13 Rheological properties of diastereomeric films at various surface pressures n.'

Surface shear viscosity qs/ milli surface poise Monolayer

C-15: 3,3'

C- 15: 6,6'

C- 15:9,9'

C-15: 12,12'

ll/dyn cm-'

2.5 5.0 7.5 10.0 12.5 15.0 2.5 5.0 7.5 10.0 12.5 15.0 2.5 5.0 7.5 10.0 12.5 15.0 2.5 5.0 7.5 10.0 12.5 15.0

meso

0.70 f 0.06 0.83 f 0.08 0.78 f 0.03 0.80 f 0.04 0.78 f 0.03

(f) 0.66 f 0.03 0.84 f 0.04 0.83 f 0.03 0.79 0.05 0.80 f 0.05

C

C

0.69 f 0.03 0.68 f 0.02 0.70 f 0.02 0.82 f 0.01

0.70 0.03 0.75 f 0.03 0.88 f 0.03 0.93 f 0.03

d

d

d

d

1.19 f 0.13 1.06 f 0.11 0.93 f 0.08 0.85 0.02 0.90 f 0.03 0.91 f 0.01

0.96 f 0.06 0.92 f 0.03 0.97 f 0.02 1.00 f 0.03 1.13 f 0.10 1.18 f 0.02 1.07 f 0.11 0.70 f 0.03 0.86 f 0.07 0.95 f 0.04 1.09 f 0.05 1.29 f 0.05

b b b b b b

AAGi,,,b/cal mol ( f)-meso 208 f 74 243 f 76 254 f 44 199 f 59 210 f 55 181 f 48 182 f 39 238 f 41 136 f 38 361 f 87 322 f 74 402 f 32 346 f 57 354 f 64 271 f 39

Reprinted with permission from Harvey et al., 1988. Copyright 1988 American Chemical Society. Free energies of activation for viscous flow, AGj,,,, calculated from surface viscosities qs. 'These films were too unstable to allow viscosity measurements. These films flowed too fast for accurate viscosity measurements.

a

resulted in expanded monolayer films. The expanded nature of the film may then allow for greater entanglement as measured by the free energies of activation to viscous flow. Furthermore, for all diacid pairs the ( k )-isomer has a higher AGf,,, than its corresponding meso-isomer. On the basis of previous arguments, it is reasonable to relate this difference to the stereochemically dependent conformation at the air-water interface, which in some manner leads to different degrees of intermolecular entanglement. These four pairs of diastereomeric two-chain surfactants provide an unprecedented opportunity to compare the effects of stereochemistry on

121

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

intramolecular interactions in monolayers. The monolayer films showed divergent behavior as a function of the stereochemistry at the chiral centers. The energetics of compression for meso-diastereomers were reminiscent of two single-chain fatty acids joined together. The extra work required to bring the stereochemically directed chains of the ( f )-isomer into a standing position reflect their preferred orientation. Rheological data show that definite differences exist in the resistance of the films to flow as a function of stereochemistry. These differences reflect the conformations of ( * ) - and meso-isomers as they sit at the air-water interface. What is much harder to elucidate is the effect of stereochemistry on intermolecular interactions. How does changing the stereochemistry at one chiral center affect interactions between diastereomers? Ab initio molecular orbital calculations have been used to address the problem of separating stereochemically dependent inter- and intra-molecular interactions in diastereomeric compounds (Craig et al., 1971). For example, diastereomeric compounds such as 2,3-dicyanobutane exhibit significant energetic dependence on intramolecular configuration about their chiral centers. So far, however, little experimental attention has been focused on this problem. Separation of intermolecular interactions. The keto-linked diastereomeric surfactants present a novel opportunity to compare not only the effect of stereochemistry on intermolecular interactions, but also the effect of carbonyl linkage position and overall fatty acid chain length. It is reasonable to assume that chiral discrimination exists between diastereomers in these monolayers as has been shown for several enantiomeric systems (Stewart and Arnett, 1982; Arnett et al., 1982); the problem is to separate these effects from the intramolecular contributions to film compression energetics. Accordingly, we have undertaken a Goodrich analysis of the excess free energy of mixing of mixtures of C-15:6,6’ and C-15:9,9 (p. 112), and two analogous diacid diastereomeric pairs constructed from chains of 12 and 18 carbons, 6,8-dihexyl-7-oxotridecanedioic (C-12: 6,6) and 6,8-didodecyl-7oxotridecanedioic acids (C-18:6,6). Goodrich’s original derivation and treatment for obtaining excess free energies of mixing utilizes the differences between the free energies of compression of the pure film components and their mixtures from n 0 to some specified pressure as expressed by ( 17).

-

n AGgisi, = Jn-o

(A1.2

- NlAl - NA2)

(17)

Excess free energies of mixing can be used to separate the contributions of stereochemistry to intermolecular interactions from their contributions to

122

P. L. ROSE €T A L .

intramolecular interactions. If one knows the role that stereochemistry plays in mesolmeso-interactions (which can be deduced easily from the appropriate surface pressure uersus area isotherm), and one knows the role that stereochemistry plays in ( )/( f)-interactions, then one can mix meso- and ( *)-isomers in a monolayer film and separate out the contribution of mesolmeso interactions, N , A , , and ( f)/( 5 )-interactions, N,A,, from ( f )/meso-interactions, A , , , . One is then left with a property AGZix, which can be used to determine what effect, if any, stereochemistry has on intermolecular interactions. The inherent assumptions of this method have been detailed in Section 2 (p. 67). The treatment given here has accounted for these where possible and is based for the most part on the basic assumptions of Goodrich ( 1957). Figures 46 and 47 show the effect of successive incorporation of meso-diastereomer on the phase transition characteristic of the ( f )-isomer; as the concentration of meso-isomer increases, the phase transition surface pressure IT occurs at higher n. The same'result was found (Arnett et al., 1988b) for mixtures of (f)-and meso-C-15:6,6' and C-15:9,9', which are not shown here. According to the surface phase rule, this is indicative of

4)-

C-lZP,6

D

-

10

a -

t

o

I

rn

1

41

I

m

I

m

I

m

I

m

1

140

I

160

I

180

I

am

Fig. 46 Surface pressure/area isotherms for the compression cycles of ( + ) - and rneso-C-12:6,6 ketodiacids and their mixtures on a pure water subphase at 25°C carried out at a compression rate of 19.24 A2/molecule per minute: A, 0% ( f ) (or 100% meso); B, 25% ( f); C, 50% ( f); D, 75% ( f ); E = 100% ( f). Reprinted with permission from Arnett et al., 1988b.Copyright 1988 American Chemical Society.

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

123

Fig. 47 Surface pressure/area isotherms for the compression cycles of ( f ) - and meso-C-18:6,6 ketodiacids and their mixtures on a pure water subphase at 25°C carried out at a compression rate of 19.24Az/molecule per minute: A, 0% ( f ) (or 100% meso); B, 25% ( f); C, 50% (k); D, 75% ( f); E, 100% ( f ). Reprinted with permission from Arnett et al., 1988b. Copyright 1988 American Chemical Society.

mixing between the ( + ) - and meso-components at ll > II‘ (Crisp, 1949; Defay, 1932). The additive area relationship was used to monitor the mixing of ( f )/meso-surfactants below the phase change. Nonideal miscibility was seen for all sets of surfactants at the lower surface pressures, assuming a more ideal (or conversely, segregated) matrix at 15.0 dyn cm- (Arnett et al., 1988b). Coupled with the change in IT with varying meso-composition, we conclude that ( )/meso-diastereomers form nonideal mixtures for all four sets (C-12:6,6‘, C-15:6,6’, C-18:6,6‘ and C-15:9,9’). Since it has been shown that nonideal mixing occurs in the 2.5- 15.0 dyn cmrange, the excess free energies of interaction were calculated for compressions of each pure component and their mixtures to each of these surface pressures. In addition, these surface pressures are below the ESPs and/or monolayer stability limits so that dynamic processes arising from reorganization, relaxation, or film loss do not contribute significantly to the work of compression. The fact that the (*)-diastereomer is a mixture of enantiomers will have no effect on the interactions of meso- and ( f )-isomers. The interaction between diastereomers may be conceptualized easily in the symmetry

+

124

P. L. ROSE €T AL.

relationship ( 18), where R,R and S,S configurations are represented in the R,R-R,R

= S,S-S,S

# R,S-S,R = S,R-R,S

(18)

( f)-diastereomer and R,S and S,R in the meso-isomer. Since the meso-isomer

has a plane of symmetry (R,S and S,R will be indistinguishable), R,R-R,S interactions will be the same as S,S- R,S interactions. However, there will be an inherent R In 2 difference in the entropy of mixing for pure ( f )- and pure meso-compounds in a monolayer since the ( f )-compound is composed of two enantiomers. Assuming that the individual enantiomers in the ( f)-film mix ideally, the entropy of mixing is given by ASmix= - R( N R , RIn N R , R+ Ns,s In N s , s ) per mole (that is, for the racemate, where Ns,s = N R , R= 0.5, ASmix= R In 2). Parts (a)-(d) of Fig. 48 give the excess free energies of mixing for all four sets of diastereomers. The highly expanded C-12:6,6 and C-l5:9,9’ films show marked deviation from ideality (AG”‘ # 0) upon compression to 10.0 and 15.0 dyn cm- indicating a dependence of packing on stereochemistry as the film is compressed. Conversely, the C-15:6,6’ and C-18 :6,6 systems show no deviation from ideality at every surface pressure across the entire composition range. Stereochemistry appears to play little or no role in the molecular interactions between ( f )- and meso-isomers for these films or are compensating each other. These differences in AGxs can be interpreted on the basis of molecular properties by an analogy to a pair of ordinary hedge-clippers. Consider each molecule to be divided above and below the carbonyl linkage. The head-group section below the linkage group is represented by the “blades” of the hedge-clipper and the hydrocarbon tails above the linkage groups are represented by the “handles” of the hedge-clipper. The linkage carbonyl group can then be considered as the “bolt” holding everything together (Fig. 49). The meso-diastereomer is then the ordinary hedge-clipper with its plane of symmetry bisecting the bolt. If one blade and handle section trade positions about the bolt, the hedge-clippers represent the ( f )-diastereomer. One may vary molecular properties by changing the position of the carbonyl linkage, the overall length of the fatty acid units and the stereochemistry of the carbonyl bolt. These compounds then present a unique set in which all three variables are presented: (i) The case where the blade lengths are kept constant and the handle lengths are varied is represented by the C-12, C-15, and C-18 diastereomers linked at the 6,6 positions. Of this set, the short-handled C-12 system is the only one that shows excess free energies of mixing.

’,

(ii) The case of varying blade length with constant handle length is represented by the C-12:6,6’ and C-15:9,9’ systems. Both of these

125

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

300

300

r

I

I

-g -

200

p

100

.-P .E r r

0

0

-

?m .-.-x

u-

-100

O

2

3

-200

300

-

5

?m

-100-

3

-200-

::

.-

.5

5

2.5,5.0, 10.0,15.0

O

z

2.5.5.0,10.0,15.0

-2001

.

1

.

1

-

1

300

I

,;,---r---i---i---.*

u-

--d

;200 -

r

-t! m 100-

;.

-100-

-300

200 -; m .-

: --- 4--- t--- r

0,

0

1 -300

I

loo-

I

: m

-

-; 2oo-

c

vO

0 U

100 0

-100

3 -200

0

-300

W 1

.

1

.

1

.

1

-300

systems have significant deviation from ideal mixing thermodynamics. The C-12:6,6 films are more highly expanded- than those of the C-15:9,9. We saw earlier that moving the carbonyl linkage (bolt) up the hydrocarbon chain for the case of the C-15 diastereomers increases the head-group-to-head-group distance, resulting in expansion of the monolayer films and increased energetics of film compression (Harvey et al., 1988). Here, C-12:6,6, having a shorter carboxylate/carbonyl distance, forms a more expanded film, indicating a large handle contribution to film compression energetics.

126

P. L. ROSE € T A L .

meso

0

0

0 ’bolt”

Fig. 49 Schematic representation of stereochemically dependent conformations for

( & ) - and meso-C-15:9,9’ ketodiacids at the air-water interface. The portion of the

molecule most affecting intermolecular interactions is in brackets. Reprinted with permission from Arnett et al., 1988b.Copyright 1988 American Chemical Society.

(iii) The case of overall clipper size remaining constant and varying the position of the bolt is represented by the C-15:6,6 and C-15:9,9’ systems. This variation results in a change of both blade and handle length. Both systems show an effect of stereochemistry on intermolecular interactions. It must be kept in mind that the Goodrich treatment separates out contributions to intermolecular interactions that arise from film expansion. The differences in film expansion are a reflection of conformation and are accounted for in the pure meso- and (k)-films. However, since ( f )- and rneso-film components do interact, the intramolecular contribution to film compression may be altered. This effect would arise from conformational perturbations as molecules interact, thereby precluding complete separation of inter- and intra-molecular contributions to the thermodynamics of compression. However, these complicating factors can be mitigated by comparing several molecules with varying structures, as has been carried out in this instance. The repeating theme seen in these comparisons is that the handle portion of the molecules plays a significant role in intermolecular interactions between meso- and ( f )-isomers.At low surface pressures, the carboxylate head groups

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

127

and carbonyl linkage should be interacting with the water subphase, leaving the hydrocarbon handles free to interact with each other (Fig. 49). It is, therefore, not surprising that the shorter-handle (i.e., six-carbons) system express stereochemically dependent ( )/meso-interactions through excess free energies of mixing. The van der Waals interactions of the short hydrocarbon chains must occur at points closer to the chiral bolt section than those of their longer-chain counterparts. The steric contribution to packing is preserved as the mixed films are compressed. The systems with long handles have their interactions occurring at points further removed from the chiral bolt. Any stereochemical effect on intermolecular interaction is lost through the large number of degrees of freedom represented in intervening methylene units. Therefore, the mixing thermodynamics serve to extract the limits of the effects of chirality on intermolecular interactions. Unlike the case of stearoylserine methyl ester and most other systems presented here, the chiral molecular recognition between the diastereomers occurs in a monolayer phase that is stable and fluid.

*

Amide-linked diacids The effect of stereochemistry on inter- and intra-molecular interactions for a series of straight chain fatty acids linked by a carbonyl group at various positions along the hydrocarbon chains has already been shown. The lowest energy conformations for ( * ) - and rneso-ketodiacids at the air-water interface resulted in very different behavior in their energetics of compression. Much more work was required to overcome the intramolecular conformational forces of the (*)-isomer in order to bring the hydrocarbon chains into collinearity. In contrast, the lowest energy confirmation of the meso-isomer already had the hydrocarbon chains side-by-side, resulting in a much lower work of compression. The amide-diacids presented here represent the case where the carbonyl linkage group has been replaced by an amide group. What effect then, does going from a carbonyl linkage to an amide linkage have on the energetics of compression as a function of the stereochemistry at the points of attachment? Although, unfortunately, the assignment of stereochemistry could not be made, several interesting comparisons between the keto- and amide-linked diacids are apparent. The diastereomers are differentiated by their affinity for a C-18 reverse phase HPLC column. Accordingly, the terms “first eluting” (FE) and “second eluting” (SE) are used to reflect their relative HPLC retention times. The l l / A isotherms for ( * ) - and meso-C-12:6.6’ keto diacids were given in Fig. 46. As was discussed, there are differences between the compression and expansion cycles for films cast from these diastereomers, which reflect

P. L. ROSE E T A .

I28

0

40

80

120 160 d2/moIecuIe

200

240

Fig. 50 Surface pressure/area isotherms for the compression and expansion cycles of first (solid line) and second (dashed line) eluting C-12:6,6-A amide diacids on a pure water subphase at ( A ) 2 5 T , (B) 30"C, and (C) 35"C, carried out at a compression rate of 19.24 Az/molecule per minute. Arrows indicate the direction of compression and expansion.

their conformations at the air-water interface. Both the meso- and ( * ) isomers form highly expanded monolayers at low surface pressures with lift-off areas of approximately 300 and 500 A2/molecule, respectively. In contrast, the first and second eluting diacids, 6-( 6-dodecanoiccarbamoy1)dodecanoic acids (C-12:6,6-A), form more highly condensed monolayers than either diastereomer of the keto diacids (Fig. 50A). There is an additional attraction between amide-linked molecules, which is absent in the interaction of keto diacids. The proclivity of the amide group for hydrogen bonding (to amide groups on neighbouring diacids) could lend itself to intermolecular interactions in monolayers cast from amide-linked diacids resulting in a condensed film. Conversely, hydrogen bonding of the amide group to water might also lead to greater expansion of the film. Comparison of the compression and expansion cycles for C- 15:6,6 diacids linked with a carbonyl group (Fig. 4 1) and those linked with an amide group (Fig. 51A) show that the C-15 amide diacids, 6 4 6-pentadecanoiccarbamoy1)pentadecanoic acids (C- 15:6,6-A), also form much more highly condensed monolayers than do their keto counterparts. Addition of three methylene

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

129

Fig. 51 Surface pressure/area isotherms for the compression and expansion cycles of first (solid line) and second (dashed line) eluting C-t5:6,6-A amide diacids on a pure water subphase at (A)2 5 T , ( B ) 30"C, and (C) 35"C,carried out at a compression rate of 19.24 A2/molecule per minute. Arrows indicate the direction of compression and expansion.

groups to the hydrocarbon chains above the amide linkage increases chain-chain, van der Waals interactions to the point where there is no measurable effect of stereochemistry on the energetics of compression; the compression and expansion cycles are almost coincident. The stereochemically dependent conformations manifested in the films cast from the C-12:6,6-A compounds are masked in the films cast from the C-15 analog due to their condensed nature. This, presumably, is the result of intermolecular hydrogen bonding interactions at the amide bridges and chain-chain interactions along the hydrocarbon chains, which hold the molecules in a tightly packed arrangement. An increase in the temperature of the C-l5:6,6-A film should perturb intermolecular hydrogen bonding interactions, possibly resulting in any stereochemically dependent conformations being expressed in the energetics of compression. Figures 51(A-C) give the compression and expansion cycles for the two isomers of the C-l5:6,6'-A diacids at 25,30 and 35°C. At 30"C,the energetics of compression of the second eluting C-l5:6,6'-A diacid are similar to those of the second eluting C-12:6,6-A diacid at 25°C. Raising the temperature causes a weakening of the intermolecular interactions in the monolayer spread

130

P. L. ROSE E T A L .

from the SE C-15: 6,6‘-A compound. The stereochemically directed conformation seen in the FE C-12 diacid at 25°C and masked in the SE C-15 diacid at 25°C is now expressed at 30°C for the second eluting C-15 diacid as a change in the energetics of resisting compression (expansion of the monolayer at low surface pressures). The stereochemically directed conformation of the F E isomer affords stronger intermolecular associations. The behavior of the molecules in a monolayer film is relatively insensitive to the temperature of the subphase over the same range of temperatures where the SE isomer is quite sensitive. It is seen from Table 14 that the FE isomer spread less spontaneously from the bulk crystal to a monolayer state. This is an indication that associations in the bulk crystal are stronger for the FE isomer than for the SE. In addition, the entropy of spreading is lower for the FE isomer, indicating a more ordered and conversely a more tightly packed film or a less ordered crystal. It cannot be ruled out at this time that the absence of a difference in the n/A isotherms for the FE and SE isomers-of the C-15:6,6-A diacids is the result of a stereochemically dependent intramolecular amide-carboxylate hydrogen bonding interaction leading to conformations at the air- water interface that are indistinguishable on the Langmuir film balance. If this were the case, raising the temperature of the subphase could result in weakening preferentially any inter, or intra-molecular hydrogen bonds, giving the results seen in Figs. 51(A-C). At this point, it is not known whether temperaturedependent energetics of compression are the result of inter- or intra-molecular hydrogen bonding. It has been shown by Harvey et al. (1989) that incorporation of palmitic acid into a monolayer spread from stearoylserine methyl ester (SSME) breaks up intermolecular SSME interactions. The palmitic acid acts as a two-dimensional diluent. Figures 52( A-C) give the l l / A isotherms for mixtures of FE and SE C-l5:6,6’-A with palmitic acid. Dilution of the monolayer cast from the second eluting isomer with 15 mol% palmitic acid separates the diacid molecules from one another on the water surface and perhaps allows for the expression of their stereochemically dependent conformations. The mixed film ( 15% palmitic acid/85% C-15:6,6-A) expands at low II and behaves in much the same manner as the single-component monolayer (C- 15:6,6’-A) behaves at 30°C. Addition of 15 mole% palmitic acid into a monolayer cast from the FE C-15 diacid has little effect on its energetics of compression, indicating a stronger intermolecular interaction afforded by its stereochemically dependent conformation at the air-water interface. Incorporation of higher mole percent palmitic acid in films spread from both isomers causes an expansion of the films beyond that obtained from simple additivity relationships. At all mole fractions of palmitic acid, the

Table 14 Equilibrium spreading pressures lTe and surface excess free energies, entropies, and enthalpies of spreading for first and second eluting C-15:6,6-A amide diacids.

A e / A Zmolecule-

lTe/dyncm-'

AG:/cal mol-

Temp./K

1st eluting

2nd eluting

1st eluting

2nd eluting

298 303 308 313

0.20 f 0.18 0.55 f 0.21 0.58 f 0.22 0.89 f 0.28

0.41 f 0.36 1.38 f 1.08 4.44 f 0.57 5.70 f 0.36

62.8 f 1.0 56.0 f 0.6 54.4 f 0.8 58.5 f 0.5

53.9 f 1.1 55.3 f 0.5 58.9 0.8 59.5 f 1.0

AS:/calmol-'

K-'

1st eluting -18.1 -44.3 -45.4 -75.0

AH:/kcal mol-

'

1st eluting

2nd eluting

1st eluting

2nd eluting

3.4 f 0.8

32.7 f 5.0

1.0 f 0.2

9.74 f 1.52

f 11.5 f 12.1 f 12.4 f 16.8

'

2nd eluting -31.8 f 19.8

- 109.8 f 60.4 - 376.2 f 37.1

-488.4 f 27.1

132

P. L. ROSE E T A L .

A 0 -

w -

I

La-

?.

10

-

Afiwleoule

Fig. 52 Surface pressure/area isotherms for the compression and expansion cycles of mixtures of (A) 15, (B) 30, and (C) 50moIe % palmitic acid with first eluting C-15:6,6'-A amide diacid (solid line) and second eluting C-15:6,6'-A amide diacid (dashed line) on a pure water subphase at 25"C, carried out at a compression rate of 19.24A'/molecule per minute. Arrows indicate the direction of compression and expansion.

energetics of compression are different for mixtures with each isomer. These data, and those reported for the temperature dependence of C-l5:6,6'-A molecules in a monolayer film, indicate that the condensed nature of the monolayers cast from the long-chain amide diacid is due to increased chain-chain interactions of the long hydrocarbon chains and intermolecular hydrogen bonding of the amide linkages. Furthermore, the strength of hydrogen bonding interactions depends on the stereochemistry of the chiral carbons at the points of attachment of the amide group; the stereochemistry gives each isomer its own conformation, which then allows for varying degrees of intermolecular interactions. At 25"C, the isomeric conformations are masked by the condensed nature of the film. However, raising the temperature and/or adding an achiral diluent breaks up intermolecular interactions preferentially, which allows the diastereomers to express their stereochemically dependent conformations. The effects of temperature and added achiral diluent to C-15:6,6-A amide diacids can also be due to differences in intermolecular interactions based on the molecule's configuration. The arguments given above are based on the assumption that the different configurations, the stereochemistry at the chiral

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

133

carbons, for FE and SE isomers results in different conformations on the water surface. Although this seems reasonable given the data presented here, the case is more ambiguous in the absence of X-ray crystal structures than for the keto diacids. 4

Conclusions

This chapter has reported the only extensive and coordinated investigation of the effects of chirality on the properties of monolayer films spread at the air-water interface. Twenty compounds of varied headgroup and chain length have been examined carrying one and two chiral centers. In every case, all of the optical isomers-enantiomers and diastereomers-were made and their properties measured both as pure compounds and as mixed monolayers in order to compare phase changes in the films with mixed melting points of the crystals. Since the laws of symmetry require that all properties of enantiomers (except their interactions with other chiral systems) be exactly the same, these studies have profited by the application of an absolute test for the presence of impurities, a perennial problem in monolayer research. In every case, all measurements were repeated with both enantiomers. Unless the results agreed within experimental error, the compounds were purified repeatedly until they did agree. An unusually extensive battery of experimental techniques was brought to bear on these comparisons of enantiomers with their racemic mixtures and of diastereomers with each other. A very sensitive Langmuir trough was constructed for the project, with temperature control from 15 to 40°C.In addition to the familiar force/area isotherms, which were used to compare all systems, measurements of surface potentials, surface shear viscosities, and dynamic suface tensions (for hysteresis only) were made on several systems with specially designed apparatus. Several microscopic techniques, epifluorescence optical microscopy, scanning tunneling microscopy, and electron microscopy, were applied to films of stearoylserine methyl ester, the most extensively investigated surfactant. The most important series of compounds used for examining the effects of stereochemical variation in the headgroup was the long-chain (mostly stearoyl) amino acid methyl esters. These were chosen because of the commercial availability of both enantiomers, their relevance to biochemistry, and the previous demonstration by Zeelan and Havinga (1958) that such compounds produce good monolayer films. Several series of diastereomeric two-chain dicarboxylic acids, synthesized by Porter’s research group, were used to explore the effects of joining the chains at different points relative

P. L. ROSE ETAL.

134

to the hydrophilic carboxylic acid groups in the water surface. Large differences between meso- and ( )-diastereomers were interpreted readily in terms of conformational forces around the points of attachment at the linkages between the chains. With only one significant exception, the phosphatidylcholines, pronounced stereochemical discrimination was expressed in all of the properties used to compare enantiomers with their racemic mixtures or diastereomers with each other. Moreover, the degree of chiral discrimination was highly dependent on the structure of the headgroup, the temperature and degree of compression of the film. Generally, clear stereochemical recognition was only well defined when films had reached liquid-condensed states. The failure to detect any degree of chiral recognition in monolayers of phospholipids remains an interesting sideline of this study, consider their importance in cell membranes and the ubiquitous employment of chirality for providing selectivity in biochemical processes. For a number of the systems, comparisons were made between the effects of enantiomeric composition in the monolayer and corresponding meltingpoint--composition curves for the crystals. All of the latter gave clear evidence of racemic compound formation in the crystals, and this type of pattern was repeated in the monolayer properties. Hysteresis was generally observed in the compression-expansion cycles of the force-area isotherms, indicating that the timescale for relaxation of the fully compressed film back to its expanded state was slower than the movement of the barrier of the Langmuir trough. Our studies, like many others, imply that monolayers are metastable and that reversible thermodynamics can only be applied to their analysis with caution.

Acknowledgements

We are glad to acknowledge here the contributions of our many colleagues who worked on this project, especially Drs Martin Stewart, Orlean Thompson, Barbara Kinzig, Jeffrey Gold, Lynne Collins-Gold, Eric Johnson, Robert Verbiar, Patricia Zingg, Graham Whitesell, Dorla Mirajovsky, Jonathan Heath, and Mrs Marjorie Richter, and for financial support, at various stages, from the National Science Foundation, the U.S. National Institutes, AT&T and Duke University. Symbols

n Y

Surface pressure Surface tension

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

135

Pi Surface potential Number of molecules Surface dipole moment Molecular area Gas constant Temperature Mole fraction Number of degrees of freedom Number of components equilibrated throughout the system Number of components in the surface monolayer Number of bulk phases equilibrated throughout the system Number of surface phases in equilibrium with each other Free energy of mixing Phase transition surface pressure Equilibrium spreading pressure (ESP) Molecular area at lIe Excess entropy of spreading Excess enthalpy of spreading Surface shear velocity Rate of shear Canal length Canal width Viscosity of subphase Helmholtz free energy of compression Helmholtz free energy of expansion Work of compression/expansion cycle Transition temperature

References Arnett, E. M. and Thompson, 0. (1981). J. Am. Chem. SOC. 103, 968 Arnett, E. M. and Zingg, S. P. (1981). J. Am. Chem. SOC. 103, 1221 Arnett, E. M., Chao, J., Kinzig, B. J., Stewart, M. V., Thompson, 0. and Verbiar, R. J. (1982). J. Am. Chem. SOC. 104, 389 Arnett, E. M., Gold, J. M., Harvey, N., Johnson, E. A. and Whitesell, L. G. (1988a). In Advances in Experimental Medicine and Biology, Vol. 238 Amett, E. M., Harvey, N. and Rose, P. L. (1988b). Langmuir 4, 1049 Arnett, E. M., Harvey, N. G. and Rose, P. L. (1989). Acc. Chem. Res. 22, 131 Arnett, E. M., Harvey, N., Mirajovsky, D. and Rose, P. L. (1990). J . Am. Chem. SOC. 112. 3547 Bain, C. D. and Whitesides, G. M. (1988). Science 240, 62 Bienkowski, R. and Skolnick, M. (1974). J. Colloid Interface Sci. 48, 350

136

P. L. ROSE E T A L .

Brittain, W. J., Porter, N. A and Krebs, P. J. (1984). J. Am. Chem. SOC. 106, 7652 Cadenhead, D. A. (1969). Ind. Eng. Chem. 61(4), 22 Chapman, D. ( 1968). Biological Membranes. Academic Press, New York Cook, A. and Webb, W. R. (1966). Ann. Thorac. Surg. 2, 327 Craig, D. P. and Mellor, D. P. (1975). Phys. Scr. 79, 119 Craig, D. P., Radom, L. and Stiles, P. J. (1971). Proc. Roy. SOC. Lond. A343, 1 1 Cram, D. J. (1984). Proc. 29th ZUPAC Congress; Pure Appl. Chem., p. 3 Crisp, D. J. (1949). Surface Chem. ( S u p p l . Res., London), p. 17 Defay, R. ( 1932). Doctoral dissertation, Brussels Dragcevic, D., Milunovic, M. and Pravetic, V. (1986). Croatia Chim. Acta 59, 397 DuNouy, L. (1929). Science 69, 251 Eliel, E. L. (1962). Stereochemistry of Carbon Compounds, pp. 43-47. McGraw-Hill, New York Fendler, J. H. (1982). Membrane Mimetic Chemistry. John Wiley, New York Fredga, A. (1944). In The Suedberg Memorial Volume (ed. A. Tiselius and K. 0. Pederson). Almquist and Wiksell, Uppsala Furhop, J. H. and Fritsch, D. (1986). Acc. Chem. Res. 19, 130 Furhop, J. H., David, H. H., Mathieu, J., Liman, U., Winter, H. J. and Boekema, E. J. (1986). J. Am. Chem. SOC. 108, 1785 Gaines, G. L. (1966). Insoluble Monolayers at Liquid-Gas Interfaces, Chaps. 7 and 8. Wiley, New York Gaines, G. L., Tweet, A. G. and Bellamy, W. D. (1965). J. Phys. Chem. 42, 2193 Gershfeld, N. L. (1972). In Techniques of Surface Chemistry and Physics (ed. R. J. Good, R. R. Stromberg and R. L. Patrick), Vol. I. Marcel Dekker, New York Gershfeld, N. L. (1976). Ann. Reu. Phys. Chem., 27, 349 Gershfeld, N. L. (1982). J . Colloid Interface Sci. 85, 28 Gershfeld, N. L. (1986). Faraday Disc. Chem. SOC. 81, 19 Gershfeld, N. L. (1988). J . Membr. Bio. 101, 65 Gershfeld, N. L. and Pagano, R. E. (1972a). J . Phys. Chem. 76, 1231 Gershfeld, N. L. and Pagano, R. E. (1972b). J . Phys. Chem. 76, 1238 Gershfeld, N. L. and Pagano, R. E. (1972~).J. Phys. Chem. 76, 1248 Gibbs, J. W. ( 1948). Collected Works,Vol. 1, p. 219. Yale University Press, New Haven Gold, L. C. Unpublished results, Duke University Goodrich, F. C. (1957). Proc. 2nd Int. Congr. Surf: Act. 1, 85 Greene, F. D., Berwick, M. A. and Stowell, J. C. (1970). J . Am. Chem. SOC.92, 867 Guay, D. and LeBlanc, R. M. (1987). Langmuir 3, 575 Handa, T. and Nakagaki, M. (1979). Colloid Polyrn. Sci. 257, 374 Harkins, W. D. and Kirkwood, J. G. (1938). J. Phys. Chem. 6, 298 Harkins, W. D., Young, T. F. and Boyd, E. (1940). J. Chem. Phys. 8, 954 Harvey, N. G. and Arnett, E. M. (1989). Langmuir 5, 998 Harvey, N., Rose, P., Porter, N. A., Huff, J. B. and Arnett, E. M. (1988). J. Am. Chem. SOC. 110,4395 Harvey, N. G., Mirajovsky, D., Rose, P. L., Verbiar, R. and Arnett, E. M. (1989). J . Am. Chem. SOC. 111, 1115 Harvey, N. G., Rose, P. L., Mirajovsky, D. and Arnett, E. M. (1990). J. Am. Chem. SOC. 112,3547 Heath, J. G. ( 1991). Doctoral dissertation; Duke University Heath, J. G. and Arnett, E. M. (1992). J. Am. Chem. SOC. 114, in press Jacques, J., Collet, A. and Wilen, S. H. (1981). Enantiomers, Racemates and Resolutions, pp. 93-97. Wiley, New York

CHIRALITY AND MOLECULAR RECOGNITION IN MONOLAYERS

137

Jaffe, A. B., Skinner, K. J. and McBride, J. M. (1972). J . Am. Chem. SOC.94, 8510 Jarvis, N. L. (1965). J . Phys. Chem. 69, 1789 Jeffers, P. M. and Daen, J. (1965). J . Phys. Chem. 69, 2368 Johnson, E. A. (1985). Doctoral dissertation, Duke University Joly, M. (1956). J . Colloid Interface Sci. 11, 519 Keller, D. J., McConnell, H. M. and Moy, V. T. (1986). J . Phys. Chem. 90, 231 1 Langmuir, I. and Schaefer, V. J. (1937). J . Am. Chem. SOC.59, 2400 Lucassen-Reynders, E. H. (1973). J . Colloid Interface Sci. 42, 554 Lucassen-Reynders, E. H., Lucassen, J., Garett, P. R., Giles, D. and Holloway, F. (1974). In Monolayers (Advances in Chemistry Series 144) (ed. E. D. Goddard), p. 272. American Chemical Society, Washington, DC Lunquist, M. (1961). Ark. Kemi 17, 183 Lunquist, M. (1963). Ark. Kemi 21, 395 Lunquist, M. (1965). Ark. Kemi 23, 299 Malcolm, B. R. (1973). Prog. Membrane Sci. 7, 123 Malcolm, B. R. (1975). Ado. Chern. 145, 338 Manheimer, R J. and Schechter, R. S. (1970). J . Colloid Interface Sci. 32, 225 Margoni, N. (1871) Nuooo Cim. 2, 5 McConnell, H. M. and Moy, V. T. (1988). J . Phys. Chem. 92,4520 McConnell, H. M., Keller, D. K. and Gaule, H. J. (1986). J . Phys. Chem. 90, 1717 McConnell, H M., Tamm, C. K. and Weis, R. M. (1984). Proc. Natl Acad. Sci. U S A 81, 3249 Moore, B., Knobler, C. M., Broseta, D. and Rondelez, F. (1986). J . Chem. SOC., Faraday Trans. 2 82, 1753 Moore, W. . I and . Eyring, H. (1938). J. Chern. Phys. 6, 394 Morawetz, H. and Kandaniem, A. Y. (1966). J . Phys. Chem. 70, 2995 Munden, J. W. and Swarbrick, J. (1973). Biochim. Biophys. Acta 291, 344 Munden, J. W., Blois, D. W. and Swarbrick, J. (1969). J . Pharm. Sci. 58, 1308 Nagaranjan, M. K. and Shah, J. P. (1981). J . Colloid Interface Sci. 80, 7 Pagano, R. E. and Gershfeld, N. L. (1972). J . Colloid Interface Sci. 41, 31 1 Pasteur, L. (1848). Ann. Chim. 24,442 Petter, R. C., Mitchell, J. C., Brittain, W. J., McIntosh, T. J and Porter, N. A. (1983). J . Am. Chem. SOC. 105, 5700 Petterson, K. (1956). Ark. Kemi 10, 297 Pirkle, W. H and Pochapsky, T. C. (1987). J . Am. Chem. SOC. 109, 5975 Porter, N. A., Arnett, E. M., Brittain, W. J., Johnson, E. A. and Krebs, P. J. (1986a). J . Am. Chem. SOC. 108, 1014 Porter, N. A., Ok, D., Adams, C. M. and Huff, J. B. (1986b).J . Am. Chem. SOC.108,5025 Porter, N. A., Ok, D., Huff, J. B., Adams, C. M., McPhail, A. T and Kim, K. (1988). J . Am. Chem. SOC. 110, 1896 Rakshit, A. K. and Zografi, G. (1980). J . Colloid Interface Sci. 80, 474 Rosano, H. L., Yin, D-N. and Cante, C. C. (1971). J . Colloid Interface Sci. 37, 706 Sauer, B., Chen, Y-L., Zografi, G. and Yu, H. (1988). Langmuir 4, 11 Shafer, P. T. (1974). Biochim. Biophys. Acta 373,425 Shah, D. 0. and Capps, R. W. (1968). J . Colloid Interface Sci. 27, 319 Stallberg-Stenhagen, S. and Stenhagen, E. (1951). Acta. Chem. Scand. 5,481 Stewart, M. and Arnett, E. M. (1982). Top. Stereochem. 13, 195 Tachibana, T. and Hori, K. (1977). J . Colloid Interface Sci. 61, 398 Tanford, C. (1973). The Hydrophobic Efect. John Wiley, New York Tanford, C. (1989). Ben Franklin Stilled the Waves. Duke University Press

138

P. L. ROSE ET AL.

Thompson, 0. ( 1981). Doctoral dissertation, University of Pittsburgh Townsend, D. F and Buck, E. J. (1988). Langmuir 4, 938 Verbiar, R. J. ( 1983). Doctoral dissertation, Duke University Washburn, E. R. and Wakeham, H. R. R. (1938). J . Am. Chem. SOC.60, 1294 Weis, R. M. and McConnell, H. M. (1984). Nature 310, 47 Whitesides, G. M (1991). Science 254, 1312 Zeelen, F. J. (1956). Doctoral dissertation, The State University of Leiden Zeelen, F. J. and Havinga, E. (1958). Red. Trav. Chim. Pays-Bas 77, 267 Zingg, S. P. ( 1981). Doctoral dissertation, University of Pittsburgh Zingg, S. P, Arnett, E. M., McPhail, A. T., Bothner-By, A. A. and Gilkerson, W. R. (1988). J. Am. Chem. SOC. 110, 1565

Transition-State Theory Revisited W. Jom ALEERY

Physical Chemistry Laboratory, University of Oxford, South Parks Road, Oxford O X 1 3QZ, U K

1 2 3 4 5 6 7 8 9 10 11 12

Introduction 139 The classical paradox 140 An alternative derivation 140 Where is the transition state? 143 The adiabatic case 145 Multistep reactions in solution 147 Proton transfers to cyanocarbon bases 152 Extension to include xR 153 Reactions at liquid/liquid interfaces 155 Colloidal deposition 158 Free energy profiles 163 Summary 167 Acknowledgments 168 Appendix 168 References 170

1

Introduction

Most physical organic chemists use transition-state theory to model their reactions and try to devise ingenious experiments to measure the properties of the transition state. Such properties include for instance the geometry, the volume, the charge and the dipole moment. Despite the success of this approach, many physical chemists and nearly all gas-phase kineticists regard the transition-state theory with suspicion, if not outright hostility. For instance we find the following in Peter Atkins’ textbook on ‘PhysicalChemistry (Atkins, 1982): “That is not to say that the theory is complete or even very reliable.” And again he writes, “. .. the basic assumption that the rate process involves an equilibrium is often a stumbling block to its acceptance.” In this 139 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 28 ISBN 0-12-033528-X

Copyright 01993 Academic Press Limited A / / rights of reproduction in anyform reserved

140

W. J. ALBERY

chapter I am first going to discuss a number of difficulties with the theory as it is classically presented. I am then going to show how the theory can be extended to include a wide variety of different kinetic processes ranging from atom-transfer reactions through electron- and ion-transfer reactions to the kinetics of colloidal deposition. Finally I am going to discuss free energy profiles and how they should be used. 2

The classical paradox

The classical derivation of transition-state theory starts by assuming that the transition state, Xs, is in equilibrium with the reactants A and B as represented in (1). The rate of reaction, u, is given by (2)where vT is some

frequency. We are now told that Xs is committed to reaction. For instance Levine and Bernstein ( 1987) write “. . . the concept of a configuration of no return such that when the system has reached the critical configuration it will necessarily proceed to reaction and not “turn back” to the reactants region.” The paradox arises because (1) and equilibrium require a formulation as in (3), where most of Xt that is formed returns to reactants and only a small

fraction is converted to products. Commitment and (2) require exactly the reverse (4);all the Xi that is formed is committed to reaction. It is the

incompatability of (3) and (4)that quite rightly gives transition-state theory a bad name. 3

An alternative derivation

In escaping from the paradox, we have to recognize that, as shown in Fig. 1, in the region of the transition state the species on the free energy surface change from being in equilibrium with the reactants and totally uncommitted

TRANSITION-STATE THEORY REVISITED

141

G

Fig. 1 A typical free energy profile showing how the fraction committed to reaction varies from zero through at the transition state and finally reaches unity.

to reaction to having passed the point of no return and being totally committed. This is shown by the fraction f of species committed to reaction. The IUPAC glossary of terms used in physical organic chemistry (Gold, 1983) defines the transition state as being the point where f = 0.5. We now proceed to derive the rate expression by creating an intermediate X, as shown in Fig. 2; X is a species that is equally likely to form products or to return to the reactants. Applying the steady state approximation to X we obtain ( 5 ) where R is the reactant(s). u = $k,[R]

= kJX]

(5)

Fig. 2 A hypothetical intermediate that is equally likely to return to reactants or to transform into products.

142

W. J. ALBERY

+

Hence, instead of (l),we find expression (6). The factor of arises because

X is half committed to reaction and half in equilibrium with R. It is interesting to note that at equilibrium when the back reaction is included, as shown in Fig. 3, then [ X I doubles, with equal contributions coming from R and from the products, P. At equilibrium, the concentration of X is given by (7) and now has its full Boltzmann value.

Now in ( 5 ) k, can be expressed in terms of the frequency v,, shown in Fig. 4, by (8). The factor of 4 arises because it takes one-quarter of a cycle

k,

= 4v,(4) = 2v,

(8)

to escape from the minimum; the factor of arises because X is equally likely to be moving left to right or right to left. Next, as shown in Fig. 5, we squeeze

.

k-l

Fig.3 The effect of adding in the back reaction at equilibrium is to double the concentration of X.

kl k,

.

2*

X

\

Fig. 4 The rate constant k , can be related to the frequency vt.

143

TRANSITION-STATE THEORY REVISITED

Fig. 5 The hypothetical intermediate X is transformed into a transition state.

out the artificial dip to obtain a surface with a clean barrier. As we do so, by the usual arguments, vt -+ 0, and the associated partition function q$ goes to its limiting value (9), The combination of (6), (8) and (9) then gives the

usual transition-state expression (lo), where K‘t is given by (11) and K i has u =(kT/h)Ki[R]

Ki

= exp ( -

AGP/RT)

had qt. removed from the partition function product. 4

Where is the transition state?

One of the most important advantages of transition-state theory is that, as shown in ( l l ) , we describe the barrier to reaction in terms of a free energy of activation, and this allows us to include the entropic changes as well as those in potential energy. Despite this improvement, there has been controversy as to whether the transition state is situated at the maximum in potential energy E or the maximum in free energy along the reaction coordinate. After discussion of this point, Laidler ( 1969) reaches the following rather confused conclusion: “When we look at matters from this point of view we seem to be forced to the conclusion that if we think in terms of activated complexes, we should logically regard them as the species of maximumfree energy along the reaction path. It has to be admitted, however, that in view of the artificiality of the activated complex this is not a compelling argument; other ways of looking at the matter can lead to different conclusions.” Let us see if we can do better.

W. J. ALBERY

144

In Fig. 6 we show a potential energy surface which will lead to the maxima in E and G being at different places. As one moves to the right and enters the narrow gorge shown by the narrowing contours, the spacing of the energy levels increases as shown in the lower half of Fig. 6. This entropic barrier means that the maximum in G is somewhere in the gorge and not on the broad sunlit uplands on the left of the diagram. It will be, as shown in Fig. 6, where the spacing is comparable to kT. Now in Fig. 6 we show schematically why the transition state is at the maximum in G and not E . To the left of the transition state there are many levels available for the molecules. In fact very few of them will be creeping along the bottom of the valley and over

Ere /--

many levels

t"' \ 4

man/levels

only one level

Fig. 6 An energy surface where the maximum in E and the maximum in G are not at the same place.

TRANSITION-STATE THEORY REVISITED

145

the maximum in E. Near the maximum in G the system becomes much more ordered and has to proceed in the lowest level. In Fig. 6 we also show the reverse reaction, which starts in the lowest level. However, close to the maximum in G, where the spacing collapses, the system can pick up thermal energy irreversibly to populate the thermally accessible levels, and then proceed into the broad valley. This argument demonstrates that it is the maximum in G and not E that determines the location of the transition state. Finally Fig. 7 shows that, as T increases and the entropic factors become more important, the transition state will move further and further into the gorge. It is the author’s view that too little attention has been paid to the determination of the temperature variation of parameters that describe the geometry of the transition state such as the Brnnsted exponent a. 5

The adiabatic case

The argument presented above assumes that the molecules can exchange energy between themselves more rapidly than the rate at which a molecule passes across the potential energy surface. In Fig. 6, for instance, in the reverse direction, the molecules pick up the thermal energy to populate the

Position of $ shifts with increasing T

Fig. 7 The position of the transition state will depend on temperature.

146

W. J. ALBERY

J' E

i,R

i,S

R

Fig. 8 An adiabatic set of reactant and transition-state levels.

accessible levels faster than their rate of progress across the surface. We now explore the opposite situation, in which molecules move across the surface more rapidly than the processes that allow them to change the populations of the different states (Child, 1967). In the gas phase, where energy transfer takes place by intermittent collisions, we are likely to find the adiabatic case. In solution the continuous jostling of the molecules means that energy transfer to populate the accessible states is a much more efficient process. In Fig. 8 we show a set of energy levels for the adiabatic case, where we assume that, in going from R to the transition state, there is no exchange between the different levels. The third level in R must go through the third level in the transition state. The overall rate is then made up of a series of parallel reactions, each with its own individual energy of activation. So we can write (12), where [Ri] is given by (13). Substitution of (13) into (12) u=

1[RJ exp kT h (

-

6,:- Ei,R RT

gives (14) in which the partition function ql:is as defined in ( 15). We therefore

obtain the same expression for the rate of reaction. The only difference is that, as shown in Fig. 9, the partition function ql: of the transition state

TRANSITION-STATE THEORY REVISITED

147

Fig. 9 For the adiabatic transition state each level in contributing to qt has its own geometry.

describes the maximum of each individual energy level. The transition state does not have a single defined geometry. As the temperature rises, there will be increasing contributions from geometries further into the gorge, so, as in the free energy case discussed above, the transition state will appear to shift further down the gorge. It is unlikely that experiments will discriminate between the free energy and adiabatic cases. What is comforting is that the same relationship is obtained and therefore can be used for either case.

6

M u l t i s t e p reactions in solution

A reaction in solution is not a single collision but is a sequence of processes, each with its own time scale. We can identify the steps shown in Scheme 1 Encounter

> 10-9

+ Ionic atmosphere adjustment P Solvation P Atom transfer + 10-9- 10-10

10-11

10- I Z

Scheme 1 (Albery, 1975a,b). The reaction sequence starts with the diffusive encounter of the reactants as they find each other and get into the same solvent cage. At the same time the ionic atmosphere will be adjusting first to any changes in the combined charge and secondly to the charge distribution and dipole moment of the transition state. Changes in solvation follow next, and finally the atom transfer takes place. As shown, the time scale becomes progressively faster. It should be emphasized that these are typical times for a doomed molecule. They are not rate constants. As an example consider the reaction (16). For [I-] = 0.1 mol drn-,, each CH,Br molecule would have an encounter I-

+ CH,Br

+ ICH,

+ Br-

(16)

148

W. J. ALBERY

every lO-'s. Most of these encounters would not lead to reaction. A number of them, however, would have an ionic atmQspherewith a K + ion for instance opposite the bromine atom. Our work on the application of the Marcus treatment (Marcus, 1956, 1963, 1964, 1965) to methyl transfer reactions (Albery and Kreevoy, 1978; Albery, 1980) has shown the importance of transition-state solvation. For a small fraction of the arrangements with the right ionic atmosphere, there would be a solvent fluctuation that would form an incipient solvent shell for the bromide ion and collapse the solvent shell of the iodide ion. Finally a small fraction of the complexes with the right ionic atmosphere and the right transition state solvation will have enough energy to break the carbon-bromine bond. In our description of the Marcus theory of electron-transfer reactions we have found it helpful to plot the free energy change in the three dimensional picture shown in Fig. 10 (Albery, 1975c, 1980). This picture emphasizes that

+

solvent ligands etc

Fig. 10 Free energy plot for an electron-transfer reaction.

ligands etc

X

T

Fig. 11 Free energy plot for an atom-transfer reaction in solution. Diffusive motion along the solvent coordinate opens the opportunity for a favourable atom transfer.

TRANSITION-STATE THEORY REVISITED

149

the electron transfer takes place much more rapidly ( s) than the solvent motions that provide the intermediate solvation of the transition state. We can apply the same concept (Albery, 1975a,b) to the sequence of steps given in (16). In Fig. 11, we now show a free energy barrier for an atom transfer. To simplify the mathematical treatment, we have assumed for the moment that G varies linearly with x, (17), as opposed to the usual Marcus parabola.

The distance xT describes the distance along the x-coordinate over which G increases by RT. We assume that motion along the x-coordinate is diffusive. This will be true for encounters, rearrangement of the ionic atmosphere or the rotation of solvent molecules. We further assume that at some distance xs the atom-transfer reaction becomes possible with a rate constant k. The diffusive kineticequation then becomes( 18),where the step function S(xs) = 0

for x < xs and = 1 for x > xs. The boundary conditions for (18) are c = co at x = 0 and c -+ 0 as x -+ 00. The flux j is given by (19). The observed first-order rate constant kobsis then j = -D[(~),+~]=kjOmS(xt)cdx

given by (20). By taking the Laplace transform of (18) we find that j is given

by (21), with x* as in (22), where xk is given by (23). The distance xk describes

xk = (Dlk)”’

(23)

150

W. J. ALBERY

how far the molecules diffuse along the x-coordinate once they have passed xt and the reaction is switched on. This distance is similar to the thickness of an electrochemical reaction layer. The integral required in (20)is found to be (24)where ct/x* is given by (25). JOmc dx = xT( cg

+ ?)[

1 - exp

(- $)] + 3x* ( x i

-xTx~)

(24)

We can see that the results depend on the three lengths, xT, xt and xk. The length xT describes the steepness of the free energy gradient on the x-coordinate. The length xt describes how far the system has to go along the x-coordinate before the reaction switches on and the length xk describes how far the molecules penetrate along the x-coordinate once the reaction is switched on. The free energy of activation in reaching xi, AG,., is given by ( 26 1. AG,tIRT = x ~ / x ~ (26) Using the ratios of the lengths, xk/xT and x$/x,, we can now construct the case diagram shown in Fig. 12, where from (20)-(25) we find three simple solutions and the regions of the case diagram that they occupy. The insets show the rate-limiting process for each case. For Case I, k is small (xk large) and the rate-limiting process is given by k modified by the free energy required to reach xt. In the inset, k is drawing on the Boltzmann population sustained by D. For Cases I1 and 111, on the other hand, k is large and the rate-limiting process is movement along the x-coordinate. For Case 11, there is a very small free energy barrier, so the rate-limiting step is diffusion (D) through the distance x s , giving D/x$. This case will be found for diffusion-controlled reactions in solution. For Case 111, a significant free energy barrier has to be surmounted and the rate-limiting process is diffusion over the thermal distance xT modified by the free energy barrier. We can now explore in more detail the boundary xk = xT between the two activated Cases I and 111. Figure 13 shows the corresponding values of k and xT for a typical value of D of lo-’ cm2s-’. For typical values of xT we see that k has to be rather large for Case I11 to be found. For outer sphere electron transfers we shall expect to find Case 111. For most atom transfers, where there will be a barrier on the k-coordinate, we will expect to find Case I. So we can summarize in Table 1 the examples of the different cases. With respect to the outer sphere electron transfers, we should in the theoretical description use a frequency factor of D/x+ instead of the classical k T / h of transition-state theory. It is perhaps fortunate and/or fortuitous that both factors are 10” s-’.

-

TRANSITION-STATE THEORY REVISITED

151

I k exp( -AC:/RT

log(x:/xT) = log(AC:/RT)

Fig. 12 Case diagram for the three different possibilities that depend on the characteristic lengths x ~ x,k and xT.

Table 1 Types of reaction and the different cases. Case I Case I1 Case 111

Classical activated reactions Diffusion-controlled reactions Outer sphere electron transfers Simple electron transfer at electrodes Proton transfer to cyanocarbon bases

'

log( k j s - )

Fig. 13 The boundary between Cases I and I11 for D

=

lo-' em's-'.

W. J. ALBERY

152

In the treatment so far we have considered the k-step to be irreversible. This will not be true for Case I11 systems, where the exchange at k is much more rapid than diffusion along the x-coordinate. This situation is considered in more detail in the Appendix, where we show that, in keeping with our general derivation, the concentration of transition states is one half of the Boltzmann concentration and that the fraction committed to reaction is also one-half. 7

Proton transfers to cyanocarbon bases

Besides the outer sphere electron transfers, we have identified (Albery, 1975d) another class of reactions that exhibit Case I11 behaviour, and this example is proton transfer to cyanocarbon bases. These reactions were studied by Long and co-workers. First, by using tritium, they measured the fractionation factor for the tritium as it was pulled off the carbon as in [l], The results

-

""\ -T-0

R-C

NC

/

/H

0 = 0.69

G

\H

+/II

0 = 0.69

O

\

H

PI

c11

in Table 2 show that this factor is close or equal to 0.69. For comparison, the fractionation factor for L 3 0 + is 0.69, where L is either H or D (Gold, 1963; Kresge and Allred, 1963). Hence in the transition state [3] we have L 0 = 0.69

NC

\R-CNC

/

-+/

L-o\

0 = 0.69

L 0

= 0.69

c31 an L 3 0 + unit adjacent to the negatively charged carbon. In accordance with the Hammond postulate the transition state is early in the downhill direction. The solvent isotope effect for the protonation of the base was measured, and the results (Hibbert and Long, 1971) are given in Table 2. One allocates three factors of 0.69 to the L 3 0 + unit and finds that there is still a factor of 0.57-0.60 left over that has to be attributed to nonspecific sites-a "medium" effect. The rate constants for protonation of the carbon base are 10' dm3 mol-ls-l, which is two orders of magnitude slower than observed diffusion-controlled rates for protonation of carboxylate bases. So there is a barrier to the reaction. Yet the large medium effect is unusual. It can, however, be nicely explained if we allocate the system to Case 111, where the medium

-

TRANS IT I0N -STATE TH E 0 RY R EV IS IT E D

153

Table 2 Solvent isotope effects for cyanocarbon bases. Factor for transferring proton from T experiment

Base

0.69 0.71

HC( CN); (CH,),CC(CN);

k,,o/k,,o

“Medium effect’”

0.187

0.57

0.199

0.59

Calculated from ( k ~ , o / k ~ , o ) / ( 0 . 6 9 ) ~ .

effect would arise on the ratio of D-values in the frequency factor. The diffusion coefficient of D 3 0 + is some 40% less than that of H 3 0 + . So a large medium effect in the kinetic solvent isotope effect is diagnostic of Case 111. It is high time that accurate work was carried out on the solvent isotope effect of outer sphere electron-transfer reactions. 8

Extension to include xR

In the treatment so far we have assumed a linear free energy profile along the x-coordinate in Fig. 11. For the two activated cases, Cases I and 111, we now extend the treatment to include a distance xR that describes, as shown in Fig. 14, a parabolic minimum for the reactants as given by (27). The AGIRT

=

G RT

-

Fig. 14 The introduction of a parabolic bowl for the reactants.

(27)

W. J. ALBERY

154

distance xT still describes the linear free energy variation in the region where k switches on. The inclusion of xR means that we have to reconsider the integral 1c dx in (20). For the simple treatment, (28)applies. For the parabolic

minimum this integral has to be replaced by (29), where we have used (27)

j-mmcdx =

[-m2 exp ( - k)’ dx = d / ’ c 0 x R

for AG. Comparison of (28) and (29) shows that, for Case 111, instead of a frequency factor of D / x + we should more strictly use a frequency factor of D/z”~x,x~ For . Case I we find that k is modified by a factor Y , given by (30). This factor is almost certainly less than unity. The reason for this is

-1

AG$/RT

Fig. 15 Plot showing how the first order rate constant depends on ACT when Y from (33) is included.

TRANSITION-STATE THEORY REVISITED

155

that, as shown in Fig. 14, more of the x-coordinate is thermally accessible to the reactants in their minimum compared with the transition states clinging precariously to the steep slope. The inclusion of xR means that the wider or sloppier the reactant bowl, the more difficult it is to collect species in the transition state and hence the slower the reaction. For the particular case of a parabola, using the relation ( 3 1 ) , we find that

xT/xR is given by ( 3 2 ) . Substitution in ( 3 0 ) gives expression ( 3 3 ) for the

Y

= 2(RT/.nAG,,)"'

(33)

modifying factor Y. If AGxr is the main contribution to the overall free energy of activation then one can calculate the size of the corresponding first order rate constant. These values are displayed in Fig. 15. The inclusion of Y has a small but significant effect. 9

Reactions at liquid/liquid interfaces

In recent years there has been considerable interest in the kinetics of ion-transfer reactions between two liquid phases. For moderately immiscible liquids such as water/nitrobenzene, it is thought that, as shown in Fig. 16, there is an interphase region between the two liquid phases through which the ion has to diffuse where the solvent changes gradually from water to nitrobenzene; the thickness of such a layer is probably about l n m . The theory of the kinetics of this process has been developed by Samec, Kharkats and Gurevich (1986). We shall give below a simplified and more general version of their treatment. As shown in Fig. 17, they demonstrated that the

Heptane

Nitrobenzene

H2°

HZ O

Fig. 16 Two different types of reaction at liquid/liquid interfaces, the solvent extraction of copper and the transfer of NO;.

156

W. J. ALBERY

Fig. 17 Linear free energy relationship for the transfer of a series of quaternary cations.

kinetics of the transfer of a series of quaternary nitrogen cations obeyed a linear free energy relation with the slope 01 w f. Note that the rate constant plotted is k'lcms-'; the heterogeneous rate constant k' allows one to write the fluxj/mol cm-, s - l for an interfacial reaction as equal to k' times a bulk concentration c/mol cm - '. As with outer sphere electron transfers, we are dealing with the kinetics of solvent reorganization around the ion as it transfers across the interface-so it is satisfactory to find the same type of linear free energy relation. A somewhat different liquid/liquid system is found in the solvent extraction of copper. We have studied the system using heptane/water. These solvents are so immiscible that, as shown in Fig. 15, there is probably no extensive interphase region but a more sharply defined interface. We have shown that for the oxime ligand, Acorga P50 [4], written as HL, the mechanism of the reaction is as in Scheme 2 (Albery et al., 1984; Albery and Choudhery, 1988). $1 $2 $3 CUL, Heptane 2HL HL HL Interface Site s HL s CuL' s C U ~s, Site Water Cu2+ Cu2+ H + 2H 2H+

Scheme 2

()

157

TRANSITION-STATE THEORY REVISITED

We have determined the complete free energy profile for this reaction; it is shown for extraction in Fig. 18, While transition states 1 and 2 involve ligand replacement reactions, the process through transition state 3 is the passage of the CuL, complex from the interface into the organic phase. When this transition state is rate-limiting, we can either write j = k,[CuL,], where k,

-20

1 HL

CuL+

CuL2

Fig. 18 Free energy profiles for the solvent extraction of copper, where L is Acorga P50. The profile shows the free energy of a site on the liquid/liquid interface. All higher-order rate constants are reduced to first-order rate constants by using the concentrations of reactants in either phase. The free energy lost in each cycle can be seen from the difference between 0 and the lo%, 50% and 80% extraction lines on the right of the diagram. The double-headed arrows indicate the rate-limiting free energy difference.

158

W. J. ALBERY

is a first-order rate constant and [CuL,] is measured in molcm-', or we can write j = k'[CuZ+],,, where k', given by (34), is a heterogeneous rate k'

=

KadSk3

(34)

constant (in cm s-') and Kads(in cm) describes the pre-equilibrium adsorption of CuL, with respect to the aqueous Cu2+. From our previous treatment for a Case I11 system, k, is given by (35) and the adsorption equilibrium

constant

Kads by

(36). From (33)-(35) we find expression (37) for k', It is

satisfactory that this expression does not contain xR, a property of the adsorbed intermediate, but only contains parameters that are concerned with transition state 3. It is important to be able to describe the system with the rate constants for the individual steps (e.g. k,) and with the overall rate constant k'. 10

Colloidal deposition

The kinetics of the deposition of colloidal particles on to a solid surface have been measured using a rotating disc to provide controlled mass transport (Marshall and Kitchener, 1966; Clint et al., 1973). In these early studies the particles were allowed to accumulate on the disc for approximately 30 min, and then the disc was removed and the numbers of particles were counted. More recently van de Ven (Dabros and van de Ven, 1987; Varennes and van de Ven, 1987, 1988) and ourselves (Albery et al., 1985, 1990a,b) have used the hydrodynamics of the impinging jet rather than the rotating disc. In this system the colloid solution is pumped through a jet on to a microscope slide where the deposition is observed in real time. In our technique we shine light on to the edge of the slide, so that it is totally internally reflected. We then observe the scattering of the evanescent wave from the particles that are deposited on the surface. The rate of increase in scattered light intensity measures the rate of deposition. By using a slide coated with indium tin

TRANSITION-STATE THEORY REVISITED

159

oxide, we can change the electrode potential of the surface and so study the effect of electrode potential on the deposition kinetics (Albery et al., 1990b). When the particles are oppositely charged with respect to the surface, the rate of deposition is controlled by the mass transport to the surface. However, when the particles and the electrode are of the same charge, there is a substantial barrier to the deposition. Furthermore, the interplay between the electrostatic repulsion and the van der Waals attraction leads to a secondary minimum some tens of nanometres away from the surface. A typical potential energy surface is shown in Fig. 19. The secondary minimum is usually found a Debye length or so from the surface; the attractive forces are still present but the repulsive electrostatic forces are shielded. In our recent work we have shown that we can trap particles in the secondary minimum, and then by

E/kT

100

0

0.C

Elkt

-2.0 -100

-4.0

-200

-6.0

I 100

loo0 log(x/A)

Fig. 19 Typical free energy profile for the deposition of a colloid.Note the logarithmic distance scale and the divided energy scale for the relatively shallow secondary

minimum and the much deeper primary minimum.

160

W. J. ALBERY

changing the potential we can tip them into the primary minimum. Or we can change the electrolyte concentration and wash them out of the secondary minimum. We are therefore interested in the kinetics of the passage of the particles over an energy surface such as the one shown in Fig. 19. The theory of such processes was developed in an important paper by Ruckenstein and Prieve (1973). Their treatment and the later treatment of Samec (Samec et al., 1986) for the case of ion transfer is essentially the same, and we now develop a simplified and unified model for both systems using the concepts of transition-state theory. The differential equation describing the transport of the ion or the colloidal particle in the steady state is (38), Note that, when j = 0, the concentration

will obey the usual thermodynamic relation c = co exp ( - C/RT), and also note the close connection between (38) and ( 18). The concentration distribution is found by integrating (37), giving (39), where AG measures the free energy

c e r p (RET) = c ,

- i j x m e x p (RsT) d x

(39)

profile with respect to the bulk solution and c, is the concentration in that bulk solution. At x = 0 the particle will have crossed the barrier into the primary minimum and its concentration in the solution will be zero. For the case of ion transfer we assume that the ions are removed in the second liquid phase by rapid mass transport. Applying this boundary condition to (39), we find that j is given by (40), where k' is expressed by (41).

k' = D / j o mexp(AG/RT) dx It appears in (41) that one has to integrate the whole free energy profile in the region close to the surface for colloidal particles or throughout the interface region for ion transfer. Ruckenstein and Prieve ( 1973) suggested, however, that this was not necessary. One has only to carry out the integration in the vicinity of the transition state located at the maximum in the free energy. The reason for this is illustrated in Fig. 20. Below the transition state on the reactant side the species will be in thermal equilibrium and hence the variation of G is unimportant. Below the transition state on the product side

161

TRANSITION-STATE THEORY REVISITED

BOLTZMANN

Fig. 20 Why we need only worry about the top of the barrier.

the species are committed to reaction and again the variation of G is unimportant. In the region of the transition state we describe G with the simple parabolic expression (42), where x’ = x - xs and, as before, xT 2

RT

RT

describes the distance for AG to change by R T ; this is shown in Fig. 20. We then find that the integral in (41) is given by (43). Substitution in (41) then

rGt)

Jomexp(g) dx z xT exp RT

a,

exp ( - 1’) d l = d/’xT exp

rg)

(43)

gives the expression (44) for the rate constant. The expression for the rate

constant contains the free energy of activation given by the barrier height. It also includes xT, the parameter that describes the curvature at the top of the barrier. The thicker the barrier, the slower is the reaction. Table 3 Typical test of Ruckenstein and Prieve approximation. [Electrolyte]/mmol dm10

5

1 0.1

0.0 1

Percentage error

-2 -2

- 3.5

-5 -5

W. J. ALBERY

162

The success or otherwise of the Ruckenstein and Prieve approximation has been tested by A. L. Smith (unpublished) for the process of colloidal deposition. He compared results from (43) with those obtained by performing the integration of the whole free energy profile. Results are collected in Table 3. It can be seen that the errors are no larger than 5%. We can also calculate f, the fraction of species that are committed to reaction, from (45). Using (39), (43) and ( 4 9 , we obtain (46). Substitution f=1-

f=

C

c, exp(-AGIRT)

(45)

jxm exp(AG/RT) dx JOm

exp(AG/RT) dx

for AG from (42) and integration then gives (47) for x < xl or x' < 0, and

f = f erfc ( -x'/xT)

(47)

(48) for x > xy or x' > 0. Figure 21 shows a plot off against x'/x, in the

f =f [ 1

+ erf(x'/xT)]

(48)

Fig. 21 The fraction of colloidal particles committed to deposit near the transition state. Note that f = f at the transition state and that the swing from f = 0 to f = 1 occurs for GIRT 1.

-

T RANS IT I 0N - STATE T H E 0RY R EVI S IT E D

163

vicinity of the transition state. It is highly satisfactory that f = at the transition state. It is also interesting that the change in f from zero on the reactant side to unity on the product side almost all takes place within R T of the top of the barrier. It is this sharp change that ensures the success of the Ruckenstein and Prieve treatment. The power of the transition-state approach is that one does not have to describe the whole free energy profile, only the free energy of the transition state, and in this case the curvature of the barrier. Conversely we can conclude that kinetic studies only allow us to measure these parameters. So far we have derived an expression for the overall rate constant k . As with the copper system, we should also like an expression for the first-order rate constant k, describing the passage of particles from the secondary minimum over the barrier. The equilibrium constant describing the population of the secondary minimum is given by the same expression (49) as (36), where, as in Fig. 14, xRdescribes the width K,,

= nli2xR exp( - A G I R T )

(49)

of the secondary minimum. Hence we find from (44) and (49) the expression (50), where AGl now describes the free energy difference between the

secondary minimum and the top of the barrier. Note that this rate constant includes both xR,which describes the curvature of the secondary minimum, and xT, which describes the curvature of the top of the barrier. 11

Free energy profiles

In this review we have made use of free energy profiles. These are useful diagrams for depicting relative stabilities of transition states and intermediates. However, they must be used with care. It is common practice to plot these diagrams using standard free energy AGO and a standard state of 1 mol dm-3. This practice can be very misleading when second- and higher-order reactions are involved. In most cases the concentrations of the reactants will be far removed from 1 moldm-3, and hence G may be very different from Go. We have suggested (Albery and Knowles, 1977) that the proper procedure for second- and higher-order reactions is to draw the free energy profile for a stated concentration. The higher-order rate constant is then converted into a first-order rate constant. The free energy difference can then be found using the appropriate frequency factor. The stated concentration is chosen so that

164

W. J. ALBERY

the profile illustrates the kinetics of the system under the conditions of interest. For biochemical systems this may be the in uiuo concentration of a metabolite. For industrial processes it may be the concentration used in practice. For instance, in the solvent extraction of copper we obtain the series of profiles shown in Fig. 18 (Albery and Choudhery, 1988). As the extraction proceeds, the concentration of oxime ligand falls. This alters the profile and different steps become rate-limiting as the extraction proceeds. For this case the profile is illustrating the free energy change of a site at the liquid/liquid interface. Note that the profiles also show the free energy lost in each cycle of the catalytic site. A biochemical example is shown in Fig. 22 (Albery and Knowles, 1986). This shows the free energy profile for proline racemase. The enzyme interconverts D- and in pro line. The enzyme itself exists in two forms, E, and E,, and we have measured the kinetics of the enzyme interconversion reactions through transition states 7 and 8. The free energy block is drawn for the system at equilibrium so that the concentrations of D- and L-proline are equal and no free energy is lost in each cycle; the block illustrates the free energy change of the enzyme active site. The simple first-order rate constants are turned into free energies using the k T / h factor. There are two second-order rate constants in the system: k , , describing the binding of S to

Fig. 22 The free energy diagram for proline racemase, showing the effect on the free energy of the enzyme of increasing substrate concentration c. When c < cD, the system is unsaturated. When cD ic < cp, the system is saturated; and when c > cp, the system is oversaturated, with transition states 7 and 8 rate-limiting.

TRANS ITION-STATE TH EO RY R EVl SITED

165

El, and k - 3 , describing the binding of P to E,. These second-order rate constants are converted into first-order rate constants by multiplying by the substrate concentration. We now have a first-order rate constant that describes what the enzyme experiences. It can be seen that the free energy difference for k , is large at the front of the diagram where the substrate concentration is low, and small at the back of the diagram where the substrate concentration is high. At cD (the “dip switch” concentration) the free energies of El and E,S are equal. Below this concentration the most stable form of the enzyme is free enzyme and the system is unsaturated; above it the most stable form of the enzyme is bound enzyme and the system is saturated. On the right of the diagram the free energies associated with enzyme interconversion do not change with substrate concentration. Hence transition states 7 and 8 climb up with respect to transition state 2. At cp (the “peak switch” concentration) the free energies of the transition states are equal. Below this concentration, the substrate-handling steps are rate-determining; above it the enzyme interconversion steps are rate-determining. Increasing the substrate concentration above cpincreases the free energy difference between the stable bound enzyme and the rate-determining transition states 7 and 8, and indeed we have shown that increasing the substrate concentration above cp actually causes the rate to fall (Fisher et al., 1986; Albery and Knowles, 1987). We have termed this new regime “oversaturation.” In many biochemical systems the concentration of the enzyme is much smaller than that of the substrate. Hence the profiles for the enzyme and for the substrate will be very different. This point is illustrated in Fig. 23, where for a simple one substrate enzyme we see that the substrate concentration is large enough to saturate the enzyme; [S] >> K,. On the other hand, the free enzyme concentration is much smaller than K,. So the profile for the substrate shows that only a small fraction of the substrate is bound to the enzyme. It is hard for the substrate to find enzyme; it is easy for the enzyme to find substrate. We advocate labelling the free energy profiles with those species that have their concentrations fixed as shown in Fig. 23. In the profiles considered so far the reactions have been fairly irreversible. However, many biochemical systems operate close to equilibrium. It is important to realize then that the largest free energy difference that controls the rate of the reaction may not involve an intermediate and a transition state in the same catalytic cycle. As shown in Fig. 24, the largest free energy difference may involve the release of product in one cycle, and the binding of substrate and passage over the rate limiting transition state in the next cycle. We have discussed this point in more detail with respect to our work on proline racemase (Albery and Knowles, 1987). We have also found the same situation in the stripping of copper from the organic phase. The relevant free energy profiles are shown in Fig. 25. If one drew the cycle starting and

W. J. ALBERY

166

$1

ES

$2

EP

$3

S

Fig. 23 Because of the different concentrations of enzyme and substrate, the free energy profiles for the enzyme and substrate are different. The most stable form of the enzyme is bound enzyme, while the most stable form of the substrate is free substrate. ES

EP

E

ES

Fig. 24 For systems near equilibrium the rate-limiting free energy difference may be from an intermediate in cycle 1 to a transition state in cycle 2.

ending with the vacant site, one would not find the correct rate-limiting free energy difference. It can be seen that the proper use of free energy profiles can provide a useful insight into the rate-limiting step and the relative free energies of transition states and intermediates.

TRANSITION-STATE THEORY REVISITED

167

60

A G ~ kJ mol-'

50

/ 10

0

J -

-1 0

:uL*

.

-

Fig. 25 Free energy profiles for stripping copper out of the organic phase. Note that the rate-limiting step, indicated by the double-headed arrow, has the reactant and the transition state in different cycles with respect to the vacant site.

12

Summary

Finally in Table 4 we collect together the expressions for the different rate constants that we have derived for the different reactions considered in this review. It can be seen that transition-state theory, coupled with diffusion, can provide a satisfactory theoretical framework for a wide range of different reactions, providing that the correct frequency factor is used.

168

W. J. ALBERY

Table 4 Summary of different frequency factors. Homogeneous

I Atom and electron transfers with xR Ion transfer with interphase Ion transfer, no 1 interphase Colloidal deposition

kT/h

I1

111

DI.1

D/x:

Heterogeneous -

-

D/7C1l2XTXR

-

-

-

D / R li2xTxR

D / R li2xT

-

-

(k T / h ) x T / d i 2 x R

D/RXTXR

/ RxT

Acknowledgments

I am grateful to Professors J. A. Knowles, R. A. Marcus and A. L. Smith for helpful and stimulating conversations. I am also grateful to Professor Smith for carrying out the calculations reported in Table 3.

Appendix

In this appendix we consider the concentration distribution near the transition state for Case I11 in more detail. For Case I11 the rate constant k is greater than D/x+, a rate constant that describes motion along the x-coordinate. Hence, as soon as the reaction window opens at x = xs, transfer from the reactant surface to the product surface is much more rapid than further progression along the x-coordinate. This will be the case for electron-transfer reactions. But, for an electron-transfer reaction, the fast isoenergetic transfer to the product surface may be followed by an equally rapid return to the reactant surface. Hence we cannot assume that a single k event commits a species to reaction. The rapid exchange-k in one direction and k in the other-equalizes the concentrations of species at xs on the reactant surface and the product surface. Committment to reaction arises by further diffusive motion of product species down the product surface until they are out of reach of the crossing point. In this appendix we consider this situation.

TRANSITION-STATE THEORY REVISITED

169

On the reactant surface the Laplace transformation of (18) for x c xs together with (19) gives (51).Inversion then shows that at x = xt we have (52),

c~=coexp(-~)-%[l

):

x co exp( -

-exp(-z)]

--EZ D

since for Case 111, xs >> xT. In this equation the first term on the right-hand side describes the Boltzmann concentration while the second term describes the depletion from that level as a result of the irreversible removal of species to form products. On the product surface, instead of (19), we have (53), where the reversal

in sign of the xT-term signifies that movement along the x-coordinate now stabilizes the species. The Laplace transformation of (18) with this condition gives (54). E = ct - j / D s s - 1/XT

(54)

Comparison with (51) shows that the denominator on the right-hand side has the form s - l/xT as opposed to s The negative sign would give rise to unacceptable terms in exp(x/xT).Hence the boundary conditions require that (55) holds, giving the acceptable solution c = ct.

+

Substitution of (55) in (52) gives (56). In keeping with our derivation of

transition-state theory, the concentration of transition states is one-half of the Boltzmann concentration, and the fraction committed to reaction is therefore also one-half.

170

W. J. ALBERY

References Albery, W. J. (1975a). In Proton Transfer Reactions (ed. E. F. Caldin and V. Gold), p. 308. Chapman and Hall, London Albery, W. J. (1975b). Faraday Symp. 10, 142 Albery, W. J. (197%). Electrode Kinetics, pp. 95-117. Clarendon Press, Oxford Albery, W. J. (1975d). In Proton Transfer Reactions (ed. E. F. Caldin and V. Gold), pp. 294 and 311. Chapman and Hall, London Albery, W. J. (1980). Ann. Rev. Phys. Chem. 31, 225 Albery, W. J. and Choudhery, R. A. (1988). J. Phys. Chem. 92, 1156 Albery, W. J. and Knowles, J. R. (1977). Angew. Chem. Int. Ed. Engl. 16, 285 Albery, W. J. and Knowles, J. R. (1986). Biochemistry 25, 2572 Albery, W. J. and Knowles, J. R. (1987). J. Theor. Biol. 124, 2529 Albery, W. J. and Kreevoy, M. M. (1978). Adv. Phys. Org. Chem. 16, 87 Albery, W. J., Choudhery, R. A. and Fisk, P. R. (1984). Faraday Disc.,77, 53 Albery, W. J., Kneebone, G. R. and Foulds, A. W. (1985). J. Colloid Interface Sci. 108, 193 Albery, W. J., Fredlein, R. A., O’Shea, G. J. and Smith, A. L. (1990a). Faraday Disc. 90,223 Albery, W. J., Fredlein, R. A., Kneebone, J. R., O’Shea, G. J. and Smith, A. L. (1990b). Colloids Surf., 44, 337 Atkins, P. W. (1982). Physical Chemistry, pp. 978 and 983. Oxford University Press, Oxford Child, M. S. (1967). Faraday Disc. 44, 68 Clint, G. E., Clint, J. H., Corkill, J. M. and Walker, T. (1973). J. Colloid Interface Sci. 44, 121 Dabros, T. G. and van der Ven, T. G. M. (1987). Physicochemical Hydrodynamics, 8, 161. Fisher, L. M., Albery, W. J. and Knowles, J. R. (1986). Biochemistry 25, 2529 Gold, V. (1963). Proc. Chem. Soc., 141 Gold, V. (1983). Pure Appl. Chem. 55, 1281 Hibbert, F. and Long, F. A. (1971). J. Am. Chem. SOC. 93, 2836 Kresge, A. J. and Allred, A. L. (1963). J . Am. Chem. Soc. 85, 1541 Laidler, K. J. (1969). Theories of Chemical Reaction Rates, p. 78. McGraw-Hill, New York Levine, R. D. and Bernstein, R. B. (1987). Molecular Reaction Dynamics and Chemical Reactivity, p. 182. Oxford University Press, New York Marcus, R. A. (1956). J . Chem. Phys. 24,966 Marcus, R. A. (1963). J. Phys. Chem. 67, 853 Marcus, R. A. (1964). Ann. Rev. Phys. Chem. 15, 155 Marcus, R. A. (1965). J. Chem. Phys. 43, 679 Marshall, J. K. and Kitchener, J. A. (1966). J . Colloid Interface Sci. 22, 342 Ruckenstein, E. and Prieve, D. C. (1973). J. Chem. Soc. Faraday Trans. 2 69, 342 Samec, Z., Kharkats, Y. I. and Gurevich, Y. Y. (1986).J . Electroanal. Chem. 204,257 Varennes, S . and van de Ven, T. G. M. (1987). Phys. Chem. Hydrodynam. 8, 161 Varennes, S. and van de Ven, T. G. M. (1988). Colloids Surf: 33, 63

Neighbouring Group Participation by Carbonyl Groups in Ester Hydrolysis KEITH BOWDEN

Department of Chemistry and Biological Chemistry, University of Essex, muenhoe Park, Colchester, Essex C 0 4 3SQ, U K

1 Introduction 171 2 Catalysis by carbonyl groups 172 3 Intramolecular catalysis in ester hydrolysis 173 Exocyclic reactions for aromatic carboxylic esters 174 Exocyclic reactions for aliphatic carboxylic esters 187 Endocyclic reactions for carboxylic esters 191 Carbon acid participation for carboxylic esters 195 Effective molarities 198 Ring size 199 Initiating nucleophile 200 Phosphate and sulphonate esters 200 4 Implications for enzymatic catalysis 202 5 Conclusions and summary 203 References 204

1

Introduction

Neighbouring group participation, especially when it involves intramolecular catalysis, has been widely studied in various systems (Capon, 1964; Capon and McManus, 1976; Bruice and Benkovic, 1966; Kirby and Fersht, 1971). Perhaps the most intensively investigated systems in recent years have been those involving ester hydrolysis (Balakrishnan et al., 1974).These investigations have been widely considered to be of relevance to the study of the mechanisms of enzyme catalysis of related reactions. The elucidation of the factor or factors governing the rate enhancements observed in intramolecularcatalysed reactions enable conclusions to be reached relating to possible or probable efficacy of such factors in enzymic processes. The energetics of 171 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 28 ISBN 0-12-033528-X

Copyright

0 1993 Academic

Press bmired

A / / rights of reproduction in any form reserved

172

K. BOWDEN

neighbouring group participation have been reviewed and the role of entropy, potential energy and strain factors evaluated (Page, 1973). Likely factors of importance in enzymic catalysis are proximity, orientation and steric strain (Dugas and Penney, 1981). The very rapid reactions observed in systems undergoing intramolecular catalysis enable such factors to be evaluated and delineated. The first observation of a significant rate enhancement in the alkaline hydrolysis of an ester by a suitably positioned carbonyl group that was related to prior attack of hydroxide at the carbonyl group was made in 1955 (Djerassi and Lippman, 1955).However, in 1962,a more detailed mechanistic pathway was suggested which involved attack by hydroxide at an o-formyl or benzoyl group, followed by intramolecular nucleophilic attack on a benzoate ester (Newman and Hishida, 1962; Bender and Silver, 1962).

2

Catalysis by carbonyl groups

Carbonyl groups can catalyse reactions in a number of ways. Such groups as keto or formyl substituents are known to be powerful electron-withdrawing groups simply on the basis of their inductive/field and resonance effects (Bowden and Shaw, 1971; Hansch et al., 1991). However, such groups appear to act as neighbouring groups in alkaline ester reactions because of the innate ability of keto and formyl groups for facile reaction with hydroxide to produce powerful nucleophiles. Simple carbonyl compounds, such as ketones and aldehydes, can hydrate by the addition of water, as shown in (1). The equilibrium constant for

hydration ( KH)= [ RR’C(OH),]/[ RR’C=O]) has been either measured or estimated in a number of cases (Bell, 1966; Greenzaid, 1973).Typical, relevant values of KH in water at 25°C are those for acetaldehyde (lS),acetone (ca. 2x and benzaldehyde (ca. 1.1 x lo-’). The hydrates, which are gem-diols, are considerably stronger acids than simple glycols or alcohols. Both the pK,-values of the hydrates and the equilibrium constants for hydration have been correlated with substituent constants, as have the equilibrium constants K , for addition of hydroxide to substituted benzaldehyes (2)(Greenzaid, 1973; Bover and Zuman, 1973).The pK,-values of the hydrates of acetaldehyde and phthalaldehyde in water are 13.57 at 25°C and 12.08 at 20°C respectively (Bell, 1966; Bowden et al., 1979). Unfortunately, for many of the ketones considered in this study no

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

xQxo

+ OH-

2x w H ( o H ) o -

173

(2)

information regarding their hydration is available. In the absence of other information, the ketones are assumed to be inappreciably hydrated. Ketones having hydrogens on an a-carbon can enolize according to (3) R'

R'

I

RCH,C=O

KYI

I

RCH=COH

(3)

(Toullec, 1982). The values of log KL! in water at 25°C have been estimated for acetone as -7.02 and acetophenone as -6.63 (Guthrie, 1979). However, the pK,-values of the ketones ionizing from the ketone to the enolate according to (4) have been measured, using either an acidity function method

R C O C H ~ ~ R C O C H+; H +

(4)

in aqueous dimethyl sulphoxide (DMSO) or a Br~nstedrelation in water at 25"C, to be 21.7,20.0 and 20.9,19.2 for acetone and acetophenone respectively (Toullec, 1982). A complicating feature of studies of carboxylic acid and their corresponding esters, having proximate keto or formyl groups, is the occurrence of ring-chain tautomerism, as in Scheme 1 (Valters and Flitsch, 1985). The rates of conversion of the ring and chain acids ([ 11 and [2]; R' = H ) are rapid. However, both the pseudo and normal esters ([I] and [2], R' = alkyl or aryl) can be isolated in favourable circumstances. The latter esters can also be interconverted by base- or acid-catalysis under suitable conditions. OR'

Scheme 1

3

Intramolecular catalysis in ester hydrolysis

Ester hydrolysis has been a very important system for the investigation of intramolecular catalysis. The types of displacements involved have been

174

K. BOWDEN

classified as either endocyclic or exocyclic, i.e. whether or not the leaving group remains attached to the intermediate originally formed (Kirby and Fersht, 1971). For the systems under study here, an endocyclic reaction has the keto or formyl group attached to the alcohol or phenol group of the ester, whereas for an exocyclic reaction it is attached to the acidic residue. Thus the hydrolysis of 2-formylphenyl acetate undergoing intramolecular catalysis (p. 191) is an endocyclic reaction, while that of methyl 2-formylbenzoate [31 is an exocyclic reaction. The catalysis can originate via carbonyl or carbon acid participation. The carbonyl groups can be formyl or keto. The latter group can be with or without an a-carbon having enolizable hydrogens. The esters can be either carboxylates, phosphates or sulphonates.

EXOCYCLIC REACTIONS FOR AROMATIC CARBOXYLIC ESTERS

Bender and co-workers (Bender and Silver, 1962; Bender et al., 1965) studied in detail the alkaline hydrolysis of methyl o-formylbenzoate [31. Variation

of acetate, phosphate or carbonate buffer concentrations at constant pH and ionic strength had no effect on the rate. The rates of hydrolysis from pH 7.2 to 9.2 showed strict proportionality to the hydroxide concentration. The alkaline hydrolysis of methyl o-formylbenzoate was found to be one of the fastest non-enzymatic hydrolyses of a methyl ester known in aqueous solution at 25°C. The rate was considered to be lo5 times faster than expected on the basis of the substitutent effect expected from the group’s known polar and steric effects. Three mechanistic possibilities were considered on the basis of the participation of the o-formyl group by formation of the adduct. This adduct could lead to catalysis in three ways: (i) by intramolecular nucleophilic catalysis involving intramolecular attack on the ester group by the mono-anion of the hydrate [41, via a tetrahedral intermediate as in Scheme 2;

NEIGH BO U RING G RO U P PARTIC I PATI0 N I N ESTER HYDROLYSIS

c 41

c 51

175

+ OR‘ R

\c/

0

co; C71

+ HOR’ Scheme 2

(ii) by intramolecular general-base catalysis in which the monoanion removes a proton from a water molecule which is itself attacking the ester group as in Scheme 3; and

+ HOR’

+ -OR‘

+ H,O Scheme 3

K. BOWDEN

176

(iii) by intramolecular general acid-hydroxide catalysis in which the hydrate of the benzaldehyde plus a hydroxide ion act to facilitate the formation of the tetrahedral intermediate as in Scheme 4.

R

OH I OH... 0 OR' \C/ \,/ +OH-

t,:

OH R I OH ... 0 \C/ \,/OR' \OH

+ -OR

k;

+

+ HOR'

+ H,O Scheme 4

Bender et al. (1965) favoured the pathway of Scheme 2, mainly on the basis of the analogous process for the morpholine-catalysed hydrolysis also reported by them. The morpholine-catalysed hydrolysis of methyl oformylbenzoate is also extremely facile, and the intermediacy of 3morpholinophthalide [ 141 was demonstrated. Compound [ 141, has a hydroxide-catalysed hydrolysis faster than that of 3-methoxyphthalide (the methyl pseudoester of o-formylbenzoic acid) [ 151. Dahlgren and Schell

( 1967) have investigated the piperazine-catalysed hydrolysis of methyl

o-formylbenzoate, as well as the secondary amine-catalysed lactonization of this ester in dioxan. A Bronsted relation is observed for the latter reaction, giving equal to 0.33 (Henderson and Dahlgren, 1973). Bowden and Taylor (1971b) have made a comprehensive survey of intramolcular catalysis of this type in the alkaline hydrolysis of 2-aroyl and

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

177

2-acylbenzoates. Significant rate enhancements, above that expected for “normal” unassisted ester hydrolysis, occur for a number of methyl 2-acylbenzoates [ 161 as shown in Table 1. The rate ratios, relative to that Me

of methyl benzoate, are corrected for the “expected” values to give the enhanced rate ratio I,. The “expected” values are those estimated on the basis of the known steric and polar effects of ortho-substituents on the alkaline hydrolysis of methyl benzoates (Chapman et al., 1963; Iskander et al., 1966) together with the para-a-values of these acyl groups (Bowden and Shaw, 1971). The significant enhancements (i.e. >40) observed for the 2-formyl, 2-acetyl, 2-propionyl, 2-isobutyryl and 2-phenylacetyl esters clearly indicate hydrolysis involving intramolecular catalysis. The much smaller rate enhancements noted for the 2-pivaloyl and 2-benzoyl esters are not sufficient to be demanding, but only indicative, evidence. However, it was possible to correlate the rates of the 2-acyl esters (Systems 2, 3, 5-8, Table 1) using the Taft relation to give equation ( 5 ) . The E,-values used are those for R obtained log ( k l k , ) = 1.494ES+ 0.553

(5)

from the system RC0,R. No improvement was found using a, or g* in addition to E,. The variation in a, or a* for the substituents under study is, in any case, small. An analogous correlation for the hydration of carbonyl compounds has been reported (Bell, 1966). For K H this correlation is in (6).

These relations both demonstrate the inhibition of formation of the tetrahedral state which can be clearly attributed to steric crowding. Such a correlation as ( 5 ) confirms the attack at the neighbouring carbonyl group and this intramolecular catalysis for all this series. The activation parameters for the alkaline hydrolysis of these esters were also measured and are shown in Table 1. The enthalpies of activation of the 2-formyl, 2-acetyl, 2-propionyl, 2-isobutyryl and 2-pivaloyl esters are exceptionally small. These are

Table 1 Rate coefficients k,, enhanced rate ratios re, ring size and activation parameters for the alkaline hydrolysis of esters in 70% (v/v) aqueous dioxan at 20°C. System 1 2 3 4 5 6 7 8 9 10 11 12 13 14

Ester Methyl benzoate Methyl 2-formylbenzoate Methyl 2-acetylbenzoate Phenyl 2-acetylbenzoate Methyl 2-propionylbenzoate Methyl 2-isobutyrylbenzoate Methyl 2-pivaloylbenzoate Methyl 2-phenylacetylbenzoate Methyl 2-benzoylbenzoate Phenyl 2-benzoylbenzoate Methyl 9-oxofluorene-1-carboxylate Methyl trans-3-benzoylacrylate Methyl cis-3-benzoylacrylate acrylate Methyl trans-3-benzoyl-3-methyl

103k,/dm3 mo1-ls-l 8.50 271 OOO 2000000" 6 450 8 200 2 050 761 21.6 53 1 48.0 89.4* 52.3 2 050 3 390 71.8

fe

-

6 400 15oooo" 760 2 100 240 90 5.1 42 5.6 2.0 -3 -

-

82

AHXlkcal mol-'

ASi/cal mo1-lK-l

Ring sue

Ref.

13 100 4 700

- 23

-

e

-31

5 5 5 5 5 5 5 5 5 5 5

-

8 400 8 900 7 900 8 400 SO00 13 300 13 700 13 500 8 900 12 500 8 500 14 300

- 26 - 24 - 30 - 30 - 39 - 15 - 18 - 18 - 34 - 16 -

27

- 15

e

' e e e

e e e

e 9

e

-

h

5 -

h h

15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Methyl cis-3-benzoyl-3-methyl acrylate Methyl 1-naphthoate Methyl 8-formyl-1-naphthoate Methyl 8-benzoyl-1-naphthoate Methyl 5-formyl-phenanthrene-4carboxylate Methyl 8-acetyl-1-naphthoate Methyl 8-propionyl-1-naphthoate Methyl 8-isobutyryl-1-naphthoate Methyl 3-benzoylpropionate Methyl 4-benzoylbutyrate Methyl 5-benzoylvalerate Methyl benzil-2-carboxylate Ethyl 2-oxocyclohexylacetate Ethyl 2-oxocyclopentylacetate 2-Formylphenyl acetate 4-Formylphenyl acetate

53.6 2.83b 35 500 0.546b 7 470 533 98.3 16.0b 109.5' 68.0' 29.7d 1100Ooo 36.7"' 2026.' 102000000" 16 700"

-

13

400 OOO > 20 > 20 000 000

> 19000 >3500 > 570 -3 -1 1 10000 20 70 6 loo"

--

9OOO 12 900 7 500 16 100 8 400

- 34 - 26 - 26 - 19

10900 13 600 13 300 7000 9 700 10 800 2 800 12 000' 1 500' 3 500" 14 800"

22 17 - 22 -43 - 32 - 30 - 35 - 24' - 56' - 24" - 2"

-

-

26

5

h

0

I

6 6 7

L

6 6 6 5 6 7 5,6 5 5 7 -

1

I

j j

j

k k k

'

m

rn

" "

K. BOWDEN

180

associated with rather large negative entropies of activation. The solvent isotope effects were also examined for the alkaline hydrolysis of the 2-acetyl, 2-benzoyl and unsubstituted esters in 70% aqueous dioxan at 20°C and values of kH20/kD20were found to be 0.68, 0.88 and 0.87 respectively. This evidence can be considered to exclude two of the possible mechanistic paths offered (Bender et al., 1965),i.e. intramolecular general-base catalysis (Scheme 3) and general acid-hydroxide catalysis (Scheme 4). Both the latter would being equal to or greater than involve isotope effects resulting in kH20/kD20 unity. However, such an isotope effect cannot be used as a criterion for either the occurrence of this type of intramolecular catalysis or the detailed mechanistic pathway (see below). Bowden and Taylor (1971b) have studied the hydrolysis in ‘*O-enriched water and showed that methyl 2-benzoylbenzoate suffers about a single enrichment, which is mainly at the keto-carbonyl group. The mechanistic pathway shown in Scheme 2 appears to be operating for those esters of this type having intramolecular catalysis by the carbonyl group. Initial attack of hydroxide occurs at the carbonyl group to give the first tetrahedral intermediate [4], followed by intramolecular nucleophilic attack to form the second tetrahedral intermediate [ 51. The rate-determining step could be in principle k , , k2 or k, in Scheme 2. The overall rate would then be given by k,, K , k , or K , K 2 k , respectively. The possibility of a concerted process covering the formation of [S] has been considered. However, there is no positive or demanding evidence for this, and evidence for discrete tetrahedral intermediates in related systems is overwhelming. The enthalpies and entropies of activation have been tentatively used to assign the rate-determining step. The reactions having small enthalpies of activation and very negative entropies of activation were considered to indicate a rate-determining step composed of the intramolecular attack, and the systems conforming to “normal” activation parameters were considered to have a rate determining step involving addition of hydroxide to the acyl function. This has proved to be too naive. Newman and his co-workers (Newman and Hishida, 1962; Newman and Leegwater, 1968) have studied the effects of 6-methyl and 6-chloro substituents on the rates of alkaline hydrolysis of methyl 2-benzoyl- and 2-acetylbenzoates [ 171. The 6-substituted esters hydrolysed faster than the Me

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

181

corresponding unsubstituted esters, and this was interpreted as a steric effect of the 6-substituent twisting the ester group out of the plane of the aromatic ring. It was considered that this would facilitate the intramolecular step by giving rise to the preferred geometry for intramolecular attack. They were unable to explain why this effect was much greater for the 2-benzoyl than for the 2-acetyl system. Bowden and Taylor (1967,1971a) studied the alkaline hydrolysis of 3'- and 4-substituted methyl 2-benzoylbenzoates [ 181 in 70% Me

(v/v) aqueous dioxan at 30°C and the ionization of the corresponding acids in 80% (w/w) aqueous 2-methoxyethanol at 25°C. The Hammett p-values for the effect of substitution on the alkaline hydrolysis reaction are shown in Table 2. A easonable estimate of the reaction constant ratio p / p o for the transmission o substituent effects from the benzoyl to the carboxylate site is 0.3. This is in good agreement with that found for the ionization reaction, i.e. p / p o = 0.34. However, the reaction constant ratio found for the hydrolysis reaction is very much greater, i.e. p / p o = 0.94, and cannot be due to the transmission of the polar effect to the carboxylate reaction site. However, the magnitude and sign of the reaction constant for hydrolysis in the system [ 181 is in accord with the intramolecular catalytic pathway being the major route for all the 2-benzoyl esters in this study. Thus the p-value for alkaline hydrolysis of methyl benzoates under identical conditions is 2.20 (see Table 2), and the rate-determining step for this reaction is considered to be the formation of the tetrahedral intermediate on the carbonyl group that is directly bonded to the substituted phenyl group. More recently, Bhatt et al. (1979) studied the effects of substitution in both the benzoyl and benzoate rings. They confirmed the earlier finding (Bowden and Taylor, 1971a) for 3'- and 4-substitution in the 2-benzoyl group, with a p-value equal to 2.22 for 70% aqueous acetone at 30°C. The studies of the alkaline hydrolysis of methyl pseudo-2-( 3- or 4-substituted benzoy1)benzoates in 70% aqueous dioxan gave p as ca. 0.72 at 60°C (Bowden and El Kajssj, 1976) and 0.64 a t 30°C (Bhatt et al., 1977). For 4- and 5-substitution in the benzoate group as in [19], a p-value of 2.38 was found (Bhatt et al., 1979) for 70% aqueous dioxan at 30"C, as shown in Table 2, which is comparable to that of 2.20

t

182

K. BOWDEN

Table 2 Hammett reaction constants p and ring size for alkaline hydrolysis of ester systems in 70% ( v / v ) aqueous dioxan.

1 2

3 4 5 6 7

8

9 10 11 12 13 14 15 16 17 18 19

Ring size

System

p

Methyl 3- and 4-substituted benzoates at 30°C Methyl 2-( 3- and 4-substituted benzoy1)benzoates at 30°C Methyl 2-(3- and 4-substituted benzoy1)benzoates at 30°C Methyl 4- or 5-substituted 2-benzoyl benzoates at 30°C Methyl trans-3-( 3- and 4-substituted benzoy1)acrylates at 1°C Methyl cis-3-(3- and 4-substituted benzoy1)acrylates at 1°C Methyl trans-3-( 3- and 4-substituted benzoy1)-3methylacrylates at 1°C Methyl cis-3-( 3- and 4-substituted benzoy1)-3methylacrylates at 1°C Methyl 8-( 3- and 4-substitu ed benzoy1)-1-naphthoates

2.20 2.07

-

2.22"

5

I

at 60°C Methyl 7-substituted-9-oxofluorene-1-carboxylate at 30°C Methyl 3-(3- and 4-substituted benzoy1)propionates at 30°C Methyl 4-( 3- and 4-substituted benzoy1)butyrates at 30°C Methyl 5-( 3- and 4-substituted benzoy1)valerates at 30°C 3-Substituted phenyl 2-acetylbenzoates at 30°C 3-Substituted phenyl 2-benzoylbenzoates at 30°C Methyl 3'- or 4-substituted benzil-2-carboxylates at 21°C Methyl 2-( 3- or 4-substituted phenylacety1)benzoates at 24°C 4-Substituted phenyl 2-formylbenzenesulphonates at 5°C 4-Substituted phenyl 4-formylbenzenesulphonates at 5°C

2.38 0.485

Ref. d

5

d e

e

5 -

f

5

f

-

f

2.10

5

f

1.73

6

9

1.62*

5

h

0.88"

5

1

0.23"

6

1

0.0"

7

1

0.504 1.48 1.52

5

j

5

j

5,6

'

1.43

5

k

0.1'

5

1

2.56 0.759

1.59'

-

1

In 70% aqueous acetone; In 45% aqueous dioxane; In 50% aqueous dioxane. References: Bowden and Taylor (1971a); Bhatt et a!. (1979); Bowden and Henry (1971); Bowden and Last (1973a); Gershom and Raval (1982); Bhatt et al. (1984); 'Anvia and Bowden (1990); li Bowden and Malik (1991); Bhatt and Shashidhar (1986).

@

'

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

183

Me

W

P

h

X

c 191 observed for substituted methyl benzoates described above. It was also considered that the 4-dimethylamin0, 4- and 5-amino esters hydrolysed by a “dual attack” pathway, i.e. both assisted and “normal” ester hydrolysis. This was based on the “deviation” of these latter compounds from the Hammett correlations. This evidence appears suspect as such substitutents are well known to be poorly behaved in such correlations (van Bekkum et al., 1959). The evidence from leaving-group studies in these systems has been very hminating. The rates of hydrolysis of the methyl, isopropyl and diphenylmethyl 2-benzoylbenzoates and benzoates have been compared (Bowden and Taylor, 1971b). The rates of both the pairs of isopropyl and diphenylmethyl esters were almost identical. The effect of these groups on the hydrolysis of benzoate is to reduce the rate by a factor of ca. 16, whereas the effect on the 2-benzoylbenzoate is a factor of only ca. 5. These effects are mainly due to the “bulk” steric effects of these groups. The much smaller reduction observed for the 2-benzoylbenzoate system confirms the occurrence of intramolecular catalysis. If “normal” ester hydrolysis had occurred, the factor would have been expected to increase, not decrease, because of “buttressing” in the crowded ortho-substituted ester. This effect has also been confirmed by others (Bhatt et al., 1979)who studied the corresponding t-butyl esters. The effect of the latter group on the hydrolysis of the benzoate ester in 56% aqueous acetone at 25°C reduces the rate by a factor of ca. 680 compared with the methyl ester; the effect on the 2-benzoylbenzoate in 70% aqueous dioxan at 30°C is ca. 40. Thus the rate enhancement noted in Table 1 for methyl 2-benzoylbenzoate is significantly less than that noted above for the corresponding isopropyl, diphenylmethyl and t-butyl esters. Perhaps the most informative studies (Anvia and Bowden, 1990) of leaving groups are of the alkaline hydrolysis of 3-substituted phenyl 2-acetyl- and 2-benzoyl-benzoates [20]. The Hammett reaction constants p are ca. 0.50 and 1.48 for the 2-acetyl and 2-benzoyl esters respectively, as.shown in Table 2. These were compared with various limiting models in a similar manner to the “effective”charge model (Williams, 1984,1992).This comparison indicates that all simple phenyl esters of the latter types hydrolyse by a mechanism

K. BOWDEN

184

involving initial hydroxide attack at the keto carbonyl group. In the transition state for the 2-acetyl esters, negative charge is being developed on the keto-carbonyl oxygen as the hydroxide develops its bond to the carbonyl carbon. In the transition state for the 2-benzoyl esters, negative charge on the keto-carbonyl oxygen is commencing transfer to the ester carbonyl oxygen as the intramolecular attack proceeds. A very diagnostic study was made by Bowden and Henry (1971), who investigated the alkaline hydrolysis of 3‘- and 4-substituted methyl cis- and trans-3-benzoylacrylates[211 and [22], as well as the corresponding 3-methyl

1w

o

\ 1

‘c=c R

/

,CO,Me \

R=H,Me

H

c211

R=H,Me

system. The “expected” rate ratios relative to methyl acrylate for the parent cis-esters can be estimated with some confidence because of the known steric and polar effects of cis- and trans-3-substituents on the alkaline hydrolysis of methyl acrylate (Bowden, 1966), together with those of ortho-substituents on the alkaline hydrolysis of methyl benzoates (Chapman et al., 1963; Iskander et al., 1966). Significant enhancements are observed for both cis-acrylate esters as shown in Table 1. Moreover, the Hammett reaction

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

185

constants p give very clear evidence. As shown in Table 2, p-values for the alkaline hydrolysis of the trans-esters are ca. 0.6-0.7; whereas for the cis-esters they are ca. 2.1 and 2.6. The reaction constant ratios p / p o for the trans-esters are ca. 0.3, which is in good agreement with a calculated values for simple transmission of ca. 0.3 and a value based on the ionization of the corresponding acids of ca. 0.2-0.3. The reaction constants for cis- and trans-3-substituted acrylic acids have been found to be very similar (Bowden, 1965; Hogeveen, 1964). However, the alkaline hydrolysis of the cis-3benzoylacrylate esters gives a value of p about three times that of the corresponding trans-esters. The former values are in good agreement with those found for the alkaline hydrolysis of 2-( 3- and 4-substituted benzoyl)benzoates, described above. Furthermore, the activation parameters for the cis- and trans-3-benzoyl systems are in stark contrast. As shown in Table 1, the cis-esters have a much smaller enthalpy of activation and more negative entropy of activation than the trans-esters.The activation parameters of the cis-esters are very similar in character to those of the methyl 2-formyl-, 2-acetyl- and related benzoates than that of methyl 2-benzoylbenzoate. Studies using the 1,8-naphthalene and 4,5-phenanthrene templates have been made (Bowden and Last, 1970, 1973a). The rates of alkaline hydrolysis of the methyl 8-formyl-1-naphthoate [23] and 4-formylphenanthrene-4carboxylate [24] are oery fast and have dramatic enhancements over

“expected rates, as shown in Table 1. Both these ester hydrolyses show the characteristic small enthalpies of activation and rather more negative entropies of activation previously observed for related systems. A study of the alkaline hydrolysis of methyl 8-(3’- or 4-substituted benzoy1)-1naphthoates [23] also indicated rate enhancements of a more modest type, but gave a Hammett p-value equal to ca. 1.73 in 70% aqueous dioxan at 60°C, as shown in Table 2. This result is in very good agreement with those for the same reaction of both methyl 2-benzoylbenzoates and cis-3benzoylacrylates (see Table 2). A reaction constant based on transmission of polar effects to the ester carbonyl groups would be expected to be ca. 0.9. Thus the intramolecular catalytic mechanism appears to occur for all the

K. BOWDEN

186

systems described above. However, the alkaline hydrolyses of the methyl 8-acetyl-, 8-propionyl- and 8-isobutyryl-1-naphthoatesoperate by carbon acid participation (see p. 195). The 8-pivalolyl ester [23] cannot use the latter pathway. The t-butyl group appears to inhibit both the “normal” and neighbouring group pathways by its “bulk” steric effect. However, its rate of alkaline hydrolysis would indicate the likelihood of intramolecular catalysis occurring. The alkaline hydrolysis of methyl 9-oxofluorene-1-carboxylate[251 was studied by Bowden and Taylor (1971b). The relatively rapid rate of reaction,

as well as the activation parameters, shown in Table 1, indicated intramolecular catalysis by carbonyl group participation. More recently, Gershom and Raval (1982) ha e studied the rates of alkaline hydrolysis of the corresponding methyl, e hyl and isopropyl esters, as well as that of ethyl fluorene-lcarboxylate. The effect of the isopropyl group on the hydrolysis of the ester, compared with the methyl group, is to reduce the rate by a factor of ca. 12. This is much closer to that observed for the unsubstituted benzoates of ca. 16 than for the 2-benzoylbenzoates of ca. 5. The effects of 7-substitution on the alkaline hydrolysis of ethyl 9-oxofluorene-1-carboxylatehave also been studied (Gershom and Raval, 1984). The 7-substituents can be considered to be a “meta-type” in relation to the carbonyl link. This gives a Hammett p-value for this system of ca. 1.6 in 45% aqueous dioxan at 3WC, as shown in Table 2. This result is more in accord with that of the 2-benzoylbenzoate system undergoing intramolecular catalysis than that expected for “normal” hydrolysis with the transmission of polar effects to the ester group. The alkaline hydrolysis of alkyl esters of anthraquinone-1- and -2carboxylic acids [26] has also been studied (Gore et al., 1971). The rates of

t

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

I a7

hydrolysis of the anthraquinone-2-carboxylatesare significantly more rapid than those of the anthraquinone-1-carboxylates,i.e. a ratio of about 10’ in 70% aqueous dioxan at 40°C. Intramolecular catalysis would only be possible for the latter esters. It is interesting that the pattern of very low enthalpy and rather more negative entropy of activation exists for the methyl anthraquinone- 1-carboxylate, but those of the corresponding 2-esters are “normal”. However, there appears to be no demanding evidence for neighbouring group participation in the alkaline hydrolysis of the 1-esters. Recently, Bowden and Malik (1991 ) have compared the alkaline hydrolysis of substituted methyl benzil-2-carboxylates [271 and 2-phenylacetylbenzoates [28]. The rates of reaction of the benzil system are very fast and the effects 0

c271

C281

of substitution, shown in Tables 1 and 2, indicate intramolcular catalysis oia the a-carbonyl group in this system to give a 5-membered cyclic intermediate.

EXOCYCLIC REACTIONS FOR ALIPHATIC CARBOXYLIC ESTERS

An interesting example of participation from a benzoyl group in the ester [291, having a conformationally rigid benzobicyclo [3.3.11nonene structure, was reported in 1966 (Ghatak and Chakravarty, 1966). The ester [29] was rapidly hydrolysed in 1% methanolic potassium hydroxide at reflux. The esters [30], without a carbonyl group, and [31], isomeric to [29], were

188

K. BOWDEN

recovered unchanged under the same or more drastic alkaline hydrolysis conditions. The suitably orientated ester group in [29] alone can undergo intramolecular attack from the adduct of hydroxide to the benzoyl carbonyl group. The alkaline hydrolysis of methyl l-benzyl-2-acetyl-6-oxo-4a-hydroxy-cisdecahydroisoquinoline-8a-carboxylate [32] proceeds rapidly despite the

H &Me

c311

c321

sterically hindered ester group (Becker and Schneider, 1964; Becker et al., 1 65). This appears to be by intramolecular catalysis from the 6-0x0 group. t was suggested that hydroxide attacks at the less sheltered keto group to give the monoanion of the hydrate and a chair-boat conversion approximates the intramolecular catalytic group to the ester function. The enhancement of rate was ca. lo4, and the enthalpy of activation was significantly reduced compared with the deoxy ester. The 6-0x0 ester in 50% aqueous dioxan, and the deoxy ester, in both 50% and 80% aqueous dioxan, gave kinetics that were first-order in hydroxide, but in 80% aqueous dioxan the kinetics for the 6-0x0 ester were second-order in hydroxide. The alkaline hydrolysis of ethyl 2-oxocyclohexylacetate [331 and 2-oxocyclopentylacetate [ 341 has been investigated (Kemp and Mieth, 1969).

P

C0,Et

PH2 C331

$,,,

C02Et

I

C341

The rates of hydrolysis in 80% aqueous ethanol at 25°C were 60 and 199 times faster than that of ethyl cyclohexylacetate and cyclopentylacetate. These rate enhancements were considered to indicate the occurrence of participation by the neighbouring y-keto group. Furthermore, unusually small enthalpies of activation were found for these esters, as shown in Table 1.

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

189

The importance of the chain length in benzoylalkanoates has been examined by Bhatt et al. (1984). They studied the alkaline hydrolysis of methyl 3-benzoylpropionates, 4-benzoylbutyrates and 5-benzoylvalerates [351 as

ej-(CH2)n-C0,Me

n=2,3,4

C351 well as their desoxo analogues. At 30"C, the relative rates of the 0x0 to the desoxo systems in 70% aqueous acetone are ca. 3.1 (3-propionate), 2.2 (Cbutyrate) and 1.1 (5-valerate). However, there are significant differences in the activation parameters, with a significantly small enthalpy of activation and more negative entropy of activation for the 3-benzoylpropionate system, as shown in Table 1. The effects of meta- and para- substitution in the benzoyl groups have been investigated, and the p-values given are ca. 0.88 (3-propionate), 0.23 (Cbutyrate) and 0.0 (5-valerate), as shown in Table 2. Studies of the ionization of the corresponding substituted 3-benzoylpropionic and 4-benzoylbutyric acids gave p-values close to zero. Thus initial attack apparently occurs at the keto carbonyl group for the 3-benzoylpropionic system and the intramolecular step appears to be rate-determining. There is no keto participation for the 5-benzoylvalerate system and only to a very minor extent, if any, for the 4-benzoylbutyrate system. The alkaline hydrolysis of p-nitrophenyl oxodecanoates [361 has been studied in the absence and presence of cyclodextrins (Cheng et a/., 1985). These flexible keto-esters have either a 4-, 5- or 6-0x0 substituent [36], and their rates of hydrolysis can be compared with unsubstituted ester [37]. The 0

Me-(CH,),

II

-

n-C-(CH,)n-CO, n=2,3,4

I--\

Me-(CH,),-CO, W

o

2

relative rates of hydrolysis at pH 9.21 at 35°C in 50% aqueous DMSO are 4.2 (6-0xo), 2.7 (5-0x0) and 1.9 (4-0x0). For intramolecular catalysis, a five-, six- and seven-membered ring intermediate would be required. The formation

190

K. BOWDEN

of these rings are all considered to be favoured processes (Baldwin, 1976). The intramolecular catalysis appears to be less efficient than those of rigid systems because of significant ring strain and unfavourable entropy terms in the formation of cyclic intermediates of flexible keto esters. The 3oxodecanoate ester hydrolysed by an Elcb pathway, and therefore cannot be compared with the latter esters. In the presence of cyclodextrin, the hydrolysis of the four esters follows Michaelis- Menten behaviour. The rates of hydrolysis of the four esters as inclusion complexes are found to be almost identical. It would appear that the flexible keto-esters are restricted to their linear forms once they are encapsulated, and any participation of the carbonyl groups is completely inhibited. The alkaline hydrolysis of non-enolizable P-keto esters has been recently investigated by Washburn and Cook (1986) who prepared a series of 4-nitrophenyl and phenyl 4-substituted 3-0~0-2,2-dimethylbutyrates [38], 0

II

~-C-CMe,-CO,

c 381

R'

R = Me, CH,X, CHX,, CX, R = Ph, p-nitrophenyl

and studied the kinetics of their alkaline hydrolysis, as well as their '*O-labelling behaviour. The rates of reaction of the esters were measured in 50% aqueous acetonitrile and were first-order in both substrate and hydroxide. For the trifluoromethyl ester, the hydrolysis rate became pH-independent above pH 13, which suggests that the kinetically active species was the monoanion of the hydrate. A Taft equation correlation for the hydrolysis of the 4-nitrophenyl 4-substituted 3-0~0-2,2-dimethylbutyrates gave p* = 1.7, whereas the value of p* for the hydrolysis of the 4-nitrophenyl a-substituted isobutyrates was found to be 0.3. Thus the insertion of the carbonyl group between the substituent and the a-carbon markedly increased the response to the substituent. The p*-values for carbonyl hydration and ionization of corresponding hydrates are 1.7 and 1.4 respectively (Bell, 1966; Greenzaid et al., 1967). The behaviour of the phenyl esters are more complex, but they hydrolysed 102-105 times more slowly than their 4-nitrophenyl counterparts. In contrast, 4-nitrophenyl pivalate hydrolysed 56 times faster than the phenyl ester under these conditions. Esters were prepared with 180-labels both at the 4-nitrophenoxy oxygen and at the keto carbonyl oxygen. The results for the alkaline hydrolysis of these esters were only consistent with intramolcular catalysis via the monoanion of the hydrated ketone. Proton exchange in the monoanion of the hydrate was found to be faster than the intramolecular nucleophilic attack to form the tetrahedral intermediate. The gem-dimethyl effect apparently biases the reaction to the

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

191

formation of a strained four-membered ring intermediate. The extent of the iiucleophilic catalysis is a function of the propensity of the B-carbonyl to form the monoanion of the hydrate. For the 4-nitrophenyl esters, the rate-determining step appears to be the formation of the cyclic intermediate [S]. For the phenyl esters, the poorer phenoxide leaving group apparently causes competition between the intramolecular cyclization and loss of phenoxide as the rate-determining step.

ENDOCYCLIC REACTIONS FOR CARBOXYLIC ESTERS

An early, qualitative observation of the exceptional reactivity of phenyl and naphthyl acetates having proximate formyl groups was made in 1946 (Vavon and Scandel, 1946). A quantitative study of the alkaline hydrolysis of the 2-, 3- and 4-formylphenyl acetates was made by Holleck et al. (1958). The 2-formyl ester [391 was very rapidly hydrolysed compared with the 3- and 0

II

0-C-Me

4-isomers, by a factor of ca. lo4. Furthermore, the former ester also had an exceptionally small enthalpy of activation, as shown in Table 1. The alkaline hydrolyses of 2-formyl and 2-acetylphenyl trans-cinnamates [40] have been studied (Shalitin and Bernhard, 1964). The relatively rapid

R=H.Me

c401

alkaline hydrolysis of the 2-formylphenyl ester was found to be accelerated tenfold by the addition of 0.002moldm-3 KCN. It was considered that intramolecular catalysis occurred and this was probably via the conjugate base of the hydrate of the 2-formylphenyl ester. The authors suggested that the

192

K. BOWDEN

effect of addition of KCN was to give, as a reaction intermediate, the cyanohydrin (as its conjugate base). The 2-acetylphenyl ester hydrolysed ca. 40 times faster than the phenyl ester, but ca. 27 slower than the 2-formylphenyl

ester, An enhanGement of the rate wa6 also found for the 2-aFetylghenyl esters on addition of cyanide. Such enhancements could arise from the more favourable formation of an anionic cyanohydrin adduct. The catalysis then proceeds as observed for amines such as morpholine and piperazine (Bender et al., 1965; Dahlgren and Schell, 1967). A formal pathway is shown in Scheme 5 for an endocyclic system, as was shown in Scheme 2 for an exocyclic pathway. The obvious difference is that, after the formation of the monoanion of the hydrate and the intramolecular attack, the next step in Scheme 5 involves fission to give separate species containing the hydrolysed ester group and the catalytic moiety

k-2flk2

\

K,

0 R’-CO,

+

II

C-R

Scheme 5

The use of mesitoate esters in the elucidation of reaction mechanisms has been pioneered by Burrows and Topping (1969, 1970).This system has been used to suppress the competitive intermolecular reaction by steric “bulk” effects and to detect participation by the identification of the products formed. Under identical conditions (pH 11.28 at 30°C in 9.5% ethanol-water), 2-acetylphenyl mesitoate [41]is hydrolysed 130 times more readily than 4-acetylphenyl mesitoate, clearly indicating intramolecular catalysis. However, the products of hydrolysis provided no clue to the mechanism of

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

193

the participation. Further evidence was obtained by investigation of the solvolysis of 2-acetylphenyl mesitoate in methanolic sodium methoxide. All the evidence showed that the products in anhydrous methanol were the anion of mesitoic acid and the dimethyl acetal of 2-hydroxyacetophenone. This occurs through the pathways shown in Scheme 6. The alkaline hydrolysis of the unhindered ester, 2-acetylphenyl benzoate, was also shown to involve participation by the keto group and was 355 times faster than that of 2-acetylphenyl mesitoate; this which was considered to indicate the importance of the intramolecular step in the determination of the overall rate.

'Me L-4

1

1

i

Scheme 6

K. BOWDEN

194

An important mechanistic study of the alkaline hydrolysis of 2formylphenyl acetate [39] has been made by Klotz and his co-workers (Walder et al., 1978). The hydrolysis was investigated using isotopic tracer studies in H i 8 0 (96% isotropic purity). The hydrolysis of this ester in water at 25°C from pH 6.0 to 8.5 was the sum of water-catalysed and hydroxide-catalysed components. Both the water and hydroxide catalysed rates were greatly enhanced by the 2-formyl group, compared with reference phenyl acetates. The fraction of MeC 1 6 0 2 Hproduced in the hydrolysis was 0.5, and establishes that intramolecular catalysis by the hydrated aldehyde is the mechanism of both the hydroxide and water-catalysed reactions, as shown in Scheme 7. It then follows that the rate determining step for the hydrolysis reaction is the rate of hydration of the aldehyde. Furthermore, cleavage of the acyl intermediate [42] must occur by elimination as shown in Scheme 7, rather than by direct attack of water or hydroxide on the ester carbonyl group. The efficiency of the intramolecular catalysis is dependent on circumventing the back attack of the acyl intermediate by the nucleophilic phenol group, which is achieved by the rapid elimination step (k, in Scheme 7). Similar behaviour has been previ usly noted for other systems (Kirby and Fersht, 1971). Here, this is achie ed in the hydrolysis of this substrate by the rapid elimination of the acyl ntermediate. This pathway has been applied to the catalytic hydrolysis of activated phenyl esters. Thus 2-hydroxybenzaldehyde has been used as a bifunctional catalyst in the

i

pH 0

I1

0-C-Me

O H . @

\

0

0-C-Me e ! ! H /

0

I1

OH

I1

Tkl, k-,

OH 0-C-Me &-H I

I /

MeC0,H Scheme 7

OH

[42]

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

195

hydrolysis of nitrophenyl esters (Johnson et al., 1978). The phenoxide acts as a nucleophile, attacking the ester and forming an acylated intermediate, while the proximate hydrated formyl group acts as a nucleophile to deacylate the intermediate.

CARBON ACID PARTICIPATION FOR CARBOXYLIC ACIDS

The first observation of carbon acid participation in the alkaline hydrolysis of esters was reported in 1970 by Bowden and Last (1970, 1973b). The 8-acetyl-, 8-propionyl- and 8-isobutyryl-l-naphthoates [43] reacted rapidly in aqueous dioxan containing an excess of base. However, the immediate products of these reactions are the phenalene-1,3-diones [44]. The two

R =Me, CH,Me, CHMe,

C431

4e; Me,Me

diones, the unsubstituted and 2-methyl compounds, which can ionize in base, are relatively stable in this anionic form in alkaline solution. By contrast, the 2,2-dimethyl dione, which cannot ionize in base, hydrolyses relatively rapidly, with a ratio of the rate of cyclization to that of ring fission of ca. 25 at 40°C in 70% aqueous dioxan. This general pathway is shown in Scheme 8. The cyclization reaction of methyl C2:2:2'2H,I-8-acetyl-f -naphthoate was ca. 6 times slower than that of the 8-acetyl ester itself at 3VC, and this clearly identifies the ionization step as the rate-determining step. For the trideuterioester, it was possible to detect a minor reaction not proceeding via the dione. Although this minor reaction does not appear to be a very significant contribution to the total hydrolysis, the trideuterioester is the most likely substrate for such an observation since this particular acyl group has both a comparatively slow ionization and a small steric bulk. This minor

K. BOWDEN

196

H

I /I\&

-

C

0 0

\&

OMe

R'

K,

+OH-

+ k1

/I\,

C

\& OMe

0 0

R'

+H20

k-1

R \/

\c/o-

if R = H

R'

LJ +-OH + MeOH

-H;/+H+

I

u

~~

LJ\OMe + H,O

k x

R

R

R

R' H-C

-0,c

"

\c/

0

Scheme 8

process appears to be enhanced in rate and to proceed via carbonyl group participation. The base-catalysed cleavage of 2,2-dimethyl-2,3dihydrophenalene-1,3-dione is first-order in substrate and hydroxide. However, the anions of B-diketones are known to be resistant to hydrolysis, as was found for the 2-methyl and 2,3-dihydrophenalene-1,3-diones. It is apparent that steric bulk and stereochemical control of mechanism operates in the alkaline hydrolysis of methyl 8-acyl-1-naphthoates. The proximity and favourable orientation of the carbonyl group at the 8-position facilitates intramolecular catalysis from this group. However, the formation of the tetrahedral intermediate at the 8-acyl carbonyl group has distinct

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

197

spatial requirements. When the acyl group is comparatively acidic, a more favourable process is the ionization of this weak carbon acid, which produces a strong internal nucleophile. This rapidly attacks the ester carbonyl-carbon intramolecularly, assisted by the release of steric interactions in forming the cychc infermedafe. Me fti yf 2-acefyf6enzoafe a n d refa Cedesfers eviden tfy fack both the severe steric crowding and favourable orientation for attack of the carbanion on the ester group. However, in methanol and methanol-DMSO containing methoxide, such esters will undergo facile intramolecular cyclization of this type (Bowden and Chehel-Amiran, 1986). Furthermore, the reversible equilibration of normal and pseudo esters occurs rapidly in methanol containing methoxide (Bowden and El Kaissi, 1977). The former reaction involves neighbouring group participation by the carbon acid and the latter by the carbonyl group. The use of mesitoic acid esters has again been successfully employed by Burrows and Topping ( 1975) in the elucidation of intramolecular carbon acid participation. Under basic aqueous conditions, 2-acetylphenyl mesitoate [411 hydrolyses to yield mesitoic acid and 2-hydroxyacetophenone, reacting with intramolecular catalysis via the monoanion of the ketonic hydrate (see p. 192). However, in 47.5% aqueous ethanol containing potassium hydroxide, the reaction products from 1-acetyl-2-naphthyl mesitoate [45] were found

_,

to be 65% of the products of hydrolysis, mesitoic acid and l-acetyl-2naphthol, and 35% of the product of isomerization, a j?-diketone, i.e. I-( 2-hydroxy-l-naphthyl)-3-mesitylpropane1,3-dione [46]. Both the mesitoate

1461 esters react with methoxide in methanol to give the corresponding dimethyl acetals, proceeding via intramolecular nucleophilic participation of the

198

K. BOWDEN

methoxide adduct of the ketone. However, 2-acetylphenyl mesitoate with potassium t-butoxide in t-butyl alcohol isomerizes to form the corresponding B-diketone via intramolecular nucleophilic participation of the carbon acid anion. The greater steric hindrance to keto carbonyl addition in the 1-acetyl-2-naphthyl mesitoate [45], compared with the 2-acetylphenyl ester [41], partially switches the type of participation observed. Participation by the enolate anion could, in principle, act via either the anionic carb n or oxygen. The former, however, is known to be considerably the more n leophilic centre. The latter would give rise to a lactone, which would be r pidly hydrolysed under the usual conditions of excess base to give the identical final product. There is no evidence for the occurrence of such a process.

3

EFFECTIVE MOLARITIES

The effective molarity (EM) has been defined formally by Kirby (1980) as the concentration of the catalytic group required to make the intermolecular reaction proceed at the rate observed for the intramolecular process. Mechanisms of both the intramolecular and intermolecular reactions must be known and have been shown to be the same type, e.g. nucleophilic, general base, etc. Whilst it was considered difficult to obtain very accurate measurements of EM, these are not essential for many purposes. Kirby ( 1980), in a comprehensive review, did not consider neighbouring group catalysis in the hydrolysis of esters by carbonyl groups. The likely reasons would appear to be that no results were available for the intermolecular reaction and the concentration of anionic catalytic group in the intramolecular reaction is unknown. An exception to the former is the hydrolysis of p-nitrophenyl acetate catalysed by the monoanion of chloral hydrate (Gawron and Draus, 1958). A preliminary approach to the calculation of EM-values can now be made. First, the rates of the normal alkaline hydrolysis of the parent esters can be used to simulate the intermolecular reaction, i.e. with hydroxide anion used as a model for the monoanion of the hydrated ketone or aldehyde. The concentration of the latter species in the intramolecular reaction can be assumed to be that of the corresponding ketone or aldehyde and can be calculated from the results for the hydration and the ionization of hydrates. However, results for relevant aldehydes and ketones are not usually available. Thus only the result for benzaldehyde is known, with K , in water at 25°C equal to 0.18dm3mol-' (Greenzaid, 1973). The EM value can then be calculated and equals k , (ester having formyl group)/K,k, (parent ester). The alkaline hydrolysis of methyl 2-formylbenzoate and 2-formylphenyl acetate give EM values, in aqueous dioxan, of 1.8 x lo5 and 3 x lo5 mol dm-3

N E I G H BOU R I N G G ROU P PARTI C I PATION IN ESTER HYDROLYSIS

199

respectively. The magnitude of such values can be used to classify the mechanisms; on this criterion, the reactions are clearly nucleophilic in type, rather than general-base catalysis (Kirby, 1980).The above treatment assumes that the formation of the adduct is not the rate-determining step. However, the formation of the adduct for benzaldehyde has been sometimes considered to be the rate-determining step. If the above assumption is correct, it is possible to compare the effectiveness of this group with others which operate as nucleophilic neighbouring groups. The EM-values calculated above are comparable to those found for several other nucleophilic neighbouring groups (Kirby, 1980). However, the rate coefficients corrected for incomplete carbonyl hydration ( k , / K , ) are 1.5 x lo9 s-l (methyl 2-formylbenzoate) and 5.6 x 105s-' (Zformylphenyl acetate) at 20°C in aqueous dioxan and are very much greater than those of any other system reviewed by Kirby, except those for the base-catalysed lactonization of hydroxy esters, i.e. the corresponding anions, in water, of phenyl 4-hydroxypropionate (5.5 x lo5s-' at 29.8"C)and of ethyl 2-hydroxymethylbenzoate(1.25 x lo6 s-l at 30°C). A comparison of the effect of the anionic hydrated formyl group with that of a carboxylate group can be estimated and shows the former to be ca. lo8 times more reactive. This must be due to the significantly greater nucleophilicity of the former group. Such considerations are impossible for carbanion participation since the rate-determining step is the ionization process itself.

RING SIZE

The ease of ring-formation in neighbouring group participation has been studied for a wide variation of ring types (Capon and McManus, 1976). In general, these show that the main factor determining the ease is the size of the ring being formed, i.e. 3 > 4 < 5 > 6 > 7 (Kirby, 1980). The situation for the studies under review is not simple, however, since many of the rings formed are composed of portions of rigid aromatic or alicyclic ring structures. Table 1 shows the ring sizes, 5, 6 or 7, required for the formation of the intermediate arising from neighbouring group participation. Furthermore, in [381 the ring size is 4. Only in the hydrolysis of methyl benzil-2-carboxylates could the system form either a five- or six-membered ring. The evidence favours a five-membered ring structure, as found in the ring-chain tautomerism of the corresponding acid system (Bowden and Malik, 1991). It is very clear from the study of the hydrolysis of the flefible methyl propionates, butyrates and valerates, having an o-benzoyl group, that the order of ease of ring formation is 5 >> 6 > 7 (Bhatt et al., 1984). In other systems, however, with greater rigidity and/or a more ideal stereochemistry

200

K. BOWDEN

for intramolecular catalysis, five-, six- or even seven- or four-membered rings can be successfully and efficiently formed (see Tables 1 and 2).

INITIATING NUCLEOPHILE

The initiating nucleophile in the vast majority of these studies is the hydroxide anion. However, in principle, any nucleophile can add to the keto or formyl group to give rise to an anionic intermediate, which then could act as an intramolecular nucleophile and effect hydrolysis of the ester. Their relative effectiveness will depend on two factors: the relative extent of formation and the nucleophilicity of the adduct. The nucleophiles that have been investigated are hydroxide, cyanide, morpholine and piperazine. The only quantitative comparison available is that of hydroxide, morpholine and piperazine, which 1 (Bender et al., 1965; Dahlgren are effective in the order of ca. lo2: and Schell, 1967). For morpholine and piperazine this is as expected on the basis of their relative basicities. However, the expected order of increasing formation of the adducts would be cyanide > nitrogen bases > hydroxide (Hine, 1971). At this time, these results cannot be analysed further, but more work on the systems could enable the structural dependence and reactivity to be elucidated.

PHOSPHATE AND SULPHONATE ESTERS

Just as intramolecular catalysis has been observed in the hydrolysis of carboxylate esters, a variety of neighbouring groups can participate in the hydrolysis of phosphates. The alkaline hydrolysis of dimethyl phosphoacetoin [47] has found to be ca. 2 x lo6 times faster in water at 25°C than that of

trimethyl phosphate (Ramirez et al., 1962). It was suggested that catalysis occurred either via an enol or participation by the carbonyl group. The enhancement of the solvolytic displacement of p-nitrophenol from a phosphonate ester by a neighbouring ketonic carbonyl group has been

20 1

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

reported (Lieske et al., 1966). p-Nitrophenyl phenacyl methylphosphonate [48] hydrolyses very rapidly in 5% aqueous dioxan at 25"C, the reaction 0

0

II

Ph-C-CH,-O-P-O

I

Me ~481 being ca. 9OOO times faster than that of methyl p-nitrophenyl methylphosphonate under identical conditions. Both intramolecular catalysis via the hydrated ketone and the enol were considered, but evidence against the enolization mechanism has been given. Addition of KCN had only a very small effect upon the rate. The most likely path appears to be by addition of hydroxide to the carbonyl group, followed by intramolecular attack of the anion of the hydrate. Both the systems above would involve formation of a five-membered strained ring, formed between one apical and one equatorial position (Dugas and Penney, 1982). During the hydrolysis process, groups enter and leave from apical positions alone. There appear to be no problems 1 completing such operations in these systems. A recent study has conclusively extended the observations of intramolecular catalysis by carbonyl groups to the hydrolysis of aryl 2-formylbenzenesulphonates [49] (Bhatt and Shashidhar, 1986). The alkaline hydrolysis of

;"

4-substituted phenyl2- and 4-formylbenzenesulphonateshas been investigated. The 4-nitrophenyl esters gave a rate ratio of ca. 4 x lo6 for the 2- to the 4-formyl system in aqueous acetone at 256°C. The occurrence of carbonyl group participation was considered to be clearly established. The p-values for the 2- and 4-formyl systems in 50% aqueous dioxan are ca. 0.1 at 5°C and 1.59 at 46.5"C respectively. It was considered that the very small value of p observed for the 2-formyl system indicates that the cleavage of the leaving group is fast compared with the attack of hydroxide on the carbonyl group.

202

4

K. BOWDEN

Implications for enzymic catalysis

Studies of neighbouring group participation have given valuable information on factors such as proximity, strain and orientation that have been suggested to be potential sources of the special catalytic power of enzymes (Dugas and Penney, 1981). In an early report, Shalitin and Bernhard (1964) suggested that the rate enhancements observed in alkaline hydrolysis of 2-formylphenyl trans-cinnamate might be pertinent to the mechanism of enzymic hydrolysis of acyl derivatives, where a serine hydroxy group is a constituent of the catalytically active site. Bender et al. (1965) considered that the facile hydroxide and morpholinecatalysed hydrolyses of methyl 2-formylbenzoate around neutrality exhibit interesting formal analogies to the facile catalysis of ester hydrolysis by enzymes around neutrality. In chymotrypsin the acyl group of the substrate becomes covalently attached to the enzymes during the catalytic process. In the catalytic ester hydrolysis described above, the acyl group of the substrate transfers to a new group in the catalytic process. In chymotrypsin, a hydroxy group of the enzyme is the nucleophile to which the acyl group becomes attached, whereas, in the catalytic ester hydrolysis, a hydroxy group of the hydrate forms a new acyl-oxygen bond. The advantage of neighbouring group participation by the hydrated formyl or keto group in ester hydrolysis is that the latter group comes nearer to a realistic mimic of the hydroxy group of chymotrypsin than many other model systems (Dugas and Penney, 1981). Both the active site of chymotrypsin and the neighbouring group system are considered to be alkoxide or alkoxide-like in their active role. In chymotrypsin, the nucleophile is revealed by a charge-relay system, whereas, in neighbouring group participation, the nucleophile is generated by addition of hydroxide. An intermolecular reaction of this type has been previously described (Gawron and Draus, 1958). The overall hydrolysis of p-nitrophenyl acetate was catalysed by the monoanion of chloral hydrate via the acetylated hydrate and regeneration of the hydrate. The distinct possibility that a hydrated carbonyl group could behave in such a similar manner to that of the active site of a hydrolytic enzyme was considered by Bruice and Benkovic (1966). The study of both carbonyl and carbon acid participation in ester hydrolysis has been used by Bowden and Last (1971) to evaluate certain of the factors suggested for important roles in enzymic catalysis. A first model concerns a comparison of the three formyl esters and shows that the proximity of the formyl to the ester group and internal strain increase in passing along the series, 1,2-benzoate, 1,8-naphthoate and 4,5-phenanthroate. The very large rate enhancements result from the proximity of the internal nucleophile once formed and from internal strain. Strain is increased or induced by the primary

NEIGHBOURING GROUP PARTICIPATION IN ESTER HYDROLYSIS

203

addition process, but is relieved as the intramolecular attack proceeds to form a cyclic intermediate. Studies of ring-chain tautomerization in acyl carboxylic acids have shown that the ring (cyclic) tautomer is favoured when internal strain arising from steric interactions increases (Valters and Flitsch, 1985).Thus the reaction is facilitated by the induction of strain derived from the reaction between substrate and hydroxide. The model system described employs the addition of hydroxide to the formyl group to simulate the nucleophilic activity of the hydroxyl group in the unique serine residues in such enzymes. A second model concerns the effect of a proximate acetyl group on alkaline ester hydrolysis. Methyl 2-acetylbenzoate rapidly hydrolyses by neighbouring carbonyl participation, whereas methyl 8-acetyl1-naphthoate employs neighbouring carbon acid participation. The dichotomy displayed by the two esters clearly indicates the specific control of reaction afforded by the orientation and environment of the assisting group. For the 8-acetyl-l-naphthoate, the carbanion is generated with highly favourable orientation for substitution and is closely proximate to the ester carbonylcarbon. However, catalysis by the carbonyl group has distinct spatial requirements, unlike the carbon-acid process. The “crowding” is very severe in the 8-acetyl-l-nap thoate, compared with the 2-acetylbenzoate. Further, the 2-acetylbenzoate does not have the favourable orientation and close proximity for the carbon-acid path to be favoured. The carbon-acid catalytic model can be considered to simulate the carbon-carbon bond formation catalysed in aldolases and related enzymes, especially those enzymes catalysing Claisen condensations or retrocondensations ( Walsh, 1979). The latter enzymes differ from aldolases in that the nucleophilic component is the a-carbanion of an ester, which attacks a variety of carbonyl-containing substrates. Model systems for such enzymes are not common, and the model described above clearly indicates the importance of orientation and steric strain factors in such reactions.

!

5

Conclusions and summary

Intramolecular catalysis has been conclusively demonstrated in a diverse range of esters undergoing alkaline hydrolysis and having suitably orientated acyl groups. This gives rise to significantly favourable pathways involving neighbouring group participation by either the monoanion of the hydrate of the acyl carbonyl group-the usual mode-or the carbanion derived from a keto group having a-hydrogens. These pathways can “be demonstrated in a number of different ways. Rate enhancements over expected rates of hydrolysis can be demonstrated. Capon ( 1964) has considered an increase in rate of at least fivefold and preferably of fiftyfold to be necessary for

204

K. BOWDEN

identification of such an effect. Unless the enhancements are comparatively large, these must be treated with care as a criterion of mechanism. Because of the often large difference in the enthalpies of activation of model and reference systems, such comparisons can be very dependent on the reference temperature selected. Activation parameters themselves can be significant indicators, especially small enthalpies of activation. Various other criteria can be used in a diagnostic manner. Important among these are polar and steric substituent effects, leaving-group effects and "0-labelling studies. Tables 1 and 2 shows the rate coefficients,enhanced rate ratios, activation parameters and Hammett reaction constants for the ester systems discussed here. For carbonyl group participation, the rate-determining step can be either the formation of the monoanion of the hydrate or the intramolecular nucleophilic attack. This appears to depend on the system under study: the particular neighbouring group, leaving group and template involved. For systems involving formyl group participation, the hydration process has been sometimes considered to be the rate-determining step, while for those involving benzoyl group participation, the rate-determining step often appears to be the intramolecular process. References Anvia, F. and Bowden, K. (1990). J. Chem. SOC.,Perkin Trans. 2, 1805 Balakrishnan, M., Rao, G. V. and Venkatasubramanian, N. (1974). J. Sci. Znd. Res. 33, 641 Baldwin, J. E. (1976). J . Chem. SOC., Chem. Commun. 734 Becker, H. G. 0. and Schneider, J. (1964). Wiss. Z . Chem., Leuna-Merseburg 6, 278 [Chem. Abs. 62, 10 408 (19651 Becker, H. G. O., Schneider, J. and Steinleitner, H. D. (1965). Tetrahedron Lett., 3761 Bell, R. P. (1966). Adu. Phys. Org. Chem. 4, 1 Bender, M. L. and Silver, M. S. (1962). J . Am. Chem. SOC.84, 4589 Bender, M. L., Reinstein, J. A., Silver, M. S. and Mikulak, R. (1965). J. Am. Chem. SOC. 81,4545 Bhatt, M. V. and Shashidhar, M. S. (1986). Tetrahedron Lett., 2165 Bhatt, M. V., Rao, K. S. and Rao, G. V. (1977). J . Org. Chem. 42, 2697 Bhatt, M. V., Rao, G. V. and Rao, K. S. (1979). J. Org. Chem. 44,984 Bhatt, M. V., Ravindranathan, M., Somayaji, V. and Rao, G . V. (1984). J. Org. Chem. 49, 3170 Bover, W. J. and Zuman, P. (1973). J. Chem. SOC.,Perkin Trans. 2, 786 Bowden, K. (1965). Can. J. Chem. 43, 3354 Bowden, K. (1966). Can. J . Chem. 44, 661 Bowden, K. and Chehel-Amiran, M. (1986). J. Chem. SOC., Perkin Trans. 2, 2031, 2035, 2039 Bowden, K. and El Kasissi, F. A. (1976). J. Chem. SOC., Perkin Trans. 2, 526 Bowden, K. and El Kaissi, F. A. (1977). J. Chem. SOC., Perkin Trans. 2, 1927 Bowden, K. and Henry, M. P. (1971). J. Chem. SOC. ( B ) , 156

NEIGH BO U R I N G G ROU P PARTI C I PAT1ON IN ESTER HYDROLYSIS

205

Bowden, K. and Last, A. M. (1970). J . Chem. SOC., Chem. Commun., 1315 Bowden, K. and Last, A. M. (1971). Can. J . Chem. 49, 3887 Bowden, K. and Last, A. M. (1973a). J . Chem. SOC., Perkin Trans. 2, 345 Bowden, K. and Last, A. M. (1973b). J . Chem. SOC., Perkin Trans. 2, 351 Bowden, K. and Malik, F. P. (1992). J . Chem. SOC., Perkin Trans. 2, 5 Bowden, K. and Shaw, M. J. (1971). J . Chem. SOC. ( B ) , 161 Bowden, K. and Taylor, G. R. (1967). J . Chem. SOC., Chem. Commun., 1112 Bowden, K. and Taylor, G. R. (1971a). J . Chem. SOC. ( B ) , 145 Bowden, K. and Taylor, G. R. (1971b). J . Chem. SOC. ( B ) , 149 Bowden, K., El Kaissi, F. A. and Nadvi, N. S. ( 1979).J . Chem. SOC.,Perkin Trans. 2,642 Bruice, T. C. and Benkovic, S. J. (1966). Bioorganic Mechanisms, Vol. 1. Benjamin, New York Burrows, H. D. and Topping, R. M. (1969). J . Chem. Soc., Chem. Cornmun., 904 Burrows, H. D. and Topping, R. M. (1970). J . Chem. SOC. ( B ) , 1323 Burrows, H. D. and Topping, R. M. (1975). J . Chem. Soc., Perkin Trans. 2, 571 Capon, B. (1964). Quart. Rev. 18, 45 Capon, B. and McManus, S. P. (1976). Neighbouring Group Participation, Vol. 1. Plenum Press, New York Chapman, N. B., Shorter, J. and Utley, J. H. P. (1963). J . Chem. Soc., 1291 Cheng, X-E., Hui, Y-Z., Gu, J-H. and Jiang, X-K. (1985). J . Chem. Soc., Chem. Commun., 71 Dahlgren, G. and Schell, D. M. (1967). J . Org. Chem. 32, 3200 D j e r e C . and Lippman, A. E. (1955). J . Am. Chem. SOC.77, 1825 Dugas, H. and Penney, C. (1981). Bioorganic Chemistry. A Chemical Approach to Enzyme Action, Chaps. 3 and 4. Springer-Verlag, New York Gawron, 0. and Draus, F. (1958). J . Am. Chem. SOC. 80, 5392 Gershom, H. R. and Raval, D. A. ( 1982). J . Inst. Chemists (India) 54, 23 (Chem. Abs. 96, 198 784) Gershom, H. R. and Raval, D. A. (1984). J . Inst. Chemists (India) 56, 33 (Chem. Abs. 101, 109 973) Ghatak, U. R. and Chakravarty, J. (1966). J . Chem. SOC.,Chem. Commun., 184 Gore, P. H., Rahim, A. and Walter, D. N. (1971). J . Chem. SOC.( B ) , 202 Greenzaid, P. (1973). J . Org. Chem. 38, 3164 Greenzaid, P., Luz, Z., and Samuel, D. (1967). J . Am. Chem. SOC. 89, 749 Guthrie, J. P. (1979). Can. J . Chem. 57, 797 Hansch, C., Leo, A. and Taft, R. W. (1991). Chem. Reo. 91, 165 Henderson, G. H. and Dahlgren, G. (1973). J . Org. Chem. 38, 754 Hine, J. (1971). J . Am. Chem. SOC.93, 3701 Hogeveen, H. (1964). Rev. Trav. Chim. 83, 813 Holleck, L., Melkonian, G. A, and Rao, S. B. (1958). Naturwissenschafen 18,438 Iskander, Y., Tewfik, R. and Wasif, S. (1966). J . Chem. SOC. (a),424 Johnson, R. S., Walder, J. A. and Klotz, I. M. (1978). J . Am. Chem. SOC.100, 5159 Kemp, K. C. and Mieth, M. L. (1969). J . Chem. SOC., Chem. Commun., 1260 Kirby, A. J. and Fersht, A. R. (1971). Prog. Bioorg. Chem. 1, 1 Kirby, A. J. (1980). Ado. Phys. Org. Chem. 17, 183 Lieske, C. N., Miller, E. G., Zeger, J. J. and Steinberg, G. M. 41966). J . Am. Chem. SOC. 88, 188 Newman, M. S. and Hishida, S. (1962). J . Am. Chem. SOC.84, 3582 Newman, M. S. and Leegwater, A. L. (1968). J . Am. Chem. SOC.90,4410 Page, M. I. (1973). Chem. SOC.Rev. 2, 295

206

K. BOWDEN

Ramirez, F., Hansen, B. and Desai, N. B. (1962). J. Am. Chem. SOC.84, 4588 Shalitin, Y. and Bernhard, S. A, (1964). J . Am. Chem. SOC.86, 2291 Toullec, J. (1982). Adu. Phys. Org. Chem. 18, 1 Valters, R. E. and Flitsch, W. (1985). Ring-Chain Tautomerism. Plenum Press, New York van Bekkum, H., Verkade, P. E. and Wepster, B. M. (1959). Rec. Trau. Chim. 78,815 Vavon, G. and Scandel, J. (1946). Compt. Rend. 223, 1144 Walder, J. A., Johnson, R. S. and Klotz, I. M. (1978). J . Am. Chem. SOC.100, 5156 Walsh, C. (1979). Enzymatic Reaction Mechanisms, Chap. 23. W. H. Freeman, San Francisco Washburn, W. N. and Cook, E. R. (1986). J . Am. Chem. SOC.108, 5962 Williams, A. (1984). Acc. Chem. Res. 17,425 Williams, A. (1992). Adu. Phys. Org. Chem. 27, 1

Electrophilic Bromination of Carbon-Carbon Double Bonds: Structure, Solvent and Mechanism MARIE-FRANCOISE RUASSE Institut de Topologie et de Dynamique des Systkmes de l’universitk Paris 7, associk au C N R S - U R A 34, 1 rue Guy de la Brosse, 75005 Paris, France

Introduction 208 Methods for obtaining reliable bromination rate constants 21 1 Kinetic rate equations 212 Kinetic techniques for bromination 214 Bromine-olefin charge transfer complexes as essential intermediates in bromination 216 The ionic intermediates: bridged bromonium ions or open B-bromocarbocations 220 Experimental observations 221 Theoretical calculations 224 Kinetic data and bromine bridging in transition states and intermediates 225 Product data and bromine bridging from stereo- and regio-chemistry 234 Kinetic substituent effects 243 Polar effects of alkyl groups 243 Steric effects of alkyl groups 246 Kinetic effects of aryl substituents 252 Selectivity relationships and transition-state shifts in arylolefin bromination 256 Early transition states in enol ether halogenation 263 Solvent effects and solvation in bromination 267 Kinetic solvent isotope effects 268 The Y,, scale for bromination 270 Values of mBr in alkene bromination: nucleophilic and electrophilic assistance by protic solvents 272 Values of mBr and transition-state shifts in the bromination of conjugated olefins 274 Bromine-catalysed bromination in non-protic and halogenated solvents 276 Solvation, the driving force of electrophilic bromination 278 The reversible formation of bromonium ions 279 Return in halogenated solvents 280 Return in protic solvents 282 207 ADVANCES IN PHYSICAL ORGANIC CHEMISTRY VOLUME 28 ISBN 0-12-033528-X

Copyright 0 1993 Academic Press Limited A / / rights of reproduction in any / o m reserved

M:F.

208

RUASSE

8 Concluding remarks 285 Acknowledgements 286 References 287

1

Introduction

The electrophilic bromination of ethylenic compounds, a reaction familiar to all chemists, is part of the basic knowledge of organic chemistry and is therefore included in ever chemical textbook. It is still nowadays presented as a simple two-step, tra s-addition involving the famous bromonium ion as the key intermediate. T is mechanism was postulated as early as the 1930s by Bartlett and Tarbell (1936) from the kinetics of bromination of transstilbene in methanol and by Roberts and Kimball(l937) from stereochemical results on cis- and trans-2-butene bromination. According to their scheme (Scheme 1), bromo-derivatives useful as intermediates in organic synthesis

i

\ C=C / / \

+ Br,

slow

Br /+\

,C-C

/

\

\

+ Br-

fast

I

-C-Br I Br-C-

I

fast

I I NU-CI

+Nu-

-C-Br

Scheme 1

can be obtained with a high degree of diastereoselectivity from olefin and bromine, two readily available reagents. However, bromination is rarely interesting from a synthetic viewpoint (Okabe et al., 1982; Ueno et al., 1982; Rodriguez et al., 1984; Castaldi et al., 1986) since it is not as selective as Scheme 1 would suggest. This contrast between practical and conceptual approaches probably arises from the fact that detailed mechanistic studies are rather recent. Moreover, most of these studies are related to the first, ionization, steps whereas data on the last, product-forming, step are still scarce. Consequently, it remains difficult, or even impossible, to control the stereo-, regio- and chemo-selectivity of this addition. Despite much work (Sergeev et al., 1973; De la Mare, 1976; Schmid and Garratt, 1977), it has taken a long time to obtain reliable rate constants

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

209

readily interpretable in terms of mechanism. This is because bromination is both a very fast reaction, the kinetics of which are not easy to monitor in the absence of specially devized techniques, and a variable reaction, highly sensitive to the double bond substituents and to the medium (solvent, salts, etc.). It is only in the last two decades that large sets of data on kinetic substituent and solvent effects and on product selectivities, which can be related to the general background of physical organic chemistry,have become available. The objective of this chapter is not to repeat the reviews of others (Freeman, 1975;De la Mare and Bolton, 1982;Schmid, 1989;Ruasse, 1990)of the large body of relevant data, but to analyse the present status of the bromination mechanism (Scheme 2) and how it depends on the substituents and on the

Scbeme 2

solvent. Some features, the occurrence and structure of ionic intermediates, the involvement of bromine-olefin charge transfer complexes on the reaction pathway, for example, are now well established. For others, in particular return, there is no conclusive and extended evidence but only isolated data, the interest of which has to be examined in relation to the mechanism of analogous reactions, such as hydration and s N 1 nucleophilic substitutions. What is retained nowadays of the initial mechanism (Scheme 1) is the occurrence of a cationic intermediate. But bromine bridging is not general, and its magnitude depends mainly on the double bond substituents (Ruasse, 1990).For example, when these are strongly electron-donating, i.e. able to stabilize a positive charge better than bromine, P-bromocarbocations are the bromination intermediates. The flexibility of transition state and intermediate stabilization puts bromination between hydration via carbocations and sulfenylation via onium ions. As the understanding of the ionic intermediates has progressed, advantage has been taken of the fact that bromination, like sN1 heterolysis, is a carbocation-forming reaction. Kinetic data on this addition have therefore been used to examine in detail how the basic concepts of physical organic chemistry work as regards transition-state shifts with reactivity (Ruasse et al., 1984). Bromination lends itself particularly well to the quantitative application of the BEMA HAPOTHLE (acronym for Bell, Marcus, Hammond, Polanyi, Thornton and Leffler; Jencks, 1985). In particular, it has been possible to evaluate the transition-state dependence on the solvent and substituents. The major disadvantage that bromination shares with many

210

M:F.

RUASSE

other reactions is that its ionic intermediate is so reactive that thermodynamic data on its stability in solution are not available. However, if this difficulty can be overcome, the intrinsic kinetic contribution to the energy barrier of some brominations can be distinguished from the thermodynamic one. The role of the solvent in bromination has also been extensively studied (Ruasse and Motallebi, 1991). It is obvious that solvation is the essential driving force of this addition since it is extremely difficult to form a bromonium ion from bromine and ethylene in the gas phase whereas it is a very fast reaction in solution. The main contribution comes from assistance to bromide ion d e p r e in the transition state either by a second bromine molecule, leading to the very stable tribromide anion in halogenated or non-protic solvents, or by solvation, when protic solvents are involved. The magnitude of the kinetic solvent effects is thus directly proportional to the charge developed on the leaving bromide at the transition state; such solvent effects therefore provide a convenient measure of the progress of the reaction at the top of the kinetic barrier. In addition to electrophilic participation, solvents can also assist positive charge development nucleophilically, via a mechanism analogous to that postulated in S,2 (intermediate) solvolysis. This effect of nucleophilic solvents is not very important energetically, but it has some unexpected consequences on the stereo- and chemo-selectivity of bromination via carbocations. Significant recent modifications of the mechanism in Scheme 1 concern the demonstration that bromine-olefin charge transfer complexes (CTCs) are active intermediates on the reaction pathway and the possibility that ionic intermediates are formed reversibly. The existence of bromine-olefin CTCs was shown a long time ago (Dubois and Garnier, 1967b).The mechanistic consequences of this finding have been extensively discussed (Olah, 1975; Kochi, 1988), but not until 1985 was it proved that these CTCs are essential intermediates on the reaction pathway, as shown in Scheme 2 (Bellucci et al., 1985a). However, only a few equilibrium constants ( K )for their formation are available. It is still too early to be sure that these K-values are, or are not, significantly solvent- and/or substituentdependent. The same question still arises regarding return in bromination. Reversible formation of several bromonium ions has been shown to occur for various olefins under specific reaction conditions (Strating et al., 1969; Brown et al., 1984, 1990; Bellucci et al., 1987, 1990; Ruasse et al., 1991). Some characteristics of return emerge from these data, but the situation is still by no means clear. In what follows, we present first the experimental conditions under which it is possible to obtain data relevant to the study of the bromination mechanism, and the present evidence for the occurrence of charge transfer complexes and cationic intermediates on the reaction pathway. Once these

ELECTROPHILIC BROMINATION OF C=C

21 1

DOUBLE BONDS

preliminaries have been established, the following sections are devoted to kinetic substituent and solvent effects on this cation-forming halogenation; these effects are discussed in terms of kinetic selectivity, transition-state shifts and solvation. Finally, the problem of return, which raises questions about the nature of the rate-limiting step, is addressed from the few available results. To conclude, we shall take a brief look at the several points that still remain obscure: competition between free bromine and the so-called tribromide ion, the nature of the last, product-forming, step, etc.

2

P

Metho s for obtaining reliable bromination rate constants

The bromination of ethylenic compounds is in most cases a very fast reaction. Half-lives of typical olefins are given in Table 1. Most of them are very short. In order to obtain extended and meaningful kinetic data, it has been necessary to find suitable reaction conditions and to design specific kinetic techniques. This was not done until 1960-1970. As a consequence, kinetic approaches to the bromination mechanism are relatively recent as compared with those to solvolytic reactions, for example.

Table 1 Bromination half-lives" of some ethylenic compounds at concentrationb and at 25°C. TCE'.~ trans-Cinnamic acid Ally1 chloride Stilbene 1-Pentene Styrene Cyclohexene trans-MeCH=CHEt Me,C=CHMe Me,C=CMe, EtOCH=CH,

-

8.3 hk.' 8.4 mino -

10 s' 1.2 so -

AcOH' 1 hf.8 1 1 hi 15 h" 1.2 minP 2 min" -

660 msO 23 ms' 2.1 ms' 10 psu

M

reagent

MeOH"

12 hh 1.6 hj 1.5 min" 2.6 sp 860 msq 90 mss 63 ms' I10 psl I 0 ps* 4.5 psu

Calculated from bromination rate constants measured in the given references. [Br2] = [Ol]. - s-'; ~ second-order in bromine. From k in M - ' s - ' ; first-order in bromine. In 75% aqueous acetic acid. @DeYoung and Berliner (1977). Schmid et al. (1977b). 'Zhang (1981). 'Dubois and Bienvenue-Goetz (1968a). Ir In 1,2-dichloroethane, where the rate is about the same as in TCE. 'Bellucci et al. (1987). '"Yates and MacDonald (1973). " Bartlett and Tarbell (1936). "Modro et al. (1977). pGarnier and Dubois (1968). Ruasse and Dubois (1975). 'Bellucci et al. (1985b). 'Dubois and Fresnet (1973). 'Ruasse and Zhang (1984). " Ruasse (1985).

' 1,1,2,2-Tetrachloroethane.dFrom k in M

M.-F. RUASSE

212

KINETIC RATE EQUATIONS

Bromination can be a second-, third- or higher-order reaction, first-order in olefin but first-, second- or higher-order in bromine. Most of the early kinetic studies were focused on this complex situation (De la Mare, 1976). It is now M are necessary to obtain known that bromine concentrations less than simple or workable kinetic equations. This limit varies slightly with the solvent; for instance, in methanol lo-' M bromine leads to convenient rate ~ the highest equations (Rothbaum et al., 1948) but in acetic acid 1 0 - 3 is that can be used (Yates et al., 1973). In halogenated solvents, olefin bromination is second-order in bromine (1) (Modro et al., 1977; Bellucci et al., 1980). Moreover, even when small dCBrzl - k3[Ol][Br2]2 dt bromine concentrations are used, the bromination kinetics may be difficult to investigate if the solvent is not adequately purified. Adventitious traces of unknown catalytic species provoke very complex and irreproducible kinetic signals. Due to this complication, most of the old published data are unreliable. Nevertheless, good rate constants can be obtained when the solvents are pure (Schmid ec al., 1972; Bellucci et al., 1981). In particular, rate data in 1,2-dichloroethane, DCE (Bellucci et al., 1985b), and in 1,1,2,2tetrachloroethane, TCE (Modro et al., 1977), are now readily available. In protic solvents, bromination is first-order in bromine but the rate law (2) also includes terms related to the bromide concentration. In these media,

since solvent-incorporated products are formed, bromide ions are liberated during the course of the reaction. The electrophilic tribromide species is then produced according to the well known, fast equilibrium (3) (Bienvenue-Goetz Br,

+ Br- PK Br;

(3)

et al., 1980; Ruasse et al., 1986). Consequently, during a kinetic run, addition

of tribromide competes increasingly with that of free bromine, concentration [BrJf. This complication was resolved very early by adding bromide in excess with respect to analytical bromine ([Br2Ian= [Br2If + [Br;]) in order

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

213

to maintain the [Br21f/[ Br;] ratio constant throughout the reaction (Bartlett and Tarbell, 1936; Dubois and Bienvenue-Goetz, 1968a). The rate constants of free bromine and tribromide additions are obtained from the experimental constant, kobs, by using (4)to describe the kinetic effect of bromide ion. This

equation, first established by Bartlett and Tarbell ( 1936)and thereafter widely applied (Dubois and Huynh, 1968; Dubois and Garnier, 1967a; Rolston and Yates, 1969a), has been found to be applicable in a large variety of protic solvents (acetic acid, water, methanol, ethanol and their aqueous mixtures). The rate constant k is related unambiguously to the free bromine addition to alkene but the constant involved in the bromide-dependent term, denoted by kBrl, cannot be interpreted straightforwardly. Its mechanistic significance is complex because it can be associated with several pathways, namely, tribromide addition, salt effect on the free bromine reaction and/or bromideassisted bromine addition, these processes being kinetically indistinguishable (Dubois and Bienvenue-Goetz, 1968a). Empirical relationships ( 5 ) between the experimental rate constants and the k-values extrapolated to zero bromide concentrations have been log k

=a

log kobs + b

(5)

occasionally found (Dubois and Bienvenue-Goetz, 1968a) for limited sets of analogous alkenes. For example, relationships of this form have been obtained for the bromination in methanol of uncrowded alkenes (Dubois and Bienvenue-Goetz, 1968a), styrenes (Ruasse et al., 1978), stilbenes (Ruasse and Dubois, 1972) and 1,l-diphenylethylenes (Dubois et al., 1972a). The a-coefficientsare generally close to unity, but the b-values can vary significantly from one olefinic series to another. The possibility of using kobs instead of k in structure-reactivity correlations has also been discussed (Dubois and Huynh, 1968)in terms of k/kB,, ratios and K-values. It is, therefore, possible to avoid the tedious experimental work of measuring bromide ion effects systematically to obtain k. However the procedure must be applied carefully and critically to avoid erroneous extrapolations. It is noteworthy that in protic solvents most of the bromination rate constants used in mechanistic studies are k, the rate constant for free bromine addition only, that is, for a pathway from which any contribution of bromide is excluded since the involvement of this ion is taken into account in the kBrf term.

214

M -F. RUASSE

KINETIC TECHNIQUES FOR BROMINATION

In halogenated and protic solvents, it is possible to obtain mechanistically significant rate constants by using bromine concentrations below M. But before 1960-1970 and even later, kinetic experiments were carried out using iodometric bromine titrations, which can be handled only for concentrations higher than this limit. In this latter range, the kinetic rate law generally exhibits several terms of higher order than second in bromine. Most work in the early period was devoted to understanding the complex rate equation. This historical, confused situation has been reviewed by De la Mare and Bolton (1982). Before modern kinetic techniques became available, reliable second order rate constants in methanol or in acetic acid were obtained for the reaction of weakly reactive olefins, such as stilbene (Bartlett and Tarbell, 1936), unsaturated carboxylic acids or ally1 chloride and acetate (Nozaki and Ogg, 1942).To enlarge the reactivity range, some workers used competitive kinetic techniques with large reagent concentrations (Bergmann et al., 1972). It was shown, when it later became possible to compare data obtained from direct and competitive techniques, that both lead to the same results only if the M. At higher concentrations, the reagent concentrations are below relative reactivities of two alkenes measured competitively are generally smaller than those obtained from direct kinetic experiments (Mouvier et al., 1973).The reactivity ratios are highly dependent on the alkene concentration. pair is For example, the rate ratio of the l-hexene/4,4-dimethyl-l-pentene about 5.8 when calculated from the directly measured rate constants, but only 1.8 if obtained in experiments where both alkenes are lo-' M in methanol at 25°C. Other significant data are shown in Fig. 1. As a consequence, rate constants obtained by competitive techniques cannot be considered as reliable. Bromination data became accessible over a large reactivity range when it became possible to follow low bromine concentrations. All the modern kinetic techniques are based on the fact that, since bromination is a second- or third-order reaction, bromination half-lives of a few milliseconds to several seconds can be obtained by working at very low reagent concentrations. For example, second-order rate constants as high as lo9 M - ' s - ' can be readily measured if the reagent concentrations are 1 0 - 9 ~ the , half-life of the bromine-olefin mixture then being 1s. Very low bromine concentrations are followed either spectrophotometrically or electrochemically. In halogenated solvents, only spectroscopic determinations are possible. The bromine extinction coefficient at its maximum (A,, = 400-450nm) is about 150-250~-'cm-', so that the workable ~ et al., 1985a). Taking into concentrations are not less than l O P 4 (Bellucci

ELECTROPHILIC BROMINATION OF C=C

01

1

ldLM

I

I

10.)~

215

DOUBLE BONDS

16%

,

-

[A]=[B]

Fig. 1 Concentration dependence of rate-constant ratios measured by competitive brominations (data from Grosjean et al., 1973).

account this limitation and the present performance of stopped-flow equipment, the highest bromination rate constants available in halogenated ~ range. solvents are in the lo5 M - s-’ In protic solvents it is possible to obtain kinetic data for more reactive alkenes by following tribromide rather than free bromine, although the reaction half-lives are shorter than those in halogenated media (see Table 1). Since the presence of bromide ions in large concentrations (0.05-0.5 M ) is necessary to obtain mechanistically significant rate constants, tribromide is generally the major bromine species in these media in which its formation constant is high (16, 92, 177 and 400 in water, acetic acid, methanol and ethanol respectively) (Bienvenue-Goetz et al., 1980). The tribromide extinction coefficients (20 000-40000 M - cm- ’) at the absorption maximum (280-300nm) are much greater than those of free bromine (Dubois and Herzog, 1963); it is therefore possible to follow analytical bromine concenM. Under these conditions, bromination rate constants trations as low as of about lo5- lo6 M - s- are accessible by spectroscopic techniques (Dubois and Garnier, 1967b). Other methods use the electrochemical properties of the bromine-tribromide couple. In these techniques, very small bromine concentrations are first produced by quantitative electrolysis of a bromide added to the reaction medium (Poupard et al., 1983). After adding the alkene, the bromine uptake is followed either potentiometrically or amperometrically. In the “concentrostat” technique (Dubois et al., 1965, 1973a), the bromine concentration

’

’ ’

M.-F. RUASSE

216

Potentiometry

:

Concentrostat

8

,

8

,

I

,

I

'

I

'

I

' I

II I I I I

I

Spect rophotornet r y

I I 2

Arnperometry

I I I

Fig. 2 Scope of kinetic methods for bromination.

is maintained constant during the kinetic run by automatically compensating its consumption by the olefin. With this method, rate constants up to 5 x lo5 M - s~ - l have been obtained in methanol. In potentiometry, the variation of the potential of a Pt electrode relative to a calomel reference electrode represents the time-dependent bromine concentration. Available [BrJ is about 2 x 10-5-10-4 M; pseudo-first-order ~ conditions ([Ol] >> [BrJ) have to be used. Rate constants up to lo4 M - s-l can thus be obtained (Atkinson and Bell, 1963; Dubois et al., 1968). Finally, the most advanced method is couloamperometry. Halogen concentrations in the 1 0 - 8 - 1 0 - 9 ~range are determined by a specially devised amperometric set-up. Second-order conditions ([BrJ w 2 x [Ol]) are used so that rate constants as high as lo9 M - s~ - l can be measured (Dubois et al., 1964, 1983). With these techniques (Fig. 2), bromination rate constants in acetic acid, water, methanol, ethanol and more generally in any solvent in which a bromide is soluble to at least 0.2 M are obtained with a precision of about f2% when the method is easily applied and f5% when it is used close to its limit.

3

Bromine-olefin charge transfer complexes as essential intermediates in bromination

In agreement with Dewar's proposal (Dewar and Leplay, 1961) and by analogy with the long-established halogen-aromatic molecular complexes

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

217

(Andrews and Keefer, 1951, 1964), donor-acceptor complexes between halogens and ethylenic compounds have been postulated for a long time. The first experimental evidence for their existence was obtained spectrophotometrically in 1967 (Dubois and Garnier, 1967b); transient charge transfer bands appear in the spectra when acceptor halogens interact with donor alkenes in freons, low-polarity solvents in which bromine addition is slow. The complexes absorb in the 240-320nm range. In agreement with theory (Mulliken, 1952), a linear relationship (6) between the energy of

bromine CTCs, hvCTC,and the ionization potential I D of the alkenes is found for 14 ethylenic compounds substituted by linear alkyl groups (Me, Et, n-Pr, n-Bu). Moreover, the interaction energies are linearly related to the activation energies. At about the same time, a Russian team (Sergeev et al., 1973) published analogous data, including some equilibrium constants for the formation of these alkene-bromine complexes. After these preliminary observations, the problem was to establish whether these CTCs are essential intermediates in olefin bromination (Scheme 3). Are \

/

/

\

C=C

+Br,

1

-+

Br /+\ ,C-C /

Br-

9’

Scheme 3

they formed in a route competitive with that leading to the cationic intermediate? Or do they occur in an equilibrium step prior to bromonium ion formation? In other words do the n-complexes dissociate heterolytically to a-complexes? Several arguments based on the parallelism between substituent effects on the kinetics and on charge transfer energies tend to favor the second hypothesis (Dubois and Garnier, 1968). Gebelein and Frederick (1972)also attempted to obtain some evidence for CTC involvement in bromination from the concentration dependence of the rate constant, but they failed since it was impossible to obtain the CTC equilibrium constants with the spectroscopic techniques available at that time: Subsequently, and although no experimental proof was accessible, it was commonly agreed that CTCs are active intermediates in bromination. In particular, Olah et al. ( 1974a) concluded from nmr spectra that the bromine-

218

M:F.

RUASSE

adamantylideneadamantane adduct (Strating et al., 1969) in non-nucleophilic media is not a bromonium ion but a molecular n-complex. Taking advantage of this observation, Olah (Olah and Hockswender, 1974), continuing his work on electrophilic aromatic substitution, developed the controversial idea (Ruasse and Dubois, 1975) that the rate-lim‘ ing transition state of olefin bromination resembles either a n- or a Q-co plex when the solvent is non-polar or highly polar, respectively. Some years later, Kochi et al. (Fukuzumi and Kochi, 1981) applied their theory on electron and charge transfers to electrophilic alkene bromination (Kochi, 1988) by comparing the reactivities of various alkenes in bromination and in mercuration. Although the substituent effect trends in the two reactions are totally different, a linear relationship (7) is observed when the reactivities

“h,

[log(k/ko) - hVCTIBr, = [log(k/kO) - hVCTIHgXz

(7)

are corrected by work terms evaluated from the corresponding energies, hv,,, of their charge transfer complexes. This result fits very well Kochi’s postulate, which is illustrated by a thermochemical cycle (Scheme 4) in which the

D

+A

K

% [D+,A-]*

[D,A]

CD A-I, +9

AGc= Ahv,,

+ AGs

Scheme 4

activation free energy changes are expressed as the sum of two terms, one, AGs, related to solvation, and the other the charge transfer energy of the donor-acceptor (DA) complex which includes steric effects. Unfortunately, this approach suffers from the fact that relationship (7) is based only on alkenes with one to four linear alkyl groups, i.e. alkenes whose bromination does not exhibit any significant kinetic sensitivity to steric effects. It would be interesting, therefore, to know if (7) is also valid for compounds bearing one or more branched substituents. Moreover, the work terms evaluated from the CT energies take into account the electronic reorganization but not that of the nuclei, a microscopic event which is probably energetically important. In this respect, it is noticeable that the distance between the C-C bond and bromine decreases significantly on going from the bromine-olefin n-complex ( d = 3.0A) to the bromonium ion ( d ’ = 1.9A) (Prisette et al., 1978). Although the involvement of molecular bromine-olefin complexes as active intermediates in bromination had been widely assumed, this was demonstrated

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

219

only recently (Bellucci et al., 1985a). In contrast with the earlier CTC observation, which dealt with olefins of very reduced reactivity (as a result of eithe the solvent or the substituents), Bellucci's team investigated the brominati n of cyclohexene in 1,2-dichloroethane. By using high alkene M ) concentrations and by monitoring (lo-'- lo-' M ) and low bromine the cyclohexene-bromine CTC at 287 nm near its absorption maximum, but far from that of bromine, they were able to evaluate the spectroscopic = 5520 250 M - ' cm- at 287 nm) and the formation characteristics constant (K, = 0.47 f 0.08 M-') of this complex at 25°C. The thermodynamic parameters of CTC formation and of the third order bromination of this alkene were also measured (Scheme 5 ) . It was shown that the negative

a

A H + = -8.37kcalrnol-' AS+ = -64.0e.u.

A H = -4.60kcalmol~' AS = - 17.0e.u. Scheme 5

temperature coefficient, frequently observed for bromination in non-protic solvents, is consistent only with a bromination mechanism where the CTC is an essential intermediate (route 2-3, Scheme 3). These authors have applied their method to measure the formation constant of the CTCs arising from adamantylideneadamantane (Bellucci et al., 1989) and tetraisobutylethylene in several solvents (Brown et al., 1990).The results, together with previous less reliable data, are shown in Table 2. Since it is now established that CTCs are involved in the bromination pathway, a question about the meaning of the experimental bromination rate constant arises. These constants are not those of an elementary step but are the products of the CTC formation constants KCTC and of the rate constants ki for CTC ionization into the 0-intermediates (8).

Information on the mechanism is mainly obtained from kinetic solvent and substituent effects, i.e. from p- and m-values, as discussed below. These coefficients are therefore a composite of p- and m-values for CTC and ionization steps as shown in (9). Obviously, neither pCTCnor mCTCis available P = PCTC+ Pi m = mCTC mi

+

(9)

220

M:F.

RUASSE

Table 2 Formation constants of bromine-alkene charge-transfer complexes. Alkene

KCTCIM -

0.08

Crotonic acid" 1 -Hexeneb 4-Me- 1-penteneb Cyclohexeneb Cyclohexene' Ad=Add TIBE' TIBE" TIBE"

0.145

0.33 0.36 0.47 289 9.71 1.72 2.5

'

Solvent DCEJ Hexane Hexane Hexane DCEJ DCEJ DCEJ AcOH MeOH

Buckles and Yuk (1953). * Sergeev et al. (1973). 'Bellucci et al. (1985a). Bellucci et al. (1989). eTetraisobutylethylene; Brown et al. (1990). 1,2-Dichloroethane.

'

at present. However, it can be assumed that they are negligible with respect to pi and mi. Substituent and solvent effects are all the more important when the charge modification on going from reactants to products or to transition states is large. Despite their name, there is no charge separation in the CTCs, whereas the rate-limiting second step is an ionization process. Values of pCTC and mCTC should therefore be small compared with pi and mi respectively. Recent data on tetraisobutylethylene bromination (Table 2) support this hypothesis. On going from acetic acid to methanol, the rate constant increases by a factor of 10 whereas the CTC formation constant is hardly modified (Brown et al., 1990). Moreover, when the ring substituent effects on the bromination rates of a-methylstyrenes (Ruasse et al., 1978) are compared with those on equilibrium constants for the formation of the iodine-acetophenone CTCs (Laurence et al., 1979), it is calculated (Ruasse, 1990) that the contribution of the olefin-bromine CTC formation to the p-values obtained from kinetic data cannot be greater than 9%. Accordingly, experimental p- and m-values are generally discussed in terms of effects related to the ionization step of bromination only. 4

The ionic intermediates: bridged bromonium ions or open B- bromocarbocations

Although there is general agreement about the occurrence of ionic intermediates on the bromination pathway, there are only few direct observations of these bromocations. It is still difficult to decide whether they are bridged or open depending on the double-bond substituents.

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

221

EXPERIMENTAL OBSERVATIONS

'H and 13C nmr spectra of methyl-substituted bromonium ions in nonnucleophilic superacid media have been obtained by Olah's group (Olah et al., 1968, 1974b; Olah and White, 1969). The dependence of the chemical shifts of the three-membered ring carbon atoms on the number of methyl groups is discussed in terms of either symmetrically bridged or unsymmetrical bromonium ions in equilibrium with a pair of open B-bromocarbocations (Olah, 1975). For the parent ion, the data are in agreement with a static bridged structure. gem-Dimethyl groups produce a dissymmetry in the bridged structure and a small contribution from equilibrium (10). Additional Br

\

7c-c\

+/

=

Br /+\ +CT

7

/Br c-c/ \ \+

substitution by a third methyl on the other carbon would significantly favour ( 10). For the symmetrically substituted ethylene bromonium ions, the results

suggest that their structure is also a mixture of a bridged species and a rapidly equilibrating pair of open ions. However, this proposal was not confirmed by later work, the I3C nmr spectra of these deuteriated ions being temperature-independent (Servis and Domenick, 1987). Moreover, recent MNDO calculations show that there is no kinetic barrier between these two structures (Galland et al., 1990). Consequently, an equilibrium such as (10) cannot describe the bromocations. One conclusion from these results is that most of the charge of these ions is on the ring carbon atoms and not on the bromine. In contrast, '3C nmr spectra of ring-substituted B-bromocumyl cations [2] can be unambiguously interpreted in terms of open B-bromocarbocations, since the ring substituent effects on the chemical shifts are similar to those on the corresponding non-brominated cations [l], even for the electron-attracting p-trifluoromethyl group (Olah et al., 1972).

Bromonium ion stabilities in the gas phase have also been measured in ion cyclotron resonance experiments by Beauchamp's group (Staley et al., 1977). The heterolytic bond dissociation energies shown in Table 3 are taken as a

222

M:F.

RUASSE

Table 3 Relative stabilities" (in kcal mol-') of substituted bromonium ions and alkyl carbocations in the gas phase.b

D (R+-Br-)

R+

D (R+-Br-)

R'

0

+ 52.5

- 5.6

-4.5

- 18.8

Y

- 18.4

-

16.5

- 30.2

A

-41.4

With respect to the parent ethylenebromonium ion and expressed as heterolytic bond dissociation energies D (R+-Br-). 'Data from Staley ef a[.'(1977). 'Calculations (Galland ef al., 1990) indicate that this ion is not bridged but open.

R\

CH-CH

/

Br

P' \

Br

'

+ R-CH-CH-R'

'dr

measure of these relative stabilities. As expected, stability increases with methyl substitution. An interesting conclusion is obtained by comparing the data for the t-butyl [ 31 and the 2-bromomethyl-2-propyl [4] cations. Since, according to theoretical calculations, the second cation is probably open, a bromo-substituent should stabilize an aliphatic carbocation by about 2kcalmol-' in the absence of bridging. It is also observed that the

\>/

CH

I CH,

c37

CH,

\>/

CH

I,

CH, 1147

CH,Br

ELECTROPHILIC BROMINATION OF C=C

223

DOUBLE BONDS

three-membered ring ion is more stable than the corresponding 1-bromoethyl cation by about 1.4 kcalmol-'. These thermodynamic data are useful for checking theoretical calculations. They have also been used to estimate the solvation enthalpies of these ions by combining gas phase data and their heats of ionization in superacid media (Larsen and Metzner, 1972). The solvent significantly attenuates the effect of methyl substituents on the gas phase stability. Data on molecular structure of bromonium ions are sometimes extrapolated from that of the tribromide-adamantylideneadamantane bromonium ion pair [6] (Slebocka-Tilk et al., 1985), the only stable ionic bromination intermediate that can be isolated and whose crystal structure has been determined. Since the first observation by Strating et al. (1969), it has been established that bromine addition to adamantylideneadamantane [ 51 in

c51

C6l

non-protic solvents stops after bromonium ion formation because of complete steric inhibition of trapping by nucleophiles. Some of the most relevant geometric data on this ion are shown in Scheme 6a. An important feature of

Br

Br

the cation is that it is not symmetrical, there being a difference of 0.078 A between the two C-Br bond lengths. However, this asymmetry probably arises from constraints in the crystal packing produce& by the tribromide counter-ion and is not necessarily conserved in solution. Despite the exceptional character of this bromonium ion, these structural data are valuable for calculations.

M.-F. RUASSE

224

THEORETICAL CALCULATIONS

Early interest of theoretical chemists in bromonium ions was focused on two aspects: are they bridged or open and do they resemble n- or a-complexes? Old semi-empirical (Bach and Henneike, 1970) and recent ab initio (Hamilton and Schaefer, 1990, 1991) calculations are in complete agreement with experimental data as regards the structure of the parent ethylenebromonium ion. Its most stable structure is definitely symmetrically bridged. Early ab initio results showed that the bridged form [7] is more stable than the 2-bromoethylcation [S] and the 1-bromoethylcation [9] by 1-4 and Br /+\

,c-c

H / H

c71

H ‘\ H

+ H ,C-CH

181

Br ,Br

I

’

C+ CH, H ‘ c91

15-30 kcalmol-’ respectively (Poirier et al., 1981, 1983; Scheme 6b). These calculations indicate also a high energy barrier on going from [7] to [9] but an almost insignificant barrier between [7] and [8]. Species [S] cannot be a stable structure, since it corresponds to a very flat segment along the open/cyclic structure. interconversion potential and not to a secondary minimum (Fig. 3; pp. 226-7). These results have been recently confirmed by more elaborate ab initio methods including electron correlation and polarizability functions. Structure [7] (Scheme 6c) is found to be more stable than [9] by only 1.5 kcalmol-’. Structure [ 81 does not exist as a minimum on the potential energy surface and spontaneously collapses to [ 71. It is reasonably expected that substituents significantly modify the structure of the parent ethylenebromonium ion. However, little attention has been paid by theoretical chemists to this assumption, since it is difficult to tackle it by ab initio methods. Recently, MNDO calculations on methyl-substituted bromonium ions have been attempted (Galland et al., 1990). The similarity of the MNDO and ab initio results for the unsubstituted ion (Scheme 6d) gives some confidence in the semi-empirical method. The minimum energy profiles for the opening of the variously substituted ions are shown in Fig. 3. Increasing the number of substituents leads to a flattening of this profile, i.e. to a decrease in the energy difference between open and bridged forms. In every case the curve has only one minimum; therefore there is no equilibrium, as described by (lo), between the two limiting structures. For unsymmetrically substituted ions, the open form is slightly more stable, whereas for symmetrical ions the bridged structure is favoured.

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

225

Finally, every kind of calculation shows that there is no substantial charge on the bromine atom of bromonium ions (Cioslowski et al., 1990), in agreement with the conclusions from nmr spectra. This result is relevant to the possible reversibility of bromonium ion formation, as discussed later. Despite much discussion, there is no simple answer to the question of whether bromonium ions may be viewed as n- or o-complexes. This arises because there is no clear-cut experimental criterion for distinguishing the two types of bonding. Calculated and experimental bond lengths and angles (Scheme 6 ) have values intermediate between those expected for one or other kind of complex. Depending on the data set considered, bromonium ions have been classified as n- or a-complexes. The X-ray structure for adamantylideneadamantane bromonium ion seems to agree better with a a-complex, whereas from ab initio calculations the ethylenebromonium ion would seem to exhibit n-character. As concluded by Slebocka-Tilk et al. (1985), it is possible that “such a distinction is semantic and has no real chemical meaning.” In the early stages of bromination studies, the charge transfer complex and the subsequent bromonium ion were called bromine-alkene n- and o-complexes respectively. Although this nomenclature is not rigorous from an orbitalbonding point of view, there is at present no way of improving it.

KINETIC DATA AND BROMINE BRIDGING IN TRANSITION STATES AND INTERMEDIATES

Kinetic data can be discussed in terms of bromine bridging in ionic intermediates if the transition states of the ionization step are late. It appears that this is the case in the bromination of a wide variety of olefins, and in particular of alkenes, stilbenes and styrenes. Large p - and m-values for kinetic substituent and solvent effects (p. 253) consistent with high degrees of charge development at the transition states, are found for the reaction of these compounds. It can therefore be concluded that their transition states closely resemble the ionic intermediates. The relative magnitude of the kinetic effects of two substituents, R, and R,, on the C, and C, carbon atoms of the double bond (Scheme 7) is taken as a measure of the symmetry of the charge development and therefore of bromine bridging in the bromocations. It is assumed that in a bromonium ion the effects of R, and R, must be similar, whereas for a P-bromocarbocation, C: ,the effect of R, must be significantly greater than that of R,. Consequently, the substituent effects are analysed in terms of a multipathway scheme (Scheme 7) where open carbocations and the bridged ion are formed via discrete pathways with rate constants k,, k , and kgr respectively (Ruasse and Dubois, 1974). The rate constant k in (4) is therefore the sum of these three

M . - F . RUASSE

226

4

Br +0,19 2.01

H

H

,c ’ 1.49

Me--

1 H

c.

1.10 1.511 “Me

I

I

I

I00

50

H

0

*

4 Br +0.08

I

50

68“4

100

0

Fig. 3 Minimum energy profiles for the opening of bromonium ions (Galland et al., 1990). They are not double-well curves, i.e. bridged and open structures are not in equilibrium, whatever the number of methyl substituents.

constants corresponding to different structure-reactivity relationships. The magnitude of bridging in the intermediates is readily obtained from the k,/kB, ratios (Scheme 7 ) . Depending on the nature and the number of substituents, the multipathway mechanism is applied in various ways. k = k,

+ kp + kgr

(11)

For alkene bromination, present kinetic data show that their trmsition states are always bridged, whatever the number of alkyl groups on the double

ELECTROPHILIC BROMINATION OF C=C

Br

227

DOUBLE BONDS

+0,08

1. 8 8 v l \ \ 2 . \ 55

'

\

-c\98" 1 . 4 8C . H -3.11 1.48\

!.I1

-

'H

Me

H

4

n

.""

\

0 5 1,1

C-

H--]l.ll

1S O

\\+ C1.50\''Me

H

Me

I05''

0

t

Me

Me

Fig. 3 Continued.

bond. First of all, values consistently close to unity have been measured (p. 268) for the rn-coefficients of their Winstein-Grunwald relationships in protic solvents; for instance, using the Winstein-Grunwald Y-parameters, in is 1.16, 1.10, 1.31 and 1.40 for 1-pentene (Garnier and Dubois, 1968), trans-2-pentene, 2-t-butyl-3-methyl- 1-butene (Ruasse and Zhang, 1984) and methylideneadamantane (Ruasse and Motallebi, 1988), respectively. It is therefore commonly agreed that the transition states for bromination of alkenes are always highly charged and closely resemble the ionic intermediates. Early evidence for a symmetrical charge distribution is found in kinetic data which show that the effects of linear alkyl groups are additive, whatever their number and their relative positions (Dubois and Mouvier, 1968). This preliminary result was challenged by Bergmann et al. (1972) who analysed

M:F.

228

IBV/

R1\

Br pathway

/Ca-C<

Br,

+

R1\

H

/Ca=F

/

H

/"./

-

RUASSE

R,

R1\+

Br /

/c.--r,-

C a pathway

R2

\

+/

R1ya-C
C , pathway 2

Scheme 7

kinetic effects of linear and branched alkyl groups in terms of a twocarbocation scheme using a multiparameter equation involving nine terms. Owing to the unreliability of the competitive kinetic technique used and to the complexity of the statistical treatment, the conclusion of this work cannot be considered mechanistically significant. A more conclusive result was obtained by Bienvenue-Goetz and Dubois (1978), who measured the kinetic effect of several very varied electron-attracting non-conjugated groups R on methylethylenes [lo], [111, [ 121 and [ 131. The effect of R is expected to be either

2 c 101

<e Clll

2

Me

1121

Me9 Me

1131

the same or different in the series [lo]-[13], depending on whether the transition states are bridged or open. For example, if the charge distribution is carbocation-like, the bromination rate should be highly R-dependent for [111 but weakly so for [131. The results given in Fig. 4 show definitely that the bromination transition states of these alkenes are bromonium ion-like whatever the number and the position of the substituents. This conclusion is not in complete agreement with theoretical calculations (p. 224), which imply that the ionic intermediate derived from isobutene closely resembles an open fl-bromocarbocation, whereas those from cis- or trans-Zbutenes would be bromonium ions. This difference between experiment and calculations is not fully understood, but could be attributed to some

ELECTROPHILIC BROMINATION OF C=C

3: 0

DOUBLE BONDS

229

I

4

2

0

Fig. 4 Symmetrical charge distribution in bromonium-ion like transition states in alkene bromination (data from Bienvenue-Goftz and Dubois, 1978). The effect of the R-substituent depends neither on the number nor on the position of methyl groups on the double bond.

charge reorganization induced by solvation on going from transition state to intermediate. For aryl-substituted ethylenes, where the aryl group can be conjugated with the developing positive charge, the transition states are either carbocationlike, or bromonium-like or a mixture of these two limiting forms, depending on the electron-donating abilities of the ring substituents. This has been shown for substituted stilbene and styrene bromination, the transition states of which are late according to their m-values (1.20 and 0.96 respectively; Ruasse and Dubois, 1975). For these olefins, it has been shown that two or three pathways of Scheme 7 can compete, i.e. (11) must be considered in its totality. This implies that the structure-reactivity relationships describing the substituent effects are curved. Upward curvature has been observed in the po plots for stilbene, styrene and u-methylstilbene bromination in methanol (Ruasse and Dubois, 1972; Ruasse et al., 1978; Ruasse and Argile, 1983). In no case can these curvatures be interpreted coherently by Yukawa-Tsuno equations in terms of variable contributions of polar and resonance effects. Application of the multipathway scheme (Ruasse, 1990) leads to the results summarized in Table 4. For stilbenebromination, a markedly non-linear structure-reactivity relationship is observed (Fig. 5 ) . Detailed analysis of the kinetic effects of two substituents, X and Y, on each aromatic ring shows that the three pathways leading to the C: and C i carbocations and to the bromonium ion can

Table 4 Contribution of bridged and open forms of the ionic intermediates in the bromination of some aromatic olefins in methanol, as calculated by the multipathway scheme"

logkb Styrenes

YOC,

% C,

% Br

100 100 80 60 40 15 100 100 100 100 100 100 100 80 60 15

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 20 40 60 85 0 0 0 0 0 0 0 20 40 85

Ref.

X-C,H4-C=CHRZ

I

X

R,

R2

4-OMe 4-Me H 3-CF3 4-NO2 3,5-(CF,)z 4-OMe 4-Me H 3-CF3 3,5-(CF,), 4-OMe 4-Me H 3-CF3 3,5-(CF,)z

H H H H H H Me Me Me Me Me H H H H H

H H H H H H H H H H H Me Me Me Me Me

Rl

6.53 4.28 3.18 1.08 0.08 0.60 7.82 6.34 5.09 2.96 1.08 6.48 4.54 3.51 1.50 --

0.22

d

d d d d d d d d d

d d

d d

d d

trans-Stilbenes I

X

Y

4-OMe 4-OMe 4-Me 4-Me H 3-c1 3-CF3 3-CF3 3-CF3 4-NO2 4-OMe H H 4-CF3 3-CF3 H

H 3-CI H 4-NOz H H H 3-CF3 4-NO2 H H 4-OMe H H 4-OMe 4-CI

R H H H H H H H H H H Me Me Me Me Me Me

R 4.32 3.92 1.86 0.40 1.04 0.12 0.08 - 1.30 - 1.53 -0.71 5.95 3.61 2.63 -0.03 2.56 2.20

e

+c

+ + + +

e

+ t

+ + t

0 63 0 0

100 37 100 100 0 100

100

0

e

e

+ + + + + + 0 0 0 0 0 0

e

f

f f

f f 9 9

9 g g 9

"Ruasse (1990). b k in M - s~C 1 for free bromine addition at 25 "C. 'This pathway occurs but its magnitude cannot be calculated exactly. References: dRuasse et al. (1978); 'Ruasse and Dubois (1973); Ruasse and Dubois (1974); Ruasse and Argile (1983).

'

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

231

Fig. 5 Reactivity-structure relationship for the bromination of monosubstituted stilbenes (data from Ruasse and Dubois, 1972).The curvature shows the X-dependence of the competition between carbocation and bromonium ion pathways.

compete. For a single electron-donating substituent, 4-methoxy for example, only the CJ path is followed, whereas with two electron donors (X = 4-OH, Y = 4-OMe) both carbocations coexist. One electron-donor group associated with an electron attractor favours the open cation and two electron attractors the bridged intermediate. Owing to the fact that the three pathways must be taken into account in the analysis of the complete kinetic data set, it is not possible to obtain strictly quantitative results but only qualitative characteristics of the competition between the three paths. This has made it impossible to establish rate-product correlations like those found in styrene and a-methylstilbene bromination. For styrene brornination, the curvature in the pa plot is not obvious (Fig. 6). Before it was accepted that the structure-reactivity relationship was not strictly linear, several significantly different p-values were reported (Dubois and Schwarcz, 1964; Rolston and Yates, 1969a; Pincock and Yates, 1970; Yates et al., 1973), depending on the substituent range studied. This p-variation was understood when the effects of the ring substituents on the bromination rates of styrenes, a-methylstyrenes and trans-p-methylstyrenes were measured (Ruasse et al., 1978). The X-dependence of a- and p-methyl effects in bromination was compared with the same effect in hydration (Chang and Tidwell, 1978). Figure 7 shows a very different trend for these two electrophilic additions. The difference is consistent with a carbocationic intermediate for all the hydrations, but, for styrene bromination, with a bromonium ion increasingly competing with the bromocarbocation as X becomes more and more electron-attracting. The outcome of this competition was calculated for substituents varying from strong electron donors (4-OH) to electron attractors [3,5-( CF,),]. The results were checked by measuring the X-dependence of the stereochemistry of bromine addition (p. 239).

M

232

-F. RUASSE

Fig. 6 Reactivity-structure relationship for the bromination of ring-substituted styrenes (data from Ruasse et al., 1978). Competition between bromonium and carbocation intermediates.

0,

Ole-

brornination

1.5

,

&.OM.

- 0.5

&!Me

3-Me

4.Cl

3.CI

,

3.CF3

3.5-(CF,I2

Fig. 7 Comparison of the X-dependence of the a-methyl effect in bromination and hydration of X-substituted styrenes (data from Ruasse et al., 1978).

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

233

Effects of substituents on both rings of a-methylstilbene can be analysed in terms of competition between the two carbocations only (Fig. 8; Ruasse and Argile, 1983). Even in the presence of an electron-attracting substituent, no contribution of a bromonium ion is detected. The replacement of a hydrogen atom on the ethylenic carbon by a methyl group stabilizes the carbocations enough to preclude bromine participation. A similar situation is found on going from styrenes to a-methylstyrenes, the bromination of which involves exclusively a carbocationic intermediate whatever the ring substituent (Ruasse et al., 1978). For X-substituted a-methylstilbenes, the linear pa relationship shows that only the tertiary bromocarbocation is involved. But for the Y-monosubstituted a-methylstilbenes, there is a sharp breakdown in the pa plot, which is readily interpreted as indicating a change in the mechanism from the secondary to the tertiary pathway when Y changes from electron-attracting to electron-donating (Fig. 8). The kinetic data can be quantitatively analysed in terms of competition between the secondary and tertiary carbocations; the results are confirmed by the regiochemistry of methanol attack on these intermediates (p. 242). I n summary, bromination intermediates are bromonium ions as long as there is no conjugated substituent able to stabilize the developing charge better than the bromine atom alone. When a strongly electron-donating substituent is conjugated with this charge, the open b-bromocarbocation is favoured. The two-fold nature of the ionic intermediate can therefore be viewed as a result of the entering bromine atom and the substituents competing to stabilize the positive charge. This picture is supported by the

t

o x + 0.83ao;

Fig. 8 Reactivity-structure relationship for the bromination of (a) X- and ( b ) Y-monosubstituted cr-methylstilbenes (data from Ruasse and Argile, 1983).Intermediates are tertiary bromocarbocations, except when Y is strongly electron-donating; in this latter case, they are secondary bromocarbocations.

234

M:F.

RUASSE

bromination of some unsaturated substrates where the anchimeric participation of hydroxy or phenyl groups overwhelms bromine bridging. For example, the reaction of some unsaturated alcohols goes through five- or six-membered cyclic oxonium ions (12), rather than through bromonium ions (Williams et al., 1969; Bienvenue-Goetz et al., 1970). Phenonium ion is also suggested as the intermediate in 4-allylanisole bromination (Dubois et al., 1973b). CH,=CH(CH,),OH

+ Br,

--+

BrCH2$J,

Br-

--+

Products

(12)

H’ It is still not clear whether steric constraints can modify the magnitude of bromine bridging. There is at least one example that suggests that this is possible. Non-additive kinetic substituent effects and cis-dibromoadducts imply that the intermediate of cis-cyclooctene bromination in methanol is a b-bromocarbocation (Dubois and Fresnet, 1974). It must also be noticed that most of the previously described results were obtained in methanol. The question is therefore whether they can be transposed to other solvents. Kinetic solvent effects described below are generally large and almost independent of olefin structure. These data suggest that solvent assists bromide ion departure strongly but positive charge development in the cationic part of the transition state only weakly. As a consequence, the charge distribution, i.e. the relative contribution of the bromonium and carbocation intermediates, cannot vary significantly with the solvent. This proposal is supported by the generally small solvent dependence of bromination stereochemistry.

PRODUCT DATA AND BROMINE BRIDGING FROM STEREO- AND REGIO-CHEMISTRY

Since bromine addition to olefins leads to brominated compounds of synthetic interest, there are many studies of bromination products in the literature. Typical examples have been reviewed by Schmid and Garratt (1977). However, there are few systematic product analysis studies related to the role of the structure and the solvent in determining bromination selectivities. Bromination can exhibit stereo-, regio- and chemo-selectivity when the reaction is carried out in the presence of nucleophiles (solvent or added salt). When the ionic intermediate is a bromonium ion, a stereospecific but non-regioselective reaction is expected. In contrast, for an open bromocarbocation, the products should be formed regioselectively but not stereospecifically. These considerations were understood very early since, in fact, Roberts and Kimball ( 1937) suggested bridged ions as bromination inter-

ELECTROPHILIC BROMINATION OF C=C

235

DOUBLE BONDS

mediates because of the trans-bromine addition to cis- and trans-2-butenes. Twenty years later, by studying the chlorination of cis- and trans-di-tbutylethylenes, Fahey (1966) showed that halogen bridging is a general rule. It is expected that, owing to steric repulsions between the two branched groups, the cis-alkene would prefer to react via an open P-chlorocarbocation 1141 where free rotation can occur rather than via a chloronium ion [ 151.

c 141

c 151

However, the product in carbon tetrachloride is exclusively the antidichloroadduct, demonstrating unambiguously that chlorine bridging is essential in this disfavoured reaction. Since bromine is a better bridging group than chlorine, a fortiori bromination intermediates must be bridged whatever the steric constraints in the alkene. The generality of this finding is confirmed when bromination products of less congested alkenes are systematically studied in methanol containing 0.2 M NaBr (Ruasse and Chrktien, 1993; Table 5). Whatever the substituents, the bromination of 1,Zdisubstituted alkenes is stereospecific: the cis-alkene leads to the threodibromide and threo-solvent-incorporated product only; the adducts from the trans are systematically erythro. As concerns the regioselectivity of methoxybromide formation, it is generally claimed that Markovnikoff addition takes place. However, as shown in Table 5, this rule is only roughly followed. Monosubstituted alkenes give preferentially the Markovnikoff adduct when the substituents are not bulky; when they are branched, anti-Markovnikoff addition is preferred or is even exclusive. Cis- and trans-alkenes lead also to significant amounts of antiMarkovnikoff products. The regioselectivity seems to be determined by a subtle balance between steric and polar factors, which cannot be expressed quantitatively. In contrast, the reaction of gem-disubstituted alkenes is completely regioselective, in agreement with Markovnikoff’s rule, even when the substituents are highly branched, such as t-butyl or adamantylidene. This result agrees with MNDO calculations, which show that the ion derived from gem-dimethylethylene bromination has an open structure, but not with the kinetic results (Fig. 4). As stated before, this difference is not readily understood. Because of their complexity, the chemoselectivity data of Table 5 have not been fully interpreted either.

236

M:F.

RUASSE

Table 5 Stereo-, regio- and chemo-selectivity of alkene bromination in methanol (0.2M NaBr) at 25°C."

R,\

7

3

-

/c=c\R4

R2

DB

,/

/R3

R1,

C(OMe)-CqBr R f R4 P

MB,

3

CBr-CTOMe

MB,

R4

~

Me n-Pr i-Pr neo-Pe Me Me n-Pr n-Pr i-Pr i-Pr t-Bu t-Bu i-Pr i-Pr Me i-Pr t-Bu' Me

~~

H H H H H H H H H H H H H H Me Me Me Me

H H H H H Me H Me H Me H Me H i-Pr H H H Me

H H H H Me H Me H Me H Me H i-Pr H H H H H

"Ruasse and Chretien (1993). b T = threo; E

=

49.4 37.3 18.8 47.7

11.3 12.1 20.7 9.3 63.0Eb 59.4Tb

21.9E 21.2T 0 10.2T 0 0

35.OE 28.8T 54.9E 38.6T 56.8E 34.71 18.8E 31.OT

84.7 73.9 65.6 81.4

0 0 0 0

39.3 50.7 60.5 43.0 37.OE 40.7T 43.1E 50.OT 45.1E 5 1.2T 43.1E 65.2T 81.2E 69.OT 15.3 26.0 21.0 18.6

erythro. ' 13.4% bromoalkene.

For the methyl-substituted ethylenes, i.e. in the absence of any steric effects, there is a roughly linear relationship between the chemoselectivity and the I3Cnmr chemical shift of the most substituted carbon atom of the bromonium ions (Dubois and Chrttien, 1978). This selectivity is therefore discussed in terms of the magnitude of the charge on the carbon atom and the relative hardness of the competing nucleophiles, according to Pearson's theory (Ho, 1977). However, this interpretation does not take into account the substituent dependence of the nucleophilic solvent assistance, which must play a role in determining this chemoselectivity.

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

237

The solvent has no influence on the stereoselectivity of bromine addition to alkenes (Rolston and Yates, 1969b), but it could have some effect on the regioselectivity, since this latter depends not only on polar but also on steric effects. Obviously, it modified the chemoselectivity. For example, in acetic acid Rolston and Yates find that 2-butenes give 98% dibromides and 2% solvent-incorporated products whereas, in methanol with 0.2 M NaBr, dibromide is only about 40% and methoxybromide 60%. There are no extensive data, however, on the solvent effects on the regio- and chemoselectivity which would allow reliable predictions. More results are available on the bromination products of aromatic olejns and in particular on styrenes and stilbenes. Bromination of X-substituted styrenes (13) in acetic acid (Rolston and Yates, 1969b) and in methanol

'\\

+ Br,

RoH

XwHOR-CH,Br

"Q-

CHBr-CH,Br

+

(13)

(Ruasse et al., 1978) is totally regioselective and shows X-dependent chemoselectivity. This is partly in agreement with the kinetic data, which indicate no primary carbocation but rather a competition between the benzylic carbocation and the bromonium ion, depending on X. According to the data of Table 6, bridged intermediates would lead to more dibromide than open ions do. From these results and from those on gem-, cis- or trans-disubstituted alkenes, empirical rules have been inferred for chemoselectivity: (i) more solvent-incorporated product is formed from open than from bridged ions; (ii) methanol competes with bromide ions more efficiently than acetic acid. Table 6 Chemoselectivity ( % solvent-incorporated products) in bromination of styrenes, XC6H,CH = CH,. X 4-OMe H 4-Br 3-CI 3-NO, 4-NOZ

AcOH, 0.1 M LiBr"

MeOH, 0.2 M NaBrb

__

98

16.0

85 75

6.5

3.6

"Rolston and Yates (1969b). bRuasse ef al., (1978).

-

15

238

M:F.

RUASSE

The stereoselectivity of bromination of cis- and trans-B-methylstyrenes, where the kinetic data show competition between benzylic carbocations and bromonium ions, has been widely studied. As early as 1968, Fahey and Schneider observed that cis- and trans-anetholes, 4-MeOC6H,CH= CHMe, are brominated in carbon tetrachloride with the same lack of stereoselectivity but that the parent B-methylstyrenes exhibit some selectivity. At about the same time, Rolston and Yates (1969b,c) obtained similar results in acetic acid, where the formation of 80% dibromide is observed. Finally, the dependence of the dibromide stereochemistry on the ring substituent in P-methylstyrene bromination in dichloromethane is found to agree quantitatively with the magnitude of bridging calculated from the kinetics in methanol (Ruasse et al., 1978). In contrast, in this latter solvent, the major ( > goo/,) methoxybromides are formed completely regioselectively and stereospecifically anti, whereas the minor dibromide is formed with some stereoselectivity only (Scheme 8). The same regio- and stereo-selectivity is observed

x~H(OMe)CHBrMe 100% erythro

xm

CHBr-CHBr Me

erythro/threo = 63/37 (X = 4-OMe) 100/0 [X = 3,5(CF,),]

Scheme 8

for the formation of acetoxybromides from unsubstituted 8-methylstyrene in acetic acid. It seems possible to generalize for these aromatic olefins that: (i) dibromides are formed with stereoselectivity that depends on the substituents and the solvent (Tables 7 and 8); (ii) the reaction of the cis-isomer is less stereoselective than that of the trans; (iii) the mixed adducts are formed completely regioselectively and stereospecifically whatever the solvent and the substituents. Less extensive data on the selectivities of product formation from cis- and trans-stilbene bromination are available; the stereoselectivity of the corresponding dibromides is solvent- and concentration-dependent (Table 9; Buckles et al., 1962; Heublein, 1966; Bellucci et al., 1990). Solvent-dependent lifetimes and ion-pairing of these intermediates can be responsible for the observed variations in the stereo- and chemo-selectivity. Assuming that bromonium ions and carbocations are formed in discrete pathways, the influence of these factors can be readily understood. On the one hand, bridged ions react stereospecifically whatever the medium; the

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

239

Table 7 Stereochemistry of dibromides in the bromination of p-methylstyrenes in methylenechloride: rate-stereoselectivity relationship." X trans-4-OMe trans-H cis-H trans-3-CF3 trans-3,5-(CF3),

YOErythro

% kB;

63 81 28 91 100

0 77

-

93 98

"Ruasse et al. (1978). Bromonium ion pathway calculated from kinetic data.

Table 8 Solvent-dependence of the stereochemistry of dibromides from cis- and trans-p-methylstyrene bromination. YOanti-addition from

Solvent C6H12b

CC1,b AcOH' Cl,CH-CHC12C CI,CH,' (AcO),O' C6H5N02' MeNOZb MeCNb

Ea

cis

trans

1.9 2.2 6.2 8.2 9.1 21 34.8 35.9 37.5

85 78 73 66 72 49 45 46 69

92 90 83 90 90 83 82 81 89

"Dielectric constant. bPannel and Mayr (1982). 'Rolston and Yates (1969b).

Table 9 Solvent-dependence of the stereochemistry of dibromides from cis- and trans-stilbene bromination." % anti-addition from

Solvent

cs2

CCI, C1,CCN CI,CCO,Me C6H5N02

"Heublein (1966).

cis

trans

81 77 34 51 29.5

94.5 89 82.5 79 83.5

240

M:F.

RUASSE

solvent can only modify the chemoselectivity by changing the ion-pair dissociation. On the other hand, the stereochemistry of bromide attack on the carbocations should depend on their lifetimes. In non-nucleophilic non-polar solvents such as hexane or dichloromethane, the cation is not stabilized by the medium; it is therefore short-lived enough not to undergo significant equilibration before reacting with the counter-bromide ion. As a result the highest stereoselectivities for cis-olefin brominations are found in non-polar media. In contrast, in polar solvents such as nitrobenzene, equilibrium ( 14) can be achieved before the intermediate is trapped.

Ph RA

H

€ € A P H h

(14)

Bromination is less stereoselective, and the reactions of cis- and trans-olefins tend to be stereoconvergent. The stereospecific formation of the mixed bromoadducts in protic media, such as methanol or acetic acid, could be interpreted in the light of the recent finding (Ruasse et al., 1991) that these solvents assist the formation of the ionic intermediate nucleophilically. If a solvent molecule is close to the cationic part of the transition state in the rate-limiting step, the intermediate can be trapped by this solvent molecule in a necessarily trans mode with respect to the first bromine, before the two components of the ion-pair diffuse away from each other (15). This would

s\

I

‘c=’

0

HY

-

I I

-C-Br

so-c,Br-

-+

I

X-Br

+ Br- + H +

(15)

I

inhibit any competition between solvent and bromide ions in the trapping of the intermediate and any conformer equilibrium, which would lead to non-stereoselectiveadducts. Moreover, chemoselectivityshould also be highly dependent on this solvent assistance, which would favour the solventincorporated products over the dibromides. There is at present not enough data on the magnitude of this solvent effect to understand fully its influence on the stereo- and chemo-selectivity of bromination in general. C,HC,H,Y, The bromination products of a-methylstilbenes, XC,H,C,Me= have been extensively studied in methanol as a function of X and Y (Ruasse and Argile, 1983). In agreement with kinetic data, which establish that the intermediates are exclusively tertiary and/or secondary bromocarbocations, the bromination of these olefins is completely regio-selective but hardly stereoselective at all. The two diastereoisomeric di-bromides and methoxy-

/

:,B

biE

E:

0 . Y I

242

M:F.

RUASSE

bromides are formed in the same erythrolthreo ratio (70/30), whatever X and Y, in both dichloromethane and methanol. When the CT carbocation is predicted to be the only intermediate, methanol attacks the C, carbon atom exclusively; when Y is electron-donating, the kinetics predict that the carbocation will be secondary, and this is confirmed by the finding that the sole product in methanol is the corresponding methoxy-bromide. For strongly electron-releasing Y (4-OH, 4-OMe), 1,2-dimethoxy-1,2-diarylpropanes resulting from a fast solvolytic displacement of the tertiary bromine are obtained according to Scheme 9. When secondary and tertiary pathways compete, the reaction in methanol leads to a mixture of the two regioisomeric methoxybromides in a ratio that agrees with that calculated from the kinetics (Table 10). Finally, the bromination of these olefins in methanol is also completely chemoselective,only solvent-incorporated products being obtained, even in the presence of 0.5 M NaBr. To summarize, when the kinetic data predict that only bromonium ions or only bromocarbocations are formed, the bromination products are obtained stereospecifically and regiospecifically, respectively, whatever the solvent. Olefin brominations involving open intermediates lead to more solvent-incorporated products in methanol or acetic acid than those involving bridged ions. This chemoselectivity can be interpreted in terms of the hard and soft acid and base theory (Dubois and Chrktien, 1978). Methanol assistance to intermediate formation also plays a role in determining product-selectivity (Ruasse et al., 1991). When the kinetics indicate that there is competition between bridged and open cations, neither the solvent-dependent stereoselectivity, nor the Table 10 Regiochemistry of the exclusive and non-stereoselective methoxybromide formation in the methanolic bromination of trans-a-methylstilbenes, XC,H,C( Me)=CHC,H,Y: rate-regioselectivity relationship." ~

~~~

X

Y

4-OMe H 4-CF3 H H 4-c1 3-CF3 4-CF3

H H H 4-c1 4-OMe 4-OMe 4-OMe 4-OH

% AdT

Yo AdS

% M

% kTb

% kTC

65*

0 0 0 0 0 20'

0 0 0 0 40 60 0 0

100 100 100 100 37 0 0 0

100 100 100 100

100 100 100 35 0 0 0

100 100

35 0 0 0

Ruasse and Argile (1983); see Scheme 9 for definition of AdT, Ads, M and k,. *From kinetics. From regiochemistry. Elimination products from the tertiary pathway. Elimination products from the secondary pathway. a

ELECTROPHILIC BROMINATION OF C=C

243

DOUBLE BONDS

regio- or chemo-selectivity is readily understood. Several factors (ionpairing, lifetimes of intermediates, conformational equilibration, nucleophilic assistance, etc.), whose effects cannot be investigated separately, probably influence them. More systematic work in this area, related to the selectivity of carbocation reactions, would be useful.

5

Kinetic substituent effects

POLAR EFFECTS OF ALKYL GROUPS

In the nineteen sixties, there was some confusion about the value of p*, the reaction constant for polar effects of alkyl groups on bromination rates, as a*. For 22 alkenes calculated according to Taft’s equation, log (k/ko) = p* 1 substituted by one, two or three linear alkyl groups, there is a satisfactory relation (16) between their reactivities, log k, in methanol and the sum of their log( k/ko) = - 3.22 C a* polar constants (Mouvier and Dubois, 1968). The correlation in which the rates vary over 6 log. units includes tetramethylethylene, the most reactive alkene. The additivity of substituent effects, demonstrated by ( 16), was the first kinetic evidence for a symmetrical charge distribution in the rate-limiting transition states of alkene bromination. As expected, deviations from (16) are found for alkenes bearing one or several branched groups. In order to take into account the steric effects, the E,. authors applied the extended Taft relationship, log(k/k,) = p * C a* 6 1 The resulting correlation is (17) where p*, now -5.4, is considerably more

+

log (k/ko)

=

-

5.43C a*

+ 0.96 C E ,

negative than the value in (16). Faced with this contradiction, Dubois and Bienvenue ( 1968b) chose a series of ethylenic compounds bearing single non-conjugated linear heteropolar groups whose a*-parameters cover a range wider than those of alkyl groups and whose steric parameters ( E , ) hardly vary (Table 11). The resulting p*-value in (18) is -3.1, in agreement with that found in (16). This result was later confirmed by including in the correlation ( 18) additional alkenes bearing the same heteropolar groups

M.-F. RUASSE

244

Table 11 Determination of the effects of alkyl groups on the rates of bromination of CH2= CH-CH,X in methanol" at 25°C. X n-Pr CH,OH CH,Ph CH,OEt CH,OPh CH,OCOMe CH,CI CH,CN

log k b -0.115 0.32 +0.215 +0.495 +0.8 1 0.69 + 1.05 + 1.13

+

+

- 0.36 -

-0.38 -0.19 -0.33

-0.24

"Dubois and Bienvenue-Goetz (1968b).bk in 'See Shorter (1982).

2.59 1.82 1.72 1.30 0.48 - 0.07 -0.78 - 1.50 M - s~ - l .

associated with one or two methyl groups ( Bienvenue-Goetz and Dubois, 1978). It was shown (Dubois and Bienvenue-Goetz, 1968b) that (17) and the very negative p*-value come from an artefact due to the occurrence of a relationship between o* and E , for linear alkyl groups ( E , = 2.520*). The p*-value of - 3.1 from ( 16) and (18) can therefore be considered as reliable and as a generally valid expression of the polar effect in bromination. Hyperconjugation does not seem to have much effect on alkene reactivity towards bromine, since (16) applies whatever the number of alkyl groups on the double bond. However, only cis-olefins are involved in this correlation. To include geminally substituted olefins, an additional term is necessary, as in (19) where d is unity for the gem-disubstituted compounds and zero for log k = - 3.03

1 + 0.43d + 6.89 O*

(19)

the others. This term can be associated with hyperconjugative effects (Bienvenue-Goetz and Dubois, 1978). Because of the absence of any obvious reference value, the p*-value of -3.1 is not readily discussed in terms of charge magnitude or brominebridging at the rate-limiting transition states. For alkene hydration, it is now accepted that the intermediates are carbocations (20). The corresponding structure-reactivity relationship (21) is obtained by using o,' and 0,'

log

k H + = - 12.3[0;

+ 0.60(02 + 0.080 - 0.084)] - 10.1

(21)

ELECTROPHILIC BROMINATION OF C-C

DOUBLE BONDS

245

substituent constants for R, and R,, respectively (Knittel and Tidwell, 1977), and not og-constants. In (21), the D-term related to hyperconjugation represents double bond stabilization by R, . There are more similarities between bromination and arylsulfenyl chloride addition leading to a thiiranium ion (Scheme 10).The p*-values of - 1.00 to - 1.8 for sulfenylation (Ruasse et al., 1979; Schmid et al., 1977a; Collin et al., 1969) are significantly Ar

I

Scheme 10

less negative than that for bromination, suggesting, in agreement with CNDO calculations (Bach and Henneike, 1970), that the carbon atoms of thiiranium ions are less charged than those of bromonium ions. Surprisingly, the absolute value of p* for bromination in methanol is significantly greater than that for solvent-assisted solvolysis of unbranched secondary alkyl derivatives in 20% aqueous ethanol (p* = -1.17; Bentley et al., 1981). It is difficult, however, to compare these two cation-forming reactions, since the solvent assistance to charge development in the ionization processes is very different (p. 273). The p*-value for bromination established in methanol is also valid in a variety of other protic solvents. Linear correlations between the bromination rates of unbranched alkenes in methanol and those in a 7&30 methanol-water mixture (M70) [(22); Barbier and Dubois, 19681, in pure water C(23); Bienvenue-Goetz and Dubois, 19681 and in acetic acid [(24); Ruasse and Zhang, 1984) are observed. An approximately linear relationship (25) between log

kM70

+

= 0.90 log ~ M ~ O H2.6

(22)

the rates in methanol and those in tetrachloroethane (TCE), where the kinetics are respectively first- and second-order in bromine, is also found (Modro et al., 1977). This almost insignificant solvent sensitivity of p* in bromination is surprising in view of the solvent dependence of p* in solvolysis (Bentley et al., 1981), where the reaction constant changes from -2.1 to -4.28 and to -9.1 on going from acetic acid to water and hexafluoroisopropyl alcohol,

246

M.-F. RUASSE

respectively. This different behaviour is not fully understood (p. 273) but suggests that bromination data in non-nucleophilic fluorinated alcohols would be useful. The kinetic effect of trans-S-alkyl or heteropolar groups R on bromination of styrenes, PhCH=CHR, has also been investigated and compared with the same R-effects on ethylene (Bienvenue-Goetz and Dubois, 1975). A fairly linear relationship (26) between the two sets of data is observed. The

attenuation of the polar R-effect can be attributed to differences in charge distribution at the rate-limiting transition states. Those for styrenes are carbocation-like [ 161 whereas in the alkene series they are bromonium ion-like [17]. It has been argued (Schmid et al., 1977a) that the p*o* correlations on the same olefins challenge this interpretation.

STERIC EFFECTS OF ALKYL GROUPS

Steric effects of alkyl groups are more difficult to describe quantitatively, since, in any electrophilic addition, up to four substituents on a double bond can interact with the entering electrophile and with each other. After the failure of the extended Taft analysis of bromination data for mono- to tri-substituted alkenes (p. 243), Grosjean et al. (1976) chose a series of tetraalkylethylenes, Me,C=CMeR,, some R, being branched substituents, and found an acceptable relationship (27) using Taft’s steric parameters only. A more general correlation (28), including additional tetrasubstituted ethylenes

1

log ( k / k o ) = 1.29 E: with linear groups only, was obtained when Hancock’s E: parameters (Hancock et al., 1961) were used. It is difficult to understand why the steric effects are additive, when they are generally non-additive (Shorter, 1972,1982;

ELECTROPHILIC BROMINATION OF C=C

247

DOUBLE BONDS

Gallo, 1983), and how the polar effects disappear totally; this result is probably fortuitous and again due to the relationship between (r* and E,. It must therefore be concluded that parameter scales are inadequate to describe the kinetic influence of alkyl groups in bromination and in electrophilic additions in general. Qualitatively speaking, it is clear that retarding steric effects very easily outweigh the accelerating polar effects. For example, the highest bromination rate is found for tetramethylethylene, since replacing one methyl by an ethyl slows the reaction slightly. Ultimately, the accumulation of steric effects leads to totally unreactive alkenes, such as tetraneopentylethylene (Andersen et al., 1985) and tetraisopropylethylene (Langler and Tidwell, 1975). Moreover, it appears from the data of Table 12 that the retarding effect of a branched group depends on the other double bond substituents; the more crowded the ethylenic bond, the higher the steric retardation caused by a t-butyl group for example. It has been claimed (Grosjean et al., 1976) that interactions between the entering bromine and substituents control the bromination rates; in addition, Table 12 shows that substituent-substituent interactions also play a role in determining the reactivity. These latter interactions between non-bonded groups could lock the branched alkyl groups in conformations which would depend on the rest of the alkene. However MM calculations show (Ruasse et al., 1990) that the favoured conformations of R are the same in tetraalkyl- and adamantyl-ethylenes. It is possible, however, that conformational changes occur on the charged transition states, which are more congested than the ground states. Data on the conformations of R in the corresponding bromonium ions would be useful in this respect. Whatever the case, topological treatments, such as DARC-PELCO analysis (Dubois et al., 1981b) or the branching equation of Charton (Charton and Charton, 1973) may be more suitable for describing large sets of bromination data. Table 12 Alkene reactivity" towards bromine: dependence of steric retardation by R o n the double-bond crowding.

H Me Et i-Pr t-Bu neo-Pe

0.67 2.60 2.69 2.51 1.98 1.78

4.57 6.11 6.13 5.66 5.22

5.02

6.11 7.15 7.13 6.27 6.23 4.61

5.62 6.20 6.08 4.92 3.20 3.96

"log k in MeOH at 25°C with k in M - ' s - ~ for bromine addition. bDubois and Mouvier (1968). 'Grosjean et a/. (1976). dRuasse et al. (1990).

248

M:F.

RUASSE

The high sensitivity of bromination to steric effects has also been discussed in terms of open or bridged rate-limiting activated complexes (Schmid and Tidwell, 1978). Since transition states for hydration and sulfenylation are unambiguously open and bridged respectively, it was suggested that the structure of bromination transition states can be determined by comparing the kinetic effects of alkyl groups on the three electrophilic additions. Unfortunately, the corresponding log/log correlations are so poor that they are inconclusive. This disappointing result is not surprising in view of the balance between polar, steric and hyperconjugative effects, which is noticeably different for the three reactions (Ruasse et al., 1979). Sulfenylation is much less sensitive to polar effects than bromination (p&, = -3.10 and pXrscl = - 1.0). Hyperconjugation is important in hydration but not in bromination (p. 245). Finally, the steric requirements of the hydronium ion, bromine and arylsulphenyl chloride differ widely. In particular, the steric contribution of two t-butyl groups depends on their relative positions and also on the entering electrophile (Scheme 11;R = t - Bu). As a consequence,

HAHBr R

cis

R

HAR Br R

trans

H

gem

Scheme 11

the k,is/k,,ons-ratios, which have sometimes been used to estimate the magnitude of steric effects (Chang and Tidwell, 1978), cannot be interpreted independently of the electrophile. In addition to the fact that steric crowding can slow the reaction by hindering bromine approach to the double bond, it appears now that bulky substituents can modify the bromination mechanism by inhibiting nucleophilic solvent assistance to ionization of the CTC and/or nucleophilic trapping of the ionic intermediates. Assistance to the rate-limiting ionization step by

ELECTROPHILIC BROMINATION OF C=C

2

3

L

5

DOUBLE BONDS

249

logk,MeOH8

Fig. 9 Comparison of polar and steric effects of alkyl groups on bromination rates branched (0) and adamantyl (A) alkenes in acetic acid and in methanol of linear (a), (Ruasse and Zhang, 1984; Ruasse et al., 1990). Polar effects are identical in both solvents [full line, eq. (24)], but steric effects differ. Deviations of branched alkenes are attributed to steric inhibition of nucleophilic solvation by methanol.

nucleophilic solvents was first suggested by systematic deviations of crowded alkenes from the correlation between the rates of bromination of linear olefins in methanol and in the less nucleophilic acetic acid (Fig. 9; Ruasse and Zhang, 1984). Since the bromination of highly congested unsaturated compounds cannot be assisted by the solvent, due to inhibition of solvent approach to the double bond, their rates are smaller than those for the assisted reaction of unencumbered ethylenic substrates. The more crowded the double bond, the less important the solvent participation is (p. 272). Consequently, part of the retardation induced by branched alkyl groups results from inhibition of solvent assistance. In particular, the large steric effect observed in adamantylidenealkanes (Fig. 9; Ruasse et al., 1990) can be interpreted in these terms. The rate of the product-forming step is generally considered to be very fast, since it is an anion-cation reaction. However, bulky substituents can also slow this last step by making nucleophilic attack on the bromonium ion difficult. The most famous example of steric inhibition of.product formation is the bromination (29) of adamantylideneadamantane in carbon tetrachloride (Strating et al., 1969). Bromine adds to this alkene, i.e. the electrophile can Br;

M.-F. RUASSE

250

Br, Br-, C / \

k Products

C’ / \

AGh AG A k - i << k,

B

kN < k - i

L L RC RC Fig. 10 Free energy profiles of bromination: (A) when bromonium ion formation is irreversible and ( B ) when it is reversible.

approach the congested double bond, but the reaction stops at the bromonium ion, bromide attack on this ion being totally inhibited. It therefore appears that steric effects must be discussed not only in terms of bromonium ion formation but also of bromonium ion trapping. When the last, productforming, step is significantly slower than the first step, the bromonium ion can be formed reversibly, as shown in Fig. 10. This has been found recently in the bromination of highly congested alkenes such as tetraisobutylethylene (TIBE [ 181; Brown et al., 1990) and adamantylidenealkanes (Ad=CRR’ [ 191; Ruasse et al., 1991).

TIBE

[I81

Ad=CRR

c 191

There are only a few studies of the bromination products of congested alkenes. Such products generally consist of the corresponding allylic bromoderivatives, which are consistent with /%proton elimination by the counter-ion from the bromonium ion. For example, the ionic bromination of octamethylcyclopentene in CC14 leads exclusively to 1,2-di(bromomethyl)hexamethylcyclopentene as in Scheme 12 (Mayr et al., 1986). Bromine addition (30) to

ELECTROPHILIC BROMINATION OF C=C

251

DOUBLE BONDS

Br -HBr

d

/

J

- G-G Br -

CH,Br

-HBr

Scheme 12

1,2-dimethyldi-t-butylethylenein CC1, affords the 3,4-di(bromomethyl)2,2,5,5-tetramethylhexene-3,in agreement with a mechanism similar to that of Scheme 12 (Lenoir, 1978). It is also noteworthy that isopropylideneadamantane is chlorinated by t-butyl hypochlorite (31) in the allylic position (Meijer et al., 1982).

To sum up, the rate retardation attributed to steric effects of bulky alkyl groups can arise from substituent-electrophile, substituent-substituent and substituent-solvent interactions in the first ionization step of the reaction and also from substituent-nucleophile interactions in the product-forming step. It is therefore not surprising that the usual structure-reactivity correlations or even simpler log/log relationships cannot satisfactorily describe the kinetic effects of alkyl groups in the electrophilic bromination of alkenes. ,

M:F.

252

RUASSE

KINETIC EFFECTS OF ARYL SUBSTITUENTS

Early values of p + for the ring-substituent effect in the -4.5 range were obtained for styrene bromination in methanol (Dubois and Schwarcz, 1964) and in acetic acid (Rolston and Yates, 1969a). The magnitude of p + , closely similar to that for t-cumyl chloride methanolysis, -4.82, was one of the first pieces of evidence for open P-bromocarbocations as intermediates in this electrophilic addition. It is noteworthy that the same value was obtained with different substituents, in methanol with electron-donors and in acetic acid with electron-acceptors. This was not considered inconsistent, since at that time the reaction constant of the a+-defining reaction was not known to exhibit significant solvent dependence (Stock and Brown, 1963).However, the absolute magnitude of p + for styrene bromination in acetic acid became higher and higher (-4.21, Rolston and Yates, 1969a; -4.71, Pincock and Yates, 1970; -4.83, Yates et al., 1973) as the substituent set was extended toward the more negative a+-constants; in methanol an upward curvature characteristic of a mechanistic change was found for the complete p+a+relationship including strong electron donors (4-OH) and acceptors [ 33(CF,),] (Ruasse et al., 1978). The quantitative analysis of structure-reactivity relationships and the measurements of significant p-values for ring substituent effects on arylolefins must take into account several related problems: (i) the change of the intermediates from open carbocations to bridged bromonium ions as the aromatic ring is less able to stabilize the positive charge on the benzylic carbon atom; (ii) a variable balance between polar and resonance effects which can be expressed by Yukawa-Tsuno (32) or Young-Jencks (33) equations (Yukawa et al., 1966; Young and Jencks, 1979) and which results log(k/ko) = pn(an+ R Ao') log(k/ko) = pnan+ p r ACT+

(32) (33)

partly from ring rotation out of the plane of the double bond or of the vacant orbital of the developing carbocation, and partly from resonance saturation; (iii) a solvent effect on p, which would arise from transition-state shifts as the stability of the intermediate changes; and, finally, (iv) a dependence of p on the other substituents attached to the double bond, which can be handled in terms of interaction coefficients.

ELECTROPHILIC BROMINATION OF C=C

253

DOUBLE BONDS

The question of bridged and/or open intermediates has been considered in Section 4, where the data on kinetic substituent effects were discussed with the help of the multipathway scheme (Scheme 7) to determine the relative importance of bromonium and carbocation paths. It is not straightforward to obtain significant p-values for each of them from the complex pa relationship (34) and (35) corresponding to this scheme for an a,fl-R,,R,

(34)

disubstituted olefin. The general procedure involves, first, the search for limiting situations, where it can be reasonably assumed that only one pathway contributes, in drder to estimate the reaction constant range, and, secondly, an iterative adjustment that selects the p-values of (35) most closely corresponding to the experimental data. Values collected in Table 13 have been obtained by this analysis for styrenes ( ArCH=CH,) and stilbenes (ArCH=CHAr'). For these latter series it has been assumed that the balance Table 13 Values of p and m ' for ring-substituent and solvent effects in the bromination of aromatic olefins trans-Ar-C( R)=CHR' in methanol at 25°C. R

R'

logkx=H

pnb

H Me H OMe

H H Me H

3.18 5.14 3.51 9.32

-4.80' -4.26' -4.72 - 1.58

H Me

Ph Ph

0.63 2.65

Ph

H

4.62

P'

PS

Styrenes -4.8 -2.8 -4.7 0.0 Stilbenes -1.6 -5.52 -5.4 -1.7 -4.0 -4.87 1,l-Diphenylethylenes -2.9 -3.579

pB

mW-G

Ref.

-1.8

0.96' 1.01f

h

-

-1.7

-

h

h

0.45

i

-

1.20' 1.03f

k

-

1.04f

I

-

-1.0

i

p-values calculated from the multipathway mechanism; mw_,-values calculated using the Y-parameters of the Winstein-Grunwald scale. bpn and p' for polar and resonance effects respectively, calculated using the Young-Jencks equation, when p" = p', p" = p ' . 'In AcOH, p" = -5.7 (Bodrikov et al., 1973). dRuasseand Dubois (1975). 'In AcOH, p" = -5.76. 'Ruasse and Lefebvre (1984). % AcOH, p" = -4.45. References: *Ruasse et al. (1978); 'Ruasse and Dubois (1984); JRuasse and Dubois (1974); 'Ruasse and Argile (1983); 'Dubois et al. (1972a).

2 54

M.-F. RUASSE

between polar and resonance effects does not differ from that in the reaction defining o+,i.e. there is no need to use a Young-Jencks equation. For styrene bromination in methanol, application of the dual parameter equation (33) does not statistically improve the initial p+a+ correlation (Ruasse et al., 1978); for stilbenes, (33) gives - 1.9 and - 5.9 for p" and pr respectively, two senseless values from a mechanistic point of view since resonance effects would be anomalously high as compared with analogous reactions (Ruasse and Dubois, 1972). In contrast, polar and resonance effects must be separated in order to analyze the data for a-substituted arylolefins [ArC(R)=CHR with R # HI . Their bromination involves open carbocation intermediates only. Resonance effects cannot be fully developed at the transition states, since the aromatic ring is not in the same plane as that of the developing carbocation, because of steric constraints. Accordingly, application of (33) gives p' < p". Attenuation of resonance arises mainly from stereochemical factors, at least in the monosubstituted 1,l-diphenylethylene [201 and a-methylstilbene [21] series; the p'/p" ratios can be related to the dihedral angle between the substituted phenyl ring and the plane of the ethylenic bond.

While p* for the polar effects of alkyl groups does not vary with the solvent, p" for the polar effects of aromatic substituents is solvent-dependent. For example, there is a fairly linear log/log relationship between the bromination rates of styrenes in acetic acid and methanol (36) with a slope higher than unity, to be compared with 0.99 in (24). The p+-value for styrenes

in acetic acid is therefore -5.7. This agrees fairly well with that measured for the reaction of the same olefins in the same solvent but with N bromosuccinimide instead of free bromine (Bodrikov et al., 1973). Similar results have also been obtained for a-methylstyrenes and 1,l-diphenylethylenes, their p-values being significantly more negative in acetic acid than in methanol (Ruasse and Lefebvre, 1984). This solvent-sensitivity of p" and p' can be attributed to the fact that nucleophilic solvent assistance is

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

255

substituent-dependent and that it increases as the charge delocalizing ability of the phenyl ring decreases. However, it can also arise from transition-state shifts due to variable solvent stabilization of the carbocationic intermediates, as discussed below. Finally, as shown in Table 13, p for an aromatic ring is also strongly dependent on the other substituents at the double bond; it varies from - 1.6 to - 5.5 on going from a-methoxystyrenes to stilbenes. This variation, which is related to the well-known non-additivity of multiple substituent effects, and contrasts with what is observed for alkene bromination, is discussed in the next paragraph, devoted to substituent interaction and selectivity relationships in bromination. The p"- and p'-values for electrophilic bromine additions to arylolefins are in the same range as those for other reactions via analogous benzylic carbocations. However, generally the comparisons are only qualitative because of significant differences in the experimental conditions and in the mechanisms. For example, as has already been mentioned, the reaction constant of t-cumyl chloride methanolysis is -4.82 (Okamoto et al., 1958), i.e. slightly higher than that for a-methylstyrene bromination in methanol, where the intermediate resembles that in the solvolysis of cumyl derivatives (Scheme 13).

p+=p"=pr=-4.82

p"= -4.26; p'= -2.11

Scheme 13

The three-pathway bromination of stilbenes can interestingly be compared with the dehydration of 1,2-diarylethanols (Noyce et al., 1968), which unambiguously takes place through two CI- and p-aryl carbocations. The ratios of the two reaction constants, p , ' / p g , are very similar (Table 14), despite large differences in solvents and in the nature of the encounter complexes formed in the step preceeding the ionization. Numerous p-values for various electrophilic additions to styrene itself are available (Schmid and Garratt, 1977). Strictly speaking, the reaction constants measure only the sensitivity of the reaction to substituent effects; they depend at the same time on the solvent, on the position of the transition state on the reaction coordinate (charge magnitude) and on the way in which substituent effects are transmitted (charge location). In particular, the observed trend of p-values for the chlorination (-3.22; Yates and Leung, 1980), bromination (- 5.7; Ruasse et al., 1978) and sulfenylation (- 2.41 ;

256

M:F.

RUASSE

Table 14 Values of p for carbocation-forming dehydration" and brominationb of 1-aryl-2-phenylethane derivatives.

P,'

Dehydration

Bromination

R

H

Br

P,' PS

- 3.11 - 1.00 3.8

- 5.52 - 1.6 3.5

I Ps

Noyce et al. (1968). bRuasseand Dubois (1972).

Kharasch and Orr, 1956) of styrenes in acetic acid results from a combination of the last two factors. Bromination and chlorination go through transition states in which the benzylic carbon atom is charged; the differences between their p-values probably reflect differences in the magnitude of this charge. Sulfenylation, on the other hand, goes through a thiiranium ion, and the attenuation of p is partly due to the greater distance between the charge and the ring substituents; in fact, p for sulfenylation should be compared with pBr for bromination via bromonium ions. For these reasons, p-values for different electrophilic additions must be compared with caution, and, most frequently, the results of these comparisons cannot be interpreted quantitatively.

SELECTIVITY RELATIONSHIPS AND TRANSITION STATE SHIFTS IN ARYLOLEFIN BROMINATION

Non-additivity of substituent effects

It is well known that the kinetic effects of several substituents on one or two aromatic rings are not additive. This is exemplified in the bromination of 1,l-diphenylethylenes, stilbenes and a-methylstilbenes. The presence of a substituent, particularly one capable of electron donation by resonance, on the aromatic ring so alters the charge distribution at the transition state that the second substituent in the other ring then interacts with a charge different from that which would prevail if the substituent were alone. This is expressed

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

257

by (37), where the reaction constant for the effect of Y in the presence of a px = pH

+ qa,

(37)

fixed substituent X, p x , is related to the o-constant of X. Equation (37) is mathematically deduced from (38), which describes the non-additive effect

in terms of different reaction constants for each of the two substituents, X and Y. The general structure-reactivity relationship for multiple substituent effects, (39), is readily obtained by combining (37) and (38). Non-additivity is therefore taken into account by the last term, which involves the coefficient q, described as the interaction constant (Lee, 1992).

Because of the complexity of the kinetic data for bromination when several pathways compete, this treatment has only been applied to olefins that react by the cationic pathway. The dependence of p x for a substituent Y on the other substituent X in the three types of diarylethylenes is shown in Table 15. The p-value for the variable group depends on the fixed substituent more in the 1,l-diarylethylene series (Fig. 11)(Hegarty et al., 1972)than for stilbenes and a-methylstilbenes (Argile and Ruasse, 1983). This is expected, since, for the first olefin, the two aryl rings are conjugated with the developing charge, whereas for the others, the second substituent, insulated from this charge by an incipient bromomethylene group, interacts more weakly with the first substituent. Consequently, the q-value, the interaction constant calculated by (37), is more negative for 1,l-diarylethylenes. Data for the ethanolysis of benzhydryl chlorides (Nishida, 1967), a reaction closely related to diphenylethylene bromination, can also be analysed using (37)-( 39). The resulting pH- and q-values, -4.39 and -2.2, are in the same range as those for electrophilic addition in methanol ( - 3.57 and - 1.55). The treatment of non-additivity has also been applied to a large variety of multiple substituent effects on various reactions (Argile et al., 1984) and, in particular, to the bromination of X,Y-disubstituted benzenes where two substituents on the same ring interact strongly (Dubois et al., 1972b); the interaction constant q = - 7.98, associated with a very negative p-value, - 12.05, is much higher than those found for the bromination of arylolefins. An extended form of (39) has been used to analyse kinetic substituent effects on bromination of alkenes GR,C=CR,Rb, where G is a conjugatively electron-donating group and R is alkyl (Bienvenue-Goetz and Dubois, 1981).

258

M.-F. RUASSE

Table 15 Non-additivity of multiple substituent effects: p-dependence on X for a substituent Y and interaction constants in arylolefin bromination in methanol at 25°C. 1,l-Diphenylethylenes" logky=H -px

X 4-OMe 4-Me H 3-C1 3-NO2

x'

2.27 3.03 3.57 4.08 4.65

- 1.55

Stilbenes (carbocation pathway)b logky=~ -px

H 4-C1 3-C1 4-NO2

Yd

6.95 5.58 4.62 3.27 2.27

4

-0.02

-0.41 - 0.66 - 1.36

5.52 5.63 5.70 5.90

-0.51

a-Me-stilbenes (tertiary pathway)* log kX=H -Pa

4-Me H 3-CF3

2.90 2.65 1.72

4.69 4.87 5.17

4-OMe 4-Me H

5.95 4.04 2.65

1.19 1.46 1.66

'Hegarty et al. (1972). *Ruasse and Argile (1983). 'With electron-donating Y only. dWith electron-donating X only.

It is reasonably assumed that the reaction of these ethylenic substrates occurs via a carbocationic Ca pathway (40) only, because of the nature of G

ELECTROPHILIC BROMINATION OF C=C

259

DOUBLE BONDS

-5

px -4

It I

I

0

rr*

I

*0.5

Fig. 11 Non-additivity of the X- and Y-substituent effects on the bromination rates of 1,l-diarylethylenes (Hegarty et al., 1972). px for the variation of Y when X is fixed is linearly related to o;, according to eq. (37).

(-OCOR, -OR, -OH). [It is noteworthy that the bromination ofenol acetates does not involve ester participation in the rate-limiting step and goes exclusively through the usual bromocation ( Bienvenue-Goetz and Dubois, 1978).] As for the reactivity-structure correlation of alkene hydration (Koshy et al., 1979), a;- and a:-constants are used to describe the effect of 01- and P-substituents respectively. The carbocationic character, common to the transition states of both additions, results in highly unsymmetrical 01- and B-effects. The general equation obtained from the data for bromination in methanol, (41), involves two interactive terms expressing the interaction

between the two a-substituents and between j?- and a-substituents respectively. This equation describes satisfactorily the rates of 30 alkenes covering a reactivity range of 8 log. units. It is interesting to compare (41) with (21) found for the hydration of analogous ethylenic compounds. Equation (21) does not imply any interaction terms, i.e. the p-values, p a and ps, are constant throughout the complete reactivity range. However, a more detailed analysis shows slight variations in p-values for hydration, which are concealed by the statistical treatment (Table 16). The different behaviour of these two electrophilic additions, both of which go through carbocations, has been

260

M:F.

RUASSE

Table 16 Reaction-constant dependence on substituent G in electrophilic hydration" and brominationb of alkenes GR,C=CR,Rb. Bromination G CF, H Ph OCOMe OH OEt

Hydration

pa

Ps

- 21

- 11.8 - 12

-11 -

8.6

- 7.4

-6.5 - 4.2

-

11.8

- 9.2 - 5.0

- 3.2

Koshy et al. (1979). bBienvenue-Goetzand Dubois ( 198 1).

interpreted in terms of differences in transition state positions. More extensive comparison of data related to the reaction of similar olefins carried out later (p. 265) support:: this interpretation. Selectivity relationships and the Bema Hapothle in bromination

Equations (37)-( 39), where the non-additivity of multiple substituent effects is described by a cross-term, express correctly the rate data for bromination and other reactions of polysubstituted substrates. The question arises, therefore: has the interaction constant, q, any physicochemical meaning in terms of mechanism and transition state charge? To reply to this question, selectivity relationships (42)that relate the p-variation to the reactivity change and not to any substituent constant, have been considered (Ruasse et al., 1984). p x = alogkH

+b

(42)

The a-coefficient is related to q in (37) by (43). The plot corresponding to

the bromination data of Table 15, where results on styrenes, methylstyrenes and methoxystyrenes have been added, is shown in Fig. 12. Two discrete straight lines are obtained: line A (44) correlates styrenes and stilbenes; line B (45) 1,l-diphenylethylenes and methoxystyrenes. The difference between

ELECTROPHILIC BROMINATION

OF C=C DOUBLE BONDS

p = 0.46 log k H - 5.61

261

(45)

the two olefin groups, ArCR=CHR’, lies mainly in the nature of R, which is able to donate electrons by conjugation in set B but not in A. The conjugative ability of R enhances the sensitivity of p to the reactivity by a factor of about two. According to (46), reaction constants p are generally p=-

6 log k 60

considered as selectivity constants. The a-coefficients in (42) are then quantitative expressions of the effect of the reactivity-selectivity principle, RSP, as described by the Bema Hapothle (p. 209; Ruasse, 1988;Jencks, 1985). There is an inverse relationship between reactivity and selectivity insofar as both are related to shifts in the transition state position. The progress of the reaction at this transition state, a, is usually obtained from coefficients of Brransted or of other rate-equilibrium relationships which compare substituent effects on kinetics and thermodynamics. The p-values can also express this position if pk for rates and pt for equilibria of the same elementary step are available (47). Relationship (48), which gives the meaning of a as regards pk = apt

(47) (48)

Fig. 12 Selectivity-reactivity relationships in arylolefin bromination. Line A correlates the data for styrenes, stilbenes and their a-methyl derivatives, and line B those of 1,l-diarylethylenes and cc-methoxystyrenes (Tables 15 and 17).

262

M.-F. RUASSE

the transition-state position, is obtained by combining (42) and (47). Given CL = 6 log k / 6 log K , the a-coefficient from rates, ak,is related (49) to the same

coefficient from equilibria, a', and to the shift in the transition state position with the reactivity. In the absence of the a'-coefficient (p' = a' log K b), there is no direct way of separating the thermodynamic contribution to ak from the kinetic (transition state) one. However, the occurrence of two selectivity relationships in bromination [Fig. 12, (44) and (45)] affords the possibility of doing it. When R is a conjugated group, as in set B (45), thermodynamic and kinetic terms contribute to the a-coefficient, but in set A, where R is alkyl or hydrogen, the thermodynamic term is negligible with respect to the kinetic one. It can therefore be inferred that the smaller U-value corresponds to a transition-state shift with reactivity changes (6cr/6 log k ) , whereas the higher includes, in addition, the substituent effect on the stability of the intermediate (a' = 6p'/6 log K ) . This result is supported by the following two observations. Substituent effects on rates (Nishida, 1967) and equilibria (Mind1 and Vecera, 1972) of the heterolytic formation of benzhydryl cations, analogous to those obtained in the bromination of 1,l-diphenylethylenes, can be analysed in terms of selectivity relationships (50) and (51). Here ak is

+

pk = 0.52 log k - 2.10

(50)

p' = 0.34 log K - 0.44

(51)

markedly greater than a', showing elegantly that the second term of (49), i.e. a transition-state shift, is significant and adds to the first one, which is related only to the thermodynamic effects. Coming back to bromination, it is reasonable to suggest that the slope in (43), ak = 0.26, represents the transition-state shift with reactivity, whereas in (45)the a-coefficient is the sum of this effect and the thermodynamic contribution. On the other hand, transition-state positions in bromination can be evaluated from solvent effects and their Winstein-Grunwald rn-coefficients, since these latter are related mainly to the magnitude of the charge in the activated complexes (p. 274). The p- and rn-values for most olefins included either in selectivity relationship A (44) or in B (45) are compared in Table 17. The rn-value varies significantly with the reactivity as does p. Since rn-variations arise from transition-state shifts, p-variations necessarily come, at least in part, from the same effect.

ELECTROPHILIC BROMINATION OF C=C

263

DOUBLE BONDS

Table 17 Selectivity relationships: substitucnt" and solvent effects in bromination of arylolefins as indexes of transition state shifts with reactivity.

trans-Stilbene a-Me-stilbene Styrene trans-b-Me-styrene a-Me-styrene 1,l-Diphenylethylene a-OMe-styrene

-0.02 2.65 2.71 2.81 5.14 4.62 9.00

5.52 4.87 4.80 4.72 4.26 3.57 1.58

1.20' 1.03' 0.96' 1.01' 1.04' 0.45/

"For the carbocation pathway. bRuasse et al. (1984). 'Ruasse and Dubois (1975). dRuasse (unpublished results). 'Ruasse and Lefebvre ( 1984). IRuasse and Dubois (1984).

EARLY TRANSITION STATES IN ENOL ETHER HALOGENATION

Imbalance between polar and resonance efects in enol ether halogenation

Bromination rates of aliphatic enol ethers have been included in the interactive treatment of alkenes GR,C=CR,R,, with G being a conjugated group; most of them fit the multiparameter equation (41) satisfactorily. A more detailed analysis of reactivity-selectivity effects in the reaction of l-ethoxyethylene [22] and its a- and 8-methyl analogues [23] and [24] has been carried out, EtO EtO-CH=CH,

\

,C=CH,

EtO-CH=CHMe

Me/ c221

c231

c 241

taking the a- and 8-methyl rate ratios ka-Me/kHar,d kD-Me/kHas criteria of kinetic selectivity (Ruasse, 1985). Their bromination is diffusion-controlled in water but not in methanol and ethanol; iodination is merely fast and hydration is slow (Table 18). The selectivities, kar-Me/kH,towards the various electrophiles vary in rough agreement with RSP. They are similarly small for both halogenations, the rates of which differ by about 4 log. units, but this ratio is very large for hydration, the slowest addition. In bromination the a-methyl effect is noticeably smaller than observed for less reactive alkenes; ~ and for vinyl bromide for example, for styrene ( k = 8 x 10' M - s-') ( k = 1.8 x l O - ' ~ - ~ s - l ) i n m e t h a n o l , i t i 180and700respectively.Theeffect s

264

M.-F. RUASSE

Table 18 Reactivity, log k,” and selectivities, ka.Me/kH and k p M e / k H , of electrophilic additions to 1-ethoxyethylene.

EtOCH=CH, EtOC(Me)=CH, trans-EtOCH=CHMe krr-Me/kn

kp-Me/kn

Transition-stateindex‘

log k,,,b

log k,:

11.2 (8.34) 12.7 (9.53) 11.6 (8.69)

7.08 8.53 7.85 28

31 2.5

0.6

6

0.6

log kH,O+d

0.25 2.76 - 0.80

330 0.1 0.6’

a k i n ~ - l s - l in ’ water at 25°C. bValuescalculated from rates in MeOH, given in parentheses, and in EtOH using the Winstein-Grunwald relationship. ‘Ruasse (1985). dOyama and Tidwell (1976). ‘mwG-values for halogenations and Brransted cc-exponent for hydration. fKresge and Chen (1972).

of a P-methyl group on this addition is significantly smaller than that of an a-methyl group, indicating a dissymmetry in the charge distribution, as expected for a carbocationic transition state. But this dissymmetry is less important than that found for the slow vinyl bromide, for which k a - ~ e / k p -=~ e18. In addition, m-values for the two ether halogenations are markedly smaller than usual. In contrast, the Brernsted a-exponent for hydration (0.6) is in the same range as those measured for other alkenes (see Table 19), indicating that the proton is about half-transferred at the transition states. All these results show that the transition states for both halogenations with similar selectivities are unusually early whereas that of hydration is, as usual, midway on the reaction coordinate. It is concluded that the selectivitiesof electrophilic additions are not directly related to the reactivities but to the transition-state positions. Extensive comparison with similar data on the bromination and hydration of other ethylenic compounds bearing a conjugated group shows that this unexpected reactivity-selectivity behaviour can arise from an imbalance between polar and resonance effects (Ruasse, 1985). Increasing resonance in the ground state would make the transition state earlier and attenuate the kinetic selectivity more strongly than it enhances the reactivity. Hydration and halogenation probably respond differently to this imbalance.

Transition-state shifts with reactivity and selectivities in hydration and bromination of styrenes ArCY=CH2

A more extensive comparison between bromination and hydration has been carried out in the styrene series ArCY=CH,, where Y varies from H, Me

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

265

to OMe and OH, the rates of both additions increasing in this order (Ruasse and Dubois, 1984). For the two reactions, the rate-limiting step is the formation of a carbocation. In both cases, ring substituent effects are expressed satisfactorily by the Young-Jencks equation (33). The transition-state positions are given by m-values and Brarnsted exponents in bromination and hydration respectively. The overall results are collected in Table 19. The very small p- and m-values observed for the fast bromination of a-methoxystyrenes deserve comment since they are the smallest found for this electrophilic addition. The rates, almost but not quite diffusion-controlled, are amongst the highest. The sensitivity to polar effects of ring substituents is very attenuated but still significant; that to resonance is nil. These unusually low p-values for a reaction leading to a benzylic carbocation are accompanied by a very small sensitivity to the solvent. All these data support a very early transition state for this olefin series. Accordingly, for the still more reactive acetophenone enols, the bromination of which is diffusion-controlled, the usual sensitivity to substituents is annulled. For hydration (Dubois et al., 1981a), there is a significant decrease in the sensitivity to Y with increasing reactivity, but it is smaller than that for bromination. The p-values for these additions are linearly related to each other, but bromination is twice as sensitive to ring substituent effects compared with hydration (Fig. 13). However, constant Brarnsted a-exponents show that, in contrast to bromination, there is no marked shift in the transition state of hydration. It is therefore assumed that the p-variation in hydration comes only from a thermodynamic effect, related to a Y-dependent change in the stability of the intermediate, whereas in bromination, a transition-state shift adds to this latter effect, as expressed by (49)and (52), where log k, expresses the reactivity of PhCY=CH,. The second term in ( 5 2 ) is probably negligible in hydration apt --a-6Pk 6 log k, 6 log k,

+

6a pt 6 log k,

but not in bromination. This result would imply, in terms of the Marcus equation, that the intrinsic kinetic barrier is much higher in hydration than in bromination. In conclusion, bromination is a particularly attractive reaction for studying the origin of reactivity-selectivity effects in detail, since it is now well established that substituent and solvent effects arise not only from changes in the stability of the cationic intermediate but also from transition-state shifts, in agreement with the Bema Hapothle, i.e. RSP, Hammond postulate and Marcus effects.

Table 19 Substituent and solvent effects in bromination and protonation of styrenes ArCY=CH,.

H Me OMe OH

-4.80 -4.26( -3.2) - 1.85( - 1.5) 0 (0)

-4.80 -2.8 0 0

2.77 5.14 9.32 * 10

0.96 0.86 0.45 -

-3.6( -4.3) - 3.4( -4.05) -2.33( -2.89) - 1.57(- 1.9) ~~

~

- 3.4 - 2.4

-0.97 -0.75

-6.96 -4.02 1.74 3.24

-

0.6 0.6 0.6

~ _ _ _ _ _ _ _ _ _

"Ruasse and Dubois (1984). *Rate data for styrenes and a-Me-styrenes (from Durand et al., 1966) and for methoxy- and hydroxystyrenes (from Loudon and Berke, 1974). 'Values in parentheses are for bromination in water (Ruasse and Lefebvre, unpublished results). dGrunwald-Winstein coefficients for solvent effects. 'Values in parentheses are for protonation in MeOH (Toullec, 1979; Dubois et al., 1981b). IBr~instedexponents.

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

267

Y

-C 0

d

.

Me

0

.-C

5.

n

-

Me0 0. 0

I pH*[ Hydration

Fig. 13 Comparison of p values for bromination and hydration of styrenes XC,H,C( Y )=CH, with variable Y. Selectivity coefficients for bromination depend on Y twice as much as for hydration (Ruasse and Dubois, 1984).

6

Solvent effects and solvation in brornination

Along with the previous approaches for evaluating transition-state shifts from substituent effects, solvent effects have frequently been referred to as indexes of charge development in bromination. This is not surprising, since physical organic chemistry considers that the sensitivity of any reaction rate to the solvent or to a substituent arises mainly from changes in charge separation along the reaction pathway. However, specific solvent/transition state interactions make this generality particularly relevant to bromination. It was shown quite early (Garnier et al., 1971) and confirmed later (Ruasse and Motallebi, 1988),from solvent isotope effects ( k M e O H / k M e O D ) on the rates of 1-pentene bromination and from a linear relationship (53) between these AG& = 0.86 AG,,,

+ 12.3

(53)

rates and the bromide transfer energies in a series of water-methanol mixtures, that the main role of the solvent is electrophilic solvent assistance to bromide ion departure.

M:F.

268

RUASSE

In other words, the magnitude of protic solvent effects is directly related to the magnitude of the negative charge on the developing bromide ion at the transition state. Since the positive charge is necessarily equal to the negative charge, solvent effects provide a simple measurement of the charge separation. Accordingly, two kinds of experimental data, kinetic solvent isotope effects and the m-values of Winstein-Grunwald correlations, have been used to obtain transition state indexes for a large variety of ethylenic compounds. However, the results from substituent and solvent effects for any olefin are not as consistent as expected, since both are complicated by other factors; these are mainly charge delocalization for the first and nucleophilic solvent assistance for the second. KINETIC SOLVENT ISOTOPE EFFECTS

Most of the available KSIEs measured in methanol, ethanol or acetic acid and their deuteriated analogues are shown in Tables 20 and 21. They must be compared with the &- factor [ 1.35 and 1.27 in methanol and ethanol, Table 20 Solvent effects in alkene bromination: dependence of the electrophilic (KSIEs) and nucleophilic assistances ( R ) on the alkene.

KSIE" Ad=CHZd*' t-Bu-i-PrC=CHZf Me,C=CMe-i-Prf cis-MeCH=CH-t-Bu'

Branched alkenes 1.30(EtOL) 1.11 1.32(MeOL)g 1.08 1.07 1.02

0.9 1.4 1.o 1.5

5.62 2.29 6.38 4.04

PhCH,CH=CH,' n-PrCH=CH,' cis-EtCH=CHEt Me,C=CMeZf

Linear alkenes 1.37(MeOL) 0.80 1.35(MeOL) 0.92 1 . 2 3 ( A ~ o L ) ~ 0.92 0.96

8.3 6.2 4.1 3.3

1.72 2.60 4.54 7.16

0.8 0.6 0.4

7.40 5.76 3.15

f

Congested alkenes Ad=CMe,' Ad=CMe-i-Pr' Ad=CH-t-Bu'

-

1.13(EtOL) 1.29(MeOL)

0.63 0.76 0.85

Kinetic solvent isotope effect as a measure of electrophilic assistance to bromide ion departure; limiting values: &,- = 1.35 in MeOL and EtOL at 25°C. bObtained from rate data in ethanol, methanol and their aqueous mixtures using Bentley's Y, scale; its decrease corresponds to the involvement of nucleophilic assistance. 'R = ( kaqEtOH/kAcOH),,as a measure of nucleophilic solvent assistance. dModel for a limiting bromination mechanism. 'Ruasse et al. (1991). 'Ruasse and Zhang (1984). gArgile and Ruasse (to be published). hModro ef al. (1979). a

ELECTROPHILIC BROMINATION

OF C=C DOUBLE BONDS

269

Table 21 Solvent effects in bromination of conjugated olefins: transitionstate shifts and nucleophilic solvent assistance.

Styrened a-Me-styrenee a-OMe-st yrene’ Stilbened cr-Me-stilbene’ 1,l-Diphenylethylene“ EtOCH=CH,g

-

1.04(MeOL)’ -

1.02(MeOL)/ -

0.80 0.81 0.36 0.86 0.83 0.83 0.48

27 5.4 -

100 5.9 2.6 -

3.18 5.14 9.3 1 1.04 2.65 4.62 8.35

Kinetic solvent isotope effect as a measure of electrophilic solvent assistance to bromide ion departure; limiting value: & - = 1.35 in MeOL. bFrom Y,, and kinetic data in water, methanol, ethanol and their aqueous mixtures. ‘( kaqEtOH/kAcOH)Y;measurement of nucleophilic solvent assistance. dCalculated from data in Ruasse and Dubois (1975) and Ruasse et al. (1978). ‘Ruasse and Lefebvre ( 1984). ’Ruasse and Lefebvre (unpublished results). @Ruasse( 1985). a

respectively (Ruasse and Motallebi, 1988)], which is the limiting KSIE value corresponding to the effect on the solvation of a fully developed bromide ion. It is noticeable (Table 20) that for most aliphatic alkenes, the KSIEs are very high, close to & - . These results are valuable evidence for late transition states [25] in the bromination of these olefins. In contrast, the

c251 KSIEs for the reaction of aromatic olefins, 1,l-diphenylethylene and or-methylstyrene (Table 21 ) are significantly smaller; they can be related to transition states earlier than those in the aliphatic series. Unfortunately, for the reactions of highly reactive aromatic olefins or enol ethers, whose low sensitivity to solvent and substituent effects indicates very early transition states, there are not enough KSIE data to confirm this conclusion. The only series of alkenes that show anomalously low RSIEs is that of highly congested adamantylidenealkanes (last rows of Table 20). The decrease in the isotope effect, associated with particularly small m-values and the absence of nucleophilic solvent assistance, can be attributed to reversible

270

M.-F. RUASSE

formation of strained bromonium ions, the nucleophilic trapping of which is strongly inhibited (p. 273).

THE Ysr SCALE FOR BROMINATION

The second series of data on protic solvent effects in bromination that are related to transition states comprises the m-values of solvent-reactivity correlations. First, it is important to underline that Y-parameters, the solvent ionizing powers, established from solvolytic displacements, work fairly well in this electrophilic addition. This is expected since bromination, like S, 1 reactions, leads to a cation-anion pair by heterolytic dissociation of the bromine-olefin CTC, a process similar to the ionization of halogenated or ether derivatives (Scheme 14).

R-X

+

R',

X-

Scheme 14

The first Winstein-Grunwald correlations for bromination (Garnier and Dubois, 1968; Ruasse and Dubois, 1975) used the initial Y-values derived from t-butyl chloride solvolysis (Fainberg and Winstein, 1956). Since that time, it has been recognized that this Y-defining reaction does not proceed through a pure SN1 mechanism (Raber et al., 1970). More recently, Bentley and Llewellyn ( 1990) established new Y,-scales from 2-adamantyl substrates involving various leaving groups X. These scales correspond to an ionization process where the solvent interacts with the transition state electrophilically but not nucleophilically, back-side obstruction of the breaking bond by the adamantyl group inhibiting nucleophilic attack by the solvent. The analysis of these Y,-scales (Bentley et al., 1984) in terms of the Kamlet-Abboud-Taft equation (54)(Kamlet et al., 1981), which dissects the solvent effect into log (ks/k,)

= S~T*

+ uu + bp

(54)

three contributions (dipolarity, z*, hydrogen-bond donor a and hydrogenbond acceptor p) shows that Y, describes mainly the solvent dipolarity, often called the medium effect. However, a significant contribution from the hydrogen-bond donor ability cannot be excluded, since the Y,-values depend slightly on X, the leaving

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

27 1

group. As regards bromination, the Y,,-scale derived from either 1- or 2-bromoadamantane solvolysis (Bentley and Carter, 1982) is the most relevant since the leaving group is a bromide ion in both reactions. In addition to electrostatic medium effects and to electrophilic assistance described by Y, nucleophilic solvents can assist positive charge development in heterolysis of secondary, and sometimes tertiary, substrates via an S,2 (intermediate) mechanism (Bentley et al., 1981). Equation ( 5 5 ) , involving log ( k s / k s o )= rnY + IN

(55)

both ionizing power and nucleophilicity N , must therefore be considered in its totality (Schadt et al., 1976). A rigorous approach to the several roles of the solvent in ionization reactions must therefore involve rate measurements in a number of media of variable Y and N . In particular, fluorinated solvents with very low nucleophilicities must be added to the usual nucleophilic alcohols in order to evaluate the solvent involvement described by 1N.This cannot be done systematically for bromination because the rates for most olefins in trifluoroacetic acid, trifluoroethanol or hexafluoroethanol are too fast to be measured. Consequently, in order to obtain information about the several solvent contributions, the following procedure has been used (Ruasse et al., 1991). The rn,,-values are first obtained using YB,-parameters(Bentley and Carter, 1982)of solvents of closely similar nucleophilicities (H,O, EtOH, MeOH and their aqueous mixtures). These values depend both on the nucleophilic solvent involvement and on the magnitude of the transition-state charge. An increase in nucleophilic assistance and an early transition state decrease mBr, as compared with its maximum value for the unassisted formation of a bromonium ion via a late transition state. In order to determine whether smaller rn-values must be attributed to charge decrease and/or to nucleophilic solvent involvement, complementary data must be obtained. Assistance can be estimated by R-ratios (Fig. 14), which compare the rates in two solvents of similar ionizing power but different nucleophilicities, such as the acetic acid-aqueous ethanol couple (Bentley and Schleyer, 1976). When measurements are possible in trifluoroethanol, TFE, R-values can also be obtained from the TFE-aqueous methanol pair in a higher reactivity range. Transition-state charges are available from KSIEs. Consequently, for the bromination of a given olefin, most of the relevant information about solvent effects can be obtained by combining KSIE, rnBr and R . The rigorous analogy between electrophilic halogenation and solvolytic displacements was recognized only recently (Ruasse et al., 1991).Consequently, for most of the previously published solvent effects, the above procedure has not been applied systematically.Winstein-Grunwald, YW-G,parameters have

M.-F. RUASSE

272

8

6 .

PhCHLCH=CH2 rn =0.8

L .

2 .

L

-2

-I

0

I

2

3

4

Fig. 14 Typical log k / Y,, plots for assisted and unassisted alkene brominations. Allylbenzene and 1-pentene, less crowded than cis-methyl-t-butylethylene and methylideneadamantane, exhibit the smallest m-values. The points corresponding to acetic acid (0) and trifluoroethanol (A),two weakly nucleophilic solvents, are below the regression line for water, methanol, ethanol and their aqueous mixtures ( 0 )of similar nucleophilicity. In contrast, they are on the line for the branched alkenes where steric crowding inhibits nucleophilic assistance by alcoholic solvents (Ruasse et al., 1991, Ruasse and Motallebi, 1991).

been used generally to obtain mW-G coefficients. Since there is a fairly linear relationship between Yw-G and Y,, for nucleophilic aqueous ethanol and methanol mixtures, mBr can be estimated from mw-G = 1.25~1,~.But these calculated m,,-values are not compatible with the method described above, because non-nucleophilic solvents, acetic acid in particular, are generally included in the mW-G calculations. Consequently, data shown in Tables 20 and 21 are not quite those given in the references quoted; they have been recalculated in agreement with the more recent procedure.

VALUES OF m,, IN ALKENE BROMINATION: NUCLEOPHILIC AND ELECTROPHILIC ASSISTANCE BY PROTIC SOLVENTS

The first step in the interpretation of m,,-values has consisted in searching for an alkene for which nucleophilic solvent assistance is impossible. Methylideneadamantane, in which steric constraints can restrict nucleophile

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

273

approach to the developing bromonium ion, was chosen by analogy with 1- or 2-bromoadamantane, whose solvolyses define the YBr-scale(Ruasse and Motallebi, 1988). With this alkene, a limiting mBr-valueof 1.1 is measured from ethanol, methanol and their aqueous mixtures. Similarly high mBr are obtained (Table 20) when the double bond is substituted by branched groups. The solvent is not nucleophilically involved since the R-ratios are close to unity. Their transition states are very late, as shown by the large KSIEs. For less congested alkenes, values of mBr are significantly smaller; they are associated with R-values markedly greater than unity, and the transition states are still very late. Consequently, the attenuation of mBr does not arise from a smaller charge development in the activated complex but from nucleophilic participation of the solvent [26]. It is noteworthy that this specific

" /

---Br---LOS

solvent-transition-state interaction seems to be reactivity-dependent; the higher the reactivity, the smaller the assistance. This agrees with the usual interpretation of charge stabilization in terms of electronic demand; the greater the stability of a cationic species, the lower its demand on substituents or solvent to assist the charge development. In the solvolysis of secondary alkyl sulfonates, competition between nucleophilic solvation and electron donation by the substituents results in a significantly solvent-dependent p*, which varies from - 9 to - 1 on going from the non-nucleophilic hexafluoro-Zpropanol to 80% aqueous ethanol (Bentley et al., 1981). In contrast, the p*-invariance for alkene bromination in H,O, M70, MeOH and AcOH [equations (22)-(25)] seems to imply a perfect balance between the two types of charge stabilization. However, this conclusion is probably risky since the nucleophilicities of the solvents implied in (22)-(25) do not vary markedly. Data in non-nucleophilic fluorinated solvents would therefore help to fill the gap in our knowledge. Finally, the last group of alkenes in Table 20 (congested alkenes) behaves very differently as regards solvent effects. The mBr,R and even KSIEs are systematically smaller than those observed for the two prevfous series. The attenuation of these coefficients can be reasonably attributed neither to earlier transition states nor to increases in nucleophilic solvent assistance. As described in the paragraph on return (p. 279), this trend is more consistent

274

M.-F. RUASSE

with the kinetic barrier for the product-forming step being higher than that of bromonium ion formation; in other words, return could be significant in the bromination of highly congested alkenes. To summarize, for the bromination of most alkenes, transition states are late and closely resemble the bromonium ions. The m,,-values reflect also the highly charged character of the activated complexes, since they are always in a range (0.8- 1.1) that is considered in solvolysis as indicating cationic intermediates. However, as in S,2( intermediate) solvolysis, mBr varies significantly because of nucleophilic solvent assistance. The greater the assistance, the smaller the mBr-values. In particular, this latter solvent contribution strongly depends on the crowding of the double bond and disappears when there are moderately bulky substituents.

VALUES OF rnB, AND TRANSITION-STATESHIFTS IN BROMINATION OF CONJUGATED OLEFINS

In contrast with simple alkenes, mBr varies markedly with the structure of the olefin when a substituent is conjugated with the double bond (Table 21). There is a rough inverse relationship between mBr and log kWeOH (Fig. 15).

0.51,

, 3

,

I

5

,

,~

7 log kX,MeOH

Fig. 15 Reactivity dependence of rn, the Winstein-Grunwald coefficients for solvent effects, in the bromination of monosubstituted 1,l-diphenylethylenes (Ruasse and Lefebvre, 1984).

ELECTROPHILIC BROMINATION OF C=C

275

DOUBLE BONDS

This result, associated with those on substituent effects, supports previous conclusions to the effect that the position of the transition state depends on the reactivity in agreement with RSP. In particular, stabilization of the intermediate as a result of conjugation, such as that in the reaction of enol ethers, makes the transition state very early. The few available KSIEs also suggest that the transition states for aromatic series are earlier than those for alkenes. The most striking finding of Table 21 concerns nucleophilic assistance. The R-values are markedly greater than those found for simple alkenes. In particular, R for stilbene is as high as 100, whereas that of allylbenzene, whose reactivity is in the same range, is only 8. This can be interpreted in terms of a different charge distribution. Whereas the transition states for alkenes are bromonium-ion like, those from aromatics resemble carbocations. The charge in the latter case is on only one carbon atom, so that assistance by the solvent is more efficient than when it is distributed over a three-membered bromonium ring. This feature can be counterbalanced by steric restrictions to solvent approach. In this respect, it is noticeable that the reactions of cc-methyl olefins are only weakly assisted, as compared with those of the parent olefins. Finally, smaller R-ratios can also arise from charge delocalization in the aromatic ring; for 1,l-diphenylethylene, DPE, the assistance is less than for a-methylstyrene, EMS, which is of similar reactivity. All these factors, namely, intermediate stability, cationic charge distribution, delocalization by resonance and steric inhibition, can influence the values of both R and mBr. To distinguish between these several contributions, a systematic study of ring substituent effects on rn and R has been carried out in the DPE and a-MS series (Table 22; Ruasse and Lefebvre, 1984). In both cases, the solvent coefficients are observed to be substituent-dependent. The marked decrease in mBr and R on going from electron-acceptors to donors is greater in a-MS than in DPE. This suggests that charge delocalization by Table 22 Substituent and solvent effects in the bromination of 1,l-diphenylethylenes, DPE, and a-methylstyrenes, a-MS." 4-CF3

3-CF3

H

4-OMe

6.5 0.90 2.59

-

-

2.6 0.83 4.62 5.4 0.81 5.14

1.5 0.49 6.95 1.7 0.54 7.82

-

-

a

Ruasse and Lefebvre (1984).

-

34.2 1.03 2.96

276

M.-F. RUASSE

the second phenyl group of DPE contributes significantly to the attenuation of nucleophilic assistance and therefore of mBr. Moreover, the solventindependence of p* found for bromination of simple alkenes does not exist for conjugated olefins; for example, the p"-values for a-MS bromination are -5.76 and -4.26 in acetic acid and methanol respectively. It can be concluded that mBr depends on the magnitude of the charge at the transition state and also on its delocalization either by the substituents or by the solvent. It therefore seems difficult to separate these effects, since R, the measure of solvent assistance, depends also on the same factors. The idea that transition-state shifts contribute to m-variations is supported by substituent effects. Consequently, it would be useful to obtain p-m correlations to compare the influence of the solvent and the substituents in determining the position of the bromination transition state. BROMINE-CATALYSED BROMINATION IN NON-PROTIC AND HALOGENATED SOLVENTS

Most brominations used for synthetic purposes are carried out in halogenated solvents. There are, however, only a few mechanistic studies in these media, since it was difficult to obtain reproducible rate constants. Even now that reliable procedures have been published (Schmid et al., 1972; Bellucci et al., 1981),the upper limit of the rate constants available is about 105-106 M - ~ S - ~ since only spectroscopic techniques can be used. The most significant difference between brominations in protic and non-protic solvents concerns the kinetic law. Whereas in protic media the reaction is first-order in bromine, in halogenated media it is second-order (Bellucci et al., 1980).CTC ionization is electrophilically assisted by hydrogen bonding by a protic solvent to the leaving bromide and leads to a bromonium- bromide ion pair. In non-protic media, assistance to the bromination step is provided by a second bromine molecule, leading to a bromonium-tribromide ion pair. In other words, in protic media bromination is solvent-assisted (56) while in halogenated media it is bromine-catalysed (57).

\ /

\C' Br,

+ Br,

+

(f+>Br, Br;

(57)

,

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

277

Since no solvent molecule seems to be involved in the rate-limiting transition states of bromination in non-protic solvents, only the electrostatic character of the medium should influence the reaction rates. Consequently, there have been some attempts to correlate the kinetic solvent effects with the Kirkwood function, it being assumed that the solvent can be considered as an isotropic continuum (V'yunov and Ginak, 1981). Such attempts either failed or covered very narrow reactivity ranges. A fairly linear Kirkwood relationship was obtained recently (Bellucci et al., 1985b); it correlates the bromination rates of cyclohexene in 1,2-dichloroethane-chloroformmixtures. A variation in the dielectric constant from 4 to 10 induces a reactivity change of two orders of magnitude. Data obtained in other halogenated solvents cannot be included in this correlation, which is unique at present. Nevertheless, and despite the paucity of relevant data on both solvent and substituent effects on bromination rates in non-protic solvents, it seems reasonable to conclude that bromine addition depends mainly on the bulk medium polarity. However, this non-specific solvent property cannot be the only one responsible for the rate variations. According to (57), the main driving force for the reaction in non-protic media is the formation of a tribromide ion from bromine and the developing bromide. Kinetic (Ruasse et al., 1986) and thermodynamic (Bienvenue-Goetz et al., 1980) data on equilibrium (58) are therefore relevant to the effect of non-protic solvents on bromination rates. Br,

kr + Br- e Br; kr

Large solvent effects on the formation constant ( K = k,/k,) of the tribromide ion are observed (Table 23). The K-variation has been analysed in terms of the free energies of transfer of the bromide ion, AGtr(Br-), from acetonitrile to the appropriate solvent. Parker and Alexander ( 1968)assumed a linear relationship between AGtr(Br-) and AGf(Br;), the energy of formation of Br; according to (58), as expected if the solvation energy of neutral bromine is the same as that of the large and polarizable tribromide ion, (59). However, this assumption does not work. It was shown later that AG,(Br;)

= - AGJBr-)

+ [AG,,(Br;)

- AGtr(Br2)]

(59)

solvation of free bromine by electron-donor solvents cannot be neglected (Bienvenue-Goetz et al., 1980).A correlation (60) where this effect is expressed AG,(Br;)

= -

1.21ACt,(Br-)

+ 5.55g - 2.16

(60)

M.-F. RUASSE

Table 23 Formation constants of tribromide ion and transfer energies of bromide ionnsbat 25°C. KIM-'

Me,CO" CICH,-CH,CI~ CH,ClZb C~,CH-CHC~,~ MeCHClZb CHCl,b MeNO," MeCN" HCONMe," EtOH" MeOH" H,O"

2x >2 x >2 x 1.9 x >2 x 1.2 x 1.6 x 1 x 2x

109 107 107 105 107 105 107 107 106

x 102

1.7 x 10' 16

AGtr(Br-y

kcyclohexene

b,d

2.6

-

- 1.6

0 0.7 - 2.9 - 4.8 - 7.4

"Bienvenue et al. (1980). bBellucci et al. (1985a). 'In kcal mol-', relative to MeCN. dRate constant of cyclohexene bromination at 25°C.

by fl, the electron-donor ability of a solvent (Kamlet et al., 1981), describes the solvent dependence of K fairly well. Unfortunately, this equation is useless for halogenated media, which do not exhibit donor capability and which provide little or no solvation of the bromide ion. More relevant to the bromination mechanism could be the linear relationship (61) which has been found between the rates of Br, formation and the solvation energies of bromide in protic solvents (Ruasse et al., 1986). But again, appropriate log ( k , / k , ) = -0.07AGt,(Br-)

(61)

data are not available for non-protic solvents. It therefore turns out that a quantitative description of halogenated solvent effects on bromination rates is still inaccessible.

SOLVATION, THE DRIVING FORCE OF ELECTROPHILIC BROMINATION

Solvation is the main driving force of bromination (Ruasse and Motallebi, 199 1), since this electrophilic addition is impossible, or at least extremely 1977; difficult, in the gas phase (Angelini and Speranza, 1981; Staley et d., Sen Sharma et al., 1985). It has been calculated (Yamabe et al., 1988) that

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

279

more than 60 kcal mol- would be necessary to form a bromonium ion from ethylene and bromine; that is, the bromination rate constant of ethylene at M - ' s-'. However, in 25°C in the gas phase would be as small as water this constant is found to be 4 x lo5M - s- ' at 25°C (Bienvenue-Goetz and Dubois, 1968), which corresponds to an activation energy less than 10 kcalmol-'. In protic solvents the m,,-values discussed above show that electrostatic medium effects and electrophilic assistance to bromide ion departure are the main rate-determining factors. Free energies of bromide solvation by alcohols and water are in the range of 56-61 kcalmol-' at 25°C (Abraham and Liszi, 1978). Most of this energy contributes to the bromination rates, as shown by the high values of KSIEs (p. 268). Nevertheless, since these calculations are only approximate, it is hazardous to conclude that the medium effect is negligible. Moreover, nucleophilic protic solvents can assist positive charge development in the rate-limiting transition states. However, this contribution provides only a small to moderate acceleration, the magnitude of which depends on the double-bond substituents (p. 272). It is noteworthy that these several solvent roles are closely similar to those found in solvolytic reactions via carbocationic intermediates (Bentiey and Llewellyn, 1990). This is not surprising, since the main microscopic event of both classes of reactions is heterolysis, either of a Br-Br or a C-X bond, which seems similarly solvent-dependent. In halogenated solvents, catalysis by a second bromine molecule, which assists the Br-Br bond heterolysis, is the main driving force. The role of the solvent is electrostatic, but the absence of an extensive Kirkwood relationship suggests that there is some other kind of contribution (Bellucci et al., 1985b).

'

7

The reversible formation of bromonium ions

For a long time, it was considered that the formation of a bromonium ion from olefin and bromine is irreversible, i.e. the product-forming step, a cation-anion reaction, is very fast compared with the preceding ionization step. There was no means of checking this assumption since the usual methods-kinetic effects of salts with common and non-common ions-used in reversible carbocation-forming heterolysis (Raber et al., 1974) could not be applied in bromination, where the presence of bromide ions leads to a reacting species, the electrophilic tribromide ion. Unusual bromide ion effects in the bromination of tri-t-butylethylene (Dubois and Loizos, 1972) and a-acetoxycholestene (Calvet et al., 1983) have been interpreted in terms of return, but cannot be considered as conclusive.

280

M.-F. RUASSE

A preliminary indication that bromonium ions could be formed reversibly was provided by the reaction of adamantylideneadamantane (p. 249) leading to a highly stable bromonium-tribromide ion pair that readily releases bromine and the initial alkene (Strating et al., 1969). However, the first evidence for possible return came from the acetolysis of 2-bromocyclohexylbrosylate in the presence of bromide ions. It was shown (Brown et al., 1984) that the cyclohexylbromonium ion intermediate is able to release bromine. The drastic reaction conditions (high temperature, long duration and high bromide concentrations) cast some doubt on the generality of this observation. In fact, the analogy between the mechanisms of heterolytic nucleophilic substitutions and electrophilic bromine additions, shown by the similarity of kinetic substituent and solvent effects (Ruasse and Motallebi, 1991),tends to support Brown’s conclusion. If cationic intermediates are formed reversibly in solvolysis, analogous bromocations obtained from bromine and an ethylenic compound could also be formed reversibly. Nevertheless, return is a priori less favourable in bromination than in solvolysis because of the charge distribution in the bromocations. Return in bromination implies that the counter-ion, a bromide ion in protic solvents, attacks the bromine atom of the bromonium ion rather than a carbon atom (see [27]). Now, it is known (Galland et al., 1990) that the charge on this bromine atom is very small in bridged intermediates and obviously nil in p-bromocarbocations [281. \ /

\cI+ ,Br,n Br

\ \

/

‘C + c\ +\

I1

9

Br-

-C-Br

I

RETURN IN HALOGENATED SOLVENTS

Brown’s result was supported by later experiments in which bromonium ions were generated by bubbling gaseous hydrobromic acid through a solution of bromohydrins in halogenated solvents. Under these conditions, bromine is eliminated as it is formed, so that the resulting alkene is observed directly (Scheme 15). This method has been applied to the bromohydrins derived from cis- and trans-stilbenes (Scheme 16) and from 5H-dibenzo [a,d] cycloheptene and -azepine systems ([29a] and [29b] respectively; Scheme 17), in which steric constraints should favour elimination (path a ) as against substitution (path b).

ELECTROPHILIC BROMINATION OF C=C

J

2

28 1

DOUBLE BONDS

Br,+)=(

1

Br

-ABr

Scheme 15

Br\

,Ph

Ph

Br Ph (12%)

Ph Scheme 16

As shown in Scheme 16, erythro-2-bromo- 1,2-diphenylethanol readily gives meso-1,2-dibromo-l,2-diphenylethaneand substantial amounts of transstilbene (Bellucci et al., 1987). The slower reaction of the threo-analogue does not lead to cis- but to trans-stilbene arising from the acid-catalysed cis-trans isomerization of the initially formed olefin. In order to avoid this complication, Bellucci et al. (1988, 1991) studied the reaction of the bromohydrins derived from cyclic compounds [291 where the conformation of the two aromatic rings is maintained by a methylene [29a] or a nitrogen bridge [29bl (Scheme 17).The corresponding dibromides [301 and ethylenic compounds [31] are obtained in ratios [30]/[31] of 9/1 and 3/7 from the reaction of [29a] and [29b] respectively, with HBr in carbon tetrachloride, i.e. there is more return from bromonium ion pair [32b] than from [32a]. This trend is attributed to differences in the extent of bromine bridging in [32] being greater when Z is NCOCl than when it is CH,. The more electron-donating is Z, the less is the bromine bridging. Moreover, the yield of [31b] from [29b] is highly solvent-dependent: 30, 50 and 70% in 1,2-dichloroethane, chloroform and tetrachloride respectively. The less polar the solvent, the less stable the ion-pair [32] and the more return there is. These results show unambiguously that bromonium ions can be attacked

282

M.-F. RUASSE

,Br. + ,Brc291

/ I

~321

a: Z=CH, b: Z=NCOCI

c311

c301 Scheme 17

by bromide at the bromine atom to give free bromine and olefin. However, in the absence of data on the magnitude of bromine bridging in [32], the importance of return in the bromination of olefins [311 cannot be determined.

RETURN IN PROTIC SOLVENTS

If the formation of bromination intermediates is reversible, the experimental rate constants obtained by following bromine uptake are not those of the first ionization steps. It is therefore important to know whether return, shown to occur in halogenated media, can also occur in protic media, in which most of the kinetic data have been measured and structure- or solvent-reactivity relationships established. To obtain data on reversibility in these media, the reactions of two kinds of congested alkenes, adamantylidenealkanes (Ruasse et al., 1991) and tetraisobutylethylene (Brown et al., 1990), have been investigated. It was assumed that crowding of the bromonium ion intermediate can favour return by inhibiting its nucleophilic trapping, i.e. by enhancing the energy barrier of the final, product-forming, step (Fig. 10). Adamantylideneadamantane bromonium ion (Strating et al., 1969) reacts by returning to the initial alkene and bromine only, whereas methylideneadamantane bromonium ion goes on to bromination products only (Ruasse and Motallebi, 1988). It was therefore hoped that the introduction of branched substituents (smaller than adamantyl) into methylideneadamantane would slow the product-forming step enough to make its barrier higher than that of return but not enough to inhibit it totally.

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

283

As said before, the comparison of solvent effects on the reaction of Ad=CRR’ (R and/or R’ # H ) [33] with those for Ad=CH, and other less

a: R = R = M e b: R=t-Bu, R ’ = H c : R=i-Pr, R’=Me

TIBE

C181

C331 crowded alkenes shows that return occurs for [33) only. The association of small m,,-values (0.8 instead of 1.1) with small kinetic solvent isotope effects, KSIEs (1.1 instead of 1.3), and small R-values (0.6 instead of 6-8; Table 20) is only observed for these congested alkenes. Small R-values are expected since nucleophilic solvent assistance is impossible. Small mBr and KSIEs, indicating a smaller than usual negative charge on the counter-bromide, agree fairly well with a mechanism involving return as shown in Fig. 10B. If the transition state of the product-forming step is of higher energy than that of the ionization step, the kinetic data and the coefficients deduced therefrom are related not only to the ionization process but also to the last step as shown in (62) and (63). Since step kN is a cation-anion reaction, there is a charge decrease on going from I, its ground state, to TSN, its transition state, so that mN is necessarily negative. Consequently, the experimental m-value is smaller than mi whose magnitude is that of alkenes which react irreversibly. The significantly smaller KSIEs are also consistent with a charge decrease on the bromide, as is expected on going from I to TSN.

k

= (ki/k-i)kN

m = mi

+ mN

(63)

Evidence for a reversibly formed bromonium ion in the bromination of [ 181 in acetic acid is provided by the unexpectedly high kinetic isotope effect, KIE (Brown et al., 1990). When the eight allylic positions of [ 181 are deuteriated, the bromination rate constant decreases by a factor of 2.3. This value is too high to be attributed to a usual secondary KIE, and is more consistent with a primary effect. This result implies that the allylic protons are involved in the rate-limiting step, which cannot therefore be the bromoniumion-forming step. It has been shown that halogenation products of congested alkenes arise from a rearrangement of their halonium ions induced by halide

284

M:F.

RUASSE

attack on allylic hydrogens (Mayr et al., 1986; Meijer et al., 1982). Although the bromination products of [33] and [18] have not been identified, the results, small KSIEs and high KIE, respectively, agree fairly well with this mechanism and with the fact that the allylic rearrangement is energetically more difficult than the usual bromonium trapping by nucleophiles. Consequently, the product-forming step is rate-limiting for the bromination of these congested alkenes. The present status of return in bromination can be summarized as follows. In protic solvents there is evidence that bromonium ions are formed reversibly when they are highly congested. In the absence of crowding, no experimental data support return but none exclude it either. When bromination intermediates are a-bromocarbocations, it is highly improbable that they are formed reversibly, since the bromine atom bears no charge. In halogenated solvents the results indicate that return can occur, even for the uncongested stilbenes. Unfortunately, its importance, as measured by the k-,/kN ratio (Fig. lo), cannot be estimated. It must be noted that Bellucci’s experiments prove only that return is possible, but do not demonstrate conclusively that it occurs in bromination, since reversibility is controlled by the relative energy levels of TS, and TSN which can be affected by the reaction conditions. Now, these conditions are not the same for nucleophilic substitution on bromohydrins and for bromine addition; in particular, the counter-ions, Br- and Br; respectively, can alter the lifetime of the intermediate and thus control its partitioning between return and nucleophilic attack. Reversible formation of the ionic intermediates results from a high energy barrier for the last, product-forming, step. It is therefore readily understood that crowding of the bromonium ion promotes return. It is more difficult to understand why this barrier is high in the absence of crowding in halogenated solvents. A reasonable interpretation based on the weakly nucleophilic character of the counter-ion, the tribromide ion, has been suggested; before reacting with the bromonium ion, tribromide has to dissociate at least partly into nucleophilic bromide and free bromine, a process that is quite slow in these solvents (Ruasse and Motallebi, 1991). A possible mechanism is proposed in (64). \C/ (bBr,

F\

\ /

+ Br,

F

T>r,

F\

\ /

Br;

T>Br, Br-, Br,

+P

(64)

F\

In conclusion, the reversibility of bromonium ion formation is at present inferred from particular experiments only; nothing allows us to conclude that this mechanistic feature is general. However, when nucleophilic trapping

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

285

of the intermediate is likely to be difficult, kinetic data must be interpreted with caution because they are not necessarily related to the ionization step alone. 8

Concluding remarks

There are many significant and coherent data on the mechanism of free bromine addition to ethylenic bonds. The relative energetic contributions of the three elementary steps of the overall reaction and their dependence on the substituents and the solvent are now well known. It appears that in most cases kinetic data are mainly related to the ionization step. The preliminary formation of bromine-olefin charge transfer complexes, equivalent to encounter complexes, does not contribute significantly to the overall kinetic barrier, and is much less affected by the substituents (Ruasse, 1990) and the solvent (Brown et al., 1990) than the subsequent cation-forming step. Apart from well-identified exceptions-non-protic solvents (Bellucci et al., 1988, 1991) and highly congested double bonds (Brown et al., 1990; Ruasse et al., 1991)-the last, product-forming, step is very fast and the ionic intermediates are formed irreversibly. Thus the more recent results do not require major alteration to the mechanism postulated in the 1930s. The same cannot be said of the mechanism postulated (Bartlett and Tarbell, 1936) for tribromide ion addition in protic (Dubois and Bienvenue-Goetz, 1968a; Dubois and Huynh, 1971) or in halogenated media (Bellucci et al., 1985b). There are still doubts about the electrophilicity of this species, since the corresponding kinetic term is probably related mainly to a bromide-assisted free bromine addition. Since, from an energetic point of view, free bromine addition can be considered as an ionization process leading from a neutral reagent to a cation-anion pair in a single elementary step, most work on this electrophilic addition can be used to understand how charge separation and stabilization occur in organic reactions. Bromination thus appears as a suitable model, and is complementary to conventional heterolytic substitutions. The data on this latter, however, are frequently less directly related to the ionization step because of complications arising, in particular, from the unresolved question of return (Cox and Maskill, 1983; Paradisi and Bunnett, 1985). In bromination, most of the usual tools of physical organic chemistry (structure-reactivity relationships involving separation of polar, resonance and steric effects, selectivity relationships, substitutes for rate-equilibrium relationships, solvent-reactivity relationships for distinguishing the medium effect and electrophilic and nucleophilic solvent contributions, etc.) have been widely developed in order to interpret kinetic data in terms of transition-state

286

M.-F. RUASSE

structure. In this context, various long-standing concepts regarding the role of substituents and solvent in promoting charge development have been quantitatively evaluated: (a) the relative contributions of conjugated and non-conjugated substituents, entering bromine, anchimerically assisting groups, etc. to the stabilization of cationic transition states and intermediates; (b) the thermodynamic and kinetic contributions to the energy barrier and the magnitude of transition state shifts, in agreement with the Hammond postulate and the Marcus equation; (c) the all-important contributions of the solvents (i) by electrophilic assistance of bromide ion departure and (ii) by nucleophilic solvation in a preassociation mechanism where the nucleophile (which traps the cationic intermediates in the last, product-forming, step) is already involved in the preceding ionization step (Jencks, 1985). The association of kinetics and stereochemistry is particularly useful for obtaining data on the structure, open or bridged depending on the substituents, of bromination intermediates. These short-lived reactive intermediates cannot be observed under the reaction conditions, but indirect kinetic methods enable us to determine the magnitude of bromine bridging on which the selectivity of product formation depends. However, much work has to be done before these intermediates are known well enough for us to understand, and control if possible, the stereo, regio- and chemo-selectivity of the bromination of any olefin. So far, most of the available data concern the two first ionization steps, but the final, product-forming, step is still inaccessible to the usual kinetic techniques. It would therefore be highly interesting to apply to bromination either the method of fast generation of reactive carbocations by pulse radiolysis (McClelland and Steenken, 1988) or the indirect method of competitive trapping (Jencks, 1980) to obtain data on the reactivity and on the life time of bromocation-bromide ion pairs that control this last step and, finally, the selectivities of the bromination products.

Acknowledgements

This work was funded by the Centre National de la Recherche Scientifique (France) and the University of Paris 7. I warmly thank Professeur J. E. Dubois, who initiated most of this work. I am grateful to Dr J. S. Lomas for fruitful discussion. Many thanks are also due to my colleagues, collaborators and students, whose names are cited in the references. I am

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

287

also grateful to Mme B. Dktry and Mr M. Simon for their help in preparing the manuscript.

References Abraham, M. H. and Liszi, J. (1978). J . Chem. SOC., Faraday Trans. I, 1604 Andersen, L., Berg, U. and Petterson, I. (1985). J . Org. Chem. 50, 493 Andrews, L. J. and Keefer, R. M. (1951). J . Am. Chem. SOC.73,462 Andrews, L. J. and Keefer, R. M. (1964). Molecular Complexes in Organic Chemistry. Holden Day, San Francisco Angelini, G. and Speranza, M. (1981). J. Am. SOC. Chem. 103, 3792 Argile, A. and Ruasse, M. F. (1983). J . Org. Chem. 48, 209 Argile, A., Ruasse, M. F and Dubois, J. E. (1984). J . Am. Chem. SOC. 106,4840 Atkinson, J. R. and Bell, R. P. (1963). J . Chem. SOC., 3260 Bach, R. D. and Henneike, H. F. (1970). J. Am. Chem. SOC.92, 5589 Barbier, G. and Dubois, J. E. (1968). J . Chim. Phys. Phys.-Chim. Biol. 65, 1989 Bartlett, P. D. and Tarbell, D. S. (1936). J . Am. Chem. SOC.58, 466 Bellucci, G., Berti, G., Bianchini, R., Ingrosso, G. and Ambrosetti, R. (1980). J . Am. Chem. SOC. 102, 7480 Bellucci, G., Berti, G., Bianchini, R., Ingrosso, G. and Yates, K. (1981). J . Org. Chem. 46, 23 15 Bellucci, G., Bianchini, R. and Ambrosetti, R. (1985a). J . Am. Chem. SOC. 107, 2464 Bellucci, G., Bianchini, R., Ambrosetti, R. and Ingrosso, G. (1985b). J . Org. Chem. 50, 3313 Bellucci, G., Chiappe, C. and Marioni, F. (1987). J . Am. Chem. SOC.109, 515 Bellucci, G., Bianchini, R., Chiappe, C., Marioni, F. and Spagna, R. (1988). J . Am. Chem. SOC. 110, 546 Bellucci, G., Bianchini, R., Chiappe, C., Marioni, F., Ambrosetti, R., Brown, R. S. and Slebocka-Tilk, H. (1989). J . Am. Chem. SOC. 111, 2640 Bellucci, G., Bianchini, R., Chiappe, C. and Marioni, F. ( 1990).J . Org. Chem. 55,4094 Bellucci, G., Chiappe, C., Marioni, F. and Marchetti, F. (1991). J . Phys. Org. Chem. 4, 387 Bentley, T. W. and Carter, G. E. (1982). J . Am. Chem. SOC. 104, 5741 Bentley, T. W. and Llewellyn, G. (1990). Prog. Phys. Org. Chem. 17, 121 Bentley, T. W. and Schleyer, P.v.R. (1976). J . Am. Chem. SOC.98, 7658 Bentley, T W., Bowen, C. T., Morten, D. H. and Schleyer, P.v.R. (1981).J . Am. Chem. SOC. 103, 5466 Bentley, T. W., Carter, G. E. and Roberts, K. (1984). J . Org. Chem. 49, 5183 Bergmann, H. J., Collin, G., Just, G., Miiller-Hagen, G. and Pritzkow, W. (1972). J . Prakt. Chem. 314, 285 Bienvenue-Goetz, E. and Dubois, J. E. (1968). Tetrahedron 24, 6777 Bienvenue-Goetz, E. and Dubois, J. E. (1975). J . Org. Chem. 40, 221 Bienvenue-Goetz, E. and Dubois, J. E. (1978). Tetrahedron 34, 2Q21 Bienvenue-Goetz, E. and Dubois, J. E. (1979). J. Chem. Res. ( S ) 192 Bienvenue-Goetz, E. and Dubois, J. E. (1981). J . Am. Ckem. SOC.103, 5388 Bienvenue-Goetz, E., Dubois, J. E., Pearson, D. W. and Williams, D. L. H. (1970). J . Chem. SOC. ( B ) , 1275

288

M:F.

RUASSE

Bienvenue-Goetz, E., Msika, R. and Dubois, J. E. (1980). J . Chim. Phys. Phys.-Chim. Biol. 77, 803 Bodrikov, I. V., Bronnikova, N. G. and Okrokova, I. S. ( 1973).Zh. Org. Khim. 9,953 Brown, R. S., Gedye, R., Slebocka-Tilk,H., Buschek, J. M. and Kopecki, K. R. (1984). J . Am. Chem. SOC.106, 4515 Brown, R. S., Slebocka-Tilk, H., Bunnett, A. J., Bellucci, G., Bianchini, R. and Ambrosetti, R. (1990). J . Am. Chem. SOC. 112, 6310 Buckles, R. E. and Yuk, J. P. (1953). J . Am. Chem. SOC.75, 5048 Buckles, R. E., Bader, J. M. and Thurmaier, R. J. (1962). J . Org. Chem. 27, 4523 Calvet, A., Josefowicz, M. and Levisalles, J. (1983). Tetrahedron 39, 103 Castaldi, G., Cavicchioli, S., Giordano, C. and Uggeri, F. (1986). Angew. Chem. Int. Ed. Engl. 25, 259 Chang, W. K. and Tidwell, T. T. (1978). J . Org. Chem. 43, 1904 Charton, M. and Charton, B. I. (1973). J . Org. Chem. 38, 1631 Cioslowski,J., Hamilton, T., Scuseria, G., Hess, B. A., Hu, J., Schaad, L. J. and Dupuis, M. (1990). J . Am. Chem. SOC. 112,4183 Collin, G., Jahnke, U., Just, G., Lorenz, G., Pritzkow, W., Rollig, M., Winguth, L., Dietrich, P., Doring, C.-E., Hauthal, H. G. and Wiedenhoft, A. (1969). J . Prakt. Chem. 311, 238 Cox, B. G. and Maskill, H. (1983). J . Chem. SOC.,Perkin Trans. 2, 1901 De la Mare, P. B. D. ( 1976).Electrophilic Halogenation. Cambridge University Press De la Mare, P. B. D. and Bolton, R. (1982). Electrophilic Additions to Unsaturated Systems, 2nd Edn. Elsevier, Oxford Dewar, M. J. S. and Leplay, A. R. (1961). J . Am. Chem. SOC. 83,4560 De Young, S. and Berliner, E. (1977). J . Am. Chem. SOC.99, 290 Dubois, J. E. and Bienvenue-Goetz, E. (1968a). Bull. SOC.Chim. Fr., 2089 Dubois, J. E. and Bienvenue-Goetz, E. (1968b). Bull. SOC. Chim. Fr., 2094 Dubois, J. E. and Chretien, J. R. (1978). J . Am. Chem. SOC. 100, 3506 Dubois, J. E. and Fresnet, P. (1973). Tetrahedron 29, 3407 Dubois, J. E. and Fresnet, P. (1974). Tetrahedron Lett., 2195 Dubois, J. E. and Garnier, F. (1967a). Bull. SOC.Chim. Fr., 4512 Dubois, J. E. and Garnier, F. (1967b). Spectr. Acta 23A, 2279 Dubois, J. E. and Garnier, F. (1968). J . Chem. Soc., Chem. Commun., 241 Dubois, J. E. and Herzog, H. (1963). Bull. SOC.Chim. Fr., 57 Dubois, J. E. and Huynh, X. Q. (1968). Bull. SOC.Chim. Fr., 1436 Dubois, J. E. and Huynh, X. Q. (1971). Tetrahedron Lett., 3369 Dubois, J. E. and LoYzos, M. (1972). C.R. SLances Acad. Sci. SLr. C 274, 1130 Dubois, J. E. and Mouvier, G. (1968). Bull. SOC. Chim. Fr., 1426 Dubois, J. E. and Schwarcz, A. (1964). Tetrahedron Lett., 2167 Dubois, J. E., Alcais, P. and Barbier, G. (1964). J . Electroanal. Chem. 8, 359 Dubois, J. E., Ropars, M. and Fresnet, P. (1965). J . Chim. Phys. 62, 856 Dubois, J. E., Uzan, R. and Alcais, P. (1968). Bull. SOC.Chim. Fr., 617 Dubois, J. E., Hegarty, A. F. and Bergmann, E. D. (1972a). J . Org. Chem. 37, 2218 Dubois, J. E., Aaron, J. J., Alcais, P., Doucet, J. P., Rothenberg, F. and Uzan, R. (1972b). J . Am. Chem. SOC.94, 6823 Dubois, J. E., Kornprobst, J. M. and Meyer, J. J. (1973a). J . Chim. Phys. 70, 1452 Dubois, J. E., Toullec, J. and Fain, D. (1973b). Tetrahedron Lett., 4859 Dubois, J. E., Al-Alaoui, M. and Toullec, J. (1981a). J . Am. Chem. SOC.103, 5393 Dubois, J. E., Lion, C. and Panaye, A. (1981b). Noun J . Chim. 5, 371 Dubois. J. E., Ruasse, M. F. and Poupard, D. (1983). J . Electroanal. Chem. 152, 85

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

289

Durand, J. P., Davidson, M., Hellin, M. and Coussemant, F. (1966). Bull. SOC. Chim. Fr., 43 Fahey, R. C. (1966). J. Am. Chem. SOC.88,4681 Fahey, R. C. and Schneider, H. J. (1968). J . Am. Chem. SOC.90,4429 Fainberg, A. H. and Winstein, S. (1956). J . Am. Chem. SOC. 78, 2770 Freeman, F. (1975). Chem. Rev. 75,439 Fukuzumi, S. and Kochi, J. K. (1981). J . Am. Chem. SOC.103, 2783 Galland, B., Evleth, E. M. and Ruasse, M. F. ( 1990).J . Chem. SOC., Chem. Commun., 898 Gallo, R. (1983). Frog. Phys. Org. Chem. 14, 115 Gamier, F. and Dubois, J. E. (1968). Bull. SOC. Chim. Fr. 3797 Gamier, F., Donnay, R. H. andDubois, J. E.( 1971). J . Chem. SOC.,Chem. Commun., 829 Gebelein, C . G. and Frederick, G. D. (1972). J . Org. Chem. 37, 221 1 Grosjean, D., Mouvier, G. and Dubois, J. E. (1973). Bull. SOC. Chim. Fr., 1735 Grosjean, D., Mouvier, G. and Dubois, J. E. (1976). J . Org. Chem. 41, 3869, 3872 Hamilton, T. P. and Schaefer, H. F. (1990). J . Am. Chem. SOC. 112, 8360 Hamilton, T. P. and Schaefer, H. F. (1991). J . Am. Chem. SOC. 113, 7147 Hancock, C. K., Meyers, E. A. and Yager, B. J. (1961). J . Am. Chem. SOC.83,421 1 Hegarty, A. F., Lomas, J. S., Wright, W. V., Bergmann, E. D. and Dubois, J. E. (1972). J . Org. Chem. 37, 2222 Heublein, G. (1966). J . Prakt. Chem. 303, 84 Ho. T. L. (1977). Hard and Soft Acids and Bases Principle in Organic Chemistry, p. 61. Academic Press, New York Jencks, W. P. (1980). Acc. Chem. Res. 13, 161 Jencks, W. P. (1985). Chem. Reu. 85, 511 Kamlet, M. J.,Abboud, J. L. H. andTaft, R. W.( 1981). Frog. Phys. Org. Chem. 13,485 Kharasch, N. and Orr, W. L. (1956). J . Am. Chem. SOC.78, 1201 Knittel, P. and Tidwell, T. T. (1977). J . Am. Chem. SOC.99, 3408 Kochi, J. K. (1988). Angew. Chem. Int. Ed. Engl. 27, 1227 Koshy, K. M., Roy, D. and Tidwell, T. T. (1979). J . Am. Chem. SOC. 101, 357 Kresge, A. J. and Chen, H. H. (1972). J . Am. Chem. SOC.94, 2818 Langler, R. F. and Tidwell, T. T. (1975). Tetrahedron Lett. 777 Larsen, J. W. and Metzner, A. V. (1972). J . Am. Chem. SOC. 94, 1614 Laurence, C., Guiheneuf, G. and Wojtkowiak, B. ( 1979). J . Am. Chem. SOC.101,4793 Lee, I. (1992). Adv. Phys. Org. Chem. 27, 57 Lenoir, D. (1978). Chem. Ber. 111,411 Loudon, G. M. and Berke, C. (1974). J . Am. Chem. SOC.96,4508 Mayr, H., Will, E., Heigh, V. W. and Shade, C. (1986). Tetrahedron 42,2519 McClelland, R. A. and Steenken, S. (1988). J . Am. Chem. SOC. 110, 5860 Meijer, E. W., Kellog, R. M. and Wynberg, H. (1982). J . Org. Chem. 47, 2005 Mindl, J. and Vecera, M. (1972). Collect. Czech. Chem. Comm. 37, 1143 Modro, A., Schmid, G. H. and Yates, K. (1977). J . Org. Chem. 42, 3673 Modro, A., Schmid, G. H. and Yates, K. (1979). J . Org. Chem. 44,4221 Mouvier, G. and Dubois, J. E. (1968). Bull. SOC. Chim. Fr., 1441 Mouvier, G., Grosjean, D. and Dubois, J. E. (1973). Bull. SOC. Chim. Fr., 1729 Mulliken, R. S. (1952). J . Am. Chem. SOC. 74, 811 Nozaki, K. and Ogg, R. A. (1942). J . Am. Chem. SOC. 64, 697 and 704 Nishida, S. (1967). J . Org. Chem. 32, 2692 Noyce, D. S., Hartter, D. R. and Miles, F. B. (1968). J . Am. Chem. SOC.90, 3794 Okabe, M., Abe, M. and Tada, M. (1982). J . Org. Chem. 47, 1775 Okamoto, Y., Inukai, T. and Brown, H. C. (1958). J . Am. Chem. SOC. 80,4972

290

M.-F. RUASSE

Olah, G. A. (1975). Halonium Ions. Wiley, New York Olah, G. A. and Hockswender, Jr, T. R. (1974). J. Am. Chem. SOC. 96, 3574 Olah, G. A. and White, A. M. (1969). J. Am. Chem. SOC.91, 5801 Olah, G. A., Bollinger, J. M. and Brinich, J. M. (1968). J. Am. Chem. SOC.90, 2587 Olah, G. A., Porter, R. D., Jeuell, C. L. and White, A. M. (1972). J. Am. Chem. SOC. 94, 2044 Olah, G. A., Schilling, P., Westerman, P. W. and Lin, H. C. (1974a). J . Am. Chem. SOC. 96, 3581 Olah, G. A., Westerman, P. W., Melby, E. G. and Mo, Y. K. (1974b). J. Am. Chem. SOC. 96, 3565 Oyama, K. and Tidwell, T. T. (1976). J . Am. Chem. SOC.98,947 Pannel, K. H. and Mayr, A. J. (1982). J. Chem. SOC., Perkin Trans. I , 2153 Paradisi, C. and Bunnett, J. F. (1985). J. Am. Chem. SOC. 107, 8223 Parker, A. J. and Alexander, R. (1968). J. Am. Chem. SOC. 90, 3313 Pincock, J. A. and Yates, K. (1970). Can. J . Chem. 48, 2944 Poirier, R. A,, Mezey, P. G., Yates, K. and Csizmadia, I. G. (1981). J. Mol. Struct. Theochem. 85, 153 Poirier, R. A., Demare, G. R., Yates, K. and Csizmadia, I. G. (1983). J. Mol. Struct. Theochem. 94, 137 Poupard, D., Ruasse, M. F. and Dubois, J. E. (1983). J. Electroanal. Chem. 152, 77 Prisette, J., Seger, G. and Kochanski, E. (1978). J . Am. Chem. SOC. 100, 6941 Raber, D. J., Bingham, R. C., Harris, J. M., Fry, J. L. and Schleyer, P. v. R. (1970). J. Am. Chem. SOC.92, 5977 Raber, D. J., Harris, J. M. and Schleyer, P. v. R. (1974). Ions and Ion Pairs in Organic Chemistry, Vol. 2 (ed. M. Szwarc), p. 247. Wiley, New York Roberts, I. and Kimball, G. E. (1937). J . Am. Chem. SOC.59, 947 Rodriguez, J., Dulctre, J. P. and Bertrand, M. (1984). Tetrahedron Lett., 527 Rolston, J. H. and Yates, K. (1969a). J . Am. Chem. SOC.91, 1483 Rolston, J. H. and Yates, K. (1969b). J . Am. Chem. SOC.91, 1477 Rolston, J. H. and Yates, K. (1969~).J. Am. Chem. SOC.91, 1469 Rothbaum, H. P., Ting, I. and Robertson, P. W. (1948). J . Chem. SOC., 980 Ruasse, M. F. (1985). Isr. J . Chem. 26, 414 Ruasse, M. F. (1988). Actualit6 Chimique, 215 Ruasse, M. F. (1990). Acc. Chem. Res. 23, 87 Ruasse, M. F. and Argile, A. (1983). J. Org. Chem. 48, 202 Ruasse, M. F. and Chretien, J. R. (1993). In preparation Ruasse, M. F. and Dubois, J. E. (1972). J. Org. Chem. 37, 1770 Ruasse, M. F. and Dubois, J. E. (1974). J . Org. Chem. 39, 2441 Ruasse, M. F. and Dubois, J. E. (1975). J. Am. Chem. SOC.97, 1977 Ruasse, M. F. and Dubois, J. E. (1984). J . Am. Chem. SOC. 106, 3230 Ruasse, M. F. and Lefebvre, E. (1984). J. Org. Chem. 49, 3210 Ruasse, M. F. and Motallebi, S. (1988). Bull. SOC.Chim. Fr., 349 Ruasse, M. F. and Motallebi, S. (1991). J. Phys. Org. Chem. 4, 527 Ruasse, M. F. and Zhang, B. L. (1984). J. Org. Chem. 49, 3207 Ruasse, M. F., Argile, A. and Dubois, J. E. (1978). J . Am. Chem. SOC. 100, 7645 Ruasse, M. F., Argile, A., Bienvenue-Goetz, E. and Dubois, J. E. (1979). J . Org. Chem. 44, 2758 Ruasse, M. F., Argile, A. and Dubois, J. E. (1984). J. Am. Chem. SOC. 106,4846 Ruasse, M. F.,Aubard, J., Galland, B. and Adenier, A. (1986).J. Phys. Chem. 90,4382 Ruasse, M. F., Motallebi, S., Galland, B. and Lomas, J. S. (1990).J . Org. Chem. 55,2298

ELECTROPHILIC BROMINATION OF C=C

DOUBLE BONDS

29 1

Ruasse, M. F., Motallebi, S. and Galland, B. (1991). J . Am. Chem. SOC.113, 3440 Schadt, F. L., Bentley, T. W. and Schleyer, P. v. R. ( 1976).J . Am. Chem. SOC.98,7667 Schmid, G. H. (1989). The Chemistry of Double-Bonded Functional Groups, Suppl. A, Vol. 2, Part 1 (ed. S. Patai'), p. 699. Wiley, New York Schmid, G. H. and Garratt, D. G. ( 1977). Chemistry of Alkenes, Vol. 3 (ed. J. Zabicky). Wiley-Interscience, New York Schmid, G. H. and Tidwell, T. T. (1978). J . Org. Chem. 43, 460 Schmid, G. H., Csizmadia, V. M., Nowland, V. J. and Garratt, D. G. (1972). Can. J . Chem. 50, 2457 Schmid, G H., Modro, A. and Yates, K. (1977a). J . Org. Chem. 42, 871 Schmid, G. H., Modro, A. and Yates, K. (1977b). J . Org. Chem. 42, 2021 Sen Sharma, D. K., Meza de Hojer, S. and Kebarle, P. (1985). J . Am. Chem. SOC. 107, 3757 Sergeev, G. B., Serguchev, Yu. A. and Smirnov, V. V. (1973). Russ. Chem. Rev. (Engl. Transl.) 42, 697 Servis, K. L. and Domenick, R. L. (1987). J . Am. Chem. SOC. 107, 7186 Shorter, J. (1972). Advances in Linear Free Energy Relationships (ed. N. B. Chapman and J. Shorter), Chap. 2. Plenum Press, London Shorter, J. ( 1982). Correlation Analysis of Organic Reactivity. Research Studies Press Wiley, New York Slebocka-Tilk, H., Ball, R. G. and Brown, R. S. (1985).J . Am. SOC.Chem. 107,4504 Staley, R. H., Wieting, R.D. and Beauchamp, J. L. ( 1977). J . Am. Chem. SOC.99,5964 Stock, L. M. and Brown, H. C. (1963). Adv. Phys. Org. Chem. 1, 35 Strating, J., Wieringa, J. H. and Wynberg, H. (1969).J.Chem. SOC.,Chem. Commun.,907 Toullec, J. (1979). Tetrahedron Lett., 3089 Ueno, Y., Chino, K., Watanabe, M., Moriya, D. and Okawara, M. (1982).J . Am. Chem. SOC.104, 5564 V'yunov, K. A. and Ginak, A. I. (1981). Russ. Chem. Rev. (Engl. Transl.) 50, 151 Williams, D. L. H., Bienvenue-Goetz, E. and Dubois, J. E. (1969). J . Chem. SOC. ( B ) , 517 Yamabe, S., Minato, T. and Inagaki, S. (1988). J . Chem. SOC.,Chem. Commun., 532 Yates, K. and Leung, H. W. (1980). J . Org. Chem. 45, 1401 Yates, K. and McDonald, R. S. (1973). J . Org. Chem. 38, 2465 Yates, K., McDonald, R. S. and Shapiro, S. A. (1973). J . Org. Chem. 38, 2460 Young, P. R. and Jencks, W. P. (1979). J . Am. Chem. SOC. 101, 3288 Yukawa, Y., Tsuno, Y. and Masawa, M. (1966). Bull. Chem. SOC. Jpn. 39, 2274 Zhang, B. L. (1981). Doctoral thesis, Universite Paris 7

This Page Intentionally Left Blank

Author Index Numbers in italic refer to the pages on which references are listed at the end of each article

Aaron, J. J., 257, 288 Abboud, J. L. K., 270, 278,289 Abe, M., 208,289 Abraham, M. H., 279, 287 Adamczak, O., 1, 43 Adams, C . M., 112, 115, 136 Adenier, A,, 212, 277, 278, 290 Al-Alaoui, M., 265, 288 Albery, W. J., 147, 148, 149, 152, 156, 158, 159, 163, 164, 165, 170 Alcais, P., 216, 257, 288 Alexander, J., 10, 14, 17, 27, 34, 35, 39, 40, 43 Alexander, R., 277, 290 Allred, A. L., 152, 170 Ambrosetti, R., 210, 212, 214, 219, 220, 250, 276, 278, 280, 282, 283, 285, 287, 288 Anderson, L., 247, 285 Andrews, L. J., 217, 287 Anet, F. L., 5, 24, 40 Angelini, G., 278, 287 Anson, F. C., 2, 41 Anton, U., 13, 15, 40 Anvia, F., 179, 182, 183, 204 Argile, A,, 209, 213, 220, 229, 230, 231, 232, 233, 237, 238, 239, 240, 242, 245, 248, 252, 253, 254, 255, 257, 258, 260, 263, 268, 269, 287, 290 Arnett, E. M., 46, 49, 55, 56, 62, 71, 72, 78, 82, 84, 86, 87, 88, 89, 90, 91, 93, 94, 96, 102, 107, 109, 110, 111, 112, 113, 114, 115, 116, 118, 121, 122, 123, 125, 126, 130, 135, 136, 138 Atkins, P. W., 20, 40, 139, 170 Atkinson, J. R., 216, 287 Aubard, J., 212, 277, 278,290 Auchter-Krummel, P., 8, 12, 22, 24, 40

Bach, R. D., 224, 245,287 Bader, J. M., 238, 288 Bain, C . D., 48, 135 Balakrishnan, M., 171, 204 Baldwin, J. E., 190, 204 Ball, R. G., 223, 225, 291 Barbier, G., 216, 245, 288, 289 Bard, A. J., 2, 41 Barrett, D., 4, 42 Bartlett, P. D., 208, 211, 213, 285, 287 Baughman, R. H., 4,40,44 Baumgarten, M., 8, 13, 15, 16, 38, 40 Beauchamp, J. L., 221, 222, 278, 291 Becker, B., 12, 13, 21, 26, 30, 32, 36, 40 Becker, H. G. O., 188, 204 Bell, R. P., 172, 177, 190, 204, 216, 287 Bellamy, W. D., 68, 136 Bellucci, G., 210, 21 1, 212, 214, 219, 220, 236, 250, 276, 277, 278, 279, 280, 281, 282, 283, 285. 287, 288 Bender, D., 8, 9, 40, 41 Bender, M. L., 172, 174, 176, 180, 192, 200, 202,204 Bensasson, R. S., 4, 42 Bentley, T. W., 245, 270, 271, 272, 279, 287, 291 Beratan, D. A., 19, 43 Berg, U., 247, 287 Bergmann, E. D., 213, 257, 258, 259, 288, 289 Bergmann, H. J., 214, 227, 287 Berke, C., 266, 289 Berliner, E., 21 1, 288 Bernhard, S. A., 191, 202, 206 Bernstein, R. B., 140, 170 Berti, G., 212, 276, 287 Bertrand, M., 208, 290 Berwick, M. A,, 107, 136 293

294

AUTHOR INDEX

Bhatt, M. V., 179, 181, 182, 183, 189, 199, 201, 204 Bianchini, R., 210, 211, 212, 214, 219, 220, 236, 250, 276, 277, 278, 279, 280, 281, 282, 283, 285,287,288 Bienkowski, R., 62, 135 Bienvenue-Goetz, E., 211, 212, 213, 215, 228, 229, 234, 243, 244, 245, 246, 248, 257, 259, 260, 277, 278, 279, 285, 287,289, 290, 291 Bingham, R. C., 270,290 Bitthin, R., 4, 41 Blois, D. W., 62, 137 Bock, H., 1, 12, 15, 18, 37, 41, 42 Bodrikov, I. V., 253, 254, 288 Boekema, E. J., 116, 136 Bohnen, A., 8, 12, 13, 14, 15, 16, 21, 26, 30, 32, 36, 38, 40, 41, 44 Bollinger, J. M., 221, 290 Bolton, J. R., 21, 22, 25, 44 Bolton, R., 209, 214, 288 Bothner-By, A. A., 102, 138 Bourn, A. J. R., 5, 24, 40 Bover, W. K., 172, 204 Bowden, C. T., 245, 271, 272, 287 Bowden, K., 172, 176, 177, 179, 180, 181, 182, 183, 185, 187, 195, 197, 199, 202, 204,205 Boyd, E., 54, 136 Bredas, J. L., 2, 40 Brinich, J. M., 221, 290 Brittain, W. J., 107, 109, 110, 111, 136 Bronikova, N. G., 253,254,288 Broseta, D., 70, 137 Brown, H. C., 252, 255, 289,291 Brown, R. S., 210, 219, 220, 223, 225, 250, 280, 282, 283,285,287,288,291 Bruice, T. C., 171, 202, 205 Buck, E. J., 62, 137 Buckles, R. E., 220, 238, 288 Bunnett, A. J., 210, 219, 220, 250, 280, 282, 283, 285, 288 Bunnett, J. F., 285, 290 Burrows, H. D., 192, 197, 205 Buschek, J. M., 210, 280, 288

Calvet, A., 279, 288 Cannon, R. D., 2, 17,41 Cante, C. C., 62, 137 Capon, B., 171, 199, 203, 205 Capps, R. W., 68, 137 Carter, G. E., 270, 271, 287, 289 Castaldi, G., 208, 288 Cavicchioli, S., 208, 288 Chakravarty, J., 187, 201, 205 Chance, R. R., 2, 4 Chang, W. K., 231, 248,288 Chanon, M., 17, 21,41 Chao, J., 71, 72, 121, 135 Chapman, D., 48, 136 Chapman, N. B., 17, 184,205 Charton, B. I., 247, 288 Charton, M., 247, 288 Chehel-Amiran, M., 197, 204 Chen, H. H., 264, 289 Chen, Y.-L., 57, 137 Cheng, X.-E., 189, 205 Chiappe, C., 210, 211, 219, 220, 236, 281, 285,287 Child, M. S., 146, I70 Chino, K., 208, 291 Choudhery, R. A,, 156, 164, 170 Chretien, J. R., 235, 236, 242, 288 Cioslowski, J., 225, 288 Clint, G. E., 158, 170 Clint, J. H., 158, 170 Closs, G. L., 4, 20, 21, 28, 29, 41, 43 Collet, A., 82, 136 Collin, G., 214, 227, 245, 287, 288 Connor, H. D., 27,44 Cook, A., 62, 136 Cook, E. R., 190, 206 Corkhill, J. M., 158, 170 Coussemant, F., 266, 289 Cox, B. G., 285,288 Craig, D. P., 102, 121, 136 Cram, D. J., 113, 136 Crentz, C., 4, 42 Crisp, D. J., 65, 67, 97, 119, 123, 136 Csizmadia, I. G., 224, 290 Csizmadia, V. M., 212, 276, 291

Cadenhead, D. A., 54, 135 Calcaterra, L. T., 20, 29, 41, 43 Caluwe, P., 27, 41, 44

Dabros, T. G., 158, 170 Dahlgren, G., 176, 192, 200, 205 Dale, J., 29, 41

AUTHOR INDEX

David, H. H., 116, 136 Davidson, M., 266, 289 Davies, E. R., 34, 41 De le Mare, P. B. D., 208, 209, 212, 214,288 De Schryver, F. C., 4, 42 De Young, S., 211, 288 Dean, J., 116, 137 Defay, R., 65, 67, 97, 119, 123, 136 Demare, G. R., 224, 290 Desai, N. B., 200, 205 Dewar, M. J. S., 216, 288 Dietrich, P., 245, 288 Dixit, V. M., 33, 43 Djerassi, C., 172, 205 Domenick, R. L., 221, 291 Donnay, R. H., 267, 289 Doring, C.-E., 245, 288 Doucet, J. P., 257, 288 Dragcevic, D., 61, 136 Draus, F., 198, 202, 205 Dreiding, A. S., 24, 32, 41 Dubois, J. F., 209, 210, 211, 212, 213, 214, 215, 216, 217, 218, 220, 225, 227, 228, 229, 230, 231, 232, 233, 234, 236, 237, 238, 239, 243, 244, 245, 246, 247, 248, 252, 253, 254, 255, 256, 257, 258, 259, 260, 263, 265, 266, 267, 269, 270, 277, 278, 279, 285, 287, 288,289,290,291 Dugas, M., 172, 201, 202, 205 Dulckre, J. P., 208, 290 DuNouy, L., 53, 136 Dupuis, M., 225, 288 Durand, J. P., 266, 289 Dustman, C. K., 12, 24, 44 Edlund, U., 1, 11, 41 Effenberger, F., 40, 42 Ehrenfreund, M., 10, 12, 13, 14, 17, 21, 26, 27, 30, 32, 36, 39, 40 El Kaissi, F. A., 172, 181, 197, 204, 205 Eliasson, B., 1, 11, 24, 41 Eliel, E. L., 82, 136 Elsenbaumer, R. L., 2, 40 Ely, G., 4,41 Endres, H., 40, 41 Enkelmann, V., 10, 40, 41, 43

295

Evleth, E. M., 221, 222, 224, 226, 280, 289 Eyring, H., 18, 41 Facchine, K. L., 12, 24, 44 Fahey, R. C., 235, 238,289 Fahnenstich, U., 8, 41 Fainberg, A. H., 270, 289 Fendler, J. H., 47, I36 Fenske, D., 12, 15,41 Fersht, A. R., 171, 174, 194, 205 Fiedler, J., 8, 9, 10, 14, 17, 25, 26, 27, 29, 30, 39, 40, 41 Fisher, L. M., 165, 170 Fisk, P. R., 156, 170 Flanagan, J. B., 2, 41 Flitsch, W., 173, 203, 206 Foulds, A. W., 158, 170 Fox, M. A., 17, 21,41 Fraenkel, G. K., 21, 22, 23, 41 Frederick, G. D., 217, 289 Fredga, A., 103, 136 Fredlein, R. A., 158, 159, 170 Freed, J. H., 21, 22, 23, 41 Freeman, F., 209, 289 Fresnet, P., 211, 215, 234, 288 Fritsch, D., 116, 136 Fry, A. J., 14, 41 Fry, J. L., 270, 290 Fukuzumi, S., 21,41, 218, 289 Furhop, J. H., 116, 136 Gaines, G. L., 48, 53, 54, 65, 68, 136 Galland, B., 210, 221, 222, 224, 226, 240, 242, 249, 250, 268, 271, 272, 277, 278, 280, 282, 285, 289,291 Gallo, R., 247, 289 Gamier, F., 210, 211, 213, 215, 217, 227, 267, 270, 289 Garratt, D. G., 208, 212, 234, 255, 276, 291 Garrett, P. R., 60, 137 Gaule, H. J., 70, 76, 89, 137 Gawron, O., 198, 202,-205 Gebelein, C. G., 217, 289 Gedye, R., 210, 280, 288 Gershfeld, N. L., 52, 53, 54, 64, 65, 66, 67, 68, 136

296

Gershom, H. R., 182, 186, 205 Gerson, F., 11, 13, 21, 24, 27, 30, 32, 33, 41, 43 Ghatak, U. R., 187, 205 Gibbs, J. W., 65, 136 Giles, D., 60, 137 Gilkerson, W. R., 102, 138 Ginak, A. I., 277, 291 Giordano, C., 208, 288 Gleiter, R., 24, 32, 41 Gokelmann, K., 10, 40, 41, 43 Gold, J. M., 62, 135 Gold, L. C., 57, 136 Gold, V., 141, 152, 170 Goodrich, F. C., 65, 67, 68, 136 Gore, P. H., 186, 205 Could, S., 4, 44 Grampp, G., 20, 33, 42 Gray, H. B., 4, 42 Green, N. J., 20, 41 Greene, F. D., 107, 136 Greenzaid, P., 172, 190, 198, 205 Grosjean, D., 214, 215, 227, 243, 246, 247,289 Grupp, A., 34, 35, 41, 43 Gu, J.-H., 189, 205 Guay, D., 68, 136 Guiheneuf, G., 220, 289 Gurevich, Y. Y., 155, 160, 170 Gust, D., 4, 42 Guthrie, J: P., 173, 205 Gutowsky, H. S., 2 1 , 4 2 Hamacher, V., 30, 32, 42 Hamilton, T. P., 224, 225, 288, 289 Hammerich, O., 13, 42 Hancock, C . K., 246, 289 Handa, T., 118, 136 Hansch, C., 172, 205 Hansen, B., 200, 205 Hanthal, H. G., 245, 288 Harding, L. O., 4, 42 Harkins, W. D., 54, 57, 136 Harrer, W., 20, 33, 42 Harriman, J. E., 27, 42 Harris, J. M., 270, 279, 290 Hartler, D. R., 255, 256, 289

AUTHOR INDEX

Harvey, N. G., 46, 49, 55, 56, 58, 62, 78, 82, 84, 86, 87, 88, 89, 90, 91, 96, 107, 113, 114, 115, 120, 122, 123, 125, 126, 130, 135, 136 Havinga, E., 78, 133, 138 Hay, P. F., 21, 42 Heath, J. G., 98, 136 Hegarty, A. F., 213, 257, 258, 259, 288, 289 Heigh, V. W., 250, 284, 289 Heilbronner, E., 1, 18, 37, 42 Heine, B., 40, 42 Heinneike, H. F., 224, 245, 287 Heinz, W., 1, 43 Heinze, J., 1 , 2, 3, 5, 11, 12, 13, 42, 43 Heitz, W., 15, 41 Hellin, M., 266, 289 Henderson, G. H., 176, 205 Henry, M. P., 179, 182, 184, 204 Herbst, H., 8, 9, 12, 13, 40, 43 Herold, B. J., 33, 42, 44 Herzog, H., 215, 288 Hess, B. A., 225, 288 Hetz, G., 20, 33, 42 Heublein, G., 238, 239, 289 Hibbert, F., 152, 170 Hine, J., 200, 205 Hishida, S., 172, 180, 205 Ho, T. L., 236, 289 Hockswender, Jr. T. R., 218, 290 Hoffmann, R., 18, 21,42 Hogen-Esch, T. E., 1, 21, 32, 42 Hogeveen, H., 183, 205 Hollaway, F., 60, 137 Holleck, L., 179, 191, 205 Holm, C . H., 21,42 Hopfield, J. J., 19, 43 Hori, K., 57, 137 Hu, J., 225, 288 Huber, M., 31, 42 Huber, W., 1, 8, 9, 10, 11, 12, 13, 14, 17, 21, 22, 24, 25, 26, 27, 29, 30, 32, 34, 35, 36, 39, 40, 41, 42, 43 Huff, J. B., 58, 112, 113, 114, 115, 120, 125, 136 Hui, Y.-Z., 189, 205 Hush, N. S., 4, 20, 21, 22, 42, 44 Huynh, X . Q., 213,285,288

AUTHOR INDEX

Inagaki, S., 278, 291 Ingrosso, G., 211, 212, 276, 277, 279, 285, 287 Inukai, T., 255, 289 Irmen, W., 10, 12, 14, 17, 27, 41 Ishhizu, K., 32, 43 Isied, S. S., 4, 42 Iskander, Y., 177, 184,205 Jacques, J., 82, 136 Jaenicke, W., 32, 42 Jaffe?, A. B., 111, 136 Jahnke, U., 245, 288 Jarvis, N. L., 59, 66, 136 Jeffers, P. M., 116, 137 Jencks, W. P., 209, 252, 261, 286, 289 Jeuell, C. C., 221, 290 Jiang, X.-K., 189, 205 Joachim, C., 21,40, 42 Johnels, D., 1, 11, 41 Johnson, E. A., 62, 107, 109, 110, 111, 135, 136 Johnson, M. D., 4, 20, 21, 28, 41 Johnson, R. S., 194,205,206 Joly, M., 59, 60, 137 Jordan, K. D., 21, 28,43 Josefowicz, M., 279, 288 Jow, T. R., 4, 44 Just, G., 214, 227, 245, 287, 288 Kamlet, M. J., 270, 278, 289 Kandaniem, A. Y., 116,137 Katz, T. J., 5, 24, 42 Kebarle, P., 278, 291 Keefer, R. M., 217, 287 Keller, D. K., 70, 76, 89, 137 Keller, H. J., 40, 41 Kellog, R. M., 251, 284, 289 Kemp, K. C., 179, 188,205 Kharach, N., 256,289 Kharkats, Y. I., 155, 160, 170 Kim, K., 112, 115,136 Kimball, G. E., 234, 290 Kindermann, I., 1, 43 Kinzing, B. J., 71, 72, 121, 135 Kirby, A. J., 171, 174, 194, 198, 199, 205 Kirkwood, J. G., 57, 136 Kirste, B., 11, 21, 25, 43

297

Kitchener, J. A., 158, 170 Klabunde, K.-U., 10, 11, 40,42, 43 Klarner, F. G., 1, 43 Klotz, I. M., 194, 205, 206 Kneebone, G. R., 158, 159, 170 Knittel, P., 245, 289 Knobler, L. M., 70, 137 Knowles, J. R., 163, 164, 165, 170 KOCh, K.-H., 8, 13, 15, 38, 40, 41, 42 Kochanski, E., 218, 290 Kochi, J. K., 20, 21, 41, 42, 210, 218, 289 Konishi, S., 30, 32, 43 Kopecki, K. R., 210, 280,288 Kornprobst, J. M., 215,253,288 Koshy, K. M., 259, 260,289 Kowert, B., 24, 32, 41 Krebs, P. J., 107, 109, 112, 115, 135, 136 Kreevoy, M. M., 148, 170 Kresge, A. J., 152, 170, 264, 289 Krohnke, C., 40, 43 Krummel, G., 8, 22, 24, 43 Kuder, J. E., 2, 44 Kurreck, H., 11, 21, 25, 31, 43 Kuznetsov, A. M., 21, 26, 43 Laidler, K. J., 143, 170 Lam, R., 2, 43 Langler, R. F., 247, 289 Langmuir, I., 49, 57, 137 Larsen, J. W., 223, 289 Last, A. M., 179, 182, 185, 195, 202, 205 Laurence, C., 220, 289 LeBlanc, R. M., 68, 136 Lee, I., 257, 289 Leegwater, A. L., 180, 205 Lefebvre, E., 253, 254, 263, 266, 269, 274, 275,290 Lenoir, D., 251, 289 Leo, A., 172,205 Leplay, A. R., 216, 288 Levine, R. D., 140, 170 Levisalles, J., 279, 288 Lewis, I. C., 3, 43 Lex, J., 1, 10, 11, 12, 17, 27, 42, 43 Liang, N., 4, 20, 21, 28, 43 Liaske, C. N., 201, 205 Liddell, P. A., 4, 42 Liman, U., 116, 136

298

Lin, Y.S., 5, 24, 40 Lin, H. C., 217, 290 Linkowsky, G. E., 12, 24, 44 Lion, C., 247, 266, 288 Lippman, A. E., 172, 180, 205 Liszi, J., 279, 287 Llewellyn, G., 270, 279, 287 LoYzos, M., 279, 288 Lomas, J. S., 247, 249, 257, 258, 259, 289,290 Long, F. A., 152, 170 Lorenz, G., 245, 288 Loudon, G. M., 266,289 Luanay, J. P.,21, 40, 42 Lubitz, W., 11, 21, 25, 43 Lucassen, J., 60, 137 Lucassen-Reynders, E. H., 60, 68, 137 Lundquist, M., 103, 137 Liittke, W., 8, 41 Luz, Z., 190,205 Maier, S., 40, 42 Maki, A. W., 27, 42 Makings, L. R., 4, 42 Malcolm, B. R., 103, 137 Malik, F. P., 179, 182, 187, 199, 205 Malmstrom, B. G., 4, 42 Maltzan, B. von, 31, 42 Manheimer, R. J., 57, 60, 137 Marchetti, F., 281, 285, 287 Marcus, R. A., 4, 18, 20, 43, 148, 170 Maresch, G. G., 10, 40, 43 Margel, S., 2, 41 Margoni, N., 60, 137 Marioni, F., 210, 211, 219, 220, 236, 281, 285,287 Marshall, J. K., 158, 170 Martin, W. A. Jr., 13, 27, 30, 32, 41 Masawa, M., 252, 291 Maskill, H., 285, 288 Mataga, N., 39, 43, 44 Mathieu, J., 116, 136 Maxfield, M., 4, 44 Mayr, A. J., 239,290 Mayr, H., 250, 284, 289 Mazur, S., 33, 43 McBride, J. M., 111, 136 McClelland, R. A., 286, 289

AUTHOR INDEX

McConnell, H. M., 21, 22, 26, 30, 43, 70, 76, 89, 137 McDiarmid, A. D., 4, 43 McDonald, R. S., 211, 212, 231, 252, 291 McManus, S. P., 171, 199, 205 McPhail, A. T., 102, 112, 115, 136, 138 Meerholtz, K., 1, 3, 5, 15, 43 Mehring, M., 10, 34, 35, 40, 41, 43 Meijer, E. W., 251, 284, 289 Melby, E. G., 221, 290 Melkonian, G. A., 179, 191, 205 Mellor, D. P., 102, 136 Metzner, A. V., 223, 289 Meyer, J. J., 215, 253, 288 Meyer, T. J., 17, 19, 43 Meyers, E. A., 246, 289 Meza de Hojer, S., 278, 291 Mezey, P. G., 224, 290 Mieth, M. L., 179, 188, 205 Miles, F. B., 255, 256, 289 Miller, E. G., 201, 205 Miller, J. R., 4, 20, 21, 28, 29, 41, 43 Milunovic, M., 61, 136 Minato, T., 278, 291 Mindl, J., 262, 289 Mirajovsky, D., 78, 82, 84, 86, 87, 88, 89, 90, 91, 130, 136 Misumi, M., 32, 43 Mo, Y. K., 221,290 Mobius, K., 30, 31, 32, 42 Modro, A., 211, 212,245, 246, 268, 289, 291 Monkenbusch, M., 40, 41 Moore, A. L., 4, 42 Moore, B., 70, 137 Moore, T. A., 4, 42 Moore, W. J., 57, 59, 137 Morawetz, H., 116, 137 Moriya, D., 208, 291 Morten, D. H., 245, 271, 272, 287 Mortenson, J., 12, 13, 43 Moshuk, G., 24, 27, 32, 41, 44 Motallebi, S., 210, 227, 240, 242, 247, 249, 267, 268, 269, 271, 272, 273, 278,280, 282, 284, 285, 291 Mouvier, G., 214, 215, 227, 243, 246, 247,289 Moy, V. T., 70, 76, 89, 137 Msika, R., 212, 215, 277, 278, 288

AUTHOR INDEX

Mullen, K., 1, 2, 4, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 21, 22, 24, 25, 26, 27, 29, 30, 32, 36, 38, 39, 40, 40, 41, 42, 43,44 Muller, B., 40, 41 Muller, U., 8, 13, 15, 16, 38, 40 Muller-Hagen, G., 214, 227, 287 Mulliken, J. W., 62, 137 Munstedt, H., 4, 41 Murray, R. W., 2, 43 Naarmann, H., 4, 13, 15, 40, 41 Nadvi, N. S., 172, 205 Naegele, D., 4, 41 Nagaki, M., 118, 136 Nagaranjan, M. K., 117, 137 Nemeto, F., 30, 32, 43 Niemczyk, M. P., 4, 44 Nishida, S., 257, 262, 289 Novais, H. M., 33, 44 Nowak, R. J., 2, 43 Nowland, V. J., 212, 276, 291 Noyce, D. S., 255, 256, 289 Nozaki, K., 214, 289 O’Shea, G. I., 158, 159, 170 Ogg, R. A., 214,289 Ok, D., 112, 115, 137 Okabe, M., 208, 289 Okada, T., 39, 43,44 Okamoto, Y., 255, 289 Okawara, M., 208, 291 Okrokova, I. S., 253, 254, 288 Olah, G. A., 210, 217, 218, 221, 290 Oliver, A. M., 27, 32, 33, 41 Onuchic, J. N., 19, 43 Orr, W. L., 256, 289 Oyama, K., 264, 290 Paddon-Row, M. N., 21, 27, 28, 32, 33, 41,43 Pagano, R. E., 66, 68, 136, 137 Page, M. I., 172, 205 Panaye, A., 247, 266, 288 Pannel, K. H., 239,290 Paradisi, C., 285, 290 Parker, A. J., 277, 290

299

Pasteur, L., 102, 137 Peake, P. M., 24, 32, 41 Pearson, D. W., 234, 287 Penfield, K. W., 20, 41 Penny, C., 172, 201, 202,205 Pepper, T., 27, 41 Petterson, I., 247, 287 Petterson, K., 82, 137 Pewitt, E. B., 4, 44 Pincock, J. A., 231, 252, 290 Piotrowiak, P., 4, 20, 21, 28, 41 Pirkle, W. H., 102, 137 Plato, M., 30, 31, 32, 42 Pochapsky, T. C., 102, 137 Poirier, D., 224, 290 Port, H., 40, 42 Porter, N. A., 58, 107, 109, 110, 111, 112, 113, 114, 115, 120, 125, 135, 136 Porter, R. D., 221, 290 Poupard, D., 215,216, 288, 290 Powers, T. A., 14, 41 Pravetic, V., 61, I36 Prieve, D. C., 160, 170 Prisette, J., 218, 290 Pritzkow, W., 214, 227, 245,287,288 Przybylski, M., 8, 9, 40, 41 Raber, D. J., 270, 279 Rabinovitz, M., 5, 43 Rader, H.-J., 12, 13, 14, 15, 17, 27, 36, 39, 40, 41 Radom, L., 121,136 Rahim, A., 186,205 Rakshit, A. K., 68, 137 Ramirez, F., 200, 206 Rao, G. V., 171, 179, 181, 182, 183, 189, 199, 204 Rao, K. S., 181, 182, 183, 204 Rao, S. B., 179, 191, 205 Rautter, 34, 35, 43 Raval, D. A., 182, 186, 205 Ravindranathan, M., 179, 182, 189, 199, 204 Reddoch, A. H., 30, 32,43 Reimers, J. R., 4, 21, 22, 44 Reinstein, J. A., 174, 176, 180, 192, 200, 202,204 Rettig, W., 39, 43, 44 Roberts, I., 234, 290

300

Roberts, K., 270, 287 Robertson, P. W., 212, 290 Rodriguez, J., 208, 290 Rollig, M., 245, 288 Rolston, J. H., 213, 231, 238, 239, 252, 290 Rondelez, F., 70, 137 Ropars, M., 215, 288 Rosano, H. L., 62, 137 Rose, P. L., 46, 49, 55, 56, 58, 78, 82, 84, 86, 87, 88, 89, 90, 91, 96, 107, 113, 114, 115, 120, 122, 123, 125, 126, 130, 135, 136 Roth, W. R., 1, 43 Rothbaum, H. P., 212,290 Rothenberg, F., 257, 288 Rougee, M., 4, 42 Roy, D., 259, 260,289 Ruasse, M. F., 209, 210, 211, 212, 213, 215, 216, 218, 220, 221, 222, 224, 225, 226, 221, 229, 230, 231, 232, 235, 236, 237, 238, 239, 240, 242, 245, 247, 248, 249, 250, 252, 253, 254, 255, 256, 257, 260, 261, 263, 264, 265, 261, 268, 269, 270, 271, 272, 273, 274, 275, 277, 278, 280, 282, 284, 285,287,288, 289,290,291 Ruckenstein, E., 160, 170 Ruppert, K., 12, 15, 41 Sakata, Y., 12, 13, 21, 26, 30, 32, 36, 40,43 Salem, L., 1, 44 Samec, Z., 155, 160, 170 Samuel, D., 190,205 Sauer, B., 57, 137 Saveant, J.-H., 13, 42 Scandel, J., 191, 206 Schaad, L. J., 225, 288 Schade, P., 8, 9, 40 Schadt, F. C., 271, 291 Schaefer, H. F., 224, 289 Schaefer, V. J., 49, 57, 137 Schechter, R. S., 57, 60, 137 Schell, D. M., 176, 192, 200,205 Schenk, R., 8, 15, 41, 44 Schilling, P., 217, 290 Schleyer, P. v. R., 245, 270, 271, 272, 279, 287, 291

AUTHOR INDEX

Schlosser, H., 40, 42 Schmid, G. H., 208, 209, 211, 212, 234, 245, 246, 248, 255, 268, 276,289,291 Schneider, H. J., 238, 289 Schneider, J., 188, 204 Schultz, F. A., 2, 43 Schiitz, J. U. von, 10, 40, 43 Schwarcz, A., 231, 252,290 Schwarz, H. A., 4,42 Schweitzer, D., 40, 41 Scuseria, G., 225, 288 Seger, G., 218,290 Sen Sharma, D. K., 278,291 Serafimov, O., 5, 42 Sergeev, G. B., 208, 217, 220, 291 Serguchev, Yu. A., 208, 217, 220, 291 Servis, K. L., 221, 291 Sethson, I., 1, 11, 41 Shacklette, L. W., 2, 4, 40, 44 Shade, C., 250, 284,289 Shafer, P. T., 103, 137 Shah, D. O., 68, 137 Shah, J. P., 117, 137 Shalitin, Y., 191, 202, 206 Shapiro, S. A., 212, 231, 252, 291 Shashidhar, M. S., 182, 204 Shaw, M.J., 112, 177,205 Shimada, K., 27, 31, 44 Shorter, J., 177, 184, 205, 244, 246, 291 Siders, P., 20, 44 Sigmund, E., 40, 42 Silver, M. S., 172, 174, 176, 180, 192, 200, 202,204 Singer, L. S., 3, 43 Skinner, K. J., 11 1, 136 Skolnick, M., 62, 135 Slebocka-Tilk, H., 210, 219, 220, 223, 225, 250, 280, 282, 283, 285, 287, 288,291 Smirnov, V. V., 208, 217, 220 Smith, A. L., 158, 159, 162, 170 Smith, T. W., 2, 44 Sohoji, M. C. B. J., 33, 42, 44 Somayaji, V., 179, 182, 189, 199, 204 Spagna, R., 281, 285,287 Speranza, M., 278, 287 Staley, R. H., 221, 222, 278, 291 Staley, S. W., 12, 24, 41, 44 Stallberg-Stenhagen, S., 103, 137 Steenken, S., 33, 42, 286, 289

AUTHOR INDEX

Steinberg, G. M., 201, 205 Steinleitner, H. D., 188, 204 Stenhagen, E., 103, 137 Steward, M. V., 46, 71, 72, 118, 121, 135,137 Stiles, P. J., 121, 136 Stock, L. M., 252, 291 Stowell, J. C., 107, 136 Strating, J., 210, 218, 223, 249, 280, 282, 291 Sullivan, P. D., 22, 44 Sutin, N., 4, 20, 42, 43 Svec, W. A., 4,44 Swarbrick, J., 62, 137 Szwarc, M., 1, 27, 31, 32, 41, 44 Tachibana, T., 57, 137 Tada, M., 208,289 Taft, R. W., 172, 205, 270, 278, 289 Tamm, C. K., 70, 76, 89, 137 Tanford, C., 47, 137 Tarbell, D. S., 208, 211, 213, 285, 28 7 Taylor, G. R., 176, 179, 180, 181, 182, 183, 205 Tewfik, R., 177, 184, 205 Thibeault, J. C., 21, 42 Thompson, O., 50, 62, 71, 72, 73, 121, 135,137 Thurmair, R. J., 238, 288 Tidwell, T. T., 231, 245, 247, 248, 259, 260, 264,288, 289,290 Ting, I., 212, 290 Topping, R. M., 192, 197, 205 Toullec, J., 173, 206, 265, 266, 288, 291 Townsend, D. F., 62, 137 Toyoda, T., 32, 43 Tsuno, Y., 252, 291 Tweet, A. G., 68, 136 Ueno, Y., 208,291 Uggeri, F., 208, 288 Ulstrup, J., 17, 21, 26, 43, 44 Umana, M., 2, 43 Unterberg, H., 8, 12, 13, 26, 30, 32, 42 Utley, J. H. P., 177, 184, 205 Uzan, R., 216, 257, 288

30 1

V'yunov, K. A., 277, 291 Valters, R. E., 173, 203, 206 Van der Anweraer, M., 4, 42 van Bekkum, H., 183, 206 van de Ven, T. G. M., 158, 170 Varennes, S., 158, 170 Vassilian, A., 4, 42 Vavon, G., 191,206 Vecera, M., 262, 289 Venkatasubramanian, N., 171, 204 Verbiar, R., 71, 72, 78, 82, 84, 86, 87, 88, 89, 90, 103, 121, 130, 135, 136, 138 Verkade, P. E., 183, 206 Wagner, M., 8, 44 Wskeham, H. R. R., 57, 138 Walder, J. A., 194, 205, 206 Walker, T., 158, 170 Walsh, C., 203, 206 Walter, D. N., 186, 205 Ward, R. L., 21, 44 Washburn, E. R., 57, 138 Washburn, W. N., 190, 206 Wasielewski, M. R., 4, 20, 21, 44 Wasif, S., 177, 184, 205 Watanahe, M., 208, 291 Webb, W. R., 62, 136 Wegener, S., 8, 44 Wegner, G., 2, 40, 43, 44 Weiringa, J. H., 210, 218, 223, 249, 280, 282, 291 Weis, R. M., 70, 76, 89, 137, 138 Weissmann, S. I., 21, 44 Wellauer, T., 27, 32, 33, 41 Wepster, B. M., 183, 206 Werner, H.-P., 10, 40, 43 Wertz, J. E., 21, 25, 44 Westerman, P. W., 217, 221, 290 Wette, M., 1, 43 White, A. M., 221, 290 Whitesell, L. G., 62, 135 Whitesides, G. M., 46, 48, 135, 138 Wiedenhoft, A., 245, 288 Wieners, G., 40, 41 Wieting, R. D., 221, 222, 278, 291

302

Wietzel, H.-P., 12, 44 Wilen, S. H., 82, 136 Will, E., 250, 284, 289 Williams, A., 183, 206 Williams, D. L. H., 234, 288, 291 Winguth, L., 245, 288 Winstein, S., 270, 289 Winter, H. J., 116, 136 Wishart, J. F., 4, 42 Wohlfarth, W., 8, 12, 13, 21, 26, 30, 32, 36, 40, 43, 44

Woitellier, S., 21, 40, 42 Wojtkowiak, B., 220, 289 Wolf, H. C., 40, 42 Wolf, J. F., 4, 44 Wolffler, F., 4, 41 Wong, C. L., 21, 41 Wright, W. V., 257, 258, 259, 289 Wychich, D., 2, 44 Wydler, C., 13, 32, 41 Wynberg, H., 210, 218, 223, 249, 251, 280, 282, 284,289,291

AUTHOR INDEX

Yager, B. J., 246, 289 Yamabe, S., 278, 291 Yao, H., 39, 43, 44 Yates, K., 211, 212, 213, 224, 231, 237, 238, 239, 245, 246, 252, 255, 268, 276,287,289,290,291 Yin, D.-N., 62, 137 Young, T. F., 54, 136 Yu, H., 57, 137 Yuk, J. P., 220, 288 Yukawa, Y., 252,291

Zeelen, F. J., 78, 133, 138 Zeger, J. J., 201, 205 Zhang, B. L., 211, 227, 245, 249. 268, 290

Zimmermann, H. E., 5, 42 Zingg, S. P., 102, 138 Zografi, G., 57, 68, 137 Ziilicke, L., 19, 44 Zuman, P., 172, 204

Cumulative Index of Authors Ahlberg, P., 19, 223 Albery, W. J., 16, 87; 28, 139 Allinger, N. I., 13, 1 Anbar, M., 7, 115 Arnett, E. M., 13, 83; 28, 45 Ballester, M., 25, 267 Bard, A. J., 13, 155 Baumgarten, M., 28, 1 Bell, R. P., 4, 1 Bennett, J. E., 8, 1 Bentley, T. W., 8, 151; 14, 1 Berg, U., 25, 1 Berger, S., 16, 239 Bernasconi, C. F., 27, 119 Bethell, D., 7, 153; 10, 53 Blandamer, M. J., 14, 203 Bowden, K., 28, 171 Brand, J. C. D., 1, 365 Brandstrom, A., 15, 267 Brinkman, M. R., 10, 53 Brown, H. C., 1, 35 Buncel, E., 14, 133 Bunton, C. A,, 22, 213 Cabell-Whiting, P. W., 10, 129 Cacace, F., 8, 79 Capon, B., 21, 37 Carter, R. E., 10, 1 Collins, C. J., 2, 1 Cornelisse, J., 11, 225 Crampton, M. R., 7, 211 Davidson, R. S., 19, 1; 20, 191 Desvergne, J. P., 15, 63 de Gunst, G. P., 11, 225 de Jong, F., 17, 279 Dosunmu, M. I., 21, 37 Eberson, L., 12, 1; 18, 79

Emsley, J., 26, 255 Engdahl, C., 19,223 Farnum, D. G., 11, 123 Fendler, E. J., 8, 271 Fendler, J. H., 8, 271; 13, 279 Ferguson, G., 1, 203 Fields, E. K., 6, 1 Fife, T. H., 11, 1 Fleischmann, M., 10, 155 Frey, H. M., 4, 147 Gilbert, B. C., 5, 53 Gillespie, R. J., 9, 1 Gold, V., 7, 259 Goodin, J. W., 20, 191 Gould, I. R., 20, 1 Greenwood, H. H., 4, 73 Hammerich, O., 20, 55 Harvey, N. G., 28,45 Havinga, E., 11, 225 Henderson, R. A., 23, 1 Henderson, S., 23, 1 Hibbert, F., 22, 113; 26, 255 Hine, J., 15, 1 Hogen-Esch, T. E., 15, 153 Hogeveen, H., 10, 29, 129 Huber, W., 28, 1 Ireland, J. F., 12, 131 Iwamura, H., 26, 179 Johnson, S. L., 5, 237 Johnstone, R. A. W., 8, 151 Jonsall, G., 19, 223 Jose, S. M., 21, 197 Kemp, G., 20, 191 Kice, J. L., 17, 65 Kirby, A. J., 17, 183 Kluger, R. H., 25, 99 Kohnstam, G., 5, 121 Korth, H.-G., 26, 131 Kramer, G. M., 11, 177 303

Kreevoy, M. M., 6,63; 16, 87 Kunitake, T., 17,435 Ledwith, A., 13, 155 Lee, I., 27, 57 Le Fivre, R. J. W., 3, 1 Liler, M., 11, 267 Long, F. A., 1, 1 Maccoll, A., 3, 91 Mandolini, L., 22, 1 McWeeny, R., 4, 73 Melander, L., 10, 1 Mile, B., 8, 1 Miller, S. I., 6, 185 Modena, G., 9, 185 More O’Ferrall, R. A., 5, 331 Morsi, S. E., 15, 63 Miillen, K., 28, 1 Neta, P., 12, 223 Nibbering, N. M. M., 241 Norman, R. 0. C., 5, 33 Nyberg, K., 12, 1 Olah, G. A., 4, 305 Page, M. I., 23, 165 Parker, A. J., 5, 173 Parker, V. D., 19, 131; 20,55 Peel, T. E., 9, 1 Perkampus, H. H., 4, 195 Perkins, M. J., 17, 1 Pittman, C. U. Jr., 4, 305 Pletcher, D., 10, 155 Pross, A., 14, 69; 21, 99 Ramirez, F., 9, 25 Rappoport, Z., 7, 1; 27, 239 Reeves, L. W., 3, 187 Reinhoudt, D. N., 17, 279 Ridd, J. H., 16, 1 Riveros, J. M., 21, 197

304

Robertson, J. M., 1, 203 Rose, P. L., 28, 45 Rosenthal, S. N., 13, 279 Ruasse, M.-F., 28, 207 Russell, G. A., 23, 271 Samuel, D., 3, 123 Sanchez, M. de N. de M., 21, 37 Sandstrom, J., 25, 1 Saveant, J.-M., 26, 1 Savelli, G., 22, 213 Schaleger, L. L., 1, 1 Scheraga, H. A., 6, 103 Schleyer, P. von R., 14, 1 Schmidt, S. P., 18, 187 Schuster, G. B., 18, 187; 22, 311 Scorrano, G., 13, 83 Shatenshtein, A. I., 1, 156

CUMULATIVE INDEX OF AUTHORS

Shine, H. J., 13, 155 Shinkai, S., 17, 435 Siehl, H.-U., 23, 63 Silver, B. L., 3, 123 Simonyi, M., 9, 127 Sinnott, M. L., 24, 113 Stock, L. M., 1, 35 Sustmann, R., 26, 131 Symons, M. C. R., 1, 284 Takashima, K., 21, 197 Ta-Shma, R., 27, 239 Tedder, J. M., 16, 51 Thatcher, G. R. J., 25, 99 Thomas, A., 8, 1 Thomas. J. M., 15, 63 Tonellato, U., 9, 185 Toullec, J., 18, 1 Tiidos, F., 9, 127 Turner, D. W., 4, 31

Turro, N. J., 20, 1 Ugi, I., 9, 25 Walton, J. C., 16, 51 Watt, C. I. F., 24, 57 Ward, B., 8, 1 Westheimer, F. H., 21, 1 Whalley, E., 2, 93 Williams, A., 27, 1 Williams, D. L. H., 19, 38 1 Williams, J. M. Jr., 6, 63 Williams, J. O., 16, 159 Williamson, D. G., 1, 365 Wilson, H., 14, 133 Wolf, A. P., 2, 201 Wyatt, P. A. H., 12, 131 Zimmt, M. B., 20, 1 Zollinger, H., 2, 163 Zuman, P., 5, 1

Cumulative Index of Titles Abstraction, hydrogen atom, from 0-H bonds, 9, 127 Acid solutions, strong, spectroscopic observation of alkylcarbonium ions in, 4, 305 Acid-base properties of electronically excited states of organic molecules, 12, 131 Acids and bases, oxygen and nitrogen in aqueous solution, mechanisms of proton transfer between, 22, 113 Acids, reactions of aliphatic diazo compounds with, 5, 331 Acids, strong aqueous, protonation and solvation in, 13, 83 Activation, entropies of, and mechanisms of reactions in solution, 1, 1 Activation, heat capacities of, and their uses in mechanistic studies, 5, 121 Activation, volumes of, use for determining reaction mechanisms, 2, 93 Addition reactions, gas-phase radical directive effects in, 16, 5 1 Aliphatic diazo compounds, reactions with acids, 5, 331 Alkyl and analogous groups, static and dynamic stereochemistry of, 25, 1 Alkylcarbonium ions, spectroscopic observation in strong acid solutions, 4, 305 Ambident conjugated systems, alternative protonation sites in, 11, 267 Ammonia, liquid, isotope exchange reactions of organic compounds in 1, 156 Anions, organic, gas-phase reactions of, 24, 1 Antibiotics, b-lactam, the mechanisms of reactions of, 23, 165 Aqueous mixtures, kinetics of organic reactions in water and, 14, 203 Aromatic photosubstitution, nucleophillc, 11, 225 Aromatic substitution, a quantitative treatment of directive effects in, 1, 35 Aromatic substitution reactions, hydrogen isotope effects in, 2, 163 Aromatic systems, planar and non-planar, 1, 203 Aryl halides and related compounds, photochemistry of, 20, 191 Arynes, mechanisms of formation and reactions at high temperatures, 6, 1 A-SE2 reactions, developments in the study of, 6, 63 Base catalysis, general, of ester hydrolysis and related reactions, 5, 237 Basicity of unsaturated compounds, 4, 195 Bimolecular substitution reactions in protic and dipolar aprotic solvents, 5, 173 Bromination, electrophilic, of carbon-carbon double bonds: structure, solvent and mechanism, 28, 207 3C N.M.R.spectroscopy in macromolecular systems of biochemical interest, 13,279 Captodative effect, the, 26, 131 Carbene chemistry, structure and mechanism in, 7, 163 Carbenes having aryl substituents, structure and reactivity of, 22, 31 1 Carbanion reactions, ion-pairing effects in, 15, 153 Carbocation rearrangements, degenerate, 19, 223 Carbon atoms, energetic, reactions with organic compounds, 3, 201 Carbon monoxide, reactivity of carbonium ions towards, 10, 29 Carbonium ions (alkyl), spectroscopic observation in strong acid solutions, 4, 305 305

306

CUMULATIVE INDEX OF TITLES

Carbonium ions, gaseous, from the decay of tritiated molecules, 8, 79 Carbonium ions, photochemistry of, 10, 129 Carbonium ions, reactivity towards carbon monoxide, 10, 29 Carbonyl compounds, reversible hydration of, 4, 1 Carbonyl compounds, simple, enolisation and related reactions of, 18, 1 Carboxylic acids, tetrahedral intermediates derived from, spectroscopic detection and investigation of their properties, 21, 37 Catalysis by micelles, membranes and other aqueous aggregates as models of enzyme action, 17, 435 Catalysis, enzymatic, physical organic model systems and the problem of, 11, 1 Catalysis, general base and nucleophilic, of ester hydrolysis and related reactions, 5,237 Catalysis, micellar, in organic reactions; kinetic and mechanistic implications, 8,271 Catalysis, phase-transfer by quaternary ammonium salts, 15, 267 Cation radicals in solution, formation, properties and reactions of, 13, 155 Cation radicals, organic, in solution, kinetics and mechanisms of reaction of, 20, 55 Cations, vinyl, 9, 135 Chain molecules, intramolecular reactions of, 22, 1 Chain processes, free radical, in aliphatic systems involving an electron transfer reaction, 23, 271 Charge density-N.M.R. chemical shift correlations in organic ions, 11, 125 Chemically induced dynamic nuclear spin polarization and its applications, 10, 53 Chemiluminescence of organic compounds, 18, 187 Chirality and molecular recognition in monolayers at the air-water interface, 28,45 CIDNP and its applications, 10, 53 Conduction, electrical, in organic solids, 16, 159 Configuration mixing model: a general approach to organic reactivity, 21, 99 Conformations of polypeptides, calculations of, 6, 103 Conjugated, molecules, reactivity indices, in, 4, 73 Cross-interaction constants and transition-state structure in solution, 27, 57 Crown-ether complexes, stability and reactivity of, 17, 279 D,O-H,O mixtures, protolytic processes in, 7, 259 Degenerate carbocation rearrangements, 19, 223 Diazo compounds, aliphatic, reactions with acids, 5, 331 Diffusion control and pre-association in nitrosation, nitration, and halogenation, 16,l Dimethyl sulphoxide, physical organic chemistry of reactions, in, 14, 133 Dipolar aprotic and protic solvents, rates of bimolecular substitution reactions in, 5, 173 Directive effects in aromatic substitution, a quantitative treatment of, 1, 35 Directive effects in gas-phase radical addition reactions, 16, 51 Discovery of the mechanisms of enzyme action, 1947- 1963, 21, 1 Displacement reactions, gas-phase nucleophilic, 21, 197 Double bonds, carbon-carbon, electrophilic bromination of structure, solvent and mechanism, 28, 171 Effective charge and transition-state structure in solution, 27, 1 Effective molarities of intramolecular reactions, 17, 183 Electrical conduction in organic solids, 16, 159 Electrochemical methods, study of reactive intermediates by, 19, 131 Electrochemistry, organic, structure and mechanism in, 12, 1

CUMULATIVE INDEX OF TITLES

307

Electrode processes, physical parameters for the control of, 10, 155 Electron spin resonance, identification of organic free radicals by, 1, 284 Electron spin resonance studies of short-lived organic radicals, 5, 23 Electron storage and transfer in organic redox systems with multiple electrophores, 28, 1 Electron-transfer reaction, free radical chain processes in aliphatic systems involving an, 23, 271 Electron-transfer reactions in organic chemistry, 18, 79 Electron transfer, single, and nucleophilic substitution, 26, 1 Electronically excited molecules, structure of, 1, 365 Electronically excited states of organic molecules, acid-base properties of, 12, 131 Energetic tritium and carbon atoms, reactions of, with organic compounds, 2, 201 Enolisation of simple carbonyl compounds and related reactions, 18, 1 Entropies of activation and mechanisms of reactions in solution, 1, 1 Enzymatic catalysis, physical organic model systems and the problem of, 11, 1 Enzyme action, catalysis by micelles, membranes and other aqueous aggregates as models of, 17, 435 Enzyme action, discovery of the mechanisms of, 1947-1963, 21, 1 Equilibrating systems, isotope effects on nmr spectra of, 23, 63 Equilibrium constants, N.M.R.measurements of, as a function of temperature, 3,187 Ester hydrolysis, general base and nucleophilic catalysis, 5, 237 Ester hydrolysis, neighbouring group participation by carbonyl groups in, 28, 171 Exchange reactions, hydrogen isotope, of organic compounds in liquid ammonia, 1, 156 Exchange reactions, oxygen isotope, of organic compounds, 2, 123 Excited complexes, chemistry of, 19, 1 Excited molecules, structure of electronically, 1, 365 Force-field methods, calculation of molecular structure and energy by, 13, 1 Free radical chain processes in aliphatic systems involving an electron-transfer reaction, 23, 271 Free radicals, identification by electron spin resonance, 1, 284 Free radicals and their reactions at low temperature using a rotating cryostat, study of, 8, 1 Gaseous carbonium ions from the decay of tritiated molecules, 8, 79 Gas-phase heterolysis, 3, 91 Gas-phase nucleophilic displacement reactions, 21, 197 Gas-phase pyrolysis of small-ring hydrocarbons, 4, 147 Gas-phase reactions of organic anions, 24, 1 General base and nucleophilic catalysis of ester hydrolysis and related reactions, 5,237 H,O-D,O mixtures, protolytic processes in, 7, 259 Halogenation, nitrosation, and nitration, diffusion control and pre-association in, 16,l Halides, aryl, and related compounds, photochemistry of, 20, 191 Heat capacities of activation and their uses in mechanistic studies, 5, 121 Heterolysis, gas-phase, 3, 91 High-spin organic molecules and spin alignment in organic molecular assemblies, 26, 179 Hydrated electrons, reactions of, with organic compounds, 7, 115

308

CUMULATIVE INDEX OF TITLES

Hydration, reversible, of carbonyl compounds, 4, 1 Hydride shifts and transfers, 24, 57 Hydrocarbons, small-ring, gas-phase pyrolysis of, 4, 147 Hydrogen atom abstraction from 0 - H bonds, 9, 127 Hydrogen bonding and chemical reactivity, 26, 255 Hydrogen isotope effects in aromatic substitution reactions, 2, 163 Hydrogen isotope exchange reactions of organic compounds in liquid ammonia, 1,156 Hydrolysis, ester, and related reactions, general base and nucleophilic catalysis of, 5, 237 Interface, the air-water, chirality and molecular recognition in monolayers at, 28,45 Intermediates, reactive, study of, by electrochemical methods, 19, 13 1 Intermediates, tetrahedral, derived from carboxylic acids, spectroscopic detection and investigation of their properties, 21, 37 Intramolecular reactions, effective molarities for, 17, 183 Intramolecular reactions of chain molecules, 22, 1 Ionization potentials, 4, 31 Ion-pairing effects in carbanion reactions, 15, 153 Ions, organic, charge density-N.M.R. chemical shift correlations, 11, 125 Isomerization, permutational, of pentavalent phosphorus compounds, 9, 25 Isotope effects, hydrogen, in aromatic substitution reactions, 2, 163 Isotope effects, magnetic, magnetic field effects and, on the products of organic reactions, 20, 1 Isotope effects on nmr spectra of equilibrating systems, 23, 63 Isotope effects, steric, experiments on the nature of, 10, 1 Isotope exchange reactions, hydrogen, of organic compounds in liquid ammonia, 1,150 Isotope exchange reactions, oxygen, of organic compounds, 3, 123 Isotopes and organic reaction mechanisms, 2, 1 Kinetics and mechanisms of reactions of organic cation radicals in solution, 20, 55 Kinetics, reaction, polarography and, 5, 1 Kinetics of organic reactions in water and aqueous mixtures, 14, 203 B-Lactam antibiotics, the mechanisms of reactions of, 23, 165 Least nuclear motion, principle of, 15, 1 Macromolecular systems of biochemical interest, 13CN.M.R. spectroscopy in, 13,279 Magnetic field and magnetic isotope effects on the products oforganic reactions, 20,l Mass spectrometry, mechanisms and structure in: a comparison with other chemical processes, 8, 152 Mechanism and structure in carbene chemistry, 7, 153 Mechanism and structure in mass spectrometry: a comparison with other chemical processes, 8, 152 Mechanism and structure in organic electrochemistry, 12, 1 Mechanisms and reactivity in reactions of organic oxyacids of sulphur and their anhydrides, 17, 65 Mechanisms, nitrosation, 19, 381 Mechanisms of proton transfer between oxygen and nitrogen acids and bases in aqueous solution, 22, 113 Mechanisms, organic reaction, isotopes and, 2, 1

CUMULATIVE INDEX OF TITLES

309

Mechanisms of reaction in solution, entropies of activation and, 1, 1 Mechanisms of reactions of a-lactam antibiotics, 23, 165 Mechanisms of solvolytic reactions, medium effects on the rates and, 14, 10 Mechanistic applications of the reactivity-selectivity principle, 14, 69 Mechanistic studies, heat capacities of activation and their use, 5, 121 Medium effects on the rates and mechanisms of solvolytic reactions, 14, 1 Meisenheimer complexes, 7, 21 1 Metal complexes, the nucleophilicity of towards organic molecules, 23, 1 Methyl transfer reactions, 16, 87 Micellar catalysis in organic reactions: kinetic and mechanistic implications, 8, 271 Micelles, aqueous, and similar assemblies, organic reactivity in, 22, 213 Micelles, membranes and other aqueous aggregates, catalysis by, as models of enzyme action, 17, 435 Molecular recognition, chirality and, in monolayers at the air-water interface, 28,45 Molecular structure and energy, calculation of, by force-field methods, 13, 1 Neighbouring group participation by carbonyl groups in ester hydrolysis, 28, 171 Nitration, nitrosation, and halogenation, diffusion control and pre-association in, 16, Nitrosation mechanisms, 19, 381 Nitrosation, nitration, and halogenation, diffusion control and pre-association in, 16, N.M.R. chemical shift-charge density correlations, 11, 125 N.M.R. measurements of reaction velocities and equilibrium constants as a function of temperature, 3, 187 N.M.R. spectra of equilibrating systems, isotope effects on, 23, 63 N.M.R. spectroscopy, I3C,in macromolecular systems of biochemical interest, 13,279 Non-planar and planar aromatic systems, 1, 203 Norbornyl cation: reappraisal of structure, 11, 179 Nuclear magnetic relaxation, recent problems and progress, 16, 239 Nuclear magnetic resonance, see N.M.R. Nuclear motion, principle of least, 15, 1 Nuclear motion, the principle of least, and the theory of stereoelectronic control, 24, 113 Nucleophilic aromatic photosubstitution, 11, 225 Nucleophilic catalysis of ester hydrolysis and related reactions, 5, 237 Nucleophilic displacement reactions, gas-phase, 21, 197 Nucleophilicity of metal complexes towards organic molecules, 23, 1 Nucleophilic substitution in phosphate esters, mechanism and catalysis of, 25, 99 Nucleophilic substitution, single electron transfer and, 26, 1 Nucleophilic vinylic substitution, 7, 1 OH-bonds, hydrogen atom abstraction from, 9, 127 Oxyacids of sulphur and their anhydrides, mechanisms and reactivity in reactions of organic, 17, 65 Oxygen isotope exchange reactions of organic compounds, 3, 123 Perchloro-organic chemistry: structure, spectroscopy and reaction pathways, 25,267 Permutational isomerization of pentavalent phosphorus compounds, 9, 25 Phase-transfer catalysis by quaternary ammonium salts, 15, 267 Phosphate esters, mechanism and catalysis of nucleophilic substitution in, 25, 99

310

CUMULATIVE INDEX OF TITLES

Phosphorus compounds, pentavalent, turnstile rearrangement and pseudorotation in permutational isomerization, 9, 25 Photochemistry of aryl halides and related compounds, 20, 191 Photochemistry of carbonium ions, 9, 129 Photosubstitution, nucleophilic aromatic, 11, 225 Planar and non-planar aromatic systems, 1, 203 Polarizability, molecular refractivity and, 3, 1 Polarography and reaction kinetics, 5, 1 Polypeptides, calculations of conformations of, 6, 103 Pre-association, diffusion control and, in nitrosation, nitration, and halogenation, 16,l Principle of non-perfect synchronization, 27, 119 Products of organic reactions, magnetic field and magnetic isotope effects on, 30, 1 Protic and dipolar aprotic solvents, rates of bimolecular substitution reactions in, 5, 173 Protolytic processes in H,O-D,O mixtures, 7, 259 Protonation and solvation in strong aqueous acids, 13, 83 Protonation sites in ambident conjugated systems, 11, 267 Proton transfer between oxygen and nitrogen acids and bases in aqueous solution, mechanisms of, 22, 113 Pseudorotation in isomerization of pentavalent phosphorus compounds, 9, 25 Pyrolysis, gas-phase, of small-ring hydrocarbons, 4, 147 Radiation techniques, application to the study of organic radicals, 12, 223 Radical addition reactions, gas-phase, directive effects in, 16, 51 Radicals, cation in solution, formation, properties and reactions of, 13, 155 Radicals, organic application of radiation techniques, 12, 223 Radicals, organic cation, in solution kinetics and mechanisms of reaction of, 20, 55 Radicals, organic free, identification by electron spin resonance, 1, 284 Radicals, short-lived organic, electron spin resonance studies of, 5, 53 Rates and mechanisms of solvolytic reactions, medium effects on, 14, 1 Reaction kinetics, polarography and, 5, 1 Reaction mechanisms, use of volumes of activation for determining, 2, 93 Reaction mechanisms in solution, entropies of activation and, 1, 1 Reaction velocities and equilibrium constants, N.M.R. measurements of, as a function of temperature, 3, 187 Reactions of hydrated electrons with organic compounds, 7, 115 Reactions in dimethyl sulphoxide, physical organic chemistry of, 14, 133 Reactive intermediates, study of, by electrochemical methods, 19, 131 Reactivity indices in conjugated molecules, 4, 73 Reactivity, organic, a general approach to: the configuration mixing model, 21, 99 Reactivity-selectivity principle and its mechanistic applications, 14, 69 Rearrangements, degenerate carbocation, 19, 223 Redox systems, organic, with multiple electrophores, electron storage and transfer in, 28, 1 Refractivity, molecular, and polarizability, 3, 1 Relaxation, nuclear magnetic, recent problems and progress, 16, 239 Selectivity of solvolyses in aqueous alcohols and related mixtures, solvent-induced changes in, 27, 239 Short-lived organic radicals, electron spin resonance studies of, 5, 53

CUMULATIVE INDEX OF TITLES

31 1

Small-ring hydrocarbons, gas-phase pyrolysis of, 4, 147 Solid-state chemistry, topochemical phenomena in, 15, 63 Solids, organic, electrical conduction in, 16, 159 Solutions, reactions in, entropies of activation and mechanisms, 1, 1 Solvation and protonation in strong aqueous acids, 13, 83 Solvent-induced changes in the selectivity of solvolyses in aqueous alcohols and related mixtures, 27, 239 Solvents, protic and dipolar aprotic, rates of bimolecular substitution-reactions in, 5, 173 Solvolytic reactions, medium effects on the rates and mechanisms of, 14, 1 Spectroscopic detection of tetrahedral intermediates derived from carboxylic acids and the investigation of their properties, 21, 37 Spectroscopic observations of alkylcarbonium ions in strong acid solutions, 4, 305 Spectroscopy, I3C N.M.R., in macromolecular systems of biochemical interest, 13,279 Spin alignment, in organic molecular assemblies, high-spin organic molecules and, 26, 179 Spin trapping, 17, 1 Stability and reactivity of crown-ether complexes, 17, 279 Stereochemistry, static and dynamic, of alkyl and analogous groups, 25, 1 Stereoelectronic control, the principle of least nuclear motion and the theory of, 24,113 Stereoselection in elementary steps of organic reactions, 6, 185 Steric isotope effects, experiments on the nature of, 10, 1 Structure and mechanisms in carbene chemistry, 7, 153 Structure and mechanism in organic electrochemistry, 12, 1 Structure and reactivity of carbenes having aryl substituents, 22, 31 1 Structure of electronically excited molecules, 1, 365 Substitution, aromatic, a quantitative treatment of directive effects in, 1, 35 Substitution, nucleophilic vinylic, 7, 1 Substitution reactions, aromatic, hydrogen isotope effects in, 2, 163 Substitution reactions, bimolecular, in protic and dipolar aprotic solvents, 5, 173 Sulphur, organic oxyacids of, and their anhydrides, mechanisms and reactivity in reactions of, 17, 65 Superacid systems, 9, 1 Temperature, N.M.R. measurements of reaction velocities and equilibrium constants as a function of, 3. 187 Tetrahedral intermediates derived from carboxylic acids, spectroscopic detection and the investigation of their properties, 21, 37 Topochemical phenomena in solid-state chemistry, 15, 63 Transition-state structure in solution, cross-interaction constants and, 27, 57 Transition-state structure in solution, effective charge and, 27, 1 Transition-state theory revisited, 28, 139 Tritiated molecules, gaseous carbonium ions from the decay of, 8, 79 Tritium atoms, energetic, reactions with organic compounds, 2, 201 Turnstile rearrangements in isomerization of pentavalent phosphorus compounds, 9, 25 Unsaturated compounds, basicity of, 4, 195 Vinyl cations, 9, 185

31 2

CUMULATIVE INDEX OF TITLES

Vinylic substitution, nucleophilic, 7, 1 Volumes of activation, use of, for determining reaction mechanisms, 2, 93 Water and aqueous mixtures, kinetics of organic reactions in, 14, 203