Metal clusters in chemistry - PDF Free Download

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications Edited by Rudi uun Eldik and Frank-Gerri...

Author: Pierre Braunstein | Luis A. Oro | Paul R. Raithby

200 downloads 2698 Views 20MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications Edited by Rudi uun Eldik and Frank-Gerrit Klamer

Further Titles of Interest B. Comils, W. A. Herrmann, R. Schlogl, C.-H. Wong (Eds.)

Catalysis from A to Z A Concise Encyclopadie 2000

ISBN 3-527-29588-X

F. Diederich, P. J. Stang (Eds.)

Metal-catalyzed Cross-coupling Reactions 1999

ISBN 3-527-29421-X

P. Braunstein, L. A. Oro, P. R. Raithby (Eds.)

Metal Clusters in Chemistry Three Volumes 1999 ISBN 3-527-29549-6

D. E. De Vos, I. F. J.Vankelecom, P. A. Jacobs (Eds.)

Chiral Catalysts Immobilization and Recycling 2000

ISBN 3-527-29952-1

II

FIZ CHEMIE Berlin (Ed.)

Dictionary o f Common Names Second, Extensively Enlarged Edition (5 Vol.) 2001 ISBN 3-527-30288-3

High Pressure Chemistry Synthetic, Mechanistic, and Su percritical Applications

Edited by Rudi van Eldik and Frank-Cerrit Klarner

@WILEY-VCH

Prof: Dr. Rudi yon Eldik University of Erlangen-Numberg Institute of Inorganic Chemistry EgerlandstraBe 1 91058 Erlangen Germany

Pro$ Dr. Frank-Cerrit Klarner University of Essen Institute of Organic Chemistry UniversitatsstraXe 5 45141 Essen Germany Cover Dr. Frank Wurche

University of Essen

This book was carefully produced. Nevertheless, editors, authors and publisher do not warrant the information contained therein to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for A catalogue record for this book is available from the British Library. Die Deutsche Bibliothek - CIP Cataloguing-in-Publication-Data A catalogue record for this publication is available from Die Deutsche Bibliothek

0WILEY-VCH Verlag GmbH, 69469 Weinheim (Federal Republic of Germany). 2002 All rights reserved (including those of translation in other languages). No part of this book may be reproduced in any form by photoprinting, microfilm, or any other means - nor transmitted or translated into machine language wrthout written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law Printed in the Federal Republic of Germany. Printed on acid-free paper. Typesetting Asco Typesetters, Hong Kong Printing betz-dmck gmbH, Darmstadt Bookbinding J. Schaffer GmbH & Co. KG,

Griinstadt ISBN 3-527-30404-5

lv

Contents

I

Basic Principles

1

Effect of Pressure on Inorganic Reactions: Introduction and Mechanistic Applications 3 Rudi van Eldik and Colin D. Hubbard

1

Introduction 3 1.1 Determination of Volumes of Activation 6 1.2 1.3 Thermal-Induced Reactions 12 Ligand Substitution Reactions 12 1.3.1 1.3.1.1 Octahedral Complexes 13 1.3.1.2 Square-Planar Complexes 19 1.3.2 Electron Transfer Reactions 21 1.3.2.1 Self-ExchangeReactions 22 1.3.2.2 Non-SymmetricalReactions 22 1.3.3 Actvation of Small Molecules 25 Addition and Elimination Reactions 30 1.3.4 1.4 Photo- and Radiation-Induced Reactions 32 1.5 Concluding Remarks 35 Acknowledgments 36 References 37 2

The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications 41 Frank Wurche and Frank-Cerrit Kliirner

2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.2.6 2.3 2.3.1

Introduction 41 Cycloadditions 45 Diels-Alder Reactions, Mechanistic Aspects 45 Complex Reactions and Synthetic Applications 56 1,3-Dipolar [3 21 Cycloadditions 64 [2 21 Cycloadditions 66 [6 41 and [ 8 21 Cycloadditions 68 Cheletropic Reactions 70 Pericydic Rearrangements 71 Sigmatropic [ 3.31 Shifts: Cope and Claisen Rearrangement

+ +

+ +

71

vi

I

Contents

2.3.2 2.3.3 2.3.4 2.3.5 2.4 2.4.1 2.4.2 2.4.3 2.5 2.6

Electrocyclic Rearrangements 74 Intramolecular Diels-Alder Reactions 76 Ene Reactions 77 The Relationship Between Activation or Reaction Volume and Ring Size 81 Free-Radical Reactions 84 Homolytic Bond Dissociations 84 Quinone Oxidations (Hydrogen Transfer Reactions) 85 Free Radical Cyclizations 87 Ionic Reactions 88 Concluding Remarks 90 Acknowledgments 91 References 91

3

High-pressure Kinetics and Highly Viscous Media Tsutomu Asano

3.1 3.2 3.3 3.4 3.5 3.6 3.7

High Pressure and Dynamic Solvent Effects 97 Selection of Reaction Systems 101 Z / E Isomerization of N-Benzylideneanilines 103 Z / E Isomerization of Push-Pull Substituted Azobenzenes 109 Z / E Isomerization of Carbocyanine Cations 124 Z / E Isomerization of DNAB in a Polymeric Medium 122 Concluding Remarks 125 Acknowledgments 126 References 126

II

Mechanistic and Synthetic Applications o f High Pressure

4

Water Exchange on Metal ions: The Effect o f Pressure Lothar Helm and Andrh E. Merbach Introduction 131

4.1 4.2 4.3 4.3.1 4.3.2 4.3.3 4.3.4 4.4 4.5

97

129 231

Concepts of Solvent Exchange Reactions 133 Water Exchange from the First Coordination Shell 136 First Row Transition Metal Ions 136 Second- and Third-Row Transition Metal Ions 142 Lanthanides 144 Main Group Elements 151 Water Exchange from the Second Coordination Shell 255 Conclusions 157 Acknowledgments 157 References 158

5

insights into Solution Chemistry from High Pressure Electrochemistry Thomas W.Swaddle

5.1 5.2

Introduction 161 Pressure Effects on the Kinetics of Self-Exchange Reactions

162

261

5.2.1 5.2.2 5.2.3 5.2.4 5.3 5.4 5.4.1 5.4.2 5.5

Principles 162 Experimental Observations with Aqueous Media 164 Experimental Observations with Non-aqueous Media 167 Approaches Involving Electrochemistry 168 Electrochemical Measurements at High Pressure 170 Homogeneous Versus Heterogeneous Electron Transfer 172 Aqueous Solutions 171 Non-aqueous Media 174 Conclusions 180 Acknowledgments 181 References 181

6

Pressure Effects on the Photoreactions o f Transition Metal Complexes

6.1 6.1.1 6.1.2 6.2 6.2.1 6.2.2 6.3 6.3.1 6.3.2 6.3.3 6.4 6.5

7

Pressure Effects on Excited State Energies 186 Pressure Effects on Excited State Kinetics 188 Unimolecular Excited State Reactions 196 Ligand Photosubstitution Reactions 196 Unimolecular Photoredox Decompositions 205 Bimolecular Pathways For Excited State Deactivation 206 Excited State Electron and Energy Transfer Reactions 207 Brarnsted Base Quenching 211 Reactions with Lewis Bases to Give an Excited State Complex Photochemically Generated Reactive Intermediates 213 Summary 218 Acknowledgments 22 9 References 22 9

212

Application of High Pressure in Transition Metal-Catalyzed Reactions

Oliver Reiser

7.1 7.1.1 7.2 7.2.1 7.2.2 7.2.3 7.3 7.4 7.5

Introduction 223 General Principles 224 Lewis Acid Catalyzed Reactions 224 Cycloadditions 224 Nucleophilic Substitution 229 Addition of Nucleophiles to Carbonyl Compounds Palladium Catalyzed Reactions 230 Rhodium-Catalyzed Hydroboration 236 Conclusion 237 References 237

8

High Pressure in Organic Synthesis: Influence on Selectivity Lutz F. Tietze and Peter L. Steck Introduction 239

8.1 8.2

184

Peter C. Ford and Leroy E. Laverman Introduction 284

Influence of High Pressure on Selectivity

243

229

239

223

viii

I

Contents

8.2.1 8.2.2 8.2.3 8.2.3.1 8.2.3.2 8.2.4 8.3

General Remarks 243 Chemo- and Regioselective Transformations 245 DiastereoselectiveTransformations 253 Reactions with Simple Diastereoslectivity 253 Reactions with Induced Diastereoselectivity 267 Enantioselective Transformations 275 Conclusion 280 Acknowledgements 281 References 281

9

High-pressure Promoted Cycloadditions for Application in Combinatorial Chemistry 284 George J. T. Kuster and Hans W. Scheeren

9.1 9.2 9.3 9.3.1

Introduction 284 High-pressure Diels-Alder Reactions on the Solid Phase 285 High-pressure Multicomponent Domino Cycloaddition Reactions 286 High Pressure-Promoted One-Pot Three-Component [4 2]/[3 21 Cycloadditions: Scope and Limitations 287 Mono Substituted Dipolarophiles 291 Di-Substituted and Cyclic Dipolarophiles 292 Novel 5,5-Membered Bi- and Tricyclic Nitroso Acetals 295 Pressure and Solvent Dependency 296 High Pressure-Promoted Domino [4 2]/[4 2]/[3 21 Cycloaddition of 2-Methoxy- 1,3-Butadieneand p-Nitrostyrene 298 High-pressure Domino [4 + 2]/[3 21 Cycloaddition Reactions on the Solid Phase 299 Conclusions and Outlook 302 References 303

9.3.1.1 9.3.1.2 9.3.2 9.3.2.1 9.3.3 9.4 9.5

+

+

+

+

+

+

10

Catalytic and Solvophobic Promotion o f High Pressure Addition Reactions 305 GerardJenner

10.1 10.2 10.2.1 10.2.2 10.2.3 10.2.4 10.3 10.3.1 10.3.2 10.3.2.1 10.3.2.2 10.3.2.3 10.3.3

Introduction 305 Chemical Activation 306 Catalysis by Lithium Salts 309 Catalysis by Lanthanides and Related Periodic Elements 312 Catalysis with Other Lewis Acids 319 Catalysis by Phosphines 320 Solvophobic Activation 321 Water and Water-like Solvents 322 Kinetic Studies in Water and Water-like Solvents 324 Michael Reactions 325 Diels-Alder Reactions 327 Effect of Addition of Micelles and Cyclodextrines in Aqueous Media Synthetic Applications 337

336

Contents

10.3.3.1 10.3.3.2 10.3.2.3 10.3.3.4 10.4

Diels-Alder Reactions 338 Michael Reactions 342 Baylis-Hillman Reactions 343 Triactivation 343 Conclusions 344 References 345

11

Future Perspectives: Applications o f High Pressure in Supramolecular Chemistry 348 Robert Rulofl Christophe Saudan, Andri E. Merbach and Frank-Cerrit Klarner Introduction 348 Biomolecules under Extreme Conditions 349

11.1 11.2 11.3 11.3.1

The Effect of Pressure on the Formation of Host-Guest Complexes 352 Inclusion of Helium Atoms or Acetonitsile Molecules in C ~ or Oa Hemicarcerand as Molecular Containers 351 11.3.2 Complexation of Cations and Cboas Guests with Crown Ethers, Cryptands, and Calixarenes as Hosts 353 11.3.3 Molecular Tweezers as Synthetic Receptors: Focussing on Volume and Entropy of Association 354 11.3.4 Formation of Host-Guest Complexes of a-Cyclodextrins with Azo Dyes: Determination of Activation and Reaction Volumes 357 11.3.4.1 Inclusion of Short Guests into the a-Cyclodextrin Cavity 357 11.3.4.2 Sequential Threading of a-Cydodextrin onto a Long Guest 362 11.3.5 Self-Assembled Multinudear Coordination Species with Chiral Bipyridine Ligand 363 Conclusion and Outlook 367 11.4 Acknowledgments 368 References 368 Ill

Chemical Reactions in Supercritical Fluids

12

Catalytic Reactions in Supercritical Fluids 372 Jason Hyde, Walter Leitner and Martyn Poliakoff

12.1 12.1.1 12.1.2 12.1.3 12.2 12.2.1 12.2.2 12.3

Introduction to Catalytic Reactions in Supercritical Fluids 372 Solvent Properties of Supercritical Fluids 372 Temperature and Pressure Relations 372 Decaffeination of Coffee via scCOz Extraction of Caffeine 374 Practical Aspects of Catalpc Reactions in scC02 375 Heterogeneously Catalyzed Reactions 375 Homogeneously Catalyzed Reactions 377 Acid-Catalyzed Continuous Flow Processes in Supercritical Fluids 379 Heterogeneously Catalyzed Alkylation Reactions 379 Heterogeneously Catalyzed Etherification and De-symmeterization Reactions 380

12.3.1 12.3.2

~

369

I

ix

x

I

Contents

12.4 12.5 12.5.1 12.5.2 12.6 12.6.1 12.6.2 12.7

Homogeneously Catalyzed C-C Coupling Reactions 381 Hydrogenation Reactions 383 Heterogeneously Catalyzed Hydrogenation Reactions 383 Homogeneously Catalyzed Hydrogenation Reactions 386 Hydroformylation Reactions 389 Homogeneously Catalysed Hydroformylation Reactions 389 Heterogeneously Catalyzed Hydroformylation Reactions 393 Closing Remarks 394 References 395

13

Application of Supercritical Fluids in the Fine Chemical Industry

398

Werner Eonrath and Reinhard Karge

13.1 13.2 13.2.1 13.2.2 13.3 13.3.1 13.3.2 13.3.3 13.3.4 13.3.5 13.4

Introduction 398 Supercritical Fluids in Separation/Purification 400 Supercritical Fluid Extraction (SFE) 400 Supercritical Fluid Chromatography (SFC) 403 Catalytic Reactions in Supercritical Fluids 405 Hydrogenation 406 Methylation 42 1 Friedel-Crafts Alkylation Reactions 412 Oxidation 415 Other Reaction Types 41 7 Concluding Remarks 419 Acknowledgments 41 9 References 41 9

14

Applications of Supercritical Water Eckhard Dinjus and Andrea Kruse Introduction 422

14.1 14.2 14.3 14.3.1 14.3.1.1 14.3.1.2 14.3.1.3 14.3.1.4 14.3.1.5 14.3.1.6 14.3.1.7 14.3.2 14.3.2.1 14.3.2.2 14.3.2.3 14.3.3

422

Physico-Chemical Properties of Water at High Temperature and Pressure and their Relation to Applications 423 Supercritical Water in Chemical Synthesis 425 Organic Reactions 425 Hydrolysis Reactions 425 Condensations 427 Diels-Alder Reactions 427 Rearrangements 428 Friedel-Crafts Reactions 429 Partial Oxidations 430 Reduction Reactions 43 I Organornetallic Reactions 43 I Heck Coupling 431 Cyclotrimerization of Alkynes 432 Hydroformylation 433 Inorganic Reactions 434

Contents

14.4 14.4.1 14.4.2 14.5 14.6

Supercritical Water in Decomposition Reactions Oxidation in Supercritical Water 434 Gasification in Supercritical Water 437 Conclusions 440 Outlook 441 References 442 Index

447

434

I

xi

I

Preface High pressure chemistry is an area that has developed a vigorous activity over the past decades. Although most of the earlier work was mainly performed in the area of organic chemistry, a major contribution from inorganic chemists over the past two to three decades resulted in the development of sophisticated instrumentation that enables the study of fast chemical reactions under high pressure. More recently, the application of supercritical fluids has received much attention especially in chemical industry. Numerous reviews have reported on the progress made in these areas over the past years. The monograph consists of fourteen contributions bascd on oral presentations at the European High Pressure Research Group Meeting held at Kloster Banz, Germany, in September 2000. The theme of the meeting was High Pressure Chemistry. It covers contributions from high pressure inorganic and organic chemistry, as well as the application of supercritical fluids in chemical synthesis and processes. The monograph is subdivided into three sections. The first three chapters are devoted to basic principles involved in the application of high pressure techniques in inorganic and organic chemistry. The subsequent eight chapters are devoted to mechanistic and synthetic applications of high pressure in inorganic, organometallic, organic, and supramolecular chemistry. The final three chapters are devoted to chemical reactions in supercritical fluids and cover catalytx reactions, applications in the fine chemical industry and the application of super critical water. All in all, the individual chapters reveal the present status of high pressure chemistry and its application in a variety of areas. The editors appreciate the co-operative support they received from the individual authors of the chapters, as well as the effective interaction with Wiley-VCH. The efforts of numerous scientific coworkers and the financial support from many funding agencies have all contributed to bringing high pressure chemistry to where it stands at present. May we all in future benefit from these developments and stimulate further activities of the next generation in this area of chemistry. Rudi van Eldik University of Erlangen-Niirnberg

Frank-Gerrit Klarner University of Essen

xiii

I

xv

List of Authors Tsutomu Asano Department of Applied Chemistry Faculty of Engineering Oita University 700 Dannoharu Oita 870-1192 Japan Werner Bonrath Vitamins and Fine Chemicals Division Chemical Process Technology F. Hoffmann-LaRoche Ltd. Grenzacherstrage 124 CH-4070 Basel Switzerland

CH-1015 Lausanne Switzerland Colin D. Hubbard Institute for Inorganic Chemistry University of Erlangen-Niirnberg Egerlandstrage 1 91058 Erlangen Germany lason Hyde School of Chemistry University of Nottingham University Park Nottingham NG7 2RD U.K.

Eckhard Dinjus Institut fur Technische Chemie Chemisch Physikalische Verfahren Forschungszentrum Karlsruhe Hermann-von-Helmholt-Platz 1 76344 Eggenstein-Leopolshafen Germany

Gerard Jenner Laboratoire de Pikzochimie Organique Universite Louis Pasteur 1: Rue Blake Pascal 67008 Strasbourg Cedex France

Rudi van Eldik Institute for Inorganic Chemistry University of Erlangen-Nurnberg Egerlandstrage 1 91058 Erlangen Germany

Reinhard Karge Vitamins and Fine Chemicals Division Chemical Process Technology F. Hoffmann-LaRoche Ltd. Grenzacherstrage 124 4070 Basel Switzerland

Peter C. Ford Department of Chemistry and Biochemistry University of California Santa Barbara CA 93106 U.S.A. Lothar Helm Institut de Chimie Minerale et Analytique Universite de Lausanne BCH

Frank-Gemt Klarner Institut fur Organische Chemie Universitat Essen Universitatsstrage 5 45141 Essen Germany Andrea h s e Institut fur Technische Chemie, Chemisch Physikalische Verfahren

xvi

I

List ofAuthors

Forschungszentrum Karlsruhe Hermann-von-Helmholtz-Platz 1 76344 Eggenstein-Leopolshafen Germany

George J. T. Kuster Department of Organic Chemistry NSR Center for Molecular Structure, Design and Synthesis University of Nijmegen Toernooiveld 1 6525 ED Nijmegen The Netherlands Leroy E. Laverman Department of Chemistry and Biochemistry University of California Santa Barbara CA 93106 U.S.A. Walter Leitner Max-Planck-Institut fur Kohlenforschung Kaiser-Wilhelm-Platz 1 45470 Miilheim an der Ruhr Lehrstuhl fur Technische Chemie und Petrolchemie RWTH Aachen Woringer Weg 1 52056 Aachen Germany Andre E. Merbach Institut de Chimie Minkale et Analytique Universiti. de h u s a n n e BCH 1015 Lausanne Switzerland Martyn Poliakoff School of Chemistry University of Nottingham University Park Nottingham NG7 2RD U.K. Oliver Reiser Institut fur Organische Chemie Universitat Regensburg UniversitatsstraBe 31 93053 Regensburg Germany

Robert Ruloff lnstitut de Chimie Mmerale et Analytique Universite dc Lausanne BCH 1015 Lausanne Switzerland Christophe Saudan Institut de Chimie Minerale et Analytique Universitk de Lausanne BCH 1015 Lausanne Switzerland Hans W. Scheeren Department of Organic Chemistry NSR Center for Molecular Structure. Design and Synthesis university of Nijmegen Toernooiveld 1 6525 ED Nijmegen The Netherlands Peter L. Steck Institut fur Organische Chemie Georg-August-Universitat Tammannstrasse 2 37077 Gottingen Germany Thomas W. Swaddle Department of Chemistry University of Calgary Calgary Alberta T2N 1N4 Canada Lutz F. Tietze Institut fur Organische Chemie Georg-August-Universitat Tammannstrasse 2 37077 Gottingen Germany Frank Wurche Institut fur Organische Chemie Universitat Essen UniversitatsstraRe 5 45141 Essen Germany

I’

I

Basic Principles

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

13

1

Effect o f Pressure on Inorganic Reactions: Introduction and Mechanistic Applications Rudi van €Idib: and Colin D. Hubbard I. I

Introduction

Chemistry literature is to a large extent concerned with preparative work and the structural and spectroscopic characterization of reaction products. The velocity of the reactions and efficiency of product formation as manifested in the reaction yield, are also of importance in synthetic studies, particularly when the products are of direct use or are intermediates in commercially relevant activities. The kinetics of reactions can be very informative in combination with other information for revealing the details of the reaction mechanism. Once a chemical reaction mechanism is fully understood, the insight gained can be used to tune the chemical process in any desired direction. The evidence for a particular mechanism is often circumstantial, and therefore kineticists try to employ the widest set of experimental variables available in an effort to interpret the resulting kinetic data in the least equivocal manner possible. The value of the mechanistic information that emerges from kinetics measurements over a series of elevated pressures for solution reactions in inorganic and organic chemistry has been realized for some time [ l - 3 ) . However, many inorganic reactions are too fast to follow using conventional instrumentation. Hence the momentum regarding investigations at high pressures vis-a-vis organic reactions was delayed somewhat until adaptation of rapid reaction techniques for operation at high pressures had been achieved, mostly in the period from 1975 to 1985. This fertile period has been recorded in reviews, in conference proceedings, and in monographs, and readers may obtain a thorough background and sense of historical development by consulting this literature [4-111. Even until quite recently, suitable instrumentation was not widely available. The purpose of this chapter is to familiarize the reader with the current status of activities in the application of hydrostatic pressure to mechanistic studies in the areas of inorganic and organometallic chemistry, as well as in the blossoming field of bioinorganic chemistry. Although the basic principles involved in high pressure kinetics for reactions in general have been the subject of many reports [12-141, some essential aspects and the most frequently used methods will be presented

4

I

7 Effect of Pressure on Inorganic Reactions: Introduction and Mechanistic Applications

here to form a basis for the subsequent chapters dealing with the effect of pressure on particular types of reactions in inorganic and organic chemistry. The parameter that is derived from high pressure kinetic experiments in solution is the difference in partial molar volume between the activated complex of transition state theory and the reactant state, and is known as the volume of activation. AVf . If the particular reaction is reversible and the system experimentally accessible, AVf for the reverse reaction can also be obtained and the difference between these two quantities results in the reaction volume, AVO. The latter quantity may also be determined by measuring the equilibrium constant ( K ) for the reaction as a function of pressure, or from the partial molar volumes of the reactants and products, derived from solution density measurements. The volume of activation itself is determined from measurements of the reaction rate constant k at different hydrostatic pressures p at a given absolute temperature T, since (? In k/?p), = -AV#/RT (R is the ideal gas constant), an equation was developed within transition state theory based upon the analogous equilibrium constant relationship, ( 3 In K/Sp), = -AVo/RT. The former equation, upon integration. can be employed to determine AVz from a plot of In k versus p . Providing the pressure is no higher than 200 MPa, in the vast majority of cases AV# is pressure independent and the plot is linear. A nonlinear behaviour is usually encountered when dealing with a compressible solvent where both the reaction and activation volume become pressure sensitive. For such cases often encountered in organic systems (see Chapter 2), where it is necessary to consider the pressure dependence of AV#, i.e. to extrapolate the data to ambient pressure, there are various treatments available for processing the primary data [5, 151. In this introductory chapter the focus will be on reactions in which there is a negligible or absence of pressure dependence of the volume of activation. In general, volume of activation data quoted in this report will refer to ambient conditions, i.e. close to room temperature, and readers are advised to consult the cited literature for more detailed information on the exact experimental conditions employed. Equilibrium and kinetic parameters obtained as a function of temperature permit the drawing of diagrams illustrating the Gibbs free energy (C), enthalpy (H) and entropy (S) changes in proceeding in the sequence reactant state/transition state/product state, and including intermediates when they are formed. Correspondingly, a volume diagram or volume profile can chart the respective volume changes along the reaction coordinate, and when appropriate actual partial molar volumes are known, on an absolute rather than a relative basis, something that cannot realized for G, H or S. Hence if reactants A and B form a product AB and no intermediates are formed, i.e. there is a single step reaction, a volume profile in which the reaction volume is, for example, negative and the volume of activation is such that the transition state is almost halfway between reactant and product states, is depicted in Fig. 1.1. As shown in Fig. 1.1,other forms of the volume profile are possible depending on the particular character of the system. Thus, in principle, a volume profile represents a simple and lucid way of describing a reaction and diagnosing the mechanism, but with the caveat that mechanistic diagnosis is uncomplicated when only intrinsic changes (changes in bond lengths, bond angles for example) occur. In

7 . 1 Introduction

15

r

I

I

..........................

A+B

Reactants

Transition State

Products

Reaction Coordinate Fig. 1.1.

Volume profile for the overall reaction A

+B

-+

AB. The activated complex is [A----B]#.

many actual reactions, when charged species are produced or neutralized during the reaction, or increases or decreases in polarity occur, then there is also a change in the volume occupied by the solvent molecules surrounding the system by virtue of an increase or decrease in (at least) the first solvation layer. Volume reduction of solvent from this source is known as electrostriction. Thus the facile interpretation of measured values of AVO or AV# can be compromised by the existence of the two contributions which are difficult to quantify. The intrinsic and solvational contributions to AVf can schematically be visualized as shown in Fig. 1.2.

A

B

(A---B)*

A- B

o+o-m-m --

Forward reaction: Reverse reaction:

A + B AB

A

AViZtr = AVittr = +

AB B

+

o +0-00- -m

Forward reaction: Reverse reaction:

-*

A' + BAB A+

Overall volume effect: AV Fig. 1.2.

+

AV:, = AV&,, =

A8 B-

*

*

= AVintr + AVsolv

Intrinsic and solvational contributions t o the volume of activation.

+

-

6

I

7 Effect of Pressure on Inorganic Reactions: Introduction and Mechanistic Applications

Clearly a reaction accelerated by pressure has a negative volume of activation and one retarded by pressure. a positive volume of activation. Most inorganic reactions that have been studied yield AVc values within the range of 30 to -30 cm' mol-'. which corresponds to retardation and acceleration respectively of a factor of about 4 at 100 MPa (1 kbar) compared to 0.1 MPa (atmospheric pressure;. In the absence of solvational contributions. positive AVf values are indicative of' the commencement of bond breakage. whereas negative AVc values are indicative of reactions in which a bond is beginning to be established upon reaching the transition state. Further classification of reaction types will be presented later. A brief account of experimental methods follows with illustrations in some cases. Thereafter thermal reactions grouped by reaction type from inorganic, organometallic and bioinorganic chemistry will be described from the perspective of the mechanistic insight gained from the application of high pressure techniques. A section on photo- and radiation-induced chemical reactions is also included. At this stage readers are referred to more detailed reports on water cxchange processes in Chapter 4,application of electrochemical techniques in Chapter 5. and photochemical processes in Chapter 6. '

1.2

Determination of Volumes of Activation

The scope of activity in the overall field in question may be gauged by the number of pertinent papers published or the number of volume parameters reported. Up to 1978 about 170 of the latter values had been published. while in the subsequent two decades approximately 1000 and 1600 values of the activation volume. respectively. have been reported [8, 16, 171. The most frequently used method of rnonitoring a rcaction in coordination chemistry is by following changes in the UV:Vis spectrum either with a conventional spcctrophotometer or with a stopped-flow instrument. For conventional time range reactions (reaction times longer than a few minutes) using Uy/Vis spectroscopy, a two-window cell (Fig. 1.3) and a pressurizable cuvette (pill-box)(Fig. 1.4)may be used for high pressure measurements 1181. The advantage of the pill-box cuvette is that pressure can be transmitted through the compression of the movable, closcly fitting cylindrical parts: it can be easily filled using a syringe needle techniquc. after which the two cylindrical parts are turned 180" to seal the cuvette. When the cell is pressurized. the two cylindrical parts move closer together as a result of the compression of the solvent used in the sample solution within the cell. and therefore the pressure from the pressurizing medium is transmitted to the sample solution. The cell is pressurized with a pressure generating system which typically consists of the components shown in Fig. 1.5. An hydraulic p u m p is used to gcnerate an oil pressure. which is then transmitted by the separator to the pressurizing medium (for instance water) used within the optical cell. Compression and expansion can be controlled with the series of mechanical valves and monitored with a pressure gauge. The type of high pressure cell in Fig. 1.3 can also be used to construct a three-

1.2 Determination of Volumes of Activation 17

I Fig. 1.3. Schematic view o f a two-window high-pressure cell: 1 - pressure plug; 2 - O-ring; 3 - reaction compartment; 4 A- and O-ring; 5 - sapphire window; 6 - pressure connection

or four-window cell which may be used for flash photolysis and pulse radiolysis applications. Technical details of these cells and methods of use may be found in recent literature [ 18-20]. The development of high pressure stopped-flow instruments opened up the possibility to study reactions in the millisecond and second time range as a function of pressure [21-271. A stopped-flow instrument is designed to enable the rapid

Fig. 1.4. Schematic presentation of a "pill-box'' optical cell for measurements i n a high pressure optical cell. The slot and hole allow the pill-box cell t o be filled and extra liquid t o be released on closing the cell.

8

I

7 €fled of Pressure on Inorganic Reactions: Introduction and Mechanistic Applications

1

7 reservoir filled with n-heptane

valve oil reservoir

high-pressure cell compartment with optical windows and pill-box cell

teflon membrane in a steel cylinder, used for pressure transmission and for separatlon of oil and n-heptane

to vacuum pump

~-

Fig. 1.5.

Typical system to generate high pressure.

mixing of two solutions containing the reactants, followed by the monitoring of the reaction progress when mixing is completed. Two high pressure versions of stopped-flow instruments are shown in Figs 1.6 and 1.7, with the difference that in the first case (Fig. 1.6) the activation of the syringes occurs by means of a motor inside the high pressure cell, whereas in the second case (Fig. 1.7) the syringes are activated from outside the cell. Activation of the sample syringes causes a flow of the two reagent solutions through a mixing jet and optical path into a receiver syringe, which is followed by the activation of the optical detection system that then monitors the reaction progress of the rapidly mixed reagents occurring in the optical path. The deadtime of the mixing process is between 2 and 10 ms. A second important method for determining AV# is by application of N M R spectroscopy. Progress in NMR instrumentation from electromagnets to superconducting magnets and higher field strengths has largely been matched by developments in construction of suitable high pressure probes for newer instruments in individual laboratories [29-38). Investigations of solvent exchange (see Chapter 4) and electron self-exchange reactions have been the principal beneficiaries of progress in high pressure NMR techniques. A typical example of an NMR high pressure probe developed in our laboratories, is shown schematically in Fig. 1.8. The operation principle of both these high pressure probes is that the NMR sample tube is placed within a high pressure cell and is pressurized with a suitable fluid by a movable stopper that transmits the pressure from the pressurizing fluid to the sample solution by moving down the NMR tube, which is controlled by the compressibility of the solvent used in the sample solution. With the aid of these high pressure probes practically all possible NMR measurements can be performed as a function of pressure up to 200-300 MPa (i.e. 2 to 3 kbar) at a fixed

7.2 Determination of Volumes of Activation

_

_

_

~

Fig. 1.6. Schematic representation of a high pressure stopped-flow unit: 1 - lid t o overall unit; 2 - outer vessel; 3 - window holder; 4 quartz windows; 5 - electric motor; 6 - motor actuator; 7 - stopped-flow unit positioning

__

rod; 8 - syringe-driving plate; 9 - drive syringe (inner); 10 - drive syringe (outer); 11 - block holding windows, mixer and syringe attachment points; 12 - mixing jet; 13 - stop syringe (outer); 14 - stop syringe (inner).

temperature. The only restriction is that the sample tube within the high pressure probe cannot be spun. The temperature-jump technique [39, 401 is frequently used to study the kinetics of rapidly equilibrating processes in solution on a microsecond time scale. This technique can only be applied to equilibria that are sensitive to temperature, such that a rapid temperature jump of a few degrees will result in a relaxation of the system to the new temperature, a process that can be followed on a micro- or millisecond time scale.

19

10

I

7 Efect of Pressure on Inorganic Reactions: lntroduction and Mechanistic Applications

_

_

b-.Thrust

_

_

_

~ rod

7

Mndow mount for the entwing light

Observatim cell

Hydraulic pressure connection Port of the temperature Sensor

L Lower tiate I

Coolant

ObservationCell Absorbance sapphirewindow Fluorescence sapphirewindow Fluorescence window mount

Fig. 1.7. Schematic representation of the commercially available Hi-Tech HPSF-56 high pressure stopped-flow unit [28].

Electrochemical methods have also been adopted for application of high pressure [41-431 (see Chapter 5). Correlations emerging from these investigations have val-

uable application in the interpretation of partial molar volume changes associated with electron transfer reactions (see Sect. 1.3.4).A potential future interest is in reactions carried out at elevated pressures in a supercritical fluid medium; in view of this a special optical cell has been developed for studying organometallic reactions initiated by flash photolysis in supercritical fluids [lo] (see Chapters 12 to 14).

The principles and instrumentation outlined above have been applied to numerous types of reactions in inorganic chemistry. A systematic treatment of the different reaction types and specific examples to illustrate the role of high pressure measurements in such studies, now follows.

7.2 Determination of Volumes of Activation

1

15

2

3 aluminum

a

4

16

17

probe jacket lower plug titanium ring

lower pressure screw

18

eroded titanium tube rnacor

19

..

PTFE

5 6 20

7

8

9

10

21

11

12

22

13

23

14

24

Fig. 1.8. Design features of a probe head for high-pressure N M R (400 MHz) measurements: 1 - O-ring; 2 - probe jacket; 3 - thermal insulation; 4 - polyvinyl chloride; 5 - O-ring; 6 - O-ring; 7 - semi-rigid coaxial cable; 8 - connection to thermostat; 9 - titanium tube; 10 - lid; 11 - screw; 12 - capacitor;

13 - capacitor holder; 14 - aluminum tube; 15 - upper plug; 16 - sample tube; 17 saddle coil; 18 - Macor; 19 - TiA16V4 vessel; 20 - lower plug; 21 - lower pressure screw; 22 - capacitor; 23 - coaxial cable; 24 capacitor holder.

I

”

12

I

7 Effect of Pressure on lnorganic Reactions: lntroduction and Mechanistic Applications

1.3

Thermal-Induced Reactions

In this section of our presentation we will focus on different types of reactions in inorganic chemistry that occur thermally. In Sect. 1.4 we will present an account of radiation-induced inorganic reactions. Photo-induced inorganic reactions are dealt with in Chapter 6. 1.3.1

Ligand Substitution Reactions

Ligand substitution reactions of metal complexes have been the topic of many mechanistic studies in coordination chemistry because of the fundamental role of such reactions in many chemical, biological and catalytic processes. For a general ligand substitution reaction as shown in Eq. (1.1),

where X is the leaving group, Y the entering ligand, and L,, the spectator ligand(s) (charges are omitted for clarity), there are basically three simple pathways: (i) the dissociative (D) process with an intermediate of lower coordination number; (ii) the associative (A) process with an intermediate of higher coordination number; (iii) the interchange (1) process, in which no intermediate of lower or higher coordination number is involved. The interchange of the ligands X and Y can be more dissociative ( I d ) or more associative (Ia) in nature, depending on whether bond breakage or bond formation is more important, respectively. These mechanisms are outlined schematically in Fig. 1.9, Such ligand substitution reactions should exhibit very characteristic AV# values depending on the degree of bond breakage or bond formation in the transition state. The most simple type of ligand substitution reaction involves the symmetrical exchange of coordinated solvent or ligand with bulk solvent or ligand molecules, respectively. [M(S),]"+

+ S'

--f

+S

[M(S),-,(S')]"'

Exchange of a unidentate solvent molecule (S) between the first coordination sphere of a solvated metal ion (M"+)and the bulk solvent (Eq. (1.2)) has been studied for cations of many elements of the Periodic Table. The incoming solvent molecule S" is denoted with an to distinguish it from the initially coordinated molecule with which it exchanges. Such reactions are very important and a knowledge of the kinetic and associated activation parameters represents important background to the understanding and tuning of substitution of a solvent by other ligands 144).The focus has frequently been; but by no means exclusively, on water as solvent. There is no reaction volume and the solvent exchange process is assumed to have zero solvational change. Thus A V f should be a direct measure of the intrinsic volume changes that occur, such that a continuous spectrum of tranQ

1.3 Thermal-lnduced Reactions

Reaction: MLnX + Y

-

MLnY

+

X

Mechanism:

D

Id

Schematic presentation of the possible ligand substitution mechanisms. In the case of the limiting D and A mechanisms, the transition states indicate the degree of bond breakage or bond formation, respectively. Fig. 1.9.

Ia

A

!

For the interchange mechanisms I d and la, the reactants are suggested t o form a precursor complex in a rapid pre-equilibration prior to the rate-determining interchange o f X and Y.

sition configurations can be envisaged, ranging from a very expanded, highly dissociative one (large and positive A V z ; rate constant significantly slowed down by pressure) to a very compact, highly associative one (large and negative A V # ; rate constant significantly accelerated by pressure). A detailed treatment of the effect of pressure on water exchange processes is given in Chapter 4;readers are referred to that chapter for more details. A typical non-symmetrical ligand substitution reaction was presented in Eq. (3.1). When X is a solvent molecule, this process is referred to as complex-formation or anation reaction, depending on whether the entering ligand Y is neutral or anionic, respectively. When X is not the solvent and Y is, then the reaction is an aquation reaction in aqueous medium or a solvolysis reaction in other solvents. Another category of substitution reactions is base hydrolysis, in which typically for a six coordinate complex, with n = 5, and X a variety of unidentate ligands, Y is either the hydroxide ion or supplies the OH- ion. Thus ligand substitution reactions encompass a wide variety of chemical reactions, and each type has received mechanistic benefit from high pressure kinetics studies. 1.3.1.1

Octahedral Complexes

Complex-formation reactions are intimately connected mechanistically with solvent exchange reactions since solvent departure may be rate determining following

14

I

I Effect of Pressure on Inorganic Reactions: Introduction and Mechanistic A p p h t i o n s

the rapid formation of an outer sphere complex or, in the case of Y being charged, an ion-pair. It is therefore not surprising that such ligand substitution reactions exhibit pressure dependencies that correlate closely with those found for solvent exchange processes [lo, 171. In the case of substitution by a ligand for one or more solvent molecules, the volume profile reveals the location of the transition state with respect to the reactant and product states, such that A V f can be related to the reaction volume AV. The formation of the mono 2,2'-bipyridine (bpy) complexes of Zn(I1) and Cd(I1) aqua ions [45, 461 provides excellent straightforward examples of both size influence on mechanistic determination and complementarity of activation volume and reaction volume measurements. An improved access to the coordinated solvent by the ligand in the outer sphere permits reaction via an I, mechanism for formation of [Cd(bpy)(H20)4]2 + , a process not possible in the formation of the analogous Zn(I1) complex, which is formed from a smaller hexaaqua ion, resulting in an Id mechanism. Figure 1.10 displays the volume profiles for these two complex-formation reactions. The mechanism proposed for the zinc complexformation is entirely consistent with recent calculations for water exchange on [Zn(Hz0)6I2' [471. Thus the corresponding volume profiles illustrate the looseness or compactness of the transition state during the ligand substitution process. Such analyses have also been applied to mechanistic questions in organometallic chemistry. In one case the volume profile for the reversible substitution of coordinated CO by P(OMe)3 on [Cr(phen)(C0)4],phen = 1,lO-phenanthroline, could be constructed on an absolute partial molar volume scale, since the partial molar volumes of all reactant and product species were determined through density measurements [48]. The volume profile in Fig. 1.11 demonstrates the significantly higher partial molar volume of the transition state and the operation of a dissociative (D) mechanism in both directions. The overall reaction volume is negative and indicates a volume decrease during the reaction. In reality the complex-formation rate constant is a composite of the precursor complex or ion-pair formation constant (KO)and the subsequent interchange rate constant (ki), i.e. k = Koki, as shown in Eqs (1.3) and (1.4). ML,X

+ Y + {ML,X

{ML,X. Y}

+

ML,Y

.Y}, &

+ X,

k,

(1.4)

The contributions from these terms frequently cannot be separated with the result that only the composite term k is known. One example for which the separation of these terms was possible, concerns the complex-formation reaction of aquacobalamin (vitamin B12). Here the usually inert Co(II1) center is labilized by the corrin ring, which induces a dissociative substitution mode. From the non-linear dependence of the observed pseudo-first-order rate constant on the entering ligand concentration for the reaction shown in Eq. (1.5), the precursor formation constant and rate-determining interchange constant can be determined, as can their pressure dependences.

7.3 Thermal-Induced Reactions 115

M=Zn

-----f--

AVO= +3.5 f 0.5

M=Cd V

Fig. 1.10. Volume profiles for the reversible formation of [M(H2O)d(bpy)l2+ complexes M = Zn(ll) and Cd(l1).

A typical volume profile for Eq. (l.S), where L = pyridine (py), is shown in Fig. 1.12, which clearly demonstrates the dissociative nature of the transition state and supports the operation of an Id mechanism [49,501. The small volume increase associated with precursor formation is ascribed to partial desolvation on forming the encounter complex. As in the case of solvent exchange reactions, the rate and mechanism of ligand substitution reactions can be systematically tuned through manipulation of steric and electronic effects. The introduction of a metal-carbon bond on an inert metal

16

I

I EfPect of Pressure on Inorganic Reactions: lntroduction and Mechanistic Appkations

E

[Cr(phen)(CO)g+ CO

+

P(OMe)31z

+13.82 0.5 +19.2 f 0.5 Cr(phen)(C0)4+ P(OMel3 ........

.c L

n

3501

-

I

Cr( phen)( C05P( OMe), calc: -5.4t1.0 exp: -4 :1 _._. ......... ' co

Reactants

'Transition State'

Fig. 1.11. Volume profile for the reversible reaction [Cr(phen)(CO)sJT P(0Me)S + [Cr(phen)(CO),P(OMe),]

Products t CO.

center such as Rh(II1) should also lead to a drastic increase in reactivity, accompanied by a possible changeover in mechanism. A system studied in detail in our group concerned the substitution behavior of rhodoxime complexes of the type trans-[Rh(dmg)2(R)H20],where dmg = dimethylglyoximate and R = CH3, CH2C1, or CH2CF3 [Sl]. Depending on the donor properties of R, the rate of complexformation with nucleophiles such as pyridine (Py), thiourea (TU), and tetramethylthiourea (TMTU) can be varied over several orders of magnitude. This is demonstrated in Fig. 1.13 for a particular entering ligand concentration for different Rs as a function of pressure. The reaction rate is largest when R = methyl, and the

+7.1

- _.

+ 16.9

Q

a, z ._ c

a

-?

-1 0

-15

LII Fig. 1.12.

Reactants

I

Precursor

I Transition stale I

Products

Reaction coordinate Volume profile for the overall reaction [B12-H20]-

t

L + [Blz-LI-

+ H2O

-

7.3 Thermal-Induced Reactions

3p

2

rMTU

PY

ru

-

-2 0

ru

50 100 15

Fig. 1.13. Plots of In kobz versus pressure for the reaction

+

trans-[Rh(drng)z(R)HzO] L ---* trans-[Rh(dmg)z(R)L] i H20.

reaction is characterized by a small positive volume of activation, typical for an Id mechanism. On decreasing the donor strength of R by the introduction of electronwithdrawing groups, the observed rate constant decreases by up to 5 orders of magnitude. In addition, there is no pressure dependence in the case of R = CH2C1, consistent with a pure interchange (I) mechanism, whereas the rate constant increases with pressure for R = CH2CF3and results in a small negative volume of activation typical for an I, mechanism. This means that the nature of the transition state can be tuned very accurately through the donor properties of R, which then determines the rate as well as the mechanism of the substitution process. This is demonstrated schematically in Fig. 1.14. Thus a decrease in donor strength slows down the substitution reaction and causes a changeover in substitution mechanism from more dissociative ( I d ) to more associative ( Ia) [ 511. Interestingly enough, whether the organic group was varied or not for a given nudeophile, all reactions studied were characterized by moderately negative entropies of activation, which by contrast demonstrates the mechanistic discrimination power of pressure versus temperature as a physical variable. The introduction of a tripodal tetradentate ligand such as tren or tmpa on the aqua Cu(I1) ion induces a changeover in the water exchange mechanism [52-541. Complex-formationreactions of these complexes exhibited a similar trend. Volume profiles for the complex-formation reactions (with pyridine) of [Cu(tren)HzO]2+ and [Cu(Me3tren)HzO]2+ (Me3tren is momomethylated on each primary amine

117

18

I

7 Efict of Pressure on inorganic Reactions: introduction and Mechanistic Applications

*

r

R = -CH3, -CH2CI, -CH2CF3 Fig. 1.14. Suggested transition state for the reaction trans[Rh(dmg)z(R)H20] t L + trans-[Rh(drng)2(R)L] H20.

+

group), for which a typical example is given in Fig. 1.15, clearly underline the compact nature of the transition state and the operation of an I, mechanism [53, 551. Increasing the steric hindrance on the tren ligand by introducing six methyl substituents on the terminal amino groups, slows down the substitution reaction of [Cu(Me6tren)H2OI2+ by about five orders of magnitude as compared to the unsubstituted tren complex, which is accompanied by a changeover in mechanism from I, to Id [55, 561. This example is an excellent illustration of how an increase in steric hindrance on the spectator ligand can cause a changeover to a more dissociative mechanism. This does not always have to be the case as an example in the following section will show.

3

8 Mg

Cu(Me3tren)(H20>2’+ py

‘F

Cu(Me3tren)(py>2’+H,C

3

0 >

a

py-Cu(Me3tren)(H20$+

3

-8

.........................

................... ....._

a

.A

5a

-2

I

I

7.3 Thermal-Induced Reactions 119

t v5

AVY

Cm3mo'-1

RCOOH

:I1

.... .. .......... . .. ...... .......................... .....,,,.,...... . ..,.,,....,. .......,....,... . ............,

/2.3

\

8.9

\

\ G -

1.7

Reaction Coordinate

$: Pd(H20)T+ RCOOH

, -

CH3CooH - 1.7 7 cn,cn2coon

+

Pd(H20)300CRt

t

Fig. 1.16. Volume profiles for the reversible reaction [Pd(HzO),]" RCOOH [Pd(HzO),(OOCR)]' H3O'

+

+

1.3.1.2 Square-Planar Complexes

Square-planar d8 complexes of Pt(I1) and Pd(I1) are, in general, accepted to undergo substitution reactions according to an associative mechanism. Recently, activation and reaction volume data were reported for complex-formation reactions of [ P ~ ( H Z O ) ~ with ] ~ +a series of organic acids 1571. A few typical volume profiles for reactions of acids are presented in Fig. 1.16, from which the compact associative nature of the transition state can be seen. A good linear correlation between AV# for the complex-formation reaction and the overall reaction volume AVO was found (see Fig. 1.17), demonstrating that the location of the transition state is controlled by the overall reaction volume. In these cases AV# and AVO are composite quantities resulting from a volume collapse due to bond formation and a volume increase due to a decrease in electrostriction as a result of charge neutralization. Since substitution reactions of square-planar complexes in most cases tend to follow an associative mechanism, a large number of complexes were synthesized in efforts to determine whether a changeover in mechanism is induced. The increase in steric hindrance caused by introducing alkyl substituents on the three N donor atoms of diethylenetriamine (dien) in going from dien to Mesdien to Etsdien, caused a decrease in the aquation rate constant for (Pd(Rsdien)Cl]+ of six orders of magnitude, but not a changeover in mechanism. The reported activation

H30+

20

I

I Ec’t

of Pressure on Inorganic Reactions: Introduction and Mechanistic Applications

Fig. 1.17. Linear correlation reported between AV’ and AVO for the reversible reaction [Pd(H20),]” LH/L +

+

+

[Pd(H20)3(L)]+/*+ H30+/H20.1 - MezSO; 2 - propionic acid; 3 - acetic acid; 4 - malonic acid; 5 - MeCN; 6 - citric acid; 7 - glycolic acid; 8 -water.

volumes remained strongly negative (between -10 and -15 cm3 mol-’) throughout the series studied and support the operation of an associative mechanism, independent of the degree of steric hindrance on the chelate ligand [S]. Two typical volume profiles for such reactions are shown in Fig. 1.18, from which the associative nature of the transition state can clearly be seen [ 51. The introduction of a metal-carbon cr bond could be an alternative way to induce an increase in lability and a possible changeover in mechanism. In the case of NCN donor complexes of Pt(II), where NCN represents ~ , G - C G H ~ ( C H ~ N van M~~)~, Koten and collaborators were able to compare the ligand substitution kinetics in water directly with the data for water exchange on [Pt(NCN)H20]+[58. 591. The AVf values for ligand substitution and water exchange are between -9 and -12 cm3 mol-’, and clearly support the associative character of the reactions. A typical volume profile for a closely related system also involving NCN donor ligands is presented in Fig. 1.19, from which the compact and associative nature of the transition state can be seen [GO]. Irrespective of the entering ligand, the AVf values are all strongly negative and support a five-coordinate transition state for the substitution process. In other complexes the introduction of a C atom as a strong cr-donor ligand, increases the reactivity of Pt(11) by six orders of magnitude compared to [ Pt( H20)4] but does not induce a changeover in mechanism. Surprisingly, there is only one example in the literature where the authors report activation volumes in support of a changeover in the substitution mechanism, i.e. from the usual associative to the unusual dissociative mechanism. The authors point out that at least two strong

’+,

7.3 Thermal-Induced Reactions

i

Reactants

i

Transition State

Products

.......................................................

______ ___ -..,-..-......-.-... J.

AV*= -13.4

V

cm'rnol-

1

1

L = 1,1,7,7-Et4dien

I

1 AV = - 1 1.9

AV*= -14.9

I

j Transition State

Reactants

i

Products

Reaction Coordinate Fig. 1.18. Volume profiles for the reversible reactions [Pd(L)Cl]+

+ H20 * [Pd(L)H20l2+ + CI-.

a-donors are needed to change the reaction mechanism [ 611. Various attempts followed to increase the donor capacity of the non-labile ligand in order to weaken the trans position. 1.3.2 Electron Transfer Reactions

A fundamental understanding of oxidation-reduction reactions is vital to the inorganic chemist in contexts ranging from energy transduction - chemical to electrical and the converse, in technical matters in corrosion processes and metallurgy, redox processes in environmental chemistry and metalloenzymes and metalloproteins involved in electron transfer. Electron-transfer reactions of transition metal complexes are accompanied by a change in the oxidation state of the metal

I

22

I

7 Effect of Pressure on Inorganic Reactions: Introduction and Mechanistic Applications

PtiN,C%)N3+ HzO

" I

.........................

N3-Pt(N,WH*O)

I

Reactants

.......................

I

Transition State

Products

Fig. 1.19. Volume profile for the reversible reaction [Pt(N,N-C)(H20)]' N, + [Pt(N,N-C)N3] H2O.

+

+

atom and the overall charge on the complex ion. This can cause both intrinsic and solvational volume changes, such that it is reasonable to expect that electrontransfer reactions should exhibit characteristic AVf values. Here we will confine ourselves (see Sect. 1.3.2.2) to electron self-exchange reactions for which considerable success in correlating experimental AV# values with theoretical calculations of the same parameters has been achieved [G2], and to some recent examples of bioinorganic electron transfer reactions. 1.3.2.1

Self-Exchange Reactions

Self-exchange reactions are of fundamental importance in the understanding of electron transfer processes, in a similar way mechanistic studies of solvent exchange reactions form the basis for understanding ligand substitution reactions. These are symmetrical processes for which the reaction volume AV is zero. The progress made recently has been concisely and thoroughly summarized [ G 2 ] . A detailed account on the effect of pressure on self-exchange reactions is given in Chapter 5. 1.3.2.2

Non-Symmetrical Reactions

For the mechanistic interpretation of activation volume data for non-symmetrical electron-transfer reactions, it is essential to have information on the overall volume change that can occur during such a process. This can be calculated from the partial molar volumes of reactant and product species when these are available. or can be determined from density measurements. Efforts have in recent years have focused on the electrochemical determination of reaction volumes from the pressure dependence of the redox potential. Tregloan and coworkers [42, 431 have demonstrated how such techniques can be employed to provide estimates of the intrinsic and solvational volume changes associated with electron-transfer reactions of transition metal complexes.

1.3 Thermal-Induced Reactions

I

23

We will restrict the discussion of non-symmetrical electron-transfer reactions to a number of systems that are ofbiological interest. In collaboration with other groups, we have investigated intermolecular and intramolecular electron-transfer reactions between various Ru complexes and cytochrome c. A variety of high pressure experimental techniques, including stopped-flow, flash-photolysis, pulse-radiolysis and voltammetry, was employed in these investigations. As the following presentation will show, remarkably good agreement was found between the volume data obtained with the aid of these different techniques, which clearly demonstrates the complementarity of these methods for the study of electron-transfer processes. Application of pulse-radiolysis techniques (see Sect. 1.4 for more details on the experimental set-up) revealed that the following intramolecular (Eqs 1.6 and 1.7) and intermolecular (Eq. 1.8) electron-transfer reactions, where cyt c represents cytochrome c, all exhibit a significant acceleration with increasing pressure. The reported volumes of activation are -17.7 f 0.9: -18.3 f 0.7, and -15.6 f 0.6 cm3 mol-', respectively, and clearly demonstrate a significant volume collapse upon reaching the transition state [ 631.

(NH3)~Ru~~-(His33)cyt c"' i (NH3)5Ru"1-(His33)cytc'l (NH3)5 RIA"-( His39)cyt clll

i

(NH3)5 Ru"'-( His39)cyt cI1

(1.6) (1.7)

At this stage there was uncertainty regarding the origin of the negative volumes of activation and their interpretation; the overall reaction volumes which would be of assistance in the analysis were not available. There was, however, data [42, 431 that suggested that the oxidation of [ Ru(N H ~ ) (2+ I ] to [ Ru(NH3)6]3+ is accompanied by a volume decrease of ca. 30 cm3 mol-l, which would mean that the activation volumes quoted above could arise mainly from volume changes associated with the oxidation of the ruthenium redox partner. In order to obtain further information on the magnitude of the overall reaction volume and the location of the transition state along the reaction coordinate, a series of intermolecular electron-transfer reactions of cyt c with pentaamminemthenium complexes was studied, where the sixth ligand on the ruthenium complex was selected in such a way that the overall driving force was low enough so that the reaction kinetics could be studied in both directions [64, 651. The substituents selected were isonicotinamide (isn), 4-ethylpyridine (etpy), pyridine (py), and 3,s-lutidine (lut).The overall reaction can be formulated as shown in Eq. (1.9). [RU"'(NH~)SL]~' + cyt c"

+ [RuT1(NH~)5L]'+ + cyt c"'

In all cases, the forward reaction was significantly decelerated by pressure, whereas the reverse reaction was significantly accelerated by presssure. The absolute values of the volumes of activation for the forward and reverse processes were indeed very similar, demonstrating that a similar rearrangement, in volume terms, occurs in order to reach the transition state. In addition, the overall reaction volume for these

24

I

1 Effect of Pressure on fnorganic Reactions: fntroduction and Mechanistic Applications

Reactants

Transition State

Products

Reaction coordinate Volume profile for the reversible electron-transfer reaction [ R u " ' ( N H ~ ) ~ L-t ] ~cyt + c ' l + [ R U " ( N H ~ ) ~ L ] ~cyt + clll.

Fig. 1.20.

+

systems could be determined spectrophotometrically by recording the spectrum of an equilibrium mixture as a function of pressure, and electrochemically by recording cyclic and differential pulse voltammograms as a function of pressure [GG] . A comparison of the AV data demonstrated the generally good agreement between the values obtained from the difference in the volumes of activation for the forward and reverse reactions, and those obtained thermodynamically. Furthermore, the values also clearly demonstrated that JAVz 1 N O.S)AVJ,i.e. the transition state lies approximately halfway between the reactant and product states on a volume basis independent of the direction of electron transfer. The typical volume profile in Fig. 1.20 presents an example of the overall picture, from which the location of the transition state can clearly be seen. At this point it is important to ask the question, what is the source of these volume changes? We argued that the major volume change arises from changes on the redox partner and not on cytochrome c itself. This was suggested by the fact that the change in partial molar volume associated with the oxidation of the Ru(11) complexes, as obtained from electrochemical and density measurements, almost fully accounted for the observed overall reaction volume. Thus the reduction of cytochrome c can only make a minor contribution towards the overall volume change. An investigation of the electrochemistry of cytochrome c as a function of pressure, using cyclic and differential pulse voltammetric techniques [GG], revealed a reaction volume of -14.0 0.5 cm3 mol-' for the reaction shown in Eq. (1.10). cyt ClIl

+ Ag(s) + C1

+

Cyt cii

+ AgCl(s)

(1.10)

A correction for the contribution from the reference electrode can be made on the basis of the data published by Tregloan et al. [43], and a series of measurements of

7.3 Thermal-lnduced Reactions

the potential of the Ag/AgCl(KC1 sat'd) electrode relative to the Ag/Ag+ electrode as a function of pressure. The contribution of the reference electrode turned out to be --9.0 0.6 cm3 mol I , from which it then followed that the reduction of cytochrome c"' is accompanied by a volume decrease of 5.0 0.8 cm3 mol-*. Thus we conclude that the observed activation and reaction volumes mainly arise from volume changes on the Ru complexes, which in turn will largely be associated with changes in electrostriction in the case ofthe ammine complexes. The oxidation of the Ru( 11) ammine cornplcxes will be accompanied by a large increase in electrostriction and almost no change in the metal-ligand bond length, whereas in the case of Co complexes a significant contribution from intrinsic volume changes associated with the oxidation of Co( 11) will partially account for the observed effects [43]. Thc available results demonstrate readily the complementarity of the kinetic and thermodynamic data obtained from stopped-flow, UV-Vis, electrochemical and density measurements, and yield a mutually consistent set of trends allowing further interpretation of the data. The overall reaction volumes determined in four different ways are surprisingly similar and underline the validity of the different methods employed. Thc volume profile in Fig. 1.20 illustrates the symmetric nature of the intrinsic and solvational reorganization in order to reach the transition state of the electron-transfer process. In these systems the volume profile is controlled by effects on the redox partner of cytochrome c, but this does not necessarily always have to be the case. The location of the transition state on a volume basis is informative regarding the "early" or "late" nature of the transition state, and therefore details of the actual electron-transfer route followed. Recent investigations on a series of intramolecular electron transfer reactions, closely related to the series of intermolecular reactions described above, revealed non-symmetrical volume profiles [ 671. (1.11)

Reactions of the type shown in Eq. ( L l l ) , where L = isonicotinamide, %ethylpyridine, 3,5-lutidine, or pyridine, all exhibited volumes of activation for the forward reaction of between $3 and +7 cm3 mol-', compared to overall reaction volumes of between +19 and +26 cm3 mol-'. This indicates that electron transfer from Fe to Ru is characterized by an "early" transition state in terms of volume changes along the reaction coordinate (see Fig. 1.21). The overall volume changes could be accounted for in terms of electrostriction effects centered around the ammine ligands on the ruthenium center. A number of possible explanations in terms of the effect of pressure on electronic and nuclear factors were offered to account for the asymmetrical nature of the volume profile [ 671. 1.3.3 Actvation of Small Molecules

Mechanistic understanding of the binding of diatomic molecules such as 02,CO and NO to ferrous hemes and hemoproteins is essential as part of a description of

I

25

26

I

7 Efect of Pressure on inorganic Reactions: introduction and Mechanistic Applications

-0

7-

-15.9

ki

5 E

b23.3

10-

[(L)RuIII/II -c). c 111/11

1f

Reaction coordinate Fig. 1.21.

Volume profile for the reversible electron-transfer reaction ( N H3)4(L) Ru "'-(His33)-Cyt cil F! (NH3)4(L) Ru "-(His33)-Cyt clll

the overall transport of such molecules in biological systems and in terms of their activation. In order to obtain further mechanistic insight, volume changes for the binding of 0 2 and CO to myoglobin, and for the binding of 0 2 to hemerythrin [ 68, 691 were determined and volume profiles were constructed. These are shown in Figs 1.22 and 1.23, respectively. For the oxygenation of the iron proteins, positive activation volumes of +13.3 1.1 and +5.2 f 0.5 cm3 mol-' for hemerythrin and myoglobin, respectively, were found. Since bond formation processes are usually characterized by a decrease in volume [8, 13, 171, the positive values were assigned to desolvation of oxygen during its entrance into the protein and to conformational changes on the protein itself. The release of oxygen is characterized by very positive activation volumes, f52 f 1 and f23.2 i 1.8 cm3 mol-', for hemerythrin and myoglobin, respectively, such that the overall reaction volume for the oxygenation process is strongly negative in both cases. The activation and reaction volumes for mononuclear myoglobin are about half of those found for binuclear hemerythrin. In the hemerythrin system, two Fe(I1) centers are oxidized to Fe(II1) during which dioxygen is reduced and bound as hydroperoxide to one Fe(II1) center. The significant volume decrease that occurs following the formation of the transition state can be ascribed to the oxidation of the Fe(I1) centers and the reduction of 0 2 to O:-. The fact that the overall volume collapse is almost double that observed for the oxygenation of myoglobin may indicate similar structural features in oxyhemerythrin and oxymyoglobin. This suggests that a description of the bonding mode as

1.3 Thermal-Induced Reactions 127

1

-lo ~

[Mb-CO]

-5.9

1

MbCO

+4.1

MbO2

Reactants

Products

Transition State

Reaction coordinate Fig. 1.22.

Volume profiles for the binding o f 0 2 and CO to myoglobin.

Fe"'-O; or Fe1"-02H ( H from histidine E7) instead of Fe"-02, is more appropriate for oxymyoglobin. A fundamental question regarding the activation of dioxygen by transition metal complexes is whether the process is controlled kinetically by ligand substitution or by electron transfer. A model system that involved the binding of dioxygen to a macrocyclic hexamethylcydam Co(11) complex to form the corresponding Co(111) superoxo species is shown in Eq. (1.12), thus modeling the first redox activation step of dioxygen; the reaction was studied in detail [70].

+0 2

[CO"(L)(HZO)~]"

[CO"'(L)(H,O)(O;)]~'

+ H2O

(L = Mebcyclam) r

[ Hr-Op]'

-

9

-z -39 Q m

a!

d Reactants Fig. 1.23.

Transition State

Volume profile for the binding o f 0

2

Product

to hemerythrin.

(1.12)

28

I

7 €@ct

-7 [(L)co. l;29

of Pressure on Inorganic Reactions: Introduction and Mechanistic A p p h t i o n s I

0

I

(L)C02+ + 0

. ....,.. . ,. ....

-

-10-

-

-15-

-

-20 ,. ....,.......................... .......................,...........,,. ............................

4 5-

Fig. 1.24.

I

I

(L)cooo2+

-

Volume profile for the reversible binding of dioxygen t o [Co"(Me~cyclarn)(H~O)~]*'

The overall reaction involves ligand substitution and electron transfer, the question being which occurs first. From the pressure dependence of the overall equilibrium constant a reaction volume of -22 cm3 mol-' was determined, which demonstrated that the displacement of a water molecule on the Co(I1) complex by dioxygen is accompanied by a significant volume collapse, probably mainly due to the oxidation of Co(I1) to Co(II1) during the overall reaction. The kinetics of the reaction could be studied by flash-photolysis,since the dioxygen complex can be photodissociated by irradiation into the charge-transfer (CT) band, and the subsequent reequilibration could be followed on the microsecond time scale. From the effect of pressure on the binding and release of dioxygen, the activation volumes for both processes could be determined. A combination of these activation volumes resulted in a value for the reaction volume that is in excellent agreement with that determined directly from the equilibrium measurements as a function of pressure. The volume profile for this reaction is given in Fig. 1.24.The small volume of activation associated with the binding of dioxygen is clear evidence for a rate-limiting interchange of ligands, dioxygen for water, which is followed by an intramolecular electron-transfer reaction between Co(I1) and O2 to form Co"'-O;, the superoxo species. It is the latter process that accounts for the large volume reduction en route to the reaction products. Thus during flash-photolysis, electron transfer in the reverse direction occurs due to irradiation into the CT band, which is followed by the rapid release of dioxygen. CO has been used in many studies as a model for the binding step component of the activation of dioxygen since CO does not undergo activation in the systems studied. The absence of subsequent electron transfer reactions simplifies the kinetic analysis and provides more mechanistic insight into the actual binding process. A typical example concerns the comparative binding of 0 2 and CO to deoxymyoglobin [68].The volume profile for the binding of 0 2 , as described above, is

1.3 Thermal-Induced Reactions

I

29

characterized by a substantial increase in volume in going from the reactant to the transition state, followed by a significant volume reduction on going to the product state. The volume profile for the binding of CO (see Fig. 1.22), however, shows a considerable volume decrease on going from the reactant to the transition state, which was ascribed to rate-determining bond formation. The reverse bond cleavage reaction is accompanied by a volume decrease, which may be related to the different bonding mode of CO compared with 0 2 . This difference in bonding mode must also account for the much smaller absolute reaction volume observed in this case. In another investigation of a metal center-diatomic molecule interaction, the binding of CO to lacunar Fe(I1) complexes was studied in detail as a function of temperature and pressure [71, 721. In such a system the high spin Fe(II) center is five-coordinate and has a vacant pocket available for the binding of CO. These systems can, therefore, be considered as ideal for the modeling of biological processes. A detailed kinetic analysis of the “on” and “off” reactions, as well as a thermodynamic analysis of the overall equilibrium reaction, enabled the construction of the energy and volume profiles for the binding of CO to [ Fe“(PhBzXy)](PF6)2;the volume profile is shown in Fig. 1.25 [ 7 3 ] .The free energy profile demonstrates the favorable thermodynamic driving force for the overall reaction, as well as the relatively low activation energy for the binding process. The entropy profile demonstrates the high degree of order in the transition state on the binding of CO.

r

F&L)

+ co

REACTION COORDINATE

Fig. 1.25.

Volume profile for the reversible binding o f CO t o a lacunar Fe(ll) complex.

30

I

7 Efict of Pressure on lnorganic Reactions: introduction and Mechanistic Applications

The large volume collapse associated with the forward reaction is very close to the partial molar volume of CO, which suggests that CO is engulfed and completely “disappears” within the ligand pocket of the complex in the transition state during partial Fe-CO bond formation. It is also known [71] that Fe”-CO bond formation is accompanied by a high-spin to low-spin conversion of the Fe(11) center. In forming the six-coordinate, low-spin Fe(11) complex, the metal moves into the plane of the equatorial nitrogen donors. Thus following the formation of the transition state for the binding of CO, there is a high-spin to low-spin change during which bond formation is completed and the metal center moves into the ligand plane. These processes account for the subsequent volume decrease accompanying the formation of the product from the transition state. The overall reaction volume of -48 cm3 mol therefore consists of a volume decrease of ca. -37 cm3 rno1-l associated with the apparent disappearance of CO into the ligand cavity, and ca. -12 cm3 mol-’ for the high-spin to low-spin transition. In a recent study the volume profile for the reversible binding of NO to metmyoglobin was determined using a combination of stopped-flow and flash-photolysis techniques [73]. The reported volume profile (see Chapter 6, Fig. 6.18) demonstrates the dissociative nature of the binding process. By way of comparison, the volume of activation found for the “on” reaction is significantly smaller than that reported for the corresponding reaction with the Fe(111) porphyrin complex [74], suggesting that the formation of the five-coordinate Fe(111) intermediate (within the concept of a dissociative mechanism) is accompanied by structural changes on the protein that result in an additional volume increase. In the case of the porphyrin systems it has recently been shown that the binding of NO is controlled by the rate and mechanism of water exchange on such complexes [75]. 1.3.4

Addition and Elimination Reactions

In principle, addition or elimination reactions involve significant bond formation or bond breakage in the transition state, such that an acceleration or deceleration is expected at elevated pressure, respectively. For instance, [2 21 cyclo-addition reactions of the type shown in Scheme 1.1were significantly accelerated by pressure, and there was almost no dependence on the polarity of the solvent [76].

+

X = Cr(CO)5, R1= Me, R2 = M e X = W(CO)5, R1 =Me, Rz = M e X = W(CO)5, R1= Ph, R2 = Et Scheme 1.1

7.3 Thermal-lnduced Reactions

1 cm3 mol-' and the solvent independence of the The average AV# of - l G process suggested that the reaction follows a nonpolar concerted, synchronous onestep mechanism. The observed pressure acceleration is very similar to that found for the insertion of dipropylcyanamide and 1-(diethy1amino)propene into the metalcarbene bond of pentacarbonyl(methoxypheny1-carbene)chromium and -tungsten (shown in Scheme 1.2) for which AVf varies between -17 and -25 cm3 mol-I [ 771. Addition reactions of a,p-unsaturated Fischer carbene complexes, as shown in Scheme 1.3, all exhibit AVf values of between -15 and -17 cm3 mol-' in acetonitrile. On decreasing the solvent polarity, AV# becomes significantly more negative and exhibits a good correlation with the solvent parameter qp (i.e. the pressure derivative of q, the polarizability of the solvent) [78].

,OMe (CO)sM=C\ Ph

,OMe (CO)sM=C\ Ph

+

PrzN-CEN

+ E$N-C=C-Me

-

N=C\ (CO)5M=C: NPrz

Me\

,c=c,

(CO)sM=C,

,OMe Ph

,OMe Ph

NEt2

M=Cr,W Scheme 1.2

Cr Ph MoPh W Ph W Me

Et Et Et Me

Scheme 1.3

It was concluded that the addition of pyrrolidine follows a two-step process with a polar transition state leading to a zwitterionic intermediate. The addition of a series of p-substituted anilines to a Fischer carbene complex as shown in Scheme

I

31

32

I

7 Effect of Pressure on Inorganic Reactions: Introduction and Mechanistic Applications

’

1.4 is characterized by AV# values between -21 and -27 cm3 mol [79], i.e. significantly more negative than for the reaction with pyrrolidine mentioned above. The second order rate constant for the addition reactions exhibitcd an excellent correlation with the basicity of the selected aniline. The observed trcnd in the activation volumes could be correlated with an “early” or “late” transition state for the fast and slow addition reactions, respectively.

6MeCN

X

X = CN, CH,CO, CI, F, H, CH,, CH,O

Q x

Scheme 1.4

We now turn to oxidative addition and reductive elimination reactions. Such reactions not only undergo significant bond formation and bond breakage, respectively, in reaching the transition state, but also undergo a change in the oxidation state and coordination number of the metal complex. These effects are expected to cause large volume changes, such that these reactions should exhibit a high pressure sensitivity. One interesting example involves the addition of Me1 to the Pd(11) complex, [ PdMez(bpy)], to form [Pd(I)Me3(bpy)],which is accompanied by a AV# value of -11.9 cm3 mol-’ 1801. A similar value was reported for the corresponding reaction with the Pt(I1) complex [81]. This value confirms the operation of an S N ~ mechanism. The reductive elimination of C ~ H from G the Pd(IV) complex forming [Pd(I)Me(bpy)] as product, yielded a AV# value of f17.5 cm3 mol-*, which is in line with the formal change in oxidation state and bond breakage. With the assumption that the transition states for the oxidative addition and reductive elimination processes have a similar partial molar volume, then an overall reaction volume of 29 cm3 mol-’ can be calculated for such reactions (see volume profile in Fig. 1.26).

1.4

Photo- and Radiation-Induced Reactions

The effect of pressure on photochemical and photo-induced reactions has been investigated in detail for many systems, and a number of reviews have appeared on this subject recently [82-84). This topic is treated in more detail in Chapter 6, and readers are referred to that chapter for more information. Pulse-radiolysis is a well-established technique for investigating the interaction of free radicals with metal complexes. For instance, with this technique metal

1.4 Photo- and Radiation-Induced Reactions

I

33

I

Reactants

state

Intermediate

Products

State

Fig. 1.26. Volume profile for the combined oxidative addition and reductive elimination reaction [PdMez(bpy)] Me1 --t [Pd(l)Me3(bpy)l [Pd(l)Me(bpy)l C2H6.

+

+

+

complexes can be prepared in unique oxidation states either as intermediates or as stable species. In addition, the formation and stability of metal-carbon c bonds through reaction with organic radicals can be studied. In many systems little is known about the details of the reaction mechanism, such that the combination of high pressure and pulse-radiolysis techniques seemed one avenue of approach to increase our understanding of such processes. Before these investigations were possible, it was essential to modify the high pressure cell shown in Fig. 1.3 in such a way that the electron beam (originating from a 5-MeV linear accelerator or a 2MeV van de Graaff accelerator) could reach the sample within the pill-box sample cell. This was not possible with the thick sapphire windows used in the standard cell. One of the windows of a four-window cell was therefore replaced by a honeycomb-type metal grid, the details of which are shown in Fig. 1.27. The ultimate wall thickness of less than 1 m m enabled the efficient entry of the electron beam into the cell. Very reproducible results could be obtained independent of the applied pressure. A review on the work performed using this technique has appeared [SS]. Two examples will be given to demonstrate the application of this combination of techniques. A typical example of a volume profile for the formation and cleavage of a Co-CHj bond is reported in Fig. 1.28 for the reaction shown in Eq. (1.13) [86].

+

+

[ C ~ " ( n t a ) ( H ~ O ) ~ 'CH3 l+ [Co1"(nta)(CH3)(H2O)]- H2O

(1.13)

34

I

I Effect of Pressure on Inorganic Reactions: Introduction and Mechanistic Applications

-r 4mm 1

t

*r

16 m m

1

L Fig. 1.27.

+17

m

4

Details o f the special window used for high pressure pulse-radiolysis studies.

The volume profile indicates an increase in partial molar volume in proceeding to the transition state, which is interpreted in terms of an Id substitution controlled binding o f the methyl radical to the Co(I1) complex. The large volume collapse that occurs subsequent to the formation of the transition state is ascribed to metalcarbon bond formation which is accompanied by oxidation of Co(I1) to Co( 111) and accompanied by a large volume contraction [86]. The reaction o f aquated Cr(I1) with 10 different aliphatic radicals R, showed a decrease in rate with increasing pressure, and resulting volumes of activation were between $3.4 and +6.3 cm3 mo1-l [87]. These data could be interpreted in terms of a water-exchange-controlled formation o f the Cr-R bond, from which it followed that water exchange on Cr(I1) must proceed according to an I d mechanism. The

-z

7

+lo-

[L(H,O)cQ:.,,OH,]*

L = nta

15-

d

?l0

3

>

0 0

LCo(H,O); + ‘R --

1.-*.-T-.-

2

-E

0

-10-

calc : -12.6 54.5 exp : -16.4 f 1.6

0

.->

-2

+18.6 f 2.0

-5-

m

3a

I

/ I

... ................ ..........

LCo(H,O)R- + H 2 0

4-

m

-15-15

Reactants

I

Transition State

I

Reaction coordinate Fig. 1.28. Volume profile for the reaction o f methyl radicals with the nitrilotriacetate complex o f Co(ll), R = CH3.

Products

1.5 Concluding Remarks

+lo-

[ (q+ c'':...R .43"2]*

+5Cr(H.$l):+ 0

'R

..............

+15.1 -5-

Cr(%O)5 R2'+ -10-

30

............................. II

Reaction coordinate Fig. 1.29. Volume profile for the reaction of an aliphatic radical with [Cr(H20),]'*, R = C(CH3)20H.

volume profile in Fig. 1.29 demonstrates this point; the large volume collapse following the transition state was assigned to Cr-R bond formation accompanied by the conversion of Cr"-R to Cr"'-R-. In general it has been found in reactions of metal complexes with free radicals, based on the observed pressure effects: that the radicals can be treated as normal nucleophiles in ligand substitution processes; the latter are often controlled by solvent exchange on the metal complex [88].

1.5 Concluding Remarks

It was the objective of this contribution to demonstrate how the application of high pressure thermodynamic and kinetic techniques can contribute to the elucidation of inorganic reaction mechanisms. The analysis of a chemical process in terms of volume changes along the reaction coordinate can help us to visualize the nature and structure of the transition state in terms of intrinsic and solvational changes in partial molar volume. In many cases the insight gained with these techniques is unique and has added a further dimension to the study of reaction mechanisms in solution. The construction of volume profiles provides a helpful visualization of details of the molecular organization and solvation of the transition state. There are many cases where it is experimentally impossible to acquire sufficient data to construct a volume profile; examples of such limitations are where subsequent reactions occur or when the reaction is irreversible as found for many electron transfer and photo- or irradiation-induced reactions. Nevertheless, the volume of activation for such processes can still be employed very successfully to obtain information on the nature of the transition state.

I

35

36

I

7 Efect of Pressure on fnorganic Reactions: fntroduction and Mechanistic Applications

The fact that the rate-determining step of a particular process exhibits a characteristic pressure dependence, creates the possibility of tuning the reactivity of particular systems via the application of moderate pressures. This can lead to the selective synthesis of particular reaction products in cases where product distribution proves to be presssure dependent, or to the optimization of design of industrial chemical processes. In some of the simplest cases (namely, solvent exchange and self-exchange reactions), the experimental data could be supported by theoretical calculations. Significant developments are expected to occur in this area, such that the theoretical optimization of transition state structures will become standard practice in mechanistic studies. Here again volume of activation data will play a crucial role, since they present an experimental measure of the intrinsic and solvational volume changes in the transition state and form a basis for comparison with theoretical predictions. It will be an ideal situation when volume profiles can be constructed for more complex reaction sequences, for instance for catalytic cycles in enzymatic processes. This will, as in the case of more simple reactions, enhance our understanding of complex chemical processes and improve our ability to tune them. The correlation of kinetic data with activation volume data for a series of closely related reactions in terms of an “early” or “late” transition state. has been reported in a number of cases. The ultimate goal will be to correlate energy and volume profiles for series of related reactions, where a low activation barrier (fast reaction) will correspond to an “early” transition state, and a high activation barrier (slow reaction) will correspond to a “late” transition state. A schematic presentation of such a correlation between the location of “early” and “late” transition states on energy and volume profiles is shown in Fig. 1.30. A three-dimensional presentation of free energy and partial molar volume changes along the reaction coordinate should represent the ultimate way to combine energy and volume profile information. In our opinion as indicated in the Introduction, when pursuing mechanistic studies, one should investigate as many chemical and physical variables as possible in order to obtain as much information as possible on the nature of the underlying reaction mechanism. Only then can the suggested mechanism come close to the “real” mechanism, which is a goal set in many mechanistic studies, but only accomplished in few. If we can in this way contribute to a better understanding of the mechanism of chemical reactions in solution, then we have fulfilled our educational and research objectives, and have made a modest contribution to existing knowledge.

Acknowledgments

The authors gratefully acknowledge financial support from the Deutsche Forschungsgemeinschaft, Fonds der Chemischen Industrie, Volkswagen Stiftung, MaxBuchner Forschungsstiftung, NATO Scientific Affairs Division and the German-

References I 3 7

7

Schematic comparison of energy and volume profiles: 1 - “early” transition state; 2 - “late” transition state. Fig. 1.30.

Israeli Foundation. The very stimulating collaboration with numerous graduate and post-graduate students, post-doctoral associates, visiting scientists and various research groups all over the world, is greatly appreciated and respected.

References 1 2 3 4

5

6

W. j. LE NOBLE, Chem. unserer Zeit 1983, 17, 152. F.-G. KLLRNER,Chem. unserer Zeit 1989, 23, 53. D. R. STRANKS, Pure Appl. Chem. 1974, 38, 303. H. KELM (Ed.), High Pressure Chemistry, Reidel, Dordrecht, 1978. R. VAN ELDIK(Ed.), Inorganic High Pressure Chemistry, Kinetics and Mechanism, Elsevier, Amsterdam, 1986. R. VAN ELDIK,J. JONAS (Eds), High Pressure Chemistry and Biochemistry, Reidel, Dordrecht, 1987.

M. K o r o w s ~ rR. , VAN ELDIK, Coord. Chem. Rev. 1989, 93, 19. 8 R. VAN ELDIK,T. ASANO,W. J. LE NOBLE, Chem. Reu. 1989, 89, 541. 9 A. E. MERBACH, J. W. Akitt, in N M R Basic Principles and Progress, vol 24, Springer Verlag, Berlin, 1990, 189. 10 (a) R. VAN ELDIK,A. E. MERBACH, Comments Inorg. Chem. 1992, 22, 341; (b) S. F. LINCOLN, A. E. MERBACH, Adu. Inorg. Chem. 1995, 42, 1; ( c ) L. HELM,A. E. MERBACH, Coord. Chem. Rev. 1999, 287, 151; (d) G. STOCHEL, R. V A N ELDIK,Coord. Chem. ReG. 1999, 187, 329; (e) T. W. SWADDLE, P. A. 7

38

I

7 Effect of Pressure on inorganic Reactions: Introduction and Mechanistic Appkations

T R E G I O A NCoord. , Chem. Rev. 1999, 187, 255. 11 R. WINIER,J. JONAS, (Eds), High Pressure Chemistry, Biochemistry and Materials Science, Kluwer, Dordrecht. 1993. 12 C. D. HUBBARD, R. VAN ELDIK. Instrum. Sci. Technol. 1995, 22. 1. 13 R. VAN ELDIK a n d C. D. HUBBARD (Eds), Chapters 2. 3 a n d 4 in Chemistry under Extreme or Non-Classical Conditions, Wiley/Spektmm, New York/I Icidelberg, 1997. 14 W. R. HOLZAPFEL, N. S. ISAACS,(Eds) High Pressure Techniques in Chemistry and Physics. Oxford University Press, Oxford, 1997. 15 C. A. ECKERT,Annu. Rev. Phys. Chem. 1972, 23, 239. 16 T. ASANO,W. J. LE NOBLE,Chem. Rev. 1978, 78, 407. 17 A. DRLIACA, c. D. HUBBARD, R. VAN ELDIK,T. ASANO,M. V. BASILEVSKY, W. J. LE NOBLE,Chem. Rev. 1998, 98, 2167. 18 M. SPITZER,F. GARTIG, R. VAN ELDIK, Rev. Sci. Instrum. 1988, 59, 2092. 19 J. F. WISHART,R. VAN ELDIK,Rev. Sci. Instrum. 1992, 63, 3224. 20 Q. JI, E. M. EYRING,R. VAN ELDIK,K. BALREDDY,S. R. GOATES,M. L. LEE, Rev. Sci. Instrum. 1995, 66, 222. 21 (a) N. TAKISAWA, M. SASAKI,F. AMITA, I. OSUGI,Chem. Lett. 1979, 671; (b) M. SASAKI,F. AMITA,J. OSUCI:Ret.. Sci. Instrum. 1979, 50, 1073. 22 K. HEREMANS, J. SNAUWAERT, J. RIJKENBERG, Reu. Sci. Instrum. 1980, 51, 806. 23 R. VAN ELDIK,D. A. PALMER,R. SCHMIIYI.,H. KELM,Inorg. Chim. &a 1981, 50, 131. 24 (a) S. FUNAHASHI, K. ISHIHARA,M. TANAKA. Inorg. Chem. 1981, 20, 51; (b) K. ISIIIHARI, s. FUNAHASHI,M. TANAKA,Reii. Sci. Instrum. 1982, 53, 1231. 25 P. J . NICHOLS, Y. DUCOMMUN, A. E. MERBACH,Inorg. Chem. 1983, 22, 3993. 26 C. BAI.NY,J. L. SALDANA, N. DAHAN, Anal. Biochem. 1984, 139, 178. 27 R. VAN ELDIK,w. GAEDE,s. WIELAND,

28

29

30

31

32

33

34

35

36

37 38

39

40 41

42

43

J. KRAFT,M. SPITZER,D. A. PALMER, Rev. Sci. Instrum. 1993, 64, 1355. P. BUGNON,G. LAURENCZY, Y. DUCOMMUN, P.-Y. SAUVAGEAT, A. E. MERBACH, R. ITH, R. TSCHANZ,M. R. BERGBAUER, E. GRELL, DELUDDA, Anal. Chem. 1996, 68. 3045. J. JONAS,Ad?. Magn. Reson. 1973, 6, 73. J. JONAS,in High Pressurc Chemistry, ( H . KELM,Ed.,) Reidel, Dordrecht, 1978. 65. J. JONAS,in High Pressure Chemistry and Biochemistry, (R. VAN ELDIK.J. JONAS,Eds), Reidel, Dordrecht, 1987, 193. A. E. MERBACH, in High Pressure Chemistq and Biochemistry, (R. VAN ELDIK,J. JONAS,Eds). Reidel, Dordrecht, 1987, 311. T. W. SWADDLE; i n High Pressure Inorganic Chemistry, Kinetics and Mechanism, (R. VAN ELDIK,Ed.), Elsevier, Amsterdam, 1986, 273. U. FREY,D. H. POWELL, A. E. MERBACH, in Dynamics of Solutions and Fluid Mixtures by NMR, (J.-J. DELPUECH, Ed.), Wiley, Chichester, 1995, 263. A. ZAHL,A. NEUBRAND, S. AYGEN,R. VAN ELDIK,Rea Sci. Instrum. 1994, 65. 882. H.-D. LUDEMANN, Polish]. Chem. 1996, 70, 387. M. M. HOFFMANN, M. S. COXRADI. Rea Sci. Instrum. 1997, 68, 159. A. CUSANELLI, L. NICUIA-DADCI. U. FREY,A. E. MERBACH,Inorg. Chem. 1997, 36, 2211. E. F. CALDIN,M. W. GRANT,B. B. /. Chem. Soc., Faraday HASINOFF, Trans. 11972, 68, 2247. M. W. GRANT,]. Chem. Soc., Faraday Trans. 11973, 69, 560. (a) H. DOINE,T. W. WHITCOMBE, T. W. SWADDLE, Can.]. Chem. 1992, 70, 81; (b) Y. Fu, T. W. SWADDLE, Chem. Commun. 1996, 1171. J , 1. SACHINIDIS, R. D. SHALDERS, P. A. TREGLOAN, /. Electroanal. Chem. Interfacial Electrochem. 1992, 327, 219. (a) J. 1. SACHINIDIS,R. D. SHALDERS, P. A. TREGLOAN, Inorg. Chem. 1994.

References 1 3 9

44 45 46 47

48 49

33, 6180; (b) J . I. SACHINIDIS, R. D. P. A. TREGLOAN, Inorg. SHALDERS, Chem. 1996, 35, 2497. J , BURGESS, Metal Ions in Solution, Ellis Horwood, Chichester, 1978. T. ITOH, Y. KITAMURA, K. YOSHITANI, Inorg. Chem. 1988, 27, 996. Y. DUCOMMUN, G . LAURENCZY.A. E. Inorg. Chem. 1988, 27, 1148. MERBACH, M. HARTMANN, T. CLARK.R. V A N ELDIK,J. Am. Chem. Soc. 1997, 119, 5867. K. SCHNEIDER, R. VAN ELDIK. Organometallics 1990, 9, 1235. G . STOCHEL, R. VAN ELDIK,Inorg. Chem. 1990, 29, 2075.

M. MEIER,R. VAN ELDIK,Inorg. Chem. 1993, 32, 2653. 51 C. DUCKER-BENFER, R. DREOS,R. VAN ELDIK,Angm. Chem. 1995, 107, 2414. 52 D. H. POWELL, P. FURRER,l'.-T. PITTET, A. E. MERBACH,]. Phys. Chem. 1995, 99, 16622. 53 D. H. POWELL, A. E. MERBACH,I. FABIAN,S. SCHINDLER, R. VAN ELDIK, Inorg. Chem. 1994, 33,4468. 54 A. NEUBRAND, F. THALER,C. D. HUBBARD, A. ZAHL,R. VAN ELDIK, J . Chem. Soc., Dalton Trans., in press. 55 F. THALER,C. D. HUBBARD, F. W. HEINEMANN, R. VAN ELDIK,S. SCHINDLER, 1. FABIAN,A. M. DITTLERKLINGEMANN, F. E. HAHN,C. ORVIG, Inorg. Chem. 1998, 37, 4022. 56 S. F. LINCOLN,A. M. HOUNSLOW, D. L. PISANIELLO, B. G. DODDRIDGE, J. H. COATES,A. E. MERBACH, D. ZBINDEN,Inorg. Chem. 1984, 23, 1090. 57 T. S H I , L. I. ELDING,Inorg. Chem. 1997, 36, 528. 58 U. FREY,D. M. GROVE,G. VAN KOTEN, Inorg. Chim. Ada 1997, 269, 322. 59 M. SCHMULLING, D. M. GROVE,G. VAN KOTEN,R. VAN ELDIK,N. A. L. SPEK,Organometallics VELDMAN, 1996, 15, 1384. 60 M. SCHMULLING, A. D. RYABOV,R. VAN ELDIK,]. Chem. Soc., Dalton Trans. 1994, 1257. 61 U. FREY,L. H E L M A. , E. MERBACH, R. ROMEO,J. Am. Chem. SOC.1989, 111,

62

T. W. SWADDLE, Can. J. Chem. 1996,

63

74, 631. J . F. WISHART,R. 31, 3986.

64

65

B. BANSCH,M. MEIER,M. M A R T I N I X , R. VAN ELDIK,C. S u , I. SUN, S. S. ISIED,J. F. WISHART,Inorg. Chem. 1994, 3 3, 4744. M. MEIER,J. SUN, J. E'. WISHART,R. VAN EI DIK,Inorg. Chem. 1996, 35, 1564.

66

67

50

8161.

V A N ELDIK,J. SUN, C. S u , S. S. ISIED,Inorg. Chem. 1992,

68 69

J . S U N ,J . F. WISHART,R. VAN ELDIK, R. D. SIIALDERS, T. W. SWADDLE, J. Am. Chem. SOC.1995, 117, 2600. J. S U N ,C. Su, M. MEIER,s. S. ISIED, J , F. WISHART,R. VAN ELDIK,Inorg. Chem. 1998, 37, 6129. H.-D. PROJAHN,R. VAN ELDIK,lnorg. Chem. 1992, 30, 3288. H.-D. PROTAHN, s. SCHINDLER, R. VAN ELDIK,D. G. FORTIER,C. R. ANDREW, A. G. SYKES,Inorg. Chem. 1995, 34, 5935.

70

71

72

M. ZHANG,R. VAN ELDIK,J. H. ESPENSON,A. BAKAC,Inorg. Chem. 1994, 33, 130. M. BUCHALOVA, P. M. WARBURTON, R. VAN ELDIK,D. H . BUSCH,]. Am. Chem. SOC.1997, 119, 5867. M. BUCHALOVA, D. H. BUSCH,R. VAN ELDIK,Inorg. Chem. 1998, 37, 1116.

73

L. E. LAVERMAN, A. WANAT,J. OSZAJCA,C. STOCHEL,P. C. FORD,R. VAN ELDIK,J. Am. Chem. SOC. 2001,

74

L. E. LAVERMAN, M. HOSHINO,P. C. FORD,J. Am. Chem. Soc. 1997, 119,

75

T. SCHNEPPENSIEPER, A. ZAHL,R. VAN ELDIK,Angew. Chem. Int. Ed. 2001, 40,

123, 285.

12663.

1678. 76

77

78 79

R. PIPOH, R. VAN ELDIK,s. L. B. WANG,W. D. WULFF,Organometallics 1992, I I, 490. K. 1. SCHNEIDER, A. NEUBRAND, R. VAN ELDIK,H. FISCHER, Organometallics 1992, 11, 267. R. P I P O H , R. VAN ELDIK,G. HENKEL, Organornetah 1993, 22, 2236. R. PIPOH,R. VAN ELDIK, Organometallics 1993, 12, 2668.

40

I

7 Effect of Pressure on Inorganic Reactions: Introduction and Mechanistic Applications 80 C. DUCKER-RENFER. R. VAK E L D I K A. .

1.

CANTY,Organometallics 1994, 13, 2412. 81 A. R. L. SKAUGE, R. D. SHALDERS, T. W. SWADDLE, Can. 1. Chem. 1996, 74, 1998. 82 P. C. FORD, D. R. CRANE,Coord. Chem. Reu. 1991, 111, 153. 83 G. STOCHEL and R. VAN ELDIK, Coord. Chem. Reu. 1997, 159, 153. 84 R. V A N ELDIIC, P. C. FORD, Adv. Photochem. 1998,24, 61.

R. V A N ELDIK, D. MEYERSTEIK, Acc. Chem. Res. 2000, 33, 207. 86 R. VAN ELDIK, 11. COHEN,D. MEYFRSTEIN, A n g m . Chem.. Int. Ed. Engl. 1991,30, 1158. 87 R. VAN ELDIK,W. GAEDE,H. COHEN. D. MEYERSTEIN, Inorg. Chem. 1992.31. 3695. 88 R. VAN ELDIK,H. COHEN.D. MEYERSTEIN, Inorg. Chem. 1994. 33, 1566. 85

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

I 2

The Effect o f Pressure on Organic Reactions: Basic Principles and Mechanistic Applications Frank Wurche a n d Frank-Gerrit K/arner$r 2.1 Introduction

Pressure in the range of 1-20 kbar strongly influences the rate and equilibrium position of many chemical reactions. (Units of pressure: 1 kbar = 100 MPa = 0.1 GPa = 14503.8 p.s.i. = 986.92 atm). Processes accompanied by a decrease in volume such as a C- C bond formation, in which the distance between two carbon atoms decreases from the van der Waals distance of ca. 3.6A to the bonding distance of ca. 1.5 A, are accelerated by pressure (volume of activation: AVz < 0) and the equilibria are shifted toward the side of products (volume of reaction: AV < 0). The reverse reaction, a homolytic bond cleavage, leads to an increase in volume ( A V t > 0, AV > 0). Pressure induces a deceleration of such a process and a shift of the equilibrium toward the side of reactants (Fig. 1.2, Chapter 1).Therefore, the effect of pressure seems to be particularly useful in controlling the course of competitive and consccutive reactions. Provided, that the activation volumes of the single reaction steps are different from each other, the application of high pressure can lead to an improvement of chemo-, regio- and stereoselectivity. Pressure also influences the physical properties of matter such as boiling and melting point, density, viscosity, solubility, dielectric constant, or conductivity. Before carrying out high pressure experiments it is important to have some knowledge of these effects. The melting points of most liquids used as common organic solvents are raised by increasing the pressure (Table 2.1). To perform a reaction in compressed solution it is necessary to use a solvent which does not solidify under the chosen conditions. The pressure-induced increase of the melting points (ca. 15-20 "C per 1 kbar), however, offers the possibility of running reactions in relatively thermally stable matrices. The solubility of gases in liquids is increased and that of solids is usually decreased by raising the pressure. Therefore, the solid solute of a saturated solution may precipitate during the generation of pressure and is no longer accessible for the reaction. The viscosity of liquid increases approximately twice every kilobar. This effect is particularly important for reactions containing diffusion-controlled steps. Finally, the compressibility of liquids is usually small compared to that of

41

42

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

Tab. 2.1. Pressure-induced increase in the melting points (m.p.) of common organic solvents [I]. m.p. ("C/1 bar)

m.p. ("C)

p (kbar)

~

~~

Acetone Benzene Cyclohexane Ethanol Diethylether Dichloromethane Pentane Toluene Water

-94.8

+20

i 5.5

+114 f97

f6.5 -117.3 -116.3 -96.7 -130.0 -95.1 0.0

+25 735 +25 0 +30 -9

(8.0) (5.5) (4.0) (20.0)

(12.0) (13.0) (12.0)

(94 (1.0)

gases. It amounts to 1 kbar between 4 % (water) and 18 % (pentane) of the original volume at 1 bar approaching rapidly an upper limit at higher pressure. For that reason experiments with compressed liquids and solids are less dangerous than those with compressed gases. A detailed discussion of the effects of pressure on physical properties of matter can be found in reference [ 11. The effect of pressure on chemical equilibria and rates of reaction can be described by the well-known relationship between pressure and Gibbs' enthalpy of reaction and activation, respectively (Scheme 2.1). The volume of reaction and

+

A + B [A-----B]** A - B AV= V(A-B)-[V(A)+ V(B)] AV* = V ( [A-----B]? - [V(A) + V(B)] AV, AV' : volumes of reaction and activation

v : partial molar volume KP : Equilibrium constant at pressure p

KP : Rate constant at p AG, AG' : Gibbs enthalpy and Gibbs enthalpy of activation

Scheme 2.1.

Volumes of reaction (AV) and activation ( A V f )

2.1 Introduction

activation can be determined from the pressure dependence of the equilibrium constant and rate constant, respectively. The volume of reaction also corresponds to the difference between the partial molar volumes of reactants and products and is usually determined from these quantities experimentally. Within the scope of transition state theory the volume of activation can be, accordingly, considered to be a measure of the partial molar volume of the transition state with respect to the partial molar volumes of reactants (Scheme 2.1). Volumes of activation are experimentally determined from the pressure dependence of the rate constants. The volumes of activation and reaction are themselves also pressure dependent as shown for the volumes of activation in Scheme 2.1. There is no theory explaining this pressure dependence which would allow the volumes of activation or reaction to be determined over a large range of pressure. Therefore, several empirical equations are employed to fit the pressure dependencies of rate and equilibrium constants [ 21. The volumes of activation and reaction derived from the pressure dependence of the rate or equilibrium constants by the use of these empirical equations are generally given at p = 0. These values at p = 0 differ only by immeasurably small amounts from those at atmospheric pressure ( p z 1 bar), so that the comparison

VW, VW’ : van der Waals volume [cm3-mol-’] (intrinsic molar volume of ground or transition structures related to one mole)

V:

M

molar volume of a pure liquid [cm3.mol-’] or

V =d

partial molar volume of a solute [cm3,mo~-‘]

v = IimQ

q, q’ : packing coefficient

C+O

q=vw; V

M (g.mol-’] : molar mass of the solute d

[g.cm”] : density of the solution

do [ g . ~ m - ~ : density ] of the pure solvent

c

[mol.l-’] : concentration of the solute

Van der Waals volumes ( V w ) , partial molar volumes ( V ) , and packing coefficients (77). Scheme 2.2.

D

M --.__ 1 d-do =do

V’ $=-W

V’

c

do

I

43

44

I

2 The Efect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

with volumes of reaction, calculated from the partial molar volumes of the reactants and products determined at atmospheric pressure, is feasible. The change in the intrinsic volumes of reacting molecules, as discussed for homolytic bond formation and dissociation is only, to a minor extent, responsible for the overall change in molar volumes measured experimentally from the densities of reactants and products and from the activation volumes. The intrinsic volume of a ground or transition structure is defined by the space occupied by the van der Waals spheres and can be calculated from the Cartesian coordinates of the molecular structure resulting from experimental data, molecular mechanics, or quantum-mechanical calculations and from the van der Waals radii of the different types of atom (e.g. R,(C) = 1.80A, R,(H) = 1.17A) derived from crystallographic data [ 31. The intrinsic volumes of ground structures can also be calculated from the table of group contributions [4].The van der Waals volumes, V,, is the intrinsic volume of a ground or transition structure multiplied by Avogadro’s number. The ratio, Vw/V or V$/V#, is defined as the packing coefficient, q or q # , of a ground or transition structure. The packing coefficients calculated for simple hydrocarbons are in the range = 0.5 to 0.G [ S ] . The empty space between the single molecules can be attributed to the so-called void volume and expansion volume required for thermally-induced motion and collision of the molecules in the liquid state [ S ] . The importance of the change in packing coefficients and, hence, in the void and expansion volume for the effect of pressure on chemical reactions, in particular on cyclizations, will be discussed in the following sections. The importance of considering the whole ensemble of molecules rather than single molecules in the explanation of pressure effects can be demonstrated with reference to electrostriction. In a heterolpc bond dissociation the attractive interaction between the newly generated ions and the solvent molecules leads to a contraction of volume that is, generally, much larger than the expansion of volume resulting from the dissociation. Thus, the overall effect, called electrostriction, leads to negative volumes of activation and reaction (AV# < 0, AV < 0). Neutralization of charges releases the molecules of the solvent cage, leading to positive volumes of activation and reaction (AV# > 0, AV > 0). A similar but less pronounced trend due to the effect of electrostriction is observed for charge concentration and charge dispersal. An increase in steric crowding in the transition or product states also results in a volume contraction (AV# < 0, AV < 0). Table 2.2 provides a survey on the expected contribution of various elementary reactions to their volumes of activation. There are many excellent monographs [l,71 and reviews [8, 91 on organic reactions at high pressure. Particularly, we would like to mention the three reviews written by le Noble and Asano in 1978 [GI, and by le Noble, Asano, van Eldik et al. in 1989 and 1998, which give the most complete survey on activation and reaction volumes up to 1998 [lo]. In this chapter we will discuss the effect of pressure on pericyclic reactions, such as cycloadditions, electrocyclic and sigmatropic rearrangements focussing on mechanistic information gained by the investigation of the pressure effect on organic reactions. Furthermore, the effect of pressure on freeradical reactions, ionic reactions and a few of the most recent applications of high pressure in organic synthesis will be discussed.

2.2 Cycloadditions Possible mechanisms to account for activation volumes o f various elementary processes 161. Tab. 2.2.

Mechanistic feature

Homolyhc bond cleavagc Homolytic association Bond deformation Ionization Neutralization Charge concentration Charge dispersal Displacement Stcric hindrance Diffusion control

Contribution in cm3 mol-'

-- +5" - -5 f10 -10 -0

-20"

t20a -5a

(-) >+20

aDependent on the solvent polarity

2.2

Cycloadditions 2.2.1

Diels-Alder Reactions, Mechanistic Aspects

Many Diels-Alder reactions show a powerful pressure-induced accelcration which is often effectively utilized for synthetic purposes [ 111.The activation volumes AV+ resulting from the pressure dependence of the rate constants are usually highly negative (AV' z -25 to -50 cm3 mol-I), sometimes even more negative than the corresponding reaction volumes AV, so that the ratio 0 = AV#/AV is close to or even larger than unity. For a comparison between activation and reactions volumes it is necessary to determine both data at the samc temperature which is, however, not feasible in many cases. The measurement of the pressure dependence of the rate constants frequently requires a temperature different from that used for the determination of partial molar volumes of reactants and products (generally room temperature). Therefore, the activation volumes have to be extrapolated to room temperature or the reaction volumes, correspondingly, to the temperature of reaction. The measurement of the temperature dependence of activation volumes requires a large collection of experimental data. An example is the Diels-Alder dimerization of isoprene, which has been reported in the literature [ 121. With the modern thermostated densimeters it is much easier to determine the ternperaturc dependence of partial molar volumes and, hence, of reaction volumes. From these data El'yanov extrapolated a generally applicable equation shown in Table 2.3 (footnote a) to describe the temperature dependence of activation and reaction volumes. The temperature dependence obtained experimentally for isoprene dimerization is in accord with the El'yanov equation. Accordingly, the partial molar volumes of the pericyclic transition states, which

I

45

46

Tab. 2.3.

I

2 The Ejiect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

Volume data o f selected Did-Alder reactions showing (0= A V # : A V ) ratios smaller than unity

(0 < 1). Reaction

Solvent

0.0-

40

-23.7

AVZ"

AV2p

Ob

-22.2

-33.0

0.67

Re$

n-BuC1

n-BuBr

O+,i' -

T["C]

70

-42.0

70

-33.0

35.0

47.8

0.73

-27.5

-44.9

0.61

30

-32.7

-32

-36.7

0.87

n-BuC1

40

-30.1

-28.3

-33.5

0.84

13

MeC02Et

10

-30.2

-32.2

-36.1

0.89

15

n-BuBr

25

-37

0.85

16

CHIC11

82

0.88

17

n-BuC1

58

0.88

17

0.97

18

0.96

19

QE

E = C02CH3

o+e,-

E

31.5

-43

31.5

-34.6

~

39.5

E

(11)

y+p'o:$

MeCN

65

\

0

0

"In cm3 mol-I; AV,? determined from the pressure dependence of the rate constant at temperature AVZ determined from the temperature of the activation volumes or extrapolated dependence from AV; by using the El'yanov equation AVZ = A V , f / [ l l 4 . 4 3 lo-' K-'(T - 25 "C)![20]. "@ = A V S :AVzs.

39.1

-41.6

-34.1

-35.3

-38.8

-36.9

2.2 Cycloadditions Tab. 2.4.

I

47

Volume data of selected Diels-Alder reactions showing (0= A V t : A V ) ratios larger than unity

(0> 1). ~

Reaction

Solvent

(4)

\

[+

OMe

AVL

: AVzs.

AV25"

Ob

-36.9 -37.3 -35.9 -36.4 -35.8 -35.4 -38.1

-38.3 -35.9 -34.5

0.96 7.04 1.04

35

-38.5 -39.0 -37.5 -38.0 -37.4 -37.0 -39.8

-36.8 -35.5 -33.4

0.97 0.99 1.14

30

-41.3

-40.4

-35.5

1.14

13

35

-44.7

-42.8

-31.9

1.22

19

35

-32.0 -53.6 -45.4 -43.7 -43.7

-30.6 -51.3 -43.5 -41.2 -41.8

-32.4 -32.2 -35.5 -28.2 -30.4

0.94 1.6 1.23 1.46 1.37

22

35

-37.2

-35.6

70

-37.0

-30.8

-28.6

1.08

23

70

-41.0

-34.2

-32.3

1.06

23

70

-36.5

-30.4

70

-35.0

-29.2

0

cm3 mol-I; AV: determined from the pressure dependence of the rate constant at temperature T; AVZ determined from AVr by extrapolation with the use of the El'yanov equation A V g = AV?/[l + 4.43 lo-' K-'(T - 25 "C)] 1201. b@ =

AV;'

0

Q- 0, 0

AV;'

Re$

21

0

0

0

~

T["C]

19

23

-29.7

0.98

23

48

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

can be derived from the activation volumes within the scope of transition-state theory, are approximately equal to or even smaller than those of the corresponding cycloadducts. This finding is surprising and seems to be contradictory to the generally accepted relationship betwecn molecular structure and its volumes. In the transition state the new bonds between diene and dienophile are only partially formed. According to quantum mechanical calculations the lengths of the partially formed CJ bonds in the pericyclic transition structures are in the range between 2.1 and 2.3A [24, 251. Thus the van der Waals volumes, V,, calculated for thc pericyclic transition structures by the use of quantum-mcchanical methods are generally larger than those calculated analogously for the corresponding cycloadducts (Tables 2.7, 2.8 and Scheme 2.6). Further insight into the relationship between the volume of activation and reaction and the mechanism of pericyclic reactions such as Diels-Alder reactions could be gained with the investigation of the effect of pressure on competing [4 21 and [2 21 or [4 21 and [4 + 41 cyloadditions. In competitive reactions the difference between the activation volumes can be directly derived from the pressure-dependence of the product ratio. All [ 2 + 21 or [4 + 41 cycloadditions listed in Tables 2.5 and 2.6 doubtlessly occur in two steps via diradical intermediates and can, therefore, be used as internal standards of the activation volumes expected for stepwise processes. Thus a relatively simple measurement of the pressure dependence of the product ratio should give important information about the mechanism of the competing [4 21 Diels-Alder cycloadditions. The results of these investigations are summarized in Tables 2.5 and 2.6. In the thermal dimerization of chloroprene (l),the activation volumes of two [4 21 cycloadditions leading to 2 and 3 were found to be smaller than those of the third 14 21 and the [2 21 cycloadditions leading to 4, 5, and 6 , rcspectively. Steward 1261 explained these results in terms of concerted Diels-Alder reactions competing with stepwise [2 21 cycloadditions. According to its larger (less negative) activation volume, the third Diels-Aldcr adduct 4 should also be formed in a non-concerted process. Similarly it can be concluded from the pressure dependence of the dimerization of 1,3-cyclohexadiene (7) that the endo-Diels-Alder dimer 8 and the [6 + 41-ene product 9 are formed concertedly whilc the exo-DielsAlder adduct 10 and the [2 21 cyclodimers 11 and 12 arisc via diradical intermediates. According to the activation volume data the Diels-Alder dimerization of 1,3butadiene [39] and o-quinodimethane (Table 2.5, entries (3) and (4), respectively) fall into the same class of concerted processes as those discussed for 1 and 7. while the Diels-Alder dimerization of hexamethylbis(methy1ene)cyclopentane seems to occur in stepwise fashion. According to the activation volume data summarized in Table 2.6 only the Diels-Alder reaction of 1,3-butadiene with a-acetoxyacrylonitrile seems to proceed concertedly while all other Diels-Alder and homo-Diels-Alder adducts are probably formed in stepwise processes comparable to the corresponding competitive 12 21 cycloadditions. Stereochemical investigations of the chloroprene and 1,3-butadiene dimerization using specifically deuterated derivatives confirm the conclusions drawn from activation volume data. In the dimerization of (E)-1-deuteriochloroprene(17) the diastereomeric Diels-Alder adducts 18a-Dz and

+

+

+

+

+

+

+

+

+

+

2.2 Cycloadditions

I

49

Tab. 2.5. Activation volumes AV# (cm3 mol ’), given in parentheses, and differences in activation volumes AAV’ (cm3 molk’) of competing [4 21 and [2 21 or [4 41 cyclodimerizations.

+

14

Reaction

+

+

+ Z]-cycloadducts

[Z + 21 or [4 + 41

* CI

1

0

AAVf

Ref:

cycloadducts ( A V i j

W ’ )

0

L-+ GI

5

6

(-22)

(-22)

70.5‘C

r-

7

-

10 (-22)

9a) (-32)

8 (-28)

119’C

/’-&

600 - 5300 bar

14

13

(-38.4)

15

16

(-20.9)

(-34.0)

80 ‘C 2 l a l -

m-o’

I

30

(- 15.8)

(-15.5)

+

a[6 41-ene reaction. bAV’ (8) - AV#(11). c A V # (9) - AVf (11). dAVt(lO) -AVi(12). eAV”(14) - AVf(15). ‘AVf(14) AVt(16). ~

181>-D2(Scheme 2.3) are formed in a (59:41) ratio. The non-stereospecific course provides clear-cut evidence that this Diels-Alder dimerization proceeds in stepwise fashion, as suggested from the activation volumes most likely passing through a diradical intermediate where rotation about C-C single bonds can substantially compete with the ring-closure reaction. The [2 21 cyclodimerization leading to a mixture of 19a-D2,19b-D2,and 19c-DZalso occurs non-stereospecifically as expected.

+

50

I

2 The EIfect of Pressure on Orgonic Reactions: Basic Principles and Mechanistic Applications

Activation volumes A V f (cm3 mol '), given in parentheses, and differences in activation volumes AAV# (cm3 mol ' ) of competing [4 21 and [2 21 or [4 41 cyclodimerizations.

Tab. 2.6.

[4

Reaction

+

(1)

+

+

+

+ Z]-cycloadducts

+

[2 21 or 14 cycloadducts

NCToAc

\

1 bar-9

+ 41

AAVf

Re$

-11.5

31, 32

OAc

kba

Q;

'c

*cl

0

32, 33

;

1 bar-10 kba

=a 1 bar-7

kbar

40 "C 1 bar-4

kbar

F

F F

3

-2

0

+ cI

35, 36

,

-0.7

(6)

34,35

38

5 1 bar-5 kba

CI

The Diels-Alder dimerization of Z,Z-1,4-dideuterio-l,3-butadiene occurs with 97% cis-stereoselectivity at atmospheric pressure (1 bar) and >99 % stereospecifity at high pressure (6.8-8 kbar). These findings provide good evidence for the stereospecific Diels-Alder mechanism in competition with a small amount of the nonstereospecific reaction, which can be almost completely suppressed by high pressure, and confirms the conclusion drawn from the different activation volumes found for the [4 21 and [ 2 21 cyclodimerization of 1,3-butadiene. From the pressure dependence of product ratio (endo:exo) it can be extrapolated that the endo-Diels-Alder reaction shows a slightly more negative activation volume than the corresponding em-reaction (AAV# z -2.5 cm3 mol although both reactions are evidently pericyclic. According to these data the difference between the activation volumes of competing pericyclic and stepwise cycloadditions (via diradical intermediates) is about AAV# = AV#(pericyclic) - AV#(stepwise) % -10 cm3 mol This is only found for reactions involving non-polar diradical intermediates and is not observed for polar zwitterionic intermediates. In the latter case the volume of activation of the zwitterionic intermediate can be more negative than that of a competing pericyclic reaction due to the solvent-dependent effect of electro-

+

+

'.

2.2 Cycloadditions CI

D 18a-D, 59

D 18b-D, 41

17-D

DE : DZ : Do

94.2: 2.2 : 3.6 93.7: 2.6 : 3.7 (recovered material)

p [bar]

1 1 6800 8000 Scheme 2.3.

T[~C]

138 120 120 120

19a-D,

19b-D,

19c-D,

48

34

18

[cis : trans]

[endo : ex01

97 : 3 97 : 3 299:99:<1

56 : 44 60 : 40 73:27 71 :29

Dimerization o f chloroprene [40]and Z,Z-l,4-dideuterio-l,3-butadiene [28].

Pericyclic Diels-Alder reactions:

AV*= -25 to 4 5 cm3mol-’, A V = -30 to -45 crn3mol-’

O = A V * / A V S I (orevenzl) MV* = A V ’ (pericyclic) - A V * (stepwise) = -10 cm3mo~-’ Volumes of activation and reaction (A!-’#, AV) of pericyclic Diels-Alder reactions. Comparison of the pericyclic processes with the corresponding stepwise processes involving diradical intermediates.

Scheme 2.4.

I

51

52

I

2 The Efect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

striction. Scheme 2.5 shows the competing [4+ 21 and [2 7 21 cycloadditions of 1,l-dimethylbutadiene with tetracyanoethylene (TCNE) as one example [ 9. 351. In toluene the activation volume of the [2 + 21 cycloaddition is found to be more negative than that of the [4+ 21 Diels-Alder cycloaddition whereas in butyronitrile reverse results are found. #

-

“NCG C N CN

Toluene:

M V * = +3.5crn3.rno~-’

Butyronitrile:

M V ’ = 4 . 0crn3.rnol~’

NC CNCN

J

The solvent-dependent effect o f pressure on the competing [4 21 and (2 21 cycloadditions o f tetracyanoethylene (TCNE) and 1,2-dimethylbutadiene a t 40 “C (9, 351.

Scheme 2.5.

+

+

One question that needs to be addressed is: why are the activation volumes of pericyclic cycloadditions smaller (more negative) than those of the corresponding stepwise reactions involving diradical intermediates? In the past it was assumed that the simultaneous formation of two new 7~ bonds in a pericyclic [4 21 cycloaddition leads to a larger contraction of volume than the formation of one bond in the stepwise process. The interpretation presented [28] is limited by the scope of Eyring transition state theory where the activation volume is related to the transition state volume, as mentioned above, and does not incorporate dynamic effects related to pressure-induced changes in viscosity [41].An extensive discussion of reaction rates in highly viscous solvents can be found in Chapter 3. For the pericyclic and stepwise cycloadditions of ethene to 1,3-butadiene (the prototype of Diels-Alder reactions) the molar volumes, the van der Waals volumes, Vw, and the packing coefficients, q, of the ground and transition structures shown in Table 2.7 were calculated following the method of Nakahara et al. [44]in order to uncover the effect of different bonding on the transition-state volumes. The packing coefficient, y ~ , of cyclohexene is significantly larger than that of the three isomeric hexadienes. Generally, y~ is found to be larger for cyclic compounds than for the corresponding acyclic ones. From the data listed in Table 2.7 the van der Waals volume of the Diels-Alder reaction [28, 44,451 can be calculated to be,

+

2.2 Cycloadditions

I

53

Molar volumes, V, van der Waals Volumes, V W , and packing coefficients, ti, calculated for acyclic and cyclic ground and transition structures needed for the explanation of the pressure effect on the Diels-Alder reaction of 1,3-butadiene with ethene. Tab. 2.7.

Compound

CH2zCH2 CH2 =CH-CH=CH2 CHZ=CH-CH~-CHZ-CH=CH~ CHZ-CH-CH~-CH=CH-CH~ CHz=CH-CH CH-CHz-CH3

0 7

vw/v

d

V = M/d"

VW0.b

'I =

0.6880 0.7000 0.7050

59.9c 83.2' 119.4 117.7 116.5

25.5 44.8 63.9 63.9 63.8

0.4257 0.5385 0.5354 0.5443 0.5475

0.8102

101.4

59.1

0.5829

109.1d

63.8

0.5829

118.7e

64.4

0.5424

120.4e

65.3

0.5424

i.54A

aIn cm3 mol-'. For the calculation of van der Waals volumes Cartesian coordinates resulting for ground structures from molecular mechanics calculations [42], and for transition structures from ab initio calculations [24], the following van der Waals radii were used: Rw(H) = 1.17A; Rw(C) = 1.80A. Calculated with volume increments [ 431. dWith the packing coefficient of cyclohexene ( q = 0,5829). 'Calculated with the average of the packing coefficients determined for the three isomeric hexadienes ( q = 0.5424).

+

with AVw = (59.1 - (25.5 44.8)) = -11.2 cm3 mol-', only roughly one-quarter of the experimentally accessible volume of reaction (AV = 101.4 - (59.9 + 83.2) = -41.4 cm3 mol-I). Consequently, a significant part of the observed AV results from the higher packing of the cyclic product rather than from the changes in bonding. The difference between the van der Waals volumes of activation calculated for the pericyclic and stepwise reaction (AAV: = AV: (pericyclic) - AV: (stepwise) = -6.7 - (-5.0) = -1.7 cm3 mol-l) is small and inconsistent with the experimental value (AAV# z -10 cm3 mol-'). In order to explain the experimental AAV# one has to assume [28] that the packing coefficient of the pericyclic transition state is similar to that of the cycloadduct . The difference between the activation volumes calculated for the pericyclic and stepwise cycloaddition using the packing coefficient of cyclohexene ( q = 0.5829) for the pericyclic transition structure (AVz (pericyclic) = -109.1 - (59.9 83.2) = -34.0 cm3 mol-') and the average of the packing

+

54

I

2 The Efect of Pressure on Organic Reactions: Basic Pn'nciples and Mechanistic A p p h t i o n s

coefficients of the three hexadienes = 0.5424) for the acyclic transition structure of the stepwise cycloaddition (AV#(stepwise)= -120.4 - (59.9 + 83.2) = 22.7 cm3 mol-') is, with AAV# = -11.3 cm3 mol-', in good accord with the experimental findings. Therefore, the analysis of activation volumes seems to provide important information regarding whether the transition-state geometry is cyclic or acyclic. The second question that needs to be addressed is, why are the activation volumes of some pericyclic Diels-Alder reactions smaller (more negative) than their reaction volumes so that the ratio 0 = AVf/AV > 1, This surprising result could be confirmed by three independent studies. In the two earlier studies Jenner et al. and Isaacs et al. found the ratio AV'/AV in the Diels-Alder reaction of furan with acrylonitrile [4G] and that of N-benzoylpyrrolewith N-phenylmaleic imide [ 471 to be larger than unity (0 = 1.06 and 1.37, respectively). The cycloadducts isolated from both reactions undergo smooth retro Diels-Alder reactions showing negative volume of activation (AVz = -2.0 and -8.3 cm3 mol respectively) in agreement with the value 0 > 1 determined for each forward reaction. In a more recent investigation F.-G. Klarner and V. Breitkopf observed that the retro Diels-Alder reactions of dihydrobarrelene and its 2-cyano derivative and of the endo-Diels-Alder adduct between dimethylfulvene and N-phenylmaleic imide are slightly decelerated by pressure showing positive volumes of activation whereas the retro Diels-Alder reactions of the endo and exo [4 + 21 cycloadducts between naphthalene and maleic anhydride and that of the exo adduct between dimethylfulvene and N-phenylmaleic imide are accelerated by pressure showing a negative volume of activation (Scheme 2.6) [48]. Grieger and Eckert [21, 491 considered two explanations of the ratio @ > 1 in the Diels-Alder reaction of isoprene with maleic anhydride: a larger dipole moment of the transition state or secondary orbital interactions which can only occur in endo Diels-Alder reactions. The findings that the difference between the activation volumes of many endo and ex0 Diels-Alder reactions is small (AAV# < 1-2 cm3 mol-l) and that the activation volume of retro Diels-Alder reaction of the endo cycloadduct between dimethylfulvene and N-phenylmaleic imide is positive and that of the retro Diels-Alder reaction of the corresponding exo cycloadduct is negative (Scheme 2.6), rule out that secondary orbital interactions are important and induce a larger contraction of the volume of the endo transition state. The following conclusions were drawn from the results obtained for the retro Diels-Alder reactions shown in Scheme 2.6. The packing of the entire ensemble consisting of solute and solvent and its reorganization during the course of reaction, and not the changes of the intrinsic molecular volumes of the reactants to the products during the course of reaction are most important for the magnitude of the activation and reaction volumes. The packing coefficients of the pericyclic transition states resemble those of the corresponding cycloadducts as already assumed for the explanation of the different activation volumes of pericyclic and stepwise cycloadditions [50]. In the retro Diels-Alder reactions showing AV# < 0, the packing coefficients of the transition states are calculated to be larger than those of the corresponding cycloadducts. This has been found particularly in the relatively polar systems bearing cyclic anhydride or imide functions (Scheme 2.6).

',

2.2 Cycloadditions

In these cases the size of the activation volumes obviously depends not only on the effective packing probably caused by the restriction of vibrations and rotations in the transition state, but also on the transition-state polarization enhanced by the polar groups leading to a further decrease in volume which is not observed in the less polarized cycloadducts. Blake and Jorgensen [ 511 have assumed similar effects to explain the acceleration of Diels-Alder reactions in water. Catalytic and solvophobic promotion of high pressure addition reactions will be discussed in Chapter 11. In so called homo Diels-Alder reactions, in which double bonds are replaced by three-membered rings or two n-bonds of the reactive 1,3-diene are bridged by an sp3-hybridized group, a powerful pressure-induced acceleration and hence highly negative activation volumes comparable to those of the ordinary Diels-Alder reactions have been observed. Examples of the pressure effect on homo-Diels-Alder reactions are shown in Scheme 2.7. The reaction of homofuran with fumaronitrile (Scheme 2.7, entry (8))[52] seems to be an exception in which the activation volume is significantly less negative than those in other cases. With the use of fumaronitrile and maleonitrile ( ( E ) - and (Z)-NC-CH= CH-CN) as dienophiles it was es-

R = H: 105.3 'C, AV*= +3.1, AV= +47.6 VwIV: 70.71118.7 q: 0.596

74.01121.8 0.608

R = CN: 105.6 ' C , AV*= +2.5, AV= +44.8

v W i v : 81.31123.3 1): 0.659

84.01125.8 0.674

87.0'C,AV*=-l.l (exoand endo),AV=+36.5(exo), AV=+36.2(endo) ex0 endo VwIV: 121.7l168.7 121.7/169.0 q: 0.721 0.720

Scheme 2.6.

ex0 endo 125.11167.6 125.11167.9 0.746 0.745

'='

Activation and reaction volumes determined experimentially and calculated van der Waals volumes and packing coefficients of retro-Diels-Alder reactions [48]. All volumes (in cm3 mol-') related t o the temperature of reaction.

=

eo

I

55

56

I

2 The Efect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

87.0 ‘C,A V L = +4.6,A V = +21.5

VwIV: 162.5/238.1 q: 0.682

165.41242.7 0.681 + e N - P h

‘0

121.7 ‘ C , AV*=-5.2, A V = +25.2

VwlV: 162.5/240.0 9: 0.677

165.4/234.8 0.704

All volumes (in cm3.mol-’) related to the temperature of reaction. Scheme 2.6.

(continued)

tablished that the reaction occurs stereospecifically with retention of configuration in the dienophiles and non-stereospecifically with respect to the ratio (endo: em). The relatively small pressure effect was explained in terms of a “late” transition state (with respect to the bond reorganization in the homofuran) in which the central cyclopropan bond has already been extensively cleaved and the bonding to the dienophile is still weak. 2.2.2 Complex Reactions and Synthetic Applications

The reaction of furanobenzocyclophane (20) with dicyanoacetylene (DCA) (Scheme 2.8) illustrates that the effect of pressure and that of a catalyst can be complementary [SS]. At 160 “C and 1 bar the reaction of 20 with DCA leads to the dicyanosubstituted furane (22) (20 %) and oxepin (24) (1 %). In a kinetically controlled reaction of 20 with DCA at 9 kbar and 20 “C the thermally unstable Diels-Aldcr adduct 21, the precursor of 22, is formed as the only product which undergoes a smooth retro Diels-Alder reaction at 1 bar and 20 “C producing the starting materials 20 and DCA. At 160 “C the retro Diels-Alder reaction of 21 leading to 22 and acetylene is obviously competitive with the non-productive reaction leading to 20

2.2 Cyc/oadd;t;ons

I

57

x=x (1)

(2)

E-CEC-E

Solvent h)

Ph-H

(cN)~c=c(cN)~ Ph-Me

T [ T ] AVr* *

(4) (6) (7)

CH2=CH-CN HCEC-E b) E-C-C-E h, c, E-NEN-E

(8)

(€)-CN-CH=CH-CN

(5)

CHC1, CIICI, CHCI, CHCI,

(CD,),CO

AVX'

Ob

Ref [53]

90

40.5

-32.0

-35.5

0.89

4o

-30.0

-32.8

-28.1 -30.7

-32.9 -36.2

0.85 [531 0.94

113

45.0

-32.4

-29.5

1.10 [17]

90 90 40.5 50

-32.0 43.5 -37.3 -28.2

-24.9 -33.8 -34.9 -25.4

-29.1 -38.8 -39.4 -29.6

0.86 0.87 0.89 0.90

70

-10

MeCN

EtOAc

A V Z ~ 'a

[54] [54] [17] [54]

[521

a) see footnotes a) and b) in Table 2.3; b) E = COzMe; c) E = C02Et Scheme 2.7.

Volume data o f selected homo-Diels-Alder reactions.

and DCA. At GO "C and 8.5 kbar or at room temperature in the presence of LiC104 as catalyst 20 reacts with DCA to produce the [ 2 21 cycloadduct 23 which isomerizes to the oxepin 24 at 160 "C. The reactions of the strained benzene derivative 25 with DCA were originally studied with the aim of synthesizing planar cyclooctatetraene derivatives which are interesting compounds with respect to the question of their antiaromaticity [ 561. The cyclooctatetraene derivative, 29 (Scheme 2.9) synthesized by photolysis of the barrelene derivative 26, however, turned out to be non-planar [57]. But the synthesis and the reactions of barrelene derivative 26 are interesting with respect to the utility of high pressure to control the course of reaction. The reaction of 25 with DCA at 1 bar and 127 "C produces the (1: 1) Diels-Alder adduct 26 (yield, 49 %) and the unexpected dark blue (2: 1) adduct 27 (yield, 14 %). At high pressure (9 kbar, 83 "C) 27 is the major product even after the low conversion of 36 % of the

+

58

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

NC

A F"

CN

k3J

- HCECH

22

21

'I'

NC

\

CN

20

23

24

Scheme 2.8. Competitive and consecutive reactions of furanobenzocyclophane 20 with dicyanoacetylene (DCA).

starting material 25. The observation, that the isolated (1: 1) adduct 26 reacts with DCA at 1 bar and 127 "C leading to 27, indicates that 26 is an intermediate in the reaction 25 + 2 DCA + 27. At 12 kbar 26 reacts with DCA readily at room temperature producing the homo-Diels-Alder adduct 28 which smoothly undergoes a rearrangement to 27 already at room temperature, so that 28 is also an intermediate in the reaction of 26 with DCA; 28 can be detected only by the use of high pressure. The dihydronaphthalene derivative 27 is an interesting compound regarding to its intense color and intramolecular dynamics [57]. Another example, which demonstrates that pressure is a useful parameter to control consecutive reactions, is the addition of DCA to benzodicyclobutene (30) which gives benzocyclobutene (31) at 1 bar and 125 "C as the only product [57]. At 11 kbar the reaction between DCA and 30 occurs readily at 53 "C producing the primary Diels-Alder adduct 31 in addition to 32. Isolated 31 rearranges to 32 slowly at 53 "C. Attempts to catalyze the reaction between 30 and DCA by means of Lewis acids such as AICl3 or TiC14 (analogous to the Diels-Alder reaction of DCA with parent benzene) failed. The cycloadditions of cyanoacetylene (Scheme 2.10) are good examples of the utility of high pressure to accelerate sluggish reactions and to control the course of complex reactions. In contrast to DCA cyanoacetylene is only a moderate dienophile reacting, for example, with 1,kyclohexadiene at 1 bar only at a temperature of ca. 100 "C at which the primary Diels-Alder adduct (2-cyanobicyclo[2.2.2]octa2,s-diene) is not thermally stable and undergoes a retro Diels-Alder reaction, producing benzonitrile and ethene [58, 591. The reaction is highly accelerated by pressure and occurs at 9 kbar and 50 "C. Under these conditions the primary Diels-Alder adduct is stable and can be isolated in good yields. A similar effect of pressure was observed in the trimerization of cyanoacetylene leading to 1,2,3- and 1,2,4-tricyanobenzenes as major products at 1 bar and 160 "C [58, 591. At 12 kbar

2.2 Cycloadditions (1) 1 bar, 127 C

25

&+& 27

I

26

v

26

20 'C, t1/2 = 20 d

20

26

acetone 29

1 bar, 125 'C: 11 kbar, 53 'C:

Scheme 2.9.

32 is the only product 31 is the major product

The effect of pressure on cycloadditions o f DCA to strained benzene derivatives

the trimerization occurs readily at 40 "C giving the thermally labile 2,3,5-tricyanoDewar-benzene which isomerizes to 1,2,4-tricyanobenzeneupon heating to a temperature L 50 "C. The results using high pressure provide good evidence that the thermal trimerization of cyanoacetylene occurs by a sequence of reactions consisting of the [2 21 cyclodimerization in the first step leading to the highly reactive 1,2-dicyanocyclobutadienefollowed by Diels-Alder cycloaddition of cyanacetylene to the cyclobutadiene producing the Dewar-benzene derivatives which aromatize to the observed benzene derivatives via orbital-symmetry forbidden electrocyclic ringopening.

+

I

59

60

I

2 The Efect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

1

.

12 kbar, 40"C,48h

+-CN

CN

NC

CN CN

16 Scheme 2.10.

:

:

6

1

Trimerization o f cyanoacetylene.

The reaction of cyanoacetylene with furan leads preferentially to the (2: 1) adduct 36 at 160 "C and 1 bar (Scheme 2.11). In analogy to the trimerization o f cyanoacetylene and the addition o f cyanoacetylene to [2, 21 paracyclophane [58, 591 the 1,2-dicyano-1,3-cyclobutadiene was assumed also to be the intermediate in this reaction. From the investigation of the pressure effect on the reactions of furan with cyanoacetylene and isolated oxanorbornadiene (33) with cyanoacetylene, both leading to the (2 :1) adducts 36 and 37, it could be concluded, that oxanorbornadiene (33) (the Diels-Alder adduct) and not the 1,2-dicyanocyclobutadieneis an intermediate in the reaction between furan and cyanoacetylene. The investigation of the pressure effect on the reaction between the furanobenzocyclophane (20) and cyanoacetylene led to similar results [55].

0

CN

33

CN 34

/CN 12 kbar,40 'C

&CN

-k I

CN 36 Scheme 2.11.

35 0

&'" /

NC

37

Addition o f cyanoacetylene to furan at various pressures.

2.2 Cycloadditions

In the past decade thc utility of high pressure has been demonstrated for many syntheses involving Diels-Alder reactions as key steps (see Chapters 9 and 11)We will describe here only fcw selected examples in order to illustrate the utility of high pressure in organic syntheses. An interesting example is the pressure-induced reaction of buckminsterfullercne CGOwith 1,3,5-cycloheptatriene [ 601. Generally C ~ reacts O as an electron-deficient dienophile or dipolarophile in numerous Diels-Alder or 1,3-dipolar cycloadditions and 1,3,5-cycloheptatriene as a diene. The reaction with CGOis a rare example where both adducts derived from the norcaradiene as well as the cycloheptatriene are observed.

Pressure-induced Diels-Alder reaction of cycloheptatriene + norcaradiene with C ~ as O dienophile.

Scheme 2.12.

The effect of pressure on hetero-Diels-Alder reactions is comparable to that on the aforementioned Diels-Alder reactions leading to carbocyclic products. The reactions are strongly accelerated by pressure showing highly negative volumes of activation (Scheme 2.13). The Diels-Alder reactions of unreactive dienes such as pyrazole [61-631 and oxazole 1641 derivatives (Scheme 2.13, entries (1-5)) are accelerated by high pressure which allows the temperature of reaction to be lowered so that the thermally labile cycloadducts and the 3,4-di-(trimethylsilyl)furane, respectively, can be isolated. The macrocycles shown in Scheme 2.14 can be synthesized by repetitive, highly stereoselective Diels-Alder reactions between the tetramethylene-substituted norbornane or 7-oxanorbornane derivatives as bis-dienes and the benzoanclated norbornadiene derivatives as bis- or tris-dienophiles, containing only 0x0 or methano bridges syn to one another. Whereas in each case the first reaction leading to the acyclic all-syn-configurated bis- or tris-diene occurs at atmospheric pressure, the subsequent inter- and intramolecular Diels-Alder reactions only succeed at high pressure. Obviously, the inter- as well as intramolecular Diels-Alder reactions are accelerated by pressure. The macrocycles are of interest in supramolecular chemistry because of their well-defined cavities of different sizes depending on the arene spacer-units. If the 0x0 (or methano) bridges are not exclusively Syn to one another either in the bis-dienophiles or bis-dienes (Scheme 2.15) or if the bis-diene is too large for the formation of a macrocycle, then the pressure-induced repetitive Diels-Alder reactions (proceeding again highly stereoselectively) produce rigid ribbon-type oligomers on a nanometer scale. The ethano-bridged bis-diene reacts less stereoselectively than the methano- or oxo-bridged bis-dienes [69]. It forms, with oxabenzonorbornadiene as the dienophile, a (2 : 1)cycloadduct and with bis-dienophiles, the ribbon-type oligomers with long chain lengths. A more flexible ribbon-type

I

62

I

2 The Efect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

0

(1)

t 14 kbar neat E = C02Me AV*= -18.8 crn3.rnol

’, A V = -30

crn3mo~-’

(4)

N--

a-12% Ar

Ar = o-X-CsH4,X = F,CI, Br, Me

1 bar, 300 ‘C: 8 kbar, 150 ‘C: Scheme 2.13.

76

:

24

only product

The effect of pressure on hetero-Diels-Alder reactions.

structure can be obtained by repetitive Diels-Alder reactions of the difuranocyclooctane as the bis-diene and dimethyl acetylenedicarboxylateas the bis-dienophile. The cage compound is formed in an undesirable side-reaction [70]. The extension of the spacer-unit of a bis-diene can be achieved by the pressure-induced DielsAlder reaction of the bis-diene with allenylchloromethylsulfone followed by basic HC1 and SO2 elimination [71]. High pressure can be also useful in tandem reactions consisting of a Diels-Alder [4 21 cycloaddition followed by a 1,3-dipolar[3 21 cycloaddition (see Chapter 10) or a Claisen-Ireland rearrangement, which is shown in Scheme 2.16. High pressure highly accelerates the Diels-Alder reaction which is the rate-determining step. Thus, pressure has a strong effect on the overall reaction [72]. The consecutive reactions of quadricyclane and norbornadiene, respectively, consisting of a homoDiels-Alder reaction with dimethylacetylenedicarboxylatefollowed by a Diels-Alder

+

+

2.2 Cycloadditions

R

X

X

Y

8-10 kbar

50-125'C yield: 2969%

rn = n = 0, R = H: X = Y = 0, X = 0, Y = CH2; X = Y = CH2, R = OMe: n = rn = 0,n = rn = 1, n = 1, rn = O

(2)

0

0

0

+ acyclic isornere (1:l) ratio 8.1 %

(3)

0

12 kbar

t

0W

0

O

Scheme 2.14. Synthesis of macrocycles by means of pressureinduced repetitive Diels-Alder reactions.

I

63

64

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

H

H 45.5 %

I"

'm+ lr m ' > SOPCH$~

(3)

//

i

1) 25 'C,15 kbar 2) KOtBU, THF

[711

\ '

Scheme 2.15. Synthesis o f ribbon-type oligomers by means of pressure-induced repetitive Diels-Alder reactions.

reactions of the primary adduct with 1,3-diene leading to polycyclic compounds are more efficient at high pressure, so that in some reactions the final bis-adducts can be observed only at high pressure [73]. 2.2.3

1,3-Dipolar [3

+ 21 Cycloadditions

1,3-Dipolar cycloadditions usually occur stereospecifically with retention of configuration in the dipolarophiles, and their rates are relatively insensitive to solvent polarity [ 7 4 . These results have been taken as an evidence for a concerted process. Only few exceptions are known where the non-stereospecific course of the reaction suggests a stepwise mechanism [75]. Activation and reaction volumes had been measured in the early 1980s for a selection of reactions with diazoalkanes, nitrones and alkyl azides as dipoles and electron-deficient or neutral alkenes as dipolarophiles. They were generally in the range of AVf = -(21 f 3) cm3 mol-' and AV = -(25 2) cm3 mol-' [76]. The relatively large ratios (AVz : AV) have been regarded as an indicator of a pericyclic mechanism. The absolute values of AV and AV# for 1,3-dipolar cycloadditions, however, are about 5 to 10 cm3 mol-', smaller than those of Diels-Alder reactions. No clear explanation has been given for this difference. But one can assume that one reason might be the dependence on ring size in the volume changes during cyclizations. According to the information ob-

2.2 Cycloadditions

X

R = H: 85 'C, 3 kbar: 88 % 100 'C, 1 bar: 100 %

CH2:

70 ' C , 3 kbar:

70 %

CHZCH2:

100 ' C , 9.5 kbar: 100 % (1 bar: 1 %)

0:

30 'C, 9.5 kbar: 86 % (1 bar: 0 %)

CHz

/I

CH-cH:

R = E: 90 'C, 1 bar: 100 % E = COzMe Scheme 2.16.

CHz: CHZCH2:

100 'C, 9.5 kbar: 66 % (1 bar: 0 %)

70 ' C , 3 kbar:

80 % (1 bar: 50 %)

100 'C, 9.5 kbar: 95 % (1 bar: 0 %)

The effect of pressure on tandem reactions.

tained later, the decrease in volume seems to be larger for the formation of a sixmembered ring rather than a five-membered ring [ 771. At 12 kbar and 25 "C benzylazide undergoes a regioselective 1,3-dipolar cycloaddition with a substituted methyl methoxymethacrylate, whereas at 1 bar and 80 "C a (1: 1) mixture of the two regioisomeric cycloadducts is formed (Scheme 2.17, entry (1))[78]. In the pressure-induced cycloaddition of nitrones with electrondeficient alkenes, alkynes and nitriles, the yields and selectivities of the cycloadducts are improved over those achieved at 1 bar and higher temperatures. The silyloxazolineresulting from the pressure-induced reaction of benzylphenylnitrone with trimethylsilylacetylene can be used for the synthesis of p-lactam antibiotics [ 791. The 1,3-dipolar cycloadditions of nitrones with electron-rich enolethers are less influenced by pressure. Moderate pressure of ca. 2 kbar and a Lewis-acid catalyst provided the optimum conditions for some of these reactions. The first step in the reaction of diazomethane with 1-phenylphosphole (Scheme 2.17, entry (3)) is certainly the addition of diazomethane to the phosphorus (R3P CHzNz -+ R3P= N-N=CH2) followed by hydrolysis leading to the highly reactive l-phenylphosphole1-oxide which reacts with diazomethane in the fashion of a 1,3-dipolarcycloaddi-

+

I

65

66

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

tion to form the monoadduct and subsequently the bis-adduct [SO]. (In the absence of water none of the cycloadducts are formed). Apparently, high pressure has a strong rate-enhancing effect on the first addition of diazomethane to the phosphole and also on the 1,3-dipolar cycloaddition, so that the overall reaction is almost completed at 12 kbar within 12 h compared to 10 day at 3-5 bar where the monoadduct is formed, preferentially.

E'

E = C02Me

(*)

80 'C, 1 bar, 30 %:

50

50

25 'C. 12 kbar, 83 %:

4

>95

~n,fi,8

SiMea 1 25'C.gkbarBn\N'o

Rn

3 - 5 bar, 10 d, conversion: 100 %

60

40

12 kbar, 12 h, conversion: 80 %

>99

Scheme 2.17.

n

Pressure-induced 1,3-dipolar cycloadditions.

2.2.4

[2

+ 21 Cycloadditions

+

[2 21 Cycloadditions involving ketene derivatives as one or both reaction partners are assumed to be rare examples of concerted [r: r,'] cycloadditions [Sl]. The activation volumes determined for the [2 + 21 cyclodimerization of diphenylketene [82] and the [2 21 cycloadditions of diphenylketene to various enolethers [S3] gave values of -30 cm3 mol-I and -22 to -52 cm3 mol-', respectively, and were highly negative. Thus, the effect of pressure leads to a powerful acceleration of these [2 21 cydoadditions comparable to that with Diels-Alder reactions, a characteristic which is useful for synthetic purposes. For example, various p-lactams can be easily synthesized by pressure-induced [2 21 cycloaddition of alkyl and aryl isocyanates and enolethers [ 841.

+

+

+

+

2.2 Cycloadditions

I

67

Dolbier and Weaver investigated the effect of pressure on the stereo- and regioselectivity in a certainly stepwise [2 + 21 cycloaddition of 1,l-difluoroallene to (Z)-p-deuteriostyrene involving a diradical intermediate (Scheme 2.18, entry (1)) [85]. In order to explain the pressure-induced increase in stereoselectivity corresponding to AAV# = A V # ( Z ) - AV#(E) = -2.6 and -2.8 cm3 mol-' (provided that there is a linear relationship between ln(Z/E) and pressure up to the very high pressure of 13 kbar), the authors concluded that at high pressure the ring-closure in the diradical intermediate leading to the (Z)-configured methylenecyclobutane derivatives, are favored over bond rotation which is a prerequisite for the formation of (E)-configuredmethylencyclobutanes. The activation volumes of stepwise [ 2 + 21 cycloadditions in non-polar systems proceeding via diradical intermediates are negative but significantly less negative than those of pericyclic Diels-Alder reactions as already mentioned. An example is the cyclodimerization of 1,3-butadiene (Table 2.5, entry (3)) [SO]. The activation volume (AVz -20.9 cm3 mol-') of the [2 t 21 cyclodimerization leading to truns-1,2-divinylcyclobutaneis less negative by -17.5 cm3 mol-' than that of the competing Diels-Alder [4 21 cycloaddition (AV" = -38.4 cm3 mol-') leading to 4-vinylcyclohexene so that the [2 + 21 cyclodimerization can be almost completely suppressed by the use of high pressure. The semicyclic 1,3-diene system of the 3methylenecyclohexene (Scheme 2.18, entry (2)) is confined strictly to a transoid conformation. Thus, its reactions are limited to the 12 21 cyclodimerization. The thermally-induced cyclodimerization of 3-methylenecyclohexene, however, cannot be observed because of the unfavorable position of the equilibrium which is on the side of the starting material even at high pressure (at 7.5-8.0 kbar and 25-60 "C: syn and anti-cyclodimers5 1 %). Starting from the photochemically accessible synand anti-cyclodimers the activation and reaction volumes of the mutual syn-anti interconversion and the ( 2 + 21 cycloreversions were determined [86]. From these data the activation volumes of the 12 21 cyclodimerizations which are not directly observable, can be calculated and are found to be comparable to those determined for the [2 + 21 cyclodimerization of chloroprene ( 2 3 "C: AVf = -22 cm3 mol-') [26], 1,3-cyclohexadiene(70.5 "C:AVf = -22 and -18 cm3 mol-' for the formation of the syn- and anti- [2 21 cyclodimer, respectively) [27], and 1,3-butadiene [SO]. The finding, that volumes of activation of the mutual syn-anti interconversion and the [2 21 cycloreversion of the cyclodimers of 3-methylene-cyclohexene dimers are positive and of the same order of magnitude, is good evidence that the cyclobutane ring-opening, leading to the corresponding diradical intermediate, occurs in the rate-determining steps of both reactions. The [ 2 21 cycloaddition of tetracyanoethene (TCNE) to vinylethers shows a powerful pressure-induced acceleration resulting in highly negative activation volumes (for example AV' = -55.0 cm3 mol-l, AV = -31.9 cm3 mol-' (25 "C, CH&) for the cycloaddition of TCNE to ethylvinylether) [87]. Detailed mechanistic studies [88] have led to the conclusion that these reactions proceed in a stepwise manner passing through interceptable dipolar intermediates. The observation, that the activation volumes are generally more negative than the corre=1

+

+

+

+

+

+

68

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

1.8 kbar 57.1 13.0 kbar 59.5

- MV':

-

AVf

@ @ @ @ @ @

+14.2 +13.1

+12.4 +13.2 -26.0(calc.) -24.2 (calc.)

: :

29.0 10.1

:

:

12.2 28.9

- 2.6

: :

1.7 1.5

- 2.8

AV +1 .o -1 .o

+38.4 +37.4 -38.4 -37.4

+

Scheme 2.18. The effect of pressure on stepwise [2 21 cycloadditions and cycloreversions (all volumes i n cm3 rnol ').

sponding reaction volumes (0= AV#/AV > 1) has been confirmed by the finding that [ 2 21 cycloreversion of a TCNE-vinylether cycloadduct is also accelerated by pressure and, hence, its activation volume is also negative [89]. Due to the effect of electrostriction the partial molar volumes of the dipolar intermediate and the polar transition states in its formation are smaller than those of the reactants (TCNE and vinylether) and the cycloadduct. This powerful effect of pressure on polar cycloadditions has been used in the synthesis of various dicyanoalkoxy-substituted cyclobutane derivatives [90].

+

2.2.5

[6

+ 41 and [8 + 21 Cycloadditions

The dependence of the orbital symmetry allowed [6

+ 41 cycloaddition of tropone

with 1,3-dienes on pressure was first studied by le Noble and Ojosipe [911 and they reported extremely small absolute values of AV# and AV. A reinvestigation by Takeshita and his coworkers [92] showed, however, that the activation and reaction

2.2 Cycloadditions

I

69

volumes of these cycloadditions are in the same order of magnitude as those of Diels-Alder reactions. Dogan [ 351 confirmed this finding with a study of the reaction between 1,3-butadiene and tropone in which a [6 41 cycloaddition competes with a 14 21 Diels-Alder reaction. The activation volume of the overall reaction was again found to be highly negative (Scheme 2.19). But the ratio between the [ 6 + 4 ] and [ 4 t 21 cycloadduct turned out to be almost pressure independent which means that the difference between the activation volumes (AAV') is almost zero and, hence, the activation volumes of both reactions are of the same value.

+

+

AV*"'

CPr-Ph; 80 'C

P+41

product ratio

(p = 0.9 kbar) (p=6.9 kbar)

10.0 10.8

Ref.

AVa'

-37.6

-36.1

[92]

-33.1

-34.9'

[92]

[4+21

: :

1 1

M V * ') = AV' [6+4] - AV' [4+2] = 4 . 3

a) in crn3.rnol-'; b) a t 80°C; c) 60°C [93];d) at 50°C. Scheme 2.19.

Cycloadditions with tropone.

+

Tropone can also react as a tetraene component in [8 21 cycloadditions induding the C=O double bond. Tropone reacts e.g. with 1,l-diethoxyethene (at 120 "C, 10 h, 1 bar) to give the corresponding [4 21, [8 21, and [6 41 cycloadducts in yields of 1.1,9.1 and 3.1 %, respectively (conversion of tropone: 16 %). At 3 kbar, 120 "C, only the 14 21 and [8 21 cycloadducts were formed in yields of 13 and 17 %, respectively (conversion of tropone: 30 %) [94]. Tropone reacts with 2,3dihydrofuran in a similar fashion leading to the corresponding [8 21 and [4 21 cycloadducts. The product ratio is again pressure dependent [95]. The heptafulvene derivative 38 shown in Scheme 2.20 undergoes a [8 21 cycloaddition leading to methyl-azulene-1-carboxylate(40) after elimination of C02 and ethanol from the undetectable primary cycloadduct, 39. The [8 21 cycloaddition competes with [4 21 cycloadditions. The study of the pressure effect on the competitive reactions has shown that the formation of the 14 21 cycloadduct (41) is reversible even at

+

+

+

+

+

+

+

+

+

+

+

70

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

10 kbar, and that 41 is not directly converted to 40. Thus, the formation of 40 can only occur via the intermediate [8 + 21 cycloadduct (39) [9G].

_/OEt

-

I..-

0

0

E= COzMe

+Ao

I

OEt

I 41

30

40 Scheme 2.20.

[8+ 21 and 14 + 21 cycloaddition of 38 with ethoxyethene 1961.

2.2.6 Cheletropic Reactions

Cheletropic reactions were defined by Woodward and Hoffmann to be processes in which two 0 bonds directed to the same atom are formed or cleaved in one step [97].The addition of a singlet carbene to an alkene is an example of a non-linear cheletropic reaction. Turro, Moss and coworkers generated phenylhalogenocarbenes (Ph-C-X: X = F, C1, Br) by flash photolysis of the corresponding diazirines and studied the pressure-dependence of their addition to tetramethylethene and (E)-2pentene at room temperature up to 2 kbar (Scheme 2.21, entry (1))[98]. The activation volumes (AVf = -10 to -18 cm3 mol-') were found to be less negative (by about 20 to 30 cm3 mol-I) than those found for Diels-Alder reactions. This result can be explained by the interdependence of the effect of pressure and ring size (uide inf7a). The addition of SO2 to 1,3-dienes (Scheme 2.21, entry (2)) is an example of a linear cheletropic reaction. The activation volume for the reaction between SO2 and 2,3-dirnethyl-1,3-butadiene was found by Isaacs and Laila to be more negative than the reaction volume of many comparable Diels-Alder reactions [ 991. For that reason one may speculate that in the rate-determining step the Diels-Alder adduct (the six-membered ring sulfinic ester) is formed followed by a rearrangement to the observed five-membered ring sulfone. The cheletropic addition of PhPBr2 to 1,3-dienes leading to the corresponding phospholes after basic HBr elimination of the primary adduct, can be dramatically accelerated by high pressure [loo]. In the analogous reaction of PhPC12 with 1-vinyl-cyclobutenethe time of reaction can be reduced from 2 months to 16 h by raising the pressure from 1 bar to 7 kbar (Scheme 2.21, entries (3) and (4))[loll.

2.3 Pericyclic Rearrangements

[951 X = F,CI. Br

AVf=-10tO-18

0 = 1.06

(3)

R'

R2

75-77%

R', R2 = H, CH3

7 kbar, 16 h, 75 % :

50

:

50

Scheme 2.21. The effect of pressure on cheletropic reactions. ( A V ' , A V in cm3 mol-'; 0 = AV#/AV).

2.3

Pericyclic Rearrangements

Many pericyclic rearrangements show a pressure-induced acceleration which is characterized by negative volumes of activation [l: 71. The effect of pressure on rearrangements is usually smaller than that on intermolecular cycloadditions and may be explained by the larger packing coefficients of the pericyclic transition states compared to those of the corresponding acyclic ground states as already discussed for the pericyclic and stepwise cycloadditions. 2.3.1

Sigmatropic [3.3] Shifts: Cope and Claisen Rearrangement

On the basis of stereochemical and kinetic investigations and quantum-mechanical calculations, most Cope rearrangements are regarded as being pericyclic processes [25, 102, 1031. The van der Waals volumes calculated for the parent 1,s-hexadiene

I

71

72

I

2 The Eflect offressure on Organic Reactions: Basic Principles and Mechanistic Applications

and the pericyclic transition state [77, 1041 are approximately the same (Scheme 2.22). This is understandable since in the symmetrical transition state bond breaking and making have proceeded to the same extent so that the effects of the two processes on the van der Waals volume compensate each other and thus no great overall effect of pressure on the Cope rearrangement is to be expected. However, it is reasonable to assume that the pericyclic transition state exhibits a larger packing coefficient than the acyclic ground state. Therefore the activation volume is expected to be negative. The activation volume of the degenerate Cope rearrangement of 1,shexadiene can be estimated to approximately -10 cm3 rno1-l if the packing coefficient determined for cyclohexene [44] is used as the unknown packing coefficient of the pericyclic transition state.

vw

63.9

63.6

V

1 1 9.4

109.1

A V ' = -0.3

A V =-10.3

All volumes are given in cm3-mol-'. The structural parameters necessary for the calculation of the van der Waals volume for the transition structure (TS) were taken from ab initio calculations [25,102]. The partial molar volume for the TS was calculated from the equation:

V (TS) = Vw (TS) / q (cyclohexene); q = Vw / V = 0.5829 (cyclohexene).

Partial molar volume ( V ) , van der Waals volume (Vw) and the volumes o f activation and reaction ( A V # , A V ) of the Cope rearrangement of 1,s-hexadiene. Scheme 2.22.

In fact, negative activation volumes of the expected size (AV# = -7 to -13 cm3 mol-') were found for several Cope rearrangements and related Claisen rearrangements (Scheme 2.23) [ 104-1081. The only exception is the activation volume of the Claisen rearrangement of the neat parent allylvinylether (Scheme 2.23. entry (4))which was determined to be AV# = -18 cm3 mol (1051. A new measurement of the pressure-dependent kinetics of this rearrangement in solution led to an activation volume of AV# = -10.3 cm3 mol-' [l09], which is of similar size to those determined for the other Cope and Claisen rearrangements. With the concept of cyclic interaction, introduced here, we can understand why the degenerate Cope rearrangement in bullvalene, investigated by Merbach, le Noble and coworkers [ 1101 with pressure- and temperature-dependent N M R spectroscopy, showed no significant pressure effect (AV+ = -0.5 cm3 mol-') (Scheme 2.23, entry (11)).As a result of the rigid bullvalene skeleton no such cyclic interaction appears in the transition state.

2.3 Pericyclic Rearrangements

'"*

[crn3.rno~-']

Ref.

4.7

[lo51

-9.7

[lo61

HO

(3)

6"

160'c 130.4'C, neat:

-18.0

128.9'C, nonane:

-10.3

:tip l ; , ; ; sT

Ph

cis, trans

-

#

meso

,-&+ Ph T S (boat)

Scheme 2.23.

-8.8

Ph

lrans. lrans

Volumes of activation (in cm3 mol ') o f Cope and Claisen rearrangements.

The utility of high pressure in the elucidation of reaction mechanisms can be also demonstrated by the example of the racemization and diastereomerization in 1,3,4,6-tetraphenyl-1,5-hexadiene (Scheme 2.24) indicating that a pericyclic Cope rearrangement competes here with a dissociative process involving free-radical intermediates [ 1111. Optically active tetraphenylhexadiene undergoes a facile racemization at temperatures just above room temperature. At 90 "C rucemic tetraphenylhexadiene shows a mutual interconversion to the meso diastereomer. Whereas the racemization may be the result of a pericyclic Cope rearrangement involving a chair-like transition state, the mutual diastereomerization cannot be explained by one or a sequence of Cope rearrangements involving chair- or boat-like transition states. The effect of pressure allowed an unambigous mechanistic conclusion. The observation, that the racemization is accelerated by pressure and, thus, exhibits a negative volume of activation, is good evidence for a pericyclic Cope mechanism in this case. The finding that the diastereomerization is retarded by pressure, ex-

I

73

74

I

of Pressure o n Organic Reactions: Basic Principles and Mechanistic Applications

2 The E'ect

Ph

-9.1

[I071

-12.8

[lo81

-13.4

[lo71

-11.3

[112]

-0.5

[110]

Ph

rac.

cis

(11)

Scheme 2.23.

(continued)

cludes a pericyclic mechanism or a cyclization by one-bond closure in the ratedetermining step and suggests a homolytic bond cleavage in the rate-determining step leading to the 1,3-diphenylallyl radicals as intermediates which can recombine to the two observed diastereomers. 2.3.2

Electrocyclic Rearrangements

The electrocyclic ring opening of heavily substituted cyclobutene derivatives which was investigated by Plieninger et al. [113],shows negative volumes of activation of

2.3 Pericyclic Rearrangements

I

75

racemization

Ph-Ph

optically active

90'~-

P h A P q-

diastereomerization

-phc Ph

Ph

Ph

meso

racernization: A V * = -7.4 cm3-mol-'

meso-,

rac:

A V " = +13.5cm3.rnol-'

rac + meso: A V * = +I 1.5 crn3.rno~-' Activation volumes of the racemization of optically active 1,3,4,6-tetraphenyl-l.S-hexadiene and the mutual interconversion o f the meso into the racemic diasteromer. Scheme 2.24.

different size depending on the substitution pattern, contrary to the expectation of positive activation volumes resulting from the ring opening. This result indicates that, for example, steric effects contribute substantially to the observed negative activation volumes, overcompensating for the increase in volume expected from ring opening. Clear-cut examples showing both effects are the isomerization of parent [114] and hexamethyl-Dewar benzene [114. 1151 leading to benzene and hexamethylbenzene, respectively (Scheme 2.25, entry (1)).The isomerization of the parent Dewar benzene is retarded by pressure (20 "C: AVf = t 5 cm3 mol-') whereas that of the hexamethyl derivative is accelerated by pressure (140 "C: AVf = -12 cm3 mol-', AV = -22 cm3 mol-'). Evidently, the steric crowding of the six methyl groups in the planar hexamethyl benzene is greater than that in the non-planar hexamethyl Dewar benzene overcompensating for the volumeincreasing effect of ring opening. In the ring opening of 2,3,5-tricyano Dewar benzene leading to 1,2,4-tricyanobenzene [ 591, the two effects obviously compensate each other so that this rearrangement is almost pressure-independent (51.3 "C: AV# zz 0 cm3 mol-'). In the transition state of the electrocyclization of (Z)-1,3,5-hexatriene to 1,3cyclohexadiene (Scheme 2.25, entry (2)), a new six-membered ring develops analogous to that in the Cope rearrangement [ 771. The electrocyclization is accelerated by pressure, showing a negative activation volume of AV# = -10.8 cm3 mol-' which is similar to those of the Cope rearrangements. From the volume data listed in Scheme 2.25 entry (2), the packing coefficient of the transition state is calculated to be approximately equal to that of the cyclic product and differs significantly from that of the acyclic reactant. This result again provides good evidence for the assumption that the packing coefficients of pericyclic transition states resemble those of the corresponding cyclic ground states.

76

I

2 The Efect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

2.24 A AV*=-10.8 AV =-14.8 Vw:

61.2

58.6

57.0

V: 118.7

107.9

103.9

: 0.516

0.543

0.549

11

Electrocyclic rearrangements of substituted Dewar benzene derivatives and (Z)-1,3,5-hexatriene. All volumes are given in c m 3 moI-’. Scheme 2.25.

2.3.3

Intramolecular Diels-Alder Reactions

In intramolecular Diels-Alder reactions, two new rings are formed. There are examples of relatively large pressure-induced accelerations comparable to the acceleration observed in intermolecular Diels-Alder reactions which can be exploited for preparative purposes [11G, 1171. All the systems studied hitherto contain polar groups and are, therefore, not very suitable for the analysis of the relationship between pressure effects and ring formation. The activation volume of the intramolecular Diels-Alder reaction appears to depend heavily on the type of solvent used as shown in Scheme 2.26, which indicates that effects other than ring formation, for example electrostriction, may also affect the size of the activation volume (see Chapter 9). In order to analyze the effect of ring size and ring number on the volume changes, the activation and reaction volumes of intramolecular Diels-Alder reactions in the non-planar pure hydrocarbon systems (E)-1,3,8-nonatriene (42) [ 1041 and (E)-1,3,9-decatriene (45),[77] were determined (Table 2.8). The packing coefficients g of the transition states are calculated to be equal to or even larger than those of the corresponding bicydic products and are comparable to the packing coefficients calculated for the electrocyclic ring closure of (Z)-lJ,S-hexatriene leading to 1,3-cyclohexadiene(Scheme 2.25, entry (2)). The absolute values of the acti-

2.3 Pericyclic Rearrangements

I

77

I

meta Ortho A V t = - 3 3 . 1 ; A V = - 3 0 (CHzClz)

A V * = - 1 3 . 4 ; A V = - 1 5 (CH3CN)

meta A V * = -32.1 (CHzC12)

AVt=-12.1

(CHFN)

Pressure dependence o f intramolecular DielsAlder reactions (AV", A V in cm3 rnol-'). Scheme 2.26.

vation volumes of intramolecular Diels-Alder reactions are approximately twice as large or even larger than those determined for the Cope rearrangements or the electrocyclization of 1,3,5-hexatriene to 1,3-cyclohexadiene (Scheme 2.25). From this it was extrapolated that each additional five- or six-membered ring formed in the rate-determining step of a pericyclic reaction contributes about -10 to -15 cm3 mol-' to its activation volume. A particularly instructive example is the thermolysis of (Z)-1,3,8-nonatriene(48) in which an intramolecular Diels-Alder reaction competes with a sigmatropic 1,shydrogen shift (Scheme 2.27) [ 1041. The use of high pressure leads to a significant acceleration of the overall reaction and allows a reversal of the selectivity. At 150 "C and 1 bar the 1,S-hydrogen shift passing through a monocyclic transition state is the preferred mechanism. At 7.7 kbar the intramolecular Diels-Alder reaction is the favored process, evidently due to its bicyclic transition state. The difference in the activation volumes of the two processes extrapolated from the pressure dependence of the product ratio (AAV# 5 -10 cm3 mol-') is comparable to that observed for pericyclic rearrangements involving monocyclic and bicyclic transition states. 2.3.4 Ene Reactions

With the concept introduced in the previous section, the pressure effects on Alder ene reactions related to sigmatropic 1,s-H shift, can also bc explained. These reactions show a powerful pressure-induced acceleration in rate. In many cases (Table

78 Tab. 2.8.

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

Volume data of the intramolecular Diels-Alder reactions of nonatriene (E)-42 and decatriene (E)-45.

Reaction

AV

Fy]+Jj) 8

-24.8

AV

-

32.0

0

0.78

,

cis- 43' 87.0 156.1 0.557

153.2-C n-hexanc

*

cis- 44 83.3 148.9 0.559

(€)- 42

vw

v

91.6 180.9 0.506 trans- 43'

87.0 156.1 0.557

-

i!;l

0.87

-45.4

0.83

-37.4

0.94

trans- 44

83.4 152.4 0.547

:4 0

4 r- '$7\'

-28.5

C D '

-37.6

-

172.5 'CI n-heptane

cis- 46' 97.7 164.4 0.594

cis- 47 93.8 156.7 0.599

(€)- 45

Vw 102.8 v 202.0 q 0.509 trans- 46'

97.7 167.0 0.585

trans- 47

93.8 164.6 0.570

All volumes are given in cm3 mol-' and related to the reaction temperatures at 153.2 and 172.5 "C, respectively, 7 = Vw/V is the packing coefficient and 0 = AVz/AV, the ratio of activation volume to reaction volume.

2.3 Pericyclic Rearrangements

I

79

150.2 ‘C, 24 h, n-Pentane:

1 bar: 80.2 % (2)-48, 13.7 % 49, 6.1 % cis-50 7.7 kbar: 5.5 % (Z)-48, 18.2 % 49, 76.4 % cis-50

M V ’ = A V F (2-48 + cis-50) - A V * = (2-48 + 49) 5 -10 crn3.rnoI-’ Scheme 2.27.

Pressure dependence o f the therrnolysis of (2)-48.

2.9, entries (1) and (2)) the absolute values of activation volumes were found to be larger than those of the corresponding reaction volumes and, hence, the ratio (AVf : AV) is larger than unity (0> 1) comparable to the Diels-Alder reactions listed in Table 2.4. This result was taken as an evidence of the pericyclic nature of the ene reactions. The van der Waals volumes of reactants, transition structure and product were calculated for the prototype of the ene reaction between propene and ethene (Scheme 2.28). The structural parameters necessary for this calculation were taken in the case of reactants and products from force-field calculations (MM2) and in the case of the transition structure from ab initio calculations [24, 251. In this case the van der Waals volume of the transition structure was calculated to be smaller than that of the product [ 1211 and, hence, the ratio (V$ : Vw) larger than unity. With the reasonable assumption that the packing coefficient of the transition structure is, at least equal to, but probably larger than that of the product, one can predict that the ratio between activation and reaction volume (AVf : AV) is also larger than unity. In the case of ene reactions between cycloalkene or alkenes and diethyl azodicarboxylate (DEAD) (e.g. Table 2.9, entry (3)), the ratio (AVz : AV) was found to be smaller than unity (0< 1).This result can be considered as an indication of a stepwise process in which a pericyclic transition state is not involved. Stephenson and Mattern, however, observed that the ratios SIR and kH/kD in the ene reaction with DEAD as enophile shown in Table 2.9 entry (4),were roughly equal, which was explained in terms of a concerted ene reaction. To explain this discrepancy Jenner and coworkers proposed a mechanism comparable to the formal ene reactions of alkenes with singlet oxygen or triazolindiones as enophiles forming in the first step three-membered rings between the alkenes and one center of the enophile prior to hydrogen transfer. To clarify the mechanism of the ene reaction with DEAD, it

80

I

2 The Efect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications Tab. 2.9.

Volume data o f selected ene reactions

n-Pro & -28.4

-27.0

1.05 118

-35.0 -31.3

-29.4

1.06

-52.0 -39.6

-35.4

1.12

-39.0

(1) n-Pr

110’C

E

(2) 52.2 ‘C

96,4 ‘c,

E

E’ E’

E’

E

S

SIR N

kH/kD N

120

U

R

3

cm3 mol-I; AV? determined from the pressure dependence of the rate constant at temperature T; AV; determined from the temperature of-the activation volumes or extrapolated by using the El‘yanov equation AVZ : AV{/[l 7 4.43 lo-’ K-’(T - 25 “C); [ZO]. b0= AV; ; AV,,.

~~~

Vwa’

35.1

25.6

54.6

55.8

-6.1

-5.3

1.15

Val

76.2

59.9

94.105

108.6

-(42-31)

-27.5

1.5- 1.13

qc)

0.4606

0.4273

0.58 - 0.52

0.5193

a In cm3 mol-’; = AV#/AV; ‘ 7 = Vw/V, for the calculation o f the partial molar volume of the transition state ( V # = V$/q#); the unknown packing coefficient 7 # was assumed to be within the range of the ene product and cyclohexane. van der Waals volumes of reactants, transition structure [24, 251 and product calculated for the prototype of the ene reaction between propene and ethene. Scheme 2.28.

2.3 Pericyclic Rearrangements

would be desirable to study the effect of pressure on the rate of the reaction studied by Stephenson and Mattern [ 1201. 2.3.5 The Relationship Between Activation or Reaction Volume and Ring Size

The first evidence for the relationship between ring size and volume data came from the observation, that the ring enlargement of cis-1,2-divinylcyclobutaneto 1,5-cyclooctadiene (AV = - 12.8 cm3 mol-l) and truns-l,2-divinylcyclobutaneto 4-vinylcyclohexene and 1,5-cyclooctadiene (AV = -9.6 and -17.4 cm3 mol-l, respectively) showed highly negative reaction volumes [ 1071. This means that the ring enlargement from a four- to a six- and finally eight-membered ring is accompanied by a substantial contraction in volume. The volumes of reaction calculated for the hypothetical cyclizations of 1-alkenes from their partial volumes (AV = V(cycloa1kane) V(1-alkene)) confirm this trend (Table 2.10) 1771. They decrease continuously from the formation of the three-membered ring (AV = -5.5 cm3 mol-') up to the formation of the 10-membered ring (AV = -32.3 cm3 mol-') and then, seem to be constant for the larger rings, whereas the van der Waals volumes of reaction (AVw) are approximately equal, with the exceptions of the formation of cyclopropane, cyclobutane and cyclopentane, and cannot explain the dramatic decrease found for the volumes of reaction. Therefore, this ring sizedependent dccrease in volume observed for the cyclizations of the 1-alkenes to the cycloalkanes results from the different packing of the cyclic and open-chained compounds rather than from the changes in their intrinsic molecular volumes. A simple explanation may come from the assumption that the empty space between the single molecules, which can be attributed to the so-called void volume and expansion volume required for thermally-induced motion and collision of the molecules in the liquid state as already pointed out by Asano and le Noble in 1973 [ 51, is reduced by the ring closure of an open chain largely due to the restriction of rotational degrees of freedom during the cyclization. Apparently, the larger the ring, the more degrees of freedom are restricted resulting in the observed ring sizedependent volume contraction. An upper limit of this effect seems to be the formation of cyclodecane. The increasing conformational flexibility in rings larger than CloHzo obviously requires a larger volume, which compensates the volume contracting effect of ring-closure so that the volume of reaction observed for the formation of these larger rings remains constant. It is interesting to note that the AV-values do not correlate with any other thermodynamic parameter such as enthalpy, entropy, or Gibbs enthalpy of reaction included in Table 2.10 for the formation of cyclopropane to cyclooctane. From these parameters the entropy of reaction should reflect best the restriction of the degrees of freedom as is assumed to occur in the explanation of volume contractions. The entropy of reaction decreases from the formation of cyclopropane to that of cyclohexane, but increases again for the formation of the larger rings, cycloheptane and cyclooctane, which can be explained with the increasing conformational flexibility of rings larger than cyclohexane. A better but not a linear correlation is found between the entropies of for~

I

81

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications Tab. 2.10. Volume of reaction, AV, van der Waals volumes of reaction AVw, enthalpies, entropies, and Cibbs enthalpy of reaction calculated for the hypothetical cyclizations o f 1-alkenes t o cycloalkanes by means of the corresponding thermodynamic parameters [76].

AVwa

AVRE

ASc

ACb

-1.7

-5.5

7.86

-7.0

9.95

-2.5

6.6

6.43

-10.3

9.50

-13.1

-9.56

-0

-3.8

-14.7

-13.46

-0

-4.4

-16.5

-19.47

-4.7

-21.2

-13.41

-19.6

-7.57

-4.9

-25.6

-9.88

-18.8

-4.28

-4.7

-30.9

-4.6

-32.3

-4.7

-32.8

-4.7

-32.3

-4.6

~-27.6

-0 --u A

"rn3 molt'. V(n-alkene) calculated by the use o f Exner increments [43]. V (cycloalkane) determined from density measurements in n.

hexane. bkcal mol-I. Ccalmolt' Ktl.

AHb

-21.0

-13.21

2.3 Pericyclic Rearrangements

I

83

-w -0

5

''..S4)

18-..

.................. ........................................................ ~

....*..

~-..(5) :: -...

.............................

12:

.................................................. 0 : L .... (6)

..,...

p

...

LC 4

.................................

,4:

Y

..-

(7)...............

-*.-....(8) *-...

10 1

,

~

l

l

l

~

l

l

l

[

.

I

,

(

.

,

I

4

mation of the cycloalkanes related to each CH2 group (ASrO(CH2)= AS;(CH,),/n) and the packing coefficients, ~(CHZ),,, which are identical to those related to each CH2 group (Figure 2.1) [77].According to the non-linear correlation the entropy parameter responds to conformational flexibility in medium-sized rings at an earlier state than the volume parameter. In the homologous sequence of n-alkanes, however, the entropy of formation per CH2 group (ASrO(CH2)= 9.5 cal mol-' K-') as well as the packing coefficient per CH2 group (q(CH2) = Vw(CH,)/V(CH2) = 10.4/16.2 = 0.64) remains constant with the increasing number o f atoms in the chain. Provided, that the activation volumes depend similarly on the ring size, the formation of larger rings should be dramatically accelerated by pressure. The intramolecular Diels-Alder reactions of (E)-1,3,8-nonatriene (42)and (E)-1,3,9decatriene (45), in which either a new five- and six-membered ring or two new sixmembered rings are formed, seems to be the first example of evidence for this assumption (Table 2.8). Furthermore, this ring-size effect explains why the activation volume of the formation of a three-membered ring in cheletropic reactions of carbenes with alkenes (Scheme 2.21) and of five-membered rings in 1,3-dipolar cycloadditions, are substantially less negative than those in the formation of sixmembered rings in Diels-Alder reactions (for example, the cheletropic addition of fluorophenylcarbene to tetramethylethylene: AVz = -17 cm3 mol-1 [98], the 1,3dipolar cycloaddition of diphenyldiazomethane to ADM: AVf = -23.2 cm3 rnol-l, AV = -26.8 cm3 mo1-l [122], and the Diels-Alder reaction o f 2,3-dimethylbutadiene to ADM: AVz = -34.1 cm3 mol-'; AV = -38.8 cm3 mol-1 [17] (ADM, methyl acetylene-dicarboxylate).

84

I

2 The E’ect

of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

2.4

Free-Radical Reactions 2.4.1

Homolytic Bond Dissociations

A volume expansion is expected for homolytic bond dissociations as already pointed out in the Introduction. This expectation has been confirmed for several homolytic bond cleavages showing positive activation volumes near AVf = +10 cm3 mol-’ [123]. The analysis of the pressure effect on the cleavage of azo compounds is however, complicated by the possibility of one- and two-bond scission processes [ 1241. The benzylic and benzhydrylic 1,Cshifts in the substituted pyridiminiumoxides (Scheme 2.29, entry (1))[ 1251 illustrate the utility of high pressure for the distinction between a pericyclic and dissociative mechanism comparable to the rearrangement of 1,3,4,G-tetraphenyl-l,S-hexadiene which has already

r,

meso+rac:AVf=+(10.7f4.8)

rac -+ meso : AVf = + (8.5 & 3.4)

AV‘ = + (37.5 zk 0.4) A H f = (44.7 i 1.9) kcalmol-’ ; A S * = (33.7 & 4.8) kcalmol-’

The effect of pressure on homolytic bond dissociations (all volumes in crn3 mol-’). Scheme 2.29.

2.4 Free-Radical Reactions

I

85

been referred to. The negative activation volume determined for the benzylic shift is good evidence for a pericyclic rearrangement whereas the positive activation volume determined for the benzhydrylic shift is in accord with the C-0 bond cleavage in the rate-determining step. The large difference between activation volumes observed for rearrangements proceeding via homolytic bond cleavage and recombination of the resulting free-radical intermediates (AVf = +6 to +13 cm3 mol-') on the one hand, and for the thiophenol-trapped dissociation of 3,4diethyl-3,4-diphenylhexane(AVz = +35.7 cm3 mol-') on the other, is notable and may be indicative of the reactions of caged or solvent-separated radical pairs, respectively [ 1261. 2.4.2

Quinone Oxidations (Hydrogen Transfer Reactions)

The oxidation of hydroarenes to arenes by quinones such as 2,3-dichlor0-5,6dicyano-1,4-quinone (DDQ) is frequently used for the synthesis of aromatic compounds. Brower et al. have already shown that the dehydrogentaion 1,4-cyclohexadiene to benzene [ 1281 or tetraline to naphthalene [ 1291 by thymoquinone is accelerated by pressure giving a negative volume of activation ((AVz = -33 (75 "C) and -28 (175 "C) cm3 mol-', respectively). A similar effect of pressure has been observed for the oxidation of leuco crystal violet with p-chloranil ((AV+ = -25 cm3 mol (21 "C) [130]. The pressure-dependent kinetic isotope effect of this reaction (29 "C: kH/kD = 11.5 (1 bar) and 8.2 (1.5 kbar)) indicates that hydrogen transfer occurs in the rate-determining step. The large k H / k D value at 1 bar and it pressure dependence was attributed to a quantum mechanical tunneling. Four mechanisms have been proposed for the quinone oxidation of hydroarenes [131]: (1) hydrogen atom transfer leading to a pair of free radicals in the ratedetermining step followed by fast subsequent reactions such as disproportionation or single-electron transfer (SET) finally producing the observed arene and hydroquinone; (2) direct hydride transfer leading to a pair of ions from which the observed products can be formed by proton transfer; ( 3 ) single-electron transfer followed by proton transfer producing the same radicals as direct hydrogen atom transfer; and (4) pericyclic hydrogen transfer which is limited to systems where vicinal C-H bonds are to be cleaved. More recent investigations by Ruchardt et al. [131] provided good evidence that the DDQ oxidation occurs via atom transfer (Scheme 2.30, mechanism (l)), comparable to other uncatalyzed transfer hydrogenations. Mechanism (1)as suggested by Ruchardt is further supported by the pressure dependence of the DDQ oxidation of various hydroarenes (Table 2.11) [132]. The finding, that the activation volumes are not significantly dependent on the solvent polarity, excludes the possibility of direct hydride transfer via mechanism (2). In this case a strong effect of electrostriction is expected due to the production of charged species in the ratedetermining step. Accordingly, the activation volume of the same reaction should be more negative in the less polar solvent (dielectric constant: E = 4.5 (MTBE), 6.03 (AcOEt), 35.9 (MeCN) [133]). However the opposite is found in experimental data

86

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

0.

OH

"'$f

+

0

R'

R'

+ OH

00

00

R" :2: C r ; " l

R'

0

Scheme 2.30. Mechanistic alternatives o f quinone dehydrogenations of hydroaromatic compounds. (1) Hydrogen atom

transfer, (2) direct hydride transfer, (3) single electron transfer, and (4) pericyclic hydrogen transfer.

(assuming that any solvent effect is indicated in these data). Pericyclic hydrogen transfer (mechanism (4)) can only occur in formal 1,3-dienes such as 1,2-dihydronaphthalene or 9,10-dihydrophenanthreneand should show a significantly more negative activation volume because of the pericyclic transition state than the hydrogen atom transfer of 1,4-dihydronaphthalene or 9,10-dihydroanthracene proceeding through acyclic transition states. But this is not the case which indicates that all reactions shown in Table 2.11 proceed with the same mechanism via hydrogen atom transfer. Further mechanistic support comes from the primary kinetic isotope effect observed for the DDQ oxidation of 9,lO-dihydroanthracene. The strong pressure dependence of the isotope effect may be attributed to a tunneling component of hydrogen transfer comparable to that observed by Isaacs et al. for the p-chloranil oxidation of leuco crystal violet [ 1301. The strongly pH-dependent activation volumes of the quinone oxidations of ascorbic acid (vitamine C) in water or methanol studied by Isaacs and van Eldik [134] are significantly less negative than those found for the oxidation of hydroarenes indicating different mechanisms. The results were explained by a rapid and reversible single electron transfer from the deprotonated anion of ascorbic acid to the quinone followed by a hydrogen atom transfer (see Scheme 2.31).

2.4 Free-Radical Reactions 187

The activation volumes in cm3 molk of the oxidation o f hydroarenes to the corresponding arenes by 2,3-dichloro-5,6-dicyano-l.4-quinone (DDQ) [132].

Tab. 2.11.

Hydroarene

T ("C) A V # (MTBE)"

1,4-Cyclohexadiene 25.1 9,lO-Dihydroanthracene 25.1 9~9,10,10-Tetradeutero-9,10-dihydroanthracene 25.1 9,9-Dimethyl-9,10-dihydr~anthracene~ 39.9 9,lO-Dihydrophenanthrene 39.9 24.9 1,4-Dihydronaphthalene 39.9 1,2-Dihydronaphthalene 64.9 Tetralin'

-24.4 -25.9 -35.7' -20.0 -25.1 -24.8 -25.3

-

A V # (MeCN/AcOEt)b

-29.5 -26.9 -

-22.9 -29.4 -28.9 -28.4 -26.7

Methyl-t-butylether. "acetomtnle and ethylacetate (1 : 1). 'kH/kD 10.8 (1bar) and 5.0 ( 3 kbar). the product is 10,1O-dimethyl-9-anthrone. ethe product is 1,2-dihydronaphthalene. a

0

OH

n

OH

R' = R2 = H : AV* = -20 (unbuffered), -16 (pH = 2), -4 (pH = 4.87) R' = CI, R2 = CN : AV' = -16 (unbuffered)

pH and pressure dependence of the oxidation of ascorbic acid (vitamine C) with quinones. Scheme 2.31.

2.4.3

Free RadicaI Cyclizations

f i e effect of pressure on the competing reactions of the free radicals 5-hexenyl and 6-heptenyl (Scheme 2.32, n = 1 and 2, respectively) were investigated in order to determine whether the cydization of free radicals showed a similar ring sizeiependence at high pressure as the cyclizations described in Sect. 2.3.5. Although the overall reaction of 5-hexenyl and 6-heptenyl is retarded by raising the pressure, :he ratio of the cyclization products (exo-trig and endo-trig) is not significantly pressure-dependent [ 1351. That means that the difference between the activation Jolumes of these cyclization reactions is nearly zero, showing no ring size depen-

88

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

dence. This finding can be explained by the assumption that in both transition states the rotational degrees of freedom of all former C-C single bonds are restricted and, hence, the expansion volumes, which were postulated to be important for the ring-size effect, do not differ from each other. I t is noteworthy that the pressure effect on the ratio between the cyclization products and the hydrogen atom transfer product ( ( k l k z ) / k 3 ) is also small. Evidently, the bimolecular hydrogen atom transfer has a similar activation volume as the monomolecular cyclizations. A11 volumes are expected to be highly negative.

Br

+

Bu3SnH AlBN

+v

Chlorobenzene I Benzene

exotrig

endotrig

i

I

Q - Q + k2) I k3

P LW

ki I kz

n = 1:

1 1000 5000

91.8 12.2 98.1 11.9 91.9 12.1

85.3 I 14.1 80.41 19.6 11.8 122.2

n=2:

1 1000 5000

82.1 I 11.3 83.41 16.6 80.1 I 19.3

20.2 119.8 16.9 183.1 16.0 I 84.0

(ki

The effect o f pressure on the cyclization of 5-hexenyl and 6-hepteny1, respectively. Transition structures of the endo-trig and exo-trig cyclization. Scheme 2.32.

2.5 Ionic Reactions

Reactions, in which ionic species are generated, show a large volume contraction and, hence a powerful pressure-induced acceleration which can be exploited for synthetic purposes. Among the nucleophilic substitutions and additions (such as

2.5 lonic Reactions

the aliphatic and aromatic SN reactions [ 1361, peptide formation (from esters) [ 1371, or the addition of nucleophiles to electron-deficient alkenes [ 1381 or oxiranes [ 139]),the Menschutltin-type S N reactions ~ are particularly worth mentioning. It is well established that the solvent-dependent activation volumes of the alkylation of pyridine derivatives are highly negative ( A V f = -20 to -50 cm3 mol-l) [GI. Stoddart et a]. used this high-pressure reaction to construct several catenanes [ 1401 and rotaxanes [ 1411 which are of interest as supramolecular devices. The [ 3lcatenanes (Scheme 2.33) can be synthesized in one step at high pressure (25 "C, 12 kbar, 5

1

1

l.DMF/12kbar/ ca. 20°C I 5 d 2. NHaPFe / H20

r\ r

0

?r? 0

0

1

1 0

0

4 PFC

0

0

iJ i i i i ?

Scheme 2.33.

0

0

0

0

Synthesis of catenanes by pressure-induced alkylation of pyridines

I

89

90

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

days) in reasonable yields of 31-33 %, whereas the yield of those catenancs in the reaction at 1 bar is only ca. 1 % and below after 9 weeks [142]. The rotaxanes shown in Scheme 2.34 were only obtained in the high-pressure reaction [141].

",l '. .,, . .

I,

I)DMF/lZkbar/20'C 2) NH4PF4/ H 2 0

+

' ~

u

'd'

t '0'

o-t

+

8 PFS

8 PFC

Scheme 2.34.

Synthesis of rotaxanes by pressure-induced alkylation of pyridines.

2.6

Concluding Remarks

It has been demonstrated that cyclizations are accompanied by a contraction of volume. The magnitude of this contraction depends on the number and size of the rings involved and is a result of the different packing coefficients of cyclic and acyclic structures rather than of the changes in their intrinsic molecular volumes during the cyclization. This effect is most important in cycloadditions and pericyclic rearrangements and explains the pressure-induced acceleration and the resultant negative activation volumes of these reactions. In reactions in which charged species are generated, the effect of electrostriction also leads to a substantial contraction of volume and, hence, to a rate enhancement at high pressure. The utility of high pressure with respect to these two effects has been described for the elucidation of reaction mechanisms and for organic synthesis.

References 191

Acknowledgments

We are grateful to the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie for financial support of our work. F.-G. K. thanks the coworkers mentioned in the references for their committed and skillful collaboration, and Ms I. Reiter and Ms H. Wo11 for their skilled assistance with the preparation of the manuscript, the tables, and schemes.

References N. S. ISAACS, Liquid High Pressure Chemistry, J o h n Wiley & Sons, New York, 1981. 2 T. ASANO,T. OKADA,J . Phys. Chem. 1984, 88, 238. 3 Y. YOSHIMURA, J. OSUGI,M. NAKAHARA,].Am. Chem. SOC.1983, 105, 5414. 4 A. BONDI,]. Chem. Phys. 1964, 68, 441. 5 T. ASANO,W. J. LE NOBLE,Rev. Phys. SOC.Jpn. 1973, 43, 82. 6 T. ASANO,W. J. LE NOBLE,Chem. Rev. 1978, 78, 407. 7 (a) W. J. LE NOBLE(Ed.), Organic High Pressure Chemistry, Elsevier, Amsterdam, 1988; (b) R. M. ACHESON, Organic Synthesis at K. MATSUMOTO, High Pressure, John Wiley & Sons, New York, 1991; (c) R. VAN ELDIK, C. D. HUBBARD,Chemistry under Extreme and Non-Classic Conditions, John Wiley & Sons, New York, 1997. 8 (a) W. J. LE NOBLE,Chem. unserer Zeit 1983, 17, 152; (b) W. J. LE NOBLE,H. KELM,Angew. Chem. Int. Ed. Engl. 1980, 19, 841-856; (c) G. JENNER,]. Chem. SOC.,Faraday Trans 11985, 81, 2437-2460; (d) K. MATSUMOTO,A. SERA,T. UCHIDA, Synthesis 1985, 1, K. 999-1027; (e) M. CIOBANU, MATSUMOTO, Liebigs Ann. Chem. 1997, M. 623-635; ( f ) K. MATSUMOTO, H. KATSURA,N. NAYASHI,T. KANEKO, UCHIDA,R. M. ACHESON,Heterocycles 1998, 47, 1135-1177; (8) F.-G. K L ~ R N E F. R , WURCHE,]. Prakt. Chem. 2000, 342, 609-636. 9 F.-G. K L ~ R N E R Chem. , unserer Zeit 1989, 23, 53. 1

VAN ELDIK,T. ASANO,W. J. LE NOBLE,Chem. Reu. 1989, 89, 549; (b) A. DRLJACA, C. D. HUBBARD, R. V A N ELDIK,T. ASANO,M. V. BASILEVSKY, W. J. LE NOBLE,Chem. Rev. 1998, 98, 2167-2289. F.-G. KLLRNER,M. K. DIEDRICH,in The Chemistry of Dimes and Polyenes, 1, Z. RAPPOPORT(Ed.), Wiley, New York, 1997, 547-617. J. RIMMELIN,G. JENNER,Tetrahedron 1974, 30, 3081. K. SEGUCHI,A. SERA,K. MARUYAMA, Bull. Chem. SOC.Jpn. 1974, 47, 2252. G. J E N N E R , J. RIMMELIN,Tetrahedron Lett. 1980, 21, 3039. J. R. MCCABE,C. A. ECKERT,1nd. Eng. Chem. Fundamentals 1973, 13, 168. C. BRUN,G. J E N N E R , Tetrahedron 1972, 28, 3113. G. JENNER,New.]. Chem. 1991, 15, 897. K. BECK,S. HUNIG,F.-G. KLLRNER,P. KRAFT,U. ARTSCHWAGER-PERL, Chem. Ber. 1987, 120, 2041. J. R. MCCABE,C. A. ECKERT,1nd. Eng. Chem. Fund. 1974, 13, 168. (a) B. S. EL'YANONV, E. M. J . Chem. SOC.,Faraday GONIKBERG, Trans 11979, 75, 172; (b) B. S. EL'YANONV, E. M. VASYLVITSKAYA, Rev. Phys. Chem. Jpn. 1980, 50, 169. R. A. GRIEGER,C. A. ECKERT,]. Am. Chem. SOC.1970, 92, 2918. R. A. GRIEGER, C. A. ECKERT,Ind. Eng. Chem. Fund. 1971, 10, 369. J. RIMMELIN, G. JENNER,H. ABDIOSKOUI,Bull. SOC.Chim. Fr. 1977, 341. Y. LI, K. N. HouK,]. Am. Chem. SOC. 1993, 115, 5414.

10 (a) R.

11

12 13 14 15

16 17

18

19 20

21 22 23 24

92

I

2 The Efect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications 25

26 27

28

29

30

31 32

33

34 35 36

37

38 39

K. N. I j O U K , Y. LI, I. D. EVANSECK, Angew. Chem. Int. Ed. Engl. 1992, 31, 682. C. A. STEWARD JR.,J. Am. Chem. SOL. 1972. 94, 635. F.-G. K L ~ R N EB. R ,M. J . DOGAN,0. ERMER,W. v. E. DOERING,M. P. COHEN,Angew. Chem. Int. Ed. Engl. 1986, 25. 110. U. DEITERS,F . C . K L ~ R N EB. R, KRAWCZYK, V. RUSTER,J. Am. Chem. SOL.1994. 116, 7646. (a) W. R. ROTH, S. B. P., Chem. Ber. 1981, 114, 3741; (b) M. BARTMANN, PhD Thesis, Ruhr-Universitat Bochum, 1980. J. BARAN,H. MAYR,V. RUSTER,F . 4 . KLXRNER, J. Org. Chem. 1989, 54, 5016. C. J. LITTLE,]. Am. Chem. SOC.1965, 87, 4020. V. RUSTER,Diploma Thesis, RuhrUniversitat Bochum, 1987; PhD Thesis, Ruhr-Universitat Bochum, 1991. (a) P. D. BARTLETT, K. E. SCHMELLER, J. Am. Chem. SOC.1968, 90, 6077; (b) J. S. SWENTON,P. D. BARTLE~T, J. Am. Chem. SOC.1968, 90, 2056. D. KAUFMANN, A. DE MEITERE, Angew. Chem. Int. Ed. Engl. 1973, 12, 159. B. M. I. DOGAN,P h D Thesis, RuhrUniversitat Bochum, 1984. M. R. D E CAMP;R. H. LEVIN,M. 1. JONES,Tetrahedron Lett. 1974, 15, 3575. (a) H . E. SIMMONS,J.Am. Chem. SOC. 1961, 83, 1657; (b) H. D. MARTIN,S. KACABU, H. J. SHIWEK,Tetrahedron Lett. 1975, 41, 3311. w. 7. LE NOBLE,R. MUKTHTAR,]. Am. Chem. SOC.1974, 96, 6191. The surprisingly small activation volume found for the formation of the very minor [4 41 cyclodimer 16 in the 1,3-buta&ene dimerization seems to be inconsistent with a stepwise mechanism. The small ratio 0 = AVvz/AV = 0.54 which is substantially smaller than that of the formation of 1 4 , O = 0.79, and almost equal to that of the formation of 15, 0 = 0.59, may indicate that 13 is formed via a stepwise [4 + 41 cycloaddition

+

passing through a (Z,Z)-configured diradical intermediate. 40 F.-G. K L ~ R N EV. R , RUSTER,B. ZIMNY, D. HOCHSTRATE. High Press. Res. 1991, 7, 133. 41 (a) L. N . KOWA,D. SCHWARZER, J. TROE,J. SCHROEDER, J. Chem. Phys. 1992, 97,4827; (b) M. A. FIRESTONE, M. VITALE,J. Org. Chem. 1981, 46, 2160. 42 PCMODEL, Serena Software, 1992. 43 0. EXNER,in Organic High Pressure Chemistry, W. J. LE NOBLF(Ed.), Elsevier, Amsterdam, 1988, 19-49. 44 Y. YOSHIMURA, J. OSUCI,M. NAKAHARA, Bull. Chem. SOC.Jpn. 1983, 56, 680. 45 R. A. FIRESTONE, G. M. SMITH,Chem. Ber. 1989, 122, 1089. 46 G. JENNER, M. PAPADOPOULOS, J. RIMMELIN,J. Org. C k m . 1983, 48, 748. 47 A. GEORGE,N. S. ISAACS,]. Chem. SOC., Perkin Trans. 2 1985, 1845. 48 F.-G. K L ~ R N EV. R , BREITKOPF,Eur. J. Org. Chem. 1999, 2757-2762. 49 R. A. GRIEGER,C. A. E C K E R T , Chem. ~. Soc., Faraday Trans 11970, G6, 2579. 50 F.-G. K L ~ R N E R B., KRAWCZYK, V. RUSTER,U. K. DEITERS,J. Am. Chem. SOC.1994, 116, 7646-7657. 51 J. F. BLAKE, W. L. JORGENSEN, /. Am. Chem. SOC.1991, 113, 7430-7435. 52 F.-G. K I ~ R N E R D., SCHROER,Chem. Ber. 1989, 122, 179. 53 G . JENNER, M. PAPADOPOULOS, Tetrahedron Lett. 1982, 23. 4333. 54 G . J E N N E R , M. PAPADOPOULOS, NouueauJ. de Chimie 1983, 7, 463. 55 v. BREITKOPF, P. BUBENITSCHEK, H. HOPF, P. G. JONES,F.-G. KLXRNER,D. SCHOMBURG, B. WITULSKI,B. ZIMNY, Liebigs Ann. Chem. 1997, 127 and unpublished results. 56 (a) G. I. FRAY,R. G. SAXTON,The Chemistry of Cyclooctatetraene and its Deriuatiues, Cambridge University Press, Cambridge, 1978 (b) P. G. WENTHOLT,D. A. HROVAT,W. T. BORDEN, W. C. LINEBERGER, Science 199G, 272, 1456-1459; (c) K. K. BALDRIDGE, J. S. SIEGEL,J. Am. Chem. SOC.2001, 123, 1755-1759; (d) A. M. MATSUURA, K. KOMATSU, J. Am. Chem.

References 193 SOC.2001, 123, 1768-1769; (e) F.-G. 68 P. R. ASHTON, I. F. STODDART.].Am. K L ~ R N EAngew. R, Chem. Int. Ed. Engl. Chem. SOC.1993, 1 15, 5422-5429. 2001 40, 3977-3981. 69 S. WEGENER, K. MULLEN, 57 F.-G. K L ~ R N E RR. , EHRHARDT. H. Macromolecules 1993, 26, 3037. R. BOESE,D. BLXSER, K. BANDMANN, 70 M. POLLMANN, K. MULLEN,].Am. N. HOUK,B. R. BENO,Chem. Eur.]. Chem. SOC. 1994, 116, 2318. 1999, 5, 2119. 71 R. M. GRAHAM, L. A. PAQUETTE,]. 58 H. HOPF,B. WITULSKI, Pure Appl. Org. Chem. 1995, 60, 5770. Chem. 1993, 65, 47. 72 J. VELKER, J.-P. ROBLIN, A. NEELS, A. 59 V. BREITKOPF,H. HOPF?F.-G. TESUORO. H. STOECKLI-EVANS, F.-G. K L ~ R N E RB. : WITULSKI, B. ZIMNY? K L ~ R N E7.6. R , GEHRKE, R. NEIER, Liebigs Ann. Chem. 1995, 613. Synlett 1999, S l , 925-929. 60 J.-F. LIN, N.KETO, A. MORI,H. 73 G. J E N N E R , High Press. Res. 1996, 15, R. ISOBE,Bull Chem. SOC. 51-57. TAKESHITA, ]pn. 1994, 67, 1507. 74 (a) R. SUSTMANN, Heterocycles 1995, 61 W. ADAM,C. BARNEVELD VAN, J.-S. 40, 1-18; (b) A. PADWA (Ed.), 1,3GEHRKE, F.-G. KLLRNER,J . Chem. SOC., Dipolar Cycloaddition Chemistry, Wiley Perkin Trans. 2 1999, 2723-2728. Interscience, New York, 1984. 62 K. EXNER,D. HOCHSTRATE; M. KELLER, 75 R. HUISGEN, G. MLOSTON, E. F.-G. KLXRNER, H. PRINZBACH, Angew. LANGIIALS,]. Am. Chem. SOC.1986, Chem. Int. Ed. Engl. 1996, 35, 2256108, 6401-6402. 2259. 76 (a) V. M. ZHULIN,Z. G. MAKAROV, M. 63 S. HUNIG,P. I
94

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications

W. R. DOLBIER, S . L. WEAVER,].Org. Chem. 1990, 55. 711. 86 (a) B. KRAWCZYK. P h D Thesis, Universitat Essen, 1996: (b) W. v. E. DOERING,J. EKMANIS, K. D. BELFIELD, Am. F.-G. KLXRNER.B. KRAWCZYK,]. Chem. SOC.(in press). 87 1. JOUANNE, H. KELM,R. HUISGEN, ]. Am. Chem. SOC.1979, 101, 151. 88 R. HUISGEN, ACC. Chem. Res. 1977, 10, 117. 89 W. 1. LE NOBLE,R. MUHKTAR,]. Am. Chem. SOC.1975, 97, 5938. 90 R. W. M. ABEN,J. GOUDRIAAN, J. M. M. SMITS,H. W. SCHEEREN, Synthesis 1993, 37. 91 W. J. LE NOBLE,B. A. OJOSIPE,]. Am. Chem. SOC.1975, 97, 5939. 92 H. TAKESHITA, S. SUGIYAMA, T. HATSUI,]. Chem. SOC.,Perkin Trans. 2 1986,1491; S . SUGIYAMA, H.TAKESHITA, Bull. Chem. SOC.Jpn. 1987, 60, 977. 93 The partial molar volumes of tropone, 1,3-butadieneand the [6 + 41 cycloadduct were measured by Dogan [35] to be V = 88.8, 83.1 and 153.2 cm3 mol-', respectively, at 21 "C and extrapolated to be 99.8, 87.9, 154.1 cm3 mol-', respectively, at 60 "C using the values give by Takeshita et al. [92] for the reaction of tropone with 2,3-dimethylbutadiene. 94 H. TAKESHITA, H. NAKASHIMA,S. SUGIYAMA, A. MORI,Bull. Chem. SOC. jpn. 1988, 61, 573. 95 Z. LI, A. MORI, H. TAKESHITA, Y. NAGANO,Chem. Express 1992, 7, 213. 96 A. MORI,Y. NUKII, H. TAKESHITA, T. NOZOE,Heterocycles 1993, 35, 863. 97 R. B. WOODWARD, R. HOFFMANN, Angew. Chem. Int. Ed. Engl. 1969, 8, 781. 98 N. J. TURRO,M. OKAMOTO, I. R. COULD,R. A. Moss, W. LAWRYNOWICZ, L. M. HADEL,J. Am. Chem. SOC.1987, 109,4973. 99 N. S. ISAACS,A. LAILA,~. Phys. Org. Chem. 1994, 7, 178. 100 N. S. ISAACS,G. N. EL-DIN,Synthesis 1989,967; Tetrahedron 1989, 45. 70837092. 101 F.-G. KIXRNER,S. YASLAK, R. DREWES, C. GESENBERG, M. PETER,Liebigs Ann. Chem. 1995, 203-210. 85

102

103

104

105 106

107

108 109

K. MOROKUMA, W. T. BORDEN.D. A. HROVAT. J . Am. Chem. SOC. 1988, 110. 4474. (a) D. A. HROVAT,B. R. BENO,H. LANGF, H.-Y. Yoo, K. N. HOUK,W. T. BORDEN,].A m . Chem. SOC.1999, 121. 10529; (b) W. v. E. DOERING,Y. WANG,]. Am. Chem. SOC.1999, 121. 10112-10118; 10967-10975. M. K. DIEDRICH,D. HOCHSTRATE, F.-G. KLXRNER,B. ZIMNY,Angew. Chem. Int. Ed. Engl. 1994, 33, 1079. C. WALLING,J. NAIMAN,]. Am. Chem. SOC.1962, 84, 2628. G. A. STASHINA,E. N. VASIL'VITSKAYA, G. D. GAMELEVICFI, B. S. EL'YANOV, E. P. SEREBRYAKOV, V. M. ZHULIN, Izu. Akad. Nauk SSSR, Ser. Khim. 1986, 329. w. V. E. DOERING,L. BIRLADEANU, K. SARMA,J. H. TELES, F.-G. K L ~ R N E R , j . 4 . GEHRKE, J . Am. Chem. SOC.1994, 116,4289-4297. P. JANSEN,Diploma Thesis, Universitat Essen, 1994. (a) F.-G. KLLRNER,D. MULLER, unpublished results; (b) D. M ~ ~ L L E R . Diploma Thesis, Universitat Essen, 1998. (c) The temperature-dependent kinetic analysis was carried out for this rearrangement in the gas phase between 142 and 173 "C leading to the Arrhenius equation log k = (11.1 0.3) - (29.3 f 0.5) kcal mol-'/RT. In the gas phase the resulting activation enthalpy and entropy, A H f = 28.5 kcal mol-' and ASf = -10.6 cal mol-I K - I , are larger than those in benzene solution (80 "C: A H # = 25.4 kcal mol-' and A S f = -15.9 cal mol-' K-l: J. BURROWS, B. K. CARPENTER, J . Am. Chem. SOC. 1981, 103, 6983). E. M. SCHULMAN, A. E. MERBACH, M. W. I. LE NOBLF. TURIN,R. WEDINGER, ]. Am. Chem. SOC.1983, 105, 3988. W. v. E. DOERING,L. BIRLADEANU, K. SARMA,G . BLASCHKE, U. SCHEIDEMANTEL, R. BOESE,J. BENETBUCHOLZ,F.-G. K L ~ R N E J.-S. R, GEHRKE,B. ZIMNY,H.-G. KORTH,]. Am. Chem. SOC.2000, 122, 193-203. 1 . 3 . GEHRKE,P h D Thesis, Universitat Essen, 1999.

+

110

111

112

References 195 113

114

115 116

117 118

119 120 121

(a) R. MUNDLICH.H. PLIENINGER, H. VOGLER,Tetrahedron 1977, 33, 2261; (b) R. MUNDLICH,H. PLIENINGER. Tetrahedron 1978, 34, 887. W. J. LE NOBLE,K. R. BROWER,C. BROWER,C. S., J. Am. Chem. SOC. 1982, 104, 3150. R. MWNDLICH,H. PLIENINGER, Tetrahedron 1976, 32, 2335. (a) L. F. TIETZE,C. Orr, K. GERKE,M. BUBACK.Angm. Chem. Int. Ed. Engl. 1993, 32, 1485; (b) S. JAROSZ,E. Tetrahedron KOZLOWSKA, A. JERZEWSKI, 1997, 31, 10775-10782; (c) Y. ARAKI, T. KONOIKE,]. Org. Chem. 1997, 62, 5299-5309; (d) A. TAHARI,D. UGUEN, Tetrahedron Lett. 1994, 23, 3945-3948; (e) G. GALLEY,M. PATzEL,]. Chem. SOC.~ Perkin Trans. 11996, 2297-2302; ( f ) L. F. TIETZE,A. SCHUFFENHAUER, Eur.]. Org. Chem. 1998, 1629-1637; (9) N. S. ISAACS,P. G. VAN D E R BECKE, /. Chem. SOC.,Perkin Trans. 2 1982, 1205; (h) M. BUBACK,j. ABELN,T. HUBSCH,C. OTT, L. F. TIETLE,Liebigs Ann. 1995, 9; (i) T. HEINER,K. MICHALSKY, K. GERKE,G. KUCHTA,M. BUBACK,A. DE MEIJERE,Synlett 1995, 355; ( 1 ) A. C. BRICKWOOD, M. G. B. DREW,L. M. HARWOOD, T. ISHIKAWA, P. MARAIS,V. MORRISON,].Chem. Soc., Perkin Trans 11999, 913-921; (k) T. HEINER,S. 1. KOZHUSHKOV, K. GERKE,T. HAUMANN, R. BOESE,A. D E MEIJERE, Tetrahedron 1996, 37, 12185-12196; (1) T . BUTZ, J. SAUER: Tetrahedron Asymm.1997, 8, 703714. M. BUBACK,K. GERKE,C. O n , L. F. TIETZE,Chem. Ber. 1994, 127, 2241. G. J E N N E RM. , PAPADOPOULOS, B. S. EL’YANONV, E. M. GONIKBERG,].Org. Chem. 1982, 47,4201. G. J E N N E RR., B E N SALEM, /. Chem. SOC.,Perkin Trans. 2 1989, 1671. L. M. STEPHENSON, D. L. MATTERN, J. Org. Chem. 1976, 41, 3614. Although the new C-C bond is only partially formed in the transition structure, the C-H-C distances are obviously both within the range of van der Waals distances, whereas in the product the non-bonding (=HzC----H-CHj- ) distance is,

122 123

124

125 126 127

128 129 130 131 132 133

according to the most stable geometry, larger than the van dcr Waals distance. G. SWIETON.J. JOUANNE, H. KELM,J. Org. Chem. 1983, 48, 1035. (a) U. B. IMASHEV, V. V. ZORIN,S. M. KALASHNIKOV, S. S. ZLOTSKII,V. M. Z H U L I N D. , L. RAKHMANKULOV, Dokl. Akad. Nauk. S S S R 1978. 242, 140; (b) A. A. LAPSHOVA, G. A. STASHINA,V. V. ZORIN,S. S. ZLOTSKII,V. M. Z H U L I N , D. L. RAKHMANKULDV, Zh. Org. Khim 1980, 16, 1251; (c) M. Y. BOTNIKOV, S. S. ZLOTSKII, V. V. ZORIN,E. K. KRAVETS, V. M. ZHULIN,D. L. RAKHMANKULOV, Izu. Akad. Nauk SSSR. Ser. a i m . 1977, 690; (d) E. V. PASTUSHENKO, M. Y. BOTNIKOV,S. S. ZLOTSKII, v. M. ZHULIN,D. L. RAKHMANKULOV, React. Kinet. Catal. Lett. 1981, 16, 195; (e) M. BUBACK,H . LENDLE,2. Naturforsch. 1979, 34a, 1482; ( f ) Y. KIMURA,Y. YOSHIMURA, M. NAKAHARA, Chem. Lett. 1987, 617. (a) R. C. NEUMAN,C. T. BERGE,G. A. BINEGAR, W. ADAM,Y. NISHIZAWA,]. Org. Chem. 1990, 55,4564; (b) R. C. NEUMAN,G. D. LOCKYER, J. Am. Chem. SOC.1983, 105, 3982; (c) R. C. NEUMAN,M. J. AMRICH,J. Org. Chem. 1980, 45, 4629; (d) V. M. ZHULIN,G. A. STASHINA,E. G. ROZANTSEV, Izu. Akad. Nauk SSSR, Ser. Khim. 1979, 977; (e) Y. OGO: M. KOTIMA,High Temp.-High Press. 1981, 13, 321. W. J. LE NOBLE,M. R. DAKA,1.Am. Chem. SOC.1978, 100, 5961. G. DIERKES,PhD Thesis, Universitat Essen (in preparation). G. KRATT, H.-D. BECKHAUS, C. RUCHARDT,Chem. Ber. 1984, 117, 1748. J. PAJAK,K. R. BROWER,]. Org. Chem. 1985, 50, 2210. K. R. BROWER,].Org. Chem. 1982, 47, 1889. N. S. ISAACS,K. JAVID,E. RANNALA,]. Chem. SOC.,Perkin Trans. 2 1978, 709. C. HOFLER,C. RUCHARDT,Liebigs Ann. Chem. 1996, 183. F. WURCHE,PhD Thesis, Universitat Essen (in preparation). C. REICHARDT, Solvents and Soli>ent E5ects in Organic Chemistry, VCH, Weinheim, 1990.

96

I

2 The Effect of Pressure on Organic Reactions: Basic Principles and Mechanistic Applications 134

N.

s. ISSACS, R. VAN ELDIK?/. Chew.

SOC., Perkin Trans. 2 1997, 1465.

C. VEKHAELEN, Diploma Thesis, Universitat Essen, 2001. 136 See e.g. (a) K. MATSUMOTO, S. HASHIMOTO, S. OTANI,]. Chem. Sac. Commun. 1991, 306; (b) T. IBATA, M.-H. SHANG,T. DEMURA,Bull. Chem SOC. Jpn. 1995, 68. 2941; (c) I. BARKPT, M. A. KERR,Tetrahedron Lett. 1999, 40, 2439. 137 (a) F.-G. K L ~ R N EU. R , KALTHOF,J. GANTE./. Prakt. Chem. 1997, 339, 359; (b) T. YAMADA, Y. OMOTE,Y. YAMANKA, T. MIYAZAWA, S. KUWATA, Synthesis 1998. 991-998; (c) T. SHIOIRI,T. IMAEDEA, Y. HAMADA, Heterocycles 1997, 46, 421-442. 138 See e.g. (a) S. PINHEIRO,A. J. GUIGNANT, D. DESMAELE, D'ANGELO,Tetrahedron Asymmetry 135

1992, 3. 1003; (b) G. TENNER, Tetrahedron Lett. 1995, 36, 233. 139 H. KOTSUKI,M. WAKAO,H. HAYAKAWA, T. SHIMANOUCHI, M. SIIITO,/. Org. Chem. 19913, 61, 8915. 140 D. B. AMABILINO, P. R. ASHTON,V. BALZANI,S. E. BOYD,A. CREDI,1. Y. LEE, S. MENZER,1. F. STODDART, M. VFNTURI,D. J. WILLIAMS, J. Am. Chem. SOC.1998. 120. 4295 and 141

142

references cited therein. R. BALLARDINI, V. BALZANI, W. A. E. DELL'ERBA, F. M. DEHAEN, RAYMO.I. F. STODDART, M. VENTURI. Eur. /. Org. Chem. 2000, 591. P. R. ASHTON,S. E. BOYD,C. G. CLAESSENS, R. E. GILLARD, S. MEWZER, J. F. STODDART, M. S. TOILEY, A. J . P. Chem. Eur. /. W ~ I I T ED. , J. WILLIAMS, 1997, 3, 788.

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

I 3

High-pressure Kinetics and Highly Viscous Media Tsutomu Asano

3.1

High Pressure and Dynamic Solvent Effects

The transition state theory (TST) may be considered to be established in 1941 by publication of a monumental book The Theory of Rate Processes [ 11. In Chapter VIII of the book, the authors discuss solution reactions and conclude “. . . that the ratedetermining step in solution is . . . the formation from the reactants of an activated complex which subsequently decomposes”. Though the authors pointed out the importance of diffusion in bimolecular reactions, they did not consider a possible break down of their two key assumptions, that is, thermal equilibrium between the initial and the transition state and neglecting recrossing, in unimolecular rate processes. The remarkable success of TST in the interpretation of kinetic effects of pressure [2] turned the attention of high-pressure kineticists away from a possible failure of TST and efforts were concentrated on the interpretation of the activation volume AV?’ obtained from pressure dependence of a rate constant k at a constant temperature (Eq. 3.1).

As a result of accumulation of thousands of activation volumes 13-51, high-pressure kinetics is now widely recognized as a powerful tool in the mechanistic investigation of reactions in solution. As long as we remain within the framework of TST however, it is not possible to study the events between the initial and the transition state. In other words, activation parameters, including activation volume, obtained in TST-valid conditions are “static” parameters concerned with the equilibrium constant for the activation step K # . Therefore, if we want to know, for example, whether solvent molecules are rearranged together with the structural transformation of the reactant molecule or the solvation shell rearrangement precedes the chemical transformation, kinetic experiments in which the conditions of TST are invalid have to be undertaken.

97

98

I

3 High-pressure Kinetics and Highly Viscous Media

rb

reaction coordinate r

Fig. 3.1. The double-well potential as a function of the reaction coordinate in Kramers’ model. The reactant moves back and forth along the reaction coordinate by Brownian motion and gradually ascends the barrier as shown by the zigzag arrows.

It is surprising that Kramers developed a comprehensive kinetic theory of chemical reactions before the establishment of TST [ 61. He modeled reactions as a movement of a particle caught in a double-well potential over an energy barrier “through the shuttling action of Brownian motion”. Figure 3.1 illustrates his model together with zigzag arrows showing schematically the movement of the particle along the reaction coordinate. The potential around the barrier top is given by

1 V(r) = Eo - -rncoi(r 2

-

rb)

2

(3.4

where rn is the mass of the particle. In this model, chemical transformations are considered to be in concert with thermal fluctuations of solvent molecules and, therefore, only one reaction coordinate was needed. At high solvent viscosities, the model predicts that the rate constant would be inversely proportional to the medium viscosity q (Eq. 3.3).

3.7 High Pressure and Dynamic Solvent Effects

In Eq. (3.3),kTsT is the rate constant expected from TST when the solvent thermal fluctuations are fast enough and the other conditions are the same. High-pressure chemists interested in kinetics were in a good position to measure the viscosity dependence of rate constants, i.e. “dynamic” solvent effects, and to examine the validity of Kramers’ prediction because solvent viscosity of organic liquids increases rapidly with increasing pressure [ 71 and pressures of several hundred McgaPascals (MPa) do not usually result in major mechanistic changes. Unfortunately, however, most of them, including this author, were strongly attracted to activation volume measurements and were satisfied with “naive” analysis of the results as criticized by Troe [ 8 ] . Troe and his coworkers were pioneers in the field of dynamic solvent effect studies and had good reasons for their criticisms. They measured rates of E / Z isomerization of stilbene and 1,4-diphenylbuta-1,3-diene (DPB) at their S1 excited state [9], typical unimolecular rate processes, in gaseous and supercritical fluids as well as in liquids at various pressures and temperatures. The results clearly demonstrated that the rate constant decreased with increasing viscosity in high density media. Similar effects were also observed by other groups. For example, Waldeck [lo, 111 and Hochstrasser [12]studied the photoisomerization of stilbenes in a series of solvents with an identical functional group but with different chain lengths. Fleming and his coworkers (131 as well as Hochstrasser [14], used high pressure as a tool to manipulate the viscosity in their study of stilbene and DPB. There were also important theoretical developments in the 1980s. Grote and Hynes elaborated Kramers’ model by introducing a concept of “frequencydependent friction” [15]. They assumed that friction felt by a reactant during the crossing of an energy barrier decreased with increasing frequency wb of the potential. A fundamentally different model was proposed by Agmon and Hopfield [lG, 171. Their model adopted two independent reaction coordinates, i.e. the chemical coordinate and the medium (solvent) coordinate. The theory was further developed by Sumi and Marcus [18, 191 and also by Basilevsky and Weinberg [20, 211. The difference between the Kramers-Grote-Hynes one-dimensional reaction coordinate (KGH) model and the Agmon-Hopfield two-dimensional reaction coordinate (AH) model can be seen clearly by comparing Figs 3.2a and b. In Fig. 3.2a, solvent rearrangement and chemical transformation cannot take place independently and the diagonal valley connecting the reactant and the product through the saddle point is the reaction coordinate in Fig. 3.1. In Fig. 3.2b, valleys of the reactant and the product are separated by a ridge with a saddle point and solvent rearrangement does not necessarily result in a movement along the chemical coordinate. If a reactant molecule tries to go over the ridge without changing its solvation shell, it will be faced with a high energy barrier. This model may be further simplified as shown in Scheme 3.1.

R

--

Scheme 3.1.

R*-

P

I

99

100

I

3 High-Pressure Kinetics and Highly Viscous Media

reactant

reactant

a

b

Fig. 3.2. Two-dimensional potential energy surface where the two coordinates are fully concerted (a) and where the solvent coordinate is almost independent o f the chemical coordinate

(b).

R and P in this scheme are the reactant and the product, respectively, and R’ represents the reactant molecules with a finite probability of crossing the energy barrier, i.e. reactant molecules in the area of the valley close to the saddle point. It was demonstrated by Sumi [ 191 that in the steady state approximation 1

1

1

kobs

kf

kTST

-- --+-

(3.4)

where the fluctuation-limited rate constant kf is a high viscosity limit of kobs in TST-invalid conditions, and may be considered to be the forward rate constant of the first step in Scheme 3.1. The validity of Eq. 3.4 was supported by exact numerical solutions obtained for the AH model [20, 211. The two models predict significantly different temperature dependence of kobs at a constant viscosity. In the KGH model, the reactant always goes over the same energy barrier in the ratedetermining step. Therefore, the isoviscous temperature dependence must be independent of the solvent viscosity inasmuch as the energy barrier does not show any significant dependence on the measure used to manipulate the viscosity, for example, the external pressure. On the other hand, in the AH model, the isoviscous activation energy is expected to decrease with increasing 7. At low viscosities where the thermal fluctuations of the solvent are fast enough to maintain the thermal equilibrium between the initial and the transition state, kf >> kTsT and

3.2 Selection of Reaction Systems

I

lo’

Thercfore, the activation energy will be approximately equal to the height of the energy barrier at the saddle point, Eo. However. as the thermal fluctuations in the solvent become slower with increasing viscosity, the rate-determining step gradually shifts to the first step in Scheme 3.1 or the movement of the reactant along the reactant valley in Figure 3.2b. As long as the viscosity is kept constant, the rate of solvent thermal fluctuations stays constant and the isoviscous temperature dependence would reflect the energy increase along the valley. Since this increase would never exceed Eo, the isoviscous activation energy is expccted to decrease with increasing viscosity in the AH model. The above discussion strongly suggests that it would bc possible to examine which model describes real reaction systems more accurately by checking the isoviscous temperature dependence of kobs at various viscosities starting from the TST-valid region. Applying high pressure seems to be practically the only one way to conduct such experiments because the solvent viscosity can be changed continuously and considerably without affecting reaction systems seriously [22]. It is true that the shape of the potential energy surface depends on the external pressure. However, the dependence can be estimated from the pressure effects on the rate constant in the TST-valid viscosity region. Fast reactions of electronically excited molecules such as photo-induced isomerization of stilbene is not suitable for this purpose, because the reaction is too fast to guarantee the validity of TST. Namely, neither kTST nor its pressure dependence can be estimated experimentally.

3.2 Selection of Reaction Systems

Reactions suitable for the kinetics aiming at the purpose mentioned above needs to satisfy one condition. The TST must be valid for a fairly wide range of pressure and temperature. Furthermore it is desirable that the reactant can be generated in situ and the reaction can be followed spectroscopically to obtain reliable rate constants. As the first set of reactions, thermal ZIE isomerization of three N-benzylideneanilines (benzaldehyde anils), i.e. N-[4-(dimethylamino) benzylidene]-4-nitroaniline (DBNA), N-[4-(dimethylamino)benzylidene]-4-ethoxycarbonylaniline (DBEA), and N-[4-(dimethylamino)benzylidene]-4-bromoaniline (DBBA), and two push-pull substituted azobenzenes, i.e. 4-(dimethylamino)-4’dimethylamino)-2-methoxy-4-nitroazobenzene nitroazobenzene ( DNAB) and 4’-( (DMNAB) as shown in Scheme 3.2 were selected. All of the forward reactions are “slow” thermal reactions of molecules in their ground So state and the half-lives were in the range of milliseconds to seconds, slow enough to guarantee the validity of TST in common solvents under “normal” conditions, because all the time scales of solvent thermal motion that may affect the reaction rate, i.e. vibrations, molecular rotations, and translational diffusion, are in nano- to picoseconds. The reaction mixture was contained in a glass/quartz bulb connected to a hypodermic syringe cut in the middle by means of a PTFE coupler [23]. The Z-isomers were generated by irradiating the solution from the

102

I

3 High-Pressure Kinetics and Highly Viscous Media

$3b

Me2N

py jfN

A

X = CH, Y = 4-NO2 X = CH, Y = 4-COOEt X = CH, Y = 4-Br X = N , Y =4-N02 X = N, Y = 2-Me0 and 4-N02

Me2N

DBNA DBEA DBBA DNAB DMNAB

Scheme 3.2.

outside of the high-pressure vessel through sapphire window(s) and their decay was monitored spectroscopically. In all of the cases studied, the reaction followed the first-order rate law and, thus, the rate constants could be determined unequivocally [24]. The compounds were stable even under UV irradiation and the measurements could be repeated when necessary. Selection of the solvent was another problem. Since it was expected that a high viscosity would be required to invalidate TST in these reactions, three viscous liquids with different functional groups were chosen, namely, 2,4-dicyclohexyl-2methylpentane (DCMP) as a non-polar solvent, glycerol triacetate (GTA) as a polar aprotic solvent, and 2-methylpentane-2,4-diol(MPD) as a protic solvent. All these solvents have a branched molecular structure and, therefore, the shear viscosity q increases much more rapidly with increasing pressure than in common solvents. The pressure effects on q of MPD are shown in Fig. 3.3. As can be seen from this figure, the pressure dependence of q could be approximately expressed by 2 0

I n 1

. a c

a

/

25'C

0

35'C

0

40'C

A

45'C

a

GOT

I

Cn 0

-

0

-1

-2

-3

0

100

200

300

400

500

P lMPa Fig. 3.3.

Pressure dependence o f viscosity o f M P D at various temperatures

3.3 Z/E lsomerization of N-Benzylideneanilines

I

103

Tab. 3.1. Shear viscosities at 0.1 MPa (rlo/Pa.s) and its pressure coefficients (n/GPa- ' ) for the three viscous liquids at various temperatures.

T/'-C

DCMP

-10

-5 0 5 10 15 20 2s 30 35 40

GTA

MPD

110

a

4%

a

70

a

1.16 0.546 0.281 0.156 0.0928 0.0584 0.0387 0.0267 0.0192 0.0142 0.0115

57.2 53.2 49.6 46.5 43.7 41.2 39.0 37.0 35.3 33.7 32.5

0.247 0.146 0.0921 0.0608 0.0419 0.0299 0.0221 0.0168 0.0131 0.0104 0.00841

46.2 41.8 37.9 34.4 31.3 28.5 25.9 23.6 21.6 19.6 17.9

0.574 0.324 0.194 0.122 0.0805 0.0553 0.0393 0.0288 0.0217 0.0167 0.0132

30.8 28.3 26.1 24.1 22.4 20.7 19.3 17.9 16.7 15.6 14.6

where qo is the viscosity at 0.1 MPa and ci is its pressure coefficient. In order to estimate the viscosities at the reaction temperatures, temperature dependence of qo and z were estimated by means of two empirical equations? Eqs 3.7 [25] and 3.8, where vo is a kinematic viscosity in centistokes at 0.1 MPa and p1-p4 are the parameters to be adjusted.

qo values

were obtained by multiplying vo with the density po. The values thus obtained are listed in Table 3.1 [ 2 G ] .

3.3

Z/E lsomerization of N-Benzylideneanilines

The isomerization of N-benzylideneanilines had been shown to be effected by the nitrogen inversion via the activated complex 3.1 where the N-phenyl group is in conjugation with the nitrogen lone pair [28]. H

\

104

I

3 High-pressure Kinetics and Highly Viscous Media

3.5

3.0

0

2.5 -

o

n

GTA AcOMe

0

2.0

I

1

I

I

I

PIMPa A comparison o f pressure effects on the rate of thermal Z / E isomerization of DBNA i n GTA and in AcOMe at 25 "C. Fig. 3.4.

The polarity of the reactant might decrease slightly during activation because of the rehybridization of the nitrogen 1291 and it, in turn,might result in a slight volume increase. On the other hand, increased rotational restriction of the phenyl group connected to the imino nitrogen in the activated complex may cause a small volume decrease. As a result, the activation volume for this isomerization is expected to be a small positive or small negative value. Kinetic measurements in various solvents [30] met this expectation. Figure 3.4 illustrates the pressure dependence of the isomerization rate of DBNA in GTA and methyl acetate (AcOMe) at the same temperature. Methyl acetate has a similar functional group to GTA and its dielectric constant (6.7 at 25 "C) is close to that of GTA (6.0 at 21 "C). As expected from the structural similarity in the two solvents, the rate constants were similar at 0.1 MPa [31]. Furthermore, the rate was little affected by an increase in pressure throughout the whole pressure range in AcOMe and at P < 200 MPa in GTA, strongly suggesting the validity of TST even at viscosities as high as 1 Pa s (= 1000 cP). However, a strong pressure-induced retardation was observed at higher pressures in the latter solvent. Since the reaction is a one-step isomerization, there is little possibility of a shift of the rate-determining step in the usual sense and this retardation is difficult to rationalize within the framework of TST. The most reasonable explanation is that the reaction was retarded by an increase in the solvent viscosity because the same retardation was not observed in AcOMe.

3.3 Z/E lsornerization of N-Benzylideneanilines

I

105

’I 0

e e 0

20’C

0

15’C

0

1O’C

A

5’C

O’C

6

I

-5‘C I

I

I

This view is supported by the results in Figs 3.5-3.7. In these figures, results at different temperatures are shown for the isomerization of DBNA in the three viscous solvents [32-341. In all of these cases, the higher the reaction temperature, the higher the pressure required to observe the pressure-induced retardations. This is reasonable because the viscosity decreases with increasing temperature and we were convinced that these retardations were in fact dynamic solvent effects observed for “slow” thermal reactions for the first time [35]. We are now in a position to examine viscosity dependence of the rate constant. The results obtained for DBNA in DCMP are plotted against viscosity in Fig. 3.8. It is obvious that the isoviscous temperature dependence stays almost constant at lower viscosities but decreases with increasing viscosity in the TST-invalid viscosity region. Similar tendencies were observed for all of the anils in all three of the viscous solvents studied and these results strongly suggest that the present reaction systems cannot be adequately described by the KGH model. The solvent coordinate has to be treated separately from the chemical coordinate as assumed in the AH model. In other words, a reorganization of the solvation shell takes place prior to crossing the energy barrier without inducing a major structural change in the reactant. If the AH model is applicable to the present reaction systems, the fluctuation-limited rate constant k f obtained from Eq. 3.4 would give a linear Arrhenius plot because solvent rearrangement requires thermal activation. Estima-

106

I

3 High-Pressure Kinetics and Highly Viscous Media

r------

3'5

PIMPa Fig. 3.6. Pressure effects on the rate of thermal Z/E isomerization of DBNA in CTA at various temperatures.

-.

2.5c1 o o o o o o o o ~ o

<'

u)

0 0

o n

0 0 0 0 0

n

0 0

0 0 0 0 0 0 0 0

m

0 -

0

0 0 0 ~ . O ~ \ A A A A A A A O ~ O O A A 0 0 o o 3 8 8 m m m m m A A A 0 8 8 8 A 8 1.5-

m

8 8

1.0-

0

10'C

0

0 'C

0

O'C

A

-5'C

8

0.5

0

-1O'C I

I

I

I

100

200

300

400

c

3.3 Z/E lsornerization of N-Benzylideneanilines

-2.0

0

2O'C

0

15'C

0

1O'C

A

5%

m

O'C

e

-5'C

I

I

I

I

0.0

2.0

4.0

6.0

t

log (q/Pa s) Fig. 3.8. Plots of log kobs against log 7 for the thermal Z/E isomerization of DBNA in DCMP at various temperatures.

tions of kTsT under TST-invalid conditions were performed by linearly extrapolating log kobs against pressure because the kinetic effects of pressure in the TST-valid region were small. The k f values thus obtained satisfied our expectation as can be seen from examples in Fig. 3.9. Although the quality of linearity vaned from one system to another, no systematic deviation from the expected linearity was observed. It has been reported by many groups that the rate of isomerization at the S1 state is inversely proportional to a fractional power of the solvent viscosity 1391. Those measurements were almost certainly performed at TST-invalid viscosities. Therefore, the observed rate constant would have been fluctuation limited. This consideration suggests a similar viscosity dependence of our kf (Eq. 3.9).

Several plots of log k f against log 9 are given in Fig. 3.10. The slopes of the plots in Fig. 3.10 were around -0.7 and, hence, smaller than unity as expected. Again, in all of the cases studied similar linearity was observed. On the basis of these results, we concluded that the AH model is a better description of our "slow" reaction systems. Now the next step is to examine whether it is possible to obtain further information on the extent of correlation of the two coordinates. Figure 3.10 gives an important hint on this point. The isoviscous tem-

I

107

108

I

3 High-Pressure Kinetics and Highly Viscous Media

0.0035

0.0034

0.0036

0.0037

0.0038

IITIK Fig. 3.9. Isobaric Arrhenius plots for the fluctuation-limited rate constant kf in the thermal Z/E isomerization of DBNA in DCMP at various pressures.

A

0 0

3h

7

r A? Y

0

0 -

20

1-

0-

-1 -

0 O C D A

0

0

25 'C

0

20 'C

0

15 'C

A

10 'C

5 'C

2

3

4

5

log( qlPa s) Fig. 3.10. Viscosity dependence of the fluctuation-limited rate constant kf in the thermal Z / E isomerization of DBNA and DBBA in CTA a t various temperatures.

3.4 Z/E lsomerization of Push-Pull Substituted Azobenzenes

perature dependence of k f is quite small. Actually it is too small to estimate the isoviscous activation energy with sufficient accuracy. This is also true of other systems. Only in the case of DBNA and DBEA in DCMP, could we estimate (with difficulty) the values to be 20-30 kJ mol-'. What these findings indicate is that the slope of the reactant valley is relatively gentle probably because the correlation of the solvent and the chemical coordinate is weak as shown in Fig. 3.2b. This is reasonable because any significant changes in solute-solvent interactions may not be required to achieve the rehybridization of the imino-nitrogen atom.

3.4

Z/E lsomerization o f Push-Pull Substituted Azobenzenes A similar but mechanistically different isomerization of two push-pull substituted azobenzenes shown in Scheme 3.2 was also studied. Although it is possible for azobenzenes to isomerize by the same nitrogen-inversion mechanism as the Nbenzylideneanilines [40], rotational isomerization via a highly dipolar activated complex 3.2 becomes predominant [41, 421 in polar solvents provided that the compound contains a strongly electron-donating dialkylamino group as in DNAB and DMNAB.

Me2N+

3.2

Because of the increased solvation, the partial molar volume of the reactant decreases in the activation step (AVf << 0). This reaction gives us an opportunity to study whether the correlation between the two coordinates becomes stronger when electrostriction is reinforced during the activation step. Figure 3.11 illustrates the effect of pressure on the isomerization rate of DNAB in common and viscous solvents. As expected from the mechanism, the reaction was strongly accelerated in aprotic as well as in protic solvents. The activation volumes were -21 and -22 cm3 mol-' in ethanol and methyl acetate, respectively. Similar pressure-induced accelerations were also observed in MPD and GTA at lower pressures. However, the direction of the pressure effect was reversed at higher pressures. Figures 3.12 and 3.13 illustrate pressure effects observed for DNAB in the two solvents. The results for DMNAB were qualitatively identical. Pressure-induced retardation clearly occurred at higher pressure when the reaction temperature was raised. The results leave little doubt about the validity of TST at low pressures and its failure at higher pressures in the rotational isomerization of push-pull substituted azobenzenes in GTA and MPD [43]. The observed rate constants are plotted against g in Figs 3.14 and 3.15. The results in GTA (Fig. 3.14) were qualitatively different from those in MPD

I

109

110

I

3 High-PressureKinetics and Highly Viscous Media

A

200

0

0

EtOH

0

MPD

0

AcOMe

600

400

1 10

800 P/M Pa

Comparisons of pressure effects on the rate of thermal Z / E isomerization of DNA6 in common and viscous solvents at 15 "C.

Fig. 3.11.

2

-

o

. 2 m

1-

so

0

v

(51

0 -

o

~

o

no

7

[,O

0 0

o o o o o o o o o o o o 0 o o o

0

A

()

A A A

A

40'C

0

25°C

0

15'C

A

-2

0

0

0 0 0

0

A

0

o

0

A

I,

-1 -

~

0

O

A A A A

0"o

~ 0

A

5'C I

I

I

I

200

400

600

800

11

3.4 Z/E isornerization of Push-Pull Substituted Azobenzenes

2.5

i n

0.5

,,I1l,

o.o[;

-5-c

-0.5

m;

I

0

100

200

300

400

500

600

1

700

PlMPa Fig. 3.13. Pressure effects on the rate o f thermal Z/E isomerization of DNAB in MPD a t various temperatures.

0

-2.5

15’C

A

5°C

I

I

I

I

0.0

2.5

5.0

7.5

log( q/Pa s) Fig. 3.14. Plots o f log kobs against log q for the thermal Z/E isomerization of DNAB i n CTA at various temperatures.

10.0

I

ll1

112

I

3 High-Pressure Kinetics and Highly Viscous Media

2-

. -

7

0)

01

n

so 0,

0 -

1-

‘m

f8

0-

-1

log( q/Pa s) Plots of log kobsagainst log q for the thermal Z/E isornerization of DNAB in MPD at various temperatures.

Fig. 3.15.

(Fig. 3.15). In GTA, the isoviscous temperature dependence almost disappears at the TST-invalid viscosity region. In MPD, however, a large proportion of the isoviscous temperature dependence still remained even at 1 x lo5 Pa s. The results in MPD suggest that the two coordinates are partially correlated and the fluctuationlimited rate constant k f would show temperature dependence even at a constant viscosity. The kobs values at TST-valid pressures were fitted to either one of the following empirical three-parameter equations [28, 441 to estimate the TST-expected rate constants.

k In 2 = a~

+ b I n ( l + cP)

(3.10)

k0.l

(3.11)

As in the N-benzylideneanilines, kf gave linear isobaric Arrhenius plots and satisfied Eq. 3.9. The values for DNAB are plotted against q in Figs 3.16 and 3.17. Obviously, the isoviscous temperature dependence of kf was small in GTA. However, significant temperature dependence was clearly seen in MPD as shown in Fig. 3.17. The isoviscous activation energies Eaf and E ~ T were ~ T estimated for DNAB and DMNAB and they are listed in Table 3.2. The E a f and E U T ~values T in Table 3.2 suggest that the reactant molecules climb more than half of the potential energy barrier during the solvent rearrangement

3.4 Z/E lsornerization of Push-pull Substituted Azobenzenes

0

-

0

r

P

0

2-

0

Y m

v

-0

0 0 00

0

1-

A

0

0 A0

0 OA

0

0-

O A

0

40’C

0

15‘C

O

25’C

0

A

5‘C

A

1

-1

2

3

I

1

1

1

4

5

6

7

8

log( q/Pa s) Fig. 3.16.

Viscosity dependence of kf in the thermal Z/E isomerization of DNAB in CTA.

0 00 v

3-

m

-0

“oe

0

35%

0

25’C

0

15’C

A

5%

0% 0 0 OOO

0

2-

-5.c

OOO 00 A ‘0 A OO A

r

m

n

A

mm

1-

‘m

~ A

B

0

0

I

1

1

2

4

6

8

log( q/Pa s) Fig. 3.17.

Viscosity dependence of kr in the thermal Z / E isornerization of DNAB in MPD.

I

113

114

I

3 High-Pressure Kinetics and Highly Viscous Media

Isoviscous activation energies calculated from kr (Eaf/kJ mol-’) and k ~ s ~ (EaTSr/kJ mol-’) for the thermal Z/E isomerization o f DNAB and DMNAB in MPD.

Tab. 3.2.

w a

10 102 103 104

5

DMNAB

DNAB

4

EaTST

56 & 6.2 50 k 4.4 45 & 3.1 45 k 0.8

69 f 0.8 69 i 1.3 69 i 3.1 69 2.0

EaTST

58 k 3.7 54 i 2.6 50 & 1.9 46 & 1.8

70 k 0.2 71 0.5 72 f 0.6 70 k 1.9

around the Z-isomer in MPD. Considering the absence of such energy increases in GTA, the hydroxyl groups on the MPD molecule may be responsible for this behavior. In order to break the nitrogen-nitrogen TC bond heterolyhcally, electron-pair donation by the amino group is essential. Therefore, desolvation of the solvent molecule hydrogen-bonded to the dimethylamino group is a prerequisite for rotation around the N-N bond and also for the preparation of a solvation shell which stabilizes the highly dipolar activated complex 3.2. Considering the fact that the magnitude of the Eaf is much larger than the normal bond energy of hydrogen bonds, mere desolvation is not enough to explain the energy increase. The slope of the reactant valley must be considerably larger in MPD than in GTA probably as a result of a partial correlation of the solvent and the chemical coordinate. In other words, rotational movement around the nitrogen-nitrogen bond takes place as the hydrogen-bonded solvent molecules are desolvated. On the other hand, the results in GTA demonstrate that solvent reorganization takes place without inducing a major structural change in the Z-isomer. From these results, it would be reasonable to expect that we would also observe a similar isoviscous temperature dependence of kobs and kf in other systems, if the solute-solvent interactions are strong enough. In order to examine this possibility, we decided to undertake kinetic measurements on some carbocyanine cations because solute-solvent interactions are much stronger in charged species than in neutral ones.

3.5

Z/E lsornerization o f Carbocyanine Cations

Carbocyanine cations are known to isomerize photochemically in their polymethinic chain. 3,3’-Diethyloxacarbocyanineiodide (DOCI) and 3,3’-diethyloxadicarbocyanine iodide (DODCI) are among the most extensively investigated compounds and the cations are believed to exist in solution in the all trans extended structure. After irradiation one of the carbon-carbon double bonds undergoes Z/E-type isomerization and unstable photoisomeric cations are formed as illustrated in Schemes 3.3 and 3.4 [45].

3.5 Z/E lsomerization of Carbocyanine Cations

I

115

DOC cation

Scheme 3.3.

DODC cation

Scheme 3.4.

Pressure effects on the rate of these thermal return processes were investigated in ethanol and MPD [46]. The results for DOC1 are shown in Fig. 3.18. The isomerization rate in ethanol clearly demonstrates that the activation volume was close to zero in this solvent. This is in accordance with the Laser-Induced

-1 0

35'C

0

25'C

0

15'C

A

5'C

-2 0

200

400

600

P lMPa

Fig. 3.18. Pressure effects o n the rate of thermal Z/E isornerization of DOC1 in MPD and ethanol at various temperatures.

116

I

3 High-pressure Kinetics and Highly Viscous Media

Optoacoustic Spectroscopy (LIOAS) measurements reported by Braslavsky and her co-workers [45]. As a result of their experiments, the authors concluded that the partial molar volume of the Z-and E-cation were close to each other. Since the reaction is a unimolecular geometrical isomerization, it is hard to imagine that the volume profile of the reaction has an extreme value and it is reasonable to observe a very small pressure effect if the reaction volume AV is close to zero. In MPD, the rate constant itself was close to that in ethanol and the pressure effect at P < 150 MPa was similarly small. The activation volumes were 0 f 1 cm3 mol-I, a clear indication of the validity of TST at lower pressures. Isomerization, however, was strongly suppressed by an increase in pressure at higher pressures. There is little doubt that applying high pressure moves the reaction into the TST-invalid region of high solvent viscosity. The results obtained for DODCI were qualitatively different from those for DOC1 as shown in Fig. 3.19. The reaction was retarded by an increase in pressure even in ethanol. The activation volume obtained by fitting these results into Eq. 3.10 was 5.8 f 0.8 cm3 mol-*. Usually log k - P plots for a reaction with a positive activation volume become concave-up (which means concave on the upper side) because the compressibility increases with the volume of the system. The plots in ethanol met this expectation and the pressure effect must have been the effect of a volume increase during

A

a

35-c

+

15'C

0

25'C

0 0

0 A

,

1.0!AA5'Cl

0.5

0

200

400

600 PlMPa

Fig. 3.19. Pressure effects on the rate of thermal Z/E isomerization of DODCI in MPD (open symbols) and ethanol (closed symbols) at various temperatures.

3.5 Z/E lsomerization of Carbocyanine Cations

activation. The observed moderately positive activation volumes were again in accordance with the results of the LIOAS measurements [45]. According to Braslavsky, the partial molar volume of the DODCI was smaller by 29 f 12 cm3 molk' in the photo-produced unstable conformer. This unexpected result was rationalized on the basis of the increasing solvation (electrostriction) caused by a charge distribution difference in the two conformers. Therefore, the reaction volume must be positive for the present thermal return and the observation of a positive activation volume is not surprising. Hara and Akimoto [ 381 studied the pressure effect on the same reaction in several I-alkanols including ethanol. In all of the solvents, the reaction was moderately retarded by pressure and the activation volume in ethanol calculated from their results was 4.3 cm3 rnol-' at 30 "C, in a rcasonable agreement with our value. However, the authors interpreted the pressure dependence as a dynamic solvent effect, namely the reaction was suppressed by an increase in the solvent viscosity caused by the applied pressure. Similarly, Fleming and his coworkers [37] measured the isomerization rate in a series of 1-alkanols at 0.1 MPa. The reaction was found to be slightly retarded with the increasing chain length (kEtOH/kDecOH = 2) and this effect was also interpreted as viscosity-induced retardation. Judging from the reaction rate (412 s-' at 25 "C and 0.1 MPa) and the different pressure dependence in ethanol and MPD (vide inju), however, we believe that TST remains valid in common 1-alkanols for this reaction. The solvent effects observed by Fleming cannot be rationalized on the basis of the bulk polarity of the solvent because the activated complex is less polar than the Z-isomer and decreasing the solvent polarity is expected to accelerate the reaction by destabilizing the reactant more than the activated complex. One possible factor which might be responsible for the retarding effect of a longer alkyl chain is the internal pressure. The internal pressure is expected to increase with increasing chain length [47, 481 and if the mechanism is the same as that for external pressure, the reaction may be slower in higher alcohols. Since the rate constant in MPD at 0.1 MPa was close to that in ethanol under the same conditions, it may be reasonably supposed that the atmospheric rate constant was, at least, close to that expected from TST. However, the log kobs- P plots were concave-down for MPD, that is, the magnitude of the pressure effect increased with increasing pressure throughout the whole pressure range. We could not detect any inflection point in the log kobs-P plots. Therefore, deviations from TST must have started at a pressure close to 0.1 MPa. The kobs values were plotted against in order to check the isoviscous temperature dependence and are shown in Figs 3.20 and 3.21. Several activation energies Euobs in isoviscous conditions were estimated from the plots in these figures and they are listed in Table 3.3. The isobaric values in ethanol are also given in Table 3.4 for comparison. Obviously, the isoviscous activation energy decreases with increasing viscosity. Since the isobaric activation energy for DODCI in ethanol shows that the height of the energy barrier is little affected by the external pressure, a decrease in the isoviscous Eaobs cannot be attributed to the pressurc dependence of the shape of the potential energy surface. The most reasonable rationalization may be that the

I

117

118

I

3 High-pressure Kinetics a n d Highly Viscous Media

2 n

9

v

o,

-0

0

35'C

0

25'C

0

15'C

A

-00%

0.5-

"00

000000000~

5 'C -5.c

0

OoOO0

0.0-

0 AAAAAAAA

OO 0

A AA 'AA

-0.5 -

AA A A

-1.o -

A A

-1.5

I 0

-2

2

5

8

log(q/Pa s) Fig. 3.20. Plots of log kobs against log q for the thermal Z/E isomerization of DOC1 in MPD a t various temperatures.

1

3.0

A

A A

0

-

A A

o

o

0

A A

0 A

15'C

A

0

0 Y

0

OO

0

5'C

A

o

0

3.5 Z/E lsornerization of Carbocyanine Cations Tab. 3.3. Isoviscous activation energies calculated from kobs (Ea,b,/kJ Z/E isornerization o f DOC1 and DODCI in MPD.

v/Pa

*

!?Gobs (DOCI)

Eaobs (DODCI)

mol-’) for the thermal

0. I

I

10

102

103

I 04

67.2 52.7

67.6 49.3

66.5 45.8

63.7 42.2

58.8

51.6

rate-determining step shifts from the point of crossing the energy barrier at low pressures/viscosities to the point of solvent reorganization around the reactant at high pressures/viscosities. If this rationalization is correct, the kf values obtained on the basis of Eq. 3.4 would give linear isobaric Arrhenius plots and obey Eq. 3.9. Since the TST-valid pressure range was wide enough in the case of DOCI, the observed rate constant kobs was extrapolated to estimate the kTsT values. However, similar extrapolation could not be performed for DODCI because the TST became invalid at pressures slightly above 0.1 MPa. Therefore, we decided to assume that the static pressure effect in MPD was identical to that observed in ethanol. The rate constants in ethanol were fitted to Eq. 3.10 and the parameters a , b, and c were determined. The kTsT under the TST-invalid conditions were estimated by applying Eq. 3.10 for ethanol to the rate constant in MPD at atmospheric pressure. The kf values obtained gave linear isobaric Arrhenius plots and they were inversely proportional to a fractional power of q. The log kf-log q plots are shown in Figs 3.22 and 3.23. The plots were satisfactorily linear and the slopes were smaller than unity. The linearity was slightly poorer in DODCI probably because k T s l values estimated on the basis of the rate constants in ethanol were used. It is obvious that kf shows temperature dependence even at the same viscosity. The isoviscous activation energies derived from kf are listed in Table 3.5 together with those derived from kTST. Judging from the E ~ ~ / E u Tratios, ~ T the correlation between the solvent and chemical coordinate in the isomerization of the carbocyanine cations was of the same order as that in the azobenzene isomerization in MPD. Considering the significant difference in the isomerization rate and in the pressure range where dynamic solvent effects were observed for the two carbocyanines, it is interesting to examine whether the dynamic solvent effect appears at a similar level in the relative time scales of the solvent and the chemical coordinate. It has earlier been shown [20, 211 that the product kTsT x q can be used in the AH model to describe the balance of the characteristic times of the reaction system and the solvent: the greater the product, the higher the anisotropy of the two time scales as

Tab. 3.4. Isobaric activation energies calculated from kobs (Ea,b,/kJ isomerization o f DODCI in ethanol.

mol-’) for the thermal Z/E

P/ MPa

0.1

150

300

450

600

Eaabs

58.9

59.6

59.9

60.0

60.4

I

119

120

I

3 High-Pressure Kinetics and Highly Viscous Media

3

-

. 5

7

0

35'C

0

25 'C 15 'C

2-

in

5 'C

m

-0

-5'C

1-

0-

-1 -

A

!

I

1

I

I

-2

0

2

4

6

-2

8

log (q/Pa s)

Fig. 3.22. Viscosity dependence of kf in the thermal Z/E isomerization of DOC1 in MPD.

3.5j

00

0

A 7

-r?

5 m

0

0

0

0

o

3.0'

-0

A.

2.5 -

n

D o 0 0 o n

0

O

A 0

O n

0

0 0 0

A

2.0 -

o A

0

0

o

0

0

35 'C 25 'C

1.5-

-1

0

0

15 'C

A

5 'C

0

A

0

A

o A

1

2

3

4

log( q/Pa s ) Fig. 3.23.

Viscosity dependence of kf in the thermal Z / E isomerization of DODCI in MPD

3.5 Z/E lsornerization of Carbocyanine Cations lsoviscous activation energies calculated from kf (€of/kJ mol-') and kTs, (EOTST/kJ mol-') for the thermal Z / € isornerization of DOC1 and DODCI in MPD.

Tab. 3.5.

1 10

lo2 10' 104

41 f 1.0 41 & 1.0 40 f 0.9 39 f 0.7

69 k 0.8 70 k 1.0 71 2 1.2 72 f 1.4

45 f 3.7 42 k 3.2 43 k 2.4 44 f 2.1

*

51 0.2 49 f 0.1 46 k 0.1 44 f 0.1

a result of the slower movement of the solvent. The plots of log(kobs/kTST)against log(kTsT ' 1 7 ) are shown in Fig. 3.24. As expected, the results obtained at different temperatures fall on the same curve when plotted in these coordinates. The quality of aggregation was higher for DOCI than for DODCI probably for the same reason that the linearity was improved in the log kf-log q plot for DOCI. It is interesting to note that the points for the two compounds fall very close to each other despite much faster isomerization and much earlier deviations from TST in DODCI. This fact strongly suggests that the extent of deviation from TST is governed mainly by the relative time scale of the solvent and

0.0

-t;;

-0.2

s; a Yo

3

-0

-0.4

-0.6

I

-0.8

A 35 T

0

-1.o

+

25'C

0 0

15'C

0

A A -1.2

0

A

5 'C I

I

I

I

I

1

2

3

4

5

lo!"Ts+')(dPa

41

Dependence o f log(k,b,/kTsr) on log(kTST . q) in the isomerization of DOC1 (open symbols) and DODCI (closed symbols) a t various temperatures. Fig. 3.24.

6

I

121

122

I

3 High-Pressure Kinetics and Highly Viscous Media

-+

< 0

<

-0.2-

n

P,

0 -

-0.4-

-0.6 -

ABm

0

-0.8-1 -

A

DBNA DBEA

0

DBBA

A

DOC1

m

A

A

DODCl ..L

I

I

I

I

0

2

4

6

I

log( kTST/s-')(v/s-') Fig. 3.25. Dependence of Iog(kobs/kTST) on log(kTST . q) in the thermal Z/E isomerization of N-benzylideneanilines and carbocyanine cations at various temperatures in MPD.

the chemical coordinate in closely related systems. This point can be seen even more clearly in Fig. 3.25 in which the points for the three N-benzylideneanilinesat various temperatures are added to Fig. 3.24. The relative rates at 0 "C and 0.1 MPa were DBBA: DBEA: DBNA = 1: 56: 1600. However, the points added cluster in a higher-anisotropy area than the carbocyanines. Similar aggregation of the points for the three N-benzylideneanilines was also observed in GTA and DCMP. Figure 3.26 shows the plots in DCMP. It may be safely concluded from these plots that the magnitude of a deviation from TST in analogous reactions is determined mostly by the relative time scale of the reaction and the solvent fluctuations. Furthermore, since the plots in Figs 3.25 and 3.26 are similar to the results of numerical simulations [20], they may be taken as evidence of the validity of the model.

3.6

Z/E lsomerization of DNAB in a Polymeric Medium Throughout our analyses, macroscopic solvent shear viscosity q has been assumed to be proportional to the microscopic friction between the reactant and the solvent

3.G Z/E lsomerization of DNAB in a Polymeric Medium

I

123

+

P -o.2-

21

X"

v

-

-0.4-

-0.6-

-0.80

DBNA

0

DBEA

-1 -

0

DBBA

0

-1.2

3

I

I

I

I

4

5

6

7

I

8

I

9

'

logf(kTsT/s-')(rl/Pas)l Fig. 3.26. Dependence of Iog(kabr/kTST) on log(kTST . 17) in the thermal Z/E isomerization of N-benzylideneanilines at various temperatures in DCMP.

during the crossing of the energy barrier. This assumption has some experimental support. For example, rotational relaxation times of neutral as well as charged species are reported to be proportional to 7 [49, SO]. In polymer media, however, it is expected that the macroscopic viscosity may not adequately describe the microscopic frictions and if this is the case, it may be considered to be indirect evidence for the proportionality between the macroscopic viscosity and the microscopic friction in monomeric solvents. In order to pursue this possibility, we measured kinetic pressure effects on the isomerization of DNAB in a commercially available silicone oil [51]. The oil used was KF-54 from Shin-Etsu Chemical and its structure is given below. The average degree of polymerization was 25 and the phenyl :methyl ratio was 1:3. ye

ph

-0-sio-phe

Ph

The pressure dependence of the viscosity of KF-54 obeyed Eq. 3.6. The rate constants observed in GTA and KF-54 at various temperatures and pressures are illustrated in Fig. 3.27.

124

I

3 High-pressure Kinetics and Highly Viscous Media

A

A

0 0

A A

15 5'C

10

It was observed that the reaction deviated from the first-order rate law at high pressures and the measurements were restricted to the pressure range in which the first-order rate law was obeyed. The results in KF-54 look similar to those in GTA. The reaction was accelerated at low pressures because of its negative activation volume. The phenyl group in the polymer chain stabilizes the rotational transition state as does benzene [41]. The reaction rate reached a maximum and then started to decrease with increasing pressure, a clear indication of a shift of the reaction from the TST-valid to the TST-invalid region. It is difficult to judge from Fig. 3.27 whether the polymeric solvent behaves differently from monomeric GTA. However, the difference can be seen clearly by plotting log kobs against log r] as shown in Fig. 3.28. The viscosity which yields the maximum reaction rate, shifts to a lower value with increasing temperature in GTA. In KF-54, however, the maximum viscosity stays almost constant. This behavior suggests the existence of a factor which affects the reaction rate but does not contribute to the macroscopic viscosity. If this second contribution to the microscopic friction decreases with increasing temperature, the behavior shown in Fig. 3.28 is to be expected. This second factor may be related to local segmental rotation of the siloxane chain. However>further experiments are necessary for detailed analysis. At any rate, the results in Fig. 3.28 may be considered as indirect evidence for the validity of our use of q as a measure of microscopic friction.

3.7 Concluding Remarks

I

125

2 0

40'C

0 4

25'C

.

0 0

15'C

so

A A

5'C

h 7

i n 1 n

v

m

0 -

0 A

-1

A

-2

-3 -2

0

2

4

6

8

10

log( q/Pa s) Fig. 3.28. Plots of log kobr against log 7 for the thermal Z/E isomerization of DNAB in CTA (open symbols) and KF-54 (closed symbols) a t various temperatures.

3.7

Concluding Remarks

Kinetics in highly viscous media is a new field of high-pressure kinetics and it gives us information that cannot be obtained by conventional activation volume measurements. From the experiments performed so far, it has now been unequivocally demonstrated that the simple KGH model which uses only one reaction coordinate does not satisfactorily explain the experimental observations, at least in slow unimolecular reactions, of molecules in their So-state. In those reactions, solvent reorganization does not necessarily result in a chemical transformation along the reaction coordinate. When the solvation of the reactant is relatively weak, the solvent molecules are rearranged prior to the chemical transformations to form a solvation shell which stabilizes the activated complex. As demonstrated by the reaction of push-pull substituted azobenzenes in GTA, this independence of the solvent coordinate remains true even if the activated complex is much more polar than the reactant. However, when specific desolvation is required in the reaction mechanism (DNAB and DMNAB in MPD) or the solvation in the initial state is strong (DOC1 and DODCI in MPD), the two coordinates are more or less correlated and the solvent molecules are rearranged together with chemical transformations. Even in such reactions, the energy increase during solvent rearrange-

126

I

3 High-Pressure Kinetics and Highly Viscous Media

mcnt does not reach the value of the activation energy observed under conditions where TST is valid. Namely the reactant goes up to the middle of the energy barrier and if further energy can be obtained from nearby solvent molecules, a large vibration along the reaction coordinate brings the molecule to the product energy surface in a “frozen” solvation shell. It is highly desirable to analyze the experimental results in more detail on the basis of the theoretical model. If the experimental results are correlated with the model quantitatively, we will have a greater understanding of the reaction system. Research along these lines is now in progress [52]. It will also be very fruitful, if calculations using molecular dynamics combined with quantum chemistry can be carried out for the present thermal isomerizations, as was the case for the photoisomerization of stilbene [ 5 3 ] .

Acknowledgments

The author is grateful for financial support from the Ministry of Education, Science, Sports and Culture of Japan, Japan Society for the Promotion of Science, and Nishida Research Fund for Fundamental Organic Chemistry. He is also grateful to coworkers in Oita University, especially to Professor Y. Ohga and Mr T. Takahashi. Helpful comments by theoretical scientists, Professors N. Weinberg (University College of the Frazer Valley), M. V. Basilevsky (Karpov Institute of Physical Chemistry) and H. Sumi (University of Tsukuba) are also greatly appreciated.

References 1 S. GLASSTONE, K. J. LAIDLER, H .

2

3 4

5

6 7

8

EYRING, The Theory of Rate Processes, McGraw-Hill, New York, 1941. S. D. HAMANN,Physico-Chemical Efects of Pressure, Buttenvorths, London, 1957. T. ASANO,W. J. LE NOBLE,Chem. Rev. 1978, 78, 407. R. VAN ELDIK,T. ASANO,W. J. LE NOBLE,Chem. Rev. 1989, 89, 549. A. DRLJACA, C. D. HUBBARD,R. VAN ELDIK,T. ASANO,M. V. BASILEVSKY, W. J. LE NOBLE,Chem. Rev. 1998, 98, 2167. H . A. KRAMERS, Physica, 1940, 7, 284. K. STEPHAN,K. LUCAS, Viscosity of Dense Fluids, Plenum, New York, 1979. j. TROE,Ber. Bunsenges. Phys. Chem. 1991, 95, 228.

9 As a review, J. SCHRODER, J. TROE, 10 11

12

13

14

15 16

Ann. Rev. Phys. Chem. 1987, 38, 163. D. M. ZEGLINSKI, D. H . W A L D E C K , ~ . Phys. Chem. 1988, 92, 692. N. S. PARK,D. H. W A L D E C KPhys. ,~. Chem. 1990, 94, 662. G. ROTHENBERGER, D. K. NEGUS, R. M. HOCHSTRASSER, J . Chem. Phys. 1983, 79, 5360. G. R. FLEMING, S. H . COURTNEY and M. W. BALK,]. Stat. Phys. 1986, 42, 83. M. LEE, G. R. HOLTOM,R. M. HOCHSTRASSER, Chem. Phys. Lett. 1985, 118, 359. As a review, J. T. HYNES,J . Stat. Phys. 1986, 42, 147. N. AGMON,J. J. HOPFIELD,]. Chem. Phys. 1983, 78, 6947.

AGMON, I. J. HOPFIELD,].Chem. Phys. 1983, 79, 2042. 18 H. SUMI,R. A. MARCUS, J . Chem. Phys. 1986. 84, 4894. 19 H. S U M I J, . Phys. Chem. 1991, 95, 3334. 20 M. V. BASILEVSKY, V. M. RYABOY,N. N. WEINBERG, J . Phys. Chem. 1990, 94, 8734. 21 M. V. BASILEVSKY, V. M. RYABOY, N.N. WEINBERG, J . Phys. Chem. 1991, 95, 5553. 22 Using a series of solvents with the same functional group is sometimes adopted as a measure to change the viscosity as discussed above [ 10-121. The viscosity range that can be covered by this method, however, is rather narrow. For example the viscosity of 1-octanol is barely 15 times larger than that of methanol. It is possible to increase the viscosity much more significantly if the alkyl chain is branched or if the number of the functional groups is increased However, any drastic change in the molecular structure can possibly alter the shape of the potential energy surface and it is very difficult to estimate the magnitude of this effect. Furthermore, it is not possible to change the viscosity continuously by changing the solvent. 23 T. ASANO,T. OKADA, ]. Phys. Chem. 1984, 88, 238. 24 For a case of a deviation from the firstorder rate law, see Sect. 3.6. 25 Standard Viscosity-Temperature Charts for Liquid Petroleum Products, D341-74, Part 23, Annual Book of ASTM Standards, ASTM, Philadelphia, 1985. 26 The high pressure viscosity of GTA showed a small upward deviation from Eq. 3.6. In order to consider this non-linearity, an empirical equation proposed by Yasutomi and his coworkers [27] was tested. Unfortunately, however, the function obtained resulted in unreasonably large high-pressure viscosities at low temperatures and, therefore, Eq. 3.6 was also adopted for GTA. 27 s. YASUTOMI,S . BAIR,W. 0. WINTER, Trans. ASME,J . Trib. 1984, 106, 291. 17 N.

28

29

30

31

32 33

34 35

36

37 38 39 40

41

42 43

T. ASANO,H. FURUTA, H.-J. R. CIMIRAGLIA, Y. TSUNO, HOFMANN, M. FUJIO,J . Org. Chem. 1993, 58,4418 and earlier papers cited therein. P. HARBERFIELD, P. M. BLOCK,M. S . Lux, J . Am. Chem. SOC.1975, 97, 5804. T. ASANO,T. OKADA, W. G. HERKSTROETER. J . Org. Chem. 1989, 54, 379. The reason for the small rate difference between the two solvents cannot be rationalized at the present level of knowledge. However, the fact that the rate in GTA was larger than that in AcOMe also supports the absence of dynamic solvent effects in GTA at low pressures. T. ASANO, H. FURUTA,H. SUMI,J . Am. Chem. SOC.1994, 116, 5545. T. ASANO,K. COSSTICK, H. FURUTA,K. MATSUO,H. SUMI,Bull. Chem. SOC. Jpn. 1396, 69, 551. T. ASANO, K. MATSUO,H. SUMI,Bull. Chem. SOC.Jpn, 1977, 70, 239. Several authors [ 36-38] have measured the pressure/solvent dependency of some unimolecular rate processes of ground-state molecules and all of the authors have interpreted the results as “dynamic” solvent effects. However, as a result of our own measurements described in this chapter, we believe that what they observed was “static” pressure/solvent effects. See also Sect. 3.5. J. JONAS, X. PENG,Ber. Bunsenges. Phys. Chem. 1991, 95, 243 and earlier papers cited therein. S. P. VELSKO,D. H. WALDECK, G. R. FLEMING,]. Chem. Phys. 1983, 78, 249. K. HAM,S. AKIMOTO,J . Phys. Chem. 1991, 95, 5811. See Tables 5 and 7 in ref. 5. R. CIMIRAGLIA, T. ASANO, H.-J. Gazz. Chim. Ital. 1996, HOFMANN, 126, 679. T. ASANO, T. OKADA,].Org. Chem. 1986, 51, 4454 and earlier papers cited therein. D.-M. SHIN, D. G. WHITTEN, J . Am. Chem. SOC.1988, 110, 5206. Measurements in DCMP were not carried out because the reaction

128

I

3 High-pressure Kinetics and Highly Viscous Media

proceeds via nitrogen inversion in non-polar media. 44 C. A. N. VIANA,J. C. R. REIS,Pure Appl. Chem. 1996, 68, 1541. 45 M. S. CIIURIO, K. P. ANCERMUND, S. E. BRASLAVSKY, J. Phys. Chem. 1994, 98, 1776 and refrences cited therein. 46 J. C. K I M , Y. OHGA,T. ASANO,N. WEINBERG, A. V. GEORGE, Bull. Chem. SOC.Jpn. 2001, 74, 103-111. 47 G . ALLEN,G. GEE,G. 1. WILSON, Polymer 1960, I , 456. 48 M. R. 1. DACK,Chem. SOC.Rev. 1975, 4, 211. 49 N. ITO, 0. KAJIMOTO, K . HAM, Chem. Phys. Lett. 2000, 318, 118.

G . B. D u n , M. I<. SIN G H A. , V. SAPRE, J . Chem. Phys. 1998: 109. 5994. 51 A. S H U K , T. TAKAHASHI. Y. OHGA,T. ASANO,H. SAITO,K. MATSUO,H.-D. L ~ D E M A N2. N ,Naturforrch. 2000. 55a, 616. 52 Y. OHGA,T. ASANO.M. V. BASILEVSKY, N. WEINBERG, 41st High-Pressure Symposium of japan, Kashiwa, Japan, 6-8 November 2000, Abstract 2C09. 53 C. D. BERWEGER, W. F. VAN GUNSTEREN, F. MBLLER-PLATHE, Angew. Chem. Int. Ed. 1999, 38. 2609. 50

II

Mechanistic and Synthetic Applications o f High Pressure

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

4

Water Exchange on Metal Ions: The Effect o f Pressure Lothar Helm a n d Andre E. Merbach*

4.1

Introduction

The solvent exchange reaction between the first and second coordination shell around a metal ion in solution is the simplest ligand substitution reaction. Environmental issues and the move towards “clean technology” is driving synthetic chemists away from organic-based solvent systems and towards water as the preferred solvent. Furthermore, water is the preferred solvent of nature, and therefore an understanding of the reactivity of hydrated metal ions is of fundamental importance for the understanding of many reactions in chemical and biological systems (Eq. 4.1). The replacement of a water molecule from the first coordination shell is an important step in complex-formationreactions of metal cations as well as in many redox processes. The general chemistry of aqua ions over the Periodic Table of elements, their synthesis, structure and reactivity have been reviewed recently by Richens [ 11. [M(HzO),]‘+

+ nHzO* * [M(HzO*)n]Zf+ nHzO

In water exchange reactions, the reagent and the product are identical, as are the ligand and the solvent: water. The Gibbs free energy change, AG, of the reaction is zero. The observed exchange rate constants cover more than 18 orders of magnitude (Fig. 4.1) from the most labile monovalent alkaline ions to the very inert trivalent transition metal ions Rh3+and Ir3+ [ 2 ] . On the fast exchange side, the mean lifetime of a water molecule in the first coordination shell of Cs+ was estimated to 200 ps [ 31. Cs’ is the largest monovalent metal cation and has therefore the lowest surface charge density. At the other extreme, a water molecule stays for more than 300 years in the first coordination shell of hexaaqua iridium(III), a third-row transition metal ion with a large ligand field contribution to the activation energy. The metal ions are often classified into three groups. The first group is represented by the main group metal ions, varying mainly in ionic radius and electric charge. The number of water molecules in the first coordination sphere range from 4 to 10 or greater 141.The second group consists of the d-transition metal ions, all

132

I

4 Water Exchange on Metal lons: The Effect of Pressure

<

=n,o1 s

n nin

Li'

1010

lo8

lo6

104

lo2

100 102 kn20 J s '

104

lo6

lo8

1 0 ~ ~ >

Fig. 4.1. Mean lifetimes o f a single water molecule in the first coordination sphere of a given metal ion, T H , O , and the corresponding water exchange rate constants, k H 2 0 . The filled bars indicate directly determined values, and the open bars indicate values deduced from ligand substitution studies.

hexa-coordinated with the exception of PdZ+and PtZ+which are square planar and Sc3+ that might be hepta-coordinated (51. Their exchange rate constants are very strongly linked to the occupancy of their d-orbitals. The third group involves the lanthanide and actinide ions which are eight- or nine-coordinated. Their kinetic behavior is mostly influenced by the decrease in ionic radius along the series and the subsequent change in the coordination number. Experimentally, water exchange rate constants are mainly determined from nuclear magnetic resonance measurements [ 6, 71. Other techniques are restricted to very slow reactions (classical kinetic methods using isotopic substitution) or are indirect methods, such as ultrasound absorption, where the rate constants are estimated from complex-formationreactions with sulfate [ 31. The microscopic nature of the mechanism of the exchange reaction is not directly accessible by experimental methods. In general, reaction mechanisms can be deduced by experimentally testing the sensitivity of the reaction rate to a variety of chemical and physical parameters such as temperature, pressure, or concentration. In recent years the application of classical molecular dynamics or Monte Carlo simulations as well as ab initio and DFT calculations have gained significant interest. The performance of the new generation of computers and theoretical approaches now allows calculations, which can provide deeper microscopic insight

4.2 Concepts of Solvent Exchange Reactions

into water exchange processes, to be undertaken. Classical simulations of hydrated metal ions in solution mainly rely on painvise additive, effective interaction potentials obtained from quantum mechanical calculations or from empirical optimization of thermodynamic values like hydration energies or radial distribution functions. This approach limits these simulations to metal ions with filled atomic orbitals, mainly to main group ions and those transition metal ions, where ligand field effects can be neglected (e.g. lanthanides). The advantage of the classical methods is that relatively large systems, containing up to lo3 water molecules can s). Exchange processes can be followed directly be simulated over several ns ( by inspecting trajectories, molecular coordinates and velocities as a function of time. Full quantum mechanical calculations using ab initio self-consistent field (SCF) methods or density functional theory (DFT) are computationally much more demanding and, therefore, mostly restricted to ion-water clusters containing one metal ion and a limited number of water molecules. Different metal-water clusters that can occur along the reaction coordinate of a water exchange process have to be calculated. A comparison of energies and structural parameters such as metaloxygen bond distances in the reactants, transition states and intermediates will help to associate them to different reaction mechanisms. In gas-phase calculations, second coordination sphere, bulk solvent and anions are neglected and, furthermore, contrary to classical molecular dynamics simulations, no real time scale is provided. On the other hand, these calculations are not based on empirical pairwise interaction potentials, in fact calculations on all types of ions, including transition metals, are feasible. A promising method, developed in recent years, is the use of first principles molecular dynamics as exemplified by the Car-Parrinello technique [ 81. In these calculations the interatomic potentials are explicitly derived from the electronic groundstate within the density functional theory in local or non-local approximation. It combines quantum mechanical calculations with molecular dynamics simulations and, therefore, overcomes the limitations of both methods. Actual computers allow only simulations of aqueous solutions of about GO water molecules for several ps ( s). This limit is still at least one order of magnitude shorter than the fastest directly measured water exchange rate, k = 3.5 x lo’ s-l for [Eu(H20)8]*+,i.e. one exchange event every (8 x 3.5 x lo’ s-’)-’ = 36 ps [91. Nevertheless, several publications appeared in the late 1990s on solvated Be2+ [lo], K+ 1111 and Cu2+ [12] presenting mainly structural results. This chapter is essentially an update of our review which appeared two years ago 1131.

4.2

Concepts of Solvent Exchange Reactions

The mechanistic classification accepted for ligand substitution reactions was proposed by Langford and Gray in 1965 [14]. They divided substitution reactions into

I

133

134

I

4 Water Exchange o n M e t a l Ions: The Effect of Pressure

three categories of stoichiometric mechanisms: associative (A) where an intermediate of increased coordination number can be detected, dissociative (D) where an intermediate of reduced coordination number can be detected, and interchange ( I ) Y repwhere there is no kinetically detectable intermediate (Eqs 4.2-4.4). MX, resents an outer-sphere complex. ' '

-X

MX,

+X

+Y

MX,,-l

+Y

(D = dissociative)

(4.2)

(A = associative)

(4.3)

(I = interchange)

(4.4)

-X

MX, F'tY MX,Y MX,

+ MX,-lY

-Y

$

tX

MX,-IY

. . . Y $ MX,_lY.

. .X

Furthermore, they distinguished two major categories of intimate mechanisms: those with an associative activation mode (a), where the reaction rate is approximately as sensitive to variation of the entering group as to variation of the leaving group, and those with a dissociative mode (d), where the reaction rate is much more sensitive to the variation of the leaving group than to the variation of the entering group (Fig. 4.2). Evidently all D mechanisms must be dissociatively and all A mechanisms must be associatively activated. The interchange ( I ) mechanisms include a continuous spectrum of transition states where the degree of bond-making between the incoming ligand and the complex ranges from very substantial (I, mechanism) to negligible ( I d mechanism) [15, 161. For a solvent exchange process, the forward and backward reaction coordinates must be symmetrical. Thus, for an Id mechanism, with weak bond-making of the entering group, the leaving group is also necessarily weakly bound. Inversely, for an I, mechanism, both the entering and the leaving group must have considerable bonding to the metal (Fig. 4.3) [15]. As mentioned above, the proposition of a reaction mechanism is mainly based on the response of the reaction rate to the variation of chemical and physical pa-

evidence for an intermediate of lower coordination number stoichiometric mechanism

evidence for an intermediate of higher coorinatiion number

no evidence for an intermediate

I

D

A

intimate mechanism

A I

I

rate independent of nature of entering

I

1

rate to nature of entering

Fig. 4.2.

Classification o f solvent substitution reactions.

I I

rate dependent of nature of entering group

4.2 Concepts of Solvent Exchange Reactions

I

135

hlechanism

Rate constant

Reaction Coordinate

Metal ion with first coordination sphere Leaving solvent molecule

@

Entering solvent molecule

Fig. 4.3. Representation o f the transition state for the spectrum of solvent o r symmetrical ligand exchange processes,

rameters. The volume of activation, AV# , has become the main tool for the experimental identification of the water exchange mechanism [17-191. AVf is defined as the difference between the partial molar volume of the reactants and that of the transition state at a temperature ?; as shown in Eq. 4.5.

The observed exchange rate, k, is either slowed down or accelerated by increasing pressure, depending on the positive or negative sign of AVz, respectively. It is usually accepted that the activation volume has two contributions: an intrinsic component, AVi$ resulting from changes in bond lengths and bond angles, and an electrostrictive component, AV&, which arises from changes in electrostriction of the solvent outside the first coordination sphere. For solvent exchange reactions differences in electrostriction between the transition state and the reactant can be

136

I

4 Water Exchange on Metal Ions: The Eflect offressure

neglected, hence AVz E AV,Zt. Consequently, the activation volume will be a direct measure of the degree of bond formation or bond breaking in the transition state. The relation between the pressure-induced changes of the observed exchange rates and the underlying solvent exchange reaction mechanisms is shown in Fig. 4.3. An important issue is the prediction of activation volumes for the limiting substitution mechanisms D and A. Swaddle [20, 211 developed a semi-empirical model in which the absolute partial molar volume, AV&, of a hydrated metal ion in water is related to its ionic radius, depending on the coordination number, and the charge of the metal ion. Taking ionic radii from Shannon, this model gives similar absolute limiting values of IAV’I = 13.5 cm3 mol-’ for D and A processes on octahedral di- and trivalent 3d transition metal ions. An extensive compilation of activation and reaction volumes of organic and inorganic reactions in solution can be found in a series of reviews by Asano et al. [ 221.

4.3

Water Exchange from the First Coordination Shell 4.3.1

First Row Transition Metal Ions

Water exchange on first row di- and trivalent transition metal ions has been the subject of extensive experimental studies and has been widely reviewed [IS, 231. Table 4.1 summarizes the experimental first order rate constants and the activation parameters. The sequence of reactivity of the diualent first-row transition metal ions does not correlate with the radii of the ions and is largely independent of the reaction mechanism. The rate constants are sensitive to the electron configuration and semiquantitatively coincide with predictions based on ligand field activation energies and molecular orbital calculations [23, 33, 341. The water exchange mechanism progressively changes from I, to Id as the number of d-electrons increases and the ionic radius decreases. This change is demonstrated most evidently by the change in sign of the activation volume, AVz. The observed progressive mechanistic changeover cannot be explained only in terms of cationic size, the electronic configuration also plays an important role. For a a-bonded octahedral complex, the t2g orbitals are non-bonding, whereas the eg orbitals are anti-bonding. The gradual filling of the t2gorbitals, spread out between ligands, will electrostatically disfavor the approach of a seventh molecule towards a face of the octahedron, and therefore decrease the ease of bond making. Similarly, an increased occupancy of the eg orbitals, directed to the ligands, will enhance the bond breaking tendency. These effects, combined with the steric effects outlined above, explain the sequence of the AV# values. A special case is the very labile hydrated copper ion. Because of its electronic dg configuration, a Jahn-Teller distortion is generally assumed to elongate axially two bonds of the octahedral [ C U ( H ~ O ) ~complex ] ~ + in solution [4, 31, 351 by anal-

M3+ rM (pm) k298 t -1 \s 1 AHZ (kJ mol-I) AS# (JK-' mol ') A V Z (cm3mol-') Mechanism Ref.

64 5.0 x 102 49.0 -27.8 8.9

1,

27

A, 1,

26

v3+

3dZ

Ti3' 67 1.8 x 105 43.4 +1.2 -12.1

3d'

28

1,

Cr3+ 61 2.4 x 10 108.6 +11.5 -9.6

79 8.7 x 10' 61.8 -0.4 -4.1 1, 24

V2+

3d'

Mn3' 64

cr2' 80

3d4

Id

25

1,

25 Fe3+ 64 1.6 x l o 2 64.0 +12.1 -5.4 1, 29, 30

78 4.4 x 106 41.4 721.2 +3.8

Fe2+

3d6

83 2.1 x 107 32.9 +5.7 -5.4

Mn2+

3dS

Id

25

25

Ni2+ 69 3.2 x 104 56.9 f32.0 +7.2

3d8

Id

74 3.2 x lo6 46.9 +37.2 +6.1

C02'

3d7

Rate constants and activation parameters for water exchange on first row transition metal ions [ M ( H 2 0 ) 6 ] " +

M2+ rM (pm) k2" (s-') AHf (kJ mol-I) AS# (1 K-' mol I ) AV# (cm3mol-I) Mechanism Ref.

Tab. 4.1.

73 5.7 x 109 11.6 -19.2 +2.0 I 12, 31

cu2+a

3d9

32

Id

'+

Ga 62 4.0 x l o 2 67.1 +30.1 -5.0

Zn2' 74

3d'O

a

3.

P

138

I

4 Water Exchange on Metal lons. The Effect of Pressure

ogy with its X-ray structure [3G]. The tetragonal distortion reorients randomly very rapidly so that the lifetime of a given distortion (51 5.1 x 10- l 2 s at 298 K) is much shorter than the lifetime of a given coordinated HzO molecule (rH20 = 2.3 x lo-'' s at 298 K). The longer axial bonds facilitate water exchange through a d-activation mode which is consistent with the small positive activation volume (Table 4.1) [37]. Pasquarello et al. [12] have shown recently by a combined experimental neutron diffraction and a theoretical first-principles molecular dynamics study that there is strong evidence for a five-fold coordination of the Cu(I1) aqua ion in aqueous solution. The simulation revealed that the [Cu(H20)5I2+complex undergoes frequent transformations between square pyramidal and trigonal bipyramidal configurations which weaken Cu-water bonds. The ''0 NMR data [31] can successfully be re-interpreted in terms of five-fold coordinated copper with interconversion (rl 5 ps) between square pyramidal and trigonal bipyramidal forms, yielding a water exchange rate of (5.7 0.2) x 109 s-'. The small positive activation volume of f2.0 f 1.5 cm3 mol-' is also fully compatible with an interchange process for water exchange on [Cu(H2O)5I2+. Water exchange on trivalent first-row transition metal ions shows a similar mechanistic behavior as that of the divalent analogues. [Cr(H20)6I3+is the most inert (Table 4.1) as a consequence of its large ligand field activation energy reflected in the large AH1 value. AVi decreases from -12.1 cm3 mol-' ([Ti(H20)6]3+)to -5.4 cm3 mo1-l ([Fe(HzO)6I3+). The activation volume for Ti3+ approaches the limiting value of -13.5 cm3 mol for an A mechanism [20], hence the operation of an A mechanism can be assigned. As the occupancy of the t2g orbitals and in parallel the electronic repulsion towards an entering ligand increase, the activation volumes become less negative, indicating a changeover to less associative, and finally a dissociative activation mode (for Ga3+)as a consequence of the filling of the o anti-bonding e l orbitals. A general problem arising with trivalent metal ions is the occurrence of the conjugate base species, [M(H20)50HI2+, already kinetically active at quite high [H+] concentrations. In all cases studied (M = Cr, Fe, Ga) the lability of the coordinated water is highly increased and the activation volumes are all positive (f2.7, +7.0, +G.2 cm3 mol-' for [Cr(H2O)5OHl2+,[Fe(Hz0)50HI2+, and [Ga(H20)5OHI2+,respectively, Table 4.2). It is worth noting that the associative activation mode found for [Fe(H20)6I3- changes to a dissociative one for [Fe(H20)50H]2 t . In recent years, several authors published ab initio or DFT calculations on gasphase hydrated metal ions. The main goal of the studies was to obtain microscopic insight into water exchange on octahedral complexes. Rotzinger [ 40-431 carried out calculations on di- and trivalent first-row transition metal ions from Sc3+ to Zn2+ using ab initio methods at the Hartree-Fock or CAS-SCF level. All pertinent transition states and intermediates that could occur along the reaction coordinate of an associative, a concerted and a dissociative water exchange process were calculated. In his study, reactants, products, and intermediates were characterized throughout by the absence of any imaginary vibrational frequency, whereas transition states always show a single imaginary frequency [40]. As an approximation, he neglected in the first studies second sphere and bulk water as well as anions. The 1

-

kobs = kl i kz/[H';

[M(H20)50H12+

= kl

rM

(pm)

-1.1 4.1 1.8x 10-4

lo-'

1.4 x

k:98a (m s-l)

AV:

28

Ref

T

ko~K~/iH+].

Mechanism

-3.8 12.7 I

(s-')

AVZ ( c m 3 mol-') AV;,, ( c m 3 mol-')

kOH

P K

( c m 3 mol-l)

Mechanism

AVf ( c m 3 mol-')

61 2.4 x -9.6 1,

k,2g8 (s-')

cr (111)

29, 30

Id

f7.0

f7.8 2.9 1.2 x 105 +0.8

1.2 x 103

1.6 x lo2 -5.4 1,

64

Fe(I11)

39

I

I 38

f1.2 3.5 4.2 x -0.2 f1.5

1.5 x

lo-$

2.2 x 10-9 -4.1 1,

66.5

Rh(111)

-2.1 2.7 5.9 x 10-4 -3.0 +0.9

1.1x 10-6

1,

3.5 x 10-6 -8.3

68

Ru(l1l)

5.7

0.2 4.45 5.6 10-7 1.5 -t 1.3 1 2 ~

1.4 x lo-"

1,

1.1 x 10-10

68

Ir(111)

Rate constants and activation volumes for water exchange o n trivalent hexa-aqua transition metal ions and their conjugate base species.

[M(Hro)6I3+

Tab. 4.2.

32

Id

+6.2

+l.S

f7.8 3.9 1.1x 105

1.4 x 10'

Id

+5.0

4.0 x lo2

62

Ca(lf1)

W

4

-...

VI

z -

6-

4

a 9'

5 6

3.

2

%

S

is-?

Y

$

lu

P

140

I

4 Water Exchange on Metal Ions: The Effect of Pressure

301+m I. mechansim lV(OH2)612'

A mechansim [Ti(0H2)d3'

D or (IJ mechansim "i(OH2)d2'

E A kJ mol-' 60

50 40

t

+61.9...... AG$czp

+43.0

......AG',.,

20

10

0

I'

Q 0 0 @ Q @ Q

000

00000

A

J W

-2 AV'~~,,= -12.1 cm3 mol-'

AV'.,,

= -4.1 cm3mol-'

AV'.,,

= +7.2 cm3mol-'

reliability of his calculations was proved by the agreement of the computed activation energies AE# with the experimental AGf and AHf values (Fig. 4.4). Calculation of the partial molar volumes of the reactants, the transition state and the intermediate is difficult and Rotzinger calculated differences of the sums of all metal-oxygen bond lengths, AC(M-0), which are related to the activation volumes. Figure 4.4 shows that the sign of AC(M-0) is negative and agrees with that of the experimental AVf . For the titanium transition state, as Rotzinger states, reaction water entry undoubtedly dominates the activation process which can clearly be identified as an A mechanism. The transition state for {[V(H20)5. . . (H20)2]2t}f has C2 symmetry and arises from a concerted process. The hypothetical transition states involving electronic low-spin (triplet) states, {[V( H Z O ) ~ #] ,~have ~ } very high energies (242 and 244 kJ mol-') relative to the reactant, excluding spin-change in the water exchange process. The negative value of ACd(M-0) for the transition state shows that bond formation occurs to a larger extent than bond breakage. For water exchange on [Ni(H20)6l2+,only a transition state for dissociative activation could be calculated. The species {[ Ni( H20)s . (HzO).(HzO)]2t}i has C, symmetry and the incoming and leaving water molecules are easily distinguishable. The '

4.3 Water Exchange from the First Coordination Shell

subsequent intermediate [ Ni(H20)5.(H20)2]2+ has ClYsymmetry and is basically a square pyramidal penta-coordinate species, hydrated by two water molecules. The energy difference between transition state and intermediate is small for this dissociative process. The calculated sum of the metal-oxygen bonds increased for both the transition state and the intermediate as compared to the reactant state, leading to a positive of value AC(M-0). An improvement of the calculations is the introduction of dynamic and static electron correlation in the computation of the energies and the second coordination sphere together with bulk solvent as a dielectric continuum in the calculation of energies and geometries [43]. This extended model was applied to the H2Oexchange in [V(H20)(1]2 [Mn(H20)6I2+,and [Fe(H20)6I2+.For each hexaaqua ion associative and dissociative mechanisms were investigated. Starting from a second sphere water molecule forming two hydrogen bonds with the first sphere water molecules, activation energies in agreement with experimental data were obtained. For water exchange in V(I1) and Fe(II), computational data suggest an a and a d activation, respectively, in accordance with experimental data. The model did not allow distinction between a and d activation for Mn(I1) on the basis of energies. The attribution of an associatively activated exchange was made in this case using changes in the sign of the bond length. In summary, the results obtained by Rotzinger on first-row transition metals provided the following picture [41]. Hexacoordinated Sc3+,Ti3+ and V3+ react via an associative A mechanism with relatively long-lived intermediates. The dissociative pathway is only possible for water exchange on Ni2+,Cu2+ and Zn" . For the elements in the middle of the 3d series both associative (IJA) and dissociative (D) pathways can occur. The structure and water exchange mechanism of hexahydrated Ti(III), its hydrolysis, and the water exchange mechanism of analogous hydroxo-aqua complexes have also been studied using DFT calculations by Hartmann et al. [44]. It was found that the water exchange reaction of this complex follows a (almost) limiting A mechanism with an energy of activation of 66.1 kJ mol-'. Only a weak structural influence, indicated by a slight increase in the mean value of the T i - 0 bond lengths of water molecules in the first coordination sphere, is observed when the hydroxo ligand is formed. The water exchange reactions of the corresponding hydroxo-aqua complexes [Ti(H2O),(OH)]2+ and [Ti(H20)5(0H)]'+H20, respectively, were found to proceed via limiting D mechanisms. The energies of activation for the exchange of the water molecule in the trans-position to the hydroxo ligand were calculated to be only 41.0 and 30.1 kJ mol-', respectively. The hydration of the zinc(11) ion was also studied by Hartmann et al. [ 451. Structures and hydration energies were computed using DFT for [ Zn( H z O ) ~ '] with n = 1-6 and for [Zn(H20),]2+.rnH20with n = 5 and rn = 1 , 2 and n = 6 and rn = 1. The structures with water molecules in the second coordination sphere revealed two types of complexes having either one or two hydrogen bonds between first and second sphere water molecules. In agreement with Rotzinger [41]a dissociative D mechanism was found to be most likely for the water exchange in [Zn(HzO)6I2+. +,

I

141

142

I

4 Water Exchange on Metal lons: The Effect of Pressure

Rate constants and activation parameters for water exchange on second- and third. row transition metal ions.

Tab. 4.3.

M ll+ Geometry

4dS

4d6

Ru3+ Octahedral

RuZ+ Rh’t 1 ~ 13 Pd2+ Pt2+ Octahedral Octahedral Octahedral Square Square planar planar 73 66.5 68 64 60 1.8 x 2.2 x 10-9 1.2 x 10-9 5.6 x 10’ 3.9 x 10-4 131.2 87.8 49.5 130.5 89.7 fl6.l +29.3 -26.0 -9.0 +2.1 -0.4 -4.1 -2.2 -4.6 -5.7 Id, D Ia 1, 1, Ia 2 38 39 46 47

68 rM (pm) k2’* (s-I) 3.5 x 10-6 AH# (k]mol I ) 89.8 ASf (1 K-’ mol-I) -48.2 AVf (cm3 mol-’) -8.3 Mechanism 1, Ref. 38

4d6

5d4

4d8

5d8

4.3.2 Second- and Third-Row Transition Metal lons

Experimental data on water exchange on second- and third-row transition metal ions are much more scarce (Table 4.3). For the low-spin t!g hexaaqua ions [Ru(H z o ) ~ ] [ Rh(HzO),] 3+ and [ Ir(H 2 0 ) ~ ] one would a priori predict a dissociative activation mode. The filled t2g orbitals, spread out between ligands, electrostatically disfavor the approach of a seventh molecule towards a face or edge of the octahedron and therefore decrease the ease of bond making. In a first study of the mechanism of substitution in [Ru(H20)~]”;it was shown that the rate constants for the anation reactions of C1-, Br-, and I- were very similar, indicating identical steps to reach the transition state (i.e. dissociation of H 2 0 )[48]. Later, this study was extended to a large variety of ligands possessing various charges and nucleophilicities, and it was clearly demonstrated that the rate determining step of the monocomplex-formation reactions was independent of the nature of the entering ligand. An Id mechanism was therefore suggested for the substitution reactions in [ Ru(H20)6]2+ [49]. However, a variable pressure study of water exchange on this ion gave an activation volume close to zero (AV# = -0.4 cm3 mol-’) and was therefore interpreted as an interchange I mechanism without a predominant “a” or “d’ character [38]. Attempts to compute, at the restricted Hartree-Fock (SCF) level using the selfconsistent reaction field model (SCRF), a transition state for an I mechanism failed and only a D mechanism with a five-coordinate intermediate could be computed [SO]. The impossibility of calculating a transition state for the Id mechanism does not mean that it does not exist: it is just not available within the model in which the second coordination sphere is not treated by quantum chemistry. Therefore, the experimental and theoretical results available today suggest that water exchange on [Ru(H20)6I2+proceeds via a “d activation”, either the Id or D mechanism. This is also in agreement with the high pressure study of the formation of [ Ru(H20)5(DMF)]2+ from the aqua ion, for which the volume of the transition

’+,

’+,

4.3 Water Exchange from the First Coordination Shell

state lies between the volume of the reactant and the volume of the products and where an I d mechanism has been assigned [Sl]. [ Rh(H20)6]3+ and its third-row analog [ Ir(H20)6] ' are the most inert aqua ions known up to now [2].The lifetime of a water molecule in the first coordination shell of Ir3- is 9.1 x lo9 s (at 298 K), which corresponds to about 300 years. The negative activation volumes of Rh3+ and Ir3+ ions suggest the operation of an associative I, exchange mechanism (Table 4.3).Theoretical calculations have allowed the characterization of the seven-coordinate transition state {[Rh(OH2)5. . . (OH2)2l3+}+ corresponding to the I, exchange mechanism for Rh3+ [SO]. Why do Rh3+ and Ir3+ on one hand and Ru2' on the other, undergo water exchange via disparate mechanisms although they are isoelectronic? The charges are different, affecting particularly the M - 0 bond strength. This is manifested by the calculated activation energies for the exchange reactions via the D mechanism: 71.9 kJ mol-' for Ru2+ and 136.6 kJ mol-' for Rh3+ [SO] (Fig. 4.5). Therefore, the strong Rh"'-0 bonds allow water exchange in [ Rh(Hz0)6]3T to proceed via the I, pathway, whereas the same reaction on [ Ru(H20)6]2 + , which has considerably weaker Ru"-0 bonds, follows the I d or D mechanisms.

160 + I 22.3 +114.8

:I. 00000

-2 -1

......AG'.,

A 000

Fig. 4.5. Energies calculated with a polarizable continuum model, differences of the sums of all metal-oxygen bond lengths, AC(M-0), and energy profiles for water exchange on rhodium(ll1) and ruthenium(l1) hexaaqua ions.

+83.0-----5 +7 1.9

00000

I

143

144

I

4 Water Exchange on Metal Ions The Effect of Pressure

The low-spin t& [ R U ( H Z O ) ~is ] ~four ’ orders of magnitude more labile than the tig [ R u ( H ~ o ) 6 ] ~[38] ’ and, according to the same arguments developed above, exchanges water by an I, mechanism. The knowledge of the difference in lability of [ Ru(H20)6]2- and [ Ru(H20)6]3- led to the unequivocal identification of an outersphere mechanism operating in the electron exchange between these two aqua complexes [38]. Water exchange rates of the three M3+ complexes were also measured for the deprotonated species [M(H20)s(H,0)]2f.As for their first row analogs, a substantial acceleration of the water exchange was observed compared to the hexa-aqua complex: k,, = 5.9 x 10- s-l ([Ru(H20),(OH)12+),k,, = 4.2 x lo-’ s-l ([Rh(H20),(OH)]2f),and k,, = 5.9 x s-l ([Ir(H20)5(0H)12+), and also a changeover towards a more dissociative activation or less associative activation mode. The vast majority of square-planar dXtransition metal ions, such as palladium(I1) and platinum(11) undergo ligand substitution through associatively activated mechanisms [ 5 2 ] . Water exchange on [Pd(H20)4l2+[4G] and [Pt(H20)4I2+1471 were studied by I7O N M R (Table 4.3). The mechanistic interpretation of the activation volumes is complicated by the square-planar geometry of the complexes. The axial sites may be occupied by non-bonded water molecules. Bond formation between the metal center and such a water molecule leading to a five-coordinate transition state may result in rather small volume changes. It is therefore difficult to distinguish from AVf values between an associative interchange I, and a limiting associative A mechanism. Density functional theory has been used by Deeth et al. [ 531 to model water exchange on square-planar [ Pd(HzO),] 21 and [ Pt(H z O ) ~2 L]. The calculations strongly suggest that water exchange at these square-planar metal centers proceeds via an associative mechanism, in full agreement with experimental assignments [4G, 471. The agreement in experimental and calculated activation enthalpy is better than 10 kJ mol-* for an I, mechanism, whereas it differs by more than 100 kJ mol-I for an Id mechanism. 4.3.3 Lanthanides

The 15 trivalent lanthanide ions, which may be collectively denoted by Ln3+, represent the most extended series of chemically similar metal ions. The progressive filling of the 4f orbitals from La3+ to Lu3+is accompanied by a smooth decrease in ionic radius with an increase in atomic number as a consequence of the increasingly strong nuclear attraction of the electrons in the diffuse f orbitals. These orbitals are shielded by the surrounding filled 5s and 5 p orbitals, leading to very small crystal field splittings in the lanthanide complexes. Thus the coordination properties of the Ln3+ ions mainly depend upon the steric and electrostatic nature of the ligands. The coordination of the lanthanide(II1) ions in aqueous solution was certainly one of the most controversial questions in lanthanide chemistry. Now it is well established from neutron scattering [54, 551, X-ray scattering [5G-58] and absorption fine structure [ 591, density [GO] and spectrophotometric (611 techniques that the lighter ions ( La3+-Nd3+) are predominantly nine-coordinate and the

4.3 Water Exchange from the First Coordination Shell

heavier ions ( Gd3+-Lu3+) are predominantly eight-coordinate species. The ions Sm3+-Eu3+ exist in equilibria between eight- and nine-coordinate species (Eq. 4.6).

+

[Ln(H2O)sI3+ H2O $ [Ln(H20)9]3+ The change in coordination number is reflected in the absolute partial molar volumes, V2b,, determined in aqueous LnC13 solutions (Fig. 4.6). The calculated absolute partial molar volumes using the semi-empirical model of Swaddle [20, 211 agree well with [ Ln( Hz0)9] 3+ at the beginning and with [ Ln(H20)8]3-' at the end of the Ln3+ series. Rate constants for water exchange on lanthanide(II1) ions can be determined by oxygen-17 NMR relaxation and chemical shift measurements [62, 631. Unfortunately, only eight-coordinate heavy lanthanides (Gd3+to Yb3+) could be studied (Fig. 4.6) due to the less important chemical shift induced by the early lanthanide ions. Powell's results [G4]at high magnetic field allowed the determination of the lower limits for water exchange rates at 298 K on [Nd(H20)9] and [ Pr( H20)9] 3 t . The water exchange rates are closely correlated with the rate, kfq8, of the interchange between an inner sphere water molecule and a SO:- ion from the outer sphere coordination (Fig. 4.6) [65]. The values show a maximum in the middle of the series, corresponding to the crossover region in CNI from nine to eight. For the heavy lanthanides, in addition to variable temperature, variable pressure measurements were performed so that the activation volumes for the exchange process could be determined. The activation volumes are all negative [G3], indicating an associatively activated water exchange mechanism (Fig. 4.6). This observation can be explained in terms of the coordination equilibrium observed in the middle of the series. For the heavy lanthanides, the octa ion is the lowest energy species and the ennea-aqua ion represents the intermediate (or transition state) in the associatively activated water exchange reaction. The exchange rates become faster towards the middle of the series as the difference in the energy between octa- and enneaaqua ions decreases. For the mid-series lanthanides, the octa- and ennea-aqua species are in equilibrium. The transition from one species to the other requires relatively little energy, so that these lanthanides should have the fastest water exchange rates of the series. For the light lanthanides, the ennea-aqua ion is now the lowest energy species. One can expect, therefore, that the octa-aqua ion will be the intermediate (or transition state) in the exchange process, which will proceed via a dissociatively activated mechanism. One would expect to observe an increase in water exchange rates and positive activation volumes from La3+ to Nd3+. Unfortunately, so far it has not proved possible to demonstrate experimentally these predictions for the light lanthanide ions. The absence of crystal field splitting and the shielding of the f orbitals makes classical molecular dynamics (MD) simulations feasible. Hence, for solvents without covalent bonding like water, ion-solvent forces can be well represented by simple Coulomb and van der Waals terms. The polarization of water molecules in the first coordination shell due to the strong electric field of the trivalent ions has

'

I

145

146

I

4 Water Exchange on Metal Ions: The Effect ofpressure

-45

1

I

I

I

I

I

I

I

I

I

I

I

I

I

I

I

\€i-coordinate

l ! 9-coordinate

E

h

''

-2 -

-

-4 -6-

-

-8-

-

1

1

1

1

1

1

1

1

1

1

1

1

l

I

I

to be taken into account [GG]. Figure 4.7 shows the distribution of the number of water molecules for three distance regions of three trivalent ions, designated as Nd3+, Sm3+ and Yb3+ obtained from classical MD simulations [GG]. The experimentally observed change in the first shell coordination number CN, from 9 for Nd3+ to 8 for Yb3+could be reproduced. The result for Sm3+reveals that the observed CNI of 8.5 indeed corresponds to an equilibrium between an octa- and enneaaqua ion. Only 9 % of all configurations have a ninth water molecule between the two spheres (Fig. 4.7). An advantage of computer simulations is that they can deliver microscopic

4.3 Water Exchange from the First Coordination Shell

I

147

I

3.1-3.78,

0.0-3.1A 3.1-5.48, '

(1st

1

shell)

(2nd shell)

0.8

.-0 5

0.6

2g 0.4 a

0.2 0

1 0.8

L. Z 0.6

R.52

5g 0.4

17.13

P,

0.2 0 1

-%3 0.60.8 .-

s 0

k 0.4 0.2

0

5

10

15

20

25

number of water molecules Fig. 4.7. Distribution of the number o f H20 molecules in three distance regions for the Simulations, Nd, Sm, and Yb within the polarizable water model. The inserted numbers are the mean number of water molecules in each distance range.

structural pictures of the systems studied. A detailed analysis of the first coordination shell [67] showed that for the light lanthanide ions (Nd3+),the nine water molecules form a tricapped trigonal prism similar to that found by solid state X-ray diffraction studies on [Ln(H20)0]3+ with counterions CF3SO7, EtOSO;, and BrO; [68]. The eight H 2 0 molecules in contact with the smaller heavy lanthanide ions (n3+) form a square antiprismatic geometry with equal ion-oxygen distances. Hydration shells from simulations of Sm3+ are actually a mixture of equal amounts of eight- and nine-fold coordinated ions adopting the geometry of a square antisprim or a tricapped trigonal prism, respectively [67]. Figure 4.8 (bottom) gives an example of the trajectories of the water molecules in the first and

148

I

4 Water Exchange on Metal Ions: The Effect of Pressure

-

--

116

0'50

ili 300

320

340

360 tlps

380

400

Fig. 4.8. Ion-oxygen distance for a 100-ps interval simulation for Sm (bottom). The plot illustrates the equilibrium between eight- and nine-fold coordination. The volume V,,, that is enclosed by the solvent-accessible surface of the [Sm(Hz0)8/9]3+aqua ion is shown as a function of time (top).

second hydration shells for a 100-psinterval and illustrates the equilibrium between the two coordinations. The reaction volume, AVO, for this equilibrium can be estimated using Swaddle's semi-empirical model [20], as about --13.0 cm3 mol-'. From the M D simulations, volumes included in the solvent-accessible Connolly surface [G9], VSa,, can be calculated (Fig. 4.8, top). With V,,, (CN, = 9) = 112.0 cm3 mo1-l and V,,, (CNI = 8) = 101.2 cm3 mol-', a reaction volume of -7.2 cm3 mol-' was calculated [70]. MD simulations performed at different densities ( p = 0.9, 1.0 and 1.1 g cm-3 corresponding to pressures of -244, 0.1 and 188 MPa), showed a change in the mean coordination number from 8.26 to 8.62 due to an increase of configurations with nine coordinated water molecules in the first shell with increasing density (Fig. 4.9). The slope of a plot of ln(K) versus pressure gives a AVO of -8.3 cm3 mol-l [70]. Both volumes calculated in different manners

4.3 Water Exchange from the First Coordination Shell

1

3.1-3.7A

0.0-3.1A (1st shell)

133'i2L$i

0.8 0.6

a 0.4

0.2

0.8

P=1.013 g/cm' 8.48

0.6

c

0.4 4.

0.2

0.8 0.6

a 0.4 0.2

L 0

5

10 15 20 number of water molecules

25

Fig. 4.9. Distribution of the water molecules in three distance ranges around the S m 3 ' ion at three different overall densities. The inserted numbers are the mean numbers of water molecules in each distance range.

from MD simulations agree well with the reaction volume of -10.9 cm3 mol-' measured experimentally for [ Ce(H20)8]3f H 2 0 F? [ Ce(Hz0)9]3+ [ 711. From molecular dynamics simulations, information on water exchange is available for the three regions of coordination. The simulation qualitatively confirmed the lability maximum of the first hydration shell found in the middle of the series [70]. From the trajectories obtained from the simulations, all transitions of water molecules between the first hydration shell and the bulk were identified and grouped into pairs of water molecules that together form a coupled exchange [70]. Similar to the hydration equilibrium of Sm3+, volume profiles for water exchange processes can be obtained using Connolly's method. Volume changes with

+

I

149

150

I

4 Water Exchange on Metal Ions: The Efict of Pressure

5 4

20 3 f “E0

. m 1

$ 0 -1.

I

-

128

‘re1 ’PS Fig. 4.10. Mean activation volume profile AVsas(t)from the water exchanges on Nd3’- (CN = 9, top) and Yb3+ (CN = 8, bottom) after averaging the positive and the negative halves of the profile.

respect to the mean volume of an aqua ion in equilibrium plus one isolated water molecule are shown for 10-ps intervals in Fig. 4.10 1701. For [Nd(H20)9]3L (Fig. 4.10 top) the dissociation of a water molecule from the primary hydration shell corresponds to an increase in volume. The picture that emerges from the volume profile for the water exchange on [Yb(H20)s]3+ is more complicated. The entering of a water molecule from the second hydration shell into the region between the first and second spheres leads to a slight increase in volume. This feature of the volume profile indicates that the square antiprismatic octa-aqua ion prepares itself geometrically for the reception of a ninth water molecule before compacting to the transition state that is a tricapped trigonal prism. The only easily accessible divalent lanthanide is europium(I1). Eu2+ is isoelectronic with Gd3+ and very similar in size to Sr2+. Since water exchange on the

4.3 Water Exchangefrom the First Coordination Shell

heavier group 2 elements cannot be directly measured, Eu" may provide insight as a probe in the chemistry of Ca" and Sr". A simultaneous fit of nuclear and electronic relaxation rate data results in k::8 = 4.4 x 10' for the water exchange rate with activation parameters AHf = 15.7 kJ mol-' and A S # = -7.0 J K-' mol-', AV# = -11.3 cm3 mol-' for the volume of activation for exchange, and 2 i 9 8 = 16.3 ps for the rotational correlation time 19, 721. The volume of activation is very large and corresponds to a limiting associative mechanism. This suggests that water exchange for an eight-coordinate ELI& (typically square antiprismatic) would proceed through a nine-coordinate transition state (tricapped trigonal prism), similar to that for [Gd(H20)8]3i[73]. However, the water exchange rate for ELI;) is 5-6 times faster than that for Gd&. 4.3.4 Main Group Elements

There are only a few main group metal ions amenable to a detailed mechanistic study of water exchange: Be2+-,Mg2+,A13+, Ga3+ and In3+. They provide the opportunity to study the influence of size and charge on the mechanism without the complicating effects of the variation of the electronic occupancy in the d-orbitals. All alkali ions and CaZ+,Sr2+and Ba2+ are very labile as a consequence of their relatively low surface charge density. The only direct experimental data on water exchange in some of these ions comes from incoherent quasi-elastic neutron scattering (IQENS) [74-761. IQENS has an observation time scale t&,s 1ns and allows for the calculation of limits for ion to water-proton binding times zi (Table 4.4). Mean lifetimes of H2O in the first shell of Ca2+ and Sr2+ can be estimated to -0.2 ns from the chemically similar EuZLion (see Sect. 4.3.2). The tetrahedrally [ 77-81] coordinated beryllium( 11) and the octahedrally [82-87] coordinated magnesium( 11) show relatively slow water exchange rates from the first coordination sphere and can be measured directly by NMR techniques (Table 4.5). The water exchange reaction on beryllium(I1) is characterized by the most negative activation volume observed for a water exchange process (AV# = -13.6

-

Tab. 4.4. Ion-water proton binding times,

T , ,obtained from IQENS experiments on concentrated aqueous ionic solutions at -298 K [76].

q

< 0.1 ns

ri

> 0.1 ns

T; > 5

Li+

Zn2+

Mg2+

cs+

Nd'+ Dy'+

Al3t

Ca2+

Ni2

cu2+ F-

Cr'+

c1-

Ga3

Ic10;

Fe3'

ns

152

I

4 Water Exchange on Metal Ions: The EJect of Pressure Tab. 4.5.

Water exchange rate constants and activation parameters for water exchange on

[Be(H20)4]2+ and [ Mg(HzO) 612+.

[Be(H20)4]2+

[Mg(H20)6l2+

7.3 x l o s 6.7 x lo5

59.2

+ 8.4

49.1

+31.1

-13.6 6.7

77

A Id,

D

88

cm’ mol -I), which is close to a calculated limiting A V f = -12.9 cm3 mol-’ for an A mechanism (hexaaqua ion) [77]. The water exchange rate for [Mg(H20)6]2’ [88] lies between that of [Co(H20)6]*+and [Ni(H20)6I2+(Table 4.1) and reflects the order of ionic radii of these three ions (Co2+= 74 pm, Mg2+= 72 pm, Ni2+ = 69 pm) [89].The strongly positive activation volume (AVt = +6.7 cm’ mol-’) is intermediate between that obtained for Co2+and Ni2+. Based on the similarity between Mg2+ and these ions, an Id or D mechanism is proposed for H 2 0 exchange. Furthermore, the complex-formation rates [90] are close to the water exchange rate. which is indicative of a metal-water bond rupture as the rate determining step. The trivalent cations A13+, Ga3+ and In3+ have closed shell configurations and differ strongly in their ionic radii (53, 62 and 80 pm, respectively). They form octahedral [M(H20)~]’-subunits in the solid alums CsM1”[SO4I2.12H201911. In aqueous solution the coordination numbers are also six as revealed by X-ray diffraction for Al(II1) [92-941 and In(II1) [95-971, and by NMR measurements for Al(II1) [97-991, Ga(II1)[97, 100-1021 and In(II1) [97]. Rate constants and activation parameters for water exchange in Al(II1) [lo31 and Ga(II1) 132, 1041 have been determined from kinetic 1 7 0 NMR experiments (Table 4.6). Water exchange processes in [AI(H20)6]’+and [Ga(HzO),] are characterized by positive AVt values, which are, however, substantially smaller than anticipated for a D mechanism, and, therefore, an Id mechanism was proposed. Unfortunately, the water exchange rate for In3+ was too fast to be measured, and thus precluded the measurement of an activation volume. The new limits for water exchange rate for In3$ established by using Tb’’ as a paramagnetic shift reagent [105], are about three orders of magnitude higher than the previous estimates [ 1061. Kowall [lo51 modeled water exchange mechanisms on Al’+, Ga’+, and In’+ using ab initio calculations at the Hartree-Fock level. For all three ions, transition states and intermediates have been obtained for a dissociative pathway with structures similar to those found by Rotzinger. The energies of the dissociative transition state are in excellent agreement with the activation enthalpies, AHf, measured for [Al(H20)6]’+ and [Ga(H20)6I3+ (Table 4.6) [103, 321. The energy difference, A!$, between the calculated intermediate and its corresponding transition state is relatively large for Al’’ and Ga’ t , but smaller for In3+.The computation of transition states for an associative exchange mechanism was only successful for In3+.The structures ofthe transition state and the intermediate for [ In(H20)7] are the same as those for [Ti(H2O)7I3’ [40]. Furthermore, the associative pathway requires a much lower activation energy (AEl = $29.1 kJ mol-l) than dissocia-

’+

’+

4.3 Water Exchangefrom the First Coordination Shell Tab. 4.6.

Selected properties of reactants/products, transition states, and intermediates.

Species

Symmetry

AEf a

(kj mol-lj [A1(OH2)6.0H2]3+ linear hydrogen bond {[Al(OH2)5...OH2.OH2l3+}# dissociative transition state [Al(OH2)s.(OH2)2]3 ' pentacoordinated intermediate

CS

0.0

c,

85.4 84.7< f74.7

Experimental C2"

[Ga(OH2)6.OH2I3+ cs linear hydrogen bond {[Ga(OH2)S . . .OH2.OH2I3+}# cs dissociative transition state Experimental [G~(OH~)S~(OH~)~I~~ C2" pentacoordinated intermediate [Ga(OHds40H2)213+ C2" water adduct [Ga(OH2)s.(OH2)2I3+ c2 water adduct

0.0

[ I I I ( O H ~ ) ~ . O3+H ~ ] linear hydrogen bond {[ln(OH2)5- ..OH2.0H2I3+}# dissociative transition state [In(OHz)s.(0H&l3' pentacoordinated intermediate { [ h ~ ( o H...OH2I3+}# ~)~ associative transition state [In(OH2)7l3' heptacoordinated intermediate

CS

0.0

CS

75.2

C2"

69.6

c1

29.1

CZ

19.7

69.1 67.1 54.6

AEfb (kj mol-')

AVJ or AV,

10.7

+5.6 +5.7

( ~ m mo1-l) - ~

+7.1

14.5

+4.8 +5.0 +6.9

5.6

T4.4

17.2 10.3

+4.8 9.4

-5.2 -7.1

"Difference in energy between transition state and linear bound water adduct. Difference in energy between transition state and penta- or heptacoordinate intermediate. CAHf (kJ mol-I).

tive water exchange (AE# = +75.2 kJ mol-I). The calculated activation energy of +29.1 kJ mol-' is in good agreement with the limit I 33.1 kJ mol-l for AGtg8 obtained experimentally [lOS]. Positive activation volumes for water exchange were measured on A13+ and Ga3+ by variable-pressure NMR experiments, and a dissociative mechanism, Id, was assigned. Due to the relatively large energy differences between transition state and intermediate calculated for these clusters, a limiting D mechanism was recently proposed from the calculations [105]. The much lower energy pathway for an a activated mechanism for In3+ together with the presence of a stable seven-coordinate intermediate, allowed the assignment of an A mechanism for HzO exchange in [In(H20),] 3 + . Activation volumes were estimated from the ab initio cluster calcu-

I

153

154

I

4 Water Exchange on M e t a l Ions: T h e Effect of Pressure

8

A13+ Ga3+

6

-

4

z2

A

“E “. 0 .

> d

-2

’+ 1x

I

-4

Fig. 4.11. Volume changes (reactants - transition states intermediates) for a dissociative D water exchange (top) and an associative A water exchange (bottom).

lations using the Connolly surface [G9] (Fig. 4.11). The van der Waals radius of oxygen was adjusted to reproduce the experimental AVk for A13+ and Ga3+. The corresponding activation volume for an associative exchange process calculated for In3+ was -5.2 cm3 molt’. The change in mechanism from dissociative (A13-, Ga3+) to associative ( S C ~In3-) ~ , has already been experimentally observed for organic solvent exchange reactions [23]. Swaddle used a More-O’Ferrall type diagram, scaled to the volume as a physical parameter, to visualize the interplay between bond making and bond breaking in simple interchange processes (Fig. 4.12). Based on Swaddle’s semi-empirical model for partial molar volumes of metal ions in solution, Merbach has chosen f13.5 cm3 mol-’ as the limiting value for A and D mechanisms [lG]. The only measurable volume is, however, the activation volume, that is the volume difference between the transition state and the reactant. If, as Swaddle supposes, all water exchange reactions had to be regarded as interchange mechanisms, the transition state would lie on the dashed diagonal in Fig. 4.12a. In this event, A and D mechanisms are considered as the limiting cases for the interchange mechanistic con-

4.4 Water Exchangefrom the Second Coordination Shell

bond making 0

M(H,O),'*

-2

-4

-6

-8

-10

-12

+ 2 H,O

M(H,O)d'

+ H,O

$ Fig. 4.12. Interpretation o f volumes of activation for water exchange on aqueous [M(HzO),]'+ in terms of contributions (cm' mol-') from bond making and bond

breaking: (a) summary of volumes o f activation for metal aqua ions, (b) calculated curves for AI(III), Ga(lll), and In(lll) with use ofthe Connolly volumes.

tinuum and are, therefore, following the edges of the square diagram. However, from Rotzinger's [40,411 and Kowall's [ 1051 calculations, intermediates with increased and decreased coordination numbers have been identified and a different view of the square diagram was defined [105]. For concerted mechanisms (I), the single transition state lies, as before, on the AV diagonal axis. However, for stepwise mechanisms, the intermediate is situated on the diagonal while the two transition states lie symmetrically off the diagonal. The intermediates must not necessarily lie on the corner of the square, which means that the absolute value of the volume difference between intermediate and ground state can be significantly smaller than 1+13.51 cm3 mol-I. From calculated structures, Kowall et al. [lo51 derived volume changes for the transition states, AV:, and for intermediates, AVc. These values are then used to draw reaction trajectories into the square diagram (Fig. 4.12b). As one would expect, transition states are about half way between the reactant and the intermediate.

4.4

Water Exchange from the Second Coordination Shell

Information on the second coordination shell of water molecules around cations is much poorer than that regarding the first shell [4]. Properties of solvent molecules in this coordination shell are often very similar to those of the bulk, making their investigation extremely difficult. The analysis of radial distribution functions, g(r ) ,

I

155

156

I

4 Water Exchange on Metal Ions: The Effect of Pressure

at long r distances, is difficult and much less reliable. The main source of structural information comes from neutron scattering 11071 and X-ray diffraction [ 1081111 experiments on aqueous solutions. For cations with strong second shell solvation, such as the trivalent A13+,Cr3+,or Rh3+,information can be obtained from far infrared spectroscopy 11121. Computer simulations performed in close relation to experimental studies can deliver more insight into the microscopic structure and dynamics of solvent molecules in the second coordination shell [113, 1141. From MD simulations it can be concluded that for trivalent ions (Cr3-, lanthanides) in general, first shell water molecules form two hydrogen bonds to second shell water oxygens. In that way the number of second shell water molecules (CNII) is roughly twice that ofthe first shell coordination number (CN,). For [Cr(H20)6] a mean second shell number of 12.9 H 2 0 molecule was found and for lanthanides, CN,I-values of 17.61 (for [Nd(H20)9I3+)and 16.74 (for [Yb(H20),13+)were obtained (Fig. 4.7). In an experimental oxygen-17 NMR study Bleuzen et al. I1131 measured an exchange rate constant of k,, = 7.8 x 10' s-l (at 298 K) on Cr3+,corresponding to a lifetime of 128 ps for one water molecule in the second coordination shell. In the same study a lifetime of 144 ps was observed from the molecular dynamics simulation. Based on the simulation, the following picture for an exchange reaction between a second and a third sphere water molecule on [Cr(H20)6I3+could be obtained (Fig. 4.13). In the first instance, a water molecule (no. 4 in Fig. 4.13) enters the second coordination sphere and increases the coordination number temporarily. In a second step the first sphere water molecule (no. I), which is close to the one that is entering, rotates around its oxygen-chromium bond, then one of the hydrogen bonds formed to second sphere waters (no. 3) breaks up and a new hydrogen bond is formed with the entering water molecule. During the rotation the second hydrogen bond is maintained (no. 2 in Fig. 4.13). In the third step the water molecule which has lost its hydrogen bond to the second sphere is itself

'+,

t=-2.5ps

t=Ops

I $L

I

--\

Fig. 4.13. Visualization o f water exchange between the second coordination shell and bulk solvent on [Cr(HzO)&+ obtained from MD simulation: (1) selected first sphere water molecule; (2). (3) second sphere water molecules before exchange; (4) exchanging outer sphere water molecule.

Acknowledgments

lost. The activation mode which can be attributed to this reaction from the MD simulation, is associative. Mufioz-Paez et al. [ 1141 performed extended X-ray absorption fine structure studies on aqueous solutions of Cr3+and Zn2+. They detected second coordination shells in both cases with coordination numbers of 13.3 f 1 (Cr3+)and 11.6 f 1.5 (Zn2+). The same group performed Monte Carlo I1151 and molecular dynamic [11G, 1171 simulations of [Cr(Hz0)6I3+in dilute aqueous solutions using an ab initio Cr3+ hydrate-water interaction potential. They found second shell coordination numbers of -14 from both simulations. Furthermore, from simulations and EXAFS measurements they concluded that chloride ions are situated beyond the second hydration shell. Near infrared experiments on lanthanide( 111) ions showed no well-defined second hydration shell [118].The residence times for water molecules in the second shells around Nd, Sm and Yb are 13, 12 and 18 ps respectively [ 7 0 ] . The slightly higher value for Yb3' can be related to its smaller ionic radius, rion.

4.5

Conclusions

The results reviewed in this chapter demonstrate the usefulness of pressure variation for the elucidation of mechanisms of water exchange reactions on metal ions. Even if the distinction between a limiting and an interchange mechanism, based on the magnitude of the activation volume, is not straightforward, the activation mode, associative or dissociative, can mostly be assigned unambiguously from the sign of AVt. Due to the range of more than 18 orders of magnitude for water exchange rate constants on metal ions, a variety of experimental techniques, mainly nuclear magnetic resonance, have to be applied. High-pressure equipment, such as high-pressure NMR probes, is still not easy to obtain and is therefore not widely used. The concept of activation volumes has also become a valuable tool in studies of exchange reactions by ab initio computer calculations and in classical computer simulations. In these theoretical studies activation volumes can be estimated by bond-length variations or by calculating volume differences using Connolly surfaces. In MD simulations pressure can be applied by variation of the density of the simulated water box. In that way reaction volumes are accessible by following for instance the change in coordination number.

Acknowledgments

The authors gratefully acknowledge financial support from the Swiss National Science Foundation and the Swiss Office for Education and Science (COST Program). Furthermore, we wish to thank the large number of people who have contributed to the work performed in Lausanne.

I

157

158

I

4 Water Exchange on Metal lons: The Efect ofpressure

References

D. T. RICHENS,The Chemistry ofAqua lons, Wiley. Chichester, 1997. 2 A. CUSANELLI, u. FREY,D. .r. RICHENS,A. E. MERBACH, ]. Am. Chem. SOC.1996, 118, 6265. 3 (a) M. EIGEN,Pure Appl. Chem. 1963, 6, 97; (b) M. EIGEN,R. G. WILKINS Ado. Chem. Ser. 1965, 49, 55. 4 H. OHTAKI,T. RADNAI,Chem. Reu. 1993, 93, 1157. 5 P. SMIRNOV, H . WAKITA~ T. YAMAGUCHI, ]. Phys. Chem. 1998, B 102,4802. 6 A. E. MERBACH, J. W. AKITT, N M R : Basic Princ. Prog. 1990, 24, 189. 7 U. FREY,D. H. POWELL, A. E. MERBACH, i n Dynamics ofSolutions and Fluid Mixtures by N M R , J.-J. DELPUECH(Ed.), J o h n Wiley & Sons, Chichester, 1995. 8 R. CAR,M. PARRINELLO, Phys. Re[>. Lett. 1985, 55, 2471. 9 P. CARAVAN, A. E. MERBACH,Chem. Commun. 1997, 2147. 10 D. MAW; M. SPRIK;M. PARRINELLO, Chem. Phys. Lett. 1997, 273, 360. 11 L. M. RAMANIAH, M. BERNASCONI, M. PARRINELLO, /. Chem. Phys. 1999, 111, 1587. 12 A. PASQUARELLO, I. PETRI, P. S. SALMON, 0. PARISEL,R. CAR, E. T ~ T H , D. H. POWELL,H. E. FISCHER,L. HELM,A. E. MERBACH,Science 2001, 856. 13 L. HELM,A. E. MERBACH,Coord. Chem. Rev. 1999, 187, 151. 14 C. H. LANGFORD, H. B. GRAY,Llgand Substitution Processes, Benjamin Inc., New York, 1965. 15 A. E. MERHACH. Pure Appl. Chem. 1982, 54, 1479. 16 A. E. MERBACH,Pure Appl. Chem. 1987, 59, 161. 17 R. VAN ELDIK,Inorganic High Pressure Chemistry: Kinetics and Mechanisms, Elsevier, Amsterdam, 1986. 18 R. VAN ELDIK,Pure Appl. Chem. 1992, 64, 1439. 19 R. VAN ELDIK,A. E. MERBACH Comm. Inorg. Chem. 1992, 12, 341. 20 T. W. SWADDLE, M. K. S. MAK Can.]. Chem. 1983, 61, 473. 1

T. W. SWADDLE, Adi. Inorg. Bioinorg. Mech. 1983, 2, 95. 22 (a) T. ASANO,W. I. LE NOBLE,Chem. Rev. 1978, 78,407; (b) R. VAN ELDIK. T. ASANO,W. J. LE NOBLEChem. Re[). 1989, 89, 549; (c) A. DRLIACA, C. D. HUBBARD, R. V A N ELDIK,T. ASANO, M. V. BASILEVSKY, W. I. I.E NOBLE, Chem. Reo. 1998, 98, 2167. 23 S. F. LINCOLN, A. E. MERBACH, Adc. lnorg. Chem. 1995>42, 1. 24 Y. DUCOMMUN, D. ZBINDEN,A. E. MERBACH, Helv. Chim. A d a 1982, 65, 1385. 25 Y. DUCOMMUN, K. E. NEWMAN, A. E. MERBACH,Inorg. Chem. 1980, 19, 3696. 26 A. HUGI, L. HELM,A. E. MERBACH, lnorg. Chem. 1987, 26, 1763. 27 A. HUGI, L. HELM,A. E. MERBACH, Helu. Chim. A d a 1985, 68, 508. 28 F.-C. Xu, H. R. KROUSE, T. W. SWADDLE, Inorg. Chem. 1985, 24, 267. 29 M. GRANT,K. B. JORDAN, Inorg. Chem. 1981, 20, 55. 30 T. W. SWADDLE, A. E. MERBACH, Inorg. G e m . 1981, 20, 4212. 31 D. H. POWELL, L. HFLM,A. E. MERBACH, ]. Chem. Phys. 1991, 95, 9258. 32 D. HUGI-CLEARY, L. HELM,A. E. MERBACH, ]. Am. Chem. Soc. 1987, 109,4444. 33 F. K. MEYER,K. E. NEWMAN, A. E. MERBACH,]. Am. Chem. SOC.1979, 101, 5588. 34 6 . M. RODE,G. J. REIBNEGGER, S. FUJIWARA,].Chem. SOC.Faraday Trans. 2 1980, 1268. 35 H. OHTAKI,M. MAEDA,Bull. Chem. SOC.Jpn. 1974, 47, 2197. 36 M. MAGINI,G. LICHERI,X-Ray Difiaction oflons i n Aqueous Solutions: Hydration and Complex Formation, CRC Press, Boca Raton, 1988. 37 D. H. POWELL, P. FURRER,P.-A. PIITET, A. E. MERBACH, J. Phys. Chem. 1995, 99, 16622. 38 1. RAPAPORT, L. HELM,A. E. MERBACH, P. BERNHARD, P. LUDI, lnorg. Chem. 1988, 27, 873. 21

References LAURENCZY, I. RAPAPORT, D. ZBINDEN,A. E. MERBACH,Magn. Reson. Chem. 1991, 29. S45. 40 F. P. ROTZINGER,].Am. Chem. Soc. 1996, 118, 6760. 41 F. P. ROTZINGER,].Am. Chem. Sac. 1997, 119, 5230. 42 F. P. ROTZINGER, Chimia 1997, 51, 97. 43 F. P. ROTZINGER, Helu. Chim. Acta 2000, 83, 3006. 44 M. HARTMANN, T. CLARK,R. VAN ELDIK. 1.Phys. Chem. A 1999, 103, 9899. 45 M. HARTMANN, T. CLARK,R. VAN ELDIK,]. Am. Chem. Soc. 1997, 119, 7843. 46 L. HELM,L. I. ELDING,A. E. MERBACH,Hela Chim. Acta 1984, 67, 1453. 47 L. HELM,L. 1. ELDING,A. E. MERBACH,Inorg. Chem. 1985, 24, 1718. 48 T. W. KALLEN, J. E. EARLEY,Inorg. Chem. 1971, 10, 1149. 49 N. AEBISCHER, G. LAURENCZY, A. LUDI,A. E. MERBACH,Inorg. Chem. 1993, 32, 2810. 50 D. DE VITO, H. SIDORENKOVA, F. P. ROTZINGER, I. WEBER,A. E. MERBACH, Inorg. Chem. 2000, 39, 5547 51 N. AEBISCHER, A. E. MERBACH,Inorg. React. Mech. 1999, I , 233. 52 R. J. CROSS,Adu. Inorg. Chem. 1989, 34, 219. 53 R. J. DEETH,L. I. ELDING,Inorg. Chem. 1996, 35, 5019. 54 C. COSSY,A. C. BARNES,J. E. ENDERBY, A. E. MERBACH,]. Chem. Phys. 1989, 90, 3254. 55 C. COSSY,L. HELM,D. H. POWELL, A. E. MERBACH, New.]. Chem. 1995, 19, 27. 56 A. HABENSCHUSS, F. H. SPEDDING,]. Chem. Phys. 1979, 70, 2797. 57 A. HABENSCHUSS, F. H. SPEDDING, 1. Chem. Phys. 1979, 70, 3758. 58 A. HABENSCHUSS, F. H. SPEDDING,]. Chem. Phys. 1980, 73,442. 59 T. YAMAGUCHI, M. NOMOURA,H . WAKITA,H. OHTAKI,]. Chem . Phys. 1988, 89, 5153. 60 A. F. H. SPEDDING, L. E. SHIERS, M. A. BROWN,J. L. DERER,D. L. SWANSON, A. HABENSCHUSS, I. Chem. Eng. Data 1975, 20, 81. 39 G.

K. MIYAKAWA, Y. KAIZU,H. 1.Chem. Sac., Faraday KOBAYASHI, Trans. 11988, 84, 1517. 62 C. COSSY,L. HELM,A. E. MERBACH, Inorg. Chem. 1988, 27, 1973. 63 C. COSSY,L. HELM,A. E. MERBACH, Inorg. Chem. 1989, 28, 2699. 64 D. H. POWELL, A. E. MERBACH,Magn. Reson. Chem. 1994, 32, 739. 65 D. P. FAY,D. LITCHINSKY, N. PURDIE, 1.Phys. Chem. 1969, 73, 544. 66 T. KOWALL, F. FOGLIA,L. HELM,A. E. MERBACH,].Am. Chem. Sac. 1995, 117, 3790. 67 T. KOWALL, F. FOGLIA,L. HELM,A. E. MERBACH, 1.Phys. Chem. 1995, 99, 13078. 68 see C. COSSY,A. E. MERBACH, Pure Appl. Chem. 1988, 60, 1785 and references therein. 69 M. L. CONNOLLY, Science 1983, 221, 709. 70 TH. KOWALL, F. FOGLIA,L. HELM,A. E. MERBACH,Eur. Chem.]. 1996, 2, 285. 71 G. LAURENCZY, A. E. MERBACH,Helu. Chim. Acta 1988, 71, 1971. 72 P. CARAVAN, E. T ~ T HA. , 1.Am. ROCKENBAUER, A. E. MERBACH, Chem. Soc. 1999, 121, 10403. 73 K. MICSKEI,D. H. POWELL, L. HELM, E. BRUCHER,A. E. MERBACH,Magn. Reson. Chem. 1993, 31, 1011. 74 P. S. SALMON,]. Phys. C: Solid State Phys. 1987, 20, 1573. 75 P. S. SALMON, W. S. HOWELLS,R. MILLS,J. Phys. C: Solid State Phys. 1987, 20, 5727. 76 P. S. SALMON, P. B. LOND,Physica B 1992, 182, 421. 77 P.-A. PITTET,G. ELBAZE,L. HELM A. E. MERBACH,Inorg. Chem. 1990, 29, 1936. 78 M. ALEI,J. A. JACKSON,].Chem. Phys. 1964, 41, 3402. 79 J. FRAHM,H.-H. FULDNER,Ber. Bunsenges. Phys. Chem. 1980, 84, 173. 80 T. YAMAGUSHI, H. OHTAKI,E. SPOHR, E. P ~ L I N K ~K.SHEINZINGER, , M. M. PROBST,2. Nutuforsch. A 1986, 41A, 1175. 81 D. MA=, M. SPRIK,M. PARINELLO, Chem. Phys. Lett. 1997, 273, 360. 82 N.MATWIYOFF, H. TAUBE,]. Am. Chem. Sac. 1968, 90, 2796. 61

I

159

160

I

4 Water Exchange on Metal Ions: The Efect ofpressure

FRATIELLO, R. R. LEE,V. M. NISHIDA,R. E. SCIIUSTER, J . Chem. Phys. 1968, 48, 3705. J . ]ORGIN, P. s. KNAPP,w. L. FLINT, G. HIGHBERGER, E. R. MALINOWSKY, J Chem. Phys. 1971, 54, 178. w. BOL,G. GERRITO,c. VON PANTHALEON V A N ECK, J . Appl. Crystallog. 1970. 3, 486. R. CAMINITI,G. LICHERI.G . PICCALUGA, G. PINNA,J . Appl. Crystallogr. 1979, 12, 34. R. CAMINITI,G. CERIONI,G. CRISPONI, P. CUCCA,2. Naturforsch. A 1988, 43A, 317. A. BLEUZEN, P:A. PIlTET, L. H E L M A. , E. MERBACH,Magn. Reson. Chem. 1997, 35, 7G5. R. D. SHANNON, Acta Crystallogr. Sect A 1976, 32, 751. M. EIGEN,K. TAMM,2. Elektrochem. 1962, 66, 107. j. K. BEATTIE,S. P. BEST,B. W. SKELTON, A. H. WHITE,J . Chem. Sac. Dalton 1981, 2105. W. BOL, T. WELZEN,Chem. Phys. Lett. 1977, 49, 189. R. CAMINITI,G. LICHERI,G. PICCALUGA, G. PINNA,T. RADNAI,J. Chem. Phys. 1979, 71, 2473. R. CAMINITI,T. RADNAI,2. Naturforsch. A 1980, 35A, 1368. M. MAEDA,H. OHTAKI,Bull. Chem. SOC.Jpn. 1977, SO, 1893. R. CAMINITI,G. PASCHINA,Chem. Phys. Lett. 1981, 82, 487. A. FRATIELLO, R. E. LEE,V. M. NISHIDA,R. E. SCHUSTER, J . Chem. Phys. 1968, 48, 3705. R. E. CONNICK,D. N. FIAT,J. Chem. Phys. 1963, 39, 1349. M. ALEI,J. A. JACKSON, J . Chem. Phys. 1964, 41; 3402. T. J. SWIFT,0. G. FRITZ, F. A. STEPHENSON, J . Chem. Phys. 1967, 46, 406. D. FIAT, R. E. CONNICK,J. Am. Chem. SOC.1966, 88,4754.

83 A.

84

85

86

87

88

89 90 91

92 93

94 95 96 97

98 99 100

101

I:RATIELLO, R. E. I.EE,R. E. SCIIUSTER, Inorg. Chem. 1970, 9, 82. D. HUGI-CLEARY. L. H E L MA. , E. MERBACH,Heltr Chim. Acta 1985, 68, 545. D. FIAT, R. E. CONNICK, J . Am. Chem. SOL.1968, 90, 608. TH. KOWALL,P. CARAVAN, H. BOURGEOIS, L. H E L M ,F. P. ROTZINGER, A. E. MKRBACH, J . Am. Chem. Sac. 1998, 120, 6569. G. E. GLASS,W. B. SCHWABACHER, R. S. TOBIAS, Inorg. Chem.1968,7,2471 R. D. BROADBENT, G. W. NEILSON, M. SANDSTROM, J . Phys.: Condens. Matter 1992, 4, 639. M. C. READ.M. SANIXTROM, Acta Chem. Scand. 1992, 46, 1177. R. CAMINITI,G. LICHERI.G. PICCAGUIA, G . PINNA,/. Chem. Phys. 1978, 69, 1. W. BOL, T. WELZEN.Chem. Phys. Lett. 1977, 49, 189. M. MAGINI,J . Chem. Phys. 1980, 73, 2499. P.-A. BERGSTROM, J. LINDGREN,M. C. READ,M. SANDSTROM, J . Phys. Chem. 1991, 95, 7650. A. BLEUZEN, F. FOGLIA,E. FURET,L. HEI.M,A. E. MERBACH, J. WEBER,J . Am. Chem. SOL.1996, 118, 12777. A. MUFJOZ-P~EZ, R. R. PAPPALARDO: E. SANCHEZMARCOS,J . Am. Chem. Sac. 1995, 117, 11710. R. R. PAPPALARDO, J. M. MARTLNEL, E. SANCHEZMARCOS,J . Phys. Chem. 1996, 100, 11748. J. M. MART~NEZ, R. R. PAPPAIARDO, E. SANCHEZMARCOS,K. REFSON,S. D~Az-MORENO, A. MUEOZ-PAEz,J . Phys. Chem. B 1998, 102, 3272. J. M. MART~NEZ, J. HERNANDEZCOBOS,H. SAINT-MARTIN, R. R: PAPPAIARDO, I. ORTEGA-BIAKE, E. SANCHEZMARCOS,J . Chem. Phys. 2000, 112, 2239. P.-A. BERGSTROM, J. LINDGREN, Inorg. Chem. 1992, 31, 1529.

102 A. 103

104 105

106 107

108 109

110 111 112

113

114

115

116

117

118

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

5 Insights into Solution Chemistry from High Pressure Electrochemistry Thomas W. Swaddle 5.1 Introduction

Just as the effect of pressure P on an equilibrium constant K can be expressed in terms of a volume of reaction AV AV

=

-RT(a In KIdP),,

(5.1)

so one can express the dependence of a rate constant k upon P in terms of a volume of activation A V z :

It is commonly argued that the value of AVf for a particular reaction can give insights into the reaction mechanism, since pressure not only affects bond making or breaking, with their attendant volume changes, but also can “tune” solvent properties that may influence k. These properties include the density p, static dielectric constant E , optical dielectric constant E , ~(usually equated to the square of the refractive index n), and viscosity 7. The argument presupposes, however, that AVf can be predicted quantitatively for a particular reaction mechanism. The thrust of much of the research in our laboratory over the past two decades has been to develop methods of predicting AVz for reactions of known mechanism for which adequate theoretical models are available, and to test the predictions experimentally. The primary test case has been outer-sphere electron transfer reactions of metal complexes in solution - an ostensibly very simple yet fundamentally important class of chemical and biological processes for which a well-developed kinetic theory is available [ 1-61 and can readily be extended to encompass pressure effects on k [7-lo]. By definition, no bond making or breaking is involved in outer-sphere electron transfer processes and, as we shall see, intramolecular reorganization does not normally affect AVz for such processes appreciably, so that AV# reflects in large measure the tuning of solvent properties by pressure. A particular virtue of pressure-tuning is that kinetically relevant solvent proper-

162

I

5 Insights into Solution Chemistryfrom High Pressure Electrochemistry

ties can be varied without changing the chemical identity ofthe solvent. Efforts to learn about solvent effects on reaction rates by comparing values of k measured in a range of solvents are dogged by the possibility that the microscopic modes of solvation of the reactants may be quite different in chemically different solvents: in other words, it may not be appropriate to regard the solvent merely as a dielectric continuum. Resort to mixed solvents to vary (say) the dielectric constant of the medium is arguably an even worse choice, since it is highly unlikely that the components of a mixed solvent will be equally effective at solvating a particular reactant, and some (usually unknown) degree of fractionation of the solvent in the solvation sheath relative to the bulk liquid is almost inevitable. Much of our earlier work on pressure effects on the kinetics of electron transfer reactions focused upon the rate constants k,, and corresponding volume of activation AVL for seljkchange reactions (5.3)

which are simplest to interpret because there is no net chemical change (so that AVex and the free energy of reaction AGex are both zero). For such reactions, an acceptable understanding of AVf has been achieved [9] and forms the basis for analyzing pressure effects on outer-sphere redox processes that involve net chemical change (“cross” reactions). For this purpose, it is necessary to determine the reaction volume AV12 of the cross reaction, and this is most readily obtainable from the pressure dependence of the electrode potentials of the contributing half-reactions relative to some common electrode. Thus, we have been led ineluctably into highpressure electrochemical measurements. Furthermore, practical limitations on the number of self-exchange reactions that can be studied experimentally at high pressures has prompted us to attempt to measure the rate constants k,l and associated volumes of activation AVZ of the corresponding electrode reactions instead, and to consider the relationship between the kinetic parameters of the two classes of reactions. Important insights into the mechanisms of electrode reactions corresponding to the reaction shown in Eq. (5.3) have resulted, and comprise the main theme of this chapter. As a preparatory step, however, it is necessary first to summarize what has been learned about pressure effects on the kinetics of bimolecular self-exchange reactions.

5.2

Pressure Effects on the Kinetics o f Self-Exchange Reactions 5.2.1 Principles

The rate constant k,, for an outer-sphere self-exchange reaction may be expressed in terms of its free energy of activation AG; through the equation:

k,, = K p ~ e c ~exp(-AGz/RT) ~e,

(5.4)

5.2 Pressure Effects on the Kinetics of Self-Exchange Reactions

Here, K is the electronic transmission coefficient ( K = 1 for adiabatic electron transfer) and vex the nuclear frequency factor, whereas K& is the equilibrium constant for assembly of a precursor state and effectively includes any coulombic work and medium (Debye-Hiickel) terms [ 4,51. Following the approach taken by Stranks [7], the observed volume of activation AVZ for a simple, adiabatic, outersphere, bimolecular electron transfer reaction can be represented as AVZ

=

-RT(d In k,,/dP),

= AVI$

+ AV& + AV&, + AV&

(5.5)

in which the terms on the right-hand-side represent the respective contributions of internal reorganization of the reactants (usually taken to mean bond length changes), the reorganization of the solvent, the coulombic work of assembling the precursor complex, and the pressure dependence of the ionic activity coefficients (Debye-Huckel terms). Stranks [7] showed that AVl$ is -0.6 cm3 mol-l for most rigid complexes, that is, barely outside the experimental uncertainty in AVZ ; this is because, as the two reactants distort to reach a common nuclear configuration to allow electron transfer to occur, the increase in volume of the oxidant almost cancels the decrease in volume of the reductant. In the simplest treatment, the other terms are given by

+ = [RTz(z+ l)C11/2/(l + B U I ' ~ ~ ) ~In] ~/dP),(3 [(~ + 2 B ~ 1 ' / ~-) p)]

A V ~ = , ~[N*Z(Z l)e2/4n~oaj(a~-'/aP)T

(5.7)

AV&

(5.8)

where a is the separation between the two M centers (usually taken to be the sum of the effective radii r,, and r,,d of the oxidant and reductant), a, B, and C are the Debye-Hiickel parameters, I is the ionic strength, is the isothermal compressibility of the solvent, and the other symbols have their usual meanings in SI. It turns out that AVg, and AV&, are opposite in sign and usually fortuitously cancel each other quite closely for practical values of I(O.1to 0.5 mol L I ) , at least in water. In non-aqueous solvents of low dielectric constant, however, AV&, and AV& can become numerically very large (for example, about f 5 0 cm3 mol-' in acetone) so that the calculation of AV: becomes unstable [9, 101. Furthermore, for any realistic combination of solvent parameters, AV& is predicted to be negative and to be the dominant term in Eq. (5.5). A more detailed version of the Stranks approach would incorporate the notion due to Sutin [4, 51 that electron transfer within the precursor assemblage {MLFfl)+,ML,Z+)occurs over a reaction zone of thickness Sa rather than on "hard-sphere'' contact of the reactants. In that case, cr must be treated as pressuresensitive, compressing along with the solvent. It turns out, however, that allowance for compression of cr in Eqs (5.6) and (5.7) can be neglected for adiabatic reactions because it is almost exactly cancelled by a term AV& representing the effect of compression on the pre-exponential part of the expression for I$!ec[9] (the exponential part generates AV&,). Electron transfer, however, could be non-adiabatic -

I

163

164

I

5 Insights into Solution Chemistryfrom High Pressure Electrochemistry

that is, the electronic coupling between the precursor and successor states could be weak enough for the electronic transmission coefficient K to become significantly less than 1. In that case, an exponential dependence of k,, on 00 - u (where go is the value of u for which electron transfer would be adiabatic) would result, such that there would be a significant additional negative non-adiabatic contribution AV& to AVZ . In summary, the key predictions of Eqs (5.5)-(5.8) are: (1) that AVsfR is usually the dominant part of AV: , and (2) that AV& (and hence AV): will be negative for simple outer-sphere electron transfer reactions in solution, regardless of whether electron transfer is fully adiabatic - with the caveat that predictions of the magnitude of AVZ are unlikely to be reliable for solvents of low E. 5.2.2

Experimental Observations with Aqueous Media

At the outset, it should be noted that Eqs (5.5)-(5.8) predict that IAVZ 1 should become numerically smaller with rising pressure because the isothermal compressibilities of solvents (and therefore also the pressure derivatives of properties such as E and n [S-lo]) decrease with rising pressure. In practice, plots of In k,, against pressure usually appear to be linear functions of P within the observational uncertainty over the usual experimental pressure range of 0-200 MPa. In this chapter. theoretical values of AV: are calculated for the mid-range pressure of 100 MPa, and the experimental AVZ values are, in effect, averaged to this pressure. Figure 5.1 shows that Eqs (5.5)-(5.8) reproduce the observed AVZ well in several cases (abbreviations: Figure 5.2). Agreement is less good for the C o ( d i a m ~ a r ) ~ + / ~ + couple, although the diprotonated form conforms well to expectations despite a rather wide error bar and the very high charges (5+/4+) [14]; solutions containing the 3+/2+ couple were buffered with morpholine/CF3SOjH, and it is suspected that some interaction with the buffer was occurring (perhaps significantly, the electrochemical transfer coefficient y for this couple was only 0.27 rather than the 0.4-0.5 normally found for the complexes considered in this chapter [lG]). In contrast, Table 5.1 lists cases in which Eqs (5.5)-(5.8) appear to fail. In each instance, however, reasons for the failure are recognizable. Thus, AVZ for the conjugate base pathway for the Fe(a~l)~+/’+ self-exchange is about 12 cm3 mol-’ more positive than expected for an outer-sphere mechanism because it is evidently of the inner-sphere type involving expulsion of a water ligand from the first coordination sphere of Fe(H20);+ to form an FeeOH-Fe bridge, and the AVZ value can be accounted for rather precisely on this basis [12]. The strongly negative AV: value for the cation-independent pathway for the Mn0,”- exchange remains enigmatic, but seems to reflect the very small sizes of the reactants, which approach the molecular dimensions of solvent water [ 191; what is important is that the Na+and K+-catalyzed pathways show much more positive AVZ values. Alkali-metal cation (M’) catalysis of anion-anion self-exchange reactions is dominant to the extent that Mf-independent pathways are ordinarily undetectable for the cyanometalate couples listed in Table 5.1 [20, 211, and the M+-mediated pathways show

5.2 Pressure Eflects on the Kinetics of SeFExchange Reactions I

I

I

I

I

I

Co(ttcn)z

Co(act)

3+/2+

3+/2+

I

-12

-8

-10

-6

-4

Calculated activation volume/cm

0

-2 3

rnol

-1

Fig. 5.1. Comparison o f observed AV: values with those calculated from Eqs (5.5)-(5.8). Diagonal line represents perfect agreement. phen 1,lo-phenanthroline; other ligand abbreviations a s in Fig. 5.2. Data from references [ll-15). 7

ttcn

tacn

Fig. 5.2. Ligand systems: act = azacapten; diamsar = diarninosarcophagine; sep = sepulchrate; ttcn = 1,4,7trithiacyclononane; tacn = 1,4,7-triazacycIononane.

I

165

I

5 lnsights into Solution Chemistryfrom High Pressure Electrochemistry Tab. 5.1. Experimental values of AV: that deviate markedly from the predictions AVL (calc) of Eqs (5.5)-(5.8).

Couple

AV: (exptj (cm3 mol-’)

AV; (calcj (cm3 mol-’)

Fe(HzO)sOH’+/Fe(aq)’+ Co(en):+/*’ Co(phen):MnO;/’MnO;‘*- (K+) Mn0,”- (Na+) Fe(CN):-I4 ( K + ) O~(CN);-’~ (K+) M ~ ( c N ) ~ - /(K+) ~MO(CN);-’~- (Et4N’) W(CN)i-’4- (K’) W(CN):-l4- (Me4N+) W(CN);-l4- (Et4N’)

-11 k 0.4 -16.5 k 0.P 4 7 . 6 k 0.7 17 f 2J

-

a

-1 1 1 k1 +14.7 k 0.6

-6 -6 -6 -6

12 17 18 19 19 19 20 21 21

k 0.6

-6

21

-6

21

~

l k l +3*1 +19

-8.2

+22+1 -7.4 -6.2

k 0.5 k 0.5

10

Ref:

-5 -2 -9

-6

-6

21

-6

21

Pressure dependent - average value, 0-200 MPa.

much more positive AVZ values than are expected for simple bimolecular electron transfer. These anomalies cannot be ascribed to ion-pairing - the AV: values are too positive to be accounted for solely on the basis of pressure-induced break-up of (supposedly) more reactive ion pairs. Besides, for cation-cation couples at least, ion-pairing seems to diminish reactivity in electron transfer [22]. A clue to the origin of the relatively positive AVZ values for cyanometalate couples with alkali-metal cations M+ is that the corresponding reactions in tetraalkylammonium media are similarly catalyzed by %N: but have negative AV; of about the magnitude expected from theory (although this last fact is fortuitous because the model does not consider mediation by a third body). Since M+(aq) ions have a definite first coordination sphere of water ligands but R4N+(aq) do not, one can infer that all cations facilitate electron transfer between anions by providing a region of positive potential to assist tunneling of the electron, but that the alkalimetal cations must be at least partially dehydrated first. Indeed, the relative effectiveness of various M+ as catalysts (Li+ < Na+ < Kt < Rbf < Cs+) follows their ease of dehydration - that is, it correlates with the (negative) heats of hydration of M+ [ 231. This interpretation also provides an explanation for the lack of significant counter-ion catalysis in cation-cation outer-sphere electron transfer reactions: involvement of an anion with two reacting cations would actually hinder tunneling by the negatively charged electron electrostatically, and there is evidence that this is the case [22]. The anomalous, strongly ne ative AV: values in Table 5.1 for the low-spin/highspin Co”!/“ couples Co(en):i’2+ and Co(phen)ifi2+ (and, on the basis of indirect evidence, for tris(bipyridyl)cobalt(III/II) [24] and Co(EDTA)-/Co(HEDTA)OH; [25]) contrast sharply with those for the low-spin/low spin Co(ttcn);”’- couple

5.2 Pressure Effects on the Kinetics of Self-Exchange Reactions

and the cage complexes C ~ ( a c t ) ~ + (low-spin/low-spin), /~+ C o ( ~ e p ) ~ + /and ~+, Co(diamsarH2)'+i4+ (low-spin/high-spin)(Fig. 5.1), for which AV,, is well represented by Eqs (5.5)-(5.8). There are basically two possible explanations for the anomaly. One (which we favored rather arbitrarily in our earlier publications 18, 17, 18, 251) is that electron transfer occurs directly between high-spin Co" and lowspin Co"' but is non-adiabatic by virtue of the simultaneous large change in spin multiplicity - in other words, that the electron transfer is "spin forbidden", much as are intramolecular electronic excitations involving spin state changes. As noted in Sect. 5.2.1, non-adiabaticity implies an exponential dependence of k,, on E ( G ~- a ) , where a is a distance scaling parameter, and would contribute a markedly negative amount AV,f, = -2ap~RT/3 to AVZ; a reasonable value of a of about 19 nm-' would be required to account for the observed values. This explanation, however, fails because it is inconceivable that electron transfer in the low-spin/ high-spin CO"'/" cage complexes could be adiabatic (as is evidently the case) if that in the smaller Co(en):+'2+ complexes, from which they are derived, were nonadiabatic [ 14, 261. The alternative is that a spin-change equilibrium (presumably in the Co" species) precedes electron transfer; such equilibria are rapidly established [ 271, and, if the Co(terpy);+ (terpy = terpyridine) high-spin/low-spin equilibrium [28] is typical, can contribute AVSpin -10 cm3 mol-' to AVZ - that is, account for the deficit in AV: . The spin pre-equilibrium model, however, begs the question as to why the Co cage couples should conform to the predictions of Eqs (5.5)-(5.8) even in lowspin/high-spin cases. The answer seems to be that flexible chelate complexes such as the tris(bidentate) Co(en):+ and Co(phen):+ or even the sexadentate Co( HEDTA)OH, are subject to substantial Jahn-Teller-relateddistortions and associated volume decreases on going from high to low spin states, whereas these distortions are suppressed in the c o cage complexes. To test this idea, we have recently examined 1291 the pressure dependence of the visible absorption spectrum of aqueous Fe(diamsarH~)~+ [ 30-331, for which the high-to-low spin equilibrium constant (0.34) is uniquely suitable for this purpose (although the 0 2 sensitivity of the iron(I1) complex poses severe experimental difficulties). The low-spin form of the complex has a ligand-field absorption maximum at 569 nm, at which the absorbance of the high-spin form is negligible. No change in the compression-corrected molar absorbance at 569 nm could be detected within the experimental uncertainty over the pressure range 0-294 MPa, implying AVspin = 0.0 k 0.2 cm3 mol-'. Unfortunately, direct determination of AVL for the Fe(diamsarH2)' +I4 self-exchange reaction is not feasible, but high-pressure electrode kinetics (see below) indicate that it is close to the theoretical value of about -5 to -6 cm3 mol

-

'.

5.2.3

Experimental Observations with Non-aqueous Media

For self-exchange reactions of multiply charged complexes in non-aqueous solvents of low dielectric constant, Eqs (5.5)-( 5.8) become numerically unstable because of runaway values of the Coulombic and Debye-Huckel terms, as noted

I

167

168

I

5 fnsights into Solution Chemrstryfrom High Pressure Electrochemistry

in Sect. 5.2.1. Moreover, for multiply charged solutes, ion pairing will also become important, with unpredictable kinetic consequences (although mild retardation by ion pairing, resulting in a negative contribution to AVL, seems likely for cation-cation exchanges [ 221). If, however, one reactant is electrically neutral and the other is singly charged, the terms AVZ,, and AV& become zero, ion pairing becomes minimal (and affects only the ionic reactant), and Eqs (5.5)(5.8) can be expected to perform well. Then, since AV,< is typically small, AVL can be identified with AV&. This seems to be the case for the Ru(hfac);’- selfexchange (hfac- = hexafluoroacetylacetonate ion), for which AVL in methanol, acetone, and acetonitrile is well accounted for by either Eq. (5.6) [34] or a model that takes the size of solvent molecules into account through the Mean Spherical Approximation (MSA) [ 351. Thus, the Ru(hfac)y’- self-exchange evidently proceeds by an adiabatic outer-sphere electron transfer, without significant involvement of ion pairing, in these three solvents at least - a significant point to which we will return later - although in chloroform there were indications of ion-pairing, as might be expected in view of the very low dielectric constant of this solvent [ 341. Practical problems including solubility and redox stability have so far limited comparisons of kinetic pressure effects in aqueous and non-aqueous media to just one couple - the Fe(phen),3+/2- self-exchange reaction [ll],for which AVL was found to be -2.2 cm3 mol-’ in aqueous HzS04, in excellent agreement with the prediction of Eqs (5.5)-(5.8) (Fig. 5.1), and -5.9 cm3 mo1-l in acetonitrile (cf. -6.8 cm3 mol-l predicted using an estimated pressure dependence of the refractive index for acetonitrile [8] and ignoring possible ion pairing). As explained in Sect. 5.4.2, these results rule out any significant influence of solvent dynamics in the self-exchange reactions studied to date at high pressures. 5.2.4

Approaches Involving Electrochemistry

The AVZ data of Fig. 5.1 that are satisfactorily accounted for by Eqs (5.5)-(5.8) are fewer in number than the anomalous cases of Table 5.1. This is a rather unsatisfactory situation, even though most of the anomalies can be explained away - indeed, deviations from the predictions of Eqs (5.5)-(5.8) can often provide important mechanistic information. More AVZ data are clearly desirable, but the prospects for further successful experiments are poor. The measurements of AVZ summarized in Fig. 5.1 and Table 5.1 were obtained at high pressures by radiochemical tracer methods for the slowest reactions [12, 17, 251, N M R linebroadening techniques for the faster cases [11, 13, 15, 19-22, 341, and stopped-flow circular dichroism [ 13, 14, 181 for moderately rapid reactions of reactants that could be prepared as resolved enantiomers. There are, however, many selfexchange reactions that are inaccessible to these techniques. For example, rates of electron transfer in couples where both reactants have unpaired electrons generally cannot be studied by NMR methods, while other couples that undergo electron transfer at intermediate rates may not be resolvable into optical isomers or be amenable to radiochemical sampling procedures under pressure.

5.2 Pressure Efects on the Kinetics of Self-Exchange Reactions

In principle at least, two further indirect methods exist for determining AV.: Both involve high-pressure electrochemistry. One is the measurement of the pressure dependence of the rate constant k,] for electron transfer in a given couple at an electrode, but it is not immediately clear how k,, and the corresponding volume of activation AVZ relate to k, and AVL , respectively, for the self-exchange reaction of the same couple. This is a major theme of this chapter, and is pursued in detail below. The other method involves invocation of the "cross relation" of Marcus [5], which expresses the rate constant kl2 for the oxidation of, say, A by Bt in terms of its equilibrium constant K12 and the rate constants k l l and k22 for the respective At/A and B+/B self-exchange reactions:

where

Here, Z is the frequency factor (assumed to be the same for all the reactions) and are the work terms (assumed to be the same for the self-exchange and the forward and reverse cross reactions). For the cross reaction, K12 can be obtained from the difference AE' in the standard electrode potentials for the two self-exchange reactions: w,i

In

~ 1 = 2

~FAE'/RT

(5.11)

where n is the number of (moles of) electrons transferred (almost always 1 in the present context). The volume relationship corresponding to Eq. (5.9) [3G] is then

+ AVZ + AVi2)/2] + X

(5.12)

AV;

= [(AVZ

X

-(RT/2)(8 lnf/dP)r

(5.13)

-nF(8AEo/dP),

(5.14)

=

AV12 =

For low driving potentials AE', f can be approximated to 1 and X can be neglected. In practice, Eq. (5.12) is accurate with the inclusion of X if AE' is small or moderate, as in the reduction of aqueous Fe(H20)2+ by Co(ttcn);+ (AE' = 0.28 V) [3G] or ofCo(phen);+ by Ru(en):+ (AE' = 0.19 V) [15],but fails for large driving potentials, as in the reduction of Fe(H20);' by Co(sep)2- (AE' = 1.03 V) and of Co(bpy);+ by Co(sep)'+ (AE' = 0.66V) [24]. Where Eqs (5.12) and (5.13) fail, the Marcus cross relation itself (Eqs (5.9) and (5.10)) also fails. The emerging picture is that selfexchange reactions in water are generally adiabatic, but cross reactions may become increasingly non-adiabatic as one goes to higher and higher driving potentials AE'. Alternatively (perhaps equivalently), some of the assumptions implicit in the Marcus theory (e.g. that the precursor and successor states may be treated as harmonic oscillators) may fail at high AE'.

I

169

170

I

5 Insights into Solution Chemistryfrorn High Pressure Electrochemistry

Thus, the cross-reaction approach, judiciously applied, can deliver AV: data that are otherwise inaccessible. The converse, however, may be more important - that is, if an adequate database of AVZ values is available, then only the easily measurable pressure dependence of AEo is needed to apply Eqs (5.12)-(5.14) to predict AVZ for numerous cross reactions (algebraic summation of partial molar volumes of all the relevant solute species is an alternative route to AVI,, but such data are rarely available or relevant to practical reaction media). Comparison of predicted and experimental AV; data could then provide mechanistic insights. Thus, highpressure electrochemical measurements are central to further progress in this field.

5.3

Electrochemical Measurements at High Pressure

The techniques of electrochemistry at high pressure have recently been reviewed [37, 381, and need not be reconsidered in detail here. In essence, the electrochemical cell consists of a Teflon body containing a working electrode (usually an Au or Pt wire or a glassy carbon fiber), a counter-electrode(typically Pt wire), and a reference electrode (usually either a AgC1-coated Ag rod in a compartment containing 4 mol L-l KC1 solution, or an Ag/Agf electrode consisting of a Ag rod in AgC104 solution). The reference electrode has a Vycor frit mounted in a free piston to allow for ionic contact and compression of the solution in the reference compartment, and the cell as a whole also has a free piston to accommodate compression of the reaction medium. The cell is mounted in a pressure vessel with electrical feed-throughs that are soldered to the electrode connections. Hydrostatic pressure is applied with a suitable electrically insulating liquid (usually clean mixed hexanes) which in turn is isolated from the hydraulic oil of the pump with a free-piston separator vessel to minimize contamination. Pressure is monitored on either side of the separator with either electronic or Bourdon gauges, calibrated against a standard deadweight gauge. The temperature of the pressure vessel is controlled with thermostated water circulating in an aluminum jacket. Some 3050 min are required for thermal equilibration of the pressure vessel and its contents at 25.0 "C, either initially or after the pressure is raised or lowered, so that a cycle of pressure measurements lasts several hours and the chemical stability of the system under examination must be accordingly high. Success in high-pressure electrochemical measurements depends largely on having clean, reproducible working electrode surfaces, which are therefore polished with fine alumina and sonicated in clean solvent before assembly of the electrochemical cell. Before a sequence of measurements at each pressure, the electrodes are cleaned electrochemically by cycling the applied potential several hundred mV either side of the potential of the couple of interest (preferably to the point of hydrogen evolution, in aqueous media). Electrode potentials relative to the particular reference electrode are conveniently measured as the half-wave potential E1p in cyclic voltammetry (CV) [IS, 16, 20, 24, 36-48] or cyclic staircase voltammetry [49-51]: or by differential pulse voltamme-

5.4 Homogeneous Versus Heterogeneous Electron Transfer

try (DPV) [ 5 2 ] . In principle, electrode reaction rate constants k,, can be estimated from the peak-to-inverse-peak voltage separation in CV [ 5 3 , 541, but our experience has been that the peak separation is neither sufficiently accurately measurable nor reproducible. Good reproducibility and precision in k,l can, however, be obtained using alternating current voltammetry (ACV), which involves imposing a small ( 25 mV) AC pertubation of angular frequency w on a DC voltage ramp, and recording the maximum in-phase and gO"-out-of-phasecurrents (at voltage Emax)as the DC voltage is swept through the redox potential. From the observed Emaxand Ell2, the electrochemical transfer coefficient y can be calculated. If Em, = Ell2, as is most often the case, then y = 0.5. From the ratio of the two currents, properly corrected for uncompensated resistance, a phase angle a, can be calculated, whence (for the simplest case where y = 0.5) k,l can be obtained from the linear plot of cot a, against 40): cot a, = 1

+ ( D o112D,112/2)112~112/kel

(5.15)

where Do and DRare respectively the diffusion coefficients of the oxidized and re112 112 duced forms of the reactants. For practical purposes, Do D, may be replaced by a mean diffusion coefficient D, which can be obtained from the maximum current of a cyclic voltammogram using a couple of known D as a standard. Once thermal equilibrium is achieved at a selected pressure in a cycle of measurements, multiple cyclic voltammograms are run to check the reproducibility and integrity of the system and to measure D, and then AC voltammograms are taken repeatedly until reproducibility is obtained. The pressure dependences of Ell2, In D, and In kel then give AV,,, (the volume of reaction of the couple relative to the particular reference electrode), AV& (the volume of activation for diffusion of the reactants), and AV:, respectively (cf. Eqs (5.2) and (5.14)). It is our practice to accept only those variable-pressure cycles of measurements of E1p, D and k,l for which the initial and final low-pressure data agree to within the experimental uncertainty, although, in some isolated cases in which measurements show a slow but constant rate of drift, provisional AVZ values may be extracted by correcting for the drift. Figure 5.3 shows some typical results for Ell2 and In k,, in an aqueous system; as explained below, the pressure dependence of D for aqueous systems at 25 "C is negligible in the 0-200 MPa range. Further details of the experimental procedure and associated calculations are given elsewhere [lG, 24, 40, 41, 551.

5.4 Homogeneous Versus Heterogeneous Electron Transfer 5.4.1 Aqueous Solutions

Marcus [56-59] has shown that the free energy of activation AG; for an electrode reaction can be expected to be one-halfof that (AGZ) for the corresponding selfexchange process, so that, from Eq. (5.4), we have for adiabatic electron transfer

I

171

172

I

5 Insights into Solution Chernisttyfrom High Pressure Electrochernistrv 0.54

I

I

1

I

I

I

I

0

50

100

150

200

V Y

1.

”,

0.52

d \ 0,

6 v)

’

0.50

>

h

I-

9

0.48

-2.0

--

-2.2

I

Lo

E 0

-2.4

\ A? v

5 -2.6 -2.8

Pressure/MPa Fig. 5.3. Pressure dependences of In k,, (by ACV) and E l p (by CV) for the W(CN)i-’4- couple (2 mmol L-’) at a Pt wire electrode i n in aqueous KCI (0.5 mol L-’) a t 25.0 “C (411.

( K = 1):

(5.16) where Zex and Zel are the pre-exponential factors taking the place of K&cilexand I$&ve~ respectively in Eq. (5.4) and its equivalent for electrode reactions. This relation may be intuitively understood by recognizing that two reactant species undergo internal and solvational reorganization to a common intermediate configuration to allow electron transfer to occur in the case of (bimolecular) self-exchange in homogeneous solution, but only one has to reorganize to that configuration in the corresponding heterogeneous electrode reaction. Marcus [ 561 pointed out that “=” should be replaced by ‘ ‘ 2 ” in Eq. 5.16 if the reactant center-electrode distance is more than half the metal-metal separation (r for the homogeneous electron transfer reaction (due, perhaps, to a layer of adsorbed solvent or other molecules on the electrode surface). Hush [GO] independently arrived at conclusions

5.4 Homogeneous Venus Heterogeneous Electron Transfer

similar to those of Marcus, but cautioned that the role of image charges in the electrode is not clear in solutions with practical supporting electrolyte concentrations which substantially reduce (or almost eliminate) the diffuse double layer potential. In addition, the measured rates of electrode reactions are notoriously sensitive to the nature and history of the electrode surface. Thus, it is not surprising that plots of In k,, vs. In k,,, which might be expected to be linear with slope 0.5 on the basis of Eq. (5.16) if Zel and Z,,were approximately constant from one reaction to another, may be severely scattered with a “slope” (insofar as it can be defined) as low as 0.1 [16]. Cannon [ G l ] , however, did find a loose correlation between some In k,l and In k,, data that leveled out (either because of the incursion of diffusion control or of incompletely compensated resistance [62]) as k,, approached 1 cm SKI. In sharp contrast, a plot of AV: against AVZ for almost all aqueous reactions for which both parameters are currently available shows an excellent linear correlation of slope 0.50 f 0.02 and negligible intercept [I61 (according to a recent reinvestigation [20], the Fe(CN)i-/4- case may be anomalous and should be excluded). For convenience, we refer to this relationship as the “fifty-percent rule”. The evident equivalence between AV; and AV: / 2 implies that the pre-exponential factors Zel and Z,, for aqueous systems are effectively independent of pressure. Such is known to be the case for Z,,, from experience with the success of Eqs (5.5)-(5.8) for self-exchange reactions, and presumably it also applies to Zel because the nature of a given electrode and its surface are virtually independent of pressure, even though the pre-exponential factor may differ widely from one electrode to another. Zel could, however, be influenced by solvent dynamics, but, as demonstrated in Sect. 5.4.2, this would not be apparent in AVZ values for dilute aqueous systems at near-ambient temperatures. Little is known about the effects of pressure on the electrical double layer [G3], but compaction of the diffuse double layer by the high concentrations of supporting electrolytes used in our experiments means that double-layer contributions to AVZ can be expected to be small. In any event, the close equivalence between AV: and AV: /2, regardless of the medium, indicates that double layer effects in aqueous systems can be empirically neglected as far as pressure effects are concerned. It also implies that the metal center-electrode separation can indeed be taken to be simply one-half of 0 for the bimolecular electron transfer [56]. Thus, if there is an adsorbed layer of solute or solvent on the electrode, it either acts as a conducting surface or can otherwise be ignored for our purposes. The fifty-percent rule implies that AVZ for aqueous electrode reactions can be predicted from Eqs (5.5)-( 5.8). Thus, for the Fe(diamsarH~)’+/~+ couple, for which we saw in Sect. 5.2.2 that there is a detectable spin equilibrium but with AVspin = 0.0 _+ 0.2 cm3 rnol-’ 1291, one can predict AVZ = 0.5AV: = -2.9 cm3 mol-’ (neglecting the possibility of a small positive contribution on the order of 0.3 cm3 mol-’ from internal rearrangement). ACV measurements [29] on Fe(diamsarH2)’+I4+ (1 mmol L-’) in aqueous HC104/NaC104 (0.1/0.4 mol L-’) on a Pt electrode at 25.0 “C gave k,, = 0.08 cm s-’ and AV: = -2.4 0.5 cm3 rnol-’. These data are consistent with rapid adiabatic electron transfer from the low-spin isomer (which

I

173

174

I

5 Insights into Solution Chemistryfrom High Pressure Electrochemistry

comprises about 25 % of the total Fe), and the AVSpm and AV: values confirm the postulate in Sect. 5.2.2 that cage complexes conform to the predictions of Eqs (5.5)-( 5.8) because distortions associated with spin-state changes are suppressed. 5.4.2 Non-aqueous Media

The foregoing rather tidy picture of pressure effects on outer-sphere electron transfer in aqueous systems is based on the supposition that transition state theory (TST), upon which the Marcus approach is based, is applicable to both homogeneous and heterogeneous electron transfer reactions. In other words, it is assumed that the transition states are present in minute concentrations in equilibrium with the reactants in their initial states. This is probably a valid assumption for homogeneous (bimolecular) electron transfer, but evidence has been accumulating to the effect that solvent dynamics (solvent “friction”) may limit the rates of some (or even all) electrode reactions. In simple terms, the expectation is that coupling of solvent motions to the passage of the reactants through the transition state may act to diminish the pre-exponential factor Zel, which can then be expected to show an approximate proportionality to thefluidity of the solvent - that is, to the reciprocal of its viscosity I.‘ (The relevant solvent parameter is actually the longitudinal relaxation time TL, but Zel K ?-’ can be substituted for proportionality to r;’ for Debye solvents at least.) In that case, if kTsT is the rate constant expected for an electrode reaction on the basis of transition state theory, the rate constant kso observed under solvent dynamical control is given by

Such phenomena have been reported for electrode reactions of Cr(EDTA)-/2- and Fe(CN)i-/4- in which the viscosity of the solvent (water or dimethyl sulfoxide) is varied by the addition of dextrose or sucrose [ 64-66]. Solvent friction has also been invoked in reductions of thiophenecarboxylatopentaamminecobalt( 111) complexes [67] and of metallocenes [68, 691 at Hg electrodes in organic solvents. Of particular interest in the present context are studies by Murray et al. [70, 711 of the reduction kinetics of C~(bpy);+/~+ in a variety of organic solvents, in which k,, was found to be proportional to I/?, to l/tL,and to the mean diffusion coefficient D of the reactants over a huge range (11 orders of magnitude) of k,]. Such studies, however, rely upon changing the chemical identity of the solvent in order to vary q ,7, and D,with the attendant risk that specific solvation phenomena or (in the case of mixed solvents) differential solvation may be responsible in part for the observed effects. This particular conundrum may avoided by the use of high pressures to “tune” the solvent properties, because the viscosity of normal liquids rises sharply and exponentially with increasing pressure (Fig. 5.4). A striking exception to this generalization is the viscosity of water at nearambient temperatures. Water may be considered to consist of a mixture of transient ice-I-like structures and free water molecules; break-up of the former by

5.4 Homogeneous Versus Heterogeneous Electron Transfer 3.5

oo 3.0 m

c x .-

2.5

v)

0 V

.-

u)

>

2.0

> .c U 0)

CE

1.5

1 .o 0

50

100

150

200

Pressure/MPa Fig. 5.4. Pressure dependence o f some typical solvents, based on data from references [72] (water, 25 "C), [73] (acetonitrile and N,N-dimethylformamide, 30 "C),[74] (acetone, 30 "C). and [75] (methanol, 30 "C).

increasing pressure results in a decrease in viscosity, whereas the free molecules behave as expected for a normal liquid. Thus, at temperatures near the freezing point, v] first falls and then rises as the pressure is increased, whereas around 25 "C these opposing trends cancel almost exactly, resulting in a near-independence of q on pressure in the range 0-200 MPa considered here (Fig. 5.4). Consequently, any dependence of k,l on solvent dynamics will be reflected in a markedly positive contribution to AVZ for outer-sphere electron transfer reactions in typical organic solvents, but none at all when water is the solvent at near-ambient temperatures. The fifty-percent rule for activation volumes, then, is fortuitously valid for aqueous systems even if solvent dynamics govern Zel. Indeed, AVZ for Co(bpy)i+/2+in aqueous NaCl is strongly negative (Table 5.2), and closely similar in magnitude to AVZ for the analogous couples Co(phen):+/2i (-9.1 cm3 mol-') and Co(en):+/2i (-8.3 cm3 mol-l) [ 161. Since AVZ values for the latter two couples follow the fiftypercent rule, one can infer that the (as yet unmeasured) AVZ for Co(bpy):+'*+ is -17 f 1 cm3 mol-l. For non-aqueous systems, however, AVZ for Co(bpy):+"+ is strongly positive and comparable to AV& (= -RT(d In DlaP),), at least for the three solvents for which good, reproducible, high-pressure electrochemical measurements could be made (Table 5.2). From the Stokes-Einstein relation D = kgT/cxrq,

(5.18)

I

175

176

I

5 lnsights into Solution Chemistryfrorn High Pressure Electrochemistry Tab. 5.2.

Kinetics of the Co(bpy);”*’ electron transfer reaction at a platinum electrode”.

Solvent

~

Ell2 (mv) AV~~II k:l (cm s-l) DO c m 2 SKI) AV: (cm3 mol-l) AV& (cm3 mol-’)

~~

___

~

(125)p (26.8 f 0.9)e 0.172 6.06 -8.6 f 0.4 0.37 0.03

4f.g 15.9 0.4‘ 0.198 10.6 9.1 k 0.3 8.8 f 0.1

+

0‘ 12.3 k 1.0‘ 0.070 9.41 10.2 f 0.7 11.1 f 0.1

-52f‘ 14.0 f 0.4‘ 0.0160 1.40 12.2 k 0.9 15.6 f 0.4

aData from ref. [24]; 1.0 mmol L-’ Co; 25.0 “C; transfer coefficient y = 0.5 in all cases. b0.20 mol L-’ NaCI. <0.20 mol L-’ Bu4NC104. Propylene carbonate. Reference electrode Ag/AgCl/(4 mol t KC1). ‘Reference electrode Ag/(0.013 mol L-’ AgC104-0.20 rnol L-’ Bu4NC104-CH3CN). g-85 10 mV vs. Fe(C5H5):”.

’

where c is a constant (6 and 4,respectively, for the sticking and slipping limiting models of diffusion) and r is the hydrodynamic radius of the reactant, it follows that AV& measured from high-pressure CV may be equated to the volume of activation AVACfor viscous flow of the solvent (= RT(3 In r / / 3 P ) T ) . In the simplest solvent-dynamical interpretation, then, we can expect (5.19)

Since AV& is large relative to the expected values of AVGTST) (that is, AVZ/2, which can be either measured or calculated from Eqs (5.5)-(5.8)), we can expect AVZ and AV& to be similar in sign and magnitude if solvent friction is indeed important and Zel i s roughly proportional to l/a. Because the TST-based Eqs (5.5)-(5.8) in combination with the fifty-percent rule predict negative values for AV; in the absence of special mechanisms such as cation catalysis of anion-anion electron transfer, the observation of positive AVZ values is a very strong indication of the incursion of solvent dynamics, especially when (as in the C~(bpy):+’~+case) a negative AV; is found for the same electrode reaction in aqueous media. We have used this last criterion in an attempt to detect solvent dynamical influences in homogeneous electron transfer (self-exchange) reactions [ll].Because solvent friction is most likely to emerge when the internal reorganizational barriers to electron transfer are lowest, the reaction chosen was the Fe(phen):+’2+ self-exchange, which is a fast low-spin/low-spin reaction with negligible change in Fe-N bond lengths and also can be followed in both aqueous and acetonitrile solutions. For the reaction in aqueous H2S04,AVL was found to be -2.2 cm3 mol-’, in excellent agreement with the TST prediction (Eqs (5.5)-

5.4 Homogeneous Versus Heterogeneous Electron Transfer

(5.8)), but in acetonitrile AVL was more negative (-5.9 cm3 mol-') than for water,

again in good agreement with the predictions of TST but ruling out rate control by solvent dynamics. Solvent friction, then, does not show up even in one of the most favorable bimolecular electron transfer reactions - unfortunately, there are exceedingly few self-exchange reactions for which AVk can be measured in both aqueous and organic media. In sharp contrast, Matsumoto [7G] has studied pressure effects on the Fe(phen)i+'2t reaction at a Pt electrode in acetonitrile containing 0.5 mol L-' tetrabutylammonium perchlorate, and found AV: +15 cm3 mol l , cf. AV: = -1.G cm3 mol-' in 0.1 mol L-' aqueous NalS04 and the negative values for self-exchange given above. It may be concluded that solvent dynamics dominate electrode kinetics in non-aqueous media even when the corresponding selfexchange reactions clearly conform to the TST model. One serious obstacle to interpretation of kinetic data in non-aqueous solvents having dielectric constants substantially lower than that of water is that ion pairing is almost inevitable, at least for the more highly charged couples such as C~(bpy);+'~+. Pyati and Murray [70] considered but rejected the possibility that their attribution of the kel-y correlation for this couple to solvent dynamical influences was vitiated by ion-pairing. Nevertheless, by going to a couple with large partners, one electrically neutral and one singly charged, one can reduce ionpairing to the absolute minimum and also completely remove the Coulombic and Debye-Hiickel contributions to AV: . We have therefore examined the electrode kinetics of the Ru(hfac)!'- couple at high pressures in acetonitrile, acetone, and methanol (for which AVZ data were already available [34]), and propylene carbonate (PC; efforts to extend the investigation to other polar solvents were unsuccessful) (551. The results are summarized in Table 5.3, and can be interpreted in detail following consideration of the background to Eqs (5.17) and (5.19).

-

Tab. 5.3.

Electrochemical characteristics of Ru(hfac)y'- at a Pt wire electrodea.

Solvent

CH3CNb

(CH3)zCOc

CH30Hd

Ell2 (mVf) AVceII' k,9 ( c m s-') AV: (cm3 mol-') AV& (cm' mol-I) AVZ (cm' mol-'g) AV: (calc)h

431 f 2 -24.8 k 0.8 4.0f 0.1 8.4& 0.5 12.1 0.7 -5.5 f 0.1 9.0f 0.7

414 f 2 -19.8 & 0.2 2.44 0.07 12.4 f 0.7 15.4f 0.7 -6.1 f 0.3 12.0f 0.7

364 2 -18.3 & 0.8 4.1 & 0.1 9.9 f 0.4 14.7k 0.7 - 5.8 k 0.3 11.4 k 0.7

*

+

aData from reference r551; lRul = 2.0-3.5mmol L-I, reference electrode Ag/(O.Ol moi L''-AgC104-0.20 rnol L-' Bu4NC104-CH3CN), 25.0 "C. Supporting electrolyte 0.20mol I>-' [Bu4N]C104. c0.20rnol L-' [Bu4N]PF6. d0.50 rnol L-' [Bu4N]C104. '0.50 rnol L-' [ B u ~ N J P F ~ . 'Relative to Ag/(O.Ol mol L AgC104-0.2 mol L-' Bu.+NC104-CH,CN), g Reference [ 34). "Calculated from Eq. (5.23)for the mid-range pressure of 100 MPa.

'

*

PCe

369 f 2 -13.0 f 0.5 2.26 f 0.03 11.2 f 0.3 18.5f 0.2 -

-

I

177

178

I

5 Insights into Solution Chemistryfrorn High Pressure Electrochemistry

The question of the possible role of solvent dynamics in controlling the rates of chemical (and particularly electrochemical) reactions has been widely considered 168, 77-85]. There are basically two approaches to solvent-dynamical interpretation of the data of Table 5.3. One, originated by Kramers as long ago as 1940 [SG] and later developed by Grote and Hynes [87], regards the motion of the solvent as coupled to the reaction coordinate in a one-dimensional manner analogous to Brownian motion, assisting the reactant(s) to climb the activation barrier but: in the case of strong coupling, also hindering their passage across the barrier top, resulting in several crossings and re-crossings before moving on to product configurations. The other approach, introduced by Agmon and Hopfield [88, 891 and developed further by Sumi, Marcus and Nadler [90-921 and subsequently by Basilevsky, Ryaboy and Weinberg 193, 941, represents the activation process as a freeenergy surface with the reaction coordinate (in effect, a low-frequency intramolecular vibration of the reactant( s)) represented along one dimension and diffusive solvent motions along another coordinate orthogonal to it. These alternative views have been summarized succinctly by Asano [ 771 (see also Chapter 3 ) . For adiabatic electrode reactions with a negligibly small internal reorganization energy AG;, the Kramers-Grote-Hynes model leads to (5.20)

v,l = ~ ; ' ( A G & / ~ K R T ) ' ' ~

where AC& is the contribution of solvent reorganization to AG;. If the solvent behaves as a Debye liquid (that is, one in which molecules can be treated as spheres rotating in cavities in a continuous dielectric) of Debye relaxation time 7 D , and VM is the effective hard-sphere molar volume of the solvent molecules, q can be obtained (at least approximately) from '7 ;

=

(E/E,~)TD~= ( E / E , , , ) R T / ~ V M ~

(5.21)

Thus, the dominant variable in the pre-exponential factor of the expression for k,, (cf. Eq. (5.4)) is q, giving rise to the approximate proportionality (5.17) and so to equation (5.22): A V i = A V & . + RT(a In q / d P ) , = AV&,

+ AV&

= AVA,

+ AV&

(5.22)

A more precise version of Eq. (5.22) for the theoretical value AVZ(ca1c) of AVZ includes a term Q representing the small, largely self-canceling pressure dependences of the factors other than q in Eqs (5.19) and (5.20). For the Ru(hfac)t' couple in the solvents listed in Table 5.3, Q is only -0.3 to -0.4 cm3 mol-'. Furthermore, since AV& is the value of AV: that would be expected in the absence of solvent friction, we can use the fifty-percent rule to estimate it as AVZ /2: AV: (calc) = (AV: /2)

+ AV& + Q

(5.23)

5.4 Homogeneous Venus Heterogeneous Electron Transfer

Table 5.3 shows that AV$(calc) agrees with the experimentally observed AVZ to within the experimental uncertainty for the Debye solvents acetone and acetonitrile, and comes surprisingly close for methanol which, because intermolecular hydrogen-bonding contributes several frequencies to the apparent rL, is not considered to be a Debye liquid. Unfortunately, there are insufficient data for application of Eq. (5.23) to the Ru(hfac)y'- electrode reaction in propylene carbonate, which is also regarded as a non-Debye liquid. In any event, the implication is that the fifty-percent rule applies to volumes of activation for electrode reactions but its effect is swamped by solvent dynamical contributions. The applicability of Eq. (5.23) to the Ru(hfac)!'- electrode reaction may be questioned on the basis that AG& is only slightly less than AG&. The Sumi-Marcus approach (901, however, does cover cases in which AG$ is not negligible, and we find that replacement of the parameter Q by one derived by differentiation of Eqs (8.6) and (8.7) in reference [90] with respect to pressure, results in no significant change in AV:](calc), within the experimental uncertainty. There is also the possibility, raised by Weaver and co-workers [95], that the Ru(hfac);/- electrode reaction may be non-adiabatic (li << l),but in that case one would expect v , ~to be proportional to r;" where 1 2 0 2 0, with 0 reaching 1 only when the reaction is fully adiabatic [92, 961. In fact, the results of Table 5.3 imply that 0 is close to 1, and in any event non-adiabaticity is expected to make a substantial negative contribution to AVZ, whereas the problem is to explain why AV: is so positive 1551. Clearly, AV: information for other O/- or +/0 couples in non-aqueous media are needed to vindicate the arguments leading to Eq. (5.23), but there is very little AVZ information for such systems. For the Fe(qS-CsH5)2f/' self-exchange reaction in aceonitrile, Hunt and co-workers reported AVZ = -7 cm3 mol-l [97], but efforts [98] to obtain AV; for the electrode reaction have been thwarted by excessive drift of the apparent k,l over the time span of a series of pressure measurements. It is, however, clear that AVZ is markedly positive for the Fe(qS-C~H5)i/' electrode reaction in acetonitrile, consistent with Eq. (5.23) in at least a qualitative way. Recent studies also reveal AVZ values of +6 and +ll cm3 mol-' for the Mn(CN-chex):+/+ electrode reaction in CH3CN and acetone, respectively [ 7 G ] , whereas for the corresponding self-exchange reaction AV: = -17 and -20 cm3 mol-' [99]. For this Mn"/' couple, the interpretation of the volumes of activation is likely to be complicated by ion-pairing and medium effects, but these and all other data collected to date point to a striking qualitative difference between self-exchange and electrode reactions: the respective volumes of activation differ unambiguously in sign, and AV: for the electrode reactions shows some relation to AV&. The only likely explanation (since none of these reactions seems to be diffusion-controlled) is that the electrode reactions are subject to solvent dynamical effects, but the selfexchange reactions are not. The reason for this difference is not yet certain but may be due in large part to the fifv-percent rule, as applied to the respective free energies of activation. Thus, AG: is expected to be just one-half of A G Z , which itself is not normally very high, and the incursion of solvent dynamical effects is expected for reactions with a low internal free-energy barrier.

I

179

180

I

5 Insights into Solution Chemistryfrorn High Pressure Electrochemistry

5.5

Conclusions

For simple outer-sphere self-exchange reactions of transition-metal complexes in both aqueous and polar organic solvents, AVZ is dominated by AV& and hence is expected to be negative, regardless of whether electron transfer is fully adiabatic. In cases in which this expectation is not realized, there is usually an identifiable departure from a simple outer-sphere mechanism attributable to inner-sphere pathways, cationic catalysis of anion-anion electron transfer, or structural distortions associated with spin multiplicity changes. Thus, AVZ can serve as a rnechanistic criterion. For bimolecular outer-sphere electron transfer reactions involving net chemical change (“cross” reactions), the volume of activation AVZ can be successfully predicted from Eqs (5.12)-(5.14) if AVZ for the two constituent self-exchange couples is known (or can be estimated from Eqs (5.5)-(5.8)) and the pressure dependence of the difference AE in their electrode potentials is measured by high-pressure electrochemistry. Such predictions may fail if AE is too large, or if the reaction mechanism is other than simple outer-sphere. Thus, measurements of AVL combined with high-pressure studies of AE can also provide mechanistic criteria for redox reactions in homogeneous solution that involve net chemical change. The relationship between the rate constants k,l for an electrode reaction and k,, for the corresponding self-exchange electron transfer reaction is not obvious because k,l can be strongly influenced by the nature and history of the electrode surface and by solvent dynamic effects if present. Electrode properties, however, are not expected to be sensitive to pressures in the 0-200 MPa range. Moreover, the signature of solvent dynamical effects is a dependence of reaction rate on solvent viscosity, but the viscosity of water is effectively independent of such pressures at near-ambient temperatures. Consequently, for typical aqueous electrode reactions, AV: = O S A V . , regardless of any involvement of solvent dynamics, and so AVZ can be predicted from transition state theory (TST) according to Eqs (5.5)(5.8). On the other hand, for non-aqueous solvents, viscosities rise sharply and approximately exponentially with pressure, so that any dependence of k,l (or k,) on solvent dynamics will be reflected in a markedly positive contribution to AV: (or AVZ). For couples such as Ru(hfac)!/- in which one of the reactants is electrically neutral, ion pairing and related effects in polar organic solvents are negligible, so that AVZ in organic solvents should be given to a good approximation by the sum of a moderately negative TST contribution and a strongly positive solventdynamical contribution represented closely by the activation volume AV& for reactant diffusion. For the self-exchange reaction of Ru(hfac);/ - in organic solvents, AVZ is negative and conforms closely to predictions based on TST, so that the TST contribution to AVZ can be identified with O.SOAV2. This accounts very well for the positive AVZ values found for the Ru(hfac):’- system in organic solvents. Thus, the fifty-percent rule evidently also applies to AVZ for electrode reactions in non-aqueous media but its effect is swamped by solvent dynamical contributions,

References

so that AVZ is invariably positive. This last remark is also valid for electrode reactions of 2+/+ and 3+/2+ couples, but the possibility of ion pairing prohibits detailed interpretation other than to note that AV: runs roughly parallel to AVZE

~761. High-pressure kinetics indicate the absence of solvent dynamical effects in outer-sphere bimolecular self-exchange reactions in non-aqueous media, even in the most favorable experimentally accessible case (the Fe(phen):+” couple) - on the contrary, expectations based on TST alone are generally met, and AVZ is negative. At the same time, AVZ is invariably positive for electrode reactions in organic solvents, signaling rate control by solvent friction. Solvent dynamics dominate electrode kinetics in non-aqueous media even when the corresponding self-exchange reactions clearly conform to the TST model. In short, pressure effects reveal that electrode reactions are subject to solvent dynamical effects in non-aqueous media at least, but the corresponding self-exchange reactions are not, regardless of the solvent.

’

Acknowledgments

I thank my co-workers, whose names appear in the references, for their many superb contributions, the Natural Sciences and Engineering Research Council of Canada for financial support of our work, Dr N. Weinberg for illuminating discussions, and the Alexander von Humboldt Foundation for a Research Award.

References 1 R. A. MARCUS,]. Chem. Phys. 1956, 24, 2

3 4

5

6

7 8 9

10

966, 979; 1957, 26, 867. R. A. MARCUS,Discuss. Faraday Sac. 1960, 29, 21. R. A. MARCUS,Faraday Discuss. Chem. SOC.1982, 74, 7. N. SUTIN,Prog. Inorg. Chem. 1983, 30, 441. R. A. MARCUS,N. SUTIN,Biochim. Biophys. Acta 1985, 812, 265. N. S. H U S H , Trans. Faraday SOC.1961, 57, 557. D. R. STRANKS,Pure Appl. Chem. 1974, 38, 303. T. W. SWADDLE, Inorg. Chem. 1990, 29, 5017. T. W. SWADDLE,C a n . j . Chem. 1996, 74, 631. T. W. SWADDLE,].Mol. Liq. 1995, G5/ 66. 237.

1 1 H. DOINE,T. W. SWADDLE, Can.].

Chem. 1988, 66, 2763. D. R. STRANKS,T. W. SWADDLE, Inorg. Chem. 1990, 29,

12 W. H. JOLLEY,

1948.

H. DOINE,T. W. SWADDLE,Inorg. Chem. 1991, 30, 1858. 14 R. D. SHALDERS, T. W. SWADDLE, Inorg. Chem. 1995, 34,4815. 1 5 P. D. METELSKI, Y. Fu, K. KHAN, T. W. SWADDLE; Inorg. Chem. 1999, 13

38, 3103.

16 Y. Fu, T. W. SWADDLE,]. Am. Chem. Soc. 1997, 119, 7137. 17 W. H. JOLLEY, D. R. STRANKS, T. W. SWADDLE, Inorg. Chem. 1990, 29, 385. 18 M. R. GRACE,T. W. SWADDLE, Inorg. Chem. 1993, 32, 5597. 19 L. SPICCIA, T. W. SWADDLE, Inorg. Chem. 1987, 26, 2265.

I

181

182

I

5 lnsights into Solution Chemistryfrom High Pressure Electrochemistry ZAHL,K. VAN ELDIK, T. W. Inorg. Chem. 2002, 41, in SWADDLE, press. 21 P. D. METELSKI. T. W. SWADDLE? Inorg. Chem. 1999, 38, 301. 22 S. WHERIAND, Coord. Chem. Rei?. 1993, 123, 169. 23 Y. MARCUS, [on Properties, Marcel Delker, Inc., New York, 1997, Chapter 8. 24 Y.Fu, A. S. COLE,T. W. SWADDLE,]. Am. Chem. Soc. 1999. 121, 10410. 25 W. H. J O L I F Y , D. R. STRANKS, T. W. SWADDLE, Inorg. Chem. 1992, 31, 507. 26 J . K. BEATTIE, H. ELSBERND,Inorg. Chim. Acta 1995, 240. 641. 27 J. K. BFATrTE, Adc. Inorg. Chem. 1988, 32, 2. 28 K. A. BINSTEAD, J. K. BEATTIE, Inorg. Chem. 1986, 25, 1481. 29 C. FEUORCHUK, M. MATSUMOTO, T. W. SWADDLE, unpublished studies. 30 L. L. MARTIN, K. S. HAGEN,A. HAUSER, R. L. MARTIN,A. M. SARGESON,]. Chem. SOC.Chem. Comm. 1988, 1313. 31 L. L. MARTIN, R. L. MARTIN,K. S. MURRAY, A. M. SARGESON, Inorg. Chem. 1990, 29, 1387. 32 L. L. MARTIN, R. L. MARTIN, A. M. SARGESON, Polyhedron 1994, 13, 1969. 33 A. M. SARGESON, personal communication. 34 H. DOINE,T. W. SWADDLE, Inorg. Chem. 1988, 27, 665. 35 H. D. TAKAGI, T. W. SWADDLE, Chem. Phys. Lett. 1996, 248, 207. 36 M. R. GRACE, H. TAKAGI, T. W. SWADDLE, Inorg. Chem. 1994, 33, 1915. 37 T. W. SWADDLE, in High Pressure Techniques in Chemistry and Physics, W. B. HOLZAPFEL, N. S. ISAACS,(Eds), Oxford University Press, Oxford, 1997, p. 343. 38 T. W. SWADDLE, P. A. TREGLOAN, Coord. Chem. Rev. 1999, 187, 255. 39 H. DOINE,T. W. WHITCOMBE; T. W. Can.]. Chem. 1992, 70, 81. SWADDLE: 40 Y. Fu, T. W. SWADDLE, Chem. Comrnun. 1996, 1171. 41 Y. Fu, T. W. SWADDLE, Inorg. Chem. 1999, 38, 876. 42 M. BIIAI,2. Natuforsch. 1991, 46a, 1108. 20 A.

M. MA~SUMOTO, T. TARUMI, I. TAKAHASHI, S. FUNAHASHI, T. NODA, 2.Natuforsch. 1997. H. D. TAKAGI, 52b, 1087. 44 M. T. CR~JAAES. H. G. DRICKAMER, L. R. FAULKNER,]. Phys. Chem. 1996, 100, 16613. 45 J. GOLAS, H. G. DRICKAMER, L. R. FAULKNER, ]. Phys. Chem. 1991. 95, 10191. 46 M. T. CRUAAES, H. G. DRICKAMER, L. R. FAUI-KNER,]. Phys. Chem. 1992, 96, 9888. 47 M. T. CRUAAES, K. K. RODGERS:S. G . SLIGAR,].Am. Chem. SOC.1992, 114, 9660. 48 M. T. C R U A ~ ~H. E SG, . DRICKAMER, L. R. FAULKNER, Langmuir. 1995, I 1, 4089. 49 I. 1. SACHINIDIS, R. D. SHALDERS, P. A. TREGLOAN, 1. Electroanal. Chem. 1992, 327, 219. 50 J. I. SACIIINIDIS, R. D. SHALDERS, P. A. TREGLOAN, Inorg. Chem. 1994, 33, 6180. 51 J. I. SACIIINIDIS, R. D. SHALDERS, P. A. TREGLOAN, Inorg. Chem. 1996, 35, 2497. 52 7. S U N ,1. F. WISHART, R. VAN ELDIK, R. D. SIIAIDERS, T. W. SWADDLE,]. Am. Chem. SOC. 1995, 117, 2600. 53 R. S. NICHOLSON, I. SHAIN,Analyt. Chem. 1964, 36, 706. 54 R. S. NICHOLSON, Analyt. Chem. 1965, 37, 1351. 55 J. ZHOU,T. W. SWADDLE, Can.]. Chem. 2001, 79, 841. 56 R. A. MARCUS, Can.]. Chem. 1959, 37, 155. 57 R. A. MARCUS,].Phys. Chem. 1963, 67, 853. 58 R. A. MARCUS,].Chem. Phys. 1965, 43, 679. 59 R. A. MARCUS, Electrochim. Acta 1968, 13, 995. 60 N. S. HUSH,Electrochim. Acta 1968. 13, 1005. 61 R. D. CANNON, Electron Transfer Reactions, Buttenvorths, London, 1980, pp. 220-222. 62 D. MILNER, M. J. WEAVER,]. Electroanal. Chem. 1985, 191, 411. 63 G. HILLS,]. Phys. Chem. 1969, 73, 3591. 43

64 65 66

67 68 69

X. Z I I A N GJ., LEDDY,A. J. BARD,].

Am. Chem. Sac. 1985, 107, 3719. X. ZIIANG,H. YANG,A. J. BARD,J . Am. Chem. Sac. 1987, 109, 1918. D. E. KHOSHTARIYA. T. D. DOLIDZE.D. KRULIC,N. FATOUROS, D. DEVILLIERS, J . Phys. Chem. B 1998, 102, 7800. J. S. JAWORSKI, 2. KEBEDE, J. Electroanal. Chem. 1994, 370. 259. T. G E N N E D. ~ , F. MILNER,M. J. WEAVER, J . Phys. Chem. 1985, 89, 2787. G. E. MCMANIS,M. N. GOLOVIN, M. J. WEAVER, J . Phys. Chem. 1986, 90, 6563.

70

R. Pumr. R. W. MURRAY, J. Am. Chem.

71

M. E. WILLIAMS, J. C. CROOKER, R. PYATI,L. J. LYONS,R. W. MURRAY, J. Am. Chem. Sac. 1997, 119, 10249. J. V. SENGERS, B. KAMGAR-PARSI, J. Phys. Chem. Rt$ Data 1984, 13, 185. 0. SALMAN, P h D Thesis, University of Illinois, 1982. P. W. BRIDGMAN, The Physics ofHigh Pressure, G . Bell, London, 1931, p. 342. J. JONAS,J. A. AKAI,]. Chem. Phys.

Sac. 1996, 118, 1743.

72 73 74 75

83 84

M. MATSUMOTO, D. LAMPRECHT, M. R. NORTH,T. W. SWADDLE, Can. J . Chern.

77

T. ASANO,Pure Appl. Chem. 1999, 71,

6563.

HARRER, W. JAENICKE, J . Chem. Sac. Faraday Trans. 11987, 83, 161.

H. A. KRAMERS, Physica 1940, 7, 284. 87 R. F. GROTE,J. T. HYNES,J . Chem. Phys. 1980, 73, 2715. 88 N. AGMON,J. J. HOPFIELD, ]. Chem. Phis. 1983, 78, 6947. 89 N. AGMON,J. J. HOPFIELD, J . Chem. Phys. 1983, 79, 2042. 90 H. S U M I ,R. A. MARCUS, J . Chem. Phys. 1986, 84, 4894. 91 R. A. MARCUS,H. SUMI,J . Electroanal. Chem. 1986, 204, 59. 92 W. NADLER, R. A. MARCUS, J . Chem. Phys. 1987, 86, 3906. 93 M. V. BASILEVSKY, V. M. RYABOY, N. N. 1.Phys. Chem. 1990, 94, WEINBERG, 86

8734. 94

78

M. J. WEAVER,].Mol. Liq. 1994, 60, 57. 80 H. HEITELE,Angew. Chem. Intl. Ed. Engl. 1993, 32, 359. 81 J. T. HYNFS,]. Phys. Chem. 1986, 90,

97

79

W. R. FAWCETT,M. OPALLO,Angew. Chem. Int. Ed. Engl. 1994, 33, 2131.

M. MCGUIRE,G. MCLENDON, /. Phys. Chem. 1986, 90,2549. K. KIRCHNER,S.-Q. DANG,M. H. W. DODGEN,S. STEBLER, WHERLAND, J. P. HUNT, Inorg. Chem. 1989, 28, 3604.

98

3701. 82

M. J. WEAVER,D. K. PHELPS,R. M. NEILSON,M. N. GOLOVIN,G . E. MCMANIS,J . Phys. Chem. 1990, 94, 2949.

96

M. J. WEAVER,Chem. Rev. 1992, 92, 463.

M. V. BASILEVSKY, V. M. RYABOY,N. N. WEINBERG, J . Phys. Chem. 1991, 95, 5533.

95

2001, 79, 1864. 1691.

I.

85 G . GRAMPP,W.

1977, 66,4946. 76

W. R. FAWCETT,C. A. Foss, JR.,]. Electroanal. Chem. 1989, 270, 103. G. E. MCMANIS,M. N. GOLOVIN, M. WEAVER, J . Phys. Chem. 1986, 90,

99

D. LAMPRECHT,T. W. SWADDLE, unpublished work. R. M. NIELSON,S. WHERLAND, Inorg. Chem. 1984, 23, 1338.

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

6

Pressure Effects on the Photoreactions of Transition Metal Complexes Peter C. Ford” and Leroy

E. Laverrnan

6.1

Introduction

In this chapter, the focus will be on the application of high pressure techniques in the study of the photochemical behavior of transition metal complexes (coordination, organometallic and bio-inorganic) in solution. We will present a systematic treatment of pressure effects on the nature of excited states (ES) and on the photophysical and photochemical processes that lead to ligand substitution, electron or energy transfer and thermal reactions of reactive intermediates generated by ES reactions. Selected examples will be presented in detail to illustrate how pressure effects can provide valuable mechanistic insight when combined with other quantitative studies. Figure 6.1 is a simple Jablonski diagram that introduces the sequence of events subsequent to a spin allowed electronic excitation of a molecule. This will define the processes that must be considered when elucidating photoreaction mechanisms. Initial excitation (hv) from the electronic ground state (GS) is vertical (FranckCondon Principle), so the first electronic excited state (ES1) is initially prepared with the geometry of the ground state. Thus, ESI is prepared in a vibronic excited state, and the first deactivation process is vibronic relaxation. In Fig. 6.1, ESI is shown to decay by several competing processes including chemical reaction to form a product ( P I ) with rate constant fluorescence (q)or non-radiative decay directly to the ground state, or by non-radiative internal conversion (kit) or intersystem crossing (kist) to a lower energy state ES2, (if the ES has the same or different spin multiplicity, respectively). The quantum yield of a particular process arising from ESI (Qi) would thus be defined by the ratio of decay rate constants,

(x)

q,

If only unimolecular processes are involved then ES1 decays according to an exponential function and the lifetime of ESI is defined as

6. 7 Introduction

ES2

krl'

Fig. 6.1.

Generalized Jablonski diagram.

Therefore Eq. (6.1) can be rewritten as

If ES2 has a different spin multiplicity from that of ES1, the efficiency of its formation is the intersystem crossing quantum yield

ES2 has its own decay processes including radiative (kr, termed phosphorescence) and non-radiative deactivation (k"), unimolecular reaction to product(s) (kp) and bimolecular quenching by energy or electron transfer (b) to another species. For this model the ES2 lifetime is defined by

However, the quantum yield is defined slightly differently since the efficiency of intersystem crossing has to be taken into account.

I

185

186

I

6 Pressure Effects on the Photoreactions of Transition Metal Complexes

The purpose of outlining the decay pathways for this simple model (Fig. 6.1) is to emphasize that hydrostatic pressure can affect rates of the individual processes. Pressure can also affect energies of excited states relative to the ground state as well as relative to each other. Understanding pressure effects on photoreactions requires careful evaluation of all parameters that may be affected. Lastly it should be emphasized that the primary photoproducts formed directly from the excited states might themselves be very reactive species. These may react subsequently to give stable products by unimolecular or more complex mechanisms. Continuous photolysis experiments which use low intensity excitation generate only small steady state concentrations of excited states and reactive intermediates, so only stable products or a steady state phenomenon such as luminescence are generally observed. Flash photolysis techniques, which prepare transient species in non-steady state concentrations, often allow direct observation of such species. 6.1.1

Pressure Effects on Excited State Energies

For metal centered ligand field (LF) transitions, the magnitude of the pressure induced spectral shifts will depend markedly on the symmetries of the orbital transitions. For a hexa-coordinate transition metal complex having an approximate Oh symmetry, transitions between the t2g(d,) and the eg (do*)orbitals are energetically related to the magnitude of LF strength as represented by the octahedral splitting parameter Ao. At high pressure, A. increases owing to stronger metalligand overlap from bond compression. For example, this is reflected in the blue shifts of about +25 cm '/I00 MPa for the ligand field bands of a variety of nickel(I I) complexes [ 13. A qualitative energy level diagram for the ligand field states of an octahedral chromium(II1) ion is shown in Fig. 6.2. The dominant bands in the absorption spectrum are ground state quartet to excited state quartet transitions, since the quartet to doublet transitions are spin forbidden. Under high pressure, solid [Cr(NH3)6]C13[2] demonstrated blue-shifts for the Q1 + Qo (+53 cm-'/100 MPa) and Q2 + QObands ( t 3 6 cm-'/100 MPa). In contrast, both absorption and emission transitions between the QOground state and the lowest energy doublet state Do display very small red shifts with pressure. For example, the Do + QOemission from the Cr(II1) center of ruby (694.2 nm at ambient pressure) shifts to the red by -0.7 cm-l/lOO MPa. The sharpness of this emission line and the linear nature of its response up to 20,000 MPa has proved extremely useful as an internal calibration of very high pressures in diamond anvil cells. Modest red shifts ( - 3 to -5 cm-'/100 MPa) were seen for the doublet emission (Do --+ Qo) and absorption (Do + Qo) bands for the Cr(NH3);' ion in aqueous solution [3]. The different pressure sensitivities of the doublet and quartet bands can be rationalized in terms of orbital parentages. The quartet excited states have (t2g)2(eg)1 configurations and must involve distortions of the M-L bond lengths owing to the

6.1 Introduction

I

187

"\ QO

Fig. 6.2. Generic energy level diagram for a strongfield Cr(lll) complex. PI is phosphorescence from the double Do and F1 the fluorescence from the quartet Q1.

nature of eg orbitals. In contrast, both the ground and doublet states have ( t ~ ) configurations, ' so there is little distortion between the QOand Do states. Although emission from Cr(II1) in an octahedral field usually displays sharp lines from the spin-forbidden Do + QO transition, broad band Q1 + QO fluorescence may be seen when the field is sufficiently weak [4]. Different sensitivities of the quartet and doublet ligand field state energies to pressure are illustrated by a diamond anvil cell investigation of the 22-K emission spectrum of solid (NH4)3[CrF~] (51. At ambient pressure, broad fluorescence at ,,v = 12,984 cm-l (Eoo = 14,319 cm-') is the only feature seen. Raising P to 7.1 GPa shifts vmax to 13,605 cm-' (Eoo = 14,855 cm-'). Raising P further leads to disappearance of the fluorescence and appearance of the sharp doublet emission at 14,896 cm-'. Thus the increased pressure tunes the QOstate above the Do state. Ligand localized nn* ES are common to complexes having large aromatic ligands. Absorption and emission bands attributed to nn* states shift to the red with increasing P [6]. For example, fluorescence maxima from the ligand centered states of the tetraphenyl porphyrin complexes Mg(TPP) and Zn(TPP) display shifts ranging from -34 cm-'/100 MPa in methanol to -94 cm-'/100 MPa in chloroform with

188

I

6 Pressure Effects on the Photoreactions of Transition Metal Complexes

increasing pressure [7]. Phosphorescence bands are less pressure sensitive than fluorescence bands owing to the lower polarizability of triplet states. Metal to ligand charge transfer (MLCT) absorption and emission bands have been examined as a function of pressure for the ruthenium(I1) complex Ru(bpy)i+ and related Ru(I1) and Os(I1) species [&lo]. In acetonitrile, small red shifts were seen for MLCT absorption and emission maxima with increasing P (- -20 cm-’/ 100 MPa over 300 MPa), and it was argued that the small shifts result from pressure induced decreases in the ligand R ‘ orbitals compensated by lower d, orbital energies from increased M-L overlap [9]. Similar shifts were seen in diamond anvil pressure studies (3-7 GPa) of the emission from solid [Ru(bpy)3I2+salts [lo]. Pressure-induced solvent freezing has been shown to have remarkable effects on the spectra of the rhenium(I) complex ReBr(C0)3(phen)and the copper(I) cluster Cu414py4.The former displays strong MLCT emission 1111; the latter displays a strong emission from a “cluster centered” (CC) excited state assigned as having mixed d-s metal-centered and iodide-to-copper charge transfer character [ 121 in ambient temperature fluid benzene. For each, the emission spectrum is markedly dependent on the rigidity of the medium, a behavior termed the “rigidochromic effect” [ll].Since benzene freezes readily at relatively low pressure (Pf = 72 MPa at 25 “C), the rigidity of benzene solutions can be sharply changed by raising P above this value. The broad emission maximum of Cu414py4 :;1( = 690 nm) was unaffected by raising pressure as long as the benzene solution remained fluid, but above 72 MPa there was a dramatic shift ;in: 1 to 575 nm (Fig. 6.3) [13]. In the same context, pressure-induced freezing of a benzene solution of ReBr(C0)3phen shifted ; : 1 from 610 nm at 0.1 MPa to 546 nm at P > 72 MPa [13]. One explanation of rigidochromic effects is the reduced ability of solvent to reorient in the frozen state to facilitate the most favorable electrostatic interaction with an ES. Alternatively, ab initio and experimental results suggest that the CC” state of Cu414py4 is quite distorted from the ground state [12b]. Thus, the hypsochromic shift of solvent freezing by raising P may result from restrictions imposed by a rigid solvent cavity. A recent review [ 141 summarizes some other high pressure (diamond anvil) studies of vibrational and electronic spectra of metal complexes. 6.1.2 Pressure Effects on Excited State Kinetics

The two most common parameters measured in photochemistry are the quantum yield @i for a specific process, and the lifetime t of the excited state. The quantum yield is operationally defined as the moles of product formed (or starting species reacted) per einstein of light absorbed by the system at a particular wavelength of irradiation (&). In this context, the pressure effect on the quantum yield gives an “apparent activation volume”, i.e. AV; = -RT[d(ln @i)/dP], from a plot of In @i vs. P. To elucidate a mechanism, one must first consider the states initially formed by photoexcitation as well as other ES eventually populated by internal conversion and intersystem crossing. Although not always the case, many metal complexes, when

G. 1 Introduction

I

500

600 700 800 Wavelength (nm)

900

Fig. 6.3. Emission spectra of Cu4I4py4in benzene a t 298 K a t ambient pressure and 75 MPa. Both spectra have been normalized to an arbitrary maximum intensity of 1.O (redrawn from ref. 13).

-

excited, undergo efficient relaxation (QiSc 1.0)to a lowest energy excited state (LEES) from which the various chemical processes lead to photoproducts. In such systems, one can apply transition state theory and consider pressure effects in terms of the excited state mechanisms. For the model shown in Fig. 6.1, the quantum yield for a specific pathway would be

and the apparent activation volume would be AV! = AVL

+ AV:

(6.8)

where AV! = RT[d(ln r)/dP)T and AVL is defined above. However, it is generally preferable to use Eq. 6.7 to determine ki for each set of conditions and to determine AV! from a plot of In ki vs. P. Radiative rates: For most transition metal complexes in solution, emission quantum yields are small, thus radiative decay is only a minor component of the overall deactivation mechanism. Limited studies show pressure effects on k, to be small, a few percent over the hydrostatic pressure ranges of principal interest here and these effects can largely be attributed to solvent perturbations [15, 161. The relatively small values of k, for most luminactive metal complexes in fluid solutions, suggest that such modest changes will not have much impact on interpretations of pressure effects on lifetimes or quantum yields.

I

189

190

I

6 Pressure Efects on the Photoreactions of Transition M e t a l Complexes

Non-radiative deactivation rates: In contrast, nonradiative deactivation (kI,) often constitutes the major pathway by which metal complex excited states decay. For large molecules non-radiative deactivation is largely due to intramolecular processes determined by vibrational and electronic factors. Vibronic coupling between electronic states has been analyzed in terms of two limiting cases. In the weak coupling limit, the relative displacement of the ES potential surfaces is small. and k, is predicted to increase exponentially as the difference in energy (AE) between the ES and the GS decreases (“the energy gap law”). In the strong coupling limit, there is a large displacement of the ES surface (at least one normal mode) relative to the ground state such that the surfaces cross not far from the minimum of the higher state [ 171. Accordingly, strong coupling deactivation should show Arrhenius type temperature dependence while weak coupling should be essentially temperature independent. Strong coupling may also be associated with ES reactions, especially unimolecular reactions such as ligand substitution. In such a case there should be correlation between pressure effects seen for the reaction pathway k, and the strong coupling component of k,. Pressure can affect k, in other ways, for example, it may raise or lower an ES energy by compression of the complex or by perturbation of the solvent dielectric constant. According to the energy gap law, k, should respond to -AE exponentially [18]. Such behavior was observed by Salmon and Drickamer 1151 who investigated the MLCT phosphorescence spectra, lifetimes and quantum yields for the rhenium( I) complexes ReC1(C0)3phen and ReCl(C0)3(4,7-Ph2phen) in different solvents. By systematic variation of solvent properties with applied P, they demonstrated a linear relationship between ln(k,) and -AE indicative of a weak-coupling mechanism for non-radiative deactivation. The effects were small in dimethylformamide and acetonitrile but were larger in m-xylene, reflecting the sensitivity of the MLCT states to the greater compressibility of the latter solvent. Another example can be drawn from the emission properties of two dinuclear d8 complexes the diplatinum(I1) ion Pt,(POP);- (POP = p-q2-H2P20;-) [19] and the diiridium( I) ion Ir2pz2(COD)2 (pz = p-q2-pyrazolate, COD = q4-1,4-cyclooctadiene) [20]) (Fig. 6.4). In both complexes the metal centers have square planar coordination with the bridging ligands holding the two square planes in a cofacial configuration. The LEES has been assigned as having d(o,&) + p(obM)orbital parentage in each case leading to a stronger M-M bond in the excited state than in the ground state [21]. Both species are strong emitters in ambient temperature fluid solutions [ 221. Hydrostatic pressure up to 300 MPa had no effect on the absorption and emission spectra ;:1( = 511 nm) of Pt2(POP):- in ambient temperature aqueous solution. There was a modest decrease in the phosphorescence lifetime from z = 8.8 ps at 0.1 MPa to 7.6 ps at 300 MPa and a corresponding 13 % decrease in the phosphorescence quantum yield (@ = 0.55, @)r3O0 = 0.48). Since the intersystem crossing to the LEES was estimated to be unity in both cases, these data demonstrate that pressure has little effect on k, (Eq. 6.9) [22], consistent with the relative insensitivity of the refractive index of water to pressure [23].

G. 7

HO P

Introduction

\*/

Fig. 6.4. Drawing of Pt2(p-q2-H2P205)j- and I r 2 p ~ 2 ( C O D ) 2 (pz = p-q2-pyrazolate, COD = q4-l ,4-cyclooctadiene) (from ref.

20).

In the absence of photoreaction, the shortened lifetime at 300 MPa must be largely due to increases in the non-radiative deactivation rate constant. Since @isc 1, then:

-

(6.10)

qoo/e

This gives = 1.34 f 0.03, from which an apparent activation volume for the non-radiative deactivation path can be calculated to be -2.4 f 0.5 cm3 mol-'. In acetonitrile, a AV: value of -0.2 cm3 mol-I was determined [22]. Similar analysis of the emission properties of I r z p ~ ~ ( C 0 D demonstrated )~ again that k, was little affected by pressure to 300 MPa. The absorption maxima shifted (from 49G to 500 nm), but there was no shift in the emission band ;1 ( = 663 nm). However, unlike Pt2(POP)t-, t increased with P from 0.24 ps at 0.1 MPa to 0.42 ps at 300 MPa, thus the AVA for I r z p ~ ~ ( C 0 D has ) ~the positive value 1-4.7 cm3 mo1-l in acetonitrile [22]. That the ES of these d8-d8dinuclear complexes have shorter M-M bonds than the respective ground states suggests that k, should display a positive AV: along the deactivation trajectories. The positive AV: value for Ir2pz2(COD)2is consistent with this expectation, but the small negative AV: value for Pt2(POP)t- suggests modest contraction rather than expansion accompanying non-radiative deactivation. The contrasting photophysical behaviors may be attributed to structural differences. The volume changes between the flexible butterfly structure of Ir2pz2(COD)2 and its ES may be larger than for the tetrabridged Pt2(POP): ion. For the latter, shortening of the Pt-Pt bond may be compensated by ES lengthening of others, perhaps Pt-P bonds, thereby minimizing overall volume changes. However, an alternative (but not mutually exclusive) interpretation would be that the small AV: value for Pt2(POP):- is indicative of a weak coupling deactivation mechanism. This is indeed consistent with the temperature independence of k, for Ptz(POP)t~

I

191

192

I

G Pressure Effects on the Photoreactions of Transition Metal Complexes

under these conditions [22]. In contrast, k d for I r z p ~ ~ ( C 0 D shows ) ~ markedly temperature-dependent behavior consistent with a strong coupling mechanism due to a molecular distortion leading to deactivation. Non-radiative deactivation involving a second excited state: A somewhat different situation is presented by the ruthenium( 11) complex Ru(bpy):'. In ambient temperature, fluid solution this species shows little unimolecular photochemistry and modest emission quantum yields (mr < 0.3) [9]. Initial pressure studies on the MLCT luminescence from Ru(bpy)if in 18 "C aqueous solution saw little change [24]. However, detailed studies by Fetterolf and Offen [9, 251 demonstrated a remarkable temperature dependence for AV:, which rises from +2.9 cm3 m o t at 15 "C to f8.7 cm3 mol-' at 45 "C in acetonitrile and from -1.5 cm3 mol at 2 "C to +7.5 cm3 mol-' at 70 "C in aqueous solution [26]. This behavior is best explained in terms of the proposed mechanism [27] where a key pathway for [R~(bpy);'];~ deactivation involves thermal promotion from the MLCT LEES to a ligand field state -3600 cm-' higher (Fig. 6.5). The large AV! associated with the thermally promoted pathway may be the consequence of two contributions, a larger partial molar volume for the 'LF state than for the 3MLCTstate, and second, a AVt for the 3LF owing to a strong-coupling mechanism.

-'

Fig. 6.5. Model describing proposed mechanism for decay of the MLCTexcited state of Ru(bpy):+ and related complexes. Emission occurs from the lowest energy excited state 'MLCT. At low temperature, nonradiative

deactivation occurs predominantly from this but a t higher temperature, the state predominant pathway is via thermal promotion to the higher energy 3LF state followed by rapid deacviation (k:).

(ky)

G.1

Introduction

I

193

The Ru(phen):+ analog shows similar behavior, while the osmium(11) analog Os(phen):' does not 191. The much higher ligand field splitting for Os(I1) complexes positions the lowest 'LF state at an energy too high relative to the 'MLCT state to be a significant contributor to the non-radiative deactivation pathways. Ambient temperature emission intensities and lifetimes from Ru(bpy);+ and Ru(bpy)z(py);+ salts in solutions of poly(4-vinylpyridine) (PVP) and poly (acrylic acid) (PAA) have been studied at pressures to 7 GPa in a diamond anvil cell [lo]. For Ru(bpy):', emission intensities and lifetimes decreased monotonically by -65 % and -50 % in PAA and PVP, respectively, as P was raised from 0.1 MPa to 7 GPa. This modest effect (AV; (apparent) -1 cm3 mol-') was attributed to increases in k, according to energy gap law considerations as the 3MLCTemission shifts to lower energy. Volume changes related to photophysical properties can also be deduced from time-resolved photoacoustical calorimetry (PAC) studies [ 281. This technique has been used with the Ru(11) complexes Ru(bpy):+, cis-Ru(bpy)z(CN)z and Ru(bpy)(CN);- [29, 301. Analysis of such data was used to calculate enthalpy and volume changes induced by the formation and decay of MLCT excited states in aqueous solutions. For Ru(bpy);+ formation of the 3MLCT state is associated with a modest contraction (AVI\?LCT= -3.6 & 0.3 cm3 mol I), which was rationalized in terms of slight decreases (-0.001 A) in Ru-N bond lengths in the ES [29]. The AV for decay is the same magnitude, but of opposite sign ($3.4 f 0.4 cm3 mol-') [30]. In sharp contrast, positive values of AVMLCT (+14.9 and +5.2 cm3 mol-1 respectively) were determined for MLCT excitation of cis-Ru(bpy)z(CN)2 and Ru(bpy)(CN):- [ 301. These were rationalized in terms of reduced basicity of coordinated cyano groups in the MLCT state resulting in net desolvation. As a consequence, the AVMLCT of Ru(bpy)(CN):- was shown to be markedly pH dependent [ 3OCl.

-

Excited state tuning with pressure: The influence of pressure on metal complex excited state energies, either by changes in the intrinsic properties of the complex, or by modification of the solvation properties, provides an opportunity to tune the excited state energies. Such tuning generally has only modest effects on unimolecular photophysical kinetics over the hydrostatic pressure ranges used in most kinetics studies (0.1-400 MPa), ifthe LEES is a single state or a collection of thermally-equilibrated states with similar orbital parentages. However, much greater effects would be expected if several excited states of dissimilar orbital parentages are involved in determining the radiative and/or non-radiative pathways as described above for the Ru(bpy)p ion. ES tuning effects on photophysical properties are quite evident in the luminescence spectrum of the iridium(II1) complex ion Ir(Mephen)2Cl; (Mephen = 56dimethyl-1,lO-phenanthroline) in ambient-temperature dimethylformamide. This displays dual emission from thermally equilibrated 3MLCT and 3LF states. Increased pressure (300 MPa) leads to enhanced MLCT emission (550 nm) at the expense of the LF emission (720 nm) with little or no shift of peak maxima (Fig. 6.6) [31]. The spectral changes were attributed to shifts in the relative populations

194

I

G

Pressure Effects on the Photoreactions of Transition Metal Complexes

NANOMETERS

500 I

700 900 110 I

I

I

I

I

I

300 MPa

z

0 fn

W

a

c M-'(

I 0- 3,

Fig. 6.6. Emission spectra of Ir(Mephen)zCIz in DMF a t 0.1 MPa and 300 MPa showing the increase in the MLCTemission intensity a t the expense of the longer wavelength LF emission a t the higher P (redrawn from reference 31).

owing to partial molar volume differences between the two ES in thermal equilibrium (Fig. 6.7), since emission lifetimes and spectra are independent of the excitation wavelength. The apparent volume difference of +4.2 cm3 mol-' was calculated according to Eqs. (11) and (12) by assuming that the ratio of the radiative rate constants is pressure independent. This indicates that the LF state is larger by 4.2 cm3 mol-I than the ground state. (6.11)

AV!pp = -RT

d ln(ky/k?) dP

+ AVeq z AVeq

(6.12)

Lifetime measurements for Ir(Mephen)*Cl: in DMF solution show that pressure decreases the deactivation rates, and a linear ln(kd) vs. P plot gives the AV: value +4.0 0.2 cm3 mol-', where kd = T-'. This would suggest, according to Fig. 6.7, that deactivation is largely occurring via thermal promotion to the LF excited state followed by non-radiative decay via the pathway. High spin/low spin equilibria: Pressure effects on the distribution of electronic states have also been investigated for d6 iron(I1) complexes of the type FeL:*, where L is a polydentate nitrogen donor ligand. For certain L such as p y m (Fig. 6.8), the

6 7 Introduction

I

195

Fig. 6.7. Scheme for luminescence from two emitting states o f different orbital parentages in equilibrium.

energy difference between the high spin (quintet)/low spin (singlet) electronic configurations is small. As solids, these species may undergo a phase transition from the quintet to the more strongly colored singlet as T is lowered. In certain cases, irradiation at low T converts the singlet to the quintet which persists indefinitely under these conditions in a process termed the LIESSTeffect [32]. For some FeL? the two electronic states are in thermal equilibrium in solution (Eqs 6.13 and 6.14), and the relative concentrations of the two states can be tuned by hydrostatic pressure. Values of AV as large as +16 cm3 rno1-l have been reported, depending upon the ligand and the solvent [ 331. kis

[FeL;'] g [FeL;'] (LS) krl (HS)

(6.13) (6.14)

Fig. 6.8.

Drawing of 2(2-pyridy1)imidazole (pyim)

196

I

G Pressure Efects on the Photoreactions of Transition Metal Complexes

The pressure effects on spin relaxation dynamics for these iron(11) complexes have been examined using laser flash photolysis techniques. For Fe(pyim):+ the two spin states are in equilibrium with a K = 0.56 in 298 K acetone with a partial molar volume difference AV = +8.1 cm3 mol-' [34]. Photoexcitation (Aex = 532 nm) leads to transient bleaching of the low spin isomer's MLCT bands followed by first order relaxation to the original spectrum with a 45-ns lifetime. Transient bleaching and subsequent return of the MLCT absorption was attributed to formation of the HS isomer and subsequent spin relaxation. The pressure dependence of the relaxation lifetimes was used to determine the activation volumes of the spin relaxation rates for a variety of FeL:+ in different solvents. It was found that AV!l fell into a remarkably narrow range of values (-5.5 1 cm3 mol-l) and it was concluded that the spin crossover for these species follows a common mechanism via a transition state located midway between the high and low spin states [33]. An analogy to Fe(I1) spin isomers can be found in considering pressure effects on the relaxation rates for d8 nickel(I1) complex Ni(dppe)Clz. The latter has a tightly solvated square planar configuration in the ground state but is presumed to have a tetrahedral configuration in its triplet excited state. Flash photolysis studies on the ES relaxation rate gave large negative AV: (-10 cm3 mol-') and AS: (-70 J mol-' K-') values. These data have been interpreted as indicating a transition state for the k, pathway distorted substantially toward the ground state configuration [ 351.

6.2

Unimolecular Excited State Reactions 6.2.1

Ligand Photosubstitution Reactions

Ligand substitution (Eq. 6.15) is commonly observed as the result of photoexcitation. Two excited state reactivity models can be considered. In one the ES responsible for the photochemistry is a bound, thermally relaxed state, in which case? transition state theory treatment is relevant and the individual ki and AV: values for ES decay can be elucidated. The second is ligand labilization from an ES that is unbound along the metal-ligand coordinate, i.e. a dissociative potential energy surface, with a lifetime on the order of vibrational relaxation. Pressure effects on photoreaction efficiencies would then be largely confined to viscosity perturbations and the rigidity of solvent cages around the molecule. [L,M-X] *

+Y

+

LM-Y

+X

(6.15)

When initial excitation is followed by relaxation to a bound lowest energy ES, the quantum yield CD, for ligand substitution from that state would equal @isckst, where k, is the rate constant for ligand substitution from the LEES. Thus the activation volume AV; for k, can be evaluated if the pressure dependencies of Q S , Disc

6.2 Unimolecular Excited State Reactions

and T are separately determined. Such parameterization has been carried out for the photosubstitutions of a limited number of metal complexes; and examples of such studies of rhodium(III), ruthenium( 11) and chromium(111) complcxes and several group VI metal carbonyls will be described here. For a vibronically relaxed bound ES, ligand substitution mechanisms can be discussed in terms of models developed for analogous thermal reactions [ 361. The limiting mechanisms would be the dissociative (D) and associative (A) pathways, where the rate-limiting steps are, respectively, dissociation of the M-X bond or formation of the M-Y bond to form distinct intermediates (Eqs 6.16 and 6.17). The electronic nature of such intermediates is ambiguous, since these species may also be electronic excited states. For example, the cis to trans isomerization concomitant with the photoaquation of C1- from the Rh(II1) complex cis-Rh(NH3)qCl: was successfully explained by a model where C1- dissociation gave a pentacoordinate intermediate in a triplet LF excited state [37, 381. D:

[L,M-XI *

+Y

+

[L,M]*?

+X +Y

A:

[LnM-X]*

+Y

+

[LnM(Y)X]*?+ LnM-Y

+

L,M-Y

+X

+X

(6.16) (6.17)

Lying between the limiting A and D pathways would be the interchange mechanisms (I, or Id). These involve concerted exchange between the first and second coordination sphere without formation of discrete intermediates and with bond formation or breaking being energetically more important, respectively. A truly concerted mechanism would require that the ground and excited state surfaces cross along the reaction coordinate so that intrinsic volume changes AV& along the substitution reaction coordinate also include contributions from the AV between the ES and GS. Based on volume profiles of thermal reactions, the A@,, contribution to AV; would be expected to be negative for an A mechanism and positive for a D mechanism. Qualitatively, thermal I, and Id pathways would show similar patterns, but with the AVint contribution somewhat attenuated (see Chapter 4).The other key contribution to AV: derives from solvation changes, AV;,,, which can easily exceed AV& in magnitude for reactions with charge creation or neutralization and which may be opposite in sign. This is illustrated by Fig. 6.9, which shows the impact of solvation changes during dissociation of an anion. As noted above, the reactive ES may have a different partial molar volume than the analogous GS. Other than the photoacoustic calorimetry experiments described earlier, there is relatively little direct quantitative information on E S volumes. The magnitude of the AV between the GS and the ES will depend on the excited state orbital parentage. For example, there is little distortion upon forming the doublet LF state of an octahedral Cr(II1) complex, so both intrinsic and solvation contributions to AV should be quite small. In contrast, the quartet LF state of Cr(II1) and the triplet LF excited states of Rh(II1) complexes, both formed by metal centered eg(ohL)+ tls(nML) excitation, would be expected to have substantially positive AVint contributions, but again relatively small differences between ground and excited

I

197

198

I

G Pressure Effects on the Photoreactions ofhansition Metal Complexes

I

TS Volume profile diagram showing the contributions of intrinsic and solvational activation volumes to the observed AV; of a hypothetical ligand substitution via limiting D mechanism. Fig. 6.9.

state solvation. In contrast, since charge transfer states involve greater intramolecular dipole changes, larger solvation contributions would be expected to accompany MLCT and LMCT excitation. Rhodium(ll1) complexes: Collaborative studies between van Eldik, Ford and coworkers have led to thorough parameterization of pressure effects on photosolvolysis of the rhodium(111) halopentaammines Rh(NHj)sX2+ (Eq. 6.18) [39-451. For these systems LF excitation is followed by rapid intersystem crossing ((Disc z 1) to the lowest energy LF state 3 E from which reactive (kp),radiative (k,) and nonradiative (kn) deactivation occur competitively (Fig. 6.10) [41,461. Rate constants for individual excited state processes were calculated from phosphorescence quantum yields @, lifetimes T and quantum yields for halide (a,) and ammine (@)A) solvolysis according to ki = Q ~ Tin different solvents over a range of temperatures and pressures [47].

x

2'

2'

(6.18)

+ HZO

A

6.2 Unimolecular Excited State Reactions

I

199

Fig. 6.10. Jablonski type diagram for the lower energy ligand field states o f a CdV Rh(lll) complex such as Rh(NH9):' showing reactive (kp), radiative (k,) and nonradiative (kn) deactivation from the lowest energy triplet state.

The @x and @A values proved to be dramatically solvent dependent. Labilization of CI- from Rh(NH3)5C12' predominates in the very polar solvents water or formamide solutions; NH3 labilization predominates in methanol, DMSO and DMF [48]. These photoreactivity patterns were attributed to the much greater solvent sensitivity of the C1- substitution rates from the ES. Activation volumes for ammine and halide labilization and for non-radiative deactivation are summarized in Table 6.1. Immediately apparent from these data are the large positive AVL values for ammine substitution and the large negative AVL values for halide substitution in all solvents. In order to interpret these data, we shall compare the volume changes predicted for the limiting A and D mechanisms. The A mechanism would proceed via associative attack of solvent S to give a seven-coordinateintermediate Rh(NH3)5(X)(S)'+ ( I 1 ) followed by ligand loss to give the two products. The volume of the transition state lies along the volume profile between the reactants and I1; thus AV; would have a value between 0 cm3 mol-' and the difference V(Ii) - (V(ES) +Vi(S)). Although V(I1) and V(ES) are both unknown, they can be estimated to give a AV for the formation of 11 -15 cm3 mol-' 1421. According to this mechanism, similar negative values would be expected for both AVL and AVL. The positive AVL values found experimentally thus argue against the operation of a common associative

-

200

I

6 Pressure Effects on the Photoreactions of Transition Metal Complexes

Volumes o f activation (in c m 3 molL') calculated from pressure effects on the photoreaction/ photoluminescence properties o f the Rh(lll) complexes Rh(NH3)X2+ in various solvents (data taken from refs. [47] and [48]).

Tab. 6.1.

Complex

Solvent

Rh(NH3)5C12'

H2O Dz 0 FMA DMF DMSO HZO D20

Rh(NH3)5Br2+

AV:!.. @x

-5.2 f 0.4 -4.2 f 0.5 -4.6 f 0.7 -7.8 k 1.8 10.3 f 1.2 -9.4 f 1.5

.-

.?:

/jV$

h

12.7 k 1.2 9.5 f 1.6 4.2 f 0.9 6.3 _+ 0.9 4.4 0.9 4.6 k 0.6 3.5 f 0.5

+

d

(-3.4) 3.5 f 1.1 0.3 k 0.4 1.3 0.2 -1 k 1 (3.5) 4.1 k 0.6 ~

~

Av.? X

-8.6 k 1.6 -7.7 k 1.6 -4.9 f 1.1

-8.9 f 2.7 -6.8 k 1.6 -5.8 f 1.8

AVt

AV i

9.3 k 1.9 6.2 k 2.2 3.9 f 1.3 7.6 f 1.1 3.3 f 1.8 8.1 f 1.2 7.5 f 1.1

(-2.6) -2.6 k 1.0 0.2 0.5 0.7 0.3 -1+1 (+2.5) +2.5 k 1.2

*

mechanism for the excited state substitution pathways. In contrast, a limiting D mechanism leads to two different Rh(II1) intermediates (Eq. 6.19), one being +2 and the other $3 in electrostatic charge. A substantial AV! difference between the halide and ammine photolabilizations would be expected owing to solvation terms. The AVB,, contributions to AV; would be negative owing to charge creation as halide dissociates from Rh(NH3)5XZ+ion to form the Rh(NH3):' plus X-. However, NH3 dissociation should afford no appreciable charge creation, hence minor contributions from AV:ol would be expected.

-c A..

hv

*+ J

s

3+

.A

A

6

A..

I

2'

T

2'

.A

,,RhLA (6.19)

Figure 6.11 presents a volume profile for the photoaquation reactions of Rh(NH3)5C12+according to a limiting D mechanism. The volume changes calculated for the overall photoreactions from ground state to products are +3.9 cm3 mo1-l for ammine aquation (AVA)and -17.8 cm3 mo1-l for C1- aquation (AVcl), the difference largely attributable to charge creation upon C1- aquation. A similar AV (- 22 cm3 mol-*) might be expected to exist between the intermediate species from C1- dissociation and from NH3 dissociation. Notably, the AAVt for these pathways (AV; - AVi) was measured to be 17 cm3 mol-'. An analogous conclusion can be drawn for the Rh(NH3)5Br2+ion in aqueous solution. The difference in the overall AV values for NH3 vs. Br- photolabilization is 19.7 4.1 cm3 mol-l, while the excited state reactions displayed a AAVJ of 14.9 2.8

G.2 Unimolecular Excited State Reactions

I 20 10

1 +3.9

w

I 3 0 0

>

I

-10

TS

\ -17.8

-20

IAH values in cm3 mot-' Fig. 6.11. Volume profile diagram for the competitive photoaquation of NH3 and of CI- from Rh(NH3)SCI2+in aqueous solution via a proposed limiting dissociative mechanism.

cm3 mol-', again pointing to charge creation as defining the differences in the pressure effects for the two excited state reactions. Thus, the pressure data are consistent with earlier studies that point strongly toward a dissociative mechanism for the ligand photosubstitutions of the halopentaamminerhodium(111) complexes [36, 381. Features of the data in Table 6.1 and other data collected elsewhere [39, 42, 451 relate to the non-radiative pathways from the 3LF. For these photoactive Rh(II1) complexes, the pressure dependence of k, generally has the same sign as that for the major photoreaction pathway but a smaller absolute value. This suggests parallel character for k, and a strong-coupling contribution to k,, [49]. Such a pattern would be consistent with a reaction coordinate which approaches the ground state surface in a manner that allows partitioning between ligand dissociation and nonradiative deactivation. Chromium(ll1) complexes: Pressure effects for both the emission and photo-

reaction quantum yields under comparable conditions have now been described for several Cr(II1) complexes in fluid solution including Cr(bpy):+, Cr(NH3);+, Cr(NH3)5(NCS)'+ and the cis and trans isomers of Cr(cyclam)(NHj);+ (cyclam = 1,4,8,11-tetraazacyclotetradecane) [ 3, 50-571. There is a rich and subtle literature regarding the photochemistry of hexacoordinate Cr(111) complexes to which such pressure studies have contributed insight. The first quantitative studies of pressure effects on the photosubstitution reactions of transition metal complexes were reported by Kelm et al. for a series of Cr(NH3)5X2+ ions (X- = SCN-, C1- or Br-) in aqueous solution [SO-521 (Eq. 6.20). These workers measured photoaquation quantum yields as a function of vs. P. Notapressure and determined the apparent AVf values from plots of In bly, the AVf values were substantially negative regardless of whether X- or NH3

I

201

202

I

G Pressure Effects on the Photoreactions

Q1

of

Transition Metal Complexes

\

DO

PRODUCTS

QO

Fig. 6.12. A simplified generalized energy level diagram for the photoreactions of Cr(lll) complexes. Initial excitation leads t o formation of the Franck-Condon quartet state which is followed by relaxation t o the thermally

equilibrated quartet ( Q ,) and doublet (Do) excited states, k,, k, and k, are the decay pathways direct from the doublet, and k; and k6 are the decay pathways from Q1.

was labilized. Owing to charge creation, AVto, contributions to AVL values for halide aquation are likely to be negative regardless of the mechanism. However, the negative AVL values for ammine aquation suggest an associative mechanism

PI.

+

Cr(NH3)5X2++ HzO 5 Cr(NH3)5(HzO)3f X

(6.20)

Figure 6.12 is a generic energy level diagram for chromium(II1) ligand substitutions as possible reactive deactivation pathways. Much of the discussion regarding Cr(111) photoreaction mechanisms has been concerned with the roles of the lowest energy quartet (Q1) and doublet (Do) states. For hexa-coordinate complexes with moderate to strong ligand fields, the LEES is the Do state with a (t2g)3electronic configuration. As noted in Sect. 6.1, Do is relatively undistorted, while Q1 with a (t2g)z(es)' configuration should be considerably more distorted from the ground state structure [58]. It is this distortion that makes Q1 an attractive candidate as the excited state responsible for photosubstitution reactivity. Chromium(II1) photosubstitutions have been discussed [59] in terms of two limiting mechanisms: (1) intersystem crossing from quartet states formed by initial excitation to the doublet manifold followed by reaction from the relatively longlived Do, or (2) reaction from the quartet state(s) via thermally promoted back-

6.2 Unimolecular Excited State Reactions

intersystem crossing from the doublet and/or prompt reaction upon formation via initial excitation. The pitfalls of making too simple an interpretation of pressure effects on the photochemistry of hexacoordinate Cr(111) complexes have been discussed by Endicott and Ryu [GO], who analyzed earlier pressure studies of Cr(NH3);' and Cr(NH3)5NCS2+photochemistry and photophysics. They noted the complications induced by the potential roles of the two states and the possibility of back-intersystem crossing as well as multiple passes through the lower energy state, which make calculation of pressure effects on individual rate constants in Fig. 6.12 subject to considerable ambiguity. Nonetheless, these workers concluded that the pressure effects support an associative mechanism (A or Ia) for ligand labilization regardless of the reactive state responsible. Waltz and coworkers subsequently reexamined pressure, temperature and medium effects on the doublet emission and photoaquation of Cr(NH3);' [56]. Emission lifetimes vary with solvent, and a strong correlation was observed between kd (T-') and the solvent donor character. This was attributed to solvent interactions with the N-H bonds of the coordinated ammines. Corresponding activation energies E,(kd) and activation volumes AV: ranged from 46 kJ mol-l and f4.3 cm3 mol-' in water to 51.5 kJ mol-l and +3.4 cm3 mol-' in HMPA, similar to calculated energy and volume differences between the Do and Q1 states. On this basis, it was concluded that at room temperature and above, non-radiative deactivation occurred primarily via thermally-activated back-intersystem crossing from Do to Q1. The k,, from the latter is much faster owing to strong coupling contributions. Quantum yields for the Cr(NH3);' photosolvolysis reaction (Eq. 6.21) are not solvent sensitive (0.44 in H2O and DMF) and are independent of whether initial excitation is into the Q1 or Do states, as are the activation volumes for photoaquation (-6 cm3 mol-l). These data led Waltz et al. to conclude, in agreement with Endicott and Ryu [GO] and with Angermann et al. [SO], that pressure effects suggest excited state ligand substitution occurs via an associative mechanism.

Cr(NH3)t+

+ H2O 3Cr(NH3)5(H20)3++ NH3

(6.21)

Group VI Carbonyl complexes: In the case of substitution of neutral ligands by

neutral ligands, pressure effects can be better correlated with the intrinsic volume changes associated with the mechanism. One such study dealt with the photosubstitution reactions of the hexacarbonyls M(C0)6 (M = Cr, Mo, W) to give M(C0)sL (Eq. 6.22) and M(CO)qL2, where L is a ligand such as pyridine [ G l ] . For each M, @co decreased with increasing pressure. Under the risky assumption that k, is independent of P, the pressure dependence of @/(I - CJ)can be used to estimate AVJ for the ES reaction. The positive values derived are consistent with a D mechanism for CO release. However, laser flash photolysis techniques have shown that CO loss to form the 5-coordinate intermediate M(CO)5 occurs in less than 1 ps. For this reason, one cannot treat the ligand substitution pathway from the reactive ES in terms of mechanisms elucidated for bound excited states. Instead the positive

I

203

204

I

G Pressure Effects on the Photoreactions of Transition Metal Complexes K,,

= [MLCT']/[LF']

L 1

Fig. 6.13.

Schematic diagram representing the various reaction routes of W(C0)sL complexes.

values of AVL may reflect solvent viscosity effects on the escape of CO from the solvent caged pair [ M(C0)5, CO] generated immediately upon flash photolysis. M(CO),

+ L 5M(CO),L + CO

(6.22)

A different situation is represented by a series of pentacarbonyl complexes W(CO)sL, involving either a LF ( L = pyridine) or a MLCT state (L = 4-acetyl and 4-cyanopyridine) as the LEES [62]. Excitation is followed by rapid internal conversion-intersystem crossing to the lowest excited state, and the proposed model suggests that the observed photochemistry in the MLCT state results from back population to the LF state (Fig. 6.13). Pressure studies demonstrate positive AVf values for each W(C0)sL complex [62], and the larger effects observed for L = 4acetylpyridine and 4qanopyridine (AVL = f9.9 and f6.3 cm3 mol-') are consistent with the volume difference expected between the LF and MLCT excited states. Photosubstitutions of the M(C0)4(phen) complexes (M = Cr, Mo, W, phen = 1,lo-phenanthroline) have received attention because of possible roles played by both LF and MLCT excited states. The working model from the pentacarbonyl complexes would suggest that the photosubstitution proceeds dissociatively from the LF excited state, and that MLCT excitation is followed by thermal back population into the LF state [63]. However, it has also been argued that the MLCT states themselves may be reactive toward associative substitution owing to the more electrophilic nature of the metal center in the ES [64]. In order to probe this, the effect of pressure on Eq. (6.23) as a function of Iirr was studied [65, 661. For L = PEt3 and M = W, negative AVL values were found for 546-nm irradiation of MLCT bands in contrast to positive activation volumes AVLs for ligand substitution when LF bands were excited with 366 nm light. This argues for a D mechanism in the latter case but an A mechanism in the former. Similar behavior was seen for the molybdenum and chromium analogs [67-691.

6 2 Unimolecular Excited State Reactions

M(CO),(phen)

+ L -fac-M(CO),(L)(phen) + CO hu

(6.23)

Ruthenium(l1): Ru(bpy):+ has been the subject of more photochemical studies than any other metal complex, owing partly to interest in using this visible absorber as a “catalyst” for the conversion of radiant energy to chemical potential energy. Its MLCT excited state is relatively long lived ( r 1 1s) in ambient temperature solutions including water. This ES has strongly reducing and oxidizing character thus is susceptible to efficient electron transfer quenching by a number of substrates (see Sect. 6.3.1). Furthermore, it displays little photoactivity toward decomposition under most such conditions, so it can be used for numerous turnovers in electron transfer cycles. Nonetheless, [ Ru(bpy);’]” does undergo some photosubstitution chemistry [ 70-721. For example, in acetonitrile solution containing chloride, the following reaction is observed.

-

+

[Ru(bpy).i+]* C1-

+ CH3CN 2 R u ( ~ ~ ~ ) ~ ( C H +~ bpy CN)C~~

(6.24)

The proposed excited state substitution mechanism is that described above for non-radiative deactivation at higher T, namely, back population of a ’LF state lying above the lowest energy 3MLCT state. This state would undergo dissociation (or solvent associative displacement) of one Ru-N bond to give a monodentate bipyridine intermediate. The role of Clk is attributed to trapping of the coordination site to give Ru(bpy)z(tI’-bpy)Clt, which undergoes further dissociation of bpy. The strong temperature dependences of ligand labilization and of nonradiative deactivation have been rationalized in terms of the reversible thermal promotion to the ’LF state. Fetterolf and Offen [72] carried out pressure experiments for Ru(bpy):+ at several temperatures and solvents and in the presence of both C1- and PF, in order to evaluate this mechanism. The effect of pressure was to slow the photosubstitution with AVL (app) values +12 and +22 cm3 mo1-l determined at 15 and 60 “C, respectively. The large temperature dependence is related to the requirement for ion association between Cl- and the Ru(I1) cation leading to desolvation and a positive contribution from AVto,. A change to the non-ligating PF, counterion reduces the photosubstitution rate in CHjCN considerably and gives only the RU(~PY)~(CH~C product. N ) ~ + In this case, AVL (app) (- 10 cm3 mol-’) is temperature independent, even though the rates are quite temperature dependent. Since this parallels the values seen for k,, one can conclude that AVL (app) largely reflects the 3MLCTto 3LF photophysical process. Ru(phen)3Clz behaved in a similar manner. 6.2.2

Unimolecular Photoredox Decompositions

Charge transfer states are (formally) internal redox processes which are reversed by relaxation to the ground state. However a CT state could also decay by dissociation of the oxidized or reduced ligand resulting in net redox decomposition. An exam-

I

205

206

I

6 Pressure Eflects on the Photoreactions of Transition Metal Complexes

ple is the LMCT excitation of the cobalt(II1) ion Co(NIH3)sBr2', which undergoes photoreduction to give the labile Co2+ center (Eq. 6.25). For this reaction Kirk et al. [73] reported a AV: value of +4.8 cm3 mol-' and suggested the formation of a caged value of radical pair (Co", Br') from the LMCT excited state. Dissociation of this radical pair to products was suggested to account for the positive AVL.

2 Co2+t 5NH3 +products of Br' formation

Co(NH3)5Br2+

(6.25)

Similarly, photolysis of nitroprusside ion Fe(CN)=,NOZ- via MLCT excitation results in the oxidation of the metal and labilization of NO (Eq. 6.26) [74]. The quantum yicld for the production of Fe(CN)5S2- is moderately sensitive to the viscosity of the medium. The values of AVL are significantly positive (+7 to f 8 cm3 mol-') and point to the participation of a cage recombination mechanism. [(CN)5Fe'1(NOf)]3--% [(CN),Fe"',NO){'-

2[(CN)5Fe1"(Sol)]2-+ NO (6.26)

6.3 Bimolecular Pathways For Excited State Deactivation

Bimolecular quenching may occur by energy transfer to another acceptor or by electron or atom transfer to give an oxidized or reduced species. Other bimolecular pathways include reaction of the LEES with another species in solution to form short-lived exciplexes. In the presence of a quencher Q the ES decay rate is defined as: 5-l

= k,

+ k, + k,

+ k,[Q]

(6.27)

and kq can be calculated from (6.28) where 50 is the lifetime in the absence of quencher. Values of AV: are determined in the normal manner from the dependence of k, on P, although modest corrections of [Q] are made for compressibility, the appropriate factor amounting to a few percent over the range 0.1 to 300 MPa depending upon the solvent. The AVfOlcontribution to AV; is especially important in bimolecular quenching by electron transfer processes owing to electrostriction effects resulting from charge creation or annihilation [75]. On the other hand, it is small for energy transfer quenching unless k, approaches the diffusion limit kD. For any bimolecular pathway, pressure induced changes in the solution viscosity become crucial as k, approaches kD, given the simple relationship between kD and the viscosity 11 1761:

kD

= 8RT/2000~

(6.29)

6.3 Birnolecular Pathways For Excited State Deactivation

In this situation, AV;, will be dominated by the pressure dependence of the viscosity regardless of the mechanism. As YI, increases with P, rates slow giving positive values of AV; (and therefore of AV;). For solvents such as acetonitrile and ethanol, AVL is large (19.5 and $7.7 cm3 mol-' respectively), but for water it is small (-1 cm3 mol-l) owing to the relative incompressibility of that medium [77]. Quenching can be envisioned as the reversible formation of a precursor complex between the ES and Q followed by energy or/and electron transfer (Eq. 6.30).

(6.30)

For this model

(6.31)

When kel +ken >> k - D this simplifies to kq = k D , or in more general terms Eq. 6.31 can be rewritten as:

(6.32)

where KA = k D / k - D . 6.3.1

Excited State Electron and Energy Transfer Reactions

An example of this behavior is the bimolecular quenching of the MLCT emission from the 2,9-diphenyl-1,lO-phenanthroline (dpp) complex Cu(dpp)f by different electron and/or energy acceptors such as a series of chromium(111) acetylacetonate (acac-x) derivatives that can quench by either mechanism (Eq. 6.33) [78]. These Cr(acac-x)3complexes display a wide range of reduction potentials (-0.79 to -2.43 V in CHlClz, referenced to the Fci/Fc couple, Fc = ferrocene) dependent on the nature of the substituents. However, their doublet ligand field ES energies are relatively invariant [(12.5 0.4) x lo3 cm-l] [79]. Electronic energy transfer is favorable in each case, since the energy of the MLCT excited state 3[Cu(dpp):]* is -14.5 x lo3 cm-l. The C~(dpp)i+/~[Cu(dpp):];~ couple was estimated as -1.41 V (versus Fc+/Fc) by subtracting the MLCT excited state energy (1.80 V) from the standard potential determined for the ground state [78]. From these data Eq. (6.34) can be used to calculate the free energy change for Eq. (6.33).

I

207

208

I

G Pressure Effects on the Photoreactions of Transition Metal Complexes

3[Cu(dpp):]+c+ Cr(acac-x)3

Cu(dpp)?

+ Cr(acac-x)j

Cu(dpp): i [Cr(acac-x)3]*

(6.33)

AGZl = -{El,2(Cr111/Cr'1)- E 1 / 2 ( C ~ ' 1 / X C ~ 1 ) }

(6.34)

For the unsubstituted acetylacetonate (R = CH3, X = H), AG;, is quite positive (+1.02 V), thus electron transfer quenching would be highly unfavorable, and quenching must be due solely to energy transfer. In this case, a k, of 1.5 x lo7 M-'s-' was observed in dichloromethane [go]. Comparable rates were also seen in CHC13 and THF solutions. In all three solvents AV; is quite small (-0 cm3 mol-'). Since energy transfer leads to little or no charge creation, A&, must be small. Thus, AV!nt must also be small, indicating that neither the MLCT excited state 3[Cu(dpp)z]*nor the doublet ES formed by energy transfer to CrL3 are dramatically distorted from the respective ground state configurations. At more favorable free energy for the redox reaction, k, increases (the turning point being AG:l 0.2 V) consistent with the increasing importance of outer sphere electron transfer as a component of this quenching. Thus for Cr(tfac)3.(tfac is trifloroacetylacetonate), AG;] = 0.23 V and k, = 4.2 x lo7 M-' s-' in dichloromethane. What is much more striking is that AV: becomes significantly negative (-8.1 cm3 mol-'). This was attributed to the solvent electrostriction contribution (AVLol) from charge creation in the transition state of the electron transfer mechanism. The third type of behavior for these systems is seen when the excited state electron transfer becomes much more favorable and k, approaches the diffusion limit. For example, 3[Cu(dpp)2]$cquenching by Cr(hfac)3 (hfac = hexafloroacetylacetonate) has a calculated AGZ, value of -0.62 V, and k, = 0.94 x 10" M-'s-'. In this case, AV; is quite positive ( t 8 . 0 cm3 mol-'), reflecting the role of pressure in tuning the solvent viscosity. Figure 6.14 illustrates the contrasting pressure dependence of the respective k, values in these three quenching experiments. These data as well as k, and AV; values for other quenchers are summarized in Table 6.2. The quenching of the CC" luminescence from the copper(I) clusters Cu414py4 by a series of Cr(acac-x)3and other electron transfer and energy transfer acceptors demonstrated a similar pattern of pressure effects [Sl]. For the [Cu414py4]jrluminophore, energy transfer quenching demonstrated virtually no pressure effect for Q = Cr(acacj3 (k, = 4.7 x lo7 M-l s-l, AV; = -0.3 cm3 mol-'), while a large positive AV; (+6.2 cm3 mol-') for diffusion limited redox quenching (1.2 x 10" M-' s-l ) by Fcf. Large negative values, (-8.2 cm3 mol-') were the case for

-

G.3

Bimolecular Pathways For Excited State Deactivation

Pressure (MPa) Fig. 6.14. Plots of In(k,/k:) vs. P for quenching of '[Cu(dpp)2]>* by C r ( t f a ~ ) ~ (upper), Cr(acac)3 (middle) and Cr(hfac)3 (lower) in dichlorornethane a t 23 "C. For Cr(tfac)j, k, is dominated by an electron

transfer mechanism (A$ - -8.1 c m 3 mol-'); for Cr(acac)g by energy transfer (AV; = -0.1 c m 3 mol-'); and for Cr(hfac)3 by diffusion kd (AV; = +8.0 cm3 mol-').

quenchers such as m-dinitrobenzene, where electron transfer is favorable (AGEl = -0.41 V) but energy transfer is not and k, is significantly less than the diffusion limit. As might be expected one of the earliest pressure effect studies of the quenching of a metal complex excited state involved the aqueous [Ru(bpy)i"]" ion. Kirk and Porter [24] found that the near diffusion limited quenching of aqueous Ru(bpy):+ by o2(1.7 x 109 M-I s - 1 ) showed a very small AV; ($0.5 cm3 mol-I), which was attributed to a modest increase in the viscosity of water over the investigated pressure range. However, a similarly small AV; would have been expected anyway if this quenching were occurring via an energy transfer pathway. Fetterolf and Offen [2G] studied the effect of pressure on the reductive quenching of Ru(bpy)f+ by several aromatic amines. Quenching by N,N-dimethylaniline (DMA) in CH3CN resulted in AV! values between +1.3 and 4 2 . 9 cm3 mol-I while values of 9 to 13 cm3 mol-' were reported for the quenching by the free base benzidine and N,N,N',N'-tetra-methylbenzidine (TMB) in CH3CN and n-BuOH as solvents. These pressure effects can be interpreted in terms of the following mechanism

(6.35)

I

209

210

I

6 Pressure Efects on the Photoreactions of Transition Metal Complexes k,'s at ambient pressure and AVi's for quenching of emission from '[Cu(dpp):]* in dichloromethane (data from ref. [51]).

Tab. 6.2.

Q Cr(hfac)3' p-Dinitrobenzene Cr(tfbzac)3 Cr(W3 Cr(tc-bzac), Cr( br-dbm)g Cr(n - a ~ a c ) ~ p-Chloronitrobenzene Cr(tfac)3 Cr(tc-acac), Cr( br-acac)j Cr(dbm)3 Cr(acac),

Ace, (e")

k4

(V)"

-0.79 -1.18 -1.43 -1.43 (-1.48) (-1.53) (-1.57) -1.58 -1.64 (-1.66) (-1.89) -1.87 -2.43

-0.62 -0.23 -0.02 0.02 (0.07) (0.12) (0.16) 0.17 0.23 (0.25) (0.48) 0.46 1.02

940 1050 200 220 60 4.2 6.3 3.0 4.2 3.4 6.7 6.8 1.5

El/2

(107 M-1

AV8 4

5-1

)

(cm3 rnol-')

+8.0 +6.9 +0.8

+2.1 +3.8 -3.5 -3.8 -20.4 -8.1 -1.4 -2.4 -0.3 -0.1

aE1/l values versus the Fe(cp)T/Fe(cp)2couple in CH2C12; values in parentheses estimated from parameters as discussed in ref. 51. bAC:, = -{E1/Z(Crlll/Crll) - El/z(Cull/qCul)}.e,e is the elementary

charge. Cacac= acetylacetone; dbm = dibenzoylmethane; hfac = 1,1,1,5,5,5hexafluoro-2,4-pentanedione; tflJzac = 4,4,4-trifluorobenzoylacetone; tta = thenoyltrifluoroacetone; tc-bzac = 2-thiocyanato-l-phenyl-1,3butane-dione; br-dbm = 2-bromo-l,3-diphenyl-l,3-propanedione; tfac = l,l,l-tri-fluoro-2,4-pentanedione; n-acac = 3-nitro-2,4pentanedione; tc-acac = 3-thiocyanato-2,4-pentanedione; br-acac = 3bromo-2,4-pentanedione.

for which k, = kD in the diffusion-controlled limit and k, = k,lKA in the activation-controlled limit, where KA = kD/k-D the association constant for the precursor complex. The large positive AV; values observed for the quenching by benzidine and TMB were attributed to pressure effects on solvent viscosity in these nearly diffusion-limited reactions (kq = 3.0 x 10' and 7.6 x lo' M-' s-' , respectively, in 298 K acetonitrile at ambient pressure). For the slower quenching by DMA (k, = 9.1 x l o 7 M-' s-l ) it is likely that the small positive values for AV; largely reflect positive AV:,l contributions to k,, owing to delocalization of the positive charge. Detailed analysis suggested that changes in the dielectric constant of the medium may account .for the observed effects in the latter case [26]. Oxidative quenching by dimethylviologen (MV2+) (Eq. 6.36, k, = 9.5 x 10' M-l s-l in aqueous solution) has been reported to exhibit a modestly negative AV; (-2.6 cm3 mol-') [24]. Interestingly, recent PAC experiments [75] indicate that the overall volume change for Eq. 6.36 is positive. Thus, it appears that the transition state of the electron transfer reaction is dominated by formation of an encounter complex with a smaller volume than the separated reactants most likely due to solvation effects. Separation of the encounter complex into products has a large, positive AV (Fig. 6.15).

6.3 Birnolecular Pathways For Excited State Deactivation

(Di I

+ A')

'7-

Fig. 6.15. Volume profile for the oxidative quenching o f [Ru(bpy)3'+]" by rnethylviologen (MV'') (redrawn from reference 75).

+

[Ru(bpy):+]* MV2+ -+ Ru(bpy):+

+ MV'

(6.36)

A somewhat different type of redox quenching mechanism has been proposed for the reaction of the excited state of the Pt(I1) dimer, Ptz(p - q2-H2Pz05)2-, with organic substrates [82]. The AV; values for quenching of the triplet excited state 3Pti, by a series of benzylic compounds are between -2.6 and -5.7 cm3 mol-' in methanol. These values are consistent with an associative interaction between the benzylic H atom donor and the excited-state complex, presumably at metal atom site trans to the Pt -Pt bond, resulting in H atom abstraction according to Eq. 6.37. In contrast, 0 2 quenching of 3Pt;, is close to diffusion limited (kq = 3.3 x 10' M-' sP1 in ambient methanol) and exhibits a AV; value of +2.8 cm3 mol-' reflecting contributions from pressure-induced viscosity changes.

31't;

+ H-R + HPt2 + R.

(6.37)

6.3.2

Brnnsted Base Quenching

The trans-Rh(cyclam)(CN): ion (cyclam = 1,4,8,11 -tetraazacyclotetradecane) displays luminescence from a ligand field excited state ('LF") in room temperature, aqueous solution with a lifetime (8.1 ps) [83], several orders of magnitude longer than generally observed for rhodium( 111) amines [ 371. This emission is quenched by OH- in solution (Eq. 6.38), a process attributed to amine deprotonation [84] at rates approaching the difFusion limit (k, 10'O M-' s-l). The pressure effects on

-

I*"

212

I

6 Pressure Effects on the Photoreactions of Transition Metal Complexes

this and for analogous quenching of the 3LF;’<emission from Rh(ND3)i+(by ODin D20) were examined by DiBenedetto [39]. The resulting AV; values were positive, f4.1 0.5 and +4.0 f 0.6 cm3 mol-’, respectively. Positive AV?,s, would be predicted by two factors. For reactions which approach diffusion-limited rates, positive AVk are expected owing to the increased solvent viscosity under pressure; however, the relative incompressibility of water makes this contribution very small (<1 cm3 mol-’). A large positive contribution to AVi is also expected from the desolvation accompanying the charge neutralization resulting from proton transfer from the ES to OH-. The change in partial molar volumes AV would be quite large for complete proton transfer, since the AV between OH- and FIzO is -16 cm3 mol and desolvation of the complex ion should give a large and positive AV as well. Thus, the much smaller value of AV; likely reflects a very early transition state for the proton transfer leading to ES quenching.

’

3[trans-Rh(cyclam)(CN):] *

-ki[OH-]

trans-Rh(cyclam)(CN):

(6.38)

6.3.3 Reactions with Lewis Bases to Give an Excited State Complex

The copper(I) complex Cu(dmp): (dmp = 2,9-dimethyl-l,lO-phenanthroline) displays MLCT luminescence in ambient temperature CH2Cl2 solutions [85]. This emission has been shown to be quenched by various Lewis bases (B), and the mechanism proposed is addition of B to the MLCT state at the metal center to give an “exciplex” which decays rapidly (Eq. 6.39). The validity of this mechanism was tested by comparing, in the presence and absence of Lewis base quenchers, the pressure effects on the emission lifetimes of Cu(dmp); with those on the emission lifetimes of the bulkier 2,9-diphenyl-phen analog Cu(dpp): [ 861. The lattice ions should not be as susceptible to reaction of the copper center with B. For both ions, emission quantum yields are small ( <w3) at ambient T and unimolecular photoreactions are not observed, so the pressure sensitivity o f t reflects non-radiative deactivation mechanisms. [Cu(dmp)f]*

1

kn

+

[Cu(dmp), : B+]*

-B

Ik.

Kb[Bl

+ kr

Cu(dmP):

fast

(6.39)

[Cu(dmp), : B f ]

The application of pressure to CH2C12 solutions of either Cu(dmp): or Cu(dpp),t led to systematic decreases in T even in the absence of added Lewis bases. However, the ln(v,/.) vs. P plots showed the emission lifetimes of \Cu(dmp)t]” to be significantly more pressure sensitive, the respective AV; values being -3.4 and -1.6 cm3 mol-’ [86]. The small negative value for [Cu(dpp);]” is similar to those seen

6.4 Photochemically Generated Reactive Intermediates

I

213

for other mctal complexes decaying via a weak coupling mechanism. The more negative hV: for [ Cu(dmp):]+ is outside this range and suggests participation of an associative pathway as a component of non-radiative deactivation, i.e. the k, pathway may involve formation of an ES solvent complex. The differences between the two complexes were more pronounced with added Lewis bases which quenched [Cu(dmp)F]" but not [Cu(dpp)t]". The presence of 0.30 M CH30H in the CHzClz solutions shortened the 5 for [C~(dmp):]~~ at ambient pressure from 90 to 66 ns. Calculation of k, at various P gave AVis of -5.4 and -6.2 cm3 mol-' for [ C ~ ( d m p ) t ]quenching +~ by CH3OH and CH3CN, respectively. Although partial molar volumes of neither [Cu(dmp):]" nor of the putative ES/Lewis base complex are known, an associative pathway, such as Eq. (6.39), where k, = Kbkk, should display a negative AV;. The negative values both of AV; and of AV:, for [Cu(dmp);]" are consistent with this non-radiative deactivation mechanism.

6.4 Photochemically Generated Reactive Intermediates

An electronic excited state often reacts to give a ground state product that itself is reactive and undergoes rapid transformations to more stable species. Flash photolysis techniques like those used to probe excited states dynamics are also valuable in probing the subsequent thermal substitution, isomerization, electron-transfer, etc. reactions of such reactive intermediates. In contrast to ES reactions, the kinetics of the reactive intermediates are not dominated by photophysical processes such as non-radiative deactivation, allowing somewhat more straightforward interpretations of pressure effects. Since related studies have been described in other chapters regarding thermal reaction schemes, this section will present only a limited set of examples concerned with ligand substitution reactions of various rnetalloporphyrins in order to illustrate pressure effects on reactive intermediates. Model heme systems: The mechanisms of heme and hemoprotein reactions with small molecules such as 02,CO and NO has attracted considerable experimental attention owing to the importance of such processes in biological systems. Flash photolysis studies [ 871 have indicated that the photolabilization of L from simple heme complexes and kinetics of the resulting back reaction (Eq. 6.40) can be modeled by the intermediacy of solvent caged "contact pair". Equation (6.41) illustrates this mechanism for the thermal back reaction for the photochemically generated intermediates for a ferrous porphyrin (Por)Fe" L (Por = porphyrin)

hv >i/v (Por)M-L F== Por(M) k"[L]

(Por)Fe"

ko + L G=+ k-D

+L

{(Por)FeTT, L}

(6.40) k, k-,

(Por)Fe"-L

(6.41)

214

I

G Pressure Eflects on the Photoreactions of Transition Metal Complexes Tab. 6.3. AVr data for the bimolecular addition o f various ligands t o model heme complexes in aqueous buffer (data from ref. 1881).

Model

Ligand

AV~:~(crn'mol-')

k., ( M - ' s - ' )

MCPH MCPH PHDME PHDME PHDME

co

-19.3 f 0.4 -11.3 f 1.0 11.6 2 0.8 9.9 f 1.0 10.9 k 3.1

(1.08 f 0.2) x 107 (1.04 2 0.04) x lo8 (3.93 +_ 0.04) x lo8 (2.46 +_ 0.01) x lo8 (1.48 k 0.04) x l o 8

0 2

MeNC 'BuNC 1 Melm

According to this model, the rate for formation of (Por)Fe"-L from solvent separated (Por)Fe" and L will have the form d[(Por)FeL] = k,,[( Por)Fe"] [L] dt

(6.42)

where k,, = k,kD/(k, + k-D) if thermal dissociation of the Fe-L bond (k,,ff) is slow. This model has two limiting cases. When k, >> k-D then k,, = kD, that is, the reaction is diffusion limited. Alternatively, when k-D >> k,, then k,, = k,kD/k-" and the reaction is activation limited. As noted in the previous sections, pressure effects in the first case should show rate slowing due to increasing solvent viscosity as P is increased (a positive AV:,). The second case is more complicated, since pressure may affect both k, and the equilibrium constant (kD/k-D) for precursor complex formation. The pathways for the recombination of L with (Por)Fe" were probed using pressure effects on the kinetics of the model heme systems protoheme dimethyl ester (PHDME) and mono-chelated protoheme (MCPH) (Fig. 6.16) [88]. Values of AV:, for these systems range from -19 to +12 cm3 molk' with different neutral ligands in toluene solution (Table 6.3). The change from a large negative AVi, to a positive value was interpreted [ 881 in terms of a change in the rate-limiting step of the "contact pair" mechanism (Eq. 6.41). For L = CO and perhaps L = 02,the reaction appears to be activation limited, that is: the rate-limiting step would be formation of the Fe"-L bond (k,). In this case AV:, = AV: AVL - AVLD = AV: since the volume difference between the contact pair and two neutral species in solution should be relatively small. The AV: is negative given that bond formation in this step is also accompanied by a high to low spin state change. The positive AV:, values for L = RNC or I-MeIm were attributed to these faster reactions approaching the diffusion limit (kD), consistent with the strong pressure dependence of toluene viscosity. In a subsequent study (891, the reaction of CO with MCPH was studied as a function of pressure in a very viscous medium, 90/10 (v/v) mineral oil/toluene solution. Figure 6.17 shows a plot of log(k,,/k:n) vs. P for this system, which clearly shows the crossover in rate-determining step from bond formation (k,) at low pressures where AV:,, -10 cm3 mol-' to a diffusion-controlled process (AVL +8

+

-

-

y &

6.4 Photochemically Generated Reactive Intermediates

1 '

CO,CH,

0

H3C02C

Fig. 6.16. Structures o f the iron(l1) hemes complexes protoheme dimethyl ester (PDME) and monochelated protoheme ( M C P H ) .

cm3 mol-') on increasing the pressure to 300 MPa. These results demonstrate how the different steps in Eq. (6.41) can become rate-limiting as a function of the conditions. A model similar to Eq. (6.41) can be applied to the reactions of the diatomic ligands CO and NO with (TPPS)Fe" generated by flash photolysis of the respective complexes in aqueous solution (H2TPPS = tetra(4-sulfonatopheny1)porphine). For L = CO, the k,, value at 25 "C and ambient pressure was measured to be 3.6 x lo7 M-'s-'. The activation parameters were determined as AH& = 11 f 0.6 kJ mol-', AS:, = -64 _+ 2 J mol-' K-' and AVtn = -6.6 _+ 0.6 cm3 mol-' consistent with Fe-CO bond formation being rate limiting [go]. In contrast, the much faster reaction of (TPPS)Fe" with NO (ken = 1.5 x 10' M - l s-l ) gave activation parameters AH:,, = 24 3 kJ mol- ASin = 12 f 10 J mol-l K-' and AV:, = 5 1

-20004 0

'Oo0

P(atm)

2000

Fig. 6.17. Plot of RT In(k,,/k:") vs P for the reaction of CO with the intermediate formed by nanosecond laser flash photolysis of MCPH-CO, at low pressures (AVin -10 c m 3 mol-') and at high pressures (AV:" +8 c m 3 mol-').

- -

3000

216

I

G Pressure Ejfects on the Photoreactions of Transition Metal Complexes

cm3 mol-' indicating that this reaction is dominated by diffusional processes [90]. This is due to a much more efficient reaction of the ((Por)Fe", NO} contact pair to form the Fe-L bond as has been noted previously in flash photolysis experiments and attributed to spin orbit coupling effects [91]. The situation is considerably different for the iron( 111) analog (TPPS)Fe"l(H20);, which is hexacoordinate in aqueous solution. In this case the reaction with NO is considerably slower (kon= 4.5 x lo5 M-' s-' ) and k o K (500 s-') is sufficiently large to be measurable by the flash photolysis method as well [92]. Temperature and hydrostatic pressure effects were probed and AH$, AS$ and AVI values of 69 f 3 kJ mol-', 95 f 10 J mol-l K-' and 1 9 1 cm3 mol-' were determined for the "on" reaction and 76 f 6 kJ mol-', 60 11 J mol-' K-' and 18 f 2 cm3 mol-I, respectively, for the "off" reaction [93]. The activation parameters for the "on" reaction compare very favorably with those measured for exchange between coordinated and solvent water for aqueous solutions of (TPPS)Fe"'( H2O)z [94] and indicate that kinetics for the reaction of NO with this complex are dominated by the lability of the coordinated water. Furthermore, the large and positive values of AVin and AS:,, point to a dissociative substitution mechanism as described in Eq. (6.43).

Fe1"(Por)(H20)2+ N O

ki

$ Fe"'(Por)(H20)

+ HzO

k-1

Fe"'(Por)(H20) + N O

kz + Fe"'(Por)(HzO)(NO)

(6.43)

k-2

Hemoproteins: In the case of hemoproteins, a second geminate pair, a longer lived "separated pair", has been introduced to the model to account for configurations where L is separated from the metal center but is still somehow encaged in the protein [95].

(Por)Fe"-L

{(Por)Fe", L}

+ {(Por)Fe"llL} $ (Por)Fe" + L

(6.44)

Flash photolysis techniques were applied to study the effect of pressure on the bimolecular association rate constant for the reaction of sperm whale rnyoglobin (Mb) with different neutral ligands in aqueous solution. [88]. The results for four L are summarized in Table 6.4. Of these only the reaction with CO is characterized by a negative AVin value, in line with the bond formation process being rate limiting. The positive AV:, data for L = 0 2 , MeNC and 'BuNC were ascribed to the ratedetermining step being the ligand entering the protein pocket, accompanied by significant desolvation and presumably protein conformational changes. The effect of pressure on the escape of the ligand from the protein-separated pair, i.e. [Mb((L]+ Mb L, as measured by ultrafast laser flash photolysis gave significantly positive AV: (+11.7 f 0.1, +12.6 f 1.7 and +9.1 & 3.5 cm3 mol-' for L = CO, 0 2 and MeNC, respectively) [88]. Thus activation of the "gate" for passing these li-

+

6 4 Photochemically Generated Reactive lnterrnediates

Volumes of activation for the bimolecular addition of various ligands to deoxy myoglobin in aqueous buffer (data from ref (881). Tab. 6.4.

Ligand

A V (cm3 ~ mo/-'j

k,,

co

-10 f 0.8 5.2 i 0.5 7.8 k 1.3 8.8 k 1.0 9.3 k 0.3

(5.8 k 1.8) x (2.5 k 0.2) x (1.3 0.3) x (1.4 0.1) x (2.1 f 0.1)x

0 2

MeNC 'BuNC

(M-' s-')

lo5 107

lo7 105 103

gands between the protein pocket and the bulk solution has a positive AVt in both directions. Pressure effects on the reaction of NO with metmyoglobin (Eq. 6.45) has also been studied by laser flash photolysis and by high pressure stopped flow reactions ~961. hv /

H2O t metMb(N0)

16n

metMb(Hz0)

16.

+ NO

(6.45)

The activation parameters for k,, determined in this manner are AHA,, = 63 k 2 kJ mol-', AS!, = 55 k 8 J mol-I K-' and AV!, = 21 f 1 cm3 mol-l. Stopped-flow experiments with NO trapping techniques were found to be more reliable for the measurement of activation parameters for k,E, which were AH:E = 75 5 kJ mol- ', AS:, = 36 17 J mol-' K-' and AViE = 16 f 1 cm3 mol-'. As with the (TPPS)Fe"'(H20)z model discussed above, the mechanism of the NO reaction with metMb(H20) was concluded to be largely dissociative, for example, Eq. (6.46) [961. ki

metMb(H20) e metMbS ki

metMb5

+ H20

+ NO k-zS metMb(N0) k7

(6.46)

where metMbS is the five-coordinateintermediate formed by H20 dissociation. Figure 6.18 illustrates the volume profiles for the reactions of NO with the metMb(H20) and for the analogous reaction with (TPPS)Fe["(H z O ) ~Dissociation . of a coordinated water molecule from an octahedral metal center is expected to be accompanied by a maximum volume increase of 13 cm3 mol-l [97]. The larger value of AVin reported here for metMb suggests that the protein may undergo some structural rearrangement during the formation of the five-coordinatemetMb5.

I

217

218

I

6 Pressure Effects on the Photoreactions of Transition Metal Complexes

[metMb5 + H20 + NO]

....

.....

+21

*

*1

+16+ 1

-

.metMb(N0) + H$

metMb(HP)

+

NO

[Fe'(TPPS)(bO)

+

t+O

+

NO]

*

.......

+I8 f 1

+8 f 1

Fd"(TPPS)(bO)2+ NO

-10f2

Volume profiles for the reactions of NO with the metMb(Hz0) (upper plot) and for the analogous reaction with (TPPS) Fe ''I (HzO), (lower plot).

Fig. 6.18.

6.5

Summary

The goal of this chapter was to demonstrate the types of effects pressure may have on solution phase photochemical reactions of metal complexes. The overall photochemical reaction can be accelerated or decelerated by hydrostatic pressure, which results from a combination of the effect of pressure on the photophysical and photochemical processes involved. Thus the pressure variable adds a further dimension to the investigation of photochemical processes and assists the clarification of intimate reaction mechanisms. Examples described here demonstrate how the lifetime of excited states and even the relative populations of electronic states

References 1219

can be tuned by pressure. Furthermore, photochemical bond formation and bond cleavage processes are accelerated or decelerated by pressure, respectively, in a similar way as found for the corresponding thermal reactions. As a result, the associative or dissociative nature of such substitution reactions can be characterized. Further characterization can also be obtained for the natures of energy and electron transfer reactions of electronic excited states by examining pressure effects. The same applies to photo-induced thermal reactions, where the interpretation of the pressure dependence is less complicated by photophysical relaxation processes. However, interpretation of pressure effects on the photochemistry of metal complexes in solution is in some cases limited by information on the partial molar volume of excited state species and (as always) the difficulty of separating the contributions of intrinsic and solvational volume. Other difficulties are the fragmentary nature of much of the published pressure data and the need for a better understanding of the effect of pressure on the rates of photophysical processes. Nonetheless, hydrostatic pressure as a physical variable has provided valuable insight into the reaction pathways taken by electronic excited states formed by the photolysis of metal complexes in solution.

Acknowledgments

Pressure studies in the University of California: Santa Barbara laboratories of PCF have long been supported by the U.S. National Science Foundation.

References D. R. STEPHENS, H. G. DRICKAMER, J. Chem. Phys. 1961, 34,937. 2 R. W. PARSONS,H. G. DRICKAMER,]. Chem. Phys. 1958, 29,930. 3 S. H. LEE,W. L. WALE, D. R. DEMMER,R. T. WALTHERS, Inorg. Chem. 1985, 24, 1531. 4 G. B. PORTER,H. L. SCHLAFER, Z. Phys. Chem. 1963, 37, 109. 5 J. W. KENNEY, 111, J. W. CLYMIRE, S. F. AGNEW,]. Am. Chem. SOC.1995, 117, 1645. 6 H. W. OFFEN,in Organic Molecular Photophysics, Volume I, J. B. BIRKS (Ed.), Wiley, New York, 1973. 7 I. G. POLITIS,H. G . DRICKAMER, J. Chem. Phys. 1982, 76, 285. 8 H. YERSIN,E. GALLHUBER, Inorg. Chem. 1984, 23, 3745. 9 M. L. FETTEROLF, H. W. OFFEN,]. Phys. Chem. 1985, 89, 3220. 1

10

11

12

13 14

15

16

J. M. LANG,2. A. DREGER,H. G. DRICKAMER, 1.Phys. Chem. 1993, 97, 2289. M. WRIGHTON,D. L. MORSE,J.Am. Chem. SOC.1974, 96. 998. (a) K. R. KYLE:C. K. Ruu, J. A. P. C. FORD,J. Am. DIBENEDETTO, Chem. SOC.1991, 113, 2954; (b) M. VITALE,W. E. PALKE,P. C. FORD,J. Phys. Chem. 1992, 96, 8329. D. TRAN, J. L. BOURASSA, P. C. FORD, Inorg. Chem. 1997, 36, 439. C. M. EDWARDS, I. S. BUTLER,Coord. Chem. Rev. 2000, 199, 1-53. (a) 0. A. SALMAN,H. G. DRICKAMER, J. Chem. Phys. 1982, 77, 3337; (b) j . M. LANG,2. A. DREGER,H. G. DRICKAMER, Chem. Phys. Let. 1992, 192. 299. S. I. STRICKLER, R. A. BERG,^. Chem. Phys. 1962, 37, 814.

220

I

6 Pressure Effectson the Photoreactions of Transition Metal Complexes 17

18 19

20

21

22

23

24 25 26 27 28 29 30

31 32

(a) R. ENGLMAN.1. JORTNER, MOl. Phys. 1970, 18. 145; (b) K. FREED, J. J . Chem. Phys. 1970, 52, JORTNER, 6272; (C) G . w . ROBINSON, R. p. RROSCH,I.Chem. Phys. 1963. 38, W. SIEBRAND. 1187; (d) B. R. HENRY, in Organic Molecular Photophysics, Volume I, J. B. BIRKS(Ed.), Wiley. New York, 1973. J. V. CASPAR, T. J. MEYER. J . Phys. Chem. 1983, 87. 952. (a) S. F. RICE,H. B GRAY,^. Am. Chem. SOC.1983, 105,4571; (b) J. G. BRUMMER, G. A. CROSBY, Chem. Phys. Let. 1984. 112, 15. J. R. WINKLER, J. L. MARSHALL, T. L. NETZEL,H. B. GRAY, J . Am. Chem. Soc. 1986, 108, 2263. C. M. CHE,L. G.BUTLER, H. B. GRAY, R. M. CROOKS, W. H. WOODRUFF, J. Am. Chem. SOC. 1983, 105, 5492. M. FEITEROLF, A. E. FRIEDMAN, Y.-Y. YANG, H. OFFEN,P. C. FORD:J . Phys. Chem. 1988, 92, 3760. N. S. ISAACS, Liquid Phase High Pressure Chemistry, J. WILEYSr SONS, NY, 1981. A. D. KIRK, G. B. PORTER, J . Phys. Chem. 1980, 84, 2998. M. L. FElTEROLF, H. W. OFFEN, I. Phys. Chem. 1986, 90, 1828. M. L. FETTEROLF, H. W. OFFEN,J . Phys. Chem. 1988, 92, 3437. J. VANHOUTEN, R. I. WAITS,J . Amer. Chem. SOC.1976, 98,4853. K. S . PETERS, G. J . SNYDER, Science 1988, 241, 1053. J. L. GOODMAN, M. S. HERMAN, Chem. Phys. Lett. 1989, 163, 417. C. C. BORSARELLI, S. E. BRASLAVSKY,]. Phys. Chem. B 1998, 102, 6231-6238. (b) C. C. BORSARELLI, M. T. INDELLI, F. SCANDOLA, S . E. BRASIAVSKY,Chem. Phys. Lett. 2000, 317, 53-58. (c) C. C. S. E. BRASLAVSKY, J. BORSARELLI, Photochem. Photobio B 2000,43, 222. J. D I B E N E D E ~R.O J., WATTS,P. C. FORD,Inorg. Chem. 1984, 23, 3039. (a) S. DECURTINS, P. GUTLICH, H. KOHLER, H. SPIERING, A. HAUSER, Chern. Phys. Lett. 1984, 105, 1; (b) P. GUTLICH,A. HAUSER, H. SPIERING, Angew. Chem. Int. Ed. Engl. 1994, 33, 2024.

J. J. MCGARVEY, 1. LAWHERS. K. HEREMANS, H. TOFTLUND? Inorg. Chem. 1990, 29, 252. 34 J. DIBENEDEITO, V. ARKLE,11. A. P. C. FORD, Inorg. Chem. GOODWIN, 1985, 24, 455. 35 V. AMIR-EBRAHIMI, J. MCGARVEY, Inorg. Chim. Acta 1984, 89, L39. 36 P. c . FORD,D. WINK,J. DIBENEDETTO, Prog. Inorg. Chem. 1983, 30, 213. 37 P. C. FORD,Coord. Chem. Rev. 1982, 44, 61. 38 P. C. FORD, J. Chem. Ed. 1983, 60. 829. 39 1. DIBENEDETTO, PhD Dissertation, U . California, Santa Barbara, 1985. 40 W. WBBER, R. VAN ELDIK, H. KELM,J. DIBENEDETTO, Y. DUCOMMUN, H. OFFEN,P. C. FORD?Inorg. Chem. 1983, 22, 623 41 W. WEBER, J. DIBENEDETTO, H. OFFEN,R. VAN ELDIK, P. C. FORD, Inorg. Chem. 1984, 23, 2033. 42 J. DIBENEDETTO, P. c. FORD,Coord. Chem. Rev. 1985, 64, 361. 43 W. WEBER, R. VAN ELDIK, Inorg. Chim. Acta 1986, 11I, 129. 44 L. H. SKIBSTED, W. WEBER:R. VAN ELDIK,H. KELM, P. C. FORD.Inorg. Chem. 1983, 22, 541. 45 P. C. FORD,J. DIBENEDETTO, in Photochemistry, Photophysics of Coordination Compounds,H. YERSIN, A. VOGLER(Eds), Springer-Verlag, Berlin, 1987, pp. 295-300. 46 T. L. KELLY,J. F. ENDICOIT, J . Phys. Chem. 1972, 76, 1937. 47 M. A. BERGKAMP, J. BRANNON, D. MAGDE,R. j. WATTS,P. C. FORD,J . Amer. Chem. SOC.1979, 101, 4549. 48 M. A. BERGKAMP, R. J. WAITS,P. C. FORD,J . Amer. Chem. Soc. 1980, 102, 2627. 49 L. J. MCCLURE, P. C. FORD,]. Phys. Chem. 1992, 96, 6640. 50 K. ANGERMANN; R. VAN ELDIK,H. KELM,F. WASGESTIAN, Inorg. Chem. 1981, 20, 955. 51 K. ANGERMANN, R. VAN ELDIK, H. KELM,F. WASGESTIAN, Inorg. Chim. Acta 1981, 49, 247. 52 K. ANGERMANN, R. SCHMIDT, R. VAN ELDIK, H. KELM,F. WASGESTIAN, Inorg. Chem. 1982, 21, 1175. 33

References 1221 W. L. WALTZ.J . LILIE,S. H. LEE, Inorg. Chem. 1984, 23, 1768. 54 D. A. FRII:SEN.S. H. LEE,J. LILIE, W. L. WAITZ.L. VINCZE,Inorg. Chem. 1991, 30, 1975. 55 L. VINCZE,D. A. FRIESEN, S. P. MERZYK, W. L. WALTZ,Inorg. Chem. 1992, 31,4950. 56 D. A. FRIESEN, S. H. LEE, R. E. NASHIEM,S. P. MERZYK.W. L. WALTZ, Inorg. Chem. 1995, 34. 4026. 57 (a) A. D. KIRK. G. B. PORTER,M. A. RAMPI-SCANDOLA, Inorg. Chim. Acta 1984, 90, 161; (b) A. D. KIRK,Coord. Chem. Rev. 1981, 39, 225. 58 R. B. WILSON,E. I. SOLOMON, Inorg. Chem. 1978, 17, 1729. 59 E. ZINATO,in Concepts i n Inorganic Chemistry, Chapter 3, A. ADAMSON, P. FLEISCHAUER (Eds), Wiley, New York, 1975. 60 J. F. ENDICOTT,C. KUL RYU,Comments Inorg. Chem. 1987, 6, 91. 61 s. WIELAND, R. VAN ELDIK,]. Dhys. Chem. 1990, 94, 5865. 62 s. WIELAND, R. VAN ELDIK,D. R. CRANE,P. C. FORD,Inorg. Chem. 1989, 28, 3663. 63 D. M. MANUTA,A. J. LEES,Inorg. Chem. 1986, 25, 1354. 64 H. K. VAN DIJK, P. C. SERVAAS, D. J . STUFKENS, A. OSKAM,Inorg. Chim. Acta 1985, 104, 179. 65 S. WIELAND, R. VAN ELDIK,]. Chem. Soc., Chem. Commun. 1989, 367. 66 S. WIELAND,K. BALREDDY:R. VAN ELDIK, Organometallics 1990, 9, 1802. 67 W.-F. F u , R. VAN EIDIK,Inorg. Chim. Acta 1996, 251, 341. 68 W:F. Fu, R. VAN ELDIK,Organometallics 1997, 16, 572. 69 W.-F. Fu, R. VAN ELDIK,Inorg. Chem. 1997, 37. 1044-1050. 70 J. VAN HOUTEN,R. J. WATTS, Inorg. Chem. 1978, 17, 3381. 71 B. DURHAM,J. CASPAR,J. NAGEL,T. J. MEYER,]. Amer. Chem. SOC. 1982. 104, 4803. 72 M. L. FETTEROLF, H . W. OFFEN,Inorg. Chem. 1987, 26, 1070. 73 A. D. KIRK,C. NAMASIVAYAM, G. B. A. PORTER,M. A. RAMPI-SCANDOLA, SIMMONS, ]. Phys. Chem. 1983, 87, 3108. 53

74

75

76

77 78

79 80

81

82 83

84

85

86

87

88

89

90

91

G. STOCHEL, R. VAN ELDIK? 2. STASICKA.Inorg. Chem. 1986, 25, 3663. C. C. BORSARELLI, S. E. BRASLAVSKY, 1.Phys. Chem. A 1999, 103, 17191727. J. B. BIRKS,Photophysics ofAromatic Molecules, Wiley-Interscience. London, 1970. W. D. TURLEY, H. W. OFFEN,]. Phys. Chem. 1984, 88, 3605. R. E. GAMACHE JR., R. A. RADER, D. R. MCMILLIN, ]. Am. Chem. SOC.1985, 107, 1141. A. M. FATTA,R. L. LINTVEDT, Inorg. Chem. 1971, 10, 478. (a) D. R. CRANE? P. C. FORD,J . Am. Chem. SOC.1991, 113, 8510; (b) D. R. CRANE,P. C. FORD.Inorg. Chem. 1993, 32, 2391. A. D0SSING, S. KUDO,C. K. RYU, P. C. FORD,]. Am. Chem. Soc. 1993, 115, 5132. D. R. CRANE,P. C. FORD,]. Am. Chem. Soc. 1990, 112, 6871. D. MILLER,P. MILLER,N. A. P. KANEMCGUIRE,Inorg. Chem. 1983, 22, 3833. M. E. FRINK,D. MAGDE,D. SEXION, P. C. FORD,Inorg. Chem. 1984, 23, 1238. D. R. MCMILLIN,J. R. KIRCHHOFF, K. V. GOODWIN,Coord. Chem. Rev. 1985, 64, 83; Inorg. Chem. 1987, 26, 875. D. R. CRANE,J. DIBENEDETTO, C. E. A. D. R. MCMILLIN,P. C. FORD, PALMER, Inorg. Chem. 1988, 27, 3698. T. G. TRAYLOR, D. MAGDE,D. J. TAUBE, K. JONGEWARD, ]. Am. Chem. SOC. 1987, 109, 5864. D. J. TAUBE,H.-D. PROJAHN,R. VAN ELDIK.D. MAGDE,T. G. TRAYLOR, ]. Am. Chem. Soc. 1990, 112, 6880. T. G. TRAYLOR, J. Luo, J. A. SIMON, P. C. FORD,I . Am. Chem. Soc. 1992, 114, 4340. (a) L. E. LAVERMAN, P. C. FORD,Chem. Comm. 1999, 1843-1844 (b) L. E. LAVERMAN, P. C. FORD,]. Am. Chem. SOC.2001, 123, 11614-11622. P. A. CORNELIUS, R. M. HOCHSTRASSER, A. W. J. STEELE, ]. Mol. Biol. 1983, 163, 119-128.

222

I

6 Pressure Effects on the Photoreactions of Transition Metal Complexes 92 M. H O S I ~ I NK.O OZAWA, . H . S E K IP. , C. FORD,J . Am. Chem. SOL.1993, 115. 9568-9575. 93 L. E. LAVERMAN. M. H O S H I N OP.. C . FORD. J . Am. Chem. SOC.1997, 119. 12663-12664. 94 (a) I. J. OSTRICH, G . Lru, L. GORDON, H. W. DODGEN, J. P. HUNT,Inorg. Chem. 1980, 19, 619-621; (b) T. SCHNEPPENSIEPER, A. ZAHI.,R. V A N EI.DIK,et al., Angm. Chem. [nt. Ed. 2001, 40, 1678.

95 E. J. ROSE, R. M. H O F F M A NAm. ,~. Chem. SOL 1983, 105, 2866-2873. 96 L. E. LAVERMAN, A. WANAT,1. OSZAJCA. G . STOCIIEL, P. C. FORD,R. VAN ELDIK, J . Am. Chem. SOC. 2001, 123, 285. 97 (a) A. DRLJACA, C. D. HUBBARD, R. VAN ELDIK. R. ASANO, M. V. BASILEWSKY. W. J. LE NOBLE. Chem. Rev. 1998, 98, 2167; (b) C. STOCHEL, R. VAN ELDIK,Coord. Chem. Ref:. 1999, 187, 329.

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

7

Application of High Pressure in Transition Metal-Catalyzed Reactions Oliver Reiser 7.1

Introduction

One of the most challenging tasks in chemistry is to design efficient processes for synthetic transformations of organic molecules. Transition metal-catalyzed reactions play an especially important role from an economic and ecological point of view, which is reflected in the high research activity in this field. Besides finding new catalpc reactions that also fulfill the demand for atom economy [ 11, the optimization of already known processes is necessary to allow their application on a technical scale. Efficiency in a given catal$c process can be defined both in terms of: (1) reactivity i.e. turnover numbers (TON) and turnover frequencies (TOF = TON/time) and (2) selectivity, i.e. regioselectivity, diastereo-/enantioselectivity and chemoselectivity. Both aspects are directly dependent on the catalyst performance. Consequently, most research efforts are directed towards the development of new catalysts, and aspects of selectivity are usually the center of attention. By rationally designing ligands that create a highly defined steric environment for a catalytically active metal, spectacular results have been achieved to reach high selectivities for almost any important chemical transformation. However, many of these catalpc reactions, despite their ability to carry out a desired transformation, cannot be considered efficient due to their long reaction times and high catalyst loading. The latter point in particular, poses a considerable problem due to the high molecular weights of many catalysts; a catalyst loading of 10 mol % might be the equivalent of 100 weight % or more. The efficiency of a catalyst is generally viewed to be high if the amounts required are as little as 0.5-1 mol %, however, these values are still too high for large scale production. In theory, catalysts are indestructible vehicles with no limits to their efficiency. However, side reactions and/or catalyst poisoning in the course of a reaction causes their more or less rapid breakdown in praxis. Increasing the stability of a catalyst by designing strong coordinating ligands is in general no solution to the problem, since the catalytic activity is drastically reduced during this procedure. Next to designing new catalysts, the development of new reaction conditions is

224

I equally important for the optimization

7 Application of High Pressure in Transition Metal-Catalyzed Reactions

of catalytic processes. Besides parameters considered routinely such as reaction temperature or choice of solvent, emerging new techniques such as activation by ultrasound, microwaves or high pressure have also been considered in catalysis, although not very frequently. This chapter evaluates recent applications of the latter technique for catalysis, in relation to both selectivity as well as reactivity. 7.1.1 General Principles

The general principles under which high pressure acts favorably in organic reactions have already been described in Chapter 2. The combination of metal catalysts and high pressure has only recently been studied in a broader context. Due to the mechanistic complexity of such reactions it is difficult to predict the net effect of pressure in general terms. With the rapidly growing number of new catalytic systems being developed, it is most important to clearly evaluate the role of pressure for a given reaction system in comparison to the best methods otherwise available. Since the ability to exchange ligands at a metal center is pivotal in every catalytic process, ligand exchange is also the most important aspect to be considered when applying high pressure in transition metal-catalyzed reactions. In order to maintain catalytic activity in a reaction process, coordinatively unsaturated species must be usually generated to allow ligand exchange at the metal center. If pressure shifts the metal-ligand equilibrium towards coordinatively saturated, e.g. an 18-electron species, as would generally be expected for such addition reactions: the catalytic activity of such metal complexes will be reduced. On the other hand, ligand exchange might take place by an associative mechanism (cf. Chapter 4), and it has been demonstrated for many examples, that pressure accelerates this reaction pathway [ 2 ] .

7.2

Lewis Acid Catalyzed Reactions 7.2.1 Cycloadditions

Since Diels-Alder reactions are activated by pressure and by Lewis acids, the combination of both has been applied in transformations that were particularly difficult to achieve otherwise. There has been no conclusive evidence about how pressure affects the reactivity and selectivity in such reactions. One can assume that the formation of Lewis acid/substrate complexes are favored by pressure, although in most cases under normal pressure the equilibrium already lies strongly on the side of such complexes. Consequently, it appears reasonable to assume that the reacting species are the same, both under ambient and high pressure (cf. Chapter 11) [ 31.

7.2 Lewis Acid Catalyzed Reactions

A detailed comparison of normal and high pressure was made in the study of the reaction of cyclopentadiene (1) and the acrylate 2 in the presence of various Lewis acids (Scheme 7.1) 141. In general, the reactions were faster and the yield was increased using pressure, while the effects on selectivity were rather small. The endolexo ratio remained almost unchanged; however, it is interesting to note that the diastereoselectivity for the formation of 5 slightly increased using zinc chloride as the catalyst.

0 1

+

E

AcO

-Table 1

/=(

3

J

4

2

A O E c E (E = CO+menthyl)

Pressure Catalyst

I

E

5

Time

Yield

Ihl

["/.I

1J

exo-diastereorners

6

ztd

de [%I ~

endo ex0 [kbar]

Scheme 7.1. Influence of pressure and Lewis acid catalysis in the Diels-Alder reaction o f 1 with 2.

The activation of the Diels-Alder reaction between 7 and 8, using both pressure and Lewis acid catalysis, was recently investigated (Scheme 7.2) [5]. Such multiactivation can be beneficial since the reaction temperature and time can be reduced, as clearly demonstrated with the reaction of 8b. However, because of competing side reactions such as polymerization, which were apparently more severe when pressure is applied and in the presence of zinc chloride, the thermal process might still be advantageous. The hetero-Diels-Alder reaction between trans-l-methoxy-l,3-diene (11) with various aldehydes and ketones 12 can be achieved at 10 kbar if Eu(fod)3 is used as the catalyst (Scheme 7.3) [GI. In the absence of a Lewis acid, catalyst pressures be-

I

225

226

I

7 Application of High Pressure in Transition Metal-Catalyzed Reactions

a

7

9

10

a: R', R2, R5 = H,R3, R4 = CH3 b: R', R2, R5 = H, R3, R4 = -(CH2)2C: R2, R3, R5 = H, R' = OCH3, R4 = OSiMe3 ~~~~~~~~~~~~~

diene

temp ('C)

reaction time (h)

Pressure War)

Lewis acid

%

conversion

8a

200

216

85

71

8a

50

46

16

-

93

46

16

ZnClp

96

46

25

25

80:20

86

50

96:4

>98:2

8a

25

24

8b

195

72

8b

50

48

16

-

8b

25

48

16

ZnCI2

100

62

-

8C

195

72

100

57a

7525

8C

45

96

12

-

100

9Za

80:20

8C

25

24

16

ZnCIP

100

decomp.

a

Isolated in the for of the ketone after hydrolysis of the silyl enol ether Diets-Alder reaction of indole 7 with various dienes 8.

Scheme 7.2.

E ~ ( f o d(1 ) ~mol%) \\'

I

II

10 kbar. 50'C, 20 h

U

OMe

11

12-81%

CH3 H H

H H H

F

OMe

13

12

R'

H

yield enddexo 9 + 10 (Yo) (9110) ratio

R' C02CH3 CH2NHC02Bn CH(CH3)0SiMe$3u

-HEk CH3 Ph 2-fury1

Isomer ratio cistrans

1:l 1:l

Yield

["/.I

6:4

a1 50 35

1:l

53

3:7

15 12 17

1:l 4:6

Scheme 7.3. Hetevo-Diets-Alder reaction of 1-rnethoxy-l,3butadiene (11) with ketones and aldehydes.

7.2 Lewis Acid Catalyzed Reactions

tween 15 and 25 kbar were necessary for the reaction to proceed, while stronger Lewis acids such as zinc chloride, boron trifluoride etherate or dialkoxy aluminum chloride immediately polymerize the starting diene 11. The combination of europium catalysts and high pressure (15 kbar) makes amino aldehydes available as heterodienophiles in [4+ 21 cycloadditions (Scheme 7.4) [ 71. The reaction between 11 and N,O-protected abthreoninals 14 and threoninals 17 occurs without racemization, nevertheless, mixtures of all four possible diastereomers were obtained. Subsequent acid-catalyzed isomerization led to the more stable trans-adducts 15 and 16 or 18 and 19.

NHZ

NHZ

1) E ~ ( f o d )15 ~ , kbar

1

OMe

OMe

OMe

14a: R = BOM 14b: R = TBDPS

NHZ

15a 15b

1) E ~ ( f o d )15 ~ , kbar

11

50'C,24h f

1 :2 1:4

Scheme 7.4.

16a 16b

NHZ

NHZ

G F OMe

17a: R = 5OM 17b- R = T5DPS

NHZ

18a 18b

OMe

1:2 1: 8

19a 19b

Hetevo-Diels-Alder reaction with amino aldehydes

A very instructive study was carried out for the intramolecular Diels-Alder reaction of 20, catalyzed by the chiral titanium complex 21, which was systematically investigated in a pressure range of 1 bar to 5 kbar and is described in detail in Chapter 8 (Scheme 8.44)181. Intriguingly, concurrent with a gradual rise in pressure, the enantioselectivity of 22 increased from 4.5 % ee (1 bar) to 20.4 % ee at 5 kbar (Scheme 7.5). Although the rise in selectivity is relatively small, it is noteworthy that the differentiation of two diastereomeric transition states leading to enantiomeric products can be improved by pressure! However, these results could not be generalized for intermolecular [4 21 cycloadditions (Scheme 7.6) [9]. 25 was formed in 38 % ee at normal pressure from isoprene (23) and the oxazolidone 24, while the enantioselectivity decreased at 5 kbar to only 21 % ee. Nevertheless, this study revealed another important factor that has to be taken into account for metal-catalyzed reactions under pressure. It was argued that the decreased selectivity may in part be due to a shift of the chiral catalyst 28 to the achiral catalyst precursor 26 induced by pressure, since this

+

I

227

228

I

7 Application of High Pressure in Transition Metal-Catalyzed Reactions

ee [%I

Pressure [bar]

20

0 1 1000 2000 2800

4.5 10.1 10.7

3600

16.9

5000

20.4

6.6

I (enfj-22

Scheme 7.5

x

-

-

X

23

24

25

PhxoyoHP Ph Ph

Ti(O'Pr)2C12

+

Me

0

Ph Ph

2'PrOH -t

ph~~f~;~lc~2 Me

.,,OH

//

/\

Ph Ph

Ph Ph 26

27

28

Scheme 7.6

causes a decrease in molecularity. Indeed, high pressure 'H-NMR studies showed that the ratio of 26/28 decreases from 3.95 at 1 bar to 2.95 at 5 kbar (cf. Chapter 8). In addition to [4 21 cycloadditions, it was demonstrated that the formation of lactams by [2 2]-cycloaddition of enol ethers and isocyanates will proceed at room temperature if a combination of pressure and catalytic amounts of ZnClz is used (Scheme 7.7) [lo].

+

+

cat. ZnC12, CH3CN

O=C=N-Ph

-*

~

12 kbar, 25'2, 16 h

70-80% 29 Scheme 7.7

30

31

7.2 Lewis Acid Catalyzed Reactions

7.2.2

Nucleophilic Substitution

Nucleophilic substitutions proceeding via SN2 pathways can be activated by pressure, as has been demonstrated in many examples. In particular, the ring opening of epoxides can be initiated by pressure; and also by Lewis acid catalysis. Consequently, combining these two activation modes might lead to an even more effective way to functionalize epoxides, and indeed, this strategy has been successfully applied. The ring opening of epoxides with indole is accelerated by a combination of lanthanide catalysts and pressure, which was exploited in a synthesis of diolmycin A2 (35) (Scheme 7.8) [ll].Thus. reaction of epoxyalcohol 32 and indole (33) at 10 kbar in the presence of ytterbium(II1) triflate and water gave rise to the adduct 34 in 51 % yield. Again, the application of pressure drastically decreased the reaction time. Subsequent debenzylation then provided the desired natural product 35.

Yb(0Tf)B / H20 (5rnol%) HCS" H

32 Scheme 7.8.

10 kbar, 60'C 42 h, 51%

*

&OR

33

H

c

34: R = Bn 35:R=H

Synthesis o f diolmycin A2 (35).

7.2.3

Addition o f Nucleophiles to Carbonyl Compounds

The benefits of pressure or Lewis acid catalysis for the addition of nucleophiles to carbonyl compounds is also well established, e.g. in various aldol processes or allylation reactions. The combination of the two methods, however, has rarely been applied. A very interesting example was reported with the addition of trimethylsilylcyanide to acetophenone (36) in the presence of the chiral titanum catalyst 38 (Scheme 7.9) [12]. The reaction proceeded by activation with pressure not only with considerably improved yields, but also with significantly increased enantioselectivity. The reason for the latter remains unclear, especially in light of the study of the Diels-Alder reaction between 23 and 24 also carried out with a chiral titanium catalyst (vide supra). Unfortunately, the exact preparation of 38 was not reported, which would have allowed a better comparison of these two studies. Also, it is interesting to note that the addition of TMSCN to 36 in the presence of the 3- or 10-fold amounts of 38 proceeded with lower enantioselectivity. This may be an indication that the formation of dimers might occur at higher concentration, thereby altering the catalytic active species.

I

229

230

I

7 Application of High Pressure in Transition Metal-Catalyzed Reactions 1) TMSCN. 35 (1 rnol%)

CH~CIP,18'C. 18h

-

2) 2M HCI

36

1 bar: 23%, 7% ee 8 kbar: 93%, 60% ee

37

3? '>O . 1Pr

38 Scheme 7.9. Asymmetric cyanhydrin synthesis.

7.3 Palladium Catalyzed Reactions

For the Lewis acid-catalyzed reactions discussed in the previous chapter, the catalyst simply plays the role of rendering a substrate more reactive by coordinating to it, while the reactions can also occur - at least in principle - in the absence of the catalyst. Many transition metal-catalyzed reactions are much more complex and generally consist of more than one reaction step, each of which might individually be influenced either positively or negatively by pressure. Consequently, predicting the net effect of pressure in such reactions is difficult, which might be the reason why it is only very recently that pressure has been systematically investigated as a parameter in the area of palladium-catalyzed coupling reactions and [ 3 21 cycloadditions. It has been suggested that intermolecular incorporation, i.e. oxidative addition and complexation of a substrate by a metal should be favored, intramolecular reactions, i.e. insertion, migration and deinsertion reactions should be invariant, and extrusion reactions such as reductive elimination or decomplexation should be disfavored by pressure [ 131. However, decomplexation reactions are in most cases ligand exchange reactions, which can proceed by associative mechanisms, and indeed, there is ample evidence that ligand exchange reactions can be accelerated by pressure [ 21. A rate acceleration in the reaction of iodobenzene (39a) and ethyl acrylate (40) has been observed qualitatively [ 131: while at room temperature under atmospheric pressure no reaction occurs, ethyl cinnamate (41)is obtained in high yield if a pressure of 10 kbar is applied (Scheme 7.10). Surprisingly, diarylated acrylate 42, which arises by a second Heck reaction onto 41,is not formed at normal pressure, while at 10 kbar 42 can be obtained as the sole product. Interestingly, if bromobenzene (39b) is used instead of iodobenzene (39a), the adduct 43 was also formed, which is explained by the addition of 45 to a second molecule of ethyl acrylate and subsequent reductive elimination and double bond isomerization. In this case, high pressure seems to slow down the reductive elimination leading to 41 suf-

+

7.3 Palladium Catalyzed Reactions

ficiently enough to make 45 accessible for further transformations. The "living nature" of similar palladium species is of great utility and has been used in intramolecular cascade cyclizations and polymerizations. Ph-X 39a: X = I

C02Et

39b: X = Br

(Ph3)2PdC12(2 mol%)

Ph+C02Et

+

eCO2Et

Et3N, MeCN

41

40

1 -

Ph-Pd-X

-

11

40

Ph

PhX

39a

45

Marl

[hl

"CI

41

42

43

12

25

0

0

0

39a

10

12

25

62

2

0

39a

10-3

20

90

a0

o

0

39a

10

26

90

54

38

0

39a

10

4

0

76

0

0

39b

10

HP

C02Et d X

Yield [y

Temperature

39b

CO2Et

46

Time

140La]

-HPdX

Ph

Pressure

10.~

Ph 43

-HPdX

H'fk02Et

44

COPEt

42

PdX

40

' i("

PhyCOpEt Ph

4a

90

78

o

42

90

14

41

25 ~

[a] In DMF instead

of MeCN

Palladium catalyzed coupling of 39 and ethyl acrylate (40) under normal and high pressure conditions.

Scheme 7.10.

It was also found that the Heck reaction of certain bromoalkenes such as 47-49 will proceed at 20 "C if the reaction is carried out at 10 kbar (Scheme 7.11) [14]. It is interesting to note that even activated vinyl chloride 49 underwent a coupling reaction at 60 "C, which compares favorably with reaction temperatures of 130 "C and above used to react chloroarenes under normal pressure using the same catalyst. The cross coupling between cyclic alkenes 52 and iodobenzene 39a, which leads to the arylated alkenes 53, 54 and 55 depending on the reaction conditions, has been extensively investigated (Scheme 7.12). In a kinetic study [15] of the reaction between 52a and 39a the rate of reaction was gradually accelerated by increasing the pressure from 1 bar to 8 kbar [ krel (1 bar) = 1; krel (2 kbar) = 4; k,l (4ltbar) =

I

231

232

I

7 Application of High Pressure in Transition Metal-Catalyzed Reactions P P h

Ph

Pd(0Ac)p / PPh3 / Et3N

THF I CHjCN IZO’C

47: X = Br, R = CN 48: X = Br, R = C02Me 49: X = CI, R = C02Me

Alkene Pressure

50: R = CN 51: R = C02Me

Time

Temperature

lkbarl

[hl

[‘Cl

47

103

48

20

47

10

Product Yield

50

0 98

48

20

50

48

20

51

0

10

48

20

51

96

49

10

72

60

51

42

49

103

72

60

51

trace

48 48

[%I

Scheme 7.11. Palladium catalyzed coupling of vinylhalides 4749 with styrene under normal and high pressure conditions.

9; k,, (8 kbar) = 231, which computes to an activation volume of AVf = - 12 cm3 mol-’. This value is indeed in good agreement with activation volumes for ligand exchange reactions of palladium(0) and palladium(11) complexes (cf. Chapter 4)[2]. Moreover, for the osmium-catalyzeddihydroxylation the activation volume was also determined to be AV# = -12 cm3 mol-l, again being in the typical range for ligand exchange [lG]. Although more kinetic data are needed in order to draw definite conclusions, the effect of pressure on transition metal-catalyzed reactions might indeed be centered around its influence on the rate of the ligand exchange processes.

6

Phl (39a)/NEt3

ph

Pd(0Ac)p / PPh3*

52a: X = 0 52b: X = NC02Et 52C: X = CH2

L

53a 53b 53c

54a 54b 54c

T Ph * H

HPdX 56

Scheme 7.12

55c

57

7.3 Palladium Catalyzed Reactions Tab. 7.1.

Entry

Synthesis o f 53-55 by palladium-catalyzed arylation of 52". Pd(0Ac)z

p (kbar)

t (h)

1 (52a) 2 (52a) 3 (52a) 4 (52a) 5 (52a) 6'(52a) 7 (52b) 8 (52c) 9 (52a) 10 (52b) 11 (52c)

10-1 10-2 10-2 5 x 10-3 10-2 lo-' 5 x 10-3 10-3 s x 10-4 5 x 10-4 10-4

Yieldb

TON

TOF (h-')

270 280 10000 19800 9800 7500 17400 100000 146000 134000 770000

22 23 276 165 102 104 145 833 2030 1860 4580

I%)

(mol%)

10-3 10-3 8 8 2 8 8 8 8 8 8

12c 12c 36 120 96 72 120 120 72 72 168

27 2.8 lOOd

99e 98 75' 87 100 73 67 77

"PhI (1 equiv., 2.0 mmol), cycloalken 52 ( 3 equiv., 6.0 mmol), NEt3 ( 3 equiv., 6 mmol). Pd(0Ac)Z / PPh3 1: 2, 3 ml THF / acetonitrile 1 : 1, 60 "C (entries 1-7), 100 "C (entries 8-11). GC yield 53-55 using pentamethylbenzene as internal standard. 'Extended reaction times did not increase the yield. Isolated yield 53a: 71 %. 'Isolated yield 53a: 69 %. Without PPh3.

Another effect of pressure in these coupling reactions is a dramatic increase in the lifetime of the catalyst (Table 7.1), which is reflected in turnover numbers (TON) of up to 770,000 [17]. Moreover, even in the absence of stabilizing ligands the coupling reactions proceeded with considerable higher TON (7500) than can be reached with the catalyst Pd(OAc)2/PPh3 at normal pressure. The phosphine ligand however, played an important role in the regioselectivity of this reaction (Scheme 7.13). While at normal pressure there was little change in the ratio of 54a/53a (95 : 5 with no PPh3; 90: 10 with Pd: PPh3 1 : GO), at 10 kbar 53a became the major product (90: 10 with no PPh3; 25 :75 with Pd: PPh3 1:GO) [18].Apparently, the decomplexation of 56 by an associative substitution with PPh3 is favored by pressure. In agreement with this analysis, an increase in enantioselectivity was also found in the coupling of 2,3-dihydrofuran (52a) and phenyl nonaflat (PhONf), when Pd-BINAP was used as the chiral catalyst. Thus, at 1 bar (R)-54ais formed with 47 % ee, while at 10 kbar a substantially improved selectivity of 89 % ee was observed. Along with the increase in enantioselectivity, again the regioselectivity of the reaction had also changed and the formation of 53a increased considerably at high pressure. From corroboration of the data for regio- and enantioselectivity, it becomes clear that the initial differentiation of the enantiotopic faces of 52a by the chiral palladium complex PhPdL2"ONf is minimally influenced by pressure. The diastereomeric intermediates 58 and 59, however, are efficiently kinetically resolved by applying pressure: 58 undergoes associative ligand displacement more rapidly liberating 53a, while in 59 metal migration to GO takes place.

I

233

234

I

7 Application $High Pressure in Transition Metal-Catalyzed Reactions

fi

PhONf I NEt,

PhQ

+

Ph,,.&

Pd(0Ac)PI (R)-BINAP

52a

(S)-53a 5 (nd) 32 (5% ee)

1 bar 10 kbar

L,J

i.

HPdONf *L' 'L'

L,d

--

HPd,ONf

*Lcp

v

59

56 Y

(R)-54a 95 (47%ee) 68 (89% ee)

YdH NfO

,

60

diastereomers Scheme 7.13

A similar pressure effect on regioselectivity was reported for palladium-catalyzed 21-cycloadditions [ 191. In the reaction of the trimethylenemethane (TMM) precursor 61 with the alkene 62 the two regioisomeric cycloadducts 63 and 64 are possible: while 64 is mainly formed at 1 bar, the only product observed at 10 kbar is 63. A possible explanation of this dramatic change in selectivity could be the increased rate of the bimolecular reaction of 65 and 62 to give 63 compared to the unimolecular isomerization of the TMM complexes 65 and 66. Thus, the kinetically formed complex 65 is effectively trapped under pressure by the alkene 62. The synthesis of isoquinolines by the cyclization of 67 demonstrated once more the advantageous effect that pressure could have on palladium-catalyzed coupling reactions [20]. 68 was obtained in good yield and with good regio- and diastereoselectivity only if pressure was applied to the system. Especially noteworthy is the beneficial effect of pressure on an intramolecular reaction, since the coupling step of 69 to 70 is most likely the rate-determining step. An increase in the packing coefficient leading to a volume contraction caused by a cyclic transition and product structure, respectively, might very well be responsible for the results observed for the reaction of 67 to 68 (Scheme 7.15). With the example of Diels-Alder reactions it was elucidated that the volume contraction resulting from the increase in the packing coefficient during a cyclization is as important for the size of the reaction and activation volumes as that resulting from an intermolecular addition [ 211. The Combination of pressure and catalysis can also be used to design a new domino process. The alkenylation of aldehydes with phosphonates (HornerWadsworth-Emmons (HWE) reaction) is readily accomplished at room temperature under pressure in the presence of triethylamine as a base. These mild con[3

+

7.3 Palladium Catalyzed Reactions

62

61

63

64

1 bar, (Ph3P),+Pd,82%

1

3

10 kbar, Pd(OAc)z, (i-PrO)3P, 71%

only

-

t

t

r

62

65

62

66

Scheme 7.14

Pd(PPh& (10 mol%)

a, N, Boc

-

10 kbar, 60'2, 12 h

/i

67

68

69

70

Scheme 7.15

ditions are amenable to the Heck protocol, and consequently, the reaction of an aldehyde, a phosphonate and a aryliodide in the presence of palladium(0) and triethylamine under pressure leads to trisubstituted alkenes 74 (Scheme 7.16) [22]. Pressure proved to be beneficial not only for the alkenylation step but also for the subsequent arylation via a Heck reaction, since disubstituted alkenes are generally considerably less reactive in such coupling reactions than monosubstituted ) are formed, alkenes. However, in the case of R # Ar, mixtures of ( E ) / ( Z isomers as was also noted in Heck reactions with cinnamic esters under normal pressure conditions [ 23J.

I

235

236

I

7 Application of High Pressure in Transition Metal-Catalyzed Reactions R

RCHO

+

71

Pd(0) I NEtB Arl

(Me0)2(0)PAC02Me

+

72

t

8-10 kbar, 80‘C 74

73

yield

R

wz

Ph

79

pOMe-Ph

73

H

80

27:73

pOMe-Ph

90

70:30

Ph

56

s95:c5

H

Me Scheme 7.16.

[“I/.

Ar

pOMe-Ph

Ar ,+om

Domino Horner-Wadsworth-Emmons Heck reaction.

7.4

Rhodium-Catalyzed Hydroboration

The hydroboration of alkenes is known to be activated either by pressure or catalysis. Consequently, the combination of these techniques might open the way for the hydroboration of particularly unreactive substrates. Maddaluno et ul. recently investigated the hydroboration of some functionalized alkenes, comparing different reagents (catecholborane (CBH) versus pinacolborane (PBH)), and activation by Wilkinson’ catalyst (RhCl[PPh3I3)and pressure [24]. While bromoalkenes and allylamines were found to give the best results with CBH at ambient pressure, 2,3dihydrofuran (52a, Scheme 7.17) was hydroborated most effectively by PBH in the

52a

0.5 0 0.5

76

75

10-3 12.5 12.5

8 48 72

24:41:0:35 250:50:25 61:39:0:0

77

45 46 84 78

Scheme 7.1 7

References 1237

presence of the rhodium catalyst and a pressure of 12.5 kbar. N o reaction took place in the absence of the catalyst at ambient pressure. Pressure alone led to the hydroboration product 75, however, 77 and 78, which arise by opening the furan ring. had also formed to a large extent. Using Willtinson’s catalyst reversed in part the regioselectivity, leading to 76 as the major hydroboration product of 2,3dihydrofuran (52a), but the ring-opening product 78 was still formed as a major byproduct. When pressure and Wilkinson’s catalyst were applied, the ring-opening products could be completely suppressed, and 75 and 76 could be obtained in significantly increased yields.

7.5

Conclusion

The application of pressure in catalysis has proved to be advantageous for a number of processes. Catalyst performance can be improved in this way, leading to higher yields, better turnover numbers and rates, and increased selectivity, demonstrating that ligand exchange on catalytic active species is facile under pressure. Nevertheless, it has also become apparent that pressure can be detrimental to catalytic processes as well, suggesting that ligand exchange can be also blocked by the application of pressure. High pressure is therefore a useful parameter to be considered for catalysis. However, at the current stage of development it is difficult to make general predictions, although some rules have emerged to describe the influence of pressure on transition metal-catalyzed reactions [ 25, 261.

References B. M. TROST,Angew. Chem. lnt. Ed. Engl. 1995, 34, 259-281. 2 R. v. ELDIK,T. ASANO,W. J . LI:NOBLE, Chem. Rev. 1989, 89, 549. 3 G. J E N N E R , New]. Chem. 1997, 21, 1

4

167-170.

1085-1090.

11 H . KOTSUKI, K. TERAGUCHI, N.

N. KATAGIRI,N. WATANABE, C. KANEKO, Chem. Pharm. Bull. 1990, 38,

SHIMOMOTO, M. Octii, Tetrahedron Lett. 1996, 37, 3727-3730. 12 M. C. K. CHOI,S. S. CHAN,K. MATSUMOTO, Tetrahedron Lett. 1997, 38, 6669-6672. 13 T. SUGIHARA, M. TAKEBAYASHI, C. KANEKO, Tetrahedron Lett. 1995,

69-72.

CHATAIGNER, E. HESS,L. TOUPET, S . R. P I E ~ R E Org. , Lett. 2001, 3, 515-

5 1.

518. 6 J. JURCZAK, A. GOLEBIOWSKI, T. BAUER,Synthesis 1985, 928.

A. GOLEBIOWSKI, J. JURCZAK, Tetrahedron 1991, 47, 1037-1044. 8 L. F. TIETZE, C. O n , K. GERKE, M. BUBACK,Angew. Chem. Int. Ed. Engl. 1993, 32, 1485-1486. 7

L. F. TIETZE,C. O n , U. FREY,Liebigs Ann. 1996, 63. 10 R. W. M. ABEN,E. P. LIMBURG, H . W. SCHEEREN, High Press. Res. 1992, 1 I , 9

36,5547-5550.

K. VOIGT,U. SCHICK,F. E. MEYER, D. MEIJERE, S y n h 1994, 189. 1 5 s. HILLERS, 0. R E I S E R , Chem. Commun. 1996, 2197. 16 R. KCKER, M. NICOIAS,B. SCHMIDT, 14

A.

238

I

7 Application of High Pressure in Transition Metal-Catalyzed Reactions

17

18 19

20

21

0. RFISER,J. Chew. SOL, Perkin Trans. 2 1999, 1615. S. HILLERS,S. SARATORI, 0. REISER,J. Am. Chem. SOC. 1996, 118, 2077-2078. S. HILLERS, 0. REISER,Tetrahedron Lett. 1993, 34, 5265. B. M. TROST,J. R. P A R Q U E ~A. E ,I.. MARQUART. 1.Am. Chem. Soc. 1995, 117, 3284-3285. L. F. TIETZE,0. BURKHARDT, M. HENRICH.Liebigs Ann. 1997, 1407-1413. (a) M. K. DIEDRICH,F.-G. K L ~ R N E R , J. Am. Chem. SOC.1998, 120, 62126218; (b) Review: F . 4 . K L ~ R N E R ,

F. WURCHE,]. Prakt. Chem. 2000. 342, 609-636. 22

23

24

25 26

K. RODMANN, S. HAS-BECKER, 0. REISER, Phosphorous, Silicon Sulfur 1999, 144-146. 173-176. M. MORENO-MANAS, M. PEREZ, R. PLEIXATS,Tetrahedron Lett. 1996, 41, 7449-7452. s. COLIN,L. VAYSSE-I.UDOT, J.-P. LECOUVE, J. MADDALUNO, /. Chem. SOC Perkin 12000, 4505-4511. 0. REISER,Reo. High Press. Sci. Technol. 1998, 8, 111-120. 0. REISER,Top. Catalysis 1998, 5, 105112.

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

I 8

High Pressure in Organic Synthesis: Influence on Selectivity Lutz F. Tietze* a n d Peter 1. Steck

8.1 Introduction

The development of selective and efficient synthetic methods is one of the major goals in organic chemistry. On the other hand, these new procedures should also be compatible with our environment and preserve our resources. In the last two decades many new reactions have been established which allow chemo-, regio-, diastereo- and enantioselective transformations. In addition, efficiency has been improved significantly by introducing the domino concept, which enables the preparation of complex molecules in one process starting from simple substrates [l-41. Control of these reactions is achieved by careful selection of the reaction parameters such a solvent, catalyst, reaction temperature and reaction time. A further parameter of increasing importance is high pressure. The application of high pressure in reactions with a large negative volume of activation (AV#) has proven to be especially useful, since it will increase the rate of reaction allowing it to run at a lower temperature; examples include Diels-Alder reactions, 1,3-dipolar cycloadditions, [2 21 cycloadditions, sigmatropic rearrangements and radical polymerizations (see Table 8.1 and Table 2.2 in Chapter 2). A mathematical correlation of the reaction rate and the applied pressure in connection with the volume of activation is given in Table 8.2. Theoretically, a transformation with a AVz of -30 cm3 mol-’ can be accelerated at 1.5 GPa of pressure by a factor of 2.0 x 10Gcompared to the reaction at atmospheric pressure; however, the calculated rates are usually only accurate for pressures up to 0.2 GPa. At higher pressure the influence of increasing viscosity on dynamic effects must be taken into consideration, which would lead to a retardation of any process [5, GI. An extensive discussion of reaction rates in highly viscous solvents can be found in Chapter 3. Even though the reaction rate may in some cases also be increased by adding a Lewis acid, it is often more advantagous to use the milder high pressure conditions since many substrates and products are not only sensitive to higher temperatures but also to Lewis acids. However, several reactions are also known, where a combination of high pressure and a Lewis acid is most successful [7, 81. The reader is referred to Chapter 10 for a detailed discussion of catalytic and solvophobic-promoted high-pressure reactions.

+

239

240

I

8 High Pressure in Organic Synthesis: Influence on Selectivity Tab. 8.1.

Typical A V ’ values of organic reactions.

0 to 13 0 to -20 -5 to -10 -8 to -15 -10 to -25 -25 to -50 -35 to -50

Free radical bond cleavage sN2 reaction Formation of acetals Claisen, Cope rearrangement Free-radical polymerization Diels-Alder reaction [2 + 21-cycloaddition

Tab. 8.2.

Influence o f pressure on the rate of reaction at 25 “C

k(p)/k(O.I MPa) = exp[-AV#/RT(p

- I)]

AAVf (cm’ mol-’) P fMP4

1 3 5

7 10

+70

-10

-20

-30

0.67 0.30 0.13 0.06 0.02

1.5 3.4 7.5 17 56

2.2 11 56 280 3200

3.4 38 420 4800 180,000

One of the first exciting examples of the usefulness of high pressure application is the total synthesis of (*)-cantharidin 4, an ingredient of Spanish Fly, by Dauben et al. using a Diels-Alder reaction of 1 and 2 to give 3 (Scheme 8.1) [9]. A similar approach has been used for the synthesis of (*)-palasonin [lo, 111. Later it was shown that the reaction of 1 and 2 to give 3 can also be catalyzed by LiClO4 at atmospheric pressure; however, this was not possible in thc synthesis of palasonin [121. 0

15kbar,6h 0

85 Yo

V

1

2

RaneyNickel *

quant.

Scheme 8.1.

4 Synthesis of cantharidin.

% LS O

3

8.7 Introduction

I

241

Today high pressure is routinely used especially for sensitive compounds, for example in the total synthesis of reveromycin A described by Shimizu. The sterically-hindered tertiary alcohol 5 was treated with monoallyl succinate G under 1.5 GPa pressure to obtain the ester 7 (Scheme 8.2) [13].

5 DCC, DMAP CH2C12, 1.5 GPa r. t., 24 h, 76%

0

Me 7 Synthesis of reveromycin (DCC = dicyclohexylcarbodiimide, DMAP = N,N-dimethy/-p-aminopyridine).

Scheme 8.2.

Eguchi has shown that nitriles such as 9 can be used as dipolarophiles under high pressure in the reaction with 8 to give the cycloadducts 10 (Scheme 8.3) [14, 151.

$08 Scheme 8.3.

+

NZC-R

9

1 loo'c GPa

- -??Lo

54-90 %

Nk

R

10

1,3 dipolar cycloaddition with nitriles.

Another interesting case illustrating the utility of high pressure, has been presented by Reiser et al. showing that in the Heck reaction of dihydrofuran 11 and iodobenzene 12 to give the products 13-15, the turnover number (TON) of this catalytic process can be improved by stabilizing the catalyst (Scheme 8.4) [ l G ] (cf. Chapter 7).

242

I

8 H i g h Pressure in Organic Synthesis: Influence on Selectivity

" 11

THF/MeCN

60'C

12

13

14

15

Pd(0AC)p

Mol% 10

-'

5.10-4 Scheme 8.4.

Pwal

TON

yield ["A]

0.1

270

27

800

146.000

73

Influence of high pressure on the stability of transition metal catalysts.

Smith et al. have demonstrated the successful application of high pressure in the total syntheses of natural products. In an approach to the synthesis of (+)discondermolide, alcohol 16 was transformed under 1.28 GPa into the phosphonium salt 17 which could then be used to generate a Wittig reagent (Scheme 8.5) ~71.

16

1) PPh3, 12, PhH/Et20 2) PPh3, Pr2NEf, PhHIEtpO 1.28 GPa, 82%

17 Scheme 8.5.

Synthesis of (+)-discodermolide.

High pressure may also be helpful in the isolation of the primary products (19) in consecutive reactions such as in the cycloaddition of y-pyrones (18) which usually lose carbon dioxide to give 20 under more vigorous conditions as shown by Mark6 et al. (Scheme 8.6) (181.

8.2

influence $High

Pressure on Selectivity

I

243

HC-CC02Me

HC-CC02Me

'cMebo heat, -C02

- p r e s s ~ ~ ~ ~ O

Me&

Me

C02Me

70 Yo

18

19

+ regioisomer Scheme 8.6.

Me

C02Me

20

+ regioisomer

Diels-Alder reaction of a-pyrones

In a similar approach, Klarner et al. demonstrated that the cycloadduct 23 of the reaction of dicyanoethine (22) and benzodicyclobutene (21) could only be obtained at higher pressure (1.10 GPa, 53 "C), whereas at ambient pressure and 125 "C the ring-cleavage product 24 was formed exclusively (Scheme 8.7) [ 19, 201 (cf. Chapter 2).

NC-C-C-CN

21 Scheme 8.7.

23

24

CN

The effect of pressure on cycloadditions of DCA t o strained benzene derivatives.

On the other hand, high pressure can also be employed for the improvement of chemo-, regio-, diastereo- and enantioselectivity. In this chapter, the influence of high pressure on stereoselectivity will be the main theme, but some newer examples relating to regio- and chemoselectivity will also be presented. Several excellent books and reviews on the application of high pressure in synthesis have appeared and should be consulted for details of the pioneering work in this field [21, 221.

8.2 influence o f High Pressure on Selectivity 8.2.1 General Remarks

There are two different basic approaches in the control of selectivity in chemical transformations. In a kinetically controlled reaction the difference between the volumes of activation leading to the isomers must be considered, whereas in a thermodynamically controlled reaction the difference in the volumes of reaction is important. Clearly, a difference between the volumes of reaction does not exist for enantiomers, but also in the case of diastereomers it is usually too small to observe a

244

I

8 High Pressure in Organic Synthesis: Influence on Selectivity Tab. 8.3.

increase o f selectivity in chemical reactions under high pressure.

A. Kinetic control 1. Formation of isomers via different mechanisms: e.g. concerted versus radical or ionic cyclization (pressure effect) 2. Favorable difference of activation volumes of the reactions leading to the isomers by the same mechanism (pressure effect) 3. Favorable difference of enthalpy of the reactions leading to the isomers. Improvement of selectivity by performing the reaction at lower temperature (temperabre effect) B. Thermodynamic control 1. Favorable difference of the reaction volume of the products

significant improvement of selectivity by applying high pressure. It should be noted that a distinct decrease in the volume of reaction was found for the cyclization of I-alkenes to give cycloalkanes. This decrease is dependent on the ring size [23]. However, exploitation of this effect to change the selectivity would require a reversible reaction. Thus, the known examples of a pronounced increase in selectivity under high pressure nearly all refer to kinetically controlled reactions. The difference in volume of activation leading to the isomers can be caused either by a pressure-induced change of mechanism or by a difference between the volumes of the transition structures within the same, or at least similar, mechanism. The first case has a simple explanation: If two isomers are produced by different pathways, the compound which is formed through the mechanism that involves the transition structure with the smallest volume is preferred under high pressure. For instance, this applies when in one pathway two bonds are formed in the ratedetermining step, whereas in the other pathway only one bond is created. The second case is more complicated. Again, it can be said that the more compact transition structure is preferred under high pressure. In cycloadditions the endo transition structure (Scheme 8.8) usually has a smaller volume than the exo transition structure and is therefore stabilized under high pressure.

endo Scheme 8.8.

ex0

Transition structures in [4 -121 cycloadditions.

However, a significant effect of high pressure which correlates with a large difference in the activation volumes ( A A V f ) is only found if a pronounced steric interaction exists in the endo-transition structure. On the other hand, if the steric bulk is too high the AAVz decreases again. Therefore only a few examples are so far known in which a useful increase in selectivity is observed when applying high pressure. In several cases the application of high pressure leads at first to a decrease in selectivity. This is observed among others, in cycloadditions where the exo transition structure is preferred at atmospheric pressure. As a rule of thumb a

8.2 Influence $High Pressure on Selectivity Tab. 8.4.

Influence of pressure and AAV' on selectivity.

0.1

100

500

lo00

(cm3 mol-')

Cl/CZ

CI / c z

Cl / c z

Cl/C2

1 5 10

1.oo 1.00 1.00

1.04 1.22 1.50

1.22 2.74 7.52

1.50 7.52 56.G

AAV

synthetically useful difference of activation volume should be at least AAV4 = 3 cm3 mol-l. With increasing AAV+ the effect of high pressure on the selectivity also increases. Thus, a reaction with AAV# = 5 cm3 mol-' will show a selectivity increase from 1:1 at atmospheric pressure to 7.52 : 1 at 1.0 GPa. An important point which is often misinterpreted in high pressure chemistry, is the effect of temperature on selectivity. Thus, in many cases the observed improvement of selectivity under high pressure conditions is not due to a favorable AAV#, but results from carrying out the reaction at a lower temperature. For transformations where a large difference in reaction enthalpy (AAH # ) exists leading to the different isomers, lowering the temperature has a strong effect on selectivity. The pressure and enthalpy effect of the reaction can be cooperative but it can also be opposed. An accurate differentiation of these two influences can only be achieved by a separate determination of the AV# and A H # for the different reaction channels. Finally, in many cases it has been observed that the solvent has a large influence on selectivity in reactions carried out under high pressure. Again, the understanding of these effects is in its infancy, though some powerful calculation models have been introduced which allow the inclusion of solvents into the calculations of ground and transition structures [24]. For some reactions it was shown that the quantity of the molar volume of the substrates is strongly solvent dependent whereas the molar volumes of the transition structures seem to be less affected by the solvent [25]. The unusually large solvent effects observed in these cases is being understood in terms of a difference in solvation of the substrate in the appropriate solvents. Thus, whenever solvent effects in organic reactions are studied at various pressures, it is recommended that activation volumes and partial molar volumes of the substrates and the products are determined in order to locate the transition structure on the absolute volume scale. 8.2.2

Chemo- and Regioselective Transformations

An example for a change of mechanism in a chemoselective reaction by applying high pressure is the transformation of the benzylidene-1,3-dicarbonyl compound

I

245

246

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

25 leading to 26 in an intramolecular hetero-Diels-Alder reaction and to 27 in an intramolecular ene reaction (Scheme 8.9) [ Z G ] . At 110 "C and 100 MPa in dichloromethane a ratio of 26 to 27 of 11 : 1 was observed, whereas at 90 "C and 550 MPa the ratio was found to be 7G.3:1. Thus, higher pressure and lower temperature favors the formation of cycloadduct 26. The AAVZ value amounts to -(10.7 t 1.9) cm3 mol-I and the A A H # value to -(32.4 7.2) kJ mol-I. The large difference between the volumes of activation for the two reaction pathways can be correlated with the intrinsic contribution of AVf for the formation of a covalent bond: in the Diels-Alder reaction two single bonds are formed whereas in the ene reaction only one single bond is produced overall. The AAVf value strongly depends on the solvent and is much lower in acetonitrile [-(4.0 0.7) cm3 mol-'1 and toluene [-(5.4 f 0.5) cm3 mol-'1.

26

25

Pressure [MPa]

Selectivity (26 : 27)

75 100 320 550

19.5: 1 23.5 : 1 40.7: 1 76.3: 1

27

M V t = -(10.7& 1.9)crn3 mol-' M H t = -(32.4f 7.2)kJ mol-' Scheme 8.9. Influence of pressure on the chernoselectivity of the reaction of 25 i n dichloromethane at 90 "C.

A pronounced pressure effect was observed by Jenner et al. for the intermolecular reaction of dimethylacetylene dicarboxylate (29) and cycloheptatriene (28) to give the two products 32 and 34 in a 2 : 1 ratio at atmospheric pressure (Scheme 8.10) [27].It can be assumed that 34 is formed by a 14 + 21 cycloaddition of the valence tautomer 33, whereas 32 originates from an ene reaction of 28 and 29 to give 30 which undergoes a mutual isomerization to the valence tautomer 31 followed by a 3,3-sigmatropic rearrangement. At higher pressure 34 is found to be the only product. Since the Diels-Alder and the ene reaction were assumed to be equally influenced by high pressure, it was proposed that a promotion of the valence-bond isomerization of 28 to 33 by pressure was responsible for the observed selectivity. This assumption is, however, a contradiction to the results pre-

8.2 Influence of High Pressure on Selectivity

sented in Scheme 8.9. In agreement with these results, we suggest that at high pressure the Diels-Alder reaction of 29 with the norcaradiene 33 is favored over the ene reaction of 29 with the cycloheptatriene 28.

E

30

E

32

31

28

E = C02Me Scheme 8.10.

Reaction of cycloheptatriene and dimethylacetylene dicarboxylate.

+

A different influence of pressure on an intramolecular [4 21 cycloaddition and a 1,s-sigmatropic rearrangement is responsible for a pressure-induced increase in selectivity in the thermolysis of (Z)-1,3,8-nonatriene35 to give 36 and 37 as shown by Klamer et al. (Scheme 8.11) [28]. At 0.1 MPa the rearrangement is favored and the products 36 and 37 are formed in a 31 :69 ratio. Applying 770 MPa of pressure, the selectivity is reversed favoring the Diels-Alder product 36 in a ratio of 73 : 27. It can be assumed that 36 is formed via the bicyclic transition structure 38, whereas 37 evolves through the monocyclic transition structure 39 (cf. Chapter 2).

H 35

36 0.1 MPa 770MPa

38 Scheme 8.11.

31 73

37 69 27

39

Thermolysis o f (Z)-1,3,8-nonatriene.

I

247

248

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

A change of mechanism in a regioselective transformation using high pressure was also observed by Trost et al. for an homogeneous transition metal-catalyzed reaction [29]. In the palladium-catalyzed reaction of coumarin (44)with the isomeric ally1 acetates 40 or 41 the main product under standard thermal conditions was the cycloadduct 46,regardless of the type of the starting material; in addition a small amount of 45 was obtained (45:46= 1:lO) (Scheme 8.12). The regioselectivity was also found to be independent of the solvent and the palladium source and had only a slight dependence on the ligand. The application of high pressure increased the regioselectivity of the cycloaddition using 40 as substrate to give the cycloadduct 45 as the favored product, while the rate of the reaction was significantly decreased. The conditions could be optimized using 2.5 mol% (q3-C3HgPdC1)2 and the bidentate phosphite ligand 47, readily available from 2,4-pentanediol and PC13, in toluene/benzene at 1.5 Gpa, to give a mixture of 45 :46 in a 3.8: 1 ratio with 77 % yield. Starting from 41 the same regioisomer as in the thermal reaction was obtained under high pressure but the selectivity was improved for 45 :46 from 1 : 10 to 1 :14.T h e observed rate retardation can be understood by the effect of pressure on ligand dissociation. Since the exchange of a phosphorus ligand at the alkene is required for the reaction, the inhibition of this process at high pressure may make this the rate-determining step. To explain the change in regiochemistry one cannot argue that the different AVO values of the intermediate palladium complexes 42 or 43 are responsible, since in that case the regioselectivity should be the same for the reaction of 40 as well as 41.The simplest explanation for the observed phenomena under high pressure is a pressure-

40

I

42

I 1: 41

I

45

44

43

47 Change of mechanism through high pressure in the reaction o f coumarin 44 with the allylacetates 40 or 41. Scheme 8.12.

46

8.2 Influence of High Pressure on Selectivity

I

249

induced acceleration of the bimolecular addition of the Pd-complexes 42 and 43, respectively to coumarin (44) in comparison to the unimolecular interconversion of 42 to the more stable 43. Thus, the reaction of 42 or 43 with 44, which is thermodynamically controlled by the equilibrium between 42 and 43 at atmospheric pressure, is transformed by high pressure to a reaction kinetically controlled by the formation of either 42 or 43. A common problem in Diels-Alder reactions is the regioselectivity. It is usually controlled by the coefficients of the interacting orbitals in the transition structures and can be improved in many cases by using Lewis acids. However, the application of high pressure also has a great influence as shown for the intramolecular heteroDiels-Alder reaction of the benzylidenebarbituric acid derivative 48 (Scheme 8.13). In this transformation the ortho-adduct 49 and the meta-adduct 50 are formed.

N’

/

I 49

48

50 Intramolecular hetero-Diels-Alder reaction of the benzylidenebarbituric acid derivative 48.

Scheme 8.13.

The intramolecular cydoaddition of 48 has been studied under high pressure up to GOO MPa in various solvents and the kinetics were measured by online FT-IR spectroscopy up to 300 MPa. The overall rate coefficient k for the reaction leading to both the ortho and meta product was derived from the time dependence of the starting material and product concentrations using the modified Kezdy-Swinbourne procedure [ 30, 311 for a pseudo first-order reaction [ 321. The cycloaddition shows a pressure-dependent increase in regioselectivity in favor of the ortho-adduct 49. The studies demonstrate a large solvent effect on the activation volumes, but only a very minor effect on the difference in activation volume for the two pathways leading to the meta and ortho adduct 49 and 50 [ 2 5 ] . Formation of the ortho product 49 is favored in all solvents when going towards higher pressure. Thus, the reaction in acetonitrile shows a meta to ortho ratio of 1:5.54 at 100 MPa and 1:6.33 at 400

t -%

2'2

;

2.0

Dichloromehane

A Toluene

* THF

0 Afe(onitrile

. €

/.o

0

200

400

-

600

Pwal Fig. 8.1.

MPa, while in toluene at 100 MPa a ratio of 1:3.85 and at 400 MPa of 1:4.06 is ob0.3) cm3 mol-l, tained which corresponds to AAV# = -(2.1 f 0.3) and -(1.3 respectively (Figs 8.1, 8.2 and 8.3). From Table 8.5 it can be seen that there are two classes of solvents. In dichloromethane, tetrahydrofuran and 1-chlorobutane the activation volume is about - 30 cm3 mo1-l. This is in the range of values reported for other intramolecular DielsAlder reactions. On the other hand, in toluene and acetonitrile the activation volume is around - 15 cm3 mol-'. The different values of the activation volume for the two classes of solvent is

--

-9

A Toluene 0 Aceimilrile

h

-10

v)

Y

m

; -11 r

c

s C

-12

-

-13

-

-141, , , 0

U

,

, 50

, , , ,

,

100

, , , ,

150

, , , ,

,

200

P [MPaI Fig. 8.2.

, ,

.,

250

8.2 Influence of High Pressure on Selectivity

I

251

-7

Tab. 8.5.

Kinetic data for the intramolecular cycloaddition of 48

~

Solvent

Dichloromethane 1-Chlorobutane THF Toluene

-(2.7 -(4.6 -(2.2 -(2.5

i 0.5) f 2.3) f 1.8) f 0.3)

*

(93.5 0.8) (95.0 0.8) (94.9 k 0.8) (79.8 f 1.8)

(96.3 f 0.8) (99.6 f 0.8) (97.0 0.8) (82.4 f 1.8)

AS$ (ortho) (J mol-' K - ' )

A$: (metaj

T = 110 "C Solvent

AAS~ (J mol-' K-'J

k 9) & 10) f 8) k 11)

(J mol-' K-'J

-(116 f 9) -(116 10) -(lo6 f 9) -(137 f 11)

(I f 2) (6 f 2) (4f 2)

-(110 -(115 -(lo0 -(134

Solvent

AAV:~ (cm3/mo/j

AV: (ortho) (cm3/mo/)

A V ~(metaj (cm3 / m ~ l j

Dichloromethane 1-Chlorobutane THF Acetonitrile Toluene

-(1.5 f 0.1) -(2.1 k 0.4) -(2.0 f 0.2) -(2.1 f 0.3) -(1.3 0.3)

-(33.7 -(30.1 -(34.8 -(17.3 -(13.4

-(32.1 -(28.0 -(32.7 -(15.2 -(12.1

Dichloromethane 1-Chlorobutane THF Toluene

(5 f 1)

T = 110 "C, p = 150 MPa

~

T

7

110 "C.

f 1.2) f 2.5) f 1.6) f 4.1) f 1.5)

f 1.1) f 2.4) f 1.4) f 4.1) f 1.5)

252

I

8 High Pressure in Organic Synthesis: lnfuence on Selectivity

and Tab. 8.6. Activation volumes, partial molar volumes o f the reactant (VSubrtrate),the ortho product (Vo.Pro~ucl), of the ortho transition state (VTS)and reaction volume (AVR = VSubstrate - VTS) for the intramolecular cycloaddition of 48.

Solvent

Dichloromethane

THF Toluene Acetonitrile

Av:?

Av!:

VTS

vo-Product

A VR

(cm’ mol-‘)

(cm’ mol-’1

(cm’ mo1-l)

(cm’ mol-’)

(cm’ mol-’)

(cm’ mol-‘)

242 f 2 244f2 232 i 2 232 f 2

-33.1 -34.2 -13.4 -17.0

-23.6 -24.4 -9.5 -12.1

218 k 4 220 i 3 222 3 220 4

212 & 2 209 k 2 214 f 2 217 i 2

-30 -35 -18 -15

110 “C

VSubrtratr

i 1.2 k 1.5 f 1.5 f 4.1

20

“C

f 2.1 f 2.1 f 1.8 f 4.6

i4 22 f2 f2

surprising. An experimental error is unlikely and it is not very reasonable to assume that a change in the mechanism of the cycloaddition causes a strong variation of AV#, as such a change should also influence the selectivity, which is not observed. In addition, the polarity of the two groups of solvents does not correlate with such an assumption. In particular, toluene and acetonitrile should not belong to the same class of solvents. An explanation for this peculiar solvent effect could be obtained from the determination of the volume profiles. In acetonitrile and toluene the molar volume Vof 48 and 49 is rather similar, but significantly smaller than in dichloromethane and tetrahydrofuran (Table 8.6). On the other hand, the molar volumes of the transition structure and of the products are not greatly influenced by the solvent. The conclusion which may be drawn from these findings is that the significant differences observed for the reaction volume and the activation volume of the intramolecular cycloaddition of 48 as a function of solvent toluene and acetonitrile on one hand and dichloromethane and tetrahydrofuran on the other - are due to differences in solvation of the substrate 48. Another interesting example of the influence of high pressure on the regioselectivity in organic reactions has been observed for the Mukaiyama aldol reaction of unsaturated silyl ketene acetals (51)with aromatic aldehydes by Bellassoued, Dumas and coworkers (Scheme 8.14) [33]. The desired Tadduct 52 was the major compound up to 0.5 GPa (52:53 = 83 :17) while the preference was reversed at 1.7 GPa, making the u-adduct 53 the predominant product (52:53 = 25 :75). This pressure dependence of the regioselectivity may imply that the transition structure leading to the linear aldol product 52 is less compact than that in the formation of the branched aldol product 53.

pressure

Y C L

51 Scheme 8.14.

PhAOTMS 52

53

Influence of pressure on the Mukaiyama aldol reaction of 51.

8.2

influence of High Pressure on Selectivity

I

253

8.2.3 DiastereoselectiveTransformations 8.2.3.1 Reactions with Simple Diastereoslectivity

In the discussion of the influence of high pressure on the diastereoselectivity of chemical transformations wc will first look at simple diastereoselectivity and later at induced diastereoselectivity.Most of the work concerning the influence of high pressure on diastereoselectivity has been carried out on Diels-Alder reactions. For an understanding of the effect of pressure on these reactions, a careful analysis of the different pathways must be undertaken. It is usually accepted that in most cases Diels-Alder reactions are concerted, they can, however, also proceed via biradicals or zwitterions depending on the solvents and substrates (Scheme 8.15).

concerted

+3 /

\

.Q biradicaloid

\ 0 //I t

0

ionic Scheme 8.15.

Possible mechanisms of Diels-Alder reactions

If two diastereomers are formed by different mechanisms, a pronounced pressure effect should be observed. Narner et al. have shown that in the dimerization of 1,3-cyclohexadiene(54) at 110 "C and atmospheric pressure an endo- and an exo[4 2]cycloadduct as well as two [2 2]cydoadducts 56 and 59 and a [G 4]ene product 57 are formed in a ratio of 74.3:11.4:4.1:2.5:7.7 (Scheme 8.16). At 700 MPa a ratio of 76.9: 3.2:0.8:0.3: 18.8 according to the given rate constants was obtained. Especially interesting is the increase in selectivity of the two diastereomeric [4 + 21 cycloadducts 55 and 58 from 6.5: 1 to 24: 1 in favor of the endo-adduct 55 [34, 351. The volumes of activation were calculated to be AVIZ+2:endo = -28

+

+

+

cm3 mol-', AV,Z+21exo = -22 cm3 mol AV,:+21syn = -22 cm3 mol-', AV,:+21antr = -18 cm3 mol-' and AV$ = -32 crn3 mol-'. Due to the low AV$+2,exovalue it was assumed that the ex0 diastereomer 58 is formed via a biradical intermediate whereas the formation of the endo product occurs via a concerted mechanism. However, the activation volume leading to the ene product is the lowest observed and therefore the ratio of this compound in the product mixture is more than doubled at 700 MPa compared to atmospheric pressure.

254

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

/

8

AV* [cm3 rno~-l]: Scheme 8.16.

56 anti-[2+2]

endz4+2]

58 exo-[4+2]

59 syfl-[2+2]

-22

-22

57 [6+4]ene

Dimerisation of cyclohexadiene.

The first synthetically useful increase of diastereoselectivity by high pressure was observed for the hetero-Diels-Alder reaction of enaminoketones (60) with alkyl vinyl ethers (61)(Scheme 8.17) [36-411. This transformation is of broad synthetic value, because it allows efficient access to 3-amino sugars 136. 42, 431.

NPht

NPht

R'

R1

60 a: R1 = CCI3 b: R' = CF3 c: R1 = C02Me d: R1 = COpMenthyl

61 a: R2 = Et b: R2 = IPr c: R2= Bu d: R2 = pMOPh

OR2 62

63

Pht =

:cjs 0

Scheme 8.17.

Hetero-Diels-Alder reactions o f enarninoketones and vinyl ethers.

For a reasonable reaction rate it was necessary to introduce an electronwithdrawing group at position 2 of the l-oxa-1,3-butadiene. In order to investigate the influence of this substituent on the simple and induced diastereoselectivity, groups of different size and electron-withdrawing strength were used. Thus, a methyl ester, a menthyl ester as a chiral functionality, a trichloro, a chlorodifluoro

8.2 Influence of High Pressure on Selectivity Tab. 8.7. Kinetic data of the hetero-Diels-Alder reaction of l-oxa-l,3-butadienes (60a-d) and 61a i n dichloromethane. ~

+

60a-d 61a Ratio 62:63 at 0.1 MPa

a: CC!, 1.5:1

AVf (cm3 mol-') A H Z (kj mol-') A S Z (kj mol-') A A V # (cm3 mol-') in hept/iscdurene in CHlCN A A H Z (kJmol-') AASZ (kj mol-I)

24 63 -155

a

~-

b: CF3

~

~

c: CO, Me 6.3: I

d: COpMenthyl 6.3:1

-23 53 -162 -(3.9 f 0.5)

-24 52 -174

-

i 0.5)

-

-

-

-(5.8 I 0 . 5 )

-

-

-8 -16

-10 -10

-9 -11

-

2.8: 1

-

-(5.8 f 0.5) -(5.3

-(2.4

-

0.2)

0.4=

-

Related to the menthyl group.

and a trifluoro group have been employed as electron-withdrawing groups. The cycloaddition of enamino ketones such as GOa and ethyl vinyl ether (Gla) led to the dihydropyrans G2a and G3a as a mixture of diastereomers in very good yield. The CC13 < CCIFz < CF3 according to reactivity of GO increases in the order COzMe I the strength of the electron-withdrawinggroup (Table 8.7). With increasing pressure a strong enhancement of the reaction rate and also an increase in the diastereoselectivity was observed. In all cases, the formation of the cis-adducts (62) presumably via an endo-E-syn-transitionstructure is preferred (Scheme 8.18). The cis-product could also be formed via an exo-Z-syn transition structure; however, due to the bulky phthalimide group it seems unlikely that a (Z)-doublebond exists in the transient state. NR3R4

NR3R4 R2&

end&€-syn 64

endo-Z-anti

66 Scheme 8.18.

-

cis

exo-€-anti

-

trans

65

-

trans

exo-Z-syn

--t

cis

67 Transition structures of hetero-Diels-Alder reaction of 1-oxa-1.3-butadienes.

I

255

256

I

8 High Pressure in Organic Synthesis: lnfuence on Selectivity

The selectivity could also be enhanced by lowering the reaction temperature resulting in a favorable value of A A H f . However, decreasing the temperature alone is not feasible since the reaction rate is too low at reduced temperature. Hence. the reaction half-life for cycloadditions at 0 "C and atmospheric pressure would be about 1.5 years. However, the reaction can be run at low temperatures with a reasonable rate due to the very negative AV# , if high pressure is applied. The great synthetic utility of this approach is demonstrated in the reaction of GOa and Gla, in which the two diastereomeric adducts G2a and G3a are formed at 90 "C and atmospheric pressure in dichloromethane in a ratio of 1.67:1.00 whereas at 0 "C and GOO MPa a selectivity of 13.6: 1.0 is obtained; in acetonitrile at 0 "C and 700 MPa a 15.6:l.O ratio was observed. The data measured in dichloromethane correspond to a AAV# = -(5.8 f 0.5) cm3 mol-I and a AAHf of -8 kJ mol-' (Table 8.7). It is important to note that the AAVZ strongly depends on the nature of the electron-withdrawing group at position 2 of the l-oxa-1,3-butadiene moiety in GO. Therefore the selectivity at atmospheric pressure is lowest for the reaction of GOa and Gla in which the largest pressure-induced selectivity (AAV# = - (5.8 5 0.5) cm3 mol-') is found. The situation is reversed for the cycloaddition of GOc and Gla (AAV' = -(2.4 f 0.2) cm3 mol-l), in which the absolute value for cisltrans selectivity at ambient pressure is highest and the AAV# value is relatively small. A pressure effect on the induced diastereoselectivity using the chiral enaminoketone God is as expected, negligible with AAV# z -(0.4) cm3 mol-'. The examples described clearly demonstrate that the difference in activation volumes increases with steric hindrance. This is further corroborated by the reaction of GOa with vinyl ethers Gla-d having bulkier groups at the oxygen to give the dihydropyrans 62 and 63 (Table 8.8). For these transformations the individual rate coefficients kcis and ktVan,were determined for the first time. A comparison of the kinetic data for the systems GOa and Gla with GOc and Gla reveals that the rate coefficients k,,,, for the cycloadditions are very similar indicating that the electronic effects of the CC13 and COzMe substituents at the diene are of the same size and that the steric effects are less important for the exo transition structures (Scheme 8.18). On the other hand, steric contributions clearly influence the endo transition structures leading to the cis diastereomers under high pressure. Besides the reaction of GOa and Gla, all other reactions furnished the trans-adduct as the main product via an exo-E-anti transition structure at atmospheric pressure and 120 "C. Hence, in the reaction of GOa with Glc containing a t-butyl group a very strong steric interaction in the endo-transition structure is accompanied by both a very low value of Kcis and a large negative value of AAV#. It can be seen in Table 8.9 that there is a continuous increase of AAVf for the reaction of the enaminoketone (GOa) with the enol ethers Gla-d from AAV' = -(5.8 & 0.5) to -(6.9 & 0.7) cm3 mol-' which correlates with the steric bulk of the substituent at the oxygen. Interestingly, the highest AAV# was found for the reaction of the en01 ether Gld containing a p methoxyphenyl group. This group seems to have a greater steric demand in the high pressure reaction than the t-butyl group. The corresponding differences in activation enthalpy AAHf are between -(4.7 1.4) and -(9.9 f 1.6) kJ mol-'.

8.2 Influence of High Pressure on Selectivity

I

257

Tab. 8.8. Kinetic data for the hetero-Diels-Alder reaction o f the enaminoketone (60a) and enol ethers 61a-d i n dichlorornethane.

+

60a 61a-d Ratio 62:63 at 0.1 M P a (yield %)

AV# (cm' mol-') A H # (kJ mol l ) AS# (kj mol-') A A V # (cm' mol-') AAH# (kJ mol-') A A S (kj mol-')

a: Et 1.5:1 (94)

-

24

63 -155 -(5.8 -8 -16

0.5)

b: iPr 0.4:l (91)

c: tau

d: pMOPh

0.2:l (87)

-

-23 65 -147 -(6.5 k 0.5) -7 -18

-24 76 -138 -(6.8 f 0.8) - 10 -28

-14 76 -138 -(6.9 f 0.8) -10 -28

The choice of solvent also has a great influence on the stereoselectivity. In general, cis diastereoselectivity increases with the polarity of the solvent. At ambient pressure, the cisltrans-ratio of the products of the reaction of GOa and Gla in heptane/ isodurene is 0.9: 1 at 110 "C and 1.37: 1 at 25 "C whereas in dichloromethane a cis/ trans-ratio of 1.67: 1 at 90 "C and 2.98: 1 at 25 "C was found (Table 8.9) [44,45]. So far, the largest observed difference in the volumes of activation with A A V + = -(7.3 f 0.6) cm3 mol-' in a diastereomeric reaction was measured for the cycloaddition of GOa and dihydrofuran 68 to give the condensed dihydropyrans G9a and 70a (Scheme 8.19). At atmospheric pressure in toluene at 110 "C a cis/truns-ratio of 1.5 : 1was observed whereas for the reaction of Gob containing the smaller trifluoro group and 68 a ratio of G9b:7Ob = 5: 1 was found; however, the A A V # is only -(3.9 0.9) cm3 mol-'. In order to increase the steric bulk even further, enol ethers with an additional substituent at the double bond were used. Examples are isopropenyl methyl ether 71 and 2-methyldihydrofuran 74. In these transformations the cycloadduct with a cis orientation of the N-phthaloyl and OR group would be formed via an endo-E-syn and the corresponding trans-products via an exo-E-anti transition structure [46]. In contrast to the reactions described earlier, in the cycloaddition of GOa and 71 to give the products 72 and 73 the e m transition structure has the strongest steric Tab. 8.9. Solvent effect on the selectivity of the hetero-DielsAlder reaction of 60a and 61a.

Pressure

T ("Cl

Solvent

62 :6 3

110

Hept./Isodurene Hept./Isodurene

0.9: 1 1.37: 1 1.67: 1 2.98: 1 13.6: 1 15.6:l

(MP4 ~~

~

0.1

0.1 0.1

0.1 GOO 700

25 90 25 0.5 0

CH2C12 CHzClz CH2C12 CHjCN

258

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

PhtN

R

60

68

69

70

*

a: R = CC13

cis:frans

1.5 : 1

48 % yield

M V f = -(7.3 0.6) crn3.mol-'

b: R = CF3

cistrans

5.0 : 1

69 % yield

M V f = -(3.9 0.9)cm3mol-'

60a

60b

71

74

73

72

75 0.1 MPa, 110 ' C 160 MPa, 30 "C

76

77

1 1

5.2

2.3

Hetero-Diels-Alder reactions o f enaminoketones 60a.b with dihydrofurans 68, 74 and isopropenyl ether 71. Scheme 8.19.

hindrance due to the interaction between the electron-withdrawing group at position 2 of the l-oxa-1,3-butadiene(6Oa) and the methyl group at the enol ether (71) and should therefore be stabilized undcr high pressure. This is indeed the case if the reaction is performed in dichloromethane or acetonitrile. However, the opposite effect is observed, performing the reaction in toluene/isodurene. Nevertheless? the corresponding AAV# is rather small with values of +(2.6 f 0.2), +(0.7 f 0.3) and -(0.5 f 0.6) cm3 mol-I. When measuring the partial molar volumes of the diene and the dienophile, it was found that the solvent effects on the activation volume are primarily due to the differences in solvation of the starting materials, as described earlier for other transformations. The intrinsic part of the activation volume accounts only for about 30 % of the measured overall activation volume. The unexpectedly small AAV' can be correlated to an overload of steric bulk in the exo-E-anti transition structure. A surprising effect was observed for the cycloaddition of 60b and 2-methyldihydrofuran 74 which led to a mixture of the annu-

8.2 lnflucnce of High Pressure on Selectivity

I

259

lated and the spiro compound 76 and 77 in a 1:5.2 ratio at atmospheric pressure and 110 "C. It can be assumed that 77 is formed from the tautomer 75 with an exo-methylene group even though this compound was not observed by 13C-NMR spectroscopy in the solution of 74. At higher pressure the ratio of 76 : 77 decreased to 1: 2.3 indicating that the sterically more demanding transition structure leading to 76 is stabilized under high pressure; from the measured data a AAV# = -(10.4 1.7) cm3 mol-' was obtained. The importance of steric bulk in the transition structure in order to obtain a large AAVf can also be seen in the cycloaddition of 78 and Gla to give the cycloadducts 79 and 80 described by Boger et al. [47]. In this reaction the pressure effect on the stereoselectivity seems to be negligible since the endolexo ratio of 5.7: 1.0 was observed to be the same at 0.62 and 1.3 GPa. This is in good agreement with the lower steric demand of the oxabutadiene (78) compared to GOa.

jM: - &oEt;eo2c

Me02C

78 Scheme 8.20.

OEt

61a

Me02C

OEt

79

80

Diels-Alder reaction o f oxabutadiene 78 with 61a.

A negligible pressure effect on the diastereoselectivity was also observed for the cycloaddition of the enamine carbaldehyde (81a) carrying an electron-withdrawing group at position 3 and Gla to yield the dihydropyrans 82 and 83 (Scheme 8.21).This reaction was again studied by direct quantitative infrared spectroscopy up to 300 MPa between 45 and 95 "C in different solvents. The activation volume was found to be -(25.1 & 1.7) cm3 mol-l in dichloromethane and -(25.0 1.8) cm3 mol-I in isodurene. Thus, in this reaction solvent polarity had no influence on the pressure dependence of the rate coefficient; in addition, the ratio of the two diastereomeric products is not changed under high pressure; thus the AAV# value is very small (AAV# < 1 cm3 mol-I). On the other hand, if one uses l-oxa-1,3-butadienescontaining an additional alkyl group at position 2 of the oxabutadiene a large AAV# is observed. For compound 81a with an ethyl group at position 2, AAV' = -(5.2 0.3) cm3 mo1-l and for 81c 0.3) cm3 mol-' was observed. with an isopropyl group a AAV+ value of -(5.3 In contrast to the reaction of the enaminoketones (GO) where the cis-cycloadducts are formed via an endo-E-syn transition structure, here the cis-compounds are formed via an exo-Z-syn transition structure due to a strong hydrogen bond between the NHAc and the carbonyl group which stabilizes the (Z)-configuration in the oxabutadiene. This is again in agreement with the observed pressure effects. In the reaction of the enaminoketones, 81b and 81c, the trans-products 83b,c are obtained preferentially under high pressure. As high pressure stabilizes the endotransient, 83b and 83c must be formed via an endo-Z-anti-transition structure (Scheme 8.18) [48].

260

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

- M e o 2 c aNAc

NAc

Me02C

OEt

R

61a

81

OEt

R

OEt

82

83

CH2C12, 100 "C 50 MPa 500MPa

R = Et: 1.06 0.74

: :

1 1

50MPa 500MPa

R = IPr: 1.29 0.88

: :

1 1

a: R = H CH2C12:AV' = -(25.1

k

1.7) crn3~mol-';

Isodurene: A V f = -(25.0 f 1.8) cm3.mol-'

MV'

< -I cm3.mol-'

b: R = Et M V f = -(5.2

f

0.5) cm3.mol-'

c: R = rPr M V ' = -(5.3

f

0.4) cm3.mol-'

Scheme 8.21.

Cycloaddition o f enaminoketones and xarbaldehydes 81a-c.

A pronounced pressure effect on the diastereoselectivity was also observed recently by Collignon et al. in the synthesis of phosphono-substituted dihydrothiopyrans (85) by a hetero-Diels-Alder reaction of a a-phosphono-cq p-unsaturated dithioester (84) as l-thia-1,3-butadiene and ethyl vinyl ether (6la) as dienophile (Scheme 8.22) [49]. The diastereomeric cydoadducts 85 and 86 were obtained in 86 % yield in a cis/trans-ratio of 85 : 15 at 125 "C and nearly atmospheric pressure in a sealed tube after 2 h. Under high pressure of 1.1 GPa at 20 "C a &-/trans-ratio of 36: 64 in 90 % yield was found. It was shown that the reaction proceeded under kinetic control and moreover that the starting material did not lose its stereochemical integrity under the reaction conditions. Thus, the trans-cycloadduct must be formed via an endo-Z-anti transition structure, whereas the cis-isomer results from an exo-Z-syn transition structure (Scheme 8.18). Usually one should expect the endo-adduct (86) to be the main product under normal thermal conditions; here, however, as described before the exo-product is formed preferentially at 120 "C probably due to steric hindrance. However, in the endo-transition structure under high pressure the transient as the more compact structure is favored. The authors did not measure the kinetics for this reaction, but it can be assumed that the observed change in diastereoselectivity is not only caused by a pressure effect due to a favorable AAVz but also by a temperature effect due to a favorable AAH#. In a similar approach, reaction of the cc-carbonylated styrylphosphonates 87 with 6la yielded the corresponding dihydropyrans 88 and 89 with a low selectivity of 44 :5G in favor of the trans diastereomer (Scheme 8.22) [SO]. Under high pressure of 1.0

8.2 Influence of High Pressure on Selectivity

I

261

GPa at 20 "C in the presence of t-BuOH an excellent cis/truns selectivity of 8:92 and a yield of 95 % was observed. Again, in this case it must be assumed that the endo-Z-anti transient of the I-oxa-1,3-butadiene moiety is preferred under high pressure. Also, for this reaction a cooperative pressure and temperature effect seems to be effective. It should be noted that the dihydrothiapyrans 85/86 as well as the dihydropyrans 88/89 can be obtained very efficiently by a three-component domino-Knoevenagel-hetero-Diels-Alderreaction [ 1-41 starting from the phosphonates, a suitable aldehyde and an enol ether without preparing the oxa- and thiabutadienes separately.

9 ( E t o ) 2 p ~ r +

EtS

9

,Ar

Me 87

EtS

61a

84

(Et0)2P

_ j

OEt

'OEt 61a

EtS

85

9

+

OEt

Ar

OEt 86

9

Ar

Me 88

89

Ar = 4-N02-C6H5 Scheme 8.22. Hetero-Diels-Alder reaction of phosphonothiadiene 87 and styrylphosphonate 90 with ethylvinylether 61a.

Pressure effects on the diastereoselectivity can also be observed for intramolecular cycloadditions such as the hetero-Diels-Alder reaction of 90, even though the AAV# are smaller compared to intermolecular reactions (Scheme 8.23). The kinetics were again measured by on-line FT-IR spectroscopy and the stereoselectivity by HPLC. At atmospheric pressure in toluene under reflux the reaction of 90 led to a 5.2: 1 ratio of the diastereomeric cycloadducts 91 and 92 in 93 % yield. Increasing pressure favors the formation of the cis-adduct 91,which is probably formed via an endo-E-syn transition structure. Interestingly, in the ground state of 90 the Zconfiguration is more stable and it is therefore assumed that isomerization of the Z- to the E-double bond occurs prior to the cycloaddition [Sl]. From the slope of the plot of 1n(91/92) versus pressure, AAV# is calculated to be - ( l A 0.2) cm3 mol-' with the individual values for AV# = -(19.4 A 0.2) and AV# = -(17.9 0.G)cm3 mol-' at 343 K. Although the AV# values are relatively low in comparison with data reported for other intramolecular Diels-Alder reactions e.g.

262

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

by Isaacs et al. [ 5 2 ] , it is believed that the reaction is still concerted due to the stereointegrity of the dienophile moiety which was determined by using 13C-labeled compounds [ 5 31.

90

91

92

0.1 MPa, 110 ‘C, toluene 93 % yield 91 : 92 = 5.2 : 1

500 MPa, 70 ’C, CH2CI2, >95 % yield 91 : 92 = 6.8 : 1 AV’,,

= -(19.4 k 0.5) crn3mol-’

AVtfrans= -(17.9

MVf

0.6) crn3.mol-’

= -(1.6 f 0.2) ~ r n ~ . r n o I - ~

Scheme 8.23.

A synthetically useful pressure-induced increase of diastereoselectivity was also found for normal Diels-Alder reactions such as the cycloaddition of thc phenyl butadienes (93a-c) with the dicyano ethylenes (94-97) to yield the cis-adducts 98a-d, 100a-d and 102a-d as well as the trans-adducts 99a-d, 10la-d and 103a-d (Scheme 8.24) [ 541.

j’ ,yCN

Toluene, 110°C N : &

+

R’

\

R2

93a: R’ = Me 94: R2 = Me 93b: R’ = Et 95: R’ = Et 93c: R1 = Pr 96: R2 = /Pr 97: R2 = W U Scheme 8.24.

R’

N -k: & R2

“/R2

R’

98a-d: R1 = Me 99a-d: R’ = Me 100a-d: R’ = Et 1Ola-d: R’ = Et 102a-d: R1 = /Pr 103a-d: R‘ = /Pr

R2: a = Me, b = Et, c = /Pr, d = Wu

Diels-Alder reactions o f E-1-phenyl-l,3-butadienes with 1 , l -dicyanoethylenes.

Assuming pressure-dependent diastereoselectivity to be primarily caused by differences in steric interaction of the two diastereomeric transition structures and not by a change in reaction mechanism, an increase in -AAVf with increasing bulkiness of the substituents R1 and RZ was anticipated. The dienes 93a-c were prepared in a three-step sequence starting from trans-cinnamaldehyde by alkylation, oxidation and Wittig reaction. The cycloadditions were carried out in toluene

8.2 Influence of High Pressure on Selectivity Tab. 8.10.

Results of the Diels-Alder reaction of 93a-c with 94-97.

Entry

1 2 3 4

5 6 7 8 9 10 11 12

R2

93a + 94 93a 95 93a 96 93a 97

+ + + 93b + 94 931, + 95 931, + 96 93b + 97 93c + 94 93c + 95 93c + 96 93c + 97

Me Et

iPr tBu

Me Et

iPr tBu

Me Et

iPr tBu

Yield PA] (reaction time)

cis: trans

70 (12 h) 80 (12 h) 66 (12 h) 90(48h)

1.86:1 1.56:1 1.18:1 1 :5.56

-(1.9 k 0.1) (1.9 i 0.2) (4.3 2 0.3) -(6.4 k 0.6)

73 (12 h) 95 (12 h) 69 (24 h) 96 (48 h, 1.0 GPa)

1.17:1 1.13: 1 1: 1.5G 1 : 5.88

- - ( 1 . 3 0.1) -(1.6 k 0.1) -(2.9 k 0.3) -(3.9 & 0.3)

54 (12 h) 98 (24 h) 31 (36 h, 1.0 GPa) 70 (24 h, 1.0 GPa)

-2:l 1 : 1.89 1 :3.33 1: >99

AAVf (cm' rnol-')

-(0.7 -(1.9

0.8)

i 0.4)

-

at 110 "C for 12-48 h at atmospheric pressure to give the cyclohexenes 98-103 each as a mixture of two diastereomers (Table 8.10). Under these conditions, the reactions of the bulkier substrates 931, and 97, 93c and 96 as well as 93c and 97 to give the cycloadducts 100d/101d, 102c/103c and 102d/103d provided only low yields. However, the yields could be easily improved by running the cycloadditions under high pressure at 1.0 GPa giving yields between 31 and 98 %. The ratios obtained for the cis to trans diastereomer of 98-103 at atmospheric pressure clearly depend on the steric requirements of the substituents R' and R2. With R' and R2 being methyl or ethyl (entries 1 and 2, 5 and 6) and R' = iPr and R2 = Me (entry 9) the formation of the cis-adducts is favored. Towards increasing bulkiness of the substituents on either the diene or dienophile (R1 = iPr or/and R2 = iPr, tBu; entries 3 and 4, 7 and 8, 11 and 12) the amount of the cis product decreases significantly. Thus, the cycloaddition of the substrates 93a and 97 leads exclusively to the trans product 103d. These results are in agreement with the expectation that the cis-diastereomers are formed via an endo-E-syn transition structure which is energetically favored unless the substituents R' and R2 are too bulky. Thus, under conditions of a strong steric interaction the em-E-anti transition structure is preferred. It should be noted that the Diels-Alder reactions of 93a-c with 94-97 are kinetically controlled and that isomerization of the double bond in the butadienes 93a-c and the stereogenic centers in the products does not take place under reaction conditions. For the cycloadditions two effects should be considered, namely the steric and the electronic influence of the substituent on the Diels-Alder reaction [ 551. The cycloaddition already discussed belongs to the normal electron-demand type, for which the overlap of the LUMO of the dienophile and the HOMO of the diene dominates. Whereas steric hindrance obviously increases both in the diene and the dienophile when going from methyl to ethyl, i-propyl and t-butyl substituents, the

I

263

264

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

electronic effects are different for the diene and the dienophile. As a consequence of the interaction between the +I effect of the alkyl substituents with the relevant molecular orbitals, the reaction rate decreases from dienophile 94 to dienophile 97 using an identical diene and increases from diene 93a to diene 93c using an identical dienophile. Furthermore, the electronic effect on the reaction rate should be more obvious for the formation of the trans product, whereas the formation of the cis product should be influenced to a higher extent by steric interaction. The differences in activation volumes AAV# = AVZ - AV,f,,, were derived from the pressure dependence of the product ratios c,,/ctrans in dichloromethane solution at pressures ranging from 25 to 300 MPa. As can be seen from Table 8.10, negative AAV# values were found for the entire set of cycloadditions. According to the well-established rule that high pressure favors sterically-hindered processes, this indicates a larger extent of steric hindrance in cycloadditions that proceed via an endo transition structure, to give the cis adduct than for the reaction via an ex0 transition structure. Fully consistent with this argument is the clear increase of AAVf with steric bulkiness of the substituent at the dienophile. Thus, for the cycloaddition of 93a (R’ = Me) with 94 ( R 2 = Me), 95 (R2 = Et), 96 ( R 2 = iPr), and 97 ( R 2 = tBu) a significant increase in -AAV# from (1.9 f 0.1) to (6.4 5 0.6) cm3 mol-’ is found. The same trend was observed within the series of cycloadditions of 93b (R’ = Et) with dienophiles 94-97, where -AAV# continuously increases from (1.3 f 0.1) to (3.9 f 0.3) cm3 mol-’. Also for the cycloadditions of 93c (R1= iPr) a significant enhancement of -AAV# was found towards increasing steric bulkiness, (0.7 0.8) cm3 mol-’ for 93c 95 and (1.9 & 0.4) cm3 mol-’ for 93c 96. However, attempts to increase steric hindrance by adding even bulkier substituents on both diene and dienophile did not yield -AAV‘ values that exceeded e.g. the number found for the 93a 4-97 cycloaddition. For a particular dienophile, 95 or 96, variation of the diene from 93a to 93c led to a decrease in -AAV# although the steric requirement of the transition structures would be expected to increase. The data given in Table 8.10 clearly indicate that sterically overloading the transition structure may lead to a situation where geometries with favorable interactions as are characteristic of endo transition structures, cannot be organized. It is as a result of these arguments that a cis product is not found in the cycloaddition of 93c 97. It is thus only within a series with the diene 93a remaining unchanged that the AAV’ value increases towards larger steric hindrance which corresponds to a pressure-induced enhancement of diastereoselectivity. This was observed when varying the substituent R2 in the dienophile 94-97 by increasing its steric size. The measured differences in activation volume are fairly large in this series, with -AAVf up to (6.4 0.7) cm3 mol-’, which may allow for applications in selective synthesis. It must be assumed that the expression “endo transition structure” and perhaps also “ex0 transition structure” does not refer to well-defined species, e.g. with identical geometry at the reactive site. The activation volume data indicate that, depending on the type of substitution at the diene and the dienophile, the overlap of

+

+

+

8.2 Influence of High Pressure on Selectivity

the dominating orbitals: in particular in the endo transition structure varies. An interchange of the substituents R1 and R2 at the diene and dienophile demonstrates, e.g. by comparing the differences of the activation volumes for the reactions of 93c + 95 (AAVf = -(0.7 f 0.8) cm3 mol-’) and of 9313 96 ( A A V = -(2.9 0.3) cm3 mol-’) that the endo transition structure of the cycloaddition of 93c 95 which has the bulky i-Pr substituent as R1 in the diene, is of lower “quality”. However, we do not propose a change in the reaction mechanism, in which the “ e m adduct” is formed via a concerted and the “endo adduct” via a twostep pathway. A pronounced pressure and temperature effect on the simple diastereoselectivity was also observed by Metz et al. in the intramolecular cycloaddition of vinylsulfonic esters (104)to give the diastereomeric sulfones 105 and 106 (Scheme 8.25) [56]. In refluxing toluene at atmospheric pressure a 1:1 ratio was obtained, whereas at 1.3 GPa in CHzClz at room temperature the ratio was 1:2.3 with the endo product 106 being formed preferentially.

+

+

104 Scheme 8.25.

105

106

Intramolecular cycloaddition of vinylsulfones.

Michael additions are rather sensitive to steric hindrance; hence, an addition of a nucleophile to /I-disubstituted a,p-unsaturated esters is nearly impossible. On the other hand, the application of high pressure alleviates this limitation and even allows the formation of two adjacent quaternary centers. The reaction of methyl tertbutylcyclohexylidenebromoacetate (107)with benzylamine in refluxing methanol as described by Duhamel et al. gave a mixture of the ester 108a and the amide 108b (Scheme 8.26) [57]. At room temperature another reaction channel is opened which leads to the aziridines 109 and 110.In this reaction, a Michael addition first takes place and this is followed by an intramolecular nucleophilic substitution. At atmospheric pressure, however, a reaction time of 60 days was necessary to obtain an 82 % yield. In contrast, at 1.1 GPa the addition took less than 4 days and the diastereomeric ratio rose from 1: 1.7 at ambient pressure to 1:10 at 1.1 GPa in favor of 110. The pressure effect on photochemical cycloadditions was also investigated in a very elaborate procedure [58]. For this purpose an optical cell with sapphire windows and a laser containing a Xe/FZ/Hz gas mixture was used. Irradiation of cyclopentenone (111)with substituents of different size at C-3 ranging from hydrogen to methyl, ipropyl and t-butyl, with cyclopentene (112)and 3,3-dimethylbutene (115)gave the diastereomeric cycloadducts 113-117 (Scheme 8.27). For the reaction of Illa-d and 112 at ambient pressure and room temperature the anti-product 113 is preferred. However, with growing bulkiness of R the

I

265

266

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

BU

fBu BnNH2

MeOH reflux_ BrKCOpMe

Br*COR

107 BnNH2,

MeOH r. t.

rac- 109 Scheme 8.26.

108a : R = OMe 108b : R = NHBn

ambient or high pressure

rac- 110

Michael addition t o an a,P-unsaturated ester

k 112

111 a: H

b: Me c: d:Bu /Pr

+

8 115

Scheme 8.27.

113

;H&*

’;.x 116

p 114

+ .

117

Photochemical cycloaddition of cyclopentenones and alkenes.

amount of syn-product (114) increases. In the reaction of 1111, and 115 the [2 2]cycloadducts 116 and 117 were obtained in a 0.8: 1 ratio in 66 % yield. As expected for these reactions, the syn-product is preferred under high pressure; with increasing steric interaction in the transition structures the AAV value increases to -(2.0 f 0.3) cm3 mol-l for the cycloaddition of llld and 112 as well as for 1111,and 115. Interestingly, the AAV# value for the reaction of lllc containing an isopropyl group and 112 is slightly lower than the AAV# found for the reaction of 1111, containing an ethyl group (AAV’ = -(0.8 & 0.3) cm3 mol-’ compared to AAV# = (-1.2 f 0.3) cm3 mol-I). For the isopropyl group the steric hindrance may be lowered by a “toothed-wheel effect”.

+

8.2 Influence of High Pressure on Selectivity

I

26J

8.2.3.2

Reactions with Induced Diastereoselectivity

Under conditions of induced diastereoselectivity high pressure reactions are considered in relation to a facial selective transformation under the influence of a stereogenic element or a chiral auxiliary connected to the substrate. However, most of the published reactions are not related to a pressure-induced change in diastereoselectivity,but discuss only the rate acceleration of these reactions under high pressure. Such an example is the hetero-Diels-Alder reaction of l-methoxybutadiene (119) and N-pyruvoyl-, N-glyoxyl- and N-phenylglyoxyloyl bornane-sultams such as 118, described by Chapuis et al. (Scheme 8.28) [59]. The cycloaddition of 118a and 119 led to the cycloadducts 120 and 121 in a 6 6 : 34 ratio and 86 % yield at 1.4GPa and 50 "C. At 20 "C the reaction did not take place, and even in the presence of Lewis acids such as Eu(hfc)3,Tic14 or ZnClz the expected cycloaddition did not occur.

121 Hetero-Diels-Alder reaction of chiral ketones.

Scheme 8.28.

There are also some examples where the reactions can be promoted either by the use of a Lewis acid or by applying high pressure. It is clearly noteworthy that in several cases the application of high pressure is not only the milder procedure but also gives better selectivity. One example was described by Jarosz et al. for the intramolecular cycloadditon of 122 [GO]. With A1C13 at atmospheric pressure 123 and 124 were formed from 122 in a 2: 1 ratio, whereas at 1.5 GPa without the addition of the Lewis acid 123 was the only product (Scheme 8.29).

AIC13 high pressuret or

Bno&co2Me

BnO

OBn 122

Scheme 8.29.

[4

OBn 123

+ 21 Cycloaddition using Lewis acid or high pressure.

OBn 124

268

I

8 High Pressure in Organic Synthesis: lnfluence on Selectiuity

An investigation of the influence of high pressure versus temperature on induced diastereoselectivity was performed by Eguchi et al. for the cycloaddition of the a,/?unsaturated sulfonamide (125) bearing a C2-symmetric chiral pyrrolidine auxiliary and cyclopentadiene (126) (Scheme 8.30) [Gl].The same product ratio for 127:128 of 75 : 25 was found after 2 h at atmospheric pressure and 80 "C and after 14 h at 1.4 GPa and 40 "C. Thus, a pressure-induced diastereoselectivity does not exist in this reaction.

125

126

127

128

I Asymmetric Diels-Alder reaction o f chiral trifluoromethylated unsaturated sulfonamides.

Scheme 8.30.

For the synthesis of compactin, Konoike et al. performed an intramolecular cycloaddition of the chiral (2)-129 to give the decalin derivatives 130 and 131 1621 (Scheme 8.31).The best results were obtained under high pressure at 1.0 GPa and 20 "C resulting in 53 % of 130 and 8 % of 131. Whereas the Lewis acid-induced reaction gave a lower selectivity of 42 % of 130 and 16 % of 131, the thermal reaction showed much worse results with 20 % of 130 and 5 % of 131. However, it is not clear whether these results are due to a pressure-induced improvement of selectivity or due to an isomerization of the starting material caused by the Lewis acid or the higher temperature.

129 Scheme 8.31.

130

131

Intramolecular Diels-Alder reaction on the way to compactin.

Chiral 2-nitro-1-sulfinylalkenes undergo a [4 + 21 cycloaddition with cyclopentadiene either in the presence of a Lewis acid or under high pressure of 800 MPa at room temperature as described by Fuji et al. [ G 3 ] . In both cases complete diastereoselectivity was observed with (Z)-sulfinyl compounds such as 132; in contrast, only low diastereoselectivity was found when using the corresponding (E)-sulfinyl

8.2 Influence of High Pressure on Selectivity

I

269

compounds (Scheme 8.32). Reaction of 132 and 1,3-pentadiene (133) led to 134 in 81 % yield at 800 MPa. However, a pressure-related increase in selectivity was not observed.

132 Scheme 8.32.

133

134

Diastereoselective Diels-Alder reaction of chiral 1-(alkylsulfinyl)-2-nitroalkenes.

A pressure effect on the induced diastereoselectivity was observed by Katagiri et al. for the asymmetric Diels-Alder reaction of di-l-menthyl acetoxymethylenemalonate (135) with cyclopentadiene (Scheme 8.33) 1641. Whereas there was no reaction at atmospheric pressure even at high temperature, the transformation proceeded smoothly at 1.3 GPa and room temperature to give a mixture of the endo and exo isomers 136 and 137 together with the corresponding diastereomers in 96 % yield with an endo: exo ratio of 1: 1.8 and an induced diastereoselectivity of 3.3: 1 and 4:1, respectively. With decreasing pressure, the ratio endo: exo increased while the induced diastereoselectivity decreased for both isomers. At 0.5 GPa an endo: ex0 selectivity of 1: 1.3 with an induced diastereoselectivity for the endo-adduct of 3 : 1 and the exo-adduct of 2.6: 1 was observed. It was concluded that the negative activation volume for the exo addition is larger in magnitude than that for the endo addition which is in contrast to the usual observation. Lewis acids, such as diethylaluminum chloride, also favor the exo diastereomer 137. On the other hand, irrespective of the pressure employed, the use of zinc chloride or Yb(fod)3 as catalysts gave the diastereomers 136 and 138 with the endo diastereomer 136 predominating. With titanium tetrachloride as catalyst the reaction proceeded under atmospheric pressure at -78 "C giving a slightly lower yield than under high pressure (80 %), but with an endo: exo selectivity of 3 :1 and >99 % de for the endo isomer.

high pressure

OAc 135

126

M = I - menthyl

136

+ diastereomer

137

C02M 138 Scheme 8.33.

Asymmetric D i d - A l d e r reaction of chiral alkylidene malonates.

270

I

8 High Pressure in Organic Synthesis: lnfluence on Selectivity

The aza Diels-Alder reactions of a,P-unsaturated sulfinimines (140)represent a very efficient approach to enantiopure dihydro- and tetrahydropyridines (141) (Scheme 8.34, Table 8.11); for a reasonable reaction rate the l-aza-1,3-butadiene moiety 140 must carry an electron-withdrawing group at the 3-position 1651. The compounds are accessible in only three steps starting from commercially available substrates. Thus, the enantiopure l-aza-l,3-butadiene can be prepared from the enantiopure menthyl sulfinate with lithium hexamethyldisilazide followed by addition of acetic acid and an cc,/j'-unsaturated aldehyde. The cycloadditions of sulfinimines such as 140 run under mild conditions with high yields and excellent endo-selectivityin most cases when high pressure is applied. In these reactions two endo and two e m transition structures namely syn and anti to the sulfoxide moiety should be discussed. The cycloaddition of 140 and t-butyl vinyl ether was performed under various pressures ranging from 0.2 to 1.2 GPa. Ph

bCN

R 3 v R3

17khrr 7ri

139

Ph R3 2

,,VCN

140

Scheme 8.34.

141

Hetero-Diels-Alder reaction of sulfinirnine 140.

A large increase in the conversion rate with increasing pressure was observed which corresponds to the well-known fact that Diels-Alder reactions have a strong negative activation volume A V f . Following El'yanov's equation [ 661 the differences in the activation volumes AAV# were determined from the plot of Tab. 8.11.

Cycloaddition o f sulfinirnines under variable

pressure.

P (=pa)

Yield ("/.I

Ratio of diastereomen endo I:endo 1l:exo I:exo IJ

1.1 0.9 0.7 0.5 0.4 0.3 0.2 1.10-4

99 87 81 44 24 11 26= 4

2.77: 1.00:0.54:0.31 2.45: 1.00:0.50:0.30 2.13: 1.00:0.52:0.35 2.21:1.00:0.55:0.37 2.01: 1.00:0.52:0.52 2.01 : 1.00:0.55 :0.35 1.75: 1.00:0.50:0.36 1.46: 1.00: 0.50: 0.43

Reaction time: 4 days. AAV#(endo 1 - endo 11) = - (1.8 0.5) cm3 mol-' AAVz(endo I - e m 11) = -(2.0 f 0.4) cm3 mol-'. a

+

8.2 Influence of High Pressure on Selectivity

In(endo Ilendo I) and In(endo Ilexo I) versus Y(p). A AAVz = -(2.0 _+ 0.4) cm3 molt' was found for the formation of the endo I- and exo I-cycloadductwhich corresponds to the expected stabilization of the more crowded endo transition structure under high pressure (Scheme 8.35). Somewhat unexpected, however is the preferential formation of the endo I cyloadduct compared to endo I1 with a AAVz (endo I-endo 11) = -(1.8 & 0.5) cm3 mol-'. This would actually mean that the attack of the dienophile anti to the toluyl group is stabilized under high pressure, probably indicating that this is the sterically more hindered approach. Even though the AAV# values are significant, they are too small to allow a pronounced pressure-induced enhancement of the induced diastereoselectivity in these cycloadditions.

endo-attack anti to p-To1

Bu

endo-attack

syn to p-To1

endo I

hCN

tBuO

6

N1

0 p-To1'"Og

endo II

exo-attack anti to p-To1 ex0 I

exo-attack syn to p-To1 ex0 II Scheme 8.35.

Transition structures of the Diels-Alder reaction of sulfinimines with enol ethers.

Interesting pressure effects were also observed for the intramolecular Heck reaction of the enantiopure amides 142 obtained from amino acids by reduction to the corresponding amino alcohols, oxidation to the aldehydes and Wittig reaction followed by alkylation with halobenzylhalides [67]. The Heck reaction of 142a in the presence of Pd(Ac)z, PPh3 and TPAB predominantly led to the isoquinoline

I

271

272

I

8 High Pressure in Organic Synthesis:

influence on Selectivity

143 in good yield under atmospheric pressure; in contrast the bromo compound 1421, gave a much lower yield ( < l o %) (Scheme 8.36). However, at 1.0 GPa a 90 % yield of 143 was obtained with >99: 1 diastereoselectivity. A pronounced pressure effect on the diastercoselectivity was also observed for the similar Heck reaction of the haloarenes 144 and 145 bearing an ( E ) - or (Z)-alkenemoiety to give the isochromanes 146 and 147 in good yields [G8]. The selectivity of the reaction depends on the size of the substituent R; increasing bulkiness of R leads to a decrease in diastereoselectivity but on the other hand to a preferential formation of the vinyl substituted isochromanes. High-pressure experiments confirm the proposed mechanism and show again that bromoarenes such as 145, which are unreactive at atmospheric pressure, give good results when the cyclization is performed under high pressure. Furthermore, the induced diastereoselectivity in the cyclizations was improved significantly when the reactions were performed under high pressure. At 0.1 MPa and 80 "C 144a led to 14Ga and 147a in a 4:1 ratio togethcr with some of the double bond isomers. At 1.0 GPa and 60 "C a ratio of 2 0 : 1 of 14Ga: 147a was obtained. Similarly, reaction of 1441, containing an isopropyl group was improved at 1 GPa to a ratio of 14Gb:147b = 3 . 2 : 1 compared to a 1:l mixture at atmospheric pressure.

142 a:X=l b: X = Br

143 and diastereomers and double bond isomers

Pd(OAc)2, PPh3, KOAc,TPAB,DMF *

mR+ 9 R =/

144: X = I 145: X = Br R = a: Me,b: /Pr

/

146

147

and double bond isomers

Scheme 8.36. Stereoselective Heck reaction (TPAB = Tetrapropylammoniurn bromide)

Butz and Sauer investigated the intramolecular Diels-Alder reaction of hrfuryl fumarates 148 to synthesize oxabicyclo[2.2.l]heptene derivatives 149 and 150 (Scheme 8.37) [ 691. Four new stereocenters of defined configuration are generated in this transformation under the induction of the stereogenic center in the side arm with remarkable selectivity. The cycloaddition of 148 with various substituents R at the stereogenic center was investigated and it was found that the diastereo-

8.2 Influence $High Pressure on Selectivity

selectivity was more or less identical using either high pressure or thermal conditions, except for transformation 148a with an iPr-group. 148a gave a diastereomeric mixture of 149 and 150 in a ratio of 9O:lO at 700 MPa and 78:22 under thermal conditions. In contrast to other studies [54] which deal with endoiexo selectivity, here two exo-adducts with very similar structures and activation volumes were considered. The difference in activation volumes is calculated by semiempirical calculations to be AAV# = 0.2-0.4 cm3 mol-'.

ATor Ap

0

C02Me

Scheme 8.37.

150

149

148 a: R = iPr

Intramolecular Diels-Alder reactions o f furfuryl furnarates.

Similarly, the intramolecular Diels-Alder reaction of the chiral triene 151 with a sulfonate moiety connecting the 1,3-diene and dienophile gave the cycloadducts 153a, 1531, after hydrolysis of the primary adducts 152a, 1521, (Scheme 8.38) [70]. Only two of the four possible diastereomers were obtained. However, the yield and the diastereoselectivity are nearly independent of the reaction conditions. After the reaction at 1.0 GPa and room temperature for 18 h 153a and 153b were obtained in 72 % yield and a 78: 22 ratio whereas at 130 "C and nearly atmospheric pressure for 4 h a 67 % yield and a ratio of 82 :18 was found. The missing influence of pressure on the diastereoselectivity is not surprising, since there is no special steric interaction in either one of the two transition structures leading to the products.

OBn

151

Scheme 8.38.

152 a: 0-H b: a-H

153 a: p-H b: a-H

Intramolecular Diels-Alder reaction o f ethene sulfonates.

In contrast, the Michael addition of diphenylmethanamine (155) to the chiral crotonate 154 derived from an arylmenthol-type auxiliary, serves as one of the few examples of a pronounced positive effect of high pressure on induced diastereoselectivity as shown by Dumas et al. [71]. At 1.4 GPa the expected /I-amino ester 156, was formed with a high diastereomeric excess of 98 % de, while at atmo-

I

273

274

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

spheric pressure only 10 % de was obtained (Scheme 8.39). The same authors recently published theoretical investigations of SIBFA (sum of interactions between fragments ab initio computed) molecular mechanics calculations predicting the influence of pressure on the selectivity [72]. In these computations a relation between compressibility and atomic van der Waals radii was introduced to take into account the influence of pressure on inter- and intramolecular energies. These results give, at least at a qualitative level, an insight into the structural modifications due to pressure and are in good agreement with experimental data.

Ph2CHNH2 155, --r

MeOH

154 Scheme 8.39.

0.1 MPa, 40 'C, 5d, 50 Yo,10 Yode 1.4 GPa, 20 'C, 12h, 67 YO,98 YO de

156

Michael addition.

Carretoro et al. investigated the cycloaddition of enantiomerically pure ( S ) - p tolylsulfinyl trialkoxycarbonyl ethene (157)with cyclopentadiene (126)to give the bridged cyclohexenes 158-161 (Scheme 8.40) [73]. As expected the reactivity of this transformation is substantially increased under high pressure but also the n-facial selectivity is improved. At atmospheric pressure over 3 days the two exo-products 160 and 161 (62 and 8 %) and the endo-product 159 (30 %) were formed, whereas at 1.0 GPa for 12 h only the exo-product160 (GO %) and the endo-product159 (40 %) were obtained. SOTol

-

0 Tol;$fO2Bn+ EtO2C

CHzCh

C02Et 157

-.r. t.

CO2Et

158 (endel)

126

SOTol 160 ( e x e l ) Cycloaddition of enantiopure (S)-p-tolylsulfinyl trialkoxycarbonyl ethene with cyclopentadiene. Scheme 8.40.

+ +

C02Et SOTol C02Et C02Bn

159 (endo-2)

C02Bn C02Et SOTol C02Et

161 (exo-2)

8.2 Influence $High Pressure on Selectivity

I

275

8.2.4 EnantioselectiveTransformations

The development of methods for enantioselective synthesis is an important and widely investigated area of organic chemistry. Of particular interest is the use of chiral catalysts. In the last few years numerous chiral Lewis acids and ligands for transition metal-catalyzed reactions have been developed which allow highly enantioselective transformations [ 74, 751. There are several examples showing that enantioselectivity can also be improved by applying high pressure. However, so far a general forecast is not possible. In the cases steric effects determine the enantioselectivity of a transformation, high pressure is less often employed. In enantioselective synthesis the formation of a complex of the prochiral substrate and the chiral catalyst is usually assumed. One face of the substrate is thus shielded by the catalyst and the attack of the reagent takes place from the opposite side. Under high pressure one could argue that the sterically more demanding transition structure should be favored, which means that the reagent would also attack from the side of the catalyst leading to a decrease in enantioselectivity. For the Diels-Alder reaction of (E)-3-crotonoyl-1,3-oxazolidin-2-one (163) as dienophile with isoprene (162) in the presence of the Narasaka catalyst (164) [7G] which is obtained from natural tartaric acid, such a decrease in enantioselectivity under high pressure was observed (Scheme 8.41) [77]. Thus, the reaction leads to the two enantiomeric cycloadducts 165 and 166 with an enantiomeric excess (ee) of 38 % at atmospheric pressure in favor of the (-)-enantiomer 165. However, at 500 MPa a decrease to 21 % ee was found. One could assume that a complex 167 would be formed as an intermediate (Scheme 8.42), which is attacked by isoprene preferentially from below at ambient pressure, whereas at higher pressure an attack from above is favored. Another source of influence would be the effect of pressure on the formation of the chiral Lewis acid 164. In order to quantify the effect of pressure and also of temperature on the equilibrium of the formation of 164 derived from the chiral diol 168 and the titanium compound 169, high pressure NMR spectroscopy [78] was performed. From the spectra it was possible to determine the ratio of the free to complexed diol 168 and 164, which enabled us to determine the pressure and temperature dependence of the equilibrium constant K. However, the ratio of the chiral Lewis acid 164 to the achiral 169, which can be obtained from the ratio of 164 to 168 and the starting concentrations of 168 and 169, is important for the effect of high pressure on the enantioselectivity of the Diels-Alder reaction. As expected, it was found that high pressure increases the amount of Ti(0iPr)zClz since there is a change in molecularity in the reaction. Thus, for the equilibrium of 168 and 164 a reaction volume of AVO= +(5.4 0.9) cm3 mol-' and for the pressure dependence of the quotient of 164: 169 a AVO = +(3.4 f 0.5)cm3 mol-l was obtained. At 0.1 MPa the ratio of 169:164 = 1:3.95 and at 210 MPa a ratio of 1 :2.95 was found. Thus, the decreased selectivity in the cycloaddition of 162 and 163 in the presence of the chiral catalyst 164 could also be at least partially explained by the pressure dependence of the equilibrium between 164 and 169 favoring the achiral

276

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

Lewis acid 169. It was also shown that low temperatures increase the amount o chiral titanium Lewis acid 164. At 265.2 K a ratio of 169: 164 = 1 : 2.85 and at 320.: K of 1: 1.67 was found. Furthermore, from the NMR spectra it seems evident thai the chiral Lewis acid 164 exists as a single species even under high pressure up tc 210 MPa. Looking at the different possibilities for a positive pressure effect on the enan. tioselectivity one could assume that an increase in enantioselectivity by applying high pressure is to be expected if, (1)the substrate and the reagent both coordinate to the chiral catalyst from the same side, (2) pressure improves the formation of a complex formed from the substrate and the chiral catalyst or a chiral solvent, (3) complexes of different stoichiometry are formed, which show an improved enantioselectivity, or (4)pressure has some positive effect on the formation of the chiral catalyst from a chiral ligand and a metal compound.

Ph’

162

‘Ph

165

Pressure [MPa] [165/166] Time [h] 0.1 200 300 500

(+I

(4

163

2.21 : 1 1.85: 1 1.79: 1 1.55: 1

18 3 3 3

166

Yield [“h]

ee [“h]

42 45 57 90

38 30 28 21

M V f = (1.7 k 0.2 crn3.mol-’), 25 ’C Influence o f pressure on the enantioselectivity o f Diels-Alder reactions in the presence of a chiral Lewis acid.

Scheme 8.41.

In the discussion of pressure effects on enatioselectivity one has to differentiate again between pressure and temperature effects. As already shown for diastereoselective transformations, lowering the temperature may improve the enantioselectivity. Provided that the enantioselective transformation has a large negative AVz the reaction can be performed at lower temperature under high pressure at a reasonable rate. However, this again is not a pressure effect on the enantioselectivity but a temperature effect. One of the first reactions dealing with the influence of high pressure on enantioselective transformations was described by Kraemer and Plieninger [791. They

8.2 Influence $High Pressure on Selectivity

I

277

P h x o F h OH Me

0

+ Ti(iOPr)pClp

F=x====

+,xoH

Ph Ph

PhxoF:, Me

169

O

oA~C , 12

"x

Ph Ph

168

164

+

167 A V ~= +(3.4

0.1 MPa 210 MPa

2 iPrOH

\

0.5) crn3.rno~-',23 'C

169 : 164 = 1 : 3.95 169 : 164 = 1 : 2.95

Effect o f pressure on the equilibrium of the chiral Lewis acid derived from TADDOL and on the Diels-Alder reaction.

Scheme 8.42.

investigated the Wagner-Meenvein rearrangement of 1,1-diphenyl-2-methyl-2-(4methoxypheny1)-oxirane (170) in chiral solvents (Scheme 8.43). The reaction at atmospheric pressure only yielded a racemic mixture of ketone 172 in good yields via 171 using for instance a (-)-menthol/toluene solvent mixture and the Lewis acid FeC13 as catalyst. When applying 1.0 GPa pressure, 172 was formed with an ee = 14 %. The authors argued that the asymmetric induction is caused by an improvement of the interaction of the chiral solvent with the substrate under high pressure, which is especially true for polarized transition structures. However, in OCH3

\ / 170 Scheme 8.43.

171

172

Enantioselective Wagner-Meerwein rearrangement i n chiral solvents.

278

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

this case it seems to be more likely that in the described experiment menthol reacts with FeC13 under high pressure to give a chiral Lewis acid which may be responsible for the observed enantioselectivity. The first example of a true positive high pressure effect on the enantioselectivity was found for the intramolecular hetero-Diels-Alder reaction of the l-oxa-1,3butadiene (173) in the presence of the Narasaka catalyst (164)to give the two enantiomeric bridged cycloadducts 174 and 175 (Scheme 8.44) [80]. At atmospheric pressure the two enantiomers were formed with 4.5 % ee, whereas at 500 MPa an increase to 20.4 % ee was observed which corresponds to a A A V = -(1.7 & 0.2) cm3 mol-'. In addition, the yield was improved from SO to 89 %. It was assumed that under high pressure complexes of different stoichiometry may be formed which are more favorable towards a facial selective addition. However, a clear interpretation of the results cannot be given at this point.

(-1 173

(+I

174

175

164 Pressure [MPa]

ee [%]

Time [h]

Yield ["A]

0.1 200 300 500

4.5 10.1 16.9 20.4

31 26 24 7

50 81 95 89

M V f = -(1.7 Scheme 8.44.

* 0.2 cm3.mol-'), 23 "C, CH2CI2

Enantioselective cycloaddition in the presence of a chiral Lewis acid.

A remarkable increase in enantioselectivity achieved by applying high pressure was observed by Hirama et al. [81] for an asymmetric Baylis-Hillman reaction, in addition an enhancement of the reaction rate was also found. The Baylis-Hillman reaction refers to the condensation of acrylates and aldehydes catalyzed by tertiary amines. Using various enantiopure 2,3-disubstituted 1,4-diazabicyclo[2.2.2loctanes (DABCOs) as chiral amine bases the influence of high pressure on the reaction

8.2 Influence $High Pressure on Selectivity

was investigated. When 4-nitrobenzaldehyde (176) was reacted with methyl vinyl ketone (177) using 15 mol% of the dibenzyl substituted DABCO (179) in THF at atmospheric pressure the reaction required 3 weeks to afford S-178 with 12 % ee in 66 % yield (Scheme 8.45). Under high pressure conditions of 500 MPa the reaction was complete within 12 h and the enantiomeric excess raised to 47 % ee. 179 (15 mot%)

Hydroquinone 02N

176

177

THF, 30 'C

07N 178

179 Scheme 8.45.

Enantioselective Baylis-Hillrnan reaction.

Another example of a pronounced positive pressure effect on enantioselectivity was found by Hillers and Reiser [82] for the Heck reaction of 2,3-dihydrofuran (180) and phenylpedluorobutylsulfonate (phenylnonaflate) (181) in the presence of (R)-BINAP (cf Chapter 7) Under normal pressure and at GO "C an enantiomeric excess of 47 % ee for 183 was achieved; when 1.0 GPa was applied, the enatioselectivity for 183 was improved to 89 % ee under otherwise unchanged conditions. On the other hand the ratio of 182 to 183 was only 1 :1.6 at high pressure and 182 was obtained with an enantiomeric excess of only 5 % ee. Apparently, a very effective kinetic resolution had taken place under high pressure. It should be noted that the use of new chiral ligands for the described Heck reaction of tetrahydrofuran now allows an enantioselectivity of 9G % ee with complete regiocontrol at atmospheric pressure [83]. Pd(OAC)2/(R)-BINAP CHsCNTTHF C H ~ C N ~ 111 i:i ~HF

0+NfoD NEtiPr2

180 Scheme 8.46.

181

Phf)1a2

+

P h y J 183

Enantioselective Heck reaction.

Another interesting enantioselective transformation under high pressure was reported by E. Nakamura et al. in the construction of quaternary carbon centers by allylzincation [84]. The allyl zink reagent, 185 possessing an anionic bisoxazoline (BOX) ligand was used to deliver the allyl nucleophile to the alkene 184 regio-

I

279

280

I

8 High Pressure in Organic Synthesis: Influence an Selectivity

and enantioselectively (Scheme 8.47). The reaction proceeded only sluggishly at atmospheric pressure and 25 "C to give the sterically more congested regioisomer 186 after 72 h in low yield (19 %) but high enantiomeric excess (>99.5 %). Application of 1 GPa pressure at 25 "C improved the yields substantially up to 93 % after 12 h without lowering the selectivities.

K 2) sat. NH4CI

Et 184 Scheme 8.47.

185

186

Regio- and enantiocontrolled allylzincation.

8.3 Conclusion

The application of high pressure is a very efficient and mild way to improve the reaction rate of transformations with a large negative activation volume such as cycloadditions, sigmatropic rearrangements and radical polymerizations. These reactions can be performed at lower temperatures when high pressure, as opposed to atmospheric pressure, is applied. This is especially useful for sensitive substrates and products. In addition, running a reaction at lower temperature usually leads to an improvement of the simple and induced diastereoselectivity as well as the enantioselectivity. This effect is always found if there is a reasonable difference in the activation enthalpy which leads to the isomers. On the other hand, true pressure effects on the diastereo- and enantioselectivity of chemical transformations are also known. Thus, several reactions have been described with large differences in the activation volumes of the different pathways leading to the stereoisomers, which are synthetically useful. The largest AAVf of -(7.3 0.G)cm3 mol-l found so far, was for the hetero-Diels-Alder reaction of the enaminoketone GOa and dihydrofuran 68 to give the two diastereomers G9a and 70a via an endo-E-syn and an exo-E-anti transition structure, respectively (Scheme 8.19). In this transformation the simple diastereoselectivitycan be dramatically enhanced by applying high pressure. In contrast, for the improvement of the induced diasteroselectivity using substrates with a chiral auxiliary, the situation is much less convincing at present. Thus, the observed AAV# values reported for various transformations are nearly all smaller than 1, which is not useful from a synthetic point of view. The only example of an excellent improvement in the induced diastereoselectivity achieved so far is the Michael addition to the chiral acrylate 154 (Scheme 8.39). On the other hand impressive improvements of the enantioselectivity of chemical transformations under high pressure are also known, although the mechanism of this effect is still not clearly understood.

Acknowledgements

We thank the Deutsche Forschungsgemeinschaft, the German Ministry of Education and Research, the Fonds der chemischen Industrie, the state Lower Saxony and the VW-Foundation for financial support of our research on selectivity under high pressure conditions. We are also indebted to BASF AG Ludwigshafen, Bayer AG Leverkusen, Degussa AG Frankfurt, Schering AG Berlin and Wacker Chemie GmbH Munchen for generous gifts of chemicals.

References 1 L.

F. TIETZE,A. MODI, Med. Res. Rev.

15

2000, 20, 304-322. 2

L. F. TIETZE,F. HAUNERT, In Stimulating Concepts in Chemistry, M. SHIBASAKI: J. F. STODDART, F. VOGTLE (Eds), Wiley-VCH, Weinheim, 2000, 39-64.

3 L.

F. TIETZE,Chem. Rev. 1996, 96,

115-136. 4

L. F. TIETZE,U. BEIFUSS, (a) Angew. Chem. 1993, 105, 137-170; (b) Angew. Chem. Int. Ed. Engl. 1993, 32, 131-

5

L. NIKOWA, D. SCHWARZER, J. TROE,J. SCHROEDER, J . Chem. Phys. 1992, 97,

6

T. ASANO,K. COSSTICK, H. FURUTA, K. MATSUO, H. SUMI,Bull. Chern. SOC. Jpn. 1996, 69, 551-560. G. J E N N E R , New]. Chem. 1997, 21,

163.

4827-4835.

7

1085-1090. 8 A. C. KINSMAN,M. A. KERR,Org. Lett. 2000, 2, 3517-3520. 9

53, 11929-11936.

s. HILLERS,s. SARTORI, 0. REISER,]. Am. Chem. SOC.1996, 118, 2087-2088. 17 A. B. SMITH111, T. J. BEAUCHAMP, M. j. LA MANCHE, M. D. KAUFMANN, Y. QIU, H. ARIMOTO, D. R. JONES, K. KOBAYASHI, I.Am. Chem. SOC. 2000, 122,8654-8664. 18 I. E. MARKO, G. R. EVANS,P. SERES,I. CHELLE, Z. JANOUSEK, Pure Appl. Chem. 1996, 68, 113-122. 1 9 F.-G. KIXRNER, R. EHRHARDT, H. BANDMANN, R. BOESE,D. BIXSER, K. N. HOUK,B. R. BENO,Chem. Eur. J . 1999, 5, 2119-2132. 20 For another example see E. WINTERFELDT, v. WRAY, Chem. Ber. 16

1992, 125, 2159-2161. 21

N. G. DAUBEN, C. R. KESSEL, K. H. TAKEMURA, J. Am. Chem. SOC.1980, 102, 6893-6894.

W. G. DAUBEN, J. L. LAM, Z. R. Guo, J . Org. Chem. 1996, 61,4816-4819. 11 K. MATSUMOTO, S. HASIIIMOTO, Y. IKEMI,S. OTANI,T. UCHIDA, Heterocycles 1986, 24, 1835-1839. 12 P. A. GRIECO, I. 1. NUNES, M. D. GAUL,].Am. Chem. SOC.1990, 112, 10

4595-4596. 13

T. SHIMIZU,T. MASUDA, K. HIRAMOTO, T. NAKATA, Org. Lett. 2000,

14

Y. Yu, H. FUJITA, M. OHNO,S. EGUCHI, Synthesis 1995, 498-500.

2, 650-652.

For another example see R. M. UITTENBOGAARD, 7.-P. G. SEERDEN, H. W. SCHEEREN, Tetrahedron 1997,

22

Reviews: (a) F.-G. KIXRNER,F. I. Prakt. Chem. 2000, 342, WURCHE, C. D. 609-636; (b) A. DRLJACA, HUBBARD, R. VAN ELDIK,T. ASANO, W. J. LE NOBLE, M. V. BASILEVSKY, Chem. Rev. 1998, 98, 2167-2289; (c) K. MATSUMOTO, M. KANEKO, H. KATSURA, N. HAYASHI, T. UCHIDA, R. M. ACHESON, Heterocycles 1998, 47, 1135-1178; (d) G. JENNER, Tetrahedron 1997, 53, 2669-2694. Monographs: (a) N. S. ISAACS,Liquid High Pressure Chemistry, Wdey, New York 1991; (b) W. J. LE NOBLE(Ed.), Organic High Pressure Chemistry, Elscvier, Amsterdam 1988; (c) R. M. ACIIESON, E. K. MATSUMOTO, Organic Synthesis at High Pressure, Wiley, New

282

I

8 High Pressure in Organic Synthesis: Influence on Selectivity

23 24 25

26 27 28

York 1991; (d) R. VAN ELDIK. C. D. Chemistry under Extreme HUBBARD, and Non-Classic Conditions, Wiley, New York 1997. M. K. DIEDRICH, F.-G. I
40

M. BUBACK,W. TOST,'T.HUBscIr, E. Vog, L. I:. TIETLE,Chem. Ber. 1989,

41

L. F. TIETZE, T. HUBSCH,M. BUBACK, W. TOST,High. Press. Res. 1990. 5,

122, 1179-1186.

638-640. 42

43 44

2258. 45

46

1079-1081. 29

B. M. TROST, J. R. PARQUETTI?, A. L. MARQUART, J. Am. Chem. Soc. 1995,

117, 3284-3285. 30 F. J. KEZDY, J. JAZ,A. BRUYLANTS, Bull. SOC.Chirn. Belg. 1960, 67, 602-604. 31 E. S. SWINBOURNE, J . Chem. SOC.1960, 2371-2372. 32

L. F. TIETZE, T. HUBSCH,J. OELZE, C. Om, W. TOST,G. WGRNER, hi. BUBACK, Chem. Ber. 1992, 125, 22492269.

33

M. BELIASSOUED, E. REBOUL, F. DUMAS,Tetrahedron Lett. 1997, 38, 5631-5634.

34

F.-G. KLLRNER, B. M. J. DOGAN, 0. ERMER, W. VON E. DOERING, M. P. COHEN, (a) Angew. Chem. Int. Ed. Engl. 1986, 25, 108-110; (b) Angew. Chem. 1986, 98, 109-111.

35

36

F.-G. KLLRNER,B. KRAWCZYK,V. RUSTER, U. K. DEITERS, J. Am. Chern. SOC.1994, 116, 7646-7657. L. F. TIETZE, E. Voss, Tetrahedron Lett. 1986, 27, 6181-6184.

37

L. F. TIETZE, U. HARTFIEL, T. HUBSCH, E. VoB, K. BOGDANOWICZ-SZWED, J. WICHMANN, Liebigs Ann. Chem. 1991,

38

L. F. TIETLE, U. HARTFIEL, T. HUBSCH, E. VoR, J. WICHMANN, Chew. Ber.

275-281. 47

1991, 124,881-888. 39

L. F. TIETZE,T. HUBSCH,E. VoB, M. BUBACK, W. TOST,J. Am. Chem. SOC. 1988, 110,4065-4066.

L. F. TIErzE, G. KETTSCHAU.J. A. GEWERT, A. S C ~ I U F F E N H A U Curr. ER, Org. Chem. 1998, 2, 19-62. L. F. TIETZE, G . KETTSCHAU, Top. c u r . Chem. 1997, 189, 1-120. L. F. TIETZE, T. HUBSCH,J. OELZE, C. O n , W. TOST,6 . WORNER, M. BUBACK,Chem. Ber. 1992, 125, 2249-

48

M. BUBACK, G. KUCHTA, A. NIKLAUS, M. HENRICH, I. ROTHERT,L. F. TIETZE, Liebigs Ann. 1996, 1151-1158. Since there is no clear definition for the endo- and exo-orientation for intermolecular reactions we like to use our new classification. Usually the orientation with the substituent at the dienophile under or above the diene is called endo, however, in the case where two or more different substituents at the dienophile exist, this rule can no longer be used. We therefore suggest that the following rule be applied for intermolecular Diels-Alder reactions: The orientation ofthe dienophile with the substituent hauing the highest priority according to the Cahn-Ingold-Prelog rules lying under or aboue the dime is called endo. The opposite is called exo. For hetero Diels-Alder reactions a slight modification is necessary: The orientation ofthe dienophile with the substituent being closest to the heteroatom in the diem with the highest priority according to the Cahn-Ingold-Prelog rules lying under or above the diene is called endo. For intramolecular Diels-Alder reactions the known definition should be used thus, the orientation with the chain connecting the diene and the dienophile lying under or aboue the dime is called endo. D. L. BOGER, K. D. ROBARGE,]. Org. Chern. 1988, 53, 5793-5798. L. F. TIETZE, T. HUBSCH,C. O n , G. KUCIITA, M. BUBACK, Liebigs Ann. 1995, 1-7.

49

50

51

H. AL-BADRI,N. COLLIGNON, J. MADDALUNO A N D S. MASSON, Tetrahedron 2000, 56. 3909-3919. H. AL-BADRI, I. MADDALUNO, S. ,MASSON,h’. COLLIGNON, J . Chem. Soc., Perkin Trans. 11999, 2255-2266. L. F. TIETZE. T. BRUMBY, M. PRETOR. G. REMBERG, /. Org. Chem. 1988, 53. 810-820.

52 53

N . S. ISAACS,P. v. d. BEEKE,Tetrahedron Lett. 1982, 23, 2147-2148. L. F. TIETZE, M. BRATZ:R. MACHINEK, G. v. KIEDROWSKI, J. Org. Chem. 1986,

J ~ N N E/. R ,Chem. Soc., Perkin 2 1992, 137-142. 67

1407-1413. 68

L. F. TIETZE,M. HENRICH,A.

55

NIKLAUS,M. BUBACK,Chem. Eur. J . 1999, 5, 297-304. I. SAUER,R. SUSTMANN, (a) Angm. Chem. 1980, 92, 773-801; (b) Angew. Chem. Int. Ed. Engl. 1980, 19, 779-

70 71

B. PLIETKER,D. SENG,R. FROHLICH, P. METZ, Tetrahedron 2000, 56, 873Y. RULEV,J . MADDALUNO, G. PLE, J.-C. PIAQUEVENT,L. DUHAMEL, /. Chem. Soc., Perkin Trans. 11998,

1999, 64, 4725-4732. 73

59

60 61

M. BUBACK,1. BUNGER,L. F. TIEIZE, Chem. Ber. 1992, 125, 2577-2582. 1. KIEGIEL,C. CHAPUIS,2. URBANCZYK-LIPKOWSKA, J. JURCZAK, Helu. Chim. Act. 1998, 81, 1672-1680. s. JAROSZ,E. KOZLOWSKA, A. JEZEWSKI, Tetrahedron 1997, 53, 10775-10782. H . TSUGE,T. NAGAI,T. OKANO,S. EGUCHI,H. KIMOTO,Synlett 1996, 1106-1108.

62

Y. ARAKI,T. K O N O I K E , Org. ~ . Chem.

1997, 62, 5299-5309. 63 K. FUJI,K. TAKAKA,H. ABE,K.

64

MATSUMOTO, T. HARAYAMA, A. IKEDA, T. TAM, Y. MIWA,M. NODE,/. Org. Chem. 1994, 59, 2211-2218. N. KATAGIRI,N. WATANABE, C. KANEKO,Chem. Pharm. Bull. 1990,38, 69-72.

65

L. F. TIETZE,A. SCHUFFENHAUER, Eur.

/. Org. Chem. 1998, 1629-1637. 66

B. S. EL’YANOV, G. M. GONIKBERG, G.

J. C. CARRETERO, J. L. GAR& RUANO, L. M. M A R T ~CABREJAS, N Tetrahedron 1995, 51,8323-8332.

74

75

K. A. j 0 R G E N S E N , (a) Angav. Chem. Int. Ed. 2000, 39, 3558-3588; (b) Angm. Chem. 2000, 112, 3702-3733. D. SEEBACH, A. K. BECK,A. HECKEL, (a) Angew. Chem. Int. Ed. 2001, 40, 92-138; (b)Angew. Chem. 2001, 113, 96-142.

1397-1401. 58

F. DUMAS,C. FRESSIGNE,1. LANGLET,

c. GIESSNER-PRETTRE,/. erg. Chem.

879. 57 A.

T. BUTZA N D 1. SAUER,Tetrahedron: Asymmetry 1997, 8, 703-714. G. GALLEY. M. PATZEL./. Chem. SOL, Perkin Tans. 11996. 2297-2302. F. DUMAS,B. MEZRHAB.7. D’ANGELO, C. RICHE,A. C H I A R O N I ,Org. ~ . Chem. 1996, 61, 2293-2304.

72

806. 56

L. F. TIETZE,0. BURKHARDT, M. HENRICH.Liebigs Ann./Recueil1997, 887-891.

69

52, 1638-1640. 54

L. F. TIETZE,0. BURKHARDT, M. Ann./ Recueil 1997, H E N R I C HLiebigs ,

K. NARASAKA, I. YAMAMOTO~Tetrahedron 1992, 48, 5743-5754. 77 L. F. TIETZE,C. O n , U. FREY,Liebigs Ann. 1996,63-67. 78 U. FREY,L. H E L M A. , E. MERBACH, High Press. Res. 1990, 237-245. 79 P. KRAEMER A N D H. PLIENINGER, Tetrahedron 1978, 34, 891-896. 80 L. F. TIETZE,C. O n , K. GERKE,M. BUBACK, (a) Angew. Chem. 1993, 105, 1536-1538; (b) Angew. Chem. Int. Ed. Engl. 1993, 32, 1485-1486. 81 T. O I S I ~H. , OGURI,M. HIRAMA, Tetrahedron Asymm.1995, 6, 124176

1244. 82

S. HILLERS, 0. REISER,Tetrahedron Lett. 1993, 34, 5265-5268.

L. F. TIETZE,K. THEDE,F. SANNICOLO, Chem. Comm. 1999, 1811-1812. 84 M. NAKAMURA, T. INOUE,A. SATO,E. NAKAMURA, Org. Lett. 2000. 2, 219383

2196.

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

9

High-pressure Promoted Cycloadditions for Application in Combinatorial Chemistry CeorgeJ. T. Kuster a n d H a n s W. Scheeren" 9.1

Introduction

Combinatorial chemistry has become one of the main tools for the rapid construction of molecular libraries for use in drug discovery and in biology. Important factors for the design of these libraries are structural diversity and complexity, which have to be balanced with size and purity of the individual compounds. For the construction of these libraries, solid phase chemistry has become a basic technique 111. Over the last 10 years, a burst of publications and reviews has been focused on the translation of different types of organic reactions towards the solid phase. The difficulty of transferring certain key reactions to the solid phase has stimulated solution phase library synthesis approaches which may be more effective. Cycloaddition reactions allow straightforward and often stereoselective construction of cyclic systems, which can serve as templates for further derivatization. Therefore, cycloaddition reactions play a key role in combinational chemistry sequences. The translation of cycloaddition reactions, especially 1,3-dipolar- [21 and Diels-Alder reactions [ 31 to the solid phase has been extensively studied. Although the benefits of high pressure for all type of cycloaddition reactions (e.g. [4 21; 1,3-dipolar, 12 21) have been very well illustrated in past decades, the application of high pressure to solid phase cycloaddition reactions is still in its infancy. The main advantage of using high pressure is the extension of the scope of reactions, which leads to in an increase in structural diversity and often eliminates the need for a catalyst. Lewis acid catalysts, which are used to promote all kinds of cycloaddition reactions, can cause decomposition or polymerization of one of the reaction partners thus limiting the diversity. Combining structural diversity and complexity with synthetic efficiency has led to an increased interest in multicomponent domino reactions. A multicomponent reaction is defined as a reaction in which more than two substrates, all present together, react with each other to form a product that is derived from all components in that system. Domino reactions as classified by Tietze [2b] are processes involving two or more bond-forming transformations, which take place under the same

+

+

9.2 High-pressure Diels-Alder Reactions on the Solid Phase

I

reaction conditions, without adding additional reagents and catalysts and in which the subsequent reactions occur as a consequence of the functionality formed in the previous step. Multicomponent domino reactions seem very apt for the generation of highly diversified libraries in the liquid phase as has been reviewed recently [4]. Often one of the steps in this domino reaction is a cycloaddition. A recent review by Dax et al. [Id] gives an overview of the state of the art of multicomponent reactions in solid phase synthesis. High pressure can play an important role in extending the scope of these reactions. Despite the wealth of opportunities for high-pressure applications in solid phase cycloaddition reactions and multicomponent domino reactions, only very few processes have been studied up until now. We hope that the results presented in this chapter will stimulate application of high pressure in combinational chemistry. First, some high-pressure Diels-Alder reactions in the solid phase will be described in the following paragraphs. The main part of this chapter will be devoted to high-pressure multicomponent domino cycloaddition reactions in the liquid phase and on the solid phase.

9.2 High-pressure Diels-Alder Reactions on the Solid Phase

The first high-pressure Diels-Alder reactions on the solid phase were performed in our group with resin bound acrylate 1 [5]. The acryloyl function was coupled to the hydroxybenzyl group of the Wang resin under basic conditions (Scheme 9.1).

1 Scheme 9.1

First the stability of the Wang resin under high-pressure conditions was studied. Application of a pressure of 15 kbar for 24 h did not change the swelling properties of the resin and FT-IR analysis of the resin before and after high-pressure treatment did not show significant differences. As simple acrylates in the absence of a catalyst undergo Diels-Alder reactions with furan only at high pressure, the highpressure reaction with furan was selected as a preliminary test for the effect of high pressure on solid phase cycloaddition reactions. The reaction sequence which was followed is presented in Scheme 9.2.

1 Scheme 9.2

2

0

3

285

286

I

9 High-pressure Promoted Cycloadditionsfor Application in Combinatorial Chemistry

To avoid the retro reaction of the cycloadducts (2) (endo and exo),they were converted into epoxides with MCPBA and cleavage from the resin was then achieved with TFA. Interestingly only product 3 was isolated, formed by opening an intramolecular epoxide with the carboxylic acid group of the endo cycloadduct and acylation of the OH group with acetic anhydride. In the work-up procedure the exo carboxylic acid was removed by extraction with sodium hydrogen carbonate solution [ S ) . Resin-bound nitrostyrenes (4) (for preparation see Scheme 9.29, Sect. 9.4) underwent a high-pressure Diels-Alder reaction with 2,3-dimethylbutadiene [ 61. The resin-bound cycloadduct was reductively removed from the resin yielding cis phenyl-aminocyclohexene (7) or the cyclic amino alcohol (9) depending on the reduction route followed (Scheme 9.3). The resin-bound amino acid ester (8) was further used in the synthesis of a small library of diketopiperazines [ 71.

P

-

6

7

15 kbar

[4+21 4

6

9

Scheme 9.3

9.3 High-pressure Multicomponent Domino Cycloaddition Reactions

Apart from the high-pressure multicomponent domino cycloadditions described in this section, only a few other examples of high-pressure domino reactions are mentioned in the literature. An early example of a multicomponent domino cycloaddition reaction studied under high pressure is the Ugi reaction mentioned by Matsumoto in the book, Organic Synthesis at High Pressures. In this reaction, building blocks such as acids, amines, aldehydes and isocyanides are used [8]. Recently Reiser et al. described a new high-pressure promoted domino process for the synthesis of trisubstituted alkenes combining the Horner-Wadsworth-Emmons reaction with a Heck coupling [9] (cf. Chapter 7, Scheme 7.16). Over the past 5 years we have studied high-pressure multicomponent domino cycloaddition reactions in the liquid phase and on solid supports. Multicomponent domino cycloaddition reactions can generate mono-, di- and tricyclic systems in an efficient manner depending on the reaction components used. Application of high pressure is sometimes essential for the progress of every individual step of a

9.3 High-pressure Multicomponent Domino Cycloaddition Reactions

multicomponent reaction, leading to the required structural diversity. This will be shown by the results presented below. 9.3.1

High Pressure-Promoted One-Pot Three-Component [4 + 2]/[3 + 2) Cycloadditions:

Scope and Limitations

+

The domino [4 + 2]/[3 21 cycloaddition of an enol ether, a nitroalkene and a third alkene is a representative example of a multicomponent reaction in which a polycyclic N-containing system is formed in a single transformation [lo, 111. In this domino reaction, a nitroalkene reacts as a heterodiene with an electron-rich alkene such as an enol ether, in an inverse electron-demand Diels-Alder reaction, to form a cyclic nitronate, which then reacts with another alkene in a 1,3-dipolar cycloaddition to produce a nitroso acetal (Scheme 9.4).

heterodiene

nitronate

nitroso acetal

Scheme 9.4

Four subtypes of this domino reaction are possible, in which each cycloaddition step can be intra- or intermolecular [ 12-14]. The domino intermolecular [4 2]/ intramolecular [ 3 + 21 cycloaddition has been studied extensively, mainly by Denmark and coworkers [ 121. The domino intermolecular [4+ 2]/intermolecular [ 3 + 21 cycloaddition, which has been studied less extensively [ 131, mostly leads to a lower degree of stereoselectivity than the domino inter/intra cycloadditions. From a combinational point of view however, domino inter [4 2]/inter [3 21 cycloadditions are more attractive as a high degree of structural diversity can be obtained from simple building blocks. In general the inverse electron-demand Diels-Alder reaction is carried out using stoichiometric amounts of Lewis acid “catalysts” (SnC14, TiC14, TiC12(OR)2, MAD (methylaluminium bis(2,6-di-tert-butylyI-4-methylphenoxide) and MAPh (methylaluminium bis(2,6-diphenylphenoxide)[ 151) at low temperatures (-90 to 0 “C). Before the [3 21 cycloaddition with an electron-poor alkene can take place, the first-formed nitronate has to be separated from the Lewis acid catalyst by an aqueous work-up and chromatography [13, 161. Probably complexation of the Lewis acid catalyst to the nitronate dipole inactivates the dipole and hinders the 1,3dipolar cycloadditions from taking place [ 171.

+

+

+

+

288

I

9 High-pressure Promoted Cycloadditionsfor Application in Combinatorid Chemistry

Without a Lewis acid catalyst the (4+ 21 cycloaddition of the domino [4 2j/ [3 + 21 cycloaddition requires a large excess of reagents (30-50 equiv.) [18], and long reaction times (4-20 days) [14a, 14b, 18b, 191. Tohda and coworkers described the intermolecular [4 2]/intermolecular [3 21 cycloaddition without the use of a Lewis acid catalyst, using activated nitrostyrenes and electron-poor dipolarophiles [ 19, 201. A major advantage of the non-Lewis acid catalyzed cycloaddition is the possibility of carrying out the domino [4 + 2]/[3 + 21 cycloaddition in a one-pot fashion, since electron-poor alkenes react much faster with the nitronate formed in situ than electron-rich alkenes [ 14c, 20, 211. This multicomponent reaction then provides the nitroso acetals in a single transformation, without the need to isolate the nitronate which was formed first, prior to the 1,3-dipolar cydoaddition. In order to circumvent the use of large excesses of reagents or the need to use activated nitroalkenes and to avoid long reaction times, the one-pot [4 2]/[ 3 21 cycloaddition was studied under high-pressure conditions. Furthermore, high pressure assists in overcoming steric hindrance and thus allows the use of a larger variety of highly substituted low-activated building blocks.

+

+

+

+

EtO C0,Me

[4+2]/[3+2]

Ph 10

Ph 11

12

13

Scheme 9.5

The first example of the successful application of high pressure in the one-pot [4+ 2]/[3 + 21 cycloaddition is the reaction of ethyl vinyl ether (lo), j3-nitrostyrene (11) and methyl acrylate (12) (Scheme 9.5). At ambient pressure nitroso acetal (13) was obtained in 43 % yield, after stirring for 20 days using 30 equiv. of enol ether (10) and 37 equiv. of methyl acrylate (12) [14c-d]. At 15 kbar nitroso acetal(l3) was obtained in 62 % yield within 1 h, using only 4 equiv. of 10 and 12 [14c]. The scope of the high pressure-promoted one-pot [4+ 2]/[ 3 + 21 cycloaddition will be shown in the examples presented below. p-Methoxybenzyl vinyl ether (14)is mostly used as enol ether since it mimics chemically the site of attachment of the Wang resin and therefore it could be indicative of the reactivity of a resin-bound vinyl ether. First, the reactivity of nitrostyrenes (15a and 1%) in combination with methyl acrylate and styrene as the dipolarophile was investigated (Scheme 9.6). With methyl acrylate as the dipolarophile, nitroso acetals (17 and 18) were obtained after 18 h at 15 kbar and room temperature (RT) in 95 and 78 % yield, respectively. Both nitroso acetals (17 and 18) were isolated as a mixture of three diastereomers. The products were formed from a completely regioselective 1,3-dipolar cycloaddition, which is in agreement with reported literature data on related cycloadditions of mono-substituted acrylates with nitrones and nitronates [ 221

9.3 High-pressure Multicomponent Domino Cycloaddition Reactions

14

15a R,=H

16a R,=H

17 Rl=H, R,=CO,Me

15b R,=CH,

16b Rl=CH3

18 R,=CH,R,=CO,Me 19 R,=H, R,=Ph 20 R,=CH,R,=Ph

pMB=p-rnethoxybenzyl Scheme 9.6

Similar results were obtained with the less reactive styrene as the dipolarophile yielding the domino adducts 19 and 20 in 76 and 73 % yield respectively (15 kbar, RT, 18 h) (Scheme 9.6). In this case however, both nitroso acetals were formed as mixtures of two diastereomers (ratio 4: 1). 2,3-Dihydrofuran (21) reacted with nitrostyrene (15a) and methyl acrylate yielding the tricyclic nitroso acetal (23) in 88 % yield as a mixture of three diastereomers (15 kbar, RT, 16 h, Scheme 9.7). The use of methyl-substituted nitrostyrene (1%)resulted in the formation of nitroso acetal (24) as a mixture of five diastereomers in 84 % yield (15 kbar, 50 " C , 18 h). The formation of five diastereomeric products indicated a non-selective [4 21 cycloaddition (exo and endo). The loss of endo-selectivityin the [ 4 + 21 cycloaddition was also observed when 3,4dihydropyran (22) was used. Tricyclic nitroso acetal (25) was produced in 37 % yield (15 kbar, 50 "C, 92 h) as a mixture of diastereomers resulting from exo and endo [4 + 21 cydoadditions and synlanti exo and endo 1,3-dipolar cycloadditions [7]. It is known that 3,4-dihydropyran (22) reacts much more slowly than the fivemembered 2,3-dihydrofuran (21), mainly due to the lower ring strain. This is reflected in the moderate yield of 25 and the lack of reactivity of 3,4-dihydropyran (22) towards nitrostyrene (15b) (65 h, at 15 kbar and elevated temperatures).

+

Ph 21 n=l 22 n=2

15a R=H 15b R=CH,

Phn 23 n=l , R=H 24 n=l , R=CH, 25 n=2, R=H 26 n=2, R=CH,

Scheme 9.7

1-Alkoxycyclohexenes, even when they are highly substituted, showed a much higher reactivity in the domino cycloaddition than dihydropyrans. For example, 2,6-dimethyl-l-cyclohexenyl methyl ether (27) reacted with nitrostyrene (1Sa) and

I

289

290

I

9 High-pressure Promoted Cycloadditionsfor Application in Combinatorial Chemistry

methyl acrylate to produce nitroso acetal (28) (15 kbar, RT, 18 h) as a mixture of two diastereomers (Scheme 9.8).

kMe +P 02~i.o-+ [CO,Me

15 kbar

b

C O.”O 0

2

M

e

Ph

Ph 27

15a

28

Scheme 9.8

The stereochemical limitations can be seen in the reaction presented in Scheme 9.9, in which the tetra substituted enol ether (27) reacted with the more stericallyhindered nitroalltene (151,) leading to the nitronate (29). However, in the 1,3 dipolar cycloaddition with methyl acrylate, the nitronate dipole (29) appeared to be unreactive.

27

15b

29

30

Scheme 9.9

The behavior of the 3’-pyridyl-, 3’-indolyl- and 2’-pyrrolyl-substituted p-nitroethenes (31a-c) in the domino cycloaddition was also studied [23]. At 15 b a r , 18 h and RT, the P-nitroethenes (31a-c) reacted efficiently with p-methoxybenzyl vinyl ether (14) and methyl acrylate to form the bicyclic nitronates (32a-c) (Scheme 9.10). Compounds 32a, 321, and 32c were each obtained as mixtures of three diastereomers in 74, 59 and 53 % yield respectively. The major diastereomer in each product mixture arose from a completely endo-selective [4 21 cydoaddition and an exo,anti-selective [ 3 21 cycloaddition (anti with respect to the aryl-group). The electron-releasing 3’-indole and 2’-pyrrole substituents on the nitroethene (311, and 31c) clearly had a decelerating effect on the reaction rate. Conversion of the enol ether (14)in the formation of 321, and 32c was only 80 %, under the abovementioned reaction conditions, while complete conversion of the enol ether (14) was observed in the case where the nitroalkene was substituted with a phenyl or an electron-withdrawing 3‘-pyridine substituent.

+

+

9.3 High-pressure Muhicomponent Domino Cycloaddition Reactions

-

(C0,Me

15 kbar

p

M

B

o

C0,Me

18h

Ar

Ar

14

a A r : p

w

32a-c

31a-c

c Ar:&

Scheme 9.10

The use of styrene as the dipolarophile in the reaction with either enol ether (10 or 14) and nitroalkene (31a or 31d) readily produced the corresponding nitroso acetals (33a, 331, and 33d), each as diastereomeric mixtures in 74, 65 and 77 % yield respectively (15 kbar, RT, 18 h, Scheme 9.11). aN+.0

PY 10 R,=Et

31a R,=H

14 R,=pMB

31d R,=CH,

fph

+

-

15kbar

R~o

33a R,=Et, R,=H 33b R,=pMB, R,=H 33d R,=pMB, R,=CH,

'=

Scheme 9.11

Remarkably N-phenyl maleic imide (34) reacted completely stereoselectively in the one-pot three-component reaction with 14 and 31a resulting in the formation of nitroso acetal (35) as one single diastereomer in 90 % yield (Scheme 9.12). This diastereomer arose from a completely endo-selective [4 + 21 and a completely endoselective [3 21 cycloaddition following an anti approach with respect to the pyridyl group.

+

0

Ph

34 31a

35

Scheme 9.12

9.3.1.1

Mono Substituted Dipolarophiles

Besides methyl acrylate and styrene, the reactivity of three other mono-substituted dipolarophiles in the [ 3 21 cycloaddition with nitronate (161,) was investigated.

+

I

291

292

I Vinyl acetate

9 High-Pressure Promoted Cycloadditionsfor Application in Combinatorid Chemistry

(36) and phenyl vinylsulphone (38) reacted as neutral- and electronpoor dipolarophiles respectively, with nitronate (1Gb) to produce nitroso acetals 37 and 39 (15 kbar, RT, 18 h, Scheme 9.13). In the three-component reaction of the silyl-protected allyl alcohol (40) (neutral alkene) with p-methoxybenzyl vinyl ether (14) and nitroalltene 351, (ratio 1: 1: l),nitronate 1Gb was the only product formed (15 kbar, RT, 16 h) and unreacted silylated allyl alcohol (40) was found [7]. When the reaction mixture of (1Gb) and the unreacted silylated allyl alcohol (40) was heated at 15 kbar to 50 "C for 24 h, nitroso acetal (41) was isolated in 11 % yield. The formation of nitroso acetals 37, 39 and 41 illustrates the scope of monosubstituted dipolarophiles, with respect to their electronic character and size, in these domino cycloadditions.

L-

Ph

14

'

-1

Ph

16b R=CH,

15b

36 R=OAc

Y'-

38 R=SO,Ph

37 R=OAc 39 R=SO,Ph

40 R=CH,OSi(i-Pr),

41 R=CH,OSi(i-Pr),

Scheme 9.13

9.3.1.2 Di-Substituted and Cyclic Dipolarophiles

A nice demonstration of the powerful effect of high pressure is the one-pot formation of nitroso acetals from sterically-hindered and low-activated dipolarophiles such as ,&rawsubstituted cc,P-unsaturated esters. The di-substituted acrylates 42 and 44 reacted with enol ether (14) and nitrostyrene (1Sb) at 15 kbar and 50 "C in 18 h, producing nitroso acetals 43 and 45 as diastereomeric mixtures in yields of 82 and 74 %, respectively (Scheme 9.14). Seebach and coworkers previously reported that, although some modifications of the acrylate are tolerated (e.g. cc-substitution),8-substituted cc,b-unsaturated esters did not react with nitronates in refluxing toluene [ 241.

14 Scheme 9.14

15b

42 R=Me 44 R=Ph

43 R=Me

45 R=Ph

9.3 High-pressure Multicomponent Domino Cycloaddition Reactions

I

293

In contrast to the complete regioselectivity observed in the l$dipolar cycloaddition of nitronate 1Gb and methyl crotonate 42 or methyl cinnamate 44 shown in Scheme 9.14, the [ 3 21 cycloaddition of P-nitrostyrene (15a) and nitronate intermediate 1Ga was not completely regioselective. Regio-isomers 46 and 47 were formed in 83 % yield, as mixtures of diastereomers, in a 7:3 ratio after the high pressure-promoted domino cycloaddition of enol ether 14 with 2 equiv. pnitrostyrene (15a) (15 ltbar, RT, 18 h, Scheme 9.15). The formation of regio-isomer 46 as major product was rather unexpected, since comparable 1,3-dipolar cycloadditions of nitrones and nitroalkenes [25] showed the opposite regio-isomer to be formed predominantly. This nitroso acetal (46) was converted to 8-lactam (48) via a base-catalyzed rearrangement (Scheme 9.16). This conversion appeared applicable to different bi- and tricyclic nitroso acetals and led to the formation of a novel class of bi- and tricyclic p-lactams [26].

+

Ro”o+-o

Ph

l4

Ph

16a

15a

47

46

R=p-methoxybenzyl Scheme 9.15

46

48 R,=H, R,=Et, R,=H, R,=aryl. alkyl R, -R,=(CH,), R,=H. R,=aryl, alkyl R, -R,=(CH,),, R,=CH,, R,=aryl

Scheme 9.16

The stereochemical limits of this type of domino reaction were also studied [8]. When 1-phenyl-2-nitropropene 151, was used, the only product isolated was nitronate (1Gb). Nitroalkene 15b failed to react as a dipolarophile in the second cycloaddition with nitronate 1Gb, under even more extreme conditions (15 kbar, 50 “C, 96 h, Scheme 9.17). However, nitronate 161, reacted with 8-nitrostyrene (15a) at 15 kbar, RT within 72 h completely regio- and stereoselectively producing the bicyclic nitroso acetal49.

294

I

9 High-Pressure Promoted Cycloadditions for Application in Combinatorial Chemistry

o:~co

PMBO

I++

p M B O . , , E -

GZiz [4+21

+

15 kbar

Ph 15b

14

yNOz

Ph 16b

P+21

Scheme 9.17

The cyclic cc,p-unsaturated ketone cyclohex-2-en-1-one(50) was used as building block in the one-pot domino cycloaddition of enol ether 14 and nitrostyrene 15a. At 15 kbar and 50 "C, nitroso acetal 51 was formed in 67 % yield, whereas nitroso acetal 54a was formed as a side product (Scheme 9.18). This result indicated that the 1,3-dipolar cycloaddition is still faster with the electron-poor substituted cyclohexenone 50 than with the electron-rich mono-substituted enol ether 14. The one-pot reaction of 52 with enol ether 14 and nitrostyrene 15a merely resulted in formation of nitroso acetal 54a instead of nitroso acetal 53. The unwanted side reaction was not observed in the one-pot three-component reaction with 14 and methyl-substituted nitrostyrene 1513 and 52 (Scheme 9.19). The large difference in reactivity between the three components in both the Diels-Alder and the [ 3 21 cycloaddition resulted in the formation of 55 as the main product. The side reaction of 1Gb with 14 to form 54b was prevented, since 14 was completely consumed in the reaction with 151, to give nitronate 1Gb (15 kbar, 50 "C, 16 h). However, heating (50 "C) the reaction mixture for 76 h at 15 kbar was necessary to produce nitroso acetal 55, which was formed as a mixture of two major diastereomers (ratio 3 :1)in 69 % yield.

+

)1''

+

7+.'i2b 15 kbar-

' l o w + O-N-O

OR1

50°C,18 h

Ph l4 1%

50R,=H 52 R,=CH,

R,=p-methoxybenzyl Scheme 9.18

0

51R,=H 53Rz=CH,

54a

9.3 High-Pressure Multicomponent Domino Cycloaddition Reactions

15 kbar

Rov-+h 50 "C, 92 h

50 'C, 16 h

55

Ph

15b

16b

52

Ph

RowoR

R = p-rnethoxybenzyl

54b Ph

Scheme 9.19

A one-pot synthesis of the tetra cyclic nitroso acetal 58a and 58b was achieved using methoxycyclohexene (56)as enol ether and cyclohexenone (50) as dipolarophile (Scheme 9.20). Heating of the one-pot reaction mixture at 15 kbar for 96 h up to 50 "C produced nitroso acetal 58a and 58b as a mixture of two diastereomers (3: 1) in 5 3 % yield together with a small amount of intermediate nitronate 57.

ooMe+ 80

' ~ o - +

Me0

15kbar

50 'C,92 h 56

Ph

15b

295

15kbar

Ph

14

I

50

+ 15 Kbar 50 "C,92 h

0 endo-anti-endo

58a

endo-anti-ex0

58b

Scheme 9.20

9.3.2

Novel 5,5-Membered Bi- and Tricyclic Nitroso Acetals [23]

Variation of pressure and solvent can be an additional tool to increase structural diversity as will be shown in the example below. The reaction of p-methoxybenzyl vinyl ether (14)with 3-[(E)-2-nitro-l-ethenyl] pyridine (31a) at a pressure of 15 kbar in chloroform yielded the domino adducts 59a and 59b (Scheme 9.21), in accordance with the observations described in Scheme 9.15. Nitroalkene 31a reacted as

296

I heterodiene with enol ether

9 High-Pressure Promoted Cycloadditionsfor Application in Combinatorial Chemistry

14 and the excess of 31a reacted as a dipolarophile with the [4 t 21 adduct generated in situ, yielding nitroso acetals 59a and 59b in a one-pot reaction. Since polymerization of nitroalkene 31a was one of the major side reactions at 15 kbar, the reactivity of the two-component system was investigated at lower pressures. Indeed the polymerization of the starting compound 31a was less dominant at lower pressures (7-12 kbar), but at these pressures the formation of a new stable product was also observed (mono-adduct GO, Scheme 9.22). Its formation appeared to be in competition with the formation of the bicyclic nitroso acetals 59a and 59b, as shown in Scheme 9.21.

14

60

31a PY,fNoz

py=

1

[3+2]

+ J

59a RLNO,, 59b R’=Py,

R2=Py R2=N0,

Scheme 9.21

9.3.2.1

Pressure and Solvent Dependency

A clear pressure-dependent product ratio of the reaction shown in Scheme 9.21 was observed when chloroform was used as the solvent (Table 9.1). The ratio 59:60 (nitroso acetals 59a-b :mono-adduct GO) changed from 1: 1 at a pressure of 7 kbar to 3 : 1 at 12 kbar and 6: 1 at 15 kbar. In dichloromethane the 59: GO ratio was not influenced by pressure (1: 1 ratio from 7-12 kbar). An inverse product ratio (59:GO = 1 :8) was obtained when acetone was used as solvent. Mono-adduct GO was now formed predominantly. An attempt to further increase this ratio by applying a pressure of 15 kbar merely resulted in nitroalkene polymerization. At pressures below 8 kbar, conversion of the starting compounds was not complete, although mono-adduct 60 was always formed as the major product. At 8 kbar a similar product ratio of 59 :GO = 1:4 was obtained when tetrahydrofuran was used as the solvent instead of acetone. Clearly, both the pressure and the solvent play a crucial role in the formation of the five-membered cyclic nitronate GO. The formation of five-membered cyclic nitronate GO has been explained via dipolar intermediate I (Scheme 9.22), which is formed after attack of the enol ether at the a-position. The anion is stabilized by the neighboring 3’-pyridyl substituent.

9.3 High-Pressure Multicornponent Domino Cycloaddition Reactions Tab. 9.1.

Pressure and solvent dependent product ratio

Entry

Solvent

Pressure (kbar)

CH2C12 CHzClz CHCI3 CHC13 CIIC13

7-12 15 7 12 15 15 8 8

acetone acetone

THF a

Ratio 59:60

1:l -a

1:l

3:l 6:l -

1:s 1 :4

Mainly polymerization of nitroalkene.

Ring closure of I gives the intermediate I1 and after a 1,2 proton-shift the fivemembered cycloadduct GO is formed. Since the electron-withdrawingeffect of the 3’-pyridyl group can be compared with that of a 4-nitro-phenylgroup, nitroalkene 31e was expected to react with enol ether 14 via a similar mechanism to form the five-membered cyclic nitronate 61.Indeed at a pressure of 8 kbar the reaction afforded compound 61 as the only product (Scheme 9.22). Formation of bicyclic nitroso acetals in a similar manner to 59,in which the excess of nitroalkene reacted as dipolarophile, was not observed. Electron-releasing substituents, like 3’-indolyland 2‘-pyrrolyl-and p-methoxyphenyl, will not stabilize the negative charge at the fl-position of the dipolar intermediate and therefore they will not favor the formation of the five-membered cyclic nitronate. Indeed the reaction of 3’-indolyl- and 2’-pyrrolyl-substituted8-nitroethenes 311, and 31c with enol ether 14,at a pressure of 8 kbar in acetone, did not result in the formation of a five-membered cyclic nitronate. In both cases the starting compounds were recovered. It is supposed, as summarized in Scheme 9.21, that the [4 21 product is the kinetically controlled product and that its formation is reversible. At lower pressure the [4 21 product does not react further with the nitroalkene so that the thermodynamically more stable product GO accumulates. The novel five-membered cyclic nitronate GO appeared to be useful as the dipole in the 1,3-dipolarcycloaddition reaction with several dipolarophiles, yielding a novel class of bi- and tricyclic nitroso acetals with fused five-membered rings [ 231.

+

+

o,* 0N”

14 31a R= 3’-pyridyl 31e R = p-N0,phenyl

Scheme 9.22

60 R= 3-pyridyl 61 R = p-N0,phenyl

I

297

298

I

9 High-pressure Promoted Cycloadditionsfor Application in Combinatorial Chemistry

9.3.3

High Pressure-Promoted Domino [4 1,3-Butadiene and /?-Nitrostyrene

+ 2]/[4 + 2]/[3 + 21 Cycloaddition of 2-Methoxy-

P-Nitrostyrene (15a) can react as a dienophile in the Diels-Alder reaction with 2alkoxy butadienes producing cyclic enol ethers (Scheme 9.23). By using an excess of nitrostyrene a domino reaction should take place with the in situ-generated enol ether. /I-Nitrostyrene (15a) may react subsequently as a dienophile in the DielsAlder reaction with a 2-alkoxy butadiene, as a heterodiene in the inverse DielsAlder reaction of alkoxy cyclohexene which is formed primarily, and as a disubstituted dipolarophile in the 1,3-dipolar cycloaddition of the nitronate formed in the inverse Diels-Alder reaction. 2-Methoxy-1,3-butadiene(61) was selected for the Diels-Alder reaction, since it reacted in a completely regioselective manner with nitroalkenes.

high-pressure

Scheme 9.23

In the presence of 3 equivalents of P-nitrostyrene (15a) 2-methoxy-1,3-butadiene (61) reacted to afford the nitroso acetals 64 and 65, and at 15 kbar and 50 "C the

conversion of the starting compounds was complete within 18 h (Scheme 9.24).

61

62

64 R,=NO,, R, = Ph 65 R,=Ph, R, = NO,

63

Scheme 9.24

Products 64, 65a and 65b were formed in a ratio of 1:1:3. Similar to earlier observations [ 261, a base-catalyzed intramolecular rearrangement of regio-isomer 64 took place, which led to the formation of tricyclic p-lactam 66 (Scheme 9.25). Regio-isomer 651, epimerized at the C-3 position after base treatment, yielding compound 65c while diastereomer GSa appeared to be base stable. After chromatographic purification, p-lactam 66 and the domino-adducts 65a and 6Sc were isolated in an overall yield of 78 % and a ratio of 1:4 ( 6 6 :65a, 65c). Nitroso acetal 65a arose from a completely endo-selective [4 21 cycloaddition and an exo,antiselective [3 21 cycloaddition (anti with respect to the C-4 phenyl group). Nitroso acetal65b arose from a completely endo-selective [4 21 cycloaddition and an endo,

+

+

+

9.4 High-pressure Domino [4

+ 21/13 t 21 Cycloaddition Reactions on the Solid Phase

anti-selective [ 3 + 21 cycloaddition (anti with respect to the C-4 phenyl group) [7]. The diversity of this novel domino reaction can be further increased by using different dipolarophiles.

1 P

MeO-h

base

m

0

C i H

OZN

1

base stable base

Ph

Ph

(66)/(65a)/(65c) = 1/I /3

66

65c

Scheme 9.25

9.4

High-pressure Domino [4

+ 2j/[3 + 21 Cycloaddition Reactions on the Solid Phase

Since three building Mocks are involved in the domino cycloaddition reactions, adaptation to the solid phase can be performed either with a resin-bound enol ether, a resin-bound nitroalkene or a resin-bound dipolarophile (Scheme 9.26). Some applications of the latter two possibilities will be discussed in this section ~71.

vc Q-+ 0 Ar

dienophile

heterodiene

dipolarophile

Scheme 9.26

The solid phase synthesis of nitroso acetals via a resin-bound dipolarophile will be described first. It has already been mentioned that nitronates react much faster with electron-poor alkenes than with electron-rich alkenes. The reaction of the nitronate formed in situ with the resin-bound acrylate is therefore expected to be faster than its reaction with the enol ether in solution. An acrylate was selected as dipolarophile and coupled to the resin via an ester linkage, which allows the facile cleavage of the resin-bound nitroso acetals by several methods (hydrolysis, reduc-

I

299

9 High-Pressure Promoted Cycloadditionsfor Application in Cornbinatorial Chemistry

tion, transesterification. transamidation) and simultaneously introducing an extra site of substituent diversity. Reactions were performed with the non-activated nitrostyrenes 15a (& = H) and 1 % (R4 - CH3). the mono-, di- and tri-substituted enol ethers G7a-f and the resin-bound acrylate 1 (Scheme 9.27. Table 9.2) 1281. In order to determine the yield of this domino cycloaddition. the products were cleaved from the resin by a cyanide-catalyzed transesterification reaction in a mixture of EtjN/MeOH/benzene (0.4:2: 10) (Scheme 9.28).

67

15

1

68

Scheme 9.27

Tab. 9.2.

Nitroso acetals 69a-f p r e p a r e d on s o l l d p h a s e

Products 69

c.y. 33

Ph 42

Ph 52

Ph

44

Ph

38

Ph

37

Ph

‘Yields of purified cycloadducts based on initial loading level of the Wang Resin (1.22 mmol g ’).

(%)a

9.4 High-Pressure Domino 14 t 21/13

+ 21 Cycloaddition Reactions on the Solid Phase

I

301

R,0 R2 o.N-o

v z a

Et,N/MeOH cat. KCNIBenzenk

R3

Ar

R4

69

68

Scheme 9.28

The cycloadducts 69a-f were isolated as mixtures of diastereomers. Yields after a three-step reaction sequence were reasonable (33-52 %; not optimized [29]), even in the case of higher substituted enol ethers and nitrostyrenes (Table 9.2). Resin-bound nitroalkenes were synthesized via a Knoevenagel condensation of resin-bound nitro acetic acid with aryl and alkyl substituted aldehyde (Scheme 9.29). In this way an extra site of diversity can be introduced into the cycloaddition products of these nitroalkenes. Furthermore, the resin-bound nitroalkenes can serve as activated alkenes in other cycloaddition reactions (Diels-Alder, 1,3-dipolar cycloaddition, [2 + 21 cycloaddition) and therefore lead to the solid phase synthesis of other interesting compound classes (see also Scheme 9.3, Sect. 9.2). Formation of the resin-bound nitroalkenes 73a-e was realized in one step via a microwaveassisted condensation of aldehyde 72a-e (10 equiv.) with the resin-bound nitro acetic acid 71, followed by dehydration of the intermediate /I-nitroalcohol [GI (Scheme 9.29). THF was used as the solvent in order to obtain optimal diffusion of the aldehyde in the polystyrene resin. 0

0

71

20 min. 350 W

Scheme 9.29

As shown in Table 9.3, both (hetero) aromatic and aliphatic R substituents can be introduced via the Ihoevenagel condensation. In a parallel fashion, ethyl vinyl ether (10) and styrene (74) reacted with the resin-bound nitroalkenes 73a-e at a pressure of 15 kbar for 20 h (50 "C), after which the resin-bound domino adducts 75a-e were formed (Scheme 9.30). Afier the domino [4 21/[ 3 21 cycloadditions

+

+

Re'$: OEt

15 kbar, [4+2]/[3+2]

R 7%-e

Scheme 9.30

10

Go

0

LiAIH4,

OEt

HO

0

7%-e

Ph

Ph

76a-e

302

I

9 High-Pressure Promoted Cycloadditions for Application in Combinatorid Chemistry Tab. 9.3. Synthesis of bicyclic nitroso acetals 76a-e

a

1642

51

1645

56

1667

29

d

1642

34

e

1668

48

C

H

&

Yields of purified cycloadducts based on initial loading (0.99 mmol g-’).

FT-IR analysis of each of the resins showed the disappearance of both the C=C vibration (1640 cm-l) and the C-N02 vibration (1565 cm-l). Subsequent reduction of the ester-linkage with LiA1H4 gave the 3a-methylalcohol substituted nitroso acetals 7Ga-e. In this way a variety of bicyclic nitroso acetals 7Ga-e prepared in parallel route were obtained, each as a mixture of diastereomers in an overall yield of 29-56 % over four steps (not optimized).

9.5 Conclusions and Outlook

The results described in this chapter show that high pressure can be a powerful tool with which to generate high structural diversity in domino [4 + 2]/[3 4-21 cycloaddition reactions. In contrast, the application of Lewis acid catalysts often prevents a domino reaction, as the catalyst promotes the [4 21 cycloaddition but may retard the 1,3-dipolar cycloaddition. High pressure however, promotes both cycloadditions. High pressure also leads to the extension of the scope of these reactions giving products with structural variation at nearly all positions of the nitroso acetal skeleton. Quite surprisingly, high pressure seems to have no negative effects on the accessibility or reactivity of resin-bound reactants in solid phase cycloaddition reactions. As multiple reactions can be carried out in a standard high-pressure apparatus at the same time, high pressure can also be used as a tool in the generation of small libraries of high diversity.

+

References 1 Recent reviews on solid phase

synthesis: (a) S. BOOTH,P. H. H. H. J. C. OTTENHEIIM, D. C. HERMKES, REES, Tetrahedron 1998, 54, 15385J . Chem. 15443; (b) R. C. D. BROWN, SOC.,Perkin Trans. 1 1998, 3293-3320; (c) B. A. LORSBACH,M. J. KURTH, Chem. Reu. 1999, 99, 1549-1581; (d) S. L. DAX,J. J. MCNALLY, M. A. YOUNGMAN,C u r . Med. Chem. 1999, 6, 255-270; (e) J. VAN MAARSEVEEN, Comb. Chem. @ High Throughput Screening 1998. 1; 185-214; ( f ) C. J. ANDRES, D. J. DENHART, M. S. DESHPANDE, K. W. GILLMAN, Comb. Chem. @ High Throughput Screening 1999, 2, 191-210. 2 Reviews on solid phase cycloadditions: M. J. KURTH, (a) E. J. KANTAROWSKI, Mol Diuersity 1997, 2, 207-216; (b) L. F. TIETZE,Chem. Rev. 1996, 96, 115-136. 3 Recent publications about solid phase Diels-Alder reactions: (a) B. NIE, V. M. RorErLo, J . Phys. Chem. Solids 1997, 58, 1897-1899; (b) B. R. STRANIX, G. D. DARLING,]. org. Chem. 1997, 62,9001-9004; (c) Y. WANG,S. R. WILSON,Tetrahedron Lett. 1997, 38,4021-40224; (d) M. CRAWSHAW, N. W. HIRD,K. IRIE, K. NAGAI,Tetrahedron Lett. 1997, 38, 7115-7118; (e) J. S. PANEK,B. ZHU, Patent nr. WO 9816508, Appl. nr. WO 1997-US17940 (1997-10-01).CA NO: 128:308494; ( f ) J. D. WINKLER, W. MCCOULL, Tetrahedron Lett. 1998, 39, A. D. 4935-4938; (8) S. WENDEBORN, MESMEAKER, W. K.-D. BRILL,Synlett 1998,865-868; (h) D. CRAIG,M. J. ROBSON, S. J. SHAW,Synlett 1998, T. 1381-1383; (i) T. J. SPAREY, HARRISON, Tetrahedron Lett. 1998, 39, Y.-S. 5873-5874; (1) J. D. WINKLER, KWAK, J. Org. Chem. 1998, 63,8634D. T. 8635; (k) D. A. HEERDING, J. TAKATA, C. KWON, W. F. HUFFMAN, SAMANEN, Tetrahedron Lett. 1998, 39, 6815-6818; (1) E. M. SMITH,Tetrahedron Lett. 1999, 40; 3285-3288;

(m) C. CHEN:B. MUNOZ,Tetrahedron Lett. 1999, 40, 3491-3494; (n) A. M. BOLDI,C. R. J O H N S O N , H. 0. EISSA, Tetrahedron Lett. 1999, 40, 619-622; (0)S. SUN,W. V. MURRAY. J . Org. Chem. 1999, 64, 5941-5945; (p) K. PAULVANAN, Tetrahedron Lett. 1999, 40, 1851-1854. 4 L. F. TIETZE, A. MODI,Med. Res. Ken 2000, 20, 304. 5 R. W. M. ABEN,J. W. SCHEEREN, unpublished results. 6 G. J. T. KUSTER, J. W. SCHEEREN, Tetrahedron Lett. 2000, 41, 515-519. 7 G. J. T. KUSTER,Thesis, Nijmegen 2001. 8 K. MATSUMOTO, R. MORRIN ACHESON, Organic Synthesis At High Pressures, John Wiley & Sons, New York 1991. 9 K. BODMANN,S. HAS-BECKER, 0. REISER,Phosphorous, Silicon Sulhr 1999, 144, 173-176. 10 S. E. DENMARK, A. THORARENSEN, Chem. Rev. 1996, 96, 137-165. 11 For theoretical prediction of DielsAlder and 1,3-dipolar cydoadditions to enol ethers see: J. LIU, S. NIWAYAMA, Y. You, K. N. HOUK,J. Org. Chem. 1999, 63, 1064-1073. 12 Inter 14 2]/intra [3 21: (a) S. E. DENMARK, D. L. PARKER, J. A. DIXON, J . Org. Chem. 1997, 62, 435-436; (b) S. E. DENMARK, A. R. HURD,H. J. SACHA, J . Org. Chem. 1997, 62, 1668L. R. 1745; (c) S. E. DENMARK, J. Org. Chem. 1997, 62, MARCIN, V. 1675-1686; (d) S. E. DENMARK, GUAGNANO, J. A. DIXON,A. STOLLE, J. Org. Chem. 1997, 62,4610-4628; (e) S. E. DENMARK, J. A. DIXON,J . Org. Chem. 1997, 62, 7086-7087; ( f ) S. E. DENMARK, D. S. MIDDLETON, J. Org. Chem. 1998, 63, 1604-1618; (g) S. E. DENMARK, J. A. DIXON,J . Org. Chem. 1998, 63, 6178-6195; (h) S. E. DENMARK, M. SEIERSTAD, J . Org. Chem. 1999, 64, 1610-1619; (i) S. E. DENMARK, E. A. MARTINBOROUGII, J. Am. Chem. SOC.1999, 121, 3046-3056. 13 Stepwise inter 14 2]/inter [3 2):

+

+

+

304

I

9 High-pressure Promoted Cycloadditionsfor Application in Combinatorial Chemistv

14

15 16 17

18

19

(a) S . E. DENMARK, A. K. HuRD,]. Org. Chem. 1998, 63, 3045-3050 (b) S. E. DENMARK, B. HERBERT. J . Am. Chem. Soc. 1998, 120, 7357-7358; (c) S. E. DENMARK, M. SEIERSTAD, B. HERBERT. J . Org. Chem. 1999, 64, 884-901. Inter [4 + 2]/inter [3 + 21: (a) M. AVALOS,R. BABIANO,P. CINTAS,J. L. JIMiNEZ, I.c. PALACOIS, M. A. SILVA, Chem. Commun. 1998,459-460; (b) M. AVALOS,R. BABIANO,P. CINTAS, F. J. HIGFS,J. L. J I M ~ N EJZ. C. , PALKOIS, M. A. SILVA,]. Org. Chem. 1999, 64, 1494-1502; (c) R. M. UIITENBOOGAARD, J.-P. G. SEERDEN, J. W. SCHEEREN, Tetrahedron 1997, 53, 11929-11936; (d) E. MAROITA,P. RIGHI, G. ROSINI, Tetrahedron Lett. 1999, 39, 1041-1044. S. E. DENMARK, L. R. MARC IN,^. Org. Chem. 1993, 58, 3857-3868. S. E. DENMARK, C. B. W. SENANAYAKE, Tetrahedron 1996, 52, 11579-11600. A first example of ( c h i d ) Lewis acidcatalyzed 1,3-dipolar cydoaddition of nitrones with electron-rich ketene acetals has been reported by: (a) J.-P. G. SEERDEN, A. W. A. SCHOLTEOP REIMER,J. W. SCHEEREN, Tetrahedron Lett. 1994, 35, 4419-4422; (b) J.-P. G. SEERDEN, M. M. M. KUYPERS,J. W. SCHEEREN, Tetrahedron: Asymmetry1995, 6, 1441-1450. (a) A. BARCO,S. BENETTI,G. POLLINI, G. SPALLUTO, V. ZANIRATO,Tetrahedron Lett. 1991,32,2517;(b) M. AVALOS, R. BABIANO,P. CINTAS,F. J. HIGES, J. L. J I M ~ N E J.Z ,C. PALACOIS,M. A. SILVA,J . Org. Chem. 1996, 61, 18801882; (c) A. PAPCHIKHIN,P. AGBACK, J. PLAVEC,J. CHAITOPADHYAYA, J. Org. Chem. 1993, 58, 2874-2879. Y. TOHDA,N. YAMAWAKI, H. MATSUI, T. KAWASHIMA, M. ARIGA,Y. MORI, Bull. Chem. Soc.Jpn. 1988, 61,461-465.

Y. TOHDA,K. TANI.N.YAMAWAKI.M AXIGA, N . NISHIWAKI,K. KOTANI, E. MATSUMURA, Bull. Chem. SOC.Jpn. 1993, 66, 1222-1228. 21 A. BARCO, S . BENETTI.G . POLLINI, G. SPALLUTO, V. ZANIRATO. Tetrahedron Lett. 1991, 32, 2517. 22 (a) K. B. G. TORSSELL, in Nitrile oxides 20

Nitrones and Nitronates in Organic Synthesis, H. FEUER(Ed.), VCIi Publishers, New York, 1988; (b) E. BHEUER, H. G . AURICH,A. NIELSEN, Nitrones. Nitronates and Nitroxides. Wiley, New York. 1990. 105-111. 23 G. J. T. KUSTER,R. H. J. STEEGHS, J. W. SCHEEREN, Eur.]. Org. Chem. 2001 (in press). 24 M. A. BROOK,D. SEEBACH, Can. 1. Chem. 1987. 65, 836-850. 25 (a) Review: D. C. BLACK. R. F. CROZIER, V. C. DAVIS,Synthesis 1975, D. GREE,J. 205-221; (b) M. JOUCLA, HAMELIN,Tetrahedron 1973, 29, 2315; (c) NOUREL-DIN,A H M E DMOUKHTAR Bull. Chem. SOC.Jpn. 1986, 59, 12391244. 26 G. J. T. KUSTER,F. KALMOUA. R. D E GELDER,I. W. SCHEEREN, Chem. Commun. 1999: 855-856. 27 The synthesis and use of the resinbound enol ether in the domino cycloaddition is described in G. J . T. KUSTER,Thesis. Nijmegen, 2001. 28 Without a Lewis acid catalyst, an additional activating electron withdrawing substituent on the aromatic ring of the nitrostyrene is required for the [4 + 21 cycloaddition: Y. TOHDA,N. YAMAWAKI, H. MATSUI, T. KAWASHIMA, M. ARIGA,Y. MORI. Bull. Chem. Soc. Jpn. 1988, 61, 461465. 29 G. J. T. KUSTER,I. W. SCHEEREN, Tetrahedron Lett. 1998, 39, 36133616.

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

I

305

10

Catalytic and Solvophobic Promotion o f High Pressure Addition Reactions GerardJenner 10.1 Introduction

That pressure is a fundamental physical parameter has been known since the origin of mankind. On a time scale however, high pressure chemistry is a relatively recent field, the first measurement of a reaction under pressure being reported in 1892 by Roentgen in the acid-catalyzed inversion of sucrose 111. Although high pressure techniques have now reached the maturity in many scientific fields, the application of high pressure for synthetic purposes has been confined almost exclusively within academic spheres. One of the main reasons is probably related to the small rate constant increase in the low pressure range (0-300 MPa) even for fairly pressure-dependent reactions such as pericyclic cycloadditions. The kinetic effect is derived from the relationship of Evans and Polanyi in the transition state theory as:

AV'

= -RT a(ln k ) / a p

(10.1)

AV# is the activation volume defined as the difference in partial molar volumes when the reaction progresses from initial to transition state. The estimation of the optimal pressure was previously discussed by taking into account the possible pressure dependence of AVf [2] as well as the interrelation of pressure and temperature defined under isokinetic conditions [ 31. The relationship (Eq. (10.1))underlines that the rate constant increases exponentially with pressure. The logarithmic behavior is illustrated in Fig. 10.1 which shows the variation of the rate constant ratio k,/k, with pressure at 25 "C. As an example, let us consider a pressure of 300 MPa which is usually an upper limit for large commercial pressure vessels. At that pressure the value of k,/k, approaches 10-40 for pressuresensitive reactions such as the very useful Diels-Alder reaction (AVz -25 to -40 cm3 mol-l) [4]. More negative values for AV# may occur in ionogenic reactions such as Morita-Baylis-Hillman reactions (--50 to -70 cm3 mol-') [S].

-

306

I

70 Catalytic and Soluophobic Promotion $High Pressure Addition Reactions

10

s0Q Y

K 1

/

:: 46_1_10 2

0

200 400 600 pressure / MPa

800

Fig. 10.1. Variation of k,/k,

a s a function of pressure (AV# = -40 crn3 mol-' for the upper curve A and -25 cm3 mol-' for the lower cutve B).

Although the magnitude of these rate accelerations may be of interest for many purposes, they are generally too low to efficiently activate reluctant reactions. To remove the inherent lethargy of difficult syntheses, much higher pressures must be generated, generally in the range 800-1500 MPa, in order to stimulate sluggish reactions which otherwise result in poor yields or fail utterly at lower pressures. These very high-pressure techniques obviously imply smaller working capacities and increased cost of equipment. Keeping this limitation in mind, it may be desirable to associate pressure with another mode of activation. This coupling will be called biactivation throughout this chapter. The aim of the chapter is to give the present status of the biactivation process considered as the combination of pressure (physical activation) with (a) chemical activation and (b) physicochemical activation.

10.2 Chemical Activation

This section is not concerned with transition metal catalysis. Catalytic activation of organic reactions by Lewis acids is a wide field of investigation [GI. It is usually preferred to other activation modes merely for commodity reasons. The idea of simultaneous use of pressure and traditional Lewis acid catalysts has been recognized for some time [7]. Curiously, the method was only developed in the last decade [a]. It was observed that coupling of both activation methods was beneficial in all [4 21 cycloadditions examined. However, most classical hard Lewis acids (AlC13, TiC14, SnC14, ZnClz . . .) present a number of inherent problems such as

+

70.2 Chemical Activation

I

307

Tab. 10.1.

The effect of Lewis-acid catalysis on activation volumes of Diels-Alder reactions.

Reaction

Catalyst

Solvent

A V z (cm3 mol-'1

Reference

I I I I1 I1

None AlCI, LiCI04 None AICI, None LiCIOI None ZnC12 Wone fi(fod)J None ELI( fad),

Ether Ether Ether None (bulk) None (bulk) Ether Ether Acetonitrile Acetonitrile Chloroform Chloroform Chloroform Chloroform

-36.1 -41.7 -45.4 -28.6 -25.7 -28.8 -29.3 -38.0 -33.2 -29.6 -31.7 -27.7 -31.9

12 12 12

111

III 1v IV V V VI VI

7 7 13 13 13 13 13 13 13 13

their extreme sensitivity to water which leads to inactivation of the catalyst, and their proton affinity which usually promotes polymerization of reactants and/or products if feasible. This explains why present efforts are focused on the search for catalysts which are less aggressive and also more air and water stable. Contemporaneous homogeneous catalysis has considered lithium salts [ 91, bismuth salts [lo], and lanthanides [Ill. These catalysts have been used in recent years in conjunction with pressure (vide in~5-u). The kinetic effect of the catalyst in high-pressure reactions is of ovcnvhelming importance. Does it affect the volume of activation? If yes, in what direction? An earlier paper reported AVz -values for the catalyzed Diels-Alder reaction between isoprene and N-phenyl maleic imide (reaction I in Table 10.1)(Scheme 10.1) [121.

Scheme 10.1.

Diels-Alder reaction between isoprene and N-phenylmaleic imide

Isaacs assumed a pre-equilibrium between the dienophile and the Lewis acid before cycloaddition. This means a two-stage reaction (albeit a concerted process)

308

1 implying a more negative value for AVf. In fact, the values found are in agrerment 7 0 Catalytic and Soluophobic Promotion of High Pressure Addition Reactions

with this view. They are more negative by about G and 10 cm3 mol-' whrn the catalyst is A1C13 or LiClO4 respectively, in diethyl ether. These results were, however, not completely confirmed by a subsequent paper reporting Diels-Alder (reactions 11, 1V) and hetero-Diels-Alder cycloadditions (reactions 111, v, VI) (Scheme 10.2, Table 10.1) [13]. Comparison of AV#-values for catalyzed versus uncatalyzed reactions reveal different trends. For reactions (11. IV) catalyzed by hard Lewis acids the absolute value of the volume is slightly higher than that of the uncatalyzed cycloaddition. For reaction 111, catalyzed by LiC104 AV# is similar to that of the uncatalyzed reaction in contrast to the result obtained for reaction 1. In the case of the lanthanide catalyst used in the hetero-Diels-Alder reactions V and VI, the catalyzed cycloadditions are slightly more pressure dependent than the uncatalyzed analogs (AAV# 2-4 cm3 mol-I). Taking these results into account as well as the uncertainty of AV#, it is concluded that Lewis acid catalysis does not significantly modify the magnitude of the kinetic pressure effect on [4 21 cycloadditions meaning that Lewis acids do not alter the concerted onestep mechanism of these processes. Accordingly, since k-values may be drastically enhanced at ambient pressure under the action of a catalyst, combination of pressure and Lewis acid catalysis emerges as a potent biactivation method for DielsAlder reactions involving reactant partners bearing functional groups able to coordinate with Lewis acids.

-

+

Scheme 10.2.

Catalyzed Diels-Alder reactions.

The question of whether this method of biactivation is generally applicable to organic reactions has been adressed. For this purpose, Lewis acid-catalyzed homo-

70.2 Chemical Activation

Diels-Alder (HDA, reaction VII) and [2 + 21 cycloadditions (reactions VII and VIII) were kinetically examined at various pressures (Scheme 10.3, Table 10.2) ~31.

(reaction VII)

CO,Me

+ -C02Me Scheme 10.3.

(reaction VIII)

Homo-Diels-Alder reaction and [2 + 21 cycloaddition.

Catalyzed and uncatalyzed homo-Diels-Alder reactions behave mechanistically in the same way as Diels-Alder reactions. However, in the presence of a Lewis catalyst two competing processes, the homo-Diels-Alder reaction and the [ 2 + 21 cycloaddition, take place. The smaller absolute value of AVt observed for the [2 + 21 cycloaddition is good evidence of a stepwise process. In this case, the size of the pressure effect depends on the possibility of electrostriction due to the polarity of the transition state [ 141. 10.2.1

Catalysis by Lithium Salts

Anhydrous lithium salts are soluble in organic solutions which can be considered to be air-stable catalysts. The most common is lithium perchlorate which is generally used in solution in diethyl ether (LPDE) [9], nitrornethane [15], and dichloromethane [16]. Its catalytic role comes from the properties of the lithium cation [ 171 which gives rise to specific solute-Li+ interactions modulated by complexation to appropriate solvents and counterions [ 181. A recent proposal denies lithium catalysis and emphasizes electrostatic stabilization of the transition state by LPDE [ 191. Tab. 10.2. The effect of Lewis acid catalysis on activation volumes of homo-Diels-Alder

reactions and [2 + 21 cycloadditions. Reaction

Catalyst

Solvent

A V z (cm’ rnol-’)

VII (HDA) VII (HDA) VII ([2 21) VII ([Z 21) VIII ((2 21)

None AlCl, ZrC14 AICl, AIc13

Benzene Benzene Dichloromethane Toluene Toluene

-33 -33 (estimated)” -20 (estimated)” 20 (estimated)a -20.5

+ +

+

-

‘From measurements of AAV# -13 cm3 mol-’ found for the competitive homo-Diels-Alder reaction and [ 2 - 21 cycloaddition (131.

I

309

310

I

10 Catalytic and Solvophobic Promotion of High Pressure Addition Reactions Tab. 10.3.

Cyclodimerization of isoprene. Comparison between external and internal pressure.

Pressure (hydrostatic)

k,:k,"

Xb

350 750 900

Pressure (internal)

k, :kOc

(MP4

(MP4

20.5 55 138

1M 2M 3M

400 760

4.7 7.4

1100

7.4

Rate constants determined at 83 "C in diethyl ether "x. molarity of LPDE solubons.
a

Soon after the discovery of LPDE catalysis, an interesting debate originated from the comparison of the effects of hydrostatic (external) pressure and the so-called internal pressure [9]. Based upon the comparable yields obtained in the DielsAlder reaction eventually yielding cantharidin effected either in dichloromethane at 1500 MPa or in 5 M LPDE solution at ambient pressure, it was suggested that the internal pressure of LPDE solutions was responsible for the increased reactivity in the same way as external pressure. The situation was further complicated when some authors ascribed the rate increase in Diels-Alder reactions carried out in LPDE solutions to both effects: internal pressure and Li+ catalysis [20]. The apparent analogy between external and internal pressure p, was exploited by Kumar who calculated AVf via Eq. (10.1) using the variation of rate constants with p, [21]. However, it was clearly established that such a procedure led to unrealistic values [ 221. Determination of rate constants in the cyclodimerization of isoprene, which cannot experience Li+ catalysis, showed that the reaction followed Hildebrands theory of regular solutions including LPDE solutions. Since the reaction is a neutral electron demand Diels-Alder cycloaddition, the increase of the rate constant with increasing polarity of the medium is modest. If the internal pressure of LPDE solutions acts in the same way as hydrostatic pressure, the rate acceleration should be much higher. However, the results listed in Table 10.3 demonstrate that this is not the case. p1 must be considered as a physical property of LPDE solutions like normal solvent properties such as dielectric constant, solubility parameter etc. Catalysis by LPDE was applied to some high-pressure reactions including DielsAlder cycloadditions (entries 1-4), homo-Diels-Alder reactions (entry 5) and ene reactions (entry 6) (Table 10.4) [22]. In a general way, combination of pressure and LPDE catalysis promotes the reactions. The most sensitive are the [4 21 cycloadditions involving carbonyl bonds (Table 10.4: entries 1, 2 and 3). The Diels-Alder adduct (entry 1) is obtained in quantitative yield when the substrates are reacted in 1 M LPDE under 300 MPa. The cycloaddition of the hindered 2,6-dimethylbenzoquinoneto isoprene leads to 68 % yield under LPDE conditions whereas the system is unreactive at ambient pressure in acetone solution (entry 2). An excellent yield of the 7-oxabicyclo(2.2.11 adduct is obtained in 3 M LPDE solution by reacting furan and methyl vinyl ketone at 300 MPa (entry 3). It should be emphasized that such yields are usually obtained in organic solvents only under pressures in excess of 1000-1500 MPa. A more re-

+

70.2 Chemical Activation Tab. 10.4.

Entry

High pressure synthesis in LPDE solutions (24 h except for entry 3 which was 16 h) Medium

Reaction

T

(“C) Yields (“/.I 0.1 MPa 300 MPa

1

2 3 4

5

6

Isoprene + p-benzoquinone

Acetone 1 M LPDE Isoprenc + 2,6-dirnethylbenzoquinone Acetone 1 M LPDE Ether Furan + methyl vinyl ketone 3 M LPDE Cycloheptatriene acrylonitrile Dichloromethane 2.5 M LPDE Norbornadiene + dimethylacetylene Chloroform dicarboxylate 3 M LPDE 1-Hexene + diethyl ketomalonate Dichloromethane 3 M LPDE

+

20 20

3 69

20 20 20 20 80 80 80 80 25 25

0 19 0 4 0 0 18 30

0 40

32 100 8

68 17 83 0 11

61 82 14

58

+

sistant reaction is the [4 21 norcaradiene cycloaddition between cycloheptatriene and acrylonitrile (entry 4)since LPDE is unable to catalyze the reaction at 0.1 MPa. Exposing the reactants to 300 MPa in LPDE solution induces reactivity leading to a modest 11 % yield of cycloadduct. The homo-Diels-Alder reaction (entry 5) is only slightly influenced by LPDE at constant pressure. The ene reaction (entry 6) is promoted remarkably by LPDE. However, raising the pressure to 300 MPa leads to a modest increasing yield only. Lithium perchlorate displays explosive properties. The hazards of LPDE accelerated the search for alternative lithium compounds. Lithium trifluoromethane sulfonamide was used in some Diels-Alder reactions [ 2 3 ] . The limited access to this compound prompted the consideration of other lithium salts such as lithium tetrafluoroborate (LB) and lithium triflate (LS) (241. They have seemingly similar catalytic activity to LPDE. Table 10.5 compares the catalpc merits of these lithium Tab. 10.5. The effect o f pressure on Li--catalyzed reactionsa ~

Reaction

T PC)

Yields (%j

LBC

LPb

0.1 MPa

300 MPa

0.1 MPa

300 MPa

0.1 MPa

300 MPa

91 83d 2G 26

7

85

3 0

20

IX

20

19

68

18 14d

X

80 80

30 40

82 58

5 3

XI

LSd

a Catalyst:substrate (1: l), 24 h. There is no reaction at 0.1 MPa in ether, acetonitrile or tetrahydrofuran without lithium perchlorate. Solvent (diethyl ether). Solvent (acetonitrile). Solvent (tetrahydrofuran).

10

I

311

sgp

I compounds in high-pressure Diels-Alder reactions

70 Catalytic and Solvophobrc Promotion of High Pressure Addition Reactions

(IX-XI) corresponding to en-

tries 2, 5, 6 in Table 10.4 (Scheme 10.4).

(E = C0,Me)

Scheme 10.4. Li+-catalyzed addition reactions.

Concerning reaction IX, all three lithium compounds give good results in highpressure reactions. In fact, LB and LS are even superior to LPDE. However, generalization is problematic, since in reaction X, it is exactly the reverse, the most efficient catalyst being LPDE. The same situation prevails for reaction XI with an interesting peculiarity: although the yields in the reaction catalyzed either by LB or LS are lower than those obtained in the presence of LPDE, their pressure sensitivity is higher. In conclusion of this section, organic solutions of lithium salts are a peculiar medium due to the properties of Li+ acting as a Lewis acid. The catalytic activity depends on the presence and the nature of the functional groups, on the counterion, and, as shown in Table 10.5, the type of reaction. 10.2.2

Catalysis by Lanthanides and Related Periodic Elements

The Lewis acidity of lanthanide complexes has bcen known for a long time. It was exploited extensively in their use as NMR shift reagents, mainly Eu(fod)3. They show strong affinity toward carbonyl oxygens and, therefore, have been widely used as catalysts for cycloaddition of dienes with aldehydes [25]. Moreover, the ability of catalytic amounts of lanthanide compounds to activate coordinating nitriles as well as imines has also been recognized [ 261. In recent years lanthanide (111) complexes have demonstrated clear effectiveness in catalyzing not only hetero-Diels-Alder reactions, but also Michael, aldol, Strecker and Friedel-Crafts acylation reactions 1271. In addition to lanthanides, yttrium and scandium salts, which appear in the same column of Mendelejew periodic table, were also shown to act as Lewis acid

10.2 Chemical Activation

I

313

catalysts 1281.The counterions most recognized for their catalyhc effectiveness are Eu3+,Yb3+, Sc3+,whereas the most appropiate anions are fod (Tris[2,2-dimethylG,G,7,7,8,8,8-heptafluoro-3,5-octanedionato]) and triflate (OTf) (trifluoromethanesulfonato). A very interesting property of lanthanide and scandium triflates is their stability and conservation of their catalytic power in aqueous solutions. They present another advantage in that they can be nearly 100 % recovered after reaction work-up with water and re-used several times without loss of their catalytic properties. Very recently, a highly efficient polymer-supported scandium catalyst has been designed to operate in micellar systems [29]. This makes lanthanide and scandium triflates promising catalysts in the future development of clean and environmentally-friendly processes. The synthetic potential of lanthanide catalysis has been recognized in high pressure chemistry mainly during the last decade although earlier work was concerned with the beneficial effect of the biactivation mode to overcome steric hindrance in hetero-Diels-Alder reactions [ 301. In this way, alkyl pyruvates and aldehydes considered as dienophiles could react with l-methoxy-1,3-butadiene yielding adducts which could serve as synthons in sugar synthesis [31]. As mentioned above, the kinetic pressure effect is not altered by the presence of Lewis acid catalysts. The convergence of both effects was exploited in the synthesis of dihydropyrans through the Diels-Alder reaction between a#-unsaturated keto compounds and enol ethers (Scheme 10.5) to yield dihydropyrans substituted by alkyl or phenyl groups of variable bulkiness (Table 10.6)[32].

Scheme 10.5.

Tab. 10.6.

High pressure synthesis of dihydropyrans.

Catalytic effect of Eu(fod)3 in the high pressure synthesis of dihydropyrans (Scheme

10.5)a.

RI

H Me H H Me Me

RZ

H H Me Ph Ph Me

R3

H H H H H Me

aChlorofom, 300 MPa, 1 day, catalyst (1 %). “Chloroform, 950 MPa, 1 day.

T (“C)

Yield of dihydropyran (?6) No catalyst

With Eu[f0d)j

40 30 60 40 GO

24 12 3

G6 62

100

0

0 0

45 13

21 24”

314

I

70 Catalytic and Soluophobic Promotion $High Pressure Addition Reactions

High pressure lanthanide-catalyzed synthesis of hindered dihydropyrans (Eu(fod)l 2 %). (Scheme 10.6)

Tab. 10.7.

Rz

Rl

R4

R3

T

("C)

Yield of dihydropyran

Time (h)

(Yh)

~

iPr

Me Me iPr Me Et Me -(CH2)2-

H H Me Me H

H Me I1 Me H

Et

20 20 50 50

50

85 65 37 36 23

16 16 72 72 72

The advantage of using a catalyst is obvious. When one reactive center of the keto compound is substituted by a methyl or phenyl group (Rz) the catalyst effect is more enhanced. Mesityl oxide (R1 = RZ = R3 = Me) requires very high pressure and efficient lanthanide catalysis. This reaction involving a sterically congested unsaturated ketone is a prominent example emphasizing the requirement of biactivation to overcome steric hindrance. It should be added that, apart from the mesityl oxide reaction, a modest pressure (300 MPa) is sufficient to induce reactivity in these catalyzed reactions which were shown to proceed only at pressures in excess of 1200 MPa in an uncatalyzed version [33]. Other sterically-hindered dihydropyrans can be obtained via the biactivation method from the cydoaddition of crotonaldehydes with congested en01 ethers (Scheme 10.6, Table 10.7) [34]. Very high pressures are needed to overcome steric hindrance. Interestingly, at these pressures the reactions are stereoselective with the exclusive formation of the endo stereoisomer.

R2

k""' R3

Scheme 10.6.

Eu(fad),

+

1500 MFa

Me High pressure synthesis o f hindered dihydropyrans

An approach to the synthesis of dihydroxyvitamine D3 and analogs involves the Diels-Alder reaction of the simple unsubstituted 2-pyrone with benzyl vinyl ether (Scheme 10.7). It is well known that the reaction progresses with extreme difficulty under normal conditions. Temperature is precluded as the potential cycloadduct easily decarboxylates. The biactivation mode induces sufficient reactivity as demonstrated by Posner et al. [35]. In the presence of the chiral ytterbium catalyst

'?AoBn a oo+ CoBnC0,Me

LiOMe+

0

Scheme 10.7.

HO\+

Diels-Alder access to analogs of dihydroxyvitamin D3.

OBn

10.2 Chemical Activation

I

315

Yb(tfc)3, the racemic Diels-Alder adduct - a bicyclic lactone - can be obtained in 88 % yield after exposing the reactants to 1200 MPa for 3 days. The reaction is regio- and stereoselective. The pressure kinetic effect on the Michael reaction between methyl vinyl ketone and nitromethane was studied under various conditions including lanthanide catalysis [36] (Scheme 10.8).Table 10.8 shows some kinetic results.

0Scheme 10.8.

Michael addition of nitromethane to methyl vinyl ketone.

Tab. 10.8. Activation volumes o f the Michael reaction between methyl vinyl ketone and nitromethane (calculated for T = 20 "C).

Catalyst

Solvent

None B u ~ N F3Hz0 Eu ( fod)3

Tetrahydrofuran Tetrahydrofuran

a Reaction

A V # (cm3 m o P j

Chloroform

d v (cm3 mol-'ja

No reaction 19.7 -46.5

22 -22

volume.

Compared to the reaction carried out with a traditional base the lanthanidecatalyzed reaction is extraordinarly sensitive to pressure. It is evident that the formation of the enolate, which is the rate-determining step, is strongly promoted by pressure. Since AVz is more negative than AV, electrostriction contributes largely to AV#. This implies a major shrinkage of the reaction species along the reaction axis. The lanthanide catalyst might have a dual character. The oxophilicity of Eu3+toward the carbonyl oxygen would increase the electrophilicity of the acceptor by coordinative activation. The ketone complexed by Eu(fod)3 would subsequently be attacked by the nitronate ion. It is suggested that this process is assisted by pressure in view of the strongly negative AVz value, consistent with a large charge build-up in the transition state [ 3 6 ] . A similar situation prevails in the conjugate addition of amines to a$-ethylenic esters yielding 8-aminoesters (Scheme 10.9).

Rl

R; "H

+

-

R3

\

R3F=?02Me R5 R,/N\ R, H R4

A$-?\

R2

Scheme 10.9.

Synthesis ofp-aminoesters

,C02Me R5

70 Catalytic and Solvophobic Promotion of High Pressure Addition Reactions Tab. 10.9.

Catalytic high pressure synthesis of p-arninoestersa.

Methyl ester

Acrylate Methacrylate Methacrylate Methacrylate Crotonate Crotonate Crotonate DMA' DMAC

Amine

iPr2NH tBuNH2 Ph2CHNH2 iPr(Me)NH iPr(Me)NH iBu2NH iPr2NH iPrNH2 tBuNH2

T ("C)

Yield of /I-aminoester

(?/.I

No catalyst

With Yb(0Tf ),

50 30

13 11

80

so

34b 28

1OOb

50 50

so 50 50 50

47

55

50 61

20

47

Ob

1Ob

49

78b 13b

5b

aAcetonitrile, Yb(OTf)3 (5 %), 300 MPa, 24 h. No reaction at ambient pressure without catalyst. "950 MPa. - 3,3-dimethyl acrylate.

The determination of the volume profile indicates a significant electrostriction related to the formation of zwitterions generating activation volumes oscillating between -40 to -55 cm3 mol-' [37]. At ambient pressure the reaction is usually facile when using unsubstituted primary amines and unhindered acrylates. It does not require catalysis at all on account of the basicity of amines. However, its sensitivity to steric hindrance is so high that only a combination of high pressure and lanthanide catalysis is an efficient way to synthesize congested 8-aminoesters (Table 10.9) [38]. An important reaction is the condensation of amines with IJ-diketones or 3ketoesters yielding p-enaminoesters which are important intermediates for the synthesis of natural products (Scheme 10.10). The step preceding the formation of the enamine involves zwitterionic intermediates such as those in the addition of amines to u,P-unsaturated compounds. The electrostrictive effect contributes to the magnification of the pressure effect. Lanthanide catalysis and use of 300 MPa pressure can be an excellent way to produce enamino compounds in high yields [39] (Table 10.10).

H Scheme 10.10.

Synthesis of B-enarninoesters.

There are several reports concerning the dual action of pressure and lanthanide catalysis on Michael reactions. Kotsuki reported the case of reactions of this kind which were inaccessible under ordinary conditions, for example, the homoconjugate

70.2 Chemical Activation Tab. 10.10. Condensation of ethylacetoacetate with arnines (ambient temperature, 24 h) Ri

Yield ofl-enurnino ester ("3)

R2

~~

Ph2CH Ph iPr Pr PhCH2

H H Me

No catalyst (0.7 MPu)

No catalyst (300 MPa)

Catalyzed (300 MPaJa

40

99 23 62 16 15 67

-

4

Pr

PhCH2

-(CH2)5-

1 8 6 8

92 81 70 64 82

"Yb(OTf), ( 2 %).

addition of P-ketoesters to activated cyclopropanes (Scheme 10.11) [40]. With addition of Yb(OTf)3 (20 % molar) the reactions proceed under 800 MPa at 70 "C to afford the adducts in excellent yields (generally over 75 %) depending on the ketoester. It should be mentioned that, under standard conditions at normal pressure, the reactants are completely inactive. E

0

U & Scheme 10.11.

Michael addition of p-ketoesters t o cyclopropanones.

In a previous study Kotsuki's group reported reactions of p-ketoesters with enones or acrylates under 800 MPa (Scheme 10.12) [41]. Table 10.11 compares the yields obtained under biactivation conditions (pressure lanthanide catalysis) on one side and in the presence of a catalytic system consisting of Yb(OTf)3 silicagel at ambient pressure on the other. Both methods reveal similar efficiency, although operation under pressure reduces reaction time. However, association of pressure and lanthanide catalysis proves its usefulness when acrylates are involved as Michael acceptors. In this case, no reaction occurs with the ambient pressure method even after prolonged reaction times.

+

+

E = (C0,Et) Scheme 10.12.

Michael addition of /I-ketoesters

t o enones.

In their efforts towards the total synthesis of bioactive indole alkaloids, Kerr examined the Yb(OTf)3-catalyzed addition of indoles to Michael acceptors [42]. Scheme 10.13 shows some examples of reactions involving enones and activated

I

317

318

I

70 Catalytic and Solvophobic Promotion of High Pressure Addition Reactions Tab. 10.11.

Reaction of/{-ketoesters with Michael acceptors (E = COzEt) [41].

Ketoester

Acceptor

E

rE

Yields (%)

800 MPa"

0.7 MPab

50 (1 d)

95 ( 2 d)

81 (1 d)

0 (12 d)

E 45 (1 d)

79 ( 6 d)

88 (1 d)

0 (7 4

42 (1 d)

71 ( 2 d)

G6 (1d)

79 ( 2 d)

Qo E

rE 0

0

K-.

oo oo

'Yb(OTf), (0.1 eq) in acetonitrile, 60 "C "Yb(OTf), (0.6 eq), S i 0 2 . 20 "C

cyclopropanes. The results shown in Table 10.12 indicate that there is a notable reduction in reaction time and increase in the yield of adduct when pressure is used. There is no reaction in the absence of Yb(OTf)j.

Scheme 10.13.

Michael addition of indoles.

The ring opening of epoxides with amines was investigated under several conditions including high pressure and catalysis by ytterbium triflate (Scheme 10.14) [43].It was found that the combination of pressure and Yb(OTf)3 catalysis was

70.2 Chemical Activation Reaction of N-methylindole with Michael acceptors [42]”.

Tab. 10.12.

Michael acceptor

Yields @)

0.1 MPa

COOEt

7300 MPa

3 (5 d)

34 (4d)

38 (7 d)

72 (2 d)

0

77 (4d)

84 (18 h)

1:

10 (7 d)

67 (26 h)

37 (7 d)

56 (2 d)

D(COOEt 0

Ph

acetonitrile at ambient temperature, Yb(OTf), (2 %).

more effective than the use of either lanthanide catalysis or the high pressure method independently.

Scheme 10.14.

Aminolysis of epoxides.

10.2.3

Catalysis with Other Lewis Acids

In recent years, new catalysts have been reported to effect useful transformations. Bismuth compounds such as BiC13 or Bi(OTf )3 were used in several reactions [ 441. A recent study examined the high pressure addition of lactams to trans-l-methoxy3-trimethylsilyloxy-1,3-butadiene (Danishefsky’s diene) in the presence of Lewis acid catalysts including BiC13 which was found to be the most appropriate (Scheme 10.15, Table 10.13) [45]. The adduct is of potential importance for drug chemistry.

I

319

320

I

70 Catalytic and Solvophobic Promotion of High Pressure Addition Reactions

n

Addition of6-valerolactam to Danishefsky's diene.

Scheme 10.15.

Tab. 10.13.

Addition of Danishefsky's diene to lactatma.

Lactam

Catalyst

P WPal

Yield of adduct ("A)

b-Valerolactam 6-Valerolactam 6-Valerolactam 6-Valerolactam Pyrrolidinone c-Caprolactam

None Y b(OTf ) 3 LPDE RiC13 BiC13 BiC13

0.1-300 100 100 100 300 100

0 41 0 46 19 22

a80 "C, 15 h, in toluene; catalyst (10 %)

The addition affords linear trans 1: 1 multifunctional adducts in a chemo- and regioselective manner. The selection of optimal pressure is delicate since depending on which catalyst and lactam is used, extensive polymerization can occur. An interesting example of a Lewis acid-catalyzed intramolecular Diels-Alder reaction under pressure is provided in Scheme 10.16 [46]. Not only the yield, but also the enantioselectivity is favorably influenced. The enantiomeric excess progresses from 4.5 % at ambient pressure to 20.5 % at 500 MPa whereas the yield is increased from 50 % (31 h) to 89 % (7 h) (cf. Chapter 9.2.4, Scheme 9.44).

(I:::,

pressure, Ti* chiral catalyst

0

NAO

I

I Scheme 10.16.

Asymmetric intramolecular Diels-Alder reaction.

10.2.4 Catalysis by Phosphines

Renewed interest in the Morita-Baylis-Hillman (MBH) reaction comes from recent activation processes. Traditional tertiary amines such as DABCO serve as basic catalysts (Scheme 10.17). However, the reaction is usually sluggish unless powerful activation is used.

70.3 Soluophobic Activation

K R2 R2

H

pressure

R, Scheme 10.17.

CH2X

Baylis-Hillman dirnerization of acrylic nitriles

High pressure was used to improve reactivity. Crotonaldehyde could be reacted with aldehydes in this way under DABCO catalysis [47]. A recent report examined the phosphine-catalyzeddimerization of activated alkenes under pressure (Scheme 10.17) [48]. Unhindered acrylic compounds afford high yields of Baylis-Hillman adducts at ambient pressure whereas substituted acrylic nitriles and esters require high pressure to lead to functionalized ethylenic derivatives in fair to excellent yields (Table 10.14).

10.3

Solvophobic Activation

Optimization of chemical yields can also be achieved by physicochemical activation. It is meant as an activation mode resembling catalysis and involving physicochemical interactions between reactants and medium. The role of the solvent with respect to the course of organic reactions is of obvious importance. One of the first systematic studies reporting the kinetic effect is due to Menshutkin who determined rate constants for the addition of ethyl iodide to triethylamine [49]. This is a typical example of an ionogenic reaction in which the stability of the transition

Tab. 10.14.

Ri

M B H dimerization of acrylic compounds". Rz

X

Catalyst

Yields ofdimer (?A)

0.1 MPa ~

Me Et Et Me Me Me

H H H H H H

CN CN CN C02Me C02Me CO2Et

300 MPa

~

PBu~ PBu~ P[NMe213 PBu~ P"Me213 WMe213

"Bulk. 50 "C,1 day, phosphine (5-30 % depending on run).

7 0 8 0 0 0

100 86 96 0 38 45

I

321

322

I

10 Catalytic and Soluophobic Promotion of High Pressure Addition Reactions Tab. 10.15.

Pressure dependence o f the dielectric constant

Medium

&

n-Heptane 1,2-Dichloroethylene Carbon disulfide Acetone Methanol Water

1.89 2.27 2.65 20.70 32.63 80.2

E

(20 "C)

lo5 (2 In &l?p) (cm' . kg-'J 4.01

7.84 5.45 8.02 7.02 4.36

state is highly sensitive to the medium. There is a change of polarity of the reaction states from initial to transition state with a volume decrease known as electrostrictive shrinkage. Electrostriction can be critical to rate enhancement under pressure. In terms of kinetic parameters it is related to the pressure dependence of the dielectric constant ( a In clap). Some representative values of this quantity are listed in Table 10.15 [50]. In addition, reaction rates may be affected in other ways. It was reported, some 20 years ago, that the rate constants of Diels-Alder reactions were increased by several orders of magnitude when traditional organic solvents were replaced by water [51]. That an organic reaction proceeds faster in aqueous solution in which organic compounds are barely soluble is at first sight rather surprising. It is clear that specific interactions between transition state and water are manifested. 10.3.1 Water and Water-like Solvents

An aqueous system which is closely related to life is one of the most important experimental systems. Water is regarded as a very unusual liquid. It has anomalous thermodynamic properties due to the existence of cavities and tetrahedrallycoordinated hydrogen bonds [ 5 2 ] . Water molecules may be considered as a continuum hydrogen-bonded network undergoing permanent topological rearrangement. This results in a fluctuating system which is affected by physicochemical and physical parameters e.g. pressure. If we envisage placing hydrophobic molecules in water the transfer of such molecules from a pure state to an aqueous solution is an unfavorable process due to the large decrease of entropy which results from the reorganization of the water molecules surrounding the solute. The region separating an immiscible liquid from water is perceived as an ill-definedborder where hydrophilic species are forced to meet hydrophobic partners. The relative insolubility of reactants conditions the feasibility of the aqueous reaction in which they are involved. If one substrate is completely insoluble the reaction does not occur. In contrast, if all the substrates are totally soluble there is no extra kinetic effect except the normal effect due to the

70.3 Solvophobic Activation

polarity of water. A limited solubility of reactants in aqueous solution is, therefore, a necessary condition. The chemical outcome of an organic synthesis performed in water depends on the effective concentration of reactants which addresses the problem of heterogenity. In addition, the hydration sphere must accommodate the product for as long as it forms during the progression of the reaction since the solubility of apolar compounds is attributed to the formation of a hydrophobic hydration shell. In order to extend this type of chemistry, other media have been examined, which might also favor molecular aggregation while allowing better miscibility of organic liquids or solubilization of solid reactants. Diols, particularly ethylene glycol, and forrnamide are polar solvents and possess an extensive hydrogen-bonding network like water [ 531. Their self-association could help to solubilize solvophobic moleculcs. These media are called “water-like”, although the term should be used cautiously since they present some inherent features differing from water effects such as peculiar salting-out effects (vide infia). The fact that ethylene glycol is particularly efficient while maintaining the solubility of reactants is rationalized in terms of a model where the reactants are ordered within a micelle, either in a nstacked arrangement which occupies a smaller volume, and is probably preferred, or in an “end-on’’ arrangement with a larger micelle volume and a higher energy [ 541. The accelerative effect of water and water-like media on organic reactions is generally interpreted as hydrophobic/solvophobic interactions between reactants and medium. It is an old concept which, at first sight, seems easy to rationalize. The relative insolubility of the substrates leads to an associative effect that reduces the water-hydrocarbon interfacial area and pushes the reactive partners into close proximity [55]. At the origin, the hydrophobic effect was visualized as an aggregation phenomenon of reactants increasing their local concentration and, accordingly, promoting their combination [56]. Aggregation generates a decrease in the solvent accessible surface area, thereby minimizing unfavorable interactions between the apolar reactants and water. Another interpretation persisting until recently has considered the high cohesion energy density of water which seemingly corresponds to an applied hydrostatic pressure of about 2000 MPa [ 571. The high value was taken to vindicate the kinetic analogy with the effect of water [56]. It was suggested that the rate acceleration of organic reactions in aqueous media was due to steric compression of the activated complex, minimizing the volume of cavities in the water structure [58]. The most recent opinions consider that the rate acceleration is due to a lowering of the Gibbs energy of activation under the action of both hydrophobic interactions and hydrogen-bonding effects [ 591. Engberts used the term “enforced’ hydrophobic interactions to distinguish the hydrophobic bonding of the reactants during the activation process from “normal” hydrophobic interactions which destabilize the initial state [GO, 611. Part of the rate increase is presumed to originate from this type of interaction as the hydrophobicity of reactants is relieved during the activation step. The problem of hydrophobicity is of great complexity as it may depend

I

323

324

I not only on the nature of hydrophobic groups present in the molecules, but also on 70 Catalytic and Soivophobic Promotion of H/gh Pressure Adddon Reactions

their distance from the reaction centers. This was shown in aqueous Diels-Alder reactions involving a common diene and N-alkylmaleimide where the alkyl group in the dienophile had a variable length [62]. It was concluded that hydrophobic groups near the reaction center seem to lose their hydrophobic character in the activated complex whereas more distant groups retain their non-polar character throughout the reaction. The hydrogen-bonding effect was shown to exert a determining role, sometimes dominating as exemplified in normal electron-demand Diels-Alder reactions involving ketones which are prone to establishing hydrogen bonds with water [63]. Such effects account for the extra stabilization interaction between water and the activated complex, meaning that the polarization of the transition state is higher than in the initial state (60, 641. In addition, the activated complex in water may also have a more polar character than in other solvents. To be complete, other interpretations have also surfaced such as micellar catalysis [ 56) and dipolaritypolarizability effects [65]. Numerous organic reactions have been studied in aqueous solutions. It was observed that water is able to induce dramatic rate accelerations in Diels-Alder cycloadditions [66], benzoin condensation [67], Claisen rearrangements 1681, Mukaiyama aldol reactions [57], Michael reactions [69], Baylis-Hillman reactions [ 701, and 1,3-dipolar cycloadditions [ 711. All these reactions are characterized by negative volume changes and negative volumes of activation. It is expected that ground state destabilization in aqueous media associated with transition state stabilization is one of the determining kinetic factors. 10.3.2 Kinetic Studies in Water and Water-like Solvents

Whereas a fair number of kinetic studies have been published in relation to aqueous organic reactions (Engberts’ group [60-63, 711, Breslow’s groups [66] and others [65, 72, 73]), the paucity of piezokinetic rate determinations is remarkable. To our knowledge, the pressure effect was examined only in the aqueous Diels-Alder reaction of 9-hydroxymethylanthraceneand N-phenylmaleimide [ 121. The following paragraphs report our own results in the field (references to our papers will be given throughout below). The crucial key in determining true rate constants resides in finding homogeneous or pseudo-homogeneous conditions. Accordingly, the kinetics must be followed in large volumes of water with low concentration of reactants. We present here the pressure effect on two types of reactions in aqueous solution (Michael-like reactions and Diels-Alder cycloadditions). The kinetic results are always compared to those obtained in organic solvents. Part of the following sections (mainly figures and tables) are reproduced with permission from (i) G. Jenner, Kinetic effect of pressure on Michael and DielsAlder reactions in aqueous solutions, J. Phys. Org. Chern. 1999, 12, 619-625, copyright (1999)John Wiley & Sons, Ltd and (ii) G. Jenner, R. Ben Salem, Kinetic effects

10.3 Solvophobic Activation

in water and ethylene glycol, New /. Chew. 2000, 24, 203-207, copyright (2000) The Royal Society of Chemistry and the Centre National de la Recherche Scientifique. 10.3.2.1 Michael Reactions

The reaction of nitromethane with an acceptor such as methyl vinyl ketone can proceed in water without the assistance of a base (see Scheme 10.8 in Sect. 10.2.2) [69]. The rate constant is low at ambient pressure, but is considerably increased with the aid of pressure [3G]. The slight acidity of the nitro compound makes the nitronate a suitable base. However. the formation of the enolate requires assistance. In view of the very negative value of the volume of activation (-35.5 cm3 mo1-l) it is suggested that an electrostatic volume term must be taken into account together with a volume contribution arising from the specific interaction of water with the activated complex. The conjugate addition of tert-butylamine to acrylonitrile (Scheme 10.18)is highly sensitive to pressure with an activation volume ranging from -50 to -55 cm3 mol-' in organic solvents [37, 741. Table 10.16 shows the kinetic results as well as the pressure effect on this reaction carried out in organic solvents of increasing polarity - expressed by their cohesive energy density S2 - and in aqueous solution [75]. +

tert-BuNH2 Scheme 10.18.

Y

+

tevt-Bu

CN

C -N

Michael addition of tert-butylarnine to acrylonitrile.

Inspection of Table 10.16 raises interesting comments. The rate constants are certainly related to the polarity of the solvent, however in an unorthodox way. This is clearly depicted in Fig. 10.2. There are two groups of solvents. In weakly polar aprotic media there is minimal modification of the rate constants. There is a sharp kinetic discontinuity in the second group consisting of polar protic solvents represented by methanol, water and water-like media. Likewise, the activation volume data reflect the same trends. It is nearly constant in diethyl ether, chloroform and acetonitrile (AV# -55 cm3 mol-'), reflecting

-

Tab. 10.16.

Effect of solvent and pressure in the addition of tert-butylamine to acrylonitrile.

abp/ Solvent

s2

T/K

10' k (dm3 rnol-'

Diethyl ether Chloroform Acetonitrile Methanol Ethylene glycol Formamide Water

55 86 141 208 213 369 547

317.2 317.2 317.2 300.7 300.7 300.7 300.7

0.15 0.16 0.163 6.15 8.68 16.5 82

a

Pseudo second-order rate constant at ambient pressure.

s-'la

AV+ (cm3 mol-'1

-55 -54 -56 -35 -33 -21 -25

I

325

326

I

10 Catalytic and Solvophobic Promotion $High Pressure Addition Reactions

0 0

100 200 300 400 500 600 cohesive energy density

Fig. 10.2. Effect of polarity of the medium on the rate constant ratio i n the addition of tert-butylamine t o acrylonitrile (k, is the rate constant in diethyl ether).

a high sensitivity to pressure whereas higher values are determined in the most polar solvents. In these media AVf ranges from -35 (methanol) to -21 cm3 mol-' ( formamide). The rate-determining step in the reaction mechanism involves a nucleophilic attack on the double bond of the nitrile generating zwitterionic-like species (Scheme 10.19) which undergo rapid proton transfer. The activation volume of such reactions is a composite quantity traditionally split into a structural term, AVsf and an electrostatic term, AVZ. A V , refers in the present case to the formation of zwitterions. The charge build-up is difficult in the less polar solvents leading to low rate constant values at ambient pressure. When pressure is increased electrostriction is manifested with a sharp volume decrease reflected in AV: . Highly polar solvents promote formation of zwitterionic species leading to higher rate constants with a lower pressure dependency. Since AVZ was shown to approach -20 cm3 mol-' (formation of one N-C bond in a late transition state) [ 371, the AVf -values for the addition of tert-butylamineto acrylonitrile in water or formamide would mean that AVZ is relativelly small.

Scheme 10.19. Zwitterionic intermediate in the Michael addition of tert-butylamine

The traditional treatment regarding medium effects is the dielectric continuum model in the Kirkwood-Onsager theory [76]. This simple model assumes that the solvation energy arises from the electrostatic interaction between the solutes and

10.3 Solvophobic Activation

I

327

the medium. However, this model seems inadequate for aqueous systems since hydrogen bonding involves strong attractive forces and causes a characteristic ordering of the liquid structure [SO]. In the reaction studied in Table 10.16, the solvent effect on the rate constant and the activation volume are relevant to two factors, the electrostatic contribution reflecting a spatially long-range interaction and an additional factor arising from a change in the local solvation structure around the solutes. Diels-Alder Reactions

10.3.2.2

Kinetics at Ambient Pressure

+

The [4 21 cycloaddition is the prototypical example of a reaction showing a relative insensitivity to solvent. Table 10.17 shows kinetic studies at ambient pressure for some Diels-Alder reactions carried out in different media including water and ethylene glycol. Most reactions are normal electron-demand cycloadditions, one example involves an inverse electron-demand reaction (hexachlorocyclopentadiene (HCCP) styrene) [ 771. The data listed in Table 10.17 give clear evidence of the enormous rate acceleration in aqueous solution, in accordance with all previous kinetic reports [51,61, 721, particularly for the corresponding cyclopentadiene cycloadditions [ 731. The highest values for the kinetic ratio k,/kref (> 1000) are observed for reactions involving the least hydrophilic molecules (HCCP styrene, isoprene methyl acrylate, dimerization of isoprene). Intermediate values are found in reactions involving ketones. The aqueous Diels-Alder reaction between isoprene and acrylonitrile is the least accelerated. Such rate enhancements are not compatible with a simple polarity effect. Hydrophobic interactions are obviously the major reason for the kinetic alterations.

+

+

Tab. 10.17.

Diels-Alder reactions in water and ethylene glycol at ambient pressure.

Reaction

+

+

T (K)

Furan methyl vinyl ketone Isoprene methyl vinyl ketone Isoprene + methyl acrylate Isoprene acrylonitrile Isoprene + toluquinone Isoprene (dimerization)' HCCP styrene

+

+

+

k (dm3 mol-' s-').

ks/kref

CHZCIZ

Glycol

Water

Glycol

303.3 314.4

0.8 x 1.2 x

1.2 x 10 1.3 x

5.22 x 6.91 x

15 11

650 575

335.2 338.7 303.3 353.2 323.7

7.6 x 2.3 x 3.4 x lo-' 1.3 x lo-' 2.1 x lo-'

3.8 x 1.2 x 4.5 x 9.2 x 1.7 x

7.72 x lo-' 1.7 x lo-' 1.G8 x lo-' 1.5 x 2.5 x

5 5 13 7 8

1015 75 495 1150 1200

10 10

'

lo-' 10

Pseudo second-order rate constant at ambient pressure. "Ratio of rate constants in glycol or water and CH2C12 (reference) a

respectively. CPressure(20 MPa).

Water

328

I

70 Catalytic and Solvophobic Promotion of High Pressure Addition Reactions

The ks/kref values are much lower when the corresponding reactions arc carried out in the water-like solvent, ethylene glycol. The rate enhancements are only about one order of magnitude when compared to the rate constants in dichloromethane. The highest values of the kinetic ratio are observed in the cycloadditions involving ketones as reaction partners. In the other Diels-Alder reactions the rate enhancement is two or three times lower. The ks/kref values determined in ethylene glycol do not correspond to those found by Breslow who studied the related Diels-Alder reaction of methyl vinyl ketone with cyclopentadiene: at 20 "C, ks/kr,f (reference was isooctane) was 13 in methanol, 53 in formamide, 80 in ethylene glycol and 733 in water [78]. Regarding the results in Table 10.17 the kinetic trends observed in glycol are visibly different to those obtained in water. This casts doubt on the existence of extended solvophobic interactions in this medium. The rate acceleration is better explained by simple polarity effects and hydrogen bonding between the hydroxyl groups of the diol and the carbonyl group of the ketone or quinone in their cycloaddition with furan or isoprene. As hydrophobicity is intimately related to solubility, a good method to distinguish these effects from others is the use of additives influencing the solubility. Salts such as LiCl are salting-out agents which make hydrocarbons less soluble in water whereas lithium perchlorate or urea increase the solubility (salting-in effect) and, in doing so, lower the hydrophobic effect by directly solvating hydrocarbons [79]. The results concerning the Diels-Alder reactions described in Table 10.17 are listed in Table 10.18 [77]. The salt effect is normal for the aqueous Diels-Alder reactions shown in Table 10.18 except for the furan cycloaddition carried out in aqueous LiCl solution. Consequently, the relative rate constants determined in water are indicative of hydrophobic interactions, the k values being higher in LiCl solutions and lower in LiC104 solutions. However, there are no regular trends if ethylene glycol is used as solvent. Erratic trends were also found by Breslow in the Diels-Alder reaction of 1,3-cyclohexadieneand nitrosobenzene performed in formamide and ethylene glycol [78]. In two reactions listed in Table 10.18 the reaction proceeds faster in LiC104 glycol solutions. The discrepancy would indicate weaker solvophobic interactions for ethylene glycol. In addition, the salt effect in glycol may be due to a specific lithium cation effect [17].

Tab. 10.18.

Salt effect in Diels-Alder reactions in water and glycol at ambient pressure".

Reaction

+

T (K)

Furan methyl vinyl ketone Isoprene methyl vinyl ketone Isoprene methyl acrylate HCCD styrene

+

a

+ +

303.3 314.4 335.2 323.7

Glycol

Water None

3 M LiCl

3 M LiC104

None

3 M LiC104

1 1

0.88 1.10 1.40 1.10

0.93 0.87 0.71 0.75

-

-

1

1.18 0.90 I .09

1 1

Relative rate constants. None means no additive added.

1 1

70.3 Solvophobic Activation Tab. 10.19.

Effect of the medium on the activation volume in Diels-Alder reactions.

T (K)

Reaction

1,3-Cyclohexadiene+ methyl vinyl ketone Furan methyl vinyl ketone Isoprene methyl acrylate HCCP styrene ANT N-phenylmaleimide”

+

+

a

+

+

313.2 303.8 335.3 313.7 318.2

d V z (cm3 rnol-’)

CH2CIl

CH30H

H20

-38.0 -32.4 -38.7 -35.4 -28.6b

-35.4 nd - 38.5 -33.2 -31.4b

-32.0 -28.5 -36.7 -28.0 -36.0

ANT, 9-hydroxymethylanthracene. heptane (-28.6) and n-butanol (-31.4) [12]

The Pressure Effect

The pressure effect on Diels-Alder reactions has been widely investigated and appears in all textbooks as the most characteristic example of the synthetic utility of pressure. The values of the volume of activation range typically from -25 to -40 cm3 molk’ at ambient temperature [80]. The kinetic pressure effect in some reactions in different media is illustrated in Table 10.19. Inspection of AV# values is straightforward. In most cases the activation volume is dependent on the polarity of the solvent. The AAV# differences range from 2 to 7.4 cm3 mo1-l from the least polar solvent to water. The volume differences are smaller when the medium is changed from dichloromethane to methanol. Except for the last reaction described in ref. 12, AV# increases when the organic solvent is replaced by water. This general trend implies that aqueous Diels-Alder reactions become less pressure sensitive. This tendancy is clearly portrayed in Fig. 10.3 which shows the pressure dependence of the rate constant in the cycloaddition of HCCP with styrene.

21

. 0

Y

a

Y C

-I

0

100 pressure / MPa

200

Fig. 10.3. Pressure dependence on the relative rate constant of the Diets-Alder reaction between styrene and hexachlorocyclopentadiene (A, in dichloromethane; B, in water).

I

329

330

I

70 Catalytic and Solvophobic Promotion of High Pressure Addition Reactions Effect of the medium on rate constants in the Diels-Alder reaction of isoprene and methyl vinyl ketone ( T - 313.7 K, p = 0.1 MPa).

Tab. 10.20.

Medium

d2

k (dm3 mol-'s-'j

Heptane Chloroform Dichloromethane Nitrobenzene Nitromethane Methanol 1,3-Propanediol Ethylene glycol Water

55 86 104 110 160 208 208 213 555

7.95 9.78 1.15 2.32 4.95

x 10-8 x lo-' x 10-7 x 10-7

x 10-~ 5.47 x

5.16 x 10-7 1.28 x 6.71 x lo-'

Returning to Table 10.19, the largest AAV#-values arc observed in reactions involving methyl vinyl ketone and the apolar reaction HCCP + styrene. The solvent dependence on the activation volume in the reaction between isoprene and methyl acrylate is low, but the cycloaddition of N-phenylmaleimide to ANT shows an opposite trend as AV# decreases on changing the medium from heptane to water. The interpretation of these results is given below. A detailed solvent and pressure study was reported for the Diels-Alder reaction between isoprene and methyl vinyl ketone [75]. The medium effect on the rate constant is recorded in Table 10.20. The theory of regular solutions predicts a linear correlation between In k and the cohesive energy density [81].Plotting In(k/k,) (k, is the rate constant observed in dichloromethane) leads effectively to a linear plot (Fig. 10.4). The pressure dependence of the rate constant on polarity of the medium yields the corresponding values for the volume of activation which have been plotted against d2 (Fig. 10.5).

8 water

6 g c

J

4

2 0 heptane

-2 0

100 200 300 400 500 600 cohesive energy density

Fig. 10.4. Application o f the theory of regular solutions t o the cycloaddition of methyl vinyl ketone with isoprene (T = 313.7 K) (k, = rate constant i n dichloromethane).

10.3 Solvophobic Activation

I

331

0

-36 0 .c .-F -3% Y

0

m

-40

i 0

100 200 300 400 500 600 cohesive energy density

Fig. 10.5. Solvent effect on the volume o f activation (cm3 mol ') in the cycloaddition o f methyl vinyl ketone with isoprene ( J = 313.7 K).

The AV# values range on an approximately straight line from -41.5 to -32.6 cm3 mol-'. If AV# (aqueous) follows the general trend, one should obtain a value higher than -30 cm3 mol-'. However, this is not observed since AVf is close to -34.0 cm3 mol-' in aqueous solution. In order to discriminate between diverse activation volume terms, the partial molar volumes, and then the reaction volumes were determined (Table 10.21). Although AVz visibly depends on the polarity of the solvent, this is not actually the case for The reaction volume based on partial molar volumes is almost constant (unfortunately, it was not possible to determine for the aqueous reaction due to the absence of reproducibility). The ratio 0 = A V z / m decreases with increasing polarity of the medium whereas the virtual volume of the transition state increases. These values refer to electrostriction reflected by AVc which may have importance, particularly for the least polar solvents. In this case, IAVfl > a seemingly surprising result. However, if AVZ is taken into account, which may represent 20-25 % of the overall value of AVf, 0 would be slightly <1 in agree-

n,

m.

m

In/,

Volume profile of the Diels-Alder reaction between isoprene and methyl vinyl ketone (all volumes i n c m 3 mol-').

Tab. 10.21.

Medium

AV'

2v

V+

0"

Bulk

- 41.0

-39.0 -35.0 -32.5 33.9

nd -36.1 -37.1 --38.0 ndb

nd 141.1 145.0 151.0 ndb

-

Chloroform Methanol Ethylene glycol Water

"0 = A V f /AT. "not determined.

1.08 0.94 0.86 -

332

I

10 Catalytic and Soluophobic Promotion $High Pressure Addition Reactions

ment with the 6' values determined for apolar Diels-Alder reactions [82] and confirming for the reaction studied here, a late transition state close to product. The AV# value characterizing the aqueous cycloaddition would reflect the peculiar properties of water considered as a medium for this reaction, although no anomaly could be detected in the rate diagram shown in Fig. 10.4. The results doubtless suggest specific interactions of the transition state with water such as charge-transfer complex or hydrogen bonding or other effects. As there is a decrease in the volume of reactants in the water-hydrocarbon interfacial area, it is obvious that pressure influences the solvent-accessible surface during the activation process. The AV# values (with the exception of the ANT reaction listed in Table 10.19) are indicative of a lower pressure enhancement in water. This means that pressure would exert a salting-in effect, thereby reducing the magnitude of hydrophobic interactions. Estimation of the volume term associated with this effect is difficult since it interferes with AV: . Water affects not only kinetics but also stereoselectivity [83]. An earlier study reported the medium effect on the endo selectivity of the adduct produced by cycloaddition of methyl vinyl ketone to furan under 300 MPa (Scheme 10.20, Fig. 10.6) ~341.

COMe

endo Formation o f stereoisomers i n the Diels-Alder reaction between furan and methyl vinyl ketone.

Scheme 10.20.

water

0

100 200 300 400 500 600 cohesive energy density

Fig. 10.6. Effect o f the polarity o f the medium on endo stereoselectivity in the cycloaddition of methyl vinyl ketone to furan at 300 MPa.

exo

10.3 Solvophobic Activation

I

333

Salt effect on endo selectivity in the aqueous Diels-Alder reaction of furan with methyl vinyl ketone (300 MPa, 32 "C). Tab. 10.22.

Additive

% endo

None LiCl ( 3 M ) LiC104

62

72 60

The endo preference is observed in the least polar solvents whereas the lowest endo selectivity appears in LPDE which is a highly polar medium. Figure 10.6 resembles the rate diagram of the Diels-Alder reaction shown in Fig. 10.5. A linear correlation exists between the endo content and the cohesive energy density 6*. The linear trend, however, is not followed in aqueous solution since the endo selectivity is raised to 62 % in this medium although water is more polar than LPDE. Water favors the more compact endo transition state, a result tentatively ascribed to the operation of enforced hydrophobic effects [ 83, 851. This hypothesis is argumented by salt effects (Table 10.22). The endo selectivity reaches 72 % when LiCl (saltingout agent) is added and GO % in aqueous LiC104 solution (salting-in effect). Therefore, at least in the aqueous Diels-Alder reactions involving methyl vinyl ketone as the dienophile, there is simultaneous operation of polarity and hydrophobic effects, in harmony with the results of computational studies relative to the aqueous Diels-Alder reaction of cyclopentadiene with methyl vinyl ketone [86, 871 and the Claisen rearrangement of allylvinylether [881. According to several authors [89, 901, hydrogen-bonding effects are basically responsible for the rate enhancement in water. This was apparently confirmed by theoretical calculations establishing that the kinetic factor due to enforced hydrophobic interactions was about 10, a value which is apparently independent of substrates, the remainder stemming from hydrogen bonding [ 861. There is some contradiction with the earlier results of Engberts who observed the largest rate increase in water, not in the strong hydrogen bond-donating solvent, 2,2,2-trifluoroethanol [GO]. In contrast however, a later computational technique applied to similar DielsAlder reactions involving cyclopentadiene as the diene and either isoprene or methyl vinyl ketone as the dienophile, proved that mere hydrophobic interactions are important, particularly in the reaction where no hydrogen bonding was possible (cyclopentadiene + isoprene) [ 891. In the other reaction hydrogen-bonding interactions were found to be important so that half of the rate enhancement was ascribed to each type of interaction. A recent study proposed detecting hydrogen-bonding effects via the volume of activation [91]. The variation of AVf alone when transposing the Diels-Alder cycloaddition from an organic to the aqueous medium would probably not distinguish hydrophobic interactions from hydrogen bonding since, if pressure modifies the hydrophobic surface exposed to water, both phenomena are necessarly influenced. However, if a hydrogen-bond generator is added, such as paranitrophenol (PNP),

334

I

70 Catalytic and Soluophobic Promotion of High Pressure Addition Reactions

which is known for its ability to form hydrogen bonds with keto compounds, a change in AV# could provide useful information (Scheme 10.21) [92].

Scheme 10.21. Hydrogen bonding between PNP and methyl vinyl ketone.

The addition of PNP to a dichloromethane solution of isoprene and methyl vinyl ketone results in a slightly enhanced rate constant (x 2). whereas the enhancement in water is only 1.5. The pressure kinetics was followed in both media and the resulting activation volumes were determined (Table 10.23) [ 911. Interestingly, the ratio z defined as

z= kPNP(+j/kPNP(-j

(ratio of rate constants with ( + ) and without (-) PNP added)

is relatively independent of pressure for the reaction carried out in dichloromethane whereas it increases uniformly with pressure when the cycloaddition is performed in water (Fig. 10.7). This means nearly identical AV' values in the first case, but two different values in the aqueous reaction depending on whether PNP is added (Table 10.23). According to Table 10.23, there is an additional term AAV# of about 4 cm3 mol-' when the reaction is run in the presence of PNB. This result is ascribed to enhanced hydrogen bonding and not to any change in magnitude of the enforced hydrophobic interactions due to the presence of PNB. The increase of AVz in the

0

20

40 60 80 100 120 pressure / MPa

Fig. 10.7. Pressure dependence o f the z-ratio in the cycloaddition of methyl vinyl ketone to isoprene (T = 314 K)

10.3 Solvophobic Activation Tab. 10.23.

Effect of addition of PNP on the activation volumes for some Diels-Alder reactions.

Reaction

Isoprene + niethyl vinyl ketone 1.3-Cyclohexadienetmethyl viny! ketone Isoprene + acrylonitrile a

Medium

Dichloromethane Water Water Watcr

AV'

(k1.0 cm' mol-')"

No PNP added

With PNP

-36.4 31.7 -29.9 -33.0

-36.3 -35.5 -33.5 -34.4

Standardized for T = 298.2 K

reaction carried out in the absence of PNB in water is related to a detrimental pressure effect on hydrophobic interactions. Two other aqueous Diels-Alder reactions are shown in Table 10.23. For the cycloaddition of methyl vinyl ketone to cyclohexadiene, again AAV# approaches 4 cm3 mol-' depending on whether PNP is added. When acrylonitrile is used as the dienophile AAV# is only 1.4 cm3 molt'. Hydrogen bond formation is driven by enthalpy with negative volume changes due to the shortening of the interatomic distance [93]. As highlighted in Table 10.23, the AAV# values are small since formation of hydrogen bonds involve necessarily smaller volume changes than the formation of a chemical bond. The volume change accompanied by the formation of a hydrogen bond complex with a water molecule is distributed in the range of -3 to -G cm3 mol-' for simple systems 180,931. Methyl vinyl ketone is a polarizable molecule. The hydrogen-bonding capacity of water leads to polarization of the substituting groups of the dienophile, which is roughly comparable to the activating effect of Lewis acids. In aqueous solution the functional carbonyl group is more polarized in the activated complex than in the initial state. The present results give credibility to changes in hydrogen bonding during the activation process as the source of this increased polarization. As pressure increases polarity the hydrogen-bond effect should be enhanced. For comparison, the acrylonitrile cycloaddition in water displays a AAV' of only 1.4 cm3 mol- l , a value lying inside the experimental error range and representing about half of the corresponding value found in the Diels-Alder reaction involving methyl vinyl ketone. According to jorgensen 1871, there are two to three hydrogen bonds between water molecules and the oxygen of methyl vinyl ketone leading to a greater stabilization of the transition state by a t least 3 kcal mol-', whereas nitriles form just one hydrogen bond with water in aqueous solution. Accordingly, there is still enhanced polarization in acrylonitrile. but lower than in the corresponding aqueous methyl vinyl ketone cycloadditions with a concomitant smaller enhancement of the strength of the hydrogen bond. Interpretation

From the whole body of results shown above, the pressure effect on solvophobic interactions will result in an additional activation volume term called AV'. We suggest that the quantity AVz = CAV: be expressed as

I

335

336

I

70 Catalytic and Solvophobic Promotion of High Pressure Addition Reactions

AV‘ = AV:

AV:

+ AV:

(10.2)

A%’ takes into account both enforced hydrophobic effects (AVL) and hydrogenbonding effects (AV;): (10.3)

The miscibility and the solubility of liquids in water is normally favored by pressure due to its salting-in effect [94]. This means that pressure has a detrimental effect on hydrophobic interactions (AV; > 0 ) . On the other hand, hydrogen bonding is promoted by pressure (AVHf < 0). On that basis, the results given in Tables 10.19 and 10.21 can be interpreted as follows. (i) The reactions involving methyl vinyl ketone would be characterized by A%# < 0 meaning that IAVZ I > IAVL 1. With increasing pressure the extra acceleration effect in water is mainly due to enforced hydrogen bonding. (ii) In the cycloaddition of methyl acrylate to isoprene, the apparent insensitivity of AVz to the medium may rely on the fact that hydrophobic and electrostatic effects are equally matched. (iii) In the Diels-Alder reaction between HCCP and styrene, AV: 0 and AVZ 0 since no significant hydrogen bonding can be expected, as both reactants are apolar and interact in water according to enforced painvise hydrophobic interactions [Gl]. This leads to a large rate increase when switching the medium from dichloromethane to water (Table 10.16).The AAV’ difference would be due to AVL only (+7 cm3 mol-’) leading to a positive value for AV;. (iv) The opposite trend for AV” found in the reaction between ANT and N-phenylmaleimide can also be rationalized through Eqs (10.2) and (10.3). The reactants are fairly soluble in water resulting in much less efficient hydrophobic interactions (AVL 0). Secondly, the dienophile possesses two carbony1 bonds, so enforced hydrogen bonding could develop in water resulting in a large AV; value. Finally the more negative AV# value in water can also be accounted for by solvation of the transition state as proposed previously [12].

-

-

-

10.3.2.3

Effect o f Addition o f Micelles and Cyclodextrines in Aqueous Media

Micellar media are formed from tensioactive molecules in aqueous solution. Micellization is a manifestation of the strong self-association of water and water-like solvents [95]. Micelles are known to increase the solubilization ofwealdy polar substances in water and, as a consequence, their presence determines the magnitude of hydrophobic interactions. Micelles aggregate spontaneously in aqueous solution beyond a critical concentration which is a function of pressure [9G]. As a result, pressure may induce an extra kinetic effect on the rate of organic reactions carried out in aqueous micellar systems. Representative ionic micelles are sodium dodecyl sulfate (SDS) and tetradecyltrimethylammonium bromide (TTAB). Recent examples demonstrate the beneficial effect of the presence of surfactants in Lewis acidcatalyzed reactions, a kind of biactivation [97]. Cyclodextrins are cyclic oligomers of glucose whose interior cavity can bind substrates. This binding is mainly hydrophobic, so it is strongest in water [98]. The effect of cyclodextrins is generally beneficial provided that the cavity size fits with

10.3 Sobophobic Activation Tab. 10.24.

Effect o f additives in the Diels-Alder reaction of furan and methyl vinyl ketone"

Solvent

Additive

Yield ofadduct (%A)

Dichloromethane Methanol Water Water Water Water

None None None SDS (0.73 rnmol) TTAB (0.60 rnmol) P-Cyclodextrin (0.05 mrnol)

5.3 10.3 30.3 15.5 16.0 29.7

120 MPa, 32 "C, 22 h , furan (1 mrnol), ketone (1 mmol). total volume (2.5 mL). a

the reactant structure. Table 10.24 presents some results concerning the effect of addition of micelles and 8-cyclodextrin on yields obtained in the Diels-Alder reaction between furan and methyl vinyl ketone carried out under a pressure of 120 MPa [99]. It has been observed that both micelles retard the reaction. Such results confirm that micellar catalysis - when it exists - is not the origin of the rate enhancement in aqueous solution. The addition of /I-cyclodextrin does not improve the yield although the dextrin has a cavity large enough to host the reactants and to bring the reactive centers into close proximity. Unfortunately, a general study describing the pressure effect on the kinetics of such reactions is still lacking. 10.3.3

Synthetic Applications

Taking into account the various and complex effects of solvophobic interactions, it is of interest to examine whether solvophobic properties of water and water-like solvents may be used as driving forces for synthetic purposes at high pressure. From a preparative point of view, it is clear that the reactants must resist the action of water and be present in high concentration. This leads to heterogeneous conditions at least for aqueous solutions. These conditions also prevail at higher pressure despite the salting-in effect of pressure. The influence of reactant concentration is shown in Fig. 10.8 (Diels-Alder reaction) and in Fig. 10.9 (Michael-like reaction) [ 771. In Fig. 10.8 the yield increases rapidly at low reactant concentration and stabilizes at a given value depending on reaction conditions, pressure for example. This can be easily rationalized in terms of hydrophobic interactions which are highest at the saturation limit. Fig. 10.9 shows a different behavior. For diluted solutions of methacrylonitrile the medium is pseudohomogeneous. Again, the best results are obtained at the saturation limit. However, increasing the concentration of reactants beyond this limit lowers the yield of p-aminoester. The biactivation method (activation by pressure and solvophobic interactions) was applied to three types of bimolecular reactions [ 1001: (i) [4 21 cycloadditions, AVf range from -30to -40 cm3 mol-l; (ii) Michael-like reactions, AV# range

+

I

337

338

I

10 Catalytic and Solvophobic Promotion of High Pressure Addition Reactions

200MPa

40

L

10 0,O

0,2 0,4 0,6 0,8 dienophile / mmol

1,0

Fig. 10.8. Effect of reactant concentrations in the aqueous cycloaddition o f methyl vinyl ketone t o isoprene (T = 303.2 K) Reaction time = 20 h at 0.1 MPa and 3 h at 200 Mpa. [isoprene]:[ketone] = 1.1 :1, total volume is 2.5 mL.

from -20 to -50 cm3 mol-'; (iii) Baylis-Hillman reactions, AVf range from -60 to -80 cm3 mol-'. Diels-Alder Reactions

10.3.3.1

Table 10.25 shows the results obtained at ambient pressure for a Diels-Alder reaction carried out in water and water-like solvents [77]. As evidenced in Table 10.25, water, formamide and diols are convenient media for this cycloaddition since the reaction affords 20-30 % yield of adduct within 2 h.

40

. S

30

1

lo 0

2 4 6 8 1 methacrylonitrile/ mmol

0

Effect of reactant concentrations i n the aqueous addition of tert-butylarnine to methacrylonitrile (323.2 K, 0.1 MPa, 24 h), [amine]:[nitrile] = 1.5:1, total volume is 2.5 mL.

Fig. 10.9.

70.3 Solvophobic Activation Diels-Alder reaction of toluquinone and isoprene".

Tab. 10.25.

Solvent

Yield of adduct (%)

Acetone Methanol Ethylene glycol 1,3-Propanediol Formarnide Water

0 1 24 29 21 37

a O . l MPa, 22 "C. 2 h. quinone (0.5 M except in water - 0.1 M ) and isoprene (1 MI.

There is no reaction in acetone and methanol. This is certainly related to solvophobic interactions and hydrogen bonding which should operate between the two carbonyls of the quinone and the oxygen of water or diols. Polarity effects may also be a cause of the increased reactivity although the reaction proceeds very slowly in methanol which has a similar cohesive energy density compared to ethylene glycol. The generalization of the biactivation method to other Diels-Alder reactions is highlighted in Table 10.26. Most Diels-Alder reactions are remarkably promoted at ambient pressure in aqueous solutions, particularly those involving quinones. The yields reached magnitudes of interest for chemical synthesis rivalling those obtained under pressure in organic solvents. The effect is even stronger than that of 300 MPa pressure. The cause is very probably the cooperativity of hydrogen bonding of the two carbonyls of the quinone leading to an enhanced stabilization of the transition state. The reactions are also promoted in ethylene glycol. The solutions are usually

Tab. 10.26.

High-pressure synthesis of Diels-Alder adducts in solvophobic mediaa.

Reaction

Medium

Isoprene + p-benzoquinone (5 h) Isoprene + toluquinone (2 h) Isoprene + toluquinone (2 h) Isoprene dimethylbenzoquinone (24 hj 2,s-Dimethylfuran + p-benzoquinone (24 hj 2,s-Dimethylfuran + toluquinone (24 h) Methyl vinyl ketone + ethyl vinyl ether (24 h)

+

Water Water Ethylene glycol Water Ethylene glycol Ethylene glycol Ethylene glycol

a 293.2 K, concentration o f reactants (in equimolar amounts) between 0.95 and 1.25 M. In parentheses, yield obtained under the same conditions in organic solvents (acetone or dichloromethane depending on reaction).

Yields of adducts

(?A)

0.1 M P a

300 MPab

21 37 24 12 14 0 0

82 (3) 95 (2) 85 (2) 47 (8) 59 (0) 6 (0) 28 ( 0 )

I

339

9~

I homogeneous. For normal electron-demand cycloadditions (dienophile activated by 70 Catalytic and Soluophobic Promotion of High Pressure Addition Reactions

carbony1 groups), the yield ratio is modified in nearly comparable proportion to that in water when the pressure is varied from ambient to 300 MPa. The reactivity in the glycol is due to its extensive hydrogen-bonding network (AV,’ < 0). A thorough study was made relative to the addition reactions of furan in water and water-like media [loll. Furan is a central keystone harboring two functionalities. The 14 21 cycloaddition, if feasible, leads to some interesting chemistry based on cycloaddition-retro-Did-Alder strategies [102]. However, unlike commonplace conjugated dienes, furan shows a strong reluctance to enter a cycloaddition. The zero or modest reactivity was not explained until the late 1970s when Dauben reported an efficient high pressure route permitting use of less reactive dienophiles [ 1031. In the following years several other promoting systems such as specific Lewis acids were developed [ 1041. However, one should keep in mind the detrimental effect of most Lewis acids on furan cycloadducts which either revert to the starting material or polymerize [105]. A clean synthetic method could be achieved by activation by solvophobic interactions. Very recently, it was shown that such effects could accelerate the rate of furan cycloadditions [loll. Table 10.27 lists the chemoselectivity and the yields obtained in the addition of methyl vinyl ketone to furan under various conditions (Table 10.27, Scheme 10.22).

+

Tab. 10.27.

Addition reactions of methyl vinyl ketone to furana. Catalyst yield (y

Activation

Pressure IMP4

Medium

Physical Chemical Chemical Chemical Chemical None Physicochemical Ph ysicochemical Physicochemical

300 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1

Chloroform None Chloroform ZnI2 3 M LPDE Yb(OTI- ) 3 Acetonitrile Chloroform BiC13 None Methanol None Ethylene glycol None Formamide None Water

Total

endo (“A)

A:B:C

17

74 60 33

1oo:o:o

18 63

-

20: 32: 48 0:0:100

0 18

-

-

60 (12

1oo:o:o 1oo:o:o

60

1oo:o:o

7 15

23 23

a30” C , 16 h.

E =COMe

B Scheme 10.22.

Addition reactions of methyl vinyl ketone to furan.

100:0:0

1oo:o:o

10.3 Soluophobic Activation Tab. 10.28.

High pressure synthesis of the furan cycloadduct in solvophobic media“.

Medium

Yields of cycloadducts (“A)

endo (“A)

Water Formamide Ethylene glycol

87 97 86

61 60 61

al’ressure (300 MPa), other conditions as in Table 10.27.

The reaction pathway can lead either to the expected Diels-Alder cycloadducts A or the monoadduct B or bisadduct C resulting from a Michael-type addition (Scheme 10.22). In the case of catalysis, with the exception of LPDE and ZnIz, the acidic character of Yb(OTf)3 or BiC13 diverts the reaction along both pathways or favors the exlusive formation of Michael-type products. Such chemical behavior is not uncommon in catalyzed furan reactions [ 1061. At variance with this is the uncatalyzed high pressure cycloaddition and the reaction carried out in solvophobic media at atmospheric pressure which are particularly selective and afford the DielsAlder cydoadduct A in nearly similar yields. Interestingly, the reaction also proceeds chemoselectively in water-like solvents at ambient pressure but not in hydrocarbon solvents and methanol. In water-like solvents the reactivity cannot be ascribed to polarity effects only, since methanol and glycol have similar 6* values. Solvophobic interactions are very probably mostly responsible for the enhanced reactivity. This is supported by the similar values of the endo: ex0 ratio. The modest yields reported in Table 10.27 can be remarkably increased when the reaction is carried out at 300 MPa in water and water-like media. As shown in Table 10.28, the yields are excellent for such a moderate pressure. In comparison, these yields are reached only above 1000 MPa in hydrocarbon solvents [103],making the recommended biactivation method synthetically useful. Pressure does not alter the stereoselectivity. This is in contrast with the catalyhc effect of LPDE which induces a reversal of the endo:exo ratio (cf. Fig. 1O.G). Generalization of the biactivation method (combination of pressure and solvophobic activation) to the Diels-Alder reaction of furan with various dienophiles leads to the results listed in Table 10.29. The yields are modest to fair. The lower Tab. 10.29.

Diels-Alder reactions of furan in ethylene glycol (30 “C, 60 h)

Dienophile

Acrylonitrile Methyl acrylate Acroleine Mcthacroleine Crotonaldehyde anot determined

Yields of adducts (“A)

endo (y

0.1 MPa

300 MPa

0.1 MPa

300 MPa

4

80 72 30 32 35

58

56 32 41 nd” n&

0 2 0 0

-

-

I

341

342

I

10 Catalytic and Soluophobic Promotion of High Pressure Addition Reactions Tab. 10.30.

High-pressure synthesis in solvophobic media (24 h).’.

Reaction

Medium

+

Crotononitrile tert-BuNH2 Crotononitrile + tert-BuNH2 Crotononitrile + iPr(Me)NH Crotononitrile + iPr(Me)NH Methacrylonitrile t tert-BuNHz Methacrylonitrile PrzNH Methacrylonitrile + Pr2NH Methyl methacrylate + iPr,NH Methyl methacrylate + iPr,NH Methyl methacrylate + tert-BuNH2 Methyl methacrylate tert-BuNH2

+

+

Water Glycol Water Glycol Glycol Water Glycol Water Glycol Water Glycol

T (“C)

30 30 30 30

50 50 50 50 50 30 30

Yields (%)

0.1 MPa

300 MPa

6 5 50 26 8 19 17 0

45 ( 0 ) 21 ( 0 ) 100 (0) 67 (0) 51 ( 0 ) 95 ( 2 ) 100 (2) 0 (23)

3

0 22

2 (23) 0 (11) 26 (11)

“In parentheses, yield obtained under the same conditions in acetonitrile at 300 MPa.

reactivity of the dienophiles could be ascribed to weaker hydrogen bonding between the hydroxyl groups of glycol and the C=O or C-N group of the dienophile. 10.3.3.2 Michael Reactions

The addition of amines to @ethylenic compounds was examined in water and water-like solvents from a synthetic point of view (Table 10.30) [ 77, 1001. Table 10.30 clearly shows that water and ethylene glycol promote the Michaeltype reactions. Excellent yields of p-aminonitriles are obtained in these media when cc,P-unsaturated nitriles are reacted. In all cases the yields are higher than those reached at 300 MPa in acetonitrile solution [ 1001. Notable exceptions are observed with methyl methacrylate reactions. In these reactions no Michael adduct is formed in aqueous solution, at variance with the pressure-assisted reaction which leads to modest yields (11-23 %). Such a dichotomy in reactivity is relevant to the prevalence of the reverse reaction. p-Aminoesters undergo rapid reversion to reactants in aqueous solutions which are highly polar, while they are quite stable in acetonitrile [74]. The absence of reactivity in water persists even at 300 MPa meaning that the use of higher pressure is unable to shift the equilibrium toward the aminoester. The yields of p-aminonitriles obtained in the diol at 300 MPa are comparable to those obtained in aqueous solution. However, as shown above (Sect. 10.3.2.1), ethylene glycol is a dissociating medium promoting ionogenesis in the same way as water does. The Michael reaction between amines and acrylic esters is as sluggish in the diol as it is in water. The yield is not enhanced by pressure, suggesting again a predominant reverse reaction. As an example, 65 % of the 8-aminoester synthe-

70.3 Solvophobic Activation Tab. 10.31.

Effect o f pressure on MBH reactionsa.

Reaction

Acrylonitrile (dimerization) Acrylonitrile + acetaldehyde Acrylonitrile + propionaldehyde Phenylvinylsulfone + acetaldehyde

T (“C)

35 20

20 20

Yields (“h) No solvent (neat)

Water

0.1 MPa

300 MPa

0.1 MPa

300MPa

0 58 63 0

81 100 100 100

3 38 3

3 41

24

22

4

aCatalyst:DABCO (10 %), 24 h.

sized from methyl methacrylate and diisopropylamine reverts to the starting materials after 24 h when submitted to 300 MPa and 50 “C in ethylene glycol [77]. 10.3.2.3 Baylis-Hillman Reactions

The Baylis-Hillman (also called Morita-Baylis-Hillman) (MBH) reaction (see Sect. 10.2.4 and Scheme 10.17) is the base-catalyzed addition of keto compounds to acrylic derivatives. The catalyst is a cyclic tertiary amine such as 1,4diazabicyclo[2.2.2]octane(DABCO). Due to the generally poor yields observed, the reaction has not received sufficient attention despite the great synthetic value of the polyfunctional adducts. Among the various methods proposed to activate the reaction: pressure [ 1071 and hydrophobic effects [70] have been used. Table 10.31 presents the results obtained in some Baylis-Hillman reactions carried out under pressure in aqueous solution [ 1081. In harmony with earlier papers, pressure proves to be an excellent activation mode when the reactions are carried out under neat conditions [47, 48,1071. This is not the case for reactions run in water. In fact, pressure has no effect at all on aqueous MBH reactions, invalidating the biactivation method. In these reactions hydrophobic interactions are probably manifested, however, the lower yields observed in water result presumably from the low solubility of the catalyst in water whereas the quantitative yields obtained under neat conditions at 300 MPa should be ascribed to the large AV# values. The magnitude is consistent with a large buildup of charge during the activation process with considerable concomitant electrostriction. 10.3.3.4 Triactivation

One example of triactivation has been reported [77]. It consists of a combination of pressure, chemical activation and physicochemical activation. The most significative results are listed in Table 10.32. There are notable improvements when carrying out the reactions in ethylene glycol under 300 MPa in the presence of the lanthanide catalyst. However, the generality of such a triactivation method is not evident.

I

343

344

I

10 Catalytic and Soluophobic Promotion of High Pressure Addition Reactions Tab. 10.32.

High-pressure catalyzed synthesis o f Diels-Alder adducts in ethylene glycola.

Reaction

T (“C)

Solvent

catalyst

Methyl vinyl ketone t ethyl vinyl ether Methyl vinyl ketone t ethyl vinyl ether Crotonaldehyde + ethyl vinyl ether Crotonaldehyde + ethyl vinyl ether 2,5-Dimethylfuran + toluquinone 2,s-Dimethylfuran + toluquinone

30

Chloroform Glycol Glycol

None None Yb(OTf),

100

GO

Chloroform Glycol Glycol

None None Yb( OTf ) 3

3 40 100

20

Dichloromethane Glycol Glycol

None None Yb(OTf ) 3

0

~

~~~

Yield

(“/.I

12 28

G 26

~~

“300 MPa, 24 h. bThe catalytic runs are carried out with ytterbium triflate (2.5 % molar).

10.4 Conclusions

The art of carrying out efficient chemical transformation by coupling pressure with another activation mode represents a target of modern organic synthesis and opens a wide field of applications. The two complementary activation modes examined in this chapter are very different in nature. High pressure combined with catalysis is a powerful biactivation method. In most cases moderate pressures (up to 300 MPa) can be used in difficult reactions which would otherwise require much higher pressures to go to completion, especially when congested transition states are involved. Lanthanide catalysts are particularly suitable for high-pressure experiments as they can be easily handled in the air and recovered after reaction without alteration of their catalytic properties. In addition, the mildness of these catalysts provides access to acid-sensitive adducts securing the survival of functionality in reactants and products. Pressure and solvophobic interactions can also be a valuable method to stimulate reluctant reactions. However, the effect of pressure on organic reactions in aqueous solution is complex. For isopolar reactions such as Diels-Alder cycloadditions the acceleration effect of pressure is normally lower in water than in organic solvents. For Michael reactions this effect is higher. For Baylis-Hillman reactions the effect is neutral. However, in the latter case, as water lowers the yield of adducts, the biactivation technique is not useful. These results point to the puzzling effect of water. From a synthetic point of view, it is related to the amount of hydrophobic interactions - in relation to reactant solubility - and to the magnitude of the activation volume. Volume contributions arising from structural, solvation and hydrophobic modifications during the progression of the reaction from initial to transi-

References 1345

tion state may act in divergent ways, making the prediction of the pressure effect and the interpretation difficult. In conclusion, conducting organic reactions in aqueous solution presents many advantages (water is a cheap medium, and perfectly appropriate from environmental considerations - green chemistry - it is also compatible with biomolecules, as equilibria in water are of vital importance for biological systems). However, inherent drawbacks should be outlined (the low aqueous solubility of most organic molecules creates heterogeneous conditions, water can be aggressive toward sensitive or hydrolyzable molecules, the lower pressure dependence of rate constants may not be suitable for some syntheses). The results presented in this chapter highlight the extreme utility of high pressure as a prominent parameter in organic synthesis, possibly combined with additional activation methods.

References 1 W. C . ROENTGEN, Ann. Phys. Chem. 2

3 4

5 6

7

8

9

10

11

12

13

1892, 45, 98. W. G. DAUBEN, J. M. GERDES,G. C. LOOK, Synthesis 1986, 532-535. G. TENNER, New J . Chem. 1979, 3, 329-334. T . ASANO,W. I. LE NOBLE,Chem. Rev. 1978, 78,407-489. J. S. HILL,N. S . ISAACS, Tetrahedron Lett. 1986, 16, 5007-5010. S . KOBAYASHI,T. BUSUJIMA, S. NAGAYAMA, Chem. Eur. J . 2000, G, 3491-3494. B. E. POLING,C. A. ECKERT, Ind. Eng. Chem. Fundam. 1972, 32, 451-457. (a) R. W. ABEN,L. MINUTI,H. W. SCHEEREN, A. TATTICHI,Tetrahedron Lett. 1991, 32, 6445-6448; (b) T. A ENGLER,U. SAMPATH,D. VAN DER VELDE,F. TAKUSAGAWA, Tetrahedron 1992, 48,9399-9416. P. A. GRIECO,J. J.NUNES,M. D. GAUL,J. Am. Chem. SOC.1990, 112, 4595-4596. C. LE Roux, H. GASPARD, J. DUBAC,]. Org. Chem. 1993, 58, 1835-1839. s. KOBAYASHI, I. HACHIYA, J . Org. Chem. 1994, 59,3590-3596. N. S. ISAACS, L. MAIWMOVIC, A. L A I ~ , J. Chem. SOC., Perkin Trans. 2 1994, 495-498. G. J E N N E R , New J. Chem. 1997, 21, 1085-1090.

14 F. K. FLEISCHMANN, H . KELM,

Tetrahedron Lett. 1973, 3773-3776. 15 M. AYERBE, F. P. COSSIO,Tetrahedron 16 17

18

19

20 21 22 23

24

25 26

27

Lett. 1995, 36, 4447-4450. M . REETZ, D. N. Fox, Tetrahedron Lett. 1993, 34, 1119-1122. M. A. FORMAN,W. P. DAILEY,].Am. Chem. SOC. 1991, 113, 2761-2762. R. M. PAGNI,G. W. KABALKA, S. BAINS,M. PLESCO, J. J . BARTMFSS, J. Org. Chem. 1993, 58, 3130-3133. Y. G. SHNRLIN,D. G. MURZIN,N. A. LUZANOVA, G. G. ISHKAKOVA, V. D. KISELEV, A. 1. KONOVALOV, Tetrahedron 1998, 54, 2631-2646. G. FAITA,P. P. RIGHETTI,Tetrahedron 1995, 51, 9091-9102. A. KUMAR, J . Org. Chem. 1994, 59, 4612-4617. G. J E N N E R , R. BEN SALEM,Tetrahedron 1997, 53,4637-4648. (a) S . I. HANDY,P. A. GRIECO, C. MINEUR,L. GHOSEZ,Synlett 1995, 565-567; (b) R. TAMION,C. MINEUR, L. GHOSEZ,Tetrahedron Lett. 1995, 34, 8977-8980. G . J E N N E R , Tetrahedron Lett. 1999, 40, 49 1-494. M. BEDNARSKI, S. DANISHEFSKY, J. Am. Chem. SOC. 1983, 105, 3716-3717. S. KOBAYASHI,S. NAGAYAMA, J . Am. Chem. SOC.1997, 119, 10049-10053. S. KOBAYASHI, Synlett 1994, 689-701.

346

I

70 Catalytic and Solvophobic Promotion of High Pressure Addition Reactions

S. KOBAYASHI,Eur. 1.Org. Chem. 1999: 15-27. 29 S . NAGAYAMA, S. KOBAYASHI, Angav. Chern. Int. Ed. Engl. 2000, 39, 567569. 30 J. JURCZAK, A. GOLEBIOWSKI, T. BAUER,Synthesis 1985, 928-929. 31 (a) J. JURCZAK, T. BAUER,S. JAROSZ, Tetrahedron 1986, 42, 6477-6486; (b) A. GOLEBIOWSKI, J. IZDEBSKI,U. JACOBSSON, J. JURCZAK, Heterocycles 1986,24, 1205-1208; ( c ) 1. JURCZAK, A. GOLEBIOWSKI, J. RACZKO, Tetrahedron Lett. 1988, 29, 5975-5978; (d) A. GOLEBIOWSKI, J. JURCZAK, Tetrahedron 1991, 47, 1037-1044; J. RACZKO,U. ( e )A. GOLEBIOWSKI, JACOBSSON, J. JURCZAK, Tetrahedron 1991,47,1053-1064; ( f ) T. BAUER,C. CHAPUIS,A. JEZEWSKI, J. KOZAK,J. JURCZAK, Tetrahedron Asymm. 1996, 7, 1391-1404. 32 G. JENNER, High Press. Res. 1995, 13, 321-326. 33 G. JENNER, High Press. Sci. Technol. 1980, 836-838. 34 D. A. VANDENPUT, H. W. SCHEEREN, Tetrahedron 1995, 51, 8383-8388. 35 G. H. POSNER, H. DAI, D. S. BULL, J. K. LEE,F. EYDOUX, Y. ISHIHARA, W. WELSCH, N. PRYOR,S. PETR,]. Org. Chem. 1996, 61, 671-676. 36 G. JENNER, New J . Chem. 1999, 23, 525-529. 37 G. JENNER, New]. Chem. 1995, 19, 173-178. 38 G. JENNER, Tetrahedron Lett. 1995, 36, 233-236. 39 G. JENNER, Tetrahedron Lett. 1996, 37, 3691-3694. 40 H. KOTSUKI, K. ARIMURA, R. MARUZAWA, R. OSHIMA,Synlett 1999, 650-652. 41 H. KOTSUKI, K. ARIMURA,Tetrahedron Lett. 1997, 38, 7583-7586. 42 (a) P. HARRINGTON, M. A. KERR, Tetrahedron Lett. 1997, 38, 5949-5952; (b) P. HARRINGTON, M. A. KERR,Can. I. Chem. 1998, 76, 1256-1265. 43 M. MEGURO, N. ASAO,Y. YAMAMOTO, 1.Chem. Soc., Perkin Trans. 11994, 2597-2601. 44 H. ROBERT, B. GARRIGUES, J. DUBAC, Tetrahedron Lett. 1998, 39, 1161-1164. 28

G. JENNER, High Press. Res. 2000, 18, 239-244. 46 L. F. TIETZE,C. Om, K. GERKE, M. BUBACK, Angav. Chem. Int. Ed. Engl. 1993, 32, 1485-1486. 47 E. L. VAN ROZENDAAL, B. M. Voss, H. W. SCHEEREN, Tetrahedron 1993, 49, 6931-6936. 48 G. JENNER, Tetrahedron Lett. 2000, 41, 3091-3094. 49 M. H. ABRAHAM, Progr. Phys. Org. Chem. 1974, 11, 1-87. 50 M. KATO,I. ABE,Y. TANIGUCHI,]. Chem. Phys. 1999, 110, 11982-11986. 51 D. RIDEOUT, R. BRESLOW,].Am. Chern. Soc 1980, 102, 7816-7817. 52 E. W. LANG,H. D. LUDEMANN, Ang. Chem. Int. Ed. Engl. 1982, 21, 315329. 53 (a) J. THOMAS, D. F. EVANS,].Phys. Chem. 1970, 74, 3812-3819; (b) A. RAY,1.Am. Chem. Soc., 1969, 91, 6511-6512. 54 T. DUNAMS, W. HOEKSTRA, M. PENTALERI, D. L I O ~ ATetrahedron , Lett. 1988, 29, 3745-3748. 55 C. K. PAI, M. B. SMITH,]. Org. Chern. 1995, 60, 3731-3735. 56 P. A. GRIECO,K. YOSHIDA,P. GARNER, /. Org. Chem. 1983, 48, 3137-3139. 57 A. LUBINEAU,].Org. Chem. 1986, 51, 2142-2 144. 58 A. LUBINEAU, E. MEYER,Tetrahedron 1988, 44,6075-6070. 59 W. BLOKZIJL, J. B. ENGBERTS, Angew. Chem. Int. Ed. Engl. 1993, 32, 15451579. 60 S. Omo, W. BLOKZIJL, J. B. ENGBERTS, J . Org. Chem. 1994, 59, 5372-5376. 61 W. BLOKZIJL, M. J. BLANDAMER, J. B. 1.Am. Chem. SOC.1991, ENGBERTS, 113, 4241-4246. 62 A. MEIJER,S. Orro, J. B. ENGBERTS,]. Org. Chem. 1998, 63, 8989-8994. 63 G. K. VAN D E R WEL, J. w. WIJNEN, J. B. ENGBERTS,].Org. Chem. 1996, 61, 9001-9005. 64 F. HIBBERT, J. EMSLEY, Adv. Phys. Org. Chem. 1990, 26,255-370. 65 C. CATIVIELA, J. I. GARCIA,J. GIL, R. M. MARTINEZ, J. A. MAYORAL, L. SALVATELLA, J. S. URIETA,A. M. I.Chem. MAINAR,M. H. ABRAHAM, Soc., Perkin Trans. 2 1997, 653-660. 45

References R. BRESLOW, ACC.Chem. Res.. 1991, 24, 159-170 and references therein. 67 E. T. KOOL, R. BRESLOW.J . Am. Chem. SOC.1988, 110, 1596-1597. 68 A. LUBINEAU, J. A u c i . N. BELIANGER. S. CAILI.EBOURDIN, J . Chem. SOL, Perkin Trans. 11992, 1631-1636. 69 A. LUBIKEAU,J. AWE, Tetrahedron Lett. 1992; 33. 8073-8074. 70 J. AuGE, N. LUBIN, A. LUBINEAU,J. Org. Chem. 1994, 35, 7947-7948. 71 J. W. WIJNEN. R. A. STEINER, J. B. ENGBERTS, Tetrahedron Lett. 1995, 36, 5389-5392. 72 1. HUNT,c. D. ]OHNSON,J. Chem. Soc., Perkin Trans. 2 1991, 1051-1056. 73 N. K. SANGWAN, H. J. SCHNEIDER,J. Chem. SOC.,Perkin Trans. 2 1989, 1223-1227. 74 G. J E N N E R , Tetrahedron 1996, 52, 13557-13568. 75 G. J E N N E R , I . Phys. Org. Chem. 1999. 12, 619-625. 76 (a) J. G. KIRKWOOD, J. Chem. Phys. J. Chem. 1934, 2, 351; (b) L. ONSAGER, Phys. 1936, 58, 1586. 77 G. J E N N E R , R. BEN SALEM, NewJ. Chem. 2000, 24, 203-207. 78 R. BRESLOW, T. Guo, J. Am. Chem. SOC. 1988, 110, 5613-5617. 79 C. T. R~zzo,]. Org. Chem. 1992, 57, 6382-6383. 80 R. VAN ELDIK, T. ASANO, W. J. LE NOBLE. Chem. Rev. 1989, 89, 549-688. 81 K. WONG,C. A. ECKERT,lnd. Eng. Chem. Process Res. Dev. 1969. 8, 568573. 82 G. JENNER, in Organic High J-’ressure Chemistry, W. J. LE NOBLE(Ed.), Elsevier, 1988, 143-203. 83 C. CATIVIELA, L. GARCIA, J. A. A. AVEN07A, J. M. MAYORAL, PEREGRINA, M. A. ROY,J . Phys. Org. Chem. 1991, 4,48-52. 84 G. JENNER, Tetrahedron Lett. 1994, 35, 1189-1192. 85 R. BRESLOW,U. MAITRA,Tetrahedron Lett. 1984, 25, 1239-1240. 86 J. F. BLAKE, W. L. JORGENSEN, J. Am. Chem. Soc. 1991, 113, 7430-7432. 87 J. F. BIAICE, D. LIM, W. L. J O R G E N S E N , J . Org. Chem. 1994, 59,803-805. 88 C. J. CRAMER, D. G. TRUHIAR,].Am. Chem. SOC.1992; 114,8794-8799. 66

89 90

T. R. FURIANI,J. GAO,J . Org. Chem. 1996, 61, 5492-5497. J. W. WIJNFN,J. B. ENGBERTS,l,iebigs

Ann. 1997, 1085-1088. G. J E H N E R . R. BENSALEM, Reu. High Press. Sci. Technol. 1998, 7, 1265-1267. 92 T. R. KELLY,P. MEGHANI, V. S. EKKUNDI,Tetrahedron Lett. 1990, 31, 3381-3384. 93 E. MORILD, Adti. Protein Chem. 1981, 34, 93-166. 94 H. D. LUEDEMANN.in High Pressure Biotechnology vol. 224; C . BALNY,R. 91

95 96 97

98 99 100

101 102 103 104

105 106

107

108

K. HEREMANS, P. MASSON HAYASHI, (Eds), John Libby Eurotext Ltd., 1992, 371-379. S. TASCIOGI.~, Tetrahedron 1996, 52, 11113-11152. Y. TANIGUCHI, K. SUZUKI,Rev. Phys. Chem. Jpn 1979, 49,91-94. (a) S . Orro, J. B. ENGBERTS,J. C. KWAK,J. Am. Chem. SOC.1998, 120, 7-25; (b) S. KOBAYASHI, T. S. NAGAYAMA, H. WAKABAYASHI, Tetrahedron Lett. 1997, 38, OMAYADA, 4559-4562. R. BRESLOW, Acc. Chem. Res. 1995, 28, 146-153. G. JENNER, F. MERMAZ, unpublished results. R. BENSALEM, G. J E N N E R , Rev. High Press. Sci. Technol. 1998, 7, 12681270. G. J E N N E R EHPRG , Meeting, Kloster Banz (Germany), 2000. B. H. LIPSHUTZ,Chem. Rev. 1986, 86, 795-819. W. G. DAUBEN, H. 0. KRABBENHOFT, J . Am. Chem. SOC.1976, 98, 1992. (a) F. BRION,Tetrahedron Lett. 1982, 23, 5299-2302; (b) E. VIEIRA,P. VOGEL,Hela Chim. Acta 1983, 66, 1865-1993. G. JENNER, High Press. Res. 1993, 11, 257-262. M. AVALOS, R. BABIANO, J. L. BRAVO, P. CINTAS,J. L. JIMENEZ, J. C. Tetrahedron Lett. 1998, 39, PAIACIOS, 9301-9304. J. S. HILL,N. S. ISAACS, J . Chem. Res. (S) 1988, 330-331 (M) 1988, 26412676. G. J E N N E R , High Press. Res. 1999, IG? 243-252.

13-47

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

348

I 11

Future Perspectives: Applications of High Pressure in Supramolecular Chemistry Robert Ruloz Christophe Saudan, Andri E. Merbach" and FrankGerrit Klarner" 11.1 Introduction

Supramolecular chemistry has been defined by Lehn as chemistry beyond the molecule leading to organized entities of higher complexity that result from association of two or more chemical species held together by intermolecular forces [ 11 (Fig. 11.1). The processes of molecular recognition and self-assembly/selforganization are of central importance for the programmed synthesis of higher organized chemical systems. These processes depend on weak but specific, mostly non-covalent intermolecular interactions, such as hydrogen bonding, ion pairing, arene-arene interactions (x-x, CH-x, cation-x), in addition to the less specific van der Waals or dispersion forces. Furthermore, coordinative metal-ligand bonds [21 and solvent effects can be important for the formation of supermolecules. The solvent often plays an active role in these processes by solvating or desolvating the substrates and/or receptors during the formation of the complex. In particular, the hydrophobic effect in aqueous media can be very strong and determine the complex stability to a substantial extent. Multiple interactions are necessary to form supermolecules which are, on the one hand, sufficiently stable under normal conditions (e.g. room temperature in aqueous solution) and, on the other, sufficiently flexible to undergo conformational changes and partial or complete dissociation processes without changing the conditions dramatically. The non-covalent interactions play a key role in many biological processes such as protein folding, the bonding and catalytic transformation of substrates by enzymes, formation of and transportation of neutral and ionic species through membranes, and the expression and transfer of genetic information. Inspired by the discovery of the crown ethers by C. Pederson in 1967 the research groups of Lehn and Cram started to investigate the supramolecular properties of simple synthetic receptors which can act as models for the far more complicated biological systems. In 1987 Cram, Lehn and Pederson were awarded the Nobel prize for their work in this area. Today, the non-covalent interactions and particularly the interplay between substrate and receptor via multiple interactions are investigated by

77.2 Biomolecules under Extreme Conditions

covalent synthesis

transport and catalysis

molecular recognition s R

receptor

substrate

non-covalent

functional systems with novef properties supermolecule

new materials Fig. 11.1.

Reaction scheme of the formation of supermolecules via non-covalent synthesis.

means of experimental and theoretical methods to gain further insight into the structural orientation and the energetic consequences resulting from weak bond formations. The study of the mechanism of formation of supramolecular complexes by the use of high-pressure techniques with the aim of gaining a better understanding of non-covalent binding forces will be a major subject of this chapter.

11.2 Biomolecules under Extreme Conditions

Biomolecules for example proteins, phospholipids, and microorganisms have already been studied under extreme conditions (high temperature and/or high pressure) for a long time [3], which, however, are not the major concern of this chapter. Nevertheless, it is worth reporting on the recent study of the pressure effect on the Tendamistat folding reactions 141. Tendamistat is a small all-/I-sheet protein which follows a two-state process during unfolding and refolding under all applied conditions (Fig. 11.2a) [S]. The use of a pressurized stopped-flow instrument allowed kinetic measurements to be undertaken at pressures between 0.1 and 100 MPa over a broad range of guanidinium chloride (GdmC1) concentrations as denaturant. Global analysis of the effects of pressure and GdmCl on the folding kinetics in combination with equilibrium unfolding transitions yielded the complete volume profile for Tendamistat folding. The volume of native Tendamistat is increased by 41.4 2.0 cm3 mol-' compared with the denaturated form at pH 2.0 and 35 "C (Fig. 11.2b). This value is virtually independent of the denaturant concentration and it is similar to reaction volumes (AVO) for folding of many other small single-domain proteins [GI.Contributions to AVO may arise from packing deficiencies in the native state as well as from different solvent interactions in the native and in the unfolded protein. Although the reaction volume is virtually independent of the denaturant concentration, the activation volumes for refolding and unfolding both greatly increase with

I

349

350

I

11 future Perspectives: Applications of High Pressure in Supramolecular Chemistry

A

In 7M GdmCl

In Water (OM GdmCI)

..'E"..-".T'.. AV,* :.+5.:3

-...

f 2.8

*... AVO = +41.4 f 2.0

AV,t=+25.0*1.2

U

*

N

Fig. 11.2. (a) Two-state process o f the fast folding reaction o f the Tendamistat protein. The parameters kf and k, are the rate constants for the refolding and unfolding

U

*

N

processes, respectively. (b) Full volume profiles o f the folding reaction o f the Tendamistat protein i n absence o f denaturant (A) and i n 7 M GdrnCl (E).

the GdmCl concentration. In the absence of denaturant, the activation volumes for the refolding and unfolding reaction are of 25.0 1.2 and -16.4 1.4 cm3 mol-' (Fig. 11.2a). These results show that 60 % of the volume of the transition state is native-like and suggest that it is still partially solvated. Above 6 M GdmC1, the volume of the transition state becomes larger than the volume of the native state, thus showing a GdmC1-induced structural change of the transition state. This finding suggests contributions from both dehydration and packing deficiencies to the volume of the transition state at high denaturant concentration, because the volume increase caused by dehydration should not exceed the reaction volume. As packing deficiencies become important only when solvent molecules are excluded from major parts of the hydrophobic core, these results indicate a largely desolvated, but not yet tightly packed transition state under strongly destabilizing conditions. Obviously, the transition state becomes more product-like with increasing denaturant concentration. This observation is in agreement with the Hammond postulate,

17.3 The Effect of Pressure on the Formation of Host-Guest Complexes

which states that the structure of the transition state of a chemical reaction becomes more product-like, when the product is destabilized [7]. In summary, the data show that the folding process of the polypeptide chain goes through a point of maximum volume before it reaches the native state. In a certain region on the reaction coordinate, the solvent molecules seem to be excluded from the core but the native set of tertiary interactions is not yet formed. This observation is in agreement with the concept of a dry molten globule, which was predicted theoretically [81 and observed experimentally during the unfolding of ribonuclease A [9]. These results suggest that the exclusion of water precedes the formation of correct side-chain contacts. The strong denaturant dependence of the activation volumes argues for a strong structural plasticity of the transition state for protein folding reactions and shows that the activated state can be either solvent accessible or shielded from solvent, depending on the experimental conditions.

11.3 The Effect of Pressure on the Formation o f Host-Guest Complexes 11.3.1

Inclusion of Helium Atoms or Acetonitrile Molecules in C60 or a Hemicarcerand as Molecular Containers

Only a few studies investigating the pressure effect on the formation of host-guest complexes have been carried out hitherto. A large effect of pressure has been observed for reactions in which atoms or molecules are compressed into the cavity of a container molecule without releasing other molecules out of the cavity. The volume of such a reaction is highly negative so that high pressure favors complex formation. One example is the incorporation of noble gas atoms into the cavity of buckminster fullerene Cbo which proceeds at 250 MPa and GOO "C [lo]. Another example is the effect of pressure on the incorporation of acetonitrile molecules into and their release from the cavity of the hemicarcerand 1(Fig. 11.3). Cram and co-workers found that 1 is formed with one and two acetonitrile molecules inside the cavity of 1 in a 1:1.2 ratio during synthesis [Ill. The two complexes CH3CN@l and (CH3CN),@l show characteristic guest-signals at 6 = -1.64 and -2.16 ppm., respectively, which are 3.12 and 4.14 ppm. upfield from the normal position of the acetonitrile protons (6 = 1.98 ppm). On heating at 110 "C at atmospheric pressure (0.1 MPa) the complex (CH3CN)2@1loses one CH3CN molecule leaving the complex (CH3CN)@1with a single CH3CN inside. This irreversible extrusion of CH3CN from (CH3CN)2@1even occurs in CH3CN as solvent. From the half-life of (CH3CN)2@1of 0.5 h at 110 "C the Gibbs activation enthalpy AG# can be calculated to be 120 kJ mol-l. The mechanisms of passage of CH3CN into and out of the hemicarcerand 1 were explored with force field methods [12] and gating - a conformational change that alters the constricting binding energies of 1 - was found to control the activation energies for the passage of the guest

I

351

352

I

7 7 Future Perspectives: Applications of High Pressure in Supramolecular Chemistry

Fig. 11.3. Structures of hemicarcerand 1 and the complexes of 1 including one, two, or three molecules of acetonitrile.

acetonitrile into and out of the cavity. On heating at 110 "C under 1150 MPa pressure for three days in CH3CN as solvent, the hemicarceplex with two acetonitrile guests, (CH3CN)2@l, does not release CH3CN and a new complex showing a lH-NMR signal at 6 = -1.50 is formed at the expense of the (1:1) carceplex CH3CN@1[ 131. According to extensive force-field calculations the new complex is assigned to incorporate three acetonitrile molecules, thus having the composition (CH3CN)3@1(Fig. 11.3).At atmospheric pressure and 55 "C (CH3CN)3@l reverts back to CH3CN@1 within 1 h. The potential energy surface for the formation of and release of guest acetonitrile molecules from (CH3CN)3@lhas been explored by force-field calculations showing that the conformational gating of 1 is also important for these processes.

17.3 The Effect of Pressure on the Formation of Host-Guest Complexes

11.3.2 Complexation o f Cations and C ~ as O Guests with Crown Ethers, Cryptands, and Calixarenes as Hosts

Volumes of reaction were reported for the formation of host-guest complexes between crown ethers and the [2.2.2]cryptand as hosts and alkali metal ions (Na+, K+, Cs+) or the tee-butylammonium ion as guests [14, 151. Most of the reaction volumes, determined for the complexation of the alkali metal cations from the measurement of the partial molar volumes of the reactants and the corresponding complexes, turned out to be positive (Fig. 11.4). This finding has been explained with the assumption that the decrease in volume resulting from the host-guest association is overcompensated by the increase in volume caused by the desolvation of host and guest during complex formation. The negative volumes of reaction

M'

: Na',

KO, Cso

X Q : C P , 10

AV [crn3~mol"] +4 to +19 (MeOH) +11 to +28 (MeCN) -13 to +4 (DMSO) +9 to +17 (H20)

AV = -4.2 crn3.mol-' (MeOH) /C\4'CH3 H3c

CH,

The complexation of alkali metal cations and tertbutylammonium cation by crown ethers and by the [2.2.2]cryptand. Fig. 11.4.

I

353

354

I

7 7 Future Perspectives: Applications $High Pressure in Supramolecular Chemistty

found for a few complexations in DMSO have been explained by the effect of electrostriction leading to a looser solvation of the ions in the aprotic solvent than in the protic solvents so that the volume expansion in DMSO resulting from the solvent release during complex formation is smaller than in methanol or water. This conclusion, however, cannot explain the solvent effect of acetonitrile. The complexation of the tert-butylammonium perchlorate by 1,3-xylyl-18-crown-Sdetermined by high-pressure 13C-NMRspectroscopy in methanol (up to 150 MPa) shows a negative volume of reaction indicating that the lipophilic tert-butylammonium cation is not solvated in methanol as much as the alkali metal cations. In the complex, the primary ammonium cation is certainly hydrogen-bonded to three oxygen atoms of the crown ether leading to a substantial volume contraction which explains why the change of overall volume is negative in this case. The inclusion of buckminster fullerene C60 into the cavity of a p-benzylcalix[51arene in toluene solution reported by Isaacs et al. [lG], is associated with a dramatic partial molar volume change of AV = +195 cm3 rno1-l which is consistent with the displacement of two toluene molecules from the cavity of the calixarene during complex formation (Fig. 11.5). 11.3.3

Molecular Tweezers as Synthetic Receptors: Focussing on Volume and Entropy of Association

Recently, the syntheses and some supramolecular properties of receptors with ribbon-type concave topology have been reported [17, 181. Due to their ability to selectively bind electron-deficient aromatic and aliphatic substrates, these compounds can be regarded as molecular tweezers. This section is devoted to the temperature and pressure dependence of the host-guest complex formation of the naphthalene-spaced tweezer 2 and the diacetoxy substituted derivative 3 with the

toluene

+

P

OH p-benzylcalix[5]arene V [cm3.mol-'] 366 +- 20

789 +- 5

AV= +195 cm3.mol-' Fig. 11.5.

The inclusion of C60 into the cavity of p-benzylcalix[S]arene.

Complex

1350 +- 50

71.3 The Effect of Pressure on the Formation of Host-Guest Complexes

2:R=H 3: R = OAC

5

4

6

Fig. 11.6. The formation o f host-guest complexes between the molecular tweezers 2 and 3 and quinoid or aromatic guests 4-6.

aromatic guests 4-6 (Fig. 11.Q to determine whether there is a relationship between volume and entropy of reaction. The experimental results, summarized in Table 11.1, were obtained as follows: 'H-NMR-titration experiments were performed to determine the association constant Ka (Eq. 11.1) for the complexation between the two hosts (H) 2 and 3 and the three guests (G) 4-6 in addition to the maximum complexation-induced shift A&,a of the guest protons in the corresponding complex (GH). (11.1)

Tab. 11.1.

Host-guest interaction o f molecular tweezers 2 and 3 with aromatic guests 4, 5 and

'

6 . Experimental results o f H-NMR titration at 294 K, variable-temperature (VT) single-point NMR analysis, and high-pressure (HP) NMR study at 298 K.

Host

Guest Solvent

NMR titration

ASmax fPP4

AH

AS

P-'l

fkJ mol-'J

(J mol-' K-')

20 35 110 145 12 18 110 850

2.8 1.6/0.6 4.3 2.7 3.2 1.6/0.6 4.1 2.1

-27 -25 -11 -16 -11 -9 -8 -16

-64

K

2 2 2 2 3 3

4 5 6 6 4 5

3

G

3

6

CDC13 CDC13 CDCl3 C6D6 CDCI, CDC13 CDC13 C6D6

HP NMR

!IT NMR

-59 4-2 -13 -16 -10 +32 -1

AV (cm3 mo/-'j

A&, fPP4

-3.0 $0.6 +1.5 -1.1

2.5 1.4/0.6 4.2 2.7

-1.1

2.1

I

355

356

I

7 7 future Perspectives: Appfications of High Pressure in Supramolecular Chemistry

Since Ad,, showed no significant variation with temperature and pressure, enthalpy AH and entropy AS of reaction could be easily determined by variabletemperature single-point analyses and the volume of reaction AV by variablepressure IH-NMR studies. The thermodynamic parameters (Table 11.1)clearly indicate, that the complex formation between the two tweezer molecules 2 and 3 as hosts and the aromatic guests 4-6 is significantly controlled by the enthalpy of reaction AH. The entropies of association vary in a relatively broad range (AS = -64 to +32 J rno1-l K-'), so that the entropic contribution leads to a shift of the equilibrium either toward the side of the complex or toward the side of the reactants. The reaction volumes of all host-guest interactions have been studied [19] and show, however, only a small variation around the zero value (AV = -3.0 to +1.5 cm3 mol-I). Consequently, the decrease in volume resulting from the complexation of the guest in the cavity of the tweezer is more or less compensated by the increase in volume caused by the desolvation of host and guest (release of solvent molecules) during the complexation. There is no correlation between the volumes and entropies of reaction for the host-guest systems listed in Table 11.1. Consequently, the large differences observed for the reaction entropies cannot only be the result of the contributions coming from the association and desolvation. The entropy values are surprisingly positive in the case of the complexation of p-dicyanobenzene ( 6 ) with both tweezers 2 and 3 (in CDC13: AS = +2 and $32 J rno1-l K-*) whereas strongly negative values were obtained for the complexation of p-benzoquinone (4) and terephthalic aldehyde (5) (in CDC13: AS = -64, -59, -16 and -10 J rno1-l K-l) as expected for bimolecular associations. But the effect of bimolecular association does not show up in the corresponding reaction volumes (AV = -3.0 and +O.G cm3 mol-l, respectively) which should be more negative, provided that they only result from association. In order to explain this contradiction the mobility of the guest molecule has been assumed to be related to the magnitude of the entropy while having only a small influence on the volume of the complex. According to single-crystal structure analyses, temperature-dependent 'H-NMR studies, and quantumchemical calculations the guest 6, p-dicyanobenzene, rotates inside the tweezer cavity rapidly on the NMR time scale (Fig. 11.7).These rotations give positive contributions to the entropy and, evidently, overcompensate, together with the also positive effect of the desolvation, the negative contribution to AS resulting from the association. For the complexes between p-benzoquinone (4) or terephthalic aldehyde (5) as guest and the tweezers 2 and 3 as host, a more symmetric orientation of the guest inside the tweezer cavity has been found leading to a negative contribution to AS. From this study it was concluded that the volume of association provides important information about the desolvation of host and guest which may occur during the formation of supramolecular complexes, whereas the interpretation of the entropy of association turned out to be more difficult because - in addition to the effects resulting from association and desolvation - the conformational freedom of the guest inside the receptor cavity provides a significant contribution to AS.

11.3 The Effect of Pressure on the Formation of Host-Guest Complexes

N ,,,

(1)

Hb

..

t -

4 ha

N

N

"

6@2

H A 0

402 Fig. 11.7. (a) Possible conformational interconversion o f p-dicyanobenzene 6 inside the cavity o f tweezer 2; (1) 60" flip o f the benzene moiety, ( 2 ) 180" rotation o f the

5@2 benzene moiety around the C-N axis. (b) Section o f the structures o f complexes 4@2 and 5@2 derived from single-crystal structure analyses.

11.3.4

Formation o f Host-Guest Complexes of aCyclodextrins with Azo Dyes: Determination of Activation and Reaction Volumes

Cyclodextrins (CDs) are cyclic oligomers of a-D-glucoseand are generally described as shallow truncated cones, where the primary hydroxy rim has a reduced diameter compared to the secondary hydroxy rim. They attracted a wealth of interest in supramolecular chemistry as host molecules capable of selectively accommodating guest species via non-covalent interactions in their hydrophobic cavity. The responsible forces for the formation of those complexes are believed to be primarily hydrophobic and van der Waals interactions. The relief of CD strain energy and the release of partially hydrogen-bonded activated water molecules from the cavity are thought also to play a specific role. A novel approach in understanding host-guest interactions of CD complexes in aqueous solution by the use of high-pressure is discussed in the following two sections. 11.3.4.1

Inclusion o f Short Guests into the a-Cyclodextrin Cavity

Volume data for host-guest interactions of a-CD with diphenyl azo dyes (Fig. 11.8) were first reported by Merbach and co-workers [20]. Their kinetic study at variable temperature and NMR experiments have shown that a 1:1 complex is formed by directional binding of the dye, which is inserted by its sulfonate/sulfonamide

I

357

p

358

I

7 7 Future Perspectives: Applications of High Pressure in Supramolecular Chemistry

Dyes [Sl

Tail

Head

R3

R4

Rs

Y

7

€I

N(Et)2

H

SOj

8

Me

OH

COO-

SO<

9

Me

OH

Me

SO3-

10

Me

OH

COO-

S02NH2

a-Cyclodextrin [a-CD]

1.37 nm

OH

OH

fast

slow

Molecular structures o f t h e diphenyl azo dyes 7-10 investigated in the complex formation with a-cyclodextrin at pH ca. 6.5, and reaction scheme. Fig. 11.8.

17.3 The Effect of Pressure on the Formation of Host-Guest Complexes

I

359

9'a-CD*

9.a-CD

Fig. 11.9. Force-field-energy minimized structures of intermediate 9 . E-CD" and final product 9 . E-CD. All hydrogen atoms are omitted for clarity.

group through the wide rim of the a-CD cavity. The inclusion reactions of all dyes (S) were found to be a two-step mechanism as shown in the reaction scheme in Fig. 11.8, where the first fast step gives the intermediate S . ol-CD", which is slowly rearranged to form the final complex S . a-CD. For more detailed information about the thermodynamic and kinetic parameters obtained the reader is referred to the original publication. The 1 D 'H-NMR spectra showed two separate sets of peaks for the complexes S . E-CD'~and S . a-CD. The 2D ROESY NMR spectrum of the a-CD complex of 9 was employed to extract distance restraints for molecular modeling calculations. The minimized structures of the intermediate and the final product are illustrated in Fig. 11.9, showing that the guest is deeply inserted in the cavity of a-CD. High-pressure stopped-flow kinetic measurements were performed up to 200 MPa. The observed rate constants, kl,obs and k2+bs, as a function of pressure and aCD concentration, were fitted simultaneously according to the reaction scheme and k2,r. shown in Fig. 11.8 leading to the microscopic rate constants kl,f, k2,f, The results are summarized in the form of volume profiles in Fig. 11.10. An interpretation of the volume profiles in the context of the complexation mechanism previously discussed, offers a global view of the inclusion process. The inclusion mechanism begins with the encounter of the dye and a-CD mainly due to hydrophobic interactions followed by a partial desolvation of the entering sulfonate/sulfonamide head of the dye. The latter interacts with the two "activated' water molecules inside the cavity of the free host, and their complete release through the smaller rim of the a-CD is retarded by the primary hydroxy groups.

360

I

7 7 Future Perspectives: Applications of High Pressure in Supramolecular Chemistry

"=Lo

..-.

AV?f = -22.lf 0.4

9

..........

AVlf =-1.8+2.8

7

1t

1 .-

.................

I

AV?f = -20.9

AV$ = -21.8f 2.9

* 0.5

AV,ft = -15.8

.........._..........

I

+

t ....

..........-....

~\1,0=-ii.o+o.8

28

AV:

= -23.6

...

-"",

AVzf; = -16.l* 2.1

= -12.6 rf: 0.6

1

4....

....-....

9

AV; =+3.7*3.0

* 0.2 AV:

O=r0

f 0.6

I

AV$ = -17.3f0.7

$

-

AVi =+6.lf3.3

A5 ' = -3.6 rf: 1.1

.....................

AV,O=-7.8*0.6

.............................. AV; = -8.2

+

f 0.3

.........

I

A V i =+3.0*1.4 .......................

f.-

..7'. . .......

........::~,y,.,.,.,.-. Av& = -0.42 0.4

' " ' ~ ~ " ' '

Avzft ~ - 2 . 5 f 0 . 3

AH-

.....-

I Avir

......... =-5.5*1.3

10

S+a-CD+{

S+a-CD*}*-

S.a-CD*+

Fig. 11.10. Volume profiles for the inclusion reactions o f t h e four guest molecules (7 at 308 K, 8 and 9 at 288 K and 10 at 278 K) with a-CD i n aqueous solution.

{S.a-CD}*+

S.a-CD

7 1.3 The Effect of Pressure on the Formation of Host-Guest Complexes

This situation represents the first transition state: a squeezed arrangement inside the host induces a negative activation volume (AV:f z -21 to -24 cm3 mol-1 for dyes 7-9 with sulfonate head groups versus -8 cm3 mol-I for dye 10 with a sulfonamide head group). The subsequent intermediate S . a-CD” is characterized by a total release of the two inner water molecules and interactions of the dye head with the primary hydroxy groups of the host in a trapped-like structure having a negative reaction volume AV; z -11 to -4 cm3 mol-l. The latter interactions and concurrent tail interactions of the guest with the secondary hydroxy groups at the wider rim of the host lead to a strained conformation of the CD molecule in the second transition state and this is associated with a volume contraction (AVf: z -2 to -16 cm3 mol-l). In the final complex Sa-CD, the head of the dye is completely rehydrated as it protrudes from the primary end of the host cavity which can now adopt a released conformation. The positive reaction volume (AV; z +3 to +6 cm3 mo1-l for dyes 8-10 versus +17 cm3 mol-l for dye 7) reflects the breaking of the primary hydroxy group trap as well as the expansion of the host cavity. Due to reduced steric hindrance in dye 7, a deeper encapsulation of the guest might induce a larger expansion of the CD cavity. 11.3.4.2

Sequential Threading o f a-Cyclodextrin onto a Long Guest

The inclusion of a relatively long guest into the cavity of a-CD may be regarded as a “threading” of the latter onto the former. This view turns out to be useful, when even a second a-CD is “threaded’ onto the long guest molecule. Indeed, such a complexation process was found to take place for the interaction of dye 11 (Fig. 1l.lla) with a-CD [21]. The complete thermodynamic and kinetic parameters of the system were determined and led to the suggested mechanism shown in Fig. 11.1lb. From IH-NMR titrations of dye 11 by c(-CD,the formation of the two 1:1 complexes, S . a-CD* and S . a-CD, was observed at low concentrations of titrant. This represents a situation similar to that found for the reaction of host molecules with the above-mentioned short diphenyl azo dyes. However, only the 1:2 complexes S . (cl-CD); and S . (GC-CD)~ were detected at higher a-CD concentration. The structures of the four resulting supramolecular complexes were deduced from H 2D ROESY NMR studies. The results of the pressure dependence experiments are summarized in the volume profile shown in Fig. 11.12. The formation of Sa-CD* shows volume data similar to that for the short dyes, indicating that the same mechanism is operative for the first step of the complexation of dye 11 with cl-CD and does not depend on the length of the dye. The threading of a second a-CD onto S . a-CD* has a reaction volume AV; of -7.5 cm3 mol-l, which is significantly more negative than AV;. This greater magnitude is probably a consequence of a conformational change in either one or both CDs and most readily achieved through reorientation of the flexible primary hydroxy arm units. The reaction volume of the isomeric equilibrium between S . z - C D ~ ~ and S . a-CD (AV; = -5.1 cm3 mol-l) is consistent with a loss of degrees of freedom and a volume contraction most probably reflecting changes in conformation of a-CD and 11. The formation of S.(a-CD)zis characterized by a positive reaction volume (AV; = +8.4 cm3 mol-I) which contrasts with the negative value of

362

I

7 7 Future Perspectives: Applications of High Pressure in Supramo/ecu/ar Chemistry

a)

Head

Tail

2-@0;

/

H H3C o b " . - ;

11

1

fast step

slow step

+

1

a-CD

very slow step

fast step

+

a-CD

S.a-CD Fig. 11.11. (a) Molecular structure ofthe azo dye 11 as guest for the complexation with ctCD. (b) Mechanism for the threading reactions ofdye 11 with a-CD. The proposed structures are based on 2D NMR studies, dimension data

S.(a-CD), from crystal structures and modeling calculations. The conformations of a-CD shown were arbitrarily chosen. All hydrogen atoms are omitted for clarity.

AV;. Three probable origins of these differences were proposed: (1) the seconc a-CD in S . ( N - C D )is ~ closer to the sulfonate head of the dye and therefore the hy dration of that group may be less significant €or this complex; (2) due to hydroger bonding between the two cr-CDs, they may largely act as a single unit and thii could be of different significance in the two 2 :1 complexes, and (3) twisting o phenyl groups of 11 out of the common plane may occur to greater extent in on1

7 7.3 The Efieect of Pressure on the Formatian of Host-Guest Complexes

-[

I

363

-

AVP =-2.8*1.6

.........................................

~... ...............

1

=-22.3*0.6

I

AV. =-7.5*1.4

..........

..............................

.......

AK:r =-19.5+1.9

AVir =-16.3k1.0

1

...., . (S.a-CD*]* + a-CD

S + 2a-CD

S.a-CD* + a-CD

.................

S. (u-CD)i

{S.(a-CD)z*}*

A

T

-_ _ _ _

AK0 =-2.8+1.6

.~.~ .....................................

1

Av30

................

............ t................

1

AV,*f =-22.3f.0.6

I-

~

S + Za-CD Fig. 11.12.

1:

.......

{S.a-CD*}*+ U-CD

1

----

=-5,1+1.7 ~

-

I

.................. -.

.

T

................... .......-.........

AV& =-16.2k1.6

=-19.5*1.9

..................

...............................

...................... .

y----.----

AG = +8.4 * 2.4 !..........

AVlr =-11.1*1.4

4

...... .........

........

Sa-CD* +a-CD {S.u-CD)* +mCD

S‘U-CD +a-CD

Volume profile for the threading of a-CD by guest 11 at 288 K in aqueous solution.

complex compared to the other. Concerning the volumes of activation, they are for the first three forward and reverse reactions substantially more negative than the volumes of reaction, indicating a compression (mainly due to the electrostriction of water) of the transition states in comparison with the ground states. To sum up, the complexation of dyes with a-CDs is largely controlled by changes in hydration and van der Waals interactions, and possibly by conformational changes in both host and guest. 7 1.3.5

Self-Assembled Multinuclear Coordination Species with Chiral Bipyridine Ligand

Self-assembly of helicates with labile metal centers has recently attracted much attention with regard to the interesting properties of the species or for the further construction of molecular architectures. Von Zelewsky and coworkers studied complexation reactions between Ag+ and a series of enantiopure ligands belonging to the 5,G-CHIRAGEN family, based on (-)-5,G,-pinene bipyridine [22, 231. For example, as found by X-ray diffraction in the solid state, ligand L (Fig. 11.13) spontaneously forms a six-fold, circular monostranded helicate [Ag6L6] with predetermined chirality (Fig. 11.14). To elucidate the behavior of this hexanuclear coordination species in solution, detailed thermodynamic studies were carried out by NMR spectroscopy as a function of concentration, temperature and pressure ~41. The starting point was the observation of a splitting of the lH-NMR signals of [Ag6L6I6+in CD3CN at low temperature into two sets having unequal signal in-

S.(a-CD),

364

I

7 1 Future Perspectives: Appkations of High Pressure in Supramolecular Chemistry

Q Fig. 11.13.

Chemical structure of the enantiopure ligand L.

tensity, which was attributed to the existence of two rapidly exchanging species in solution. Addition of Ag' did not lead to any observable change in the relative proportions and chemical shifts of the two sets of signals, which is a strong indication that the ratio between Ag+ and L is the same in both species in equilibrium. Measurements of IH-NMR spectra as a function of total [Ag6L6I6+concentration and a systematic data analysis revealed the nature of this equilibrium. The resulting model, an equilibrium between the hexanuclear and a tetranuclear species, is represented by Eq. (11.2). 2[Ag6L6IGf + 3[Ag4L4I4+

(11.2)

The corresponding equilibrium constant K is expressed by Eq. (11.3).

Fig. 11.14.

Crystal structure of the

[Ag6kl6+ ion; view along its six-fold axis.

7 7.3 The Efleect ofpressure on the Formation of Host-Guest Complexes

Additional results obtained from Io9Ag-NMRexperiments were in good agreement with the proposed equilibrium model, two signals showing silver in different magnetic environments. Electrospray mass spectra showing the presence of both hexanuclear and tetranuclear species were also reported. A variable temperature 'H-NMR study of the equilibrium reaction in Eq. (11.2) was performed to determine the equilibrium constant at standard temperature K2", the reaction enthalpy AHo and the reaction entropy AS". The following values 0.8 kJ mo1-l were obtained K2'* = (8.7 & 0.7) x lo-' mol kg-l, AH' = -15.65 3 J mol-' I<-'. Consequently, the formation of the tetraand AY = -130.2 nuclear species is entropically disfavored while being enthalpically favored at low temperature. It is a surprising, counterintuitive result that the entropy decreases for a reaction in which the number of molecules increases. In order to obtain further insight into that phenomenon pressure-dependence measurements were conducted. Variable pressure 'H-NMR experiments were performed at 256.4 I< to determine the reaction volume Avo of the equilibrium between [Ag6L6I6+and [Ag4L4I4+.The pressure effect was found to be remarkably large, as shown in Fig. 11.15 by the well-separated singlets due to the protons in the para-xylydenebridge of L. Thus, a

-.

-7 8

n

9

..............:

:

, ~ ~ , ~ , ~ - , ~ - - , ~ - - ~ , " - -

-,,

6

5

4 3 (PP4

2

1

0

Fig. 11.15. (a) Variable pressure 400-MHz NMR spectra o f a solution o f [Ag6L6](PF6)6 in CD3CN at 256.4 K. (b) Zoom o f t h e NMR spectra between 5.2 and 6.3 ppm. (c) Pressure dependence o f experimental values o f In K (H) at 256.4 K. The solid line represents the calculated function.

I

365

366

I

7 7 Future Perspectives: Applications of High Pressure in Supramolecular Chemistry

relatively small pressure range (75 MPa) was sufficient to obtain AVO with good accuracy. Figure 1 1 . 1 5 ~shows the result of applying Eq. (11.4) to the data, where KO is the equilibrium constant at zero pressure and A T the compressibility

In K = In KO- (AV"P/RT) + ( A p o P 2 / R T )

(11.4)

coefficient. The reported value of AV" is -160 f 12 cm3 mol-I with mol kg-' (at T = 256.4 K) and AT = -(4.0 -t 0.6) x KO = (2.8 f 0.2) x cm3 mol-l MPa-l indicating a preferred formation of the tetranuclear complex at high pressure. Indeed, the entropy sign from the isobaric measurements was confirmed, and a similar surprising result as that for the entropy was found by the pressure-dependence study: The reaction volume of the equilibrium in Eq. (11.2) decreases when going from two hexamers to three tetramers. In order to understand this unexpected result and to approach the situation on a molecular level, model calculations were carried out considering solvation effects. and the Connolly's molecular volumes were calculated for the hexamer ( VLg,L,) tetramer (VLg4g,L,) by varying the probe radius rprobe in the range of 0 to 3.0A, as shown in Fig. 11.16.The Connolly's molecular surface of both complexes for a chosen rprobe value equal to 1.4A is illustrated in Fig. 11.17. As reported, the reaction volume was given by Eq. (11.5). AV" = 3 VLg,L,- 2 VLgGL,

(11.5)

All calculated volumes are based on the X-ray structure of [Ag&]'+ and a stmcture model of [Ag4L4I4+,which was minimized by using the X-ray bond lengths and constraints of the hexamer. Owing to a lack of accurate information about the acetonitrile solvation pattern of these complexes a relative large range of rprobe was

,3.6

r---

n

4

-,

0 r-!

1 0

0 -0

-1 Fig. 11.16. Calculated Connolly's molecular volume o f the tetramer (0),of the hexamer ( 0 )and the reaction volume ( W ) as a function o f the spherical probe radius (rprobe).

7 1.4 Conclusions and Outlook

Fig. 11.17. View o f t h e Connolly’s molecular surface (in dots) o f [Ag,bl4+ (left) and [Ag6L616+ (right) for rprobe= 1.4A.

considered. Thus, different types of interactions were covered. The resulting values for AVO calculated using Connolly’s method vary non-monotonically with a maximum of +117 and a minimum of -375 cm3 mol-’. This corresponds to a variation of the volume ratio 3V&4L4/2V2g,Ls between +2 and -G %. The observed value AVO = -160 f 12 cm3 mol-l is considered to be a consequence of the two contributions: (1)the difference in the intrinsic molar volume of three tetramers and two hexamers and (2) the difference in the solvation contribution of these complexes. Since the calculations derived by Connolly’s method revealed the possibility of either sign for the solvation effect, it could not be decided conclusively at that point which of the two contributions dominates in producing the remarkably large reaction volume observed.

11.4

Conclusion and Outlook

The investigation of the effect of pressure on the formation and self-assembly of supramolecules provides further insight into the mechanisms of these processes. The comparison between volume and entropy of reaction shows that volumes can give clear-cutevidence for the extent of the desolvation of host and guest which has to compete with the host-guest association during complex formation, whereas the magnitude of the entropy can be additionally determined by the mobility of the guest inside the host cavity. The activation volumes give valuable information on the steric demands of the transition states during complex formation. Finally, high pressure can be employed to synthesize host-guest complexes which are stable under normal conditions, simply by compressing small atoms or molecules into empty molecular containers.

I

367

368

I

I 7 Future Perspectives: Applications $High Pressure in Supramolecular Chemistry

Acknowledgments

Dr F. Wurche is thanked for his skilled assistance with the editing of this chapter. F.-G. K. is grateful to the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich SFB 425) and the Fonds der Chemischen Industrie for financial support.

References 1

2

3

4

5

6

7 8 9

10

11

12

13

J.-M. LEHN, Supramolecular Chemistry, Concepts and Perspectives, VCH’ Weinheim, 1995. F. M. TABELLION, S. R. SEIDEL,A. M. ARIF, P. J. STANG,]. Am. Chem. SOC. 2001, 123, 7740-7741 and literature cited therein. K. HEREMANS, in Chemistry under Extreme and Non-Classical Conditions, R. VAN ELDIK,C. D. HUBBARD, (Eds), Wiley-Spectmm, New York, Heidelberg, 1997, Chapter 12, 515-545. G. PAPPENBERGER, C. SAUDAN,M. BECKER, A. E. MERBACH,T. KIEFHABER,Proc. Natl. Acad. Sci. USA 2000, 97, 17-22. N. S C H ~ N B R U N N G. ER, PAPPENBERGER, M. SCHARF,J. ENGELS, T. KIEFHABER, Biochemistry 1997, 36, 9057-9065. K. HEREMANS, Annu. Rev. Biophys. Bioeng. 1982, 11, 1-21. G. S . HAMMOND,].Am. Chem. SOC. 1955, 77, 334-338. E. I. SHAKNOVICH, A. V. FINKELSTEIN, Biopolymers1989, 28, 1667-1680. T. KIEFHABER, A. M. LABHARDT, R. L. BALDWIN,Nature (Lond) 1995, 375, 513-515. M. SAUNDERS, H. A. JIMBNEZV ~ S Q U ER. Z , J. CROSS,S. MROCZKOWSKI, M. L. GROSS,D. E. GIBLIN,R. J. P O R E D A ,Am. ~ . Chem. SOC.1994, 116, 2193-2194. J. A. BRYANT,M. T. BLANDA,M. VINCENTI,D. J. CRAM,]. Am. Chem. SOC.1991, 113, 2167-2172. (a) K. NAKAMURA, K. N. HouK,]. Am. Chem. SOC.1995, 117, 1853-1854; (b) K. N. HOUK,K. NAKAMURA, C. SHEN, A. E. KEATING,Science 199G, 273,627629. G. VAYNER,J. CHEN,K. NAKAMURA,

K. N. HOUK,F.-G. KLLRNER, Manuscript in preparation. 14 H. YAMADA, M. KAZUOKA, N. MORIGUCHI,A. SERA,Bull. Chem. SOC. Jpn. 1993, 66, 3528-3529. 15 T. M. LETCHER, J. D. MERCERCHALMERS, R. L. KAY,Pure Appl. Chem. 1994, 66,419-427. 16 N. S. ISAACS, P. J. NICHOLS,C. L. RASTON,CH. A. SANDOVA, D. J. YOUNG, Chem. Commun. 1997, 1839-1840. 17 F.-G. KLLRNER,U. BURKERT, M. KAMIETH,R. BOESE,J. BENETBUCHHOLZ,Chem. Eur. ]. 1999, 5, 1700-1707. 18 F.-G. K L ~ R N E J. R ,BENKHOFF,R. BOESE,U. BURKERT,M. KAMIETH,U. NAATZ,(a) Angew. Chem. 1996, 108, 1195-1198; (b) Angm. Chem. Int. Ed. 1996,35,1130-1133. 19 R. RULOFF,U. P. SEELBACH, A. E. MERBACH, F.-G. KLLRNER,J . Phys. Org. Chem. 2002 (in press). 20 A. ABOU-HAMDAN, P. BUGNON,C. SAUDAN,P. G. LYE, A. E. MERBACH,]. Am. Chem. SOG.2000, 122, 592-602. 21 C. SAUDAN,F. A. DUNAND, A. ABOUHAMDAN,P. BUGNON,P. G. LYE,S. F. LINCOLN,A. E. MERBACH, J . Am. Chem. Soc. 2001, 123, 10290-10298. 22 0. MAMULA, A. VON ZELEWSKY, G. BERNARDINELLI, (a) Angew. Chem. 1998, 110, 301-305; (b) Angew. Chem. Int. Ed. Engl. 1998, 37, 289-293. 23 0. MAMULA, A. VON ZELEWSKY, T. BARK,G. BERNARDINELLI, (a) Angew. Chem. 1999, 111,3129-3133; (b) Angm. Chem. Int. Ed. Engl. 1999, 38, 2945-2948. 24 0. MAMULA, F. j. MONLIEN,A. PORQUET,G. HOPFGARTNER, A. E. MERBACH, A. VON ZELEWSKY, Chem. Eur.]. 2001, 7,533-539.

I

Ill

Chemical Reactions in Supercritical Fluids

369

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

I 12

Catalytic Reactions in Supercritical Fluids Jason Hyde, Walter Leitner” and Mattyn Po/iakoff* 12.1

Introduction to Catalytic Reactions in Supercritical Fluids 12.1.1

Solvent Properties of Supercritical Fluids

Supercritical fluids (SCFs) are solvent systems sufficiently different to offer interesting new opportunities for catalytic reactions. Our research groups at Nottingham University and MPI fur Kohlenforschung have performed a wide spectrum of reactions using these fluids [l-61. This chapter is not intended to be comprehensive, but rather it aims to give a general introduction to the variety of work which can be performed by the use of these solvents. Our use of the word solvents is very broad, referring usually to a single supercritical fluid but also to much more complex mixtures of fluids. However, a single supercritical fluid can be tuned to mimic the properties of a wide range of conventional solvents by merely varying the pressure [7], which is one of the more prominent advantages of supercritical fluids. This chapter is largely concerned with reactions in supercritical carbon dioxide ( s c C 0 ~ utilizing ) heterogeneous or homogeneous catalysts. However, it must be stressed that there are interesting opportunities also for reactions in other supercritical solvents such as propane or water (see Chapter 14) [8-121. An SCF can be defined as a substance heated above its critical temperature (T& and which is also compressed above its critical pressure (pc)(see Fig. 12.1) [13]. At the critical point, the fluid phase boundary between liquid and gaseous phases disappears, and the properties of the new single “supercritical” phase are best described as a combination of those of the liquid and gaseous phase. 12.1.2

Temperature and Pressure Relations

The density of SCFs can be tuned to promote either the liquid-like or gas-like properties with comparatively small pressure changes [ 141. The influence of pres-

371

372

I

12 Catalytic Reactions in Supercritical Fluids

Pressure

Lquid

Temperature Fig. 12.1. P,Tphase diagram of a single substance. The letters a-d refer t o the photographs i n Fig. 12.2.

Fig. 12.2. A series of four photographs showing the effects o f heating the fluid (ethane) through the critical point, 32.6 "C. (A) shows the distinct liquid and gas phases. As the temperature rises the meniscus becomes flatter (B) and the liquid phase expands (C). When the critical point has been reached (D), the interface between gas and liquid phase

disappears. In these photographs, a trace of the highly absorbing dye gaiazulene has been added t o the liquid phase t o emphasize the distinction between phases. Note that the dye remains i n solution even above the critical point, but that the shape of the vessel prevents complete mixing o f the colored and transparent regions of the fluid.

12.I Introduction to Catalytic Reactions in Supercritical Fluids

I

373

sure on density and on many of the other properties discussed below, is most pronounced in the region close to the critical point (the near critical or compressible region). Variation in the bulk density can have pronounced effects on chemical reactions, including the possibility to change molar fractions (i.e. concentrations) at constant volume [15]. It must be noted, however, that SCFs are highly dynamic media and local densities around a solute can be significantly different from the bulk density of the medium [16]. The dynamic viscosities of SCFs are closer to those normally found in the gaseous state. As the pressure is increased, the viscosity and diffusion ability of a supercritical liquid approach that of a liquid. However, even at high pressures (300-400 bar), viscosity and diffusion can be one to two orders of magnitude lower or higher than in liquids, respectively. Diffusion will increase with increasing temperature, whereas the viscosity decreases with a temperature increase, the opposite of that which is found for gases [17]. Furthermore, temperature can be used to control the dielectric constants of more polar fluids [18].Towards the critical point, the dielectric constant has been shown to fall slightly, and at the critical point, to drop dramatically [19], an effect that can be used to manipulate the reaction conditions for a particular reaction [20]. Therefore, the properties of gas-like difision, gas-like viscosity, and liquid-like density combined with the pressure-dependent solvating power and/or polarity have provided a strong impetus for chemical research using SCF solvents. A 3-D density, pressure, temperature diagram for C02 is shown in Fig. 12.3. This clearly shows that for a given pressure, the highest density occurs close to the critical temperature. By careful control of pressure and temperature, the density, and hence the dielectric constant and heat capacity can be controlled [21], a factor that is often crucial to reactions carried out in scCO2. The density can be predicted reasonably accurately by employing appropriate equations of state, such as those used by computer modeling programs. In a continuous reactor, at a constant flow rate, density is one of the key factors for controlling the residence time within the reactor and hence to optimize the efficiency of the process. The density can be measured by a variety of methods including, with suitable calibration, spectroscopic measurements. Catalytic reactions utilizing SCFs as a reaction medium have recently been highlighted as an industrially promising approach to “greener” processing [221. Research into the field has increased dramatically over the past decade, driven by the need to replace environmentally less acceptable solvents. Advances in the field are leading to the development of processes, which potentially could have a major impact on everyday life in the near future. In this context, the most significant existing industrial investment into SCFs has been the decaffeination of coffee by extraction with scCO2, as invented by Kurt Zosel at the MPI fur Kohlenforschung in the 1960s [23]. This process has revolutionized the market for decaffeinated coffee by removal of toxic solvents from the process. Although decaffeination does not involve a catalytic chemical reaction, it is useful to discuss briefly some features which are most pertinent to the technical application of scCO2 in any chemical process.

374

I

12 Catalytic Reactions in Supercritical Fluids

Fig. 12.3. Plot o f density variations with respect t o pressure and temperature for COz. calculated using the program

NIST-14.

12.1.3 Decaffeination of Coffee via scCOz Extraction of Caffeine

Extraction of caffeine was traditionally performed by solvent extraction w CH2C12,which frequently left residues in the coffee beans. The extraction procc was greatly refined after scC02 was found to extract caffeine selectively direc from water-swollen green coffee beans [23]. The process is shown very schem: cally in Fig. 12.4. Raw coffee beans are fed batch-wise into the vertical extraction vessel, and scC1 is circulated through it. Caffeine is sparingly soluble in the scCO2 phase, whi transports the caffeine to a water scrubber. The caffeine partitions into the wa phase and C02 is recirculated back to the extraction vessel. As the C02 is still w it makes the extraction of caffeine more favorable [ 241. It is important to stress tl the caffeine is recovered without the need for expansion (and recompression) the C02. Thus, although caffeine is only poorly soluble in scC02: the absence recompression means that the C02 can be recycled many times over the sar

72.2 Practical Aspects of Catalytic Reactions in scCOz

Fig. 12.4. Simple diagram o f the extraction o f caffeine, from coffee beans via scCOz extraction.

coffee beans to achieve efficient extraction. Many of the chemical reactions described below are somewhat different because the ratio of substrate: COZ is much higher than that in the case of decaffeination. This suggests that product recovery by decompression may be a commercially viable alternative.

12.2

Practical Aspects of Catalytic Reactions in scCOz 12.2.1

Heterogeneously Catalyzed Reactions

Heterogeneously catalyzed reactions in SCFs must, by definition, involved at least two separate phases, the reaction mixture and the solid catalyst, and frequently may contain more. There has been considerable argument as to whether a singlephase reaction mixture is needed to exploit the advantages of a supercritical fluid, but these arguments really lie out of the scope of this chapter [25]. Suffice it to say

I

375

376

I

72 Catalytic Reactions in Supercritical Fluids

Fig. 12.5. (a) Schematic diagram o f a simple continuous reactor for heterogeneously catalyzed reactions i n SCCOZ.(b) A photograph o f the actual reactor, which occupies the area within the dotted rectangle in (a). The reactor shown in (b) is 25 c m long.

that, even in those reactions where more than one liquid phase exists, the presence of scC02 or of another supercritical fluid can significantly alter the properties of the liquid phase, for example by reducing its density or increasing the solubility of permanent gases within it. A simplified continuous supercritical fluid reactor is shown in a clear schematic in Fig. 12.5, such reactors are the basis for all the apparatus used for the heterogeneously catalyzed reactions of organic substrates described in this chapter. scC02 is delivered to the top of the pre-mixer, where it is mixed with the substrates and any other reactants that are needed for the reaction. The flow of reactants will then be carried by the scCO2 over the heated catalyst bed where reaction occurs. The products and scCO2 are then passed through the apparatus to the point where they are expanded to a pressure below the critical point of C02, to induce phase separation of products [26]. The products may then be separated by gravity from the gaseous COz, which may be recycled or vented. Clearly, the exact conditions required for reactions within these types of reactors depend on the chemistry. The important point is that the use of the SCF enables the reaction conditions to be controlled more precisely. The presence of the SCF does not necessarily make the catalyst more active, nor does an SCF normally alter the catalytic cycle. But in many cases, overall conversions and selectivities have been found to be at least equal to, if not higher than those obtained by conventional routes.

72.2 Practical Aspects of Catalytic Reactions in scCO2

I

377

In most reactions the SCF probably does not alter the activity of the catalysts. However, the SCF may nevertheless give significant advantages. By reducing the viscosity of the reaction mixture, the SCF can facilitate penetration of the substrates into the meso-/micro-pores of the catalyst support [27]. In those reactions where there is a tendency for coking, the SCF can act as solvent for the coke precursors thereby keeping the surface of the catalyst clean and maintaining its activity [28]. Surface tension effects are also important when considering reactions in SCFs. As the critical phase is achieved the interfacial tension disappears between any previously present liquid phase and a remaining solid catalyst phase. This leads to a situation where the surface tension coefficient tends towards zero; and so transport of any species is not limited by this factor. The catalysts used for heterogeneous reactions in s c C 0 ~are broadly similar to those used for reactions under conventional supercritical conditions. There are some limitations on the support material which can be used. For example, the use of small reactors requires the use of relatively high temperatures to maximize the productivity of the reactor, and some support materials (for example acid resins) cannot survive very long under these conditions. Similarly, carbon catalyst supports could react with COz at high temperatures to generate CO, which may then poison the active sites on the catalyst. More importantly, reactions with carbon dioxide may degrade the support and accelerate leaching of the catalyst. 12.2.2

Homogeneously Catalyzed Reactions

In modern synthetic chemishy, transition metal coordination compounds are often used as highly efficient and selective catalysts in homogeneous solution. These molecular catalysts are able to promote reactions that would otherwise be impossible or at least require rather forcing conditions. They often show very high selectivities for the desired products and highly enantioselective reactions are possible if chiral ligands are employed. Homogeneous catalysis has already found important industrial applications ranging from bulk chemicals to highly specialized pharmaceuticals. However, the practical utility of this attractive synthetic concept is often hampered by the inherent difficulties of separating and recovering the precious catalyst from the products and other dissolved species in the reaction medium. An important opportunity for SCFs in homogeneous catalysis relates therefore to the design of integrated processes combining synthesis with separation of products and recovery of catalysts in active form. Currently, the most successful approach to catalyst immobilization with conventional solvents is the use of so-called “biphasic systems” which rely on the separation of two immiscible liquid phases. SCFs clearly offer an attractive alternative approach, because the rich phase behavior of supercritical multicomponent mixtures provides a variety of possibilities for the combination of homogeneous reactions with subsequent selective separation stages. The combination of homogeneous catalysis and SCF extraction (SFE) using

378

I

12 Catalytic Reactions in Supercritical Fluids

Fig. 12.6.

Typical equipment for batch-wise studies o f homogeneous catalysis in scC02

scCOz has been established recently for the first time; these processes have been referred to as “catalysis and extraction using supercritical solutions” (CESS) [ 41. In addition, there are some unique features that make the combination of transition metal complex catalysts and SCFs particularly attractive. The compressibility of scCOz has been shown to allow the control of inter- versus intramolecular reaction pathways [ 291. The chemical interaction of COz with N-H groups can be used to temporarily protect secondary or primary amines [29, 30). Finally, COz can be used as a building block and incorporated in the products in certain processes [31, 321. Most studies of homogeneous catalysis in scCO2 have so far been carried out in batch reactors, such as the equipment shown in Fig. 12.6. These reactors can be coupled to on-line analytical methods including GC and GC/MS [33] or IR [ 341. Multinuclear high-pressure NMR spectroscopy can be applied to investigate catalytically-active intermediates directly in the supercritical medium [ 351. Using these methodologies, detailed insight into the molecular mechanisms of homogeneous catalysis in SCCOZhas been achieved in some cases (vide infr.). The areas of both heterogeneous [ 36-41] and homogeneous catalysis [4, 42-44] in SCFs have been extensively reviewed. In the following sections, we give a brief overview of acid-catalyzed reactions in continuous reactors utilizing heterogeneous catalysts. For comparison, we also describe a selected range of C-C bond forming reactions, which have so far been mainly the domain of homogeneous catalysts in SCFs. Finally, we discuss hydrogenation and hydroformylation in scCOz under both heterogeneous and homogeneous conditions.

12.3 Acid-Catalyzed Continuous Flow Processes in Supercritical Fluids

12.3

Acid-Catalyzed Continuous Flow Processes in Supercritical Fluids 12.3.1

Heterogeneously Catalyzed Alkylation Reactions

Friedel-Crafts reactions are commonly performed both at the laboratory and industrial scale by using acid catalysts and acyl chlorides or alcohols as acylating or alkylating agents. The advantages of SCFs for these types of reactions are clear. There is less need for reactive acid catalysts such as AlC13, which are difficult to separate from the reaction mixture and are, on the whole, environmentally unfriendly [45]. Instead, solid acid catalysts can be used efficiently [46, 471. Traditional Friedel-Crafts synthesis uses low temperatures and long reaction times to control the reaction. In many cases, reactions are performed over 24-72 h, which nevertheless produces a mixture of products that need separation to recover the desired product with appropriate purity. In the case of continuous flow reactors, however, higher temperatures are required than in a batch reactor because the residence time within small-volume reactors is comparatively low; even so, high selectivities and conversions have still been achieved at these temperatures [48]. The formation of water as a by-product from the reactions does not pose a significant problem in SCF flow reactors, as aqueous phases are easily separated from the organic phase. Water appears to be flushed from the active catalytic sites by the flow of SCCOZ.Because the residence time within the reactor is short, the products are transported away from the catalyst surface before further reaction can proceed to any great extent. This gives control over the reaction; conditions can be manipulated in a SCF flow reactor to give a high selectivity for mono substitution, which is usually the desired reaction. Scheme 12.1 below shows the Friedel-Crafts alkylation of mesitylene with isopropanol. In the scheme, the product formed from SCF Friedel-Crafts alkylation is selective to the mono-product shown. There is little or no formation of the di- or tri-substituted product. The overall conversion is 50 % [48].

Scheme 12.1.

Friedel-Crafts alkylation o f mesitylene with isopropanol.

Other examples of alkylation reactions have been shown to occur with high efficiency using scCOz as a solvent system. Clark and co-workers have successfully

I

379

380

I

12 Catalytic Reactions in Supercritical Fluids

demonstrated the alkylation of 1-butene and isobutene to produce alkylated products of trimethylpentanes and dimethylhexanes [49]. The acid catalyst used in this case was sulphonated-zirconia, and production of alkylates was shown to be continuous for nearly 48 h. When the temperature was increased to allow the reaction to occur without the use of C02 (> 135 "C), the activity of the catalyst dropped dramatically due to coking. Querini et al. have recently described a process for regeneration of acid catalysts for this reaction [ 501. Acylation reactions, however, have been less successful. This limitation is probably not due to the chemistry of the SCFs, but lies with the halide species generated in situ within the reactor when traditional acid chlorides are used as reactants. Immobilized enzymes from Candida antarctica have been shown, however, to exhibit high activity in scCOz at temperatures less than 70 "C; this enzyme is capable of acylating glucose, with a suitable acyl-donor [ 511. 12.3.2

Heterogeneously Catalyzed Etherification and De-symmeterization Reactions

Another useful class of reactions which has been successfully demonstrated as a continuous SCF process are etherfication/cyclisation reactions, again using acid catalysts [52]. Examples include the formation of asymmetric esters and heterocyclic compounds. Scheme 12.2 shows the intramolecular cyclization of 1,4butanediol over Amberlyst-15 at 120 bar, 170 "C,which produces the heterocyclic compound tetrahydrofiran (THF).1,4-Butanediol has a melting point of 16 "C, and so, although not strictly required, a small amount of methanol may be added to the reactants to aid solubility during decompression of the SCF. Adding a cosolvent to the system may also improve solubility in scC02 [13].

0

Amberlyst-15 t S C C O/ ~100 bar, 175C

+H20

The dehydration/cyclization o f 1,4-butanediol has been successfully performed using scCOz and SCCJH~; under these conditions conversion t o THF is quantitative.

Scheme 12.2.

Alkyl-diols or alcohols may undergo etherification reactions over a solid acid catalyst. It has been demonstrated that simple diols may etherify either with monoor &-substitution. In Scheme 12.3, if 1,G-hexanediolis used then a high conversion is obtained for the monomethylether (> 60 %); this is thought to be due to the difference in polarity between the mono- and di-ether species [53]. Solid Acid Catalyst R

W

O

H

scC0, or scC,H8

*

In this scheme, R may represent another alcohol group, which may possibly also undergo etherification. Scheme 12.3.

R

W

O

M

e

72.4 Homogeneoudy Catalyzed C-C Coupling Reactions

It should be noted that the product of this de-symmetrization reaction, the monoether, is of particular synthetic value, enabling selective protection of one end of the symmetrical diol, thus allowing further synthetic steps to be carried out on the remaining free hydroxyl functionality. Conventionally, this is performed by high dilution techniques, which involve large quantities of solvents, both for the reaction medium and extraction of the product. Diols may also be reacted under similar conditions to form acetals and ketals [54]. Deloxan ASP catalyst at 200 bar and 150 "C in the presence of scC02 solvent can catalyze the general reaction shown in Scheme 12.4.

HoI>\+04R Deloxan ASP

SCCOZ/ 200 bar, 150C

H

HO

Scheme 12.4.

When R = Ph; conversions as high as 89 % are possible.

The formation of acetals and ketones is again an equilibrium-driven process, which requires that the scC02 flushes water away from the catalyst active sites. This, coupled with the solubility is believed to be the driving force for the selectivity.

12.4

Homogeneously Catalyzed C-C Coupling Reactions

Metal-catalyzed coupling reactions have emerged as very powerful tools to build up carbon skeletons and have found numerous applications in the synthesis of biologically active natural or non-natural compounds. Consequently, there have been several recent attempts to use scC02 as a solvent for various processes of this type. In particular, Pd-catalyzed reactions have been the focus of a number of studies, using the Heck-coupling shown in Scheme 12.5 as a common test reaction. A variety of catalytic systems was shown to work reliably in the presence of compressed C02, affording high conversions and selectivities that are comparable to those observed in conventional solvents [55-571.

Pd(0Ac)ZlPR3IEtSN -+

@R

scCOz, 9WC, 345 bar, 12h

R=CO,Me, Ph Scheme 12.5.

Heck cross-coupling in scC02.

The Heck-reaction and related Pd-catalyzed couplings are highly useful synthetic tools, but stoichiometric amounts of inorganic salts are formed inevitably

I

381

382

I

I2 Catalytic Reactions in Supercritical Fluids

from these processes. Unfortunately, there are currently very few salt-free C-C coupling reactions of similar efficiency and scope. The hydrovinylation (i.e. the codimerization of ethylene with another olefin) is a remarkable exception that can be achieved even with high enantioselectivity if chiral nickel-catalysts are used. When the chiral information was provided by the chiral catalyst shown in Scheme 12.6, a slightly better asymmetric induction was observed in COz than in CH2Clz at comparable temperatures [ 581. The choice of the proper co-catalyst was found to be essential: similar to the hydrogenation catalysts discussed below: the nickel complex was best activated and solubilized using the BARF anion (BARF = tetrakis-[3,sbis(trifluoromethyl)phenyl]borate). Recovery of the chiral nickel catalyst using the CESS procedure was only partly successful, because the nickel species rapidly deactivate in the absence of substrate. It was noted, however, that the primary hydrovinylation product is extracted from the reactor very selectively, leaving behind any by-products resulting from double or multiple incorporation of ethylene. Again, this emphasizes the potential of scCO2 for integrated processes of synthesis and product isolation.

T = 35-40°C R

Scheme 12.6.

Ni-cat./NaBARF

Enantioselective hydrovinylation of styrenes in scCO2.

The formation of cyclic compounds of various ring sizes has been achieved also uia C-C bond forming reactions in s c C 0 ~ .Five-membered cyclic ketones were formed with similar activity and selectivity when the co-catalysed Pauson-Khand reaction was carried out in scCO2 [ 591. The co-trimerization of two molecules of an alkyne with one molecule CO2 to form six-membered lactones is an interesting approach to the use of C02 simultaneously as a solvent and a reagent [GO, G l ] . Macrocyclic compounds with ring sizes up to 16 were obtained by ring closing metathesis (RCM) of dienes in excellent yields using ruthenium and molybdenum catalysts (Scheme 12.7) [29]. These products are of considerable commercial value as olfactory substances. Their solvent-free preparation and isolation via the CESS approach is therefore of particular interest.

12.5 Hydrogenation Reactions

I

383

scco2

- C2H4 d

*

0.65g mL-1 88% yield

cat:

Ring closing metathesis in scC02 leads to macrocyclic compounds at high pressure/density.

Scheme 12.7.

It is noteworthy that the formation of the cyclic products via RCM is highly dependent on the system pressure in scCO2; ring closure occurs at high pressures, whereas oligomerization prevails at low pressures [29]. This finding is counterintuitive, because classical arguments involving activation volumes AV# would predict that high pressure favors the intermolecular pathway over the intramolecular reaction channel [621. Owing to the high compressibility of the supercritical phase, however, the pressure increase leads to a concomitant large increase in bulk density. If molar fractions are considered rather than volumetric concentrations, this increasing density leads to an effective dilution of the substrate. Low bulk densities, on the other hand, may also result in locally enhanced substrate concentrations via solute/solute clustering [63]. Both effects kinetically favor ring closure over oligomerization at high density and may cooperatively help to control the reaction towards the desired product.

12.5

Hydrogenation Reactions 12.5.1 Heterogeneously Catalyzed Hydrogenation Reactions

Reduction by direct addition of hydrogen to a material is a major industrial reaction in numerous synthetic pathways to many common products and pharmaceuticals. This process frequently suffers from mass transfer limitations, and chemists have been forced to circumvent this by applying an overpressure of hydrogen over an already saturated solvent, making the process on a large scale both slow and potentially hazardous. A continuous SCF flow reactor, however, is capable of providing a method to hydrogenate organic substrates without long residence times and limitations of mass

384

I

12 Catalytic Reactions in Supercritical Fluids

transport. In practice, a flow of scCO2 can be doped with an appropriate partial pressure of hydrogen, which is fixed for a constant flow rate of organic substrate through the reactor [64].In most cases the partial pressure of hydrogen is low comparable to the partial pressure of scC02. Under these circumstances, hydrogen is completely miscible with the scCO2, which minimizes the mass-transport limitations of the solvent system. The specific heat capacity of the mixture quickly removes the exotherm released by the majority of reductive additions. This allows the reaction to proceed, on a small scale, without the formation of hotspots within the catalyst bed. A common reaction that may be used as a test of catalytic activity is the hydrogenation of cyclohexene, as shown in Scheme 12.8. The reaction conditions are mild and the reaction is extremely quick and efficient in scCO2. Pressures as low as 72 bar may be used to achieve the supercritical state, however pressures below the critical pressure of COz may also be used; cyclohexene has successfully been reduced at pressures as low as 60 bar. Temperatures as low as 40 "C are more than adequate to initiate the reaction; however, it has been noted that the reaction is able to proceed spontaneously without heating the catalyst bed at all [64].

S C C O/ ~H,

40C / 1 2 0 b ~

Deloxan AP I1 5% (Pd) Scheme 12.8. Reaction of cyclohexene in scCOz/Hz. In a continuous SCF flow reactor a high throughput with high space-time yields may be achieved. Cyclohexene may be hydrogenated with a flow rate of substrate as high as 20 m L min-' with conversions to cyclohexane over 95 %.

Although this is a simple reaction, it illustrates the principle clearly. The phase behavior has been studied in some detail [G4].This case has a particular advantage, as both the transport of reactants and products, towards and away from the active catalytic sites, is not inhibited by phase boundaries. A more interesting example, is the hydrogenation of isophorone. This is one component of feedstock used in the perfumery industry, and the selective hydrogenation of isophorone is required to produce the desired product, 3,3,5-trimethylcyclohexanone. The advantages of SCF processes over traditional synthesis of this product are apparent. Liquid phase synthesis exhibits low conversions, which is due to mass-transport limitations, whereas gas phase synthesis requires higher temperatures and multiple passes over the catalyst bed to achieve a high conversion. This yields a range of products, which due to the close boiling points, are difficult to separate. Continuous SCF synthesis has been demonstrated to yield high conversions accompanied by high selectivities [65].

12.5 Hydrogenation Reactions

I

385

In this particular example, shown in Scheme 12.9, isophorone can be hydrogenated quantitatively by the commercially-available catalyst, Deloxan AP I1 5 % (Pd), between a temperature range of 140-200 "C. Increasing the temperature, pressure and hydrogen-to-substrateratio, the selectivity of the reaction towards the desired product falls, and hydrogenation of the carbonyl moiety proceeds as a second hydrogenation step [GS].Due to the presence of a prochiral center and the nature of hydrogenation, there is no control over the enantiomeric selectivity. 0

0

Deloxan AP II5% (Pd) SCCOZ/ Hz 120 bar 140-2OOC

Scheme 12.9.

Hydrogenation of isophorone.

Typically, hydrogenation reactions in SCFs or scCO2 are performed to yield a single reaction product by controlling the stoichiometry of the hydrogen and substrate. Once a reaction has been optimized with respect to temperature and pressure, changing the hydrogen-to-substrateratio may well yield further products. The substrate rn-cresol has three possible reductive sites, which are all based on the aromatic ring. Varying the substrate: H2 ratio at a h e d pressure has been shown to control product formation. Reaction temperature may also be used to change the product composition (see Fig. 12.7). At the lowest temperature, the addition of two equivalents of hydrogen yields the rearrangement product 3-methylcyclohexanone;whilst 3 equivalents, at a slightly higher temperature (and hence activation energy) yields 3-methylcyclohexanol. Above 300 "C further reduction of the carbon-oxygen bond occurs, which produces methylcyclohexane; whereas immediately afterwards at even higher temperatures, a dehydrogenation reaction occurs, producing the final product, toluene. Commercially-availableheterogeneous catalysts have been demonstrated to catalyze the reduction of a wide range of functional groups under SCF solvent systems; these include aldehydes,and ketones, both aromatic and aliphatic, oximes, immines, Schiff bases, nitriles and epoxides [GS]. There is also the possibility of integrating the SCF flow reactors into a synthetic pathway, as many protecting groups may be removed by the reduction of the group under an atmosphere of hydrogen. The limitation to hydrogenation in continuous flow reactors utilizing scC02 is the hydrogenation of the nitro functionality; this may not be reduced to the amine group, as the formation of solid carbamates forms blockages with the reactor pathways. This may be circumvented to some degree by the use of a co-solvent such as methanol, which immediately forms a soluble salt. However, this has only had limited success. A more practical solution is to substitute scC02 for S C C ~ H ~ , which has a lower critical pressure, but is highly flammable [GS].

386

I

12 Catalytic Reactions in Supercritical Fluids

I

"-

w *-

v

rn-Cresol 3-Methylcyclohexanone 3-Methylcyclohexanol Methylcyclohexane Toluene

s

100

150

250

200 %(wall,'

300

350

400

"C

Fig. 12.7. Graphical plot o f the reaction conditions producing variations in the selectivity o f the reaction. rn-Cresol may be hydrogenated t o form various products. The graph shows the effect o f varying the temperature. (Reproduced from reference 1651;0 American Chemical Society).

12.5.2

Homogeneously Catalyzed Hydrogenation Reactions

Since the pioneering discovery of Wilkinson's catalyst, homogeneous catalysts are widely used for hydrogenation reactions. An area of particular importance is the asymmetric reduction of prochiral C=C and C=X bonds. The well-defined molecular nature of the catalytically active center allows a rational approach to the design of effective chiral ligands that enable an effective asymmetric induction during hydrogen transfer. Nowhere is the consideration of ecological and toxicological issues more important than in the synthesis of biologically active and often chiral fine chemicals such as pharmaceuticals, food additives, agrochemicals and cosmetics. Furthermore, the enantio-differentiation during a complex multi-step catalyhc cycle relies on very small energy differences of the corresponding transition states, and tunable solvents like SCFs may hold additional promise in such processes. Clearly, there is a huge potential for the application of scC02 as a reaction medium for enantioselective hydrogenation, which has been barely explored up until now.

72.5 Hydrogenation Reactions

I

387

The first successful attempts to use chiral hydrogenation catalysts in scC02 focussed on the enantioselective reduction of dehydroamino acids as shown in Scheme 12.10. [GG, 671. A very general limitation for the use of homogeneous catalysts in scCOz was noticed during these studies, as the low solubility of many established catalysts prevented their use in this medium. This problem could be ameliorated for cationic rhodium catalysts by the proper choice of the anion, such as the BARF anion (see Scheme 12.10).Catalytic activities and selectivities in scC02 were comparable and in some cases even superior to those observed in liquid organic solvents [ 661. The neutral ruthenium complexes of the famous BINAP-ligand also had to be modified for the use in scCO2. Partial hydrogenation of the naphthyl moieties increased the solubility slightly. However, the addition of perfluorinated alcohols as co-solvents was still required to achieve results similar to those obtained in conventional solvents [67]. In a very general approach, the solubility of phosphine-containing catalysts in scC02 can also be increased dramatically by the attachment of fluorinated solubilizers in the periphery of the ligand system [ 681.

R T c o o R R'

+

H~

scCOn cat.

R'

+ CFz(CFz)&HzOH (ca. 350 equiv.) Scheme 12.10.

Asymmetric hydrogenation o f dehydroaminoacids in scCO2.

The catalytic cycle of asymmetric hydrogenation is one of the best investigated mechanisms in homogeneous catalysis and provides therefore an ideal case study to assess the influence of scCO2 on a reaction pathway at the molecular level. The specific ligand framework shown in Scheme 12.11 was chosen for a detailed mechanistic comparison between the asymmetric hydrogenation in scCO2 and conventional solvents [69]. The catalmc system shown combined all three possibilities for solubility enhancement (ligand, anion, co-solvent) in order to ensure a truly homogenous system. Labeling studies using para-hydrogen and deuterium demonstrated unambiguously that the hydrogen transfer in scCO2 occurs through the same principle pathway and involves the same intermediates as in conventional solvents. In particular, the participation of formate complexes through reversible reaction of C02 with metal hydride intermediates could be ruled out.

388

I

12 Catalytic Reactions in Supercritical Fluids

Enantioselective hydrogenation o f itaconic acid as a mechanistic probe for hydrogen transfer processes in sccoz.

Scheme 12.11.

Although the principal reaction pathway is not changed with the change of solvent, there are still significant differences between the reaction in scCOz and e.g. hexane. Whereas the enantioselectivity of the given catalyst shows a rapid decrease at lower hydrogen pressures in the liquid organic solvent, it remains uniformly high in scCOz. Because of its extremely high rate, the hydrogenation becomes mass transfer-limited in organic solvents at low H2 pressures, leading to a situation where the catalyst “starves” of hydrogen and works less selectively. This situation is avoided in scCO2 owing to the better availability of hydrogen in the supercritical phase. A remarkable increase in catalyst efficiency in scC02 as compared to conventional solvents was observed for the Ir-catalyzed enantioselective hydrogenation of imines shown in Scheme 12.12 [34]. Imine hydrogenation is a key step in the commercial synthesis of (S)-metolachlor, a commercial herbicide produced by Novartis in Switzerland. The reaction is approximately zero-order with respect to substrate in COZ whereas it slows down dramatically at higher conversion in the organic solvent (Fig. 12.8). Thus, the time required for quantitative conversion is reduced by a factor of 20 when changing from the conventional to the supercritical solvent!

scco2

N/Ph

K

Ph

CH3

/h

T = 40°C p” = 200 bar H2 (30 bar)

*

ACH3 +PhACH3

Ph

chiral Ir-catalyst Scheme 12.12.

Enantioselective hydrogenation of imines using a chiral Ir-catalyst.

12.G Hydroformylation Reactions

I

389

K

.-0 E

a,

>

0 0

0

1

2

3

4

5

6

7

8

9 1 0 2 2 2 3 2 4

time [h] Fig. 12.8. Substituting scC02 for CH2C12 increases catalyst efficiency by a factor o f 20 in the Ir-catalysed hydrogenation o f im ines.

The solubility of the catalyst drops dramatically at quantitative conversion, probably because coordination of the substrate substantially increases the solubility of the catalytically active intermediates. Thus, the product can be isolated in solventand metal-free form (Ir content < 5 p.p.m.) directly from the SCF stream upon venting, whereas the catalyst remains in the reactor for further use. This combination of catalysis and extraction using supercritical solutions (CESS process) appears highly attractive for the development of sustainable processes in fine chemical synthesis. In this particular case, the secondary amine resulting from imine hydrogenation can be separated from the catalyst by supercritical fluid extraction under conditions similar to those used during the reaction. This requires the presence of phenyl groups at nitrogen, as aryl amines are less basic and therefore less prone to formation of carbamates. For dialkyl imines, formation of insoluble carbamic acids/carbamates from the product amines is a limitation, just as discussed in the heterogeneous hydrogenation of nitro compounds (vide supra).

12.6 Hydroformylation Reactions 12.6.1 Homogeneously Catalysed HydroformylationReactions

The catalytic addition of CO and Hz (synthesis gas) to olefins (hydroformylationor 0x0-synthesis) is one of the major industrial applications of homogeneous catalysis. Over G million metric tons of aldehydes or alcohols (0x0-products) are produced worldwide per year. Commodities based on the C4 oxygenates currently have

390

I

12 Catalytic Reactions in Supercritical Fluids

a market share of more than 75 %, but aldehydes with longer carbon chains or even functionalized and chiral skeletons are receiving increasing interest including applications in fine chemical synthesis. It comes as no surprise that this industrially important reaction has been investigated intensively for a possible application of scCOz. Homogeneous catalytic systems that were found to be compatible with the use of compressed C02 include unmodified cobalt [70, 711 or rhodium [33] catalysts, phosphine-modified rhodium catalysts [ 33, 68, 72-74], and even chiral catalysts for asymmetric synthesis [35, 75, 761. Unmodified rhodium catalysts are readily formed in scCO2 from simple precursor complexes such as [(CO)2Rh(acac)],[(cod)Rh(hfacac)],or [Rh6(CO)16][33].The resulting rhodium carbonyl species are highly active in this medium for a range of substrates including simple olefins, vinyl arenes and polar substrates such as ally1 acetate. Especially the reaction rates for internal C=C bonds are remarkably higher than those observed in liquid organic solvents under typical hydroformylation conditions (Scheme 12.13).

-

CO/H2 (45 bar), T = 40°C. t = 20h

*-..1;--0-,

cat.: [(cod)Rh(hfacac)] (0.1 mol%) scCOz: > 97% conversion toluene: 23% conversion

major

+

r

minor

H

Hydroformylation of internal C=C bonds using unmodified rhodium catalysts occurs a t higher rates in scCOz as compared to conventional solvents. Scheme 12.13.

Terminal olefins are currently the most important feedstocks for hydroformylation processes. Although unmodified catalysts are also very active for this class of substrates, their relatively low cherno- and regioselectivity has prevented technical application. The use of aryl phosphines as ligands for so-called modified rhodium catalysts is known to overcome these shortcomings and modified rhodium catalysts are widely used in industrial processes. However, these ligands and their complexes are not applicable in the supercritical medium due to their insufficient solubility in scCO2. Fortunately, substitution of the aryl rings with highly “ C 0 2 philic” [77] substituents such as the (CH2)2(CF2)6F group, also provides high levels of solubility for this class of catalysts [68]. For example, I-octene can be hydroformylated in scCO2 with fast rates and high regioselectivity towards the desired linear n-aldehyde with rhodium catalysts based on such ligands (Scheme 12.14). Again, the system is applicable to a wide range of substrates with a reactivity profiles similar to that observed in conventional solvents (331. One of the key problems in hydroformylation of long chain olefins using modified rhodium catalysts is the isolation of the high-boiling products and the recovery of the expensive, thermally-sensitive catalyst in active and selective form. The rich

12.G Hydroformylation Reactions

I

391

-CHO

COH2 (20 bar), T = 6 5 T , \

H9C4-

S C C O(d ~ = 0.6 g / d , ptot= 200 bar)

H9C4

*

cat.: [(cod)Rh(hfacac)] (0.1 mol%)

+ PR3 (1.O mol%)

1-octene

TOF,,,, : diso :

430 h-’ 4.5

500 h-‘

5.6

H9C4

115 h-‘

8.5

Hydroformylation of 1-octene in scCOz using “COz-philic” ligands (TOF = turnover frequency = mole product per mole rhodium per hour).

Scheme 12.14.

phase behavior of supercritical reaction mixtures together with the selective extraction properties of scC02 allows the use of the CESS procedure as a new approach to this challenge. After the hydroformylation in scCO2 is completed, the singlephase mixture can be brought into a two-phase regime by small variations of the reaction conditions. The two-phase mixture comprises a liquid phase and a compressed (“ supercritical”)gas phase. The organometallic catalyst is contained almost exclusively in the liquid phase, whereas the organic components distribute partly into the compressed gas phase. Extraction of the mixture with C02 under these conditions removes the products selectively from the reactor, leaving the catalyst behind for subsequent use in a new run with virtually no loss in conversion or selectivity (Fig. 12.9). The overall sequence results in isolation of the product in solvent-free form with effective immobilization of the catalyst. Thermal stress for product and catalyst are kept to a minimum, as the extraction temperatures are way below those required for conventional distillation of long-chain aldehydes [ 3 3 ] . Asymmetric hydroformylation is of great current interest providing for example, viable routes to important anti-inflammatory drugs like ibuprofen or naproxen [78].The chiral phosphine/phosphite ligand (R,S)-BINAPHOS(Scheme 12.15) provides outstanding levels of enantio control in this reaction, but the established protocols require application of ecologically and toxicologically hazardous organic solvents, in particular benzene [791. The perfluoroalkyl-substituted, and hence “COz-philic” ligand 3-HZF6-BINAPHOSallows asymmetric hydroformylation of vinyl arenes and other pro-chiral olefins in scCO2 with rates and enantioselectivities that are fully comparable to those of the unsubstituted BINAPHOS ligand [ 35, 761. Product isolation and catalyst immobilization is possible with C02 as the only solvent using the CESS approach. Gratifymgly, the regioselectivity towards the desired branched chiral aldehyde is even higher under the new reaction conditions (Scheme 1 2 4 , leading to a significant increase in selectivity towards the desired single stereoisomer. As in hydrogenation, high-pressure spectroscopy revealed that

392

I

72 Catalytic Reactions in Supercritical Fluids 100

80

E

Q Q

20

0 1

2

3

5

4

cycle Catalyst recycling in the hydroformylation o f 1-octene using the CESS procedure. Reaction conditions as i n Scheme 12.14.

Fig. 12.9.

the same key intermediates are formed in scCO2 as in conventional solvents. The increase in regioselectivity does not result, however, from a COz-inducedchange in the kinetics, but could be unambiguously related to the substitution pattern of the “C02-philic”ligand. H CHO

CO/H, (60 bar), T = 60T, = 16 h

Ph/--

cat: [(cod)Rh(hfacac)] (0.05 mol;) + Lig. (0.1 mol%)

-CHO

+

Ph Ar =

a

Ph

xCH3 (R)

88% iso, 94%ee

BINAPHOS, benzene

d(cHddcFd8

Lig.

Ar =

93% iso, 92%ee

3-HZF6-BINAPHOS,scC0, Asymmetric hydroformylation o f styrene i n scCO2 using 3-H2F6-BINAPHOS leads to higher overall selectivity as compared to BINAPHOS in benzene.

Scheme 12.15.

If hydroformylation is carried out in the presence of primary or secondary amines, condensation with the initially formed aldehyde and subsequent hydrogenation of the resulting imine or enamine can occur under suitable conditions. The overall reaction sequence is referred to as hydroaminomethylation. The intramo-

72. G Hydroformylation Reactions

lecular version of this sequential transformation using allylic amines would be an interesting synthetic approach to pyrrolidine compounds. In conventional solvents, however, the neighboring N-H group leads to a different reaction pathway under exclusive formation of five-membered cyclic amides. This course of reaction can be shut down very effectively if the intramolecular hydroaminomethylation of allylic amines is carried out in SCCOZas the reaction medium (Scheme 12.16) [30].

/ dioxane:

sccoz :

>99 8

/ : :

<1 92

Hydroaminomethylation o f allylic amines leads t o cyclic amides in conventional solvents, but pyrrolidines are formed preferentially in scCOz. Scheme 12.16.

The remarkable change in chemoselectivity shown in Scheme 12.16 results from the reversible interaction of secondary amines with C02 to give the corresponding carbamic acids, as shown by high-pressure NMR spectroscopy [30]. We have discussed several examples above, where this equilibrium played a role for catalyix reactions in scC02, either by deactivating or protecting the catalyst. In the present case, the reactivity of C 0 2 proves beneficial and can be exploited to control the reaction pathway. 12.6.2

Heterogeneously Catalyzed Hydroformylation Reactions

Research in hydroformylation has also been driven towards a catalyst that may be immobilized on a surface, without losing activity, but is as active as an homogeneous counterpart. Deloxan rhodium (HK I 2 % (Rh))-basedcatalysts may be used for these reactions. It has been shown that oct-1-ene can be hydroformylated in a continuous flow reactor using scCOz as a solvent. It is possible to obtain a total conversion of up to 99.3 %, however, only an actual yield of 88.5 76 aldehyde is achievable. Lower temperatures have been shown to increase the n : is0 ratio of the reaction, which can reach 2.2 : 1. There have been several reports of catalytic materials that possess promising properties, which have now been shown to work within a flow reactor system without loss of activity [go]. Perhaps this is one of the most impressive successes of SCF as solvent systems in a continuous reactor to date. The large-scale production of aldehydes by the petrochemical industry is currently based upon an aqueous extraction, which is dependent on solubility of organic compounds in water [N]. This limits the production mainIy to aldehydes from olefins with chain lengths of less than five carbons. The discovery that higher alkanes may be hydro-

I

393

394

I

72 Catalytic Reactions in Supercritical Fluids

Fig. 12.10.

The structure o f the immobilized species present in the Deloxan HK I (Rh) catalyst.

formylated selectively with a high n: is0 ratio in a scCOz medium utilizing an immobilized rhodium diphosphine complex (Fig. 12.11) provides a promising new approach to address this problem [80].The large bite angle of the catalyst is thought to be an contributing factor to both the stability and the high n: is0 ratio, which was found to be as high as 50: 1. The advantages of this particular catalyst system is that it retains its high turnover frequency and activity without modification, nor is the activity diminished by the immobilizing process or by metal leaching at high CO overpressures; a problem which was a limiting factor with most of the former catalysts for such a system [82).

12.7

Closing Remarks

The unique solvent properties of supercritical fluids hold many opportunities for new and innovative applications in catalysis. Heterogeneous and homogeneous

Fig. 12.11. Rhodium complex used t o generate an immobilized catalyst for continuous flow hydroformylation in scCO2.

References 1395

catalytically active materials can show largely improved performance under these conditions as compared to conventional gas phase or liquid phase processes. In both areas, we will also see increasing efforts to move towards continuous flow processes in fine chemical applications. Supercritical fluids are ideally suited as mobile phases for processes of this type, as they are readily removed from the product by simple pressure reduction. The favorable properties of carbon dioxide in terms of acute toxicology, environmental hazard, process safety, and material costs make this particular solvent especially attractive for “Green” synthetic processes. The acceptance of this technology in chemical synthesis is expected to increase as scC02 finds increasing application as a solvent in other chemistry-related industries also [83].

References

M. POLIAKOFF, N. J. MEEHAN, S. K. Ross, Chem. Ind. 1999, 19, 750. 2 M. POLIAKOFF, M. W. GEORGE,/.Phys. Org. Chem. 1998, 12, 589. 3 P. G . JESSOP, w . LEITNER, Chemical Synthesis Using Supercritical Fluids, Wiley-VCH, Weinheim, 1999. 4 W. LEIPNER, Top. Cur. Chem. 1999, 1

206, 107. 5

W. LEITNER, Appl. Organomet. Chem.

2000, 24, 809. 6 C. BOIX,N. POLIAKOFF, J. Chem. Soc.-Perkin Trans. 11999, 1487.

A. HOURRI,J. M. ST-ARNAUD, T. K. BOSE,Rev. Sci. Instrum. 1998, 69, 2732. 8 X. S. LANG,A. AKGERMAN, D. B. BUKUR,Ind. Eng. Chem. Res. 1995, 34, 7

72. 9

S. VAN DEN HARK,M. HARROD,P. MBLLER, /. Am. o i l Chem. SOC. 1999,

S. C. TUCKER, M. W. MADDOX,].Phys. Chem. B. 1998, 102, 2437. 17 K. K. LIONG,N. R. FOSTER, Ind. Eng. Chem. Res. 1991, 30, 1329. 18 R. C. DOWNING, Fluorocarbon 16

Refigerents Handbook, Prentice-Hall, Englewood Cliffs, NJ, 1998. 13 D. P. FERNANDEZ, Y. M. GOODWIN, J. M. H. SENGERS, A. R. H. LEVELT, /. Phys. Chem. Ref: Data. 1995, 24, 33. 20 D. C. WYNNE, M. M. OLMSTEAD, P. G. JESSOP, /. Am. Chem. Soc. 2000, 222, 7638.

SEDLBAUER, J. P. O’CONNELL, R. H. WOOD,Chem. Geol. 2000, 163, 43. 22 P. T. ANASTAS, T. C. WILLIAMSON (Eds), Green Chemistry:Frontiers in Benign Chemical Synthesis, Oxford University Press, 1998. 23 K. ZOSEL,Angew. Chem. Int. Ed. Engl. 21 J.

1978, 27, 702.

76, 1363. 10

11 12 13

14 15

T. ADSCHIRI, R. SHIBATA,T. SATO,M. WATANABE, K. ARAI,Ind. Eng. Chem. Res. 1998, 37, 2634. R. C. CRITTENDON, E. J. PARSONS, Organometallics 1994, 13, 2587. K. S. JEROME, E. J. PARSONS, Organometallics 1993, 12, 2991. A. D. M. IUPAC Recommendations. Blackwell Science Ltd., Oxford, 1997. J. C. GIDDINGS,L. MCLAREN, R. A. KELLER, Science 1968, 162, 67. L. D E V E ~ P. A ,CANU,A. BERTUCCO, K. STEINER, Chem. Eng. Sci. 1997, 52, 4163.

24

H. PEKER,M. P. SRINIVASAN, J. M. SMITH,B. J. McCoy, A I C H E ] . 1992, 38, 761.

25

N. F. LUCIEN,Chemical Synthesis Using SCFs., 1St edn, Wiley-VCH, Weinheim,

1998. E. M. RUSSICK, C. L. J. ADKINS,C. W. DYCK,ACS. Sym. Ser. 1997, 670, 255. 27 B. S. BAPTIST-NGUYEN, AICHE]. 1992, 38, 1027. 26

28

F. TRABELSI, F. STUBER,K. ABAROUDI, M. A. LARRAYOZ, F. RECASENS, J. E. SUEIRAS,Ind. Eng. Chem. Res. 2000, 39, 3666.

396

I

12 Catalytic Reactions in Supercritical Fluids 29 A. FURSTNER,D.

KOCH, K. LANGEMANN,W. LEITNER, C. SIX, Angew. Chem. Int. Ed. Engl. 1997. 36,

46 47

2466. 30

K. W I ~ M A N W. N , WISNIEWSKI, M. MYNOIT,W. LEITNER, C. L. KRANEMANN,T. RISCHE,P. EILBRACHT, S. KLUWER, J. M. ERNSTING, C. J. ELSFVIER,2001. 7,

36, 1458. 48

W. I.EITNER, Coord. Chem. Reu. 1996,

50

153. 257. 32 33 34

P. G. JESSOP, T. [KARIYA, R. NOYORI, Science 1995, 269, 1065. D. KOCH, W. LEITNER, J . Am. Chem. Soc. 1998, 120, 13398. S. KAINZ, A. BRINKMANN, W. LEITNER, A. PFALTZ. J. Am. Chem. Soc. G. FRANCIO,K. WIITMANN,W. LEITNER,/. Organomet. Chem. 2001,

135.

TSITSIMPIKOU, H. STAMATIS, V. SERETI, H. DAFLOS, F. N. KOLISIS,]. Chem. Technol. Biotechnol. 1998, 71,

309. 52

53 54

621, 130. 36

B. SUBRAMANIAM, M. A. MCHUGH: lnd. Eng. Chem. Proc. Design Deuelop.

55

1986, 25, 1. 37 38

B. SUBRAMANIAM, M. A. MCHUGH, Ind. Eng. Chem. Res. 1987, 26, 1267. K. M. DOOLEY, S. R. BRODT,F. C. KNOPF,Ind. Eng. Chem. Res. 1987, 26.

56

57

1267. 39 J . F.

BRENNECKE, A C S Symposium

Series 1993, 514, 201. 40 41 42

43

R. MARR,T. GAMSE, Chem. Eng. Process. 2000, 39, 19. J. A. CUSMANO, Chemech 1992. 48, 2. D. A, MORGENSTERN, R. M. LELACHEUR, D. K. MORITA,D. L. BORKOWSKY, S. FENG,G. H. BROWN, L. LUAN,M. G. GROSS,M. J. BURK,W. in Green Chemistry, P. T. TUMAS, ANASTAS, T. C. WILLIAMSON (Eds), Am. Chem. Soc., Washington DC, ACS Symp. Ser. 1996, 626, 132. E. DINJUS,R. FORNIKA, M. SCHOLZ, in Chemistry under Extreme or NonClassical Conditions, R. VAN ELDIK, C. D. HUBBARD (Eds),Wiley, New York,

58

59 60

61

624, 602.

64 65

C. J. LYON,B. SUBKAMANIAN, C. J. PERIERA,Abstr. Pap. Am. Chem. Soc.

66

1999, 218, 22.

W. K. GRAY,F. R. SMAII.,M. G. Ross, s. K. HITZLER, M. POLIAKOFF, /. Am. Chem. Soc. 1999, 121, 10711. F. A. MESKINS, Synthesis 1981, 501. K. NAKAGAWA, M. 0. N. IKENAGA, T. SUZUKI, T. KOBAYASHI,Chem. Commun. 1998, 1025. M. A. CARROLL, A. B. HOLMES, Chem. Commun. 1998, 1395. N. SHEZ.AD, R. S. OAKES, A. A. CLIFFORD, C. M. RAYNER,Tetrahedron Lett. 1999, 40, 2221. D. K. MORITA,D. R. PESIRI,S. A. DAVID,W. H. GLAZE, W. TUMAS, Chem. Commun. 1998, 1397. A. WEGNER, W. LEITNER,Chem. Commun. 1999,1583. N.J E O N G , s. H. HWANG,J. s. LIM,/. Am. Chem. SOC.1997, 119, 10549. M. T. RE=, W. KONEN,T. STRACK, Chimica. 1993, 47, 493. U. KREHER, S. SCHEBESTA, D. WALTHER, 2. Anorg. AIlg. Chem. 1998,

62 0. REISER,Top. Catal. 1998, 5, 105. 63 S. PANDEY, M. A. KANE,G. BAKER,

1996, 219. 44 P. G. JESSOP, T. IKARIYA,R. NOYORI, Chem. Rev. 1999, 99, 457. 45

M. C. CLARK, B. SUBRAMANIAM, lnd. Eng. Chem. Res. 1998, 37. 1243. C. A. QUERINI. Catal. Today 2000, 62,

51 C.

1999, 127,6421. 35

M. G . HITZLER, F. R. SMAIL, S. K. Ross, M. POLIAKOFF, Chem. Commun. 1998, 359.

49

4584. 31

Y. GAO.Z. N. ZHU,W.K. YUAN,Prog. Nat. Sci. 1996, 6>625. L. FAN,I. NAKAMUKA,S. ISHIDA,K. FUJIMOTO, Znd. Eng. Chem. Res. 1997.

F. V. BRIGHT,H. SEIDEL, A. FURSTNER, W. LEITNER,].Phys. Chem. B. 2002, in press. M. G. HITZLER, M. POLIAKOFF, Chem. Commun. 1997, 1667. M. G. HITZLER, F. R. SMAIL,S. K. Ross, M. POI.IAKOFF, Org. Process Res. Den 1998, 2, 137. M. J. BURK,S. FENG,M. F. GROSS,W. TUMAS,].Am. Chem. Soc. 1995, 117, 8277.

67

J. XIAO,S. C. A. NEFKENS, P. G.

T. IKARIYA, R. NOYORI, Tetrahedron Lett. 1996, 37, 2813. 68 S. KAINZ,D. KOCH,W. BAUMANN, W. LEITNER,Angew. Chem. Int. Ed. Engl. 1997, 36, 1628. 69 S. LANGE, A. BRINKMANN, P. TRAUTNER, K. WOELK, J. BARGON, W. Chirality 2000, 12, 450. LEITNER, 70 J. W. RATHKE, R. J. KLINGLER, T. R. KRAUSE, Organometallics 1991, 10, 1350. 71 Y. Guo, A. AKGERMAN, I . Supercrit. Fluids 1999, 15, 63. 72 D. R. PALO,C. ERKEY, Ind. Eng. Chem. Res. 1998, 37, 4203. 73 D. R. PALO,C. ERKEY, Organornetallics 2000, 19, 81. 74 M. F. SELLIN, D. J. COLE-HAMILTON,]. Chem. Soc. Dalton Trans. 2000, 11,1681. 75 S. KAINZ,W. LEITNER, Catal. Lett. 1998, 55, 223. JESSOP,

G. FRANCI~, W. LEITNER,Chem. Commun. 1999, 1663. 77 W. LEITNER, Nature 2000, 405, 129. 78 F. AGBOSSOU, J. F. CARPENTIER, A. MORTREUX, Chem. Rev. 1995, 95, 2485. 79 K. NOZAKI, N. SAKAI,S. MANO,T. HIGASHIJIMA, T. HORIUCHI,H. TAKAYA, 1.Am. Chem. Sac. 1997, 119, 4413. 80 N. J. MEEHAN, A. J. SANDEE, J. N. H. REEK,P. C. J. KAMER, P. VAN LEEUWEN, M. POLIAKOFF, Chem. Commun.2000, 1497. 81 E. G. KUNTZ,Chemech 1987, 17, 570. 82 M. BELLER, C. D. FROHNING,C. W. KOHLPAINTER, J. Mol. Catal. A-Chem. 1995, 104, 17. 83 J. M. DESIMONE, D. K. TAYLOR, R. CARBONELL, Ann. Rev. Energy. Environ. 2000, 25, 115. 76

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

398

I 13

Application o f Supercritical Fluids in the Fine Chemical Industry Werner Bonrath” and Reinhard Karge 13.1 Introduction

Supercritical fluids, for example scCO2, could be an alternative for use as a process solvent because supercritical fluids have advantages, for example in terms of environmental benefits (scCO2) no waste

- no damage of ozone layer . non-carcinogenic . non-toxic . nonflammable

From the viewpoint of chemical processes, the main advantages of supercritical fluids are

---

high diffusion rates easy separation of products in many cases no solvent residues use of a thermodynamically stable solvent low costs for the solvent, when appropriately recycled

The application of supercritical fluids, for example scC02, as an environmentally acceptable replacement for conventional solvents, is well documented in the industry. Based on the work of Zosel, the decaffeination of coffee and tea using scCO2 was the first industrial use of this technology [l].The advantages of supercritical fluids are not only useful in separation techniques, for cxample supercritical fluid extraction (SFE) or supercritical chromatography (SFC), their application as process solvents is well recognized [2, 31. For new industrial processes, economic and ecological aspects are closely related: production of waste is per se not an economic achievement. Therefore environmentally friendly or so-called “green chemistry” is gaining increasing recognition.

13.1 Introduction

I

399

Tab. 13.1.

The €-factor in various segments of industry 15-81,

Industry segment

Product tonnage

kg by-product/ kg product

Oil refining Bulk chemicals Fine chemicals Pharmaceuticals

106-10s 104-106 102-104 101-103

<0.1

1-5 5-50 25->lo0

In this context sc-fluids could offer sustainable contributions in a variety of potential applications. The production volume as well as the atom economy is widely used for the classification of chemical processes. The atom economy describes the number of atoms of all starting materials, which are transferred into the product [4].But these criteria include not all aspects of chemical processes, for example loss of solvents (amount of waste), type of waste and energy balance. An additional aspect to include in these criteria is the E-factor (environmental factor). The E-factor was first discussed by R. A. Sheldon and is defined as kg by-product per kg product. The E-factor is associated with type of waste, because 1 kg of sodium chloride or 1 kg of sodium cyanide or chromium oxide will have different environmental impacts as waste products [S-81. The fine chemical industry is normally based on organic chemistry and the manufacture of fine chemicals is often based on stoichiometric reactions, and therefore large amounts of waste are generated. Examples for this are the use of acids, the Friedel-Crafts reactions (acylation, alkylation), toxic chemicals like phosgene, or the application of chlorinated hydrocarbons. Since the fine chemical manufacturers are facing global competition, the production process with the best economy and ecology will result in a leading market position. Therefore, for the future of the fine chemical industry it is necessary to develop more efficient processes, which means efficient in energy and raw materials consumption and a minimum of waste production. The driving forces for process R D concerning reactions, separation/purification and new product forms are:

+

-

improve process economics (running costs and/or investment) new products, new qualities competitive advantage through superior technology health, safety, environment demonstration of the payback or potential payback

Furthermore, in many cases continuous processes with high productivity, selectivity, and conversion are superior to classical batch processes, especially when several unit operations can be combined in a single sequence.

400

I

13 Application of Supercritical Fluids in the Fine Chemfcal Industry

As mentioned earlier, supercritical fluids have a broad potential for application in new processes. In recent years a number of publications reviewed the role of supercritical fluids in technical applications 19-1 G] and more industrial applications of sc-fluids were established in the synthesis of fine chemicals. In a variety of reactions such as hydrogenations. hydroformylations and Friedel-Crafts reactions, the advantageous use of sc-fluids as solvents has been demonstrated with respect to yield, selectivities, and no work-up procedure [17]. In the following sections some aspects of (potential) applications of sc-fluids in the fine chemical industry with respect to product separation/purification and catalytic reactions are discussed. Earlier industrial applications of supercritical fluid reactions, for example the Haber-Bosch process for the synthesis of ammonia, synthesis of methanol from hydrogen and carbon monoxide, or the polymerization of ethene will not be discussed. An extensive overview on the use of sc-fluids in the synthesis of bulk chemicals is given in the book edited by Jessop and Leitner [12].

13.2

Supercritical Fluids in Separation/Purification

Following the successful commercialization of decaffeination with SCCOZ,further applications of sc-fluids in the field of separation/purification of natural products have been investigated. In addition, product purification in synthetic processes also came into focus. The major challenge was to specifically exploit the unique physical properties of supercritical fluids to solve those separation problems which are difficult by “classical” approaches. Therefore, sc-fluids have been investigated for the extraction of solids as well as liquids. In addition, supercritical fluid chromatography has been studied as an alternative to “classical” HPLC methods. 13.2.1

Supercritical Fluid Extraction (SFE)

Supercritical fluid extraction (SFE) is a separation technique that uses sc-fluids as separating solvents. Supercritical fluids can replace other solvents in many purification procedures, even in countercurrent extraction. In synthetic chemistry, SFE can be an alternative to conventional methods for purification/isolation of complex products, for example pharmaceuticals, nutraceuticals and vitamins [ 12, 181. Since SFE is still quite a “young” discipline, physical properties and basic parameters for many interesting compounds and mixtures are not yet known (in contrast to “classical” methods like distillations). Therefore, it must be pointed out that for all applications of sc-fluids the phase equilibria have to be determined properly. Unfortunately, for many technical or industrial applications of procedures based on supercritical fluids, the basic parameters are often not yet known. For industrial implementation, scale-up, miniplant, or pilot plant activities, it is absolutely necessary to have information about phase behaviour, solubility, energy balances and

73.2 Supercritical Fluids in Separation/Pur~cation

I

401

mass balances. It would also be helpful to have comparative experiments using “classical” solvents to demonstrate the benefit of supercritical fluids. In many publications these data are missing and therefore extensive research work is still necessary. As pointed out, in the isolation of products from natural sources the application of sc-fluids could be useful. Chang et al. described the separation of green tea catechins using scC02 I191 (Scheme 13.1). The separation of caffeine and polyphenols was achieved by controlling the extraction conditions. The most important compound of these polyphenols is epigallo-catechin-3gallate (EGCG). Numerous studies have demonstrated that green tea is an anti-oxidant and a cancer-preventive

OH

I

OH

OH

OH

epigallocatechin-3gallate

epicatechin

OH

p“

epicatechin-3 gallate OH

epigallocatechin Scheme 13.1.

I OH

Green tea catechins.

The extraction of interesting purine alkaloids like theobromine or theophylline using scCO2 has been described [Zl]. The isolation of b-carotene and vitamin E from crude palm oil using a scCO2 counter-current extraction, after esterification of the starting material with methanol, was described by Jungfer and Brunner [22]. Based on the measurement of the solubilities of a-tocopherol, palmitic acid, and tripalmitin, the extraction of a-tocopherol was studied [23]. This investigation

402

I

73 Appkation ofSupercritical Fluids in the Fine Chemical lndustry

I

Tocopherol Tocopherol acetate-

I

Tocopherol Tocopherol acetate end: >98 %, yield -98 %

start: -85%

1

Low thermal stress, high yield Fig. 13.1.

Purification by Supercritical Fluid Extraction (SFE)

showed that a-tocopherol could be extracted in an energy-efficient manner simultaneously with fatty acids. The purification of crude synthetic tocopherol acetate by SFE (SCCOZ as solvent) in a continuous process resulted in high quality material [24] (Fig. 13.1). The process was carried out at 323, 333 and 343 K on 13.6-m columns with a diameter of 35 mm in a pressure range from 16 to 28 MPa. In two separation steps the material was purified to a minimum content of 97 wt-%. Advantages of this procedure are the low thermal stress and a high yield. The disadvantages are located in the high reflux ratio, the low throughput, and a highenergy consumption for gas recycling. These may be by-passed by applying other gases (for example propane) or using a different separation method, for example adsorption/desorption. Another method for applying SFE is the combination of extraction and adsorption. In Fig. 13.2 the flow-sheet of an apparatus for adsorption experiments is shown. As an alternative to the energy-consuming SFE process with high-pressure expansion as the separation principle of solute and solvent, adsorption on silica under high pressure was investigated. The desired 50 % of loading could be easily reached by adsorption under high pressure. A 100 % recovery of solvent was achieved, and energy costs were decreased because of the higher pressure level in comparison to conventional separation by pressure expansion [ 251. By combination of SFE and adsorption, a formulated high quality product was gained with decreased production costs in comparison to an SFE process with conventional separation of solute solvent by pressure reduction. Poly-unsaturated fatty acids (PUFA), especially n-3 fatty acids, are well known for their health benefits [26]. Fish oil as starting material for the separation of PUFA is a cheap source, but its fractionation is difficult. Separation techniques based on urea precipitation or silver ion chelate formation are less favored because of waste

I

13.2 Supercritical Nuids in Separation/Purification

403

From Column

I

Sample

Gas flow, back to gas recycle and column

Fig. 13.2.

Combination of extraction and adsorption.

and residue problems. Rectification needs high temperatures with risk of thermal damage of the all-Z-configureddouble bonds. Exploiting the unique phase behavior of sc-gases in a sophisticated manner in a counter-current extraction process, the separation of these temperature-sensitive products could be achieved [271. The separation of CIS- and Czo-compounds,means separation by chain length, at GO "C and 14.5 MPa [28] (Fig. 13.3). Based on phase equilibrium data, a pilot plant for measuring the concentration profile along the column was built, and pilot production on a several-thousand-kg scale was undertaken (Fig. 13.4). 13.2.2

Supercritical Fluid Chromatography (SFC)

Supercritical fluid chromatography (SFC) uses sc-gases as mobile phase [ 29-32]. In contrast to classical HPLC, the recycling of the solvent is much easier. The sep-

I

-

-

PUFA- ethylesters starting with 1045% EPA and 10-15% DHA

-

-IFig. 13.3.

Purification of PUFA-ethylesters.

PUFA-ethylesters EPA>65% DHA > 65 % purity in chain length: > 90%

404

I

73 Application of Supercritical Fluids in the Fine Chemical lndustty

COOH

a-linolenic acid

COOH

eicosapentaenicacid

COOH docosapentaenoic acid

COOH

docosahexaenoic acid Scheme 13.2.

Polyunsaturated fatty acids.

aration of the compounds is achieved by pressure release. Modifiers, for example organic solvents, are only used in minor quantities and there is therefore a potential for using SFC on a production scale. In the current vitamin D3 process (see Scheme 13.3), 7-dehydrocholesterol is transformed by ultraviolet light (mercury lamp) to precalciferol. The pre-vitamin D is in equilibrium with its dosed-ring photoproducts, starting material, lumisterol, and the E/Z isomers to Tachysterol. In the next step, a thermal reversible isomerization, 1,7-hydrogen shift from C19 to Cg, is achieved by heating the solution of the photoproducts at 80 “C. After concentration and recovery of the starting material, tachysterol is removed by a Diels-Alder reaction (dieneophile maleinic anhydride). This process has the disadvantage that the “wrong” isomers are separated by the reaction, which means that it is a costly procedure and involves a loss in yield [33]. A procedure for the isolation of vitamin D3 from a mixture of steroids, for example dehydrocholesterol, lumisterol, and tachysterol using S FC (SCCOZand an optional modifier) was described by Johannsen [34] (Fig. 13.5). The advantage of this new procedure is that the “wrong” isomers are recycled. This resulted in a more efficient process and a minimum of waste. A block diagram is shown in Fig. 13.6. The experiments are carried out using scCOz (and modifier) at 100 to 200 bar, the column has the dimensions of 400 x 30 mm, at a

73.3 Catalytic Reactions in Supercritical Fluids

I

405

(4)

(8)

3.Floor Rectificationcolumn

Buffer

I,.1

(9)

m

2. Floor

I

I

Column A 35 mm

1. Floor

(2) Column diameters: 35 mm, 50 mm, 70 mm height about 13 m; pressure 5 300 bar temperature 5 150 C C02, but also ethane, propane or mixtures

Fig. 13.4.

SFE unit.

flow rate of 100 to 200 g COZ per minute. The production of 1kg product per week can be achieved.

13.3

Catalytic Reactions in Supercritical Fluids

As mentioned in the Introduction, sc-fluids find increasing use as reaction media. By exploiting the solvent properties of sc-fluids, it is possible to enhance reaction rates or improve selectivity. Especially in cases, where mass transfer problems are limiting, for example reaction rates or selectivities, the special properties of the supercritical state may circumvent these difficulties. To use the advantages of sc-fluids in reaction medium, it is necessary to have information on the phase behavior that is exhibited by the reaction mixture. Supercritical fluids have been used as solvents for various inorganic/organic reactions [35]. However, only a few examples (see Introduction) of the application of sc-fluids in supercritical fluid reactions (SFR) have been described. From an industrial point of view, the potential of many of the described reactions could not be realized, because there are not enough details for the development and transformation of the reaction to a tech-

406

I

13 Application afSupercritica1 Fluids in the Fine Chemical Industry

Lumisterol

Previtamin D3

I/

U

O

H

Tachysterol A

p

R=

%-,.,

Vitamin D3 Scheme 13.3.

Synthesis of vitamin D3.

nical scale. In a number of cases there is low solubility of substrates, products and/ or catalysts, for example in scCO2. In addition, in many cases it is necessary for the entire system to be in the supercritical state, not just the solvent. SFR could be advantageous based on the viscosity, heat- and mass-transfer properties, camed out preferably in continuous reaction systems. 13.3.1 Hydrogenation

Catalytic hydrogenation is one of the most important reactions in industry. This type of reaction can be carried out in an homogeneous or heterogeneous manner. Hydrogenation in scCO2 is based on the solubility of H2 and C02 above 304 K [ 361. Carbon dioxide itself can be hydrogenated to yield formic acid [15, 371. In this homogeneous reaction, scCO2 has found an industrial application. One of the advantages of this process is the coupling of the hydrogenation step with esterification or addition of amines followed by dehydration, resulting in N,Ndialkylformamides (see Scheme 13.4) [38, 391. Homogeneous hydrogenations in scCO2 could also be carried out without involving a reaction with the solvent. Investigations on the asymmetric hydrogenation of tiglic acid using a Ru(OAc)2-BINAP catalyst (BINAP = 2,2'-bis(dipheny1-

13.3 Catalytic Reactions in Supercritical Nuids

I

\-

, I

I1

FEED

Fig. 13.5.

Preparative SFC.

sterole (DHC)

Waste

SFC- Separation, recycling of the 'lwrong" isomers

I

407

408

I

73 Application ofSupercritical Fluids in the Fine Chemical Industry

CO,

+

Hz

\

-

Ru-cat. 80 bar H,

*

HCOZH

T=373K

Scheme 13.4.

HC0,Me

+

H,O

NEt,, MeOH T=353K

HC(O)NMe,

+

HzO

Synthesis of formic acid and derivatives.

phosphin0)-1,l-binaphtyl) yielded an enantiomeric excess (ee) comparable to that obtained in liquid organic solvents [12, 401. Problems encountered with such reactions in sc-fluids are that ligands or metal salts are not soluble in scCOz,or their solubility is too low. Using fluorinated alcohols in scCOz resulted in an increase in the ee. However, it is helpful to have data on the solubility of compounds. For example, solubility measurements on copper, yttrium, and barium acetylacetonates and hexafluoro acetylacetonates have been carried out in supercritical C02 at 423 and 447 K at a pressure range between 120 and 220 bar. The thermal behavior of these solutions has been studied, as well as the influence of ethanol as the added solvent [41]. For this purpose the solubility of several acetylacetonate metallorganic compounds in supercritical carbon dioxide at 313 K and pressures up to 62 MPa have been studied. Moreover, the solubility effect of several polar entrainers has been studied along with the effect of process conditions on properties and characteristics of crystals and coatings obtained when expanding and pyrolyzing the above-mentioned supercritical saturated mixtures (Table 13.2). It has also been found that acetone and other entrainers including water, DMF and ethanol, have no effect at all on the solubility of the three insoluble metallorganic compounds [42]. The catalybc hydrogenation of unsaturated ketone in supercritical carbon dioxide with a supported palladium catalyst, was described by Bertucco et al. [43] (Scheme 13.5). ChemicaI kinetic information was obtained using an integral-reactor and the experimental results were interpreted with an homogeneous model. Preliminary observations of phase behavior may be useful for reactor optimization purposes before any information about kinetic phenomena becomes available. The phase boundaries for the system C-Hz-COz are shown in Fig. 13.7. One should keep in mind that variation of the COz amount could be applied to reduce the two-phase region. This effect is shown in Fig. 13.8. The conclusions mentioned above were drawn for a 1 to 1 solvent ratio, but addition of C02 to the system narrows the two-phase area and results in smaller retrograde behaviour.

13.3 Catalytic Reactions in Supercritical Fluids

Solubilities of acetylacetonates [42].

Tab. 13.2.

Solid

Entrainer

p (MPaJ

T (K]

Wss

We

Fe(acacJ3 Fe(acacJ3 Fe(acac), Fe(acacJ3 Fe(acac)3 Cr(acac)j Cr(acac), Cr(acac)3 Cr(acac)3 Zr(acac)s Ni(acac)z Co(acac)z

-

15.0 30.0 30.0 30.0 40.0 30.0 30.0 30.0 40.0 30.0 61.0 62.0

313.1 313.1 313.2 313.1 313.1 313.1 313.1 313.1 313.1 313.1 313.4 313.5

0.110 0.181 0.232 0.233 0.261 0.113 0.116 0.116 0.120 0.004 0.005 0.004

-

-

Acetone Acetone

-

Acetone Acetone -

-

-

0.5 1.0

0.5 1.0

-

Wss = weight percent of considered metallorganic compound (which is at saturation) We = weight percent of entrainer

Allene Scheme 13.5.

Hydrogenation reaction o f an unsaturated ketone.

500

400

300

e! 3 w 10

e

2PO

CL

7 00

0 0

Too

200

Temperature PC] Fig. 13.7.

Phase boundaries for the ternary system CGH2-CO2.

300

400

I

409

410

I

13 Application ofSupercritical Fluids in the Fine Chemical Industry

500

400

,--i

........... ... . .

% 300 a

I....

?? 3

g

200

a" 100

0 0

100

300

200

400

Temperature ["C] Fig. 13.8.

P-T-diagram for the ternary system C-Hz-CO2

1441.

In comparison to the industrial batch process at low pressure and with mass transfer controlled, continuous supercritical hydrogenation has the advantage of accelerating the reaction rate and product formation very significantly. Small-scale laboratory experiments initially indicated an increase in product formation of approximately 500-fold. Continuous hydrogenation of organic substances like cyclohexene, acetophein scCOz using scC3Hb, has been none, and I&( methylenedioxy)-4-nitrobenzene described by Poliakoff et al. [45](see Chapter 12).The advantages of sc-fluids for controlling the reaction conditions to achieve a better selectivity was shown, but the results were not compared to those achoeved under normal pressure conditions. Heterogeneous asymmetric hydrogenation of ethyl pyruvate in S C C ~using H ~ Pt/ alumina modified with cinchonidine as the catalyst has, in comparison to the reaction in toluene, the benefit of a shorter reaction time [4G] (Scheme 13.G).

scc,n6 313 K 24 h

0 Scheme 13.6.

5 % Pt/Al*O, cinchomdine

H+,

*

+OW

0

Enantioselective hydrogenation of ethyl pyruvate.

Another approach to overcome the solubility problem and the transport limitations of hydrogen was described by Harrod et al. [47].Poliakoff used near-critical

13.3 Catalytic Reactions in Supercritical Fluids

I

411

or scC3H6 as the solvent and carried out the reaction in a continuous fixed-bed reactor. Advantages of this are high reaction rates and selectivities. From the industrial point of view, the disadvantages are problems with catalyst deactivation and the small production scale. 13.3.2 Methylation

The eight naturally-occurring substances with vitamin E activity are derivatives of 6-chromanol. There are two groups of compounds with vitamin E activity. In the first group these are derivatives from tocol, which has a saturated c16 isoprenoid side chain, and in the second group these are derivatives of tocotrienol, which has a triple unsaturated C16-side chain (Scheme 13.7) [48].

R' R' R' R'

= R2 = R3 = CH,: a-tocopherol = R3 = CH,, R2 = H: P-tocopherol = H, R2 = R3 = CH,: y-tocopherol = R2 = H, R3 = CH,: &tocopherol

R' R' R' R' Scheme 13.7. Compounds with

= R2 = R3 = CH,: a-tocotrienol = R3 = CH,, R2 = H 0-tocotrienol = H, R2 = R3 = CH,: y-tocotrienol = R2 = H, R3 = CH,: 6-tocotrienol

I

R3

vitamin E activity.

The application of semi-synthetic (R,R,R)-a-tocopherol from natural sources is exclusively restricted to the pharmaceutical, food, and cosmetic industry. The most important natural sources of vitamin E are plant oil and fats. Various processes

412

I

73 Application ofSupercritical Fluids in the Fine Chemical lndustry

for isolation/purification of tocopherols have been described [49]. The amount of atocopherol (the compound with the highest vitamin E activity) in the mixture of tocopherols isolated from soybean oil is low [49]. So, there is a need to synthesize atocopherol from non-a-tocopherols.There are a few processes available for achieving this transformation, for example hydroxymethylation with formaldehyde and hydrogenation, chloromethylation with formaldehyde and HC1 and hydrogenation, formylation with zinc cyanide and HCI, and the Mannich aminomethylation and hydrogenation [SO]. Most of these processes have severe disadvantages in the production of waste, or handling of corrosive materials. The per-methylation of non-a-tocopherols to a-tocopherol under sc or near-sc conditions, i.e. pressures of 50 to 120 bar and temperatures of 513 to 623 K using methanol (or CO and H2)and a co-solvent such as COz or Hz0, has the advantages of high yield and no waste problems [Sl]. The reaction is catalyzed by hydrotalcite catalysts, for example Mg6A1z(OH)16C034Hz0(Scheme 13.8). This methylation can be carried out in a continuous process in sc-fluids (Fig. 13.9) 13.3.3

Friedel-Crafts Alkylation Reactions

Continuous Friedel-Crafts alkylation, with high selectivity, of mesitylene and anisole with propene or propan-2-01in S C C ~ or H SCCOZ ~ using a heterogeneous polysiloxane solid acid catalyst (Deloxan, ASP I/7) is described by Poliakoff et al. [52] (see Chapter 12). No comparison was made with the continuous alkylation in a conventional solvent and it is, therefore, difficult to judge the technical potential of this approach. The influence of various catalysts on the alkylation and ring-closure reaction of trimethylhydroquinone (TMHQ) and isophytol (IP) in scCO2 and scNzO to (allrac)-a-tocopherol, was investigated by Wang et al. [S3]. Tocopherol, vitamin E, an essential food ingredient, is of great economic importance. The demand for the main product on the market, (all-rac)-a-tocopherol,is constantly increasing [541. Industrial syntheses of (all-rac)-a-tocopherolare based on the reaction of TMHQ and IP, whereby Lewis and Bronsted acids serve as catalysts [49] (Scheme 13.9). The main problems with an industrial synthesis following this route are difficult purification under high vacuum and distillation, contamination of waste water with metal ions, and corrosion caused by the acidic media. Application of scsolvents in the synthesis and purification steps is particularly attractive [SS, 561. We studied and optimized the reaction by carrying out experiments under various temperatures, pressures, concentrations, and reaction times [531. The catalysts tested, either alone or in combinations, and the results obtained are reported in Table 13.3. The yield and conversion obtained in scCOz or scN20 is comparable to that achieved using the conventional process. From Table 13.3 it follows that the highest yield is obtained in scN20 and SCCOZ in the presence of HN(SOzCF3)z or AgN(SOzCF3)z as catalyst. The experiments show that there is no significant temperature or pressure effect on the yield at

73.3 Catalytic Reactions in Supercritical Fluids

\

I

413

414

I

73 Application ofSupercritica1 Fluids in the Fine Chemical Industry

- I T -v

feed

0 starting material balance

~

thermo couple

7-

pressure equalizing valve

heat exchanger Fig. 13.9.

Set-up for continuous methylation of non-a-tocopherol.

IP

TMHQ

TMHQ = trimethyl hydroquinone JP = isophytol = c16H33

Scheme 13.9.

Synthesis of (all-roc)-a-tocopherolfrom TMHQ and IP.

the end of the reaction. The yield of (all-rac)-a-tocopherolis influenced by the undesired decomposition of IP during the reaction. Continuous feeding of IP into the reactor during the reaction (semi-batch) should improve the process. The set-up for the semi-batch experiments is shown in Fig. 13.10. The results obtained for the catalysts HN(S02CF3)z or AgN(SOZCF3)2 in batch and semi-batch experiments are shown in Fig. 13.11. There is no difference with either catalyst.

13.3 Catalytic Reactions in Supercritical Fluids

I

Synthesis of (all-rac)-c-tocopherol under various experimental conditions [53].

Tab. 13.3.

Catalyst

Solvent

T (K)

p (bar)

Yield (%%)

Amberlyst 15 Nafion NR50 H3PWiz04o AgN(cF3S02)~ HNCF3SOz)z HN(CF3SOz)z/Si02

C02 C02 COz NzO NzO COz

110 110 110 150 150 150

160 160 160 155 160 160

60.7 12.1 36.5 89.0 89.6 80.5

All yields are based on IF‘. analysis GC ISTD method.

The amount of by-products did not change with time after the reactants were nearly totally converted, and it was assumed that the by-product formation from the products could be ignored (Fig. 13.12). From Fig. 13.12 it is clear that small amounts of IP decompose because TMHQ was not completely converted. A reaction mechanism is being proposed on the basis of kinetic data to facilitate reactor analysis and design. 13.3.4 Oxidation

Supercritical water oxidation (SCWO) technology has been studied as an altemative method for destroying hazardous waste, or converting it into “harmless” Waste gas 4

I

ET Autoclave Fig. 13.10.

Pump

Synthesis o f (all-rac)-a-tocopherol, set-up for semi-batch experiments.

I. IP/Balance

415

416

I

73 Application ofSupercritica1 Fluids in the Fine Chemical Industry

025

>

0.20 1

E"

T=150 'C

I

P=160 bar C=0.2632 [mol/L] S/Cat=250 TMHQ:IP=I :I in N,O

C

;.

0.15

~

c L

m C

8 g

010-

G1 r Q

+

u

I-"

Batch with AgN(CF,SO,),

4 Batch with HN(CF,SO,),

-n-

0 00 01

Semi-batch with AgN(CF,SO,),

I

I

02

03

I

0

Time [rnin]

Fig. 13.11.

Results obtained from batch and semi-batch experiments

products, for example COz. A detailed review on the scope and limitations of various equipment has been published [ 111. Oxidation reactions of organic compounds, with the goal of introducing functional groups, is still one of the biggest challenges in organic reactions. 0.25

1

0.20 1 =

?

Y

0.15

T=102 'C P=160 bar C=0.2632 S/Cat=250[rnol/L]

c

._ c

s c

g

TMHQ:IP=I:I in N,O

0.10

8

Fi v

TMHQ

0.05

0.00 0

10

20

30

Time [min]

Fig. 13.12.

Comparison of simulated (line) and experimental data (points).

40

73 3 Catalytic Reactions in Supercritical Fluids

I

417

Several catalytic oxidation reactions in sc-fluids are described in the literature. The RuO2-catalyzed (NaI04 as oxidant) synthesis of adipic acid from cyclohexene in the two-phase system C02- H20, has been reported [57]. Alkene oxidation using tert-butylhydroperoxide (TBHP) and various catalysts in scC02 resulted in the highly selective formation of epoxides with total conversion, depending on the catalyst used [ 581. Carrying out the asymmetric Sharpless epoxidation in scCOz using Ti( TBHP, and di(isopropy1)tartrate (DIPT), yielded the epoxide in a remarkable ee of 87 % (Scheme 13.10) [58].

Ti(O'Pr),/DIPT O

213 K

,

liqco, Scheme 13.10.

Asymmetric epoxidation using Ti(OiPr)d/DIPT.

Allylic alcohols, for example geraniol, 2-methylallyl alcohol, 3,3-dimethylallyl alcohol, 3-buten-2-01,1-octen-3-01,and I-hexen-3-01, are epoxidized with tert-butyl hydroperoxide in the presence of a vanadyl salen 0x0-transfer catalyst in supercritical COz. The metal catalyst was prepared in a simple two-step, Schiff base-type reaction to form the salen ligand, followed by complexation to the vanadyl group. The use of nontoxic supercritical CO2 in the presence of the new epoxidation vanadium catalyst led to yields and diastereoselectivities that were comparable to those resulting from the use of environmentally hazardous solvents such as CHzCl2 [59]. From an industrial point of view, oxidation using oxygen is very important because no additional waste or need for the recycling of organic material from the oxidant is necessary. Oxidation in scCO2 has the advantage that C02 is inert to oxidation by 0 2 . A problem resulting from oxidation in scCO2 is the low solubility of the catalyst, thus this technique has no industrial application. 13.3.5 Other Reaction Types

Hydroformylation involves the addition of H2 and CO to a C=C bond to yield aldehydes, and is one of the most important industrial processes [60]. The first example of the homogeneous hydroformylation in scCO2 was described by Rathke [61.]. The solubility of rhodium catalysts is better in scCO2 (also in conventional solvents) compared with the cobalt systems. Leitner and Koch studied the Rhcatalyzed hydroformylation of various alkenes in scCO2 and found a higher reaction rate compared to that in organic solvents [62]. A maximum turnover frequency (TOF) of 1345 h-' could be achieved (Scheme 13.11).

418

I

13 Application ofSupercritica1 Fluids in the Fine Chemical Industry

scco,

313 - 328 K

R,'

examples for: R' = R2 = C2H5;R 1= C6HI3,R2 = H R' =Ph, R2 = H R'= CH~OAC, R2 = H Scheme 13.11. Rh-catalyzed hydroformylation in SCCOZ.

A novel catalyst, RhH(CO)(P(p-CF,Ph),)3,was synthesized for the homogeneous catalytic hydroformylation of olefins in supercritical carbon dioxide. The incorporation of p(trifluoromethy1) groups in the conventional hydroformylation catalyst, HRhCO(PPh3)3,provided enhanced solubility in supercritical carbon dioxide while maintaining catalyst activity and selectivity in the hydroformylation of 1-octene. The reaction rate showed a first-order dependence on the catalyst concentration. The total system pressure had no effect on either the reaction rate or selectivity. However, selectivity was found to depend on the concentration of the catalyst [63]. Intermediates for potential Herpes simplex vims thymidine inhibitors, a new class of antiviral agents, are 5-amino-substituted 2-fluoroarabinofuranosyl 5-ethyl pyrimidine nucleosides. Wang et al. described the direct transformation of a mesylate or tosylate intermediate to the corresponding amine in high yields (96 %) using sc-NH3. This is an interesting alternative to the azide substitution/ hydrogenation sequence usually used to introduce an amino function. A reaction mechanism has been proposed, along with the kinetics required for reactor analysis and design [ 641 (Scheme 13.12).

0

8 - 266 bar

295 - 433 K

scNH3 bH

Scheme 13.12. Ammonolysis with sc-NH3

References 13.4

Concluding Remarks *

*

-

Examples of industrial applications of supercritical fluids can be found in pharmaceutical chemistry or in the synthesis of fine chemicals. The advantages of using supercritical fluids over classical procedures, such as higher yields or selectivities, should be emphasized. Of particular importance are processes which require food-approved solvents. The use of sc-fluids in synthetic chemistry should be considered in “new” reactions and in the design of reactions.

Acknowledgments

The authors gratefully acknowledge Professor G. Brunner, Drs M. Breuninger, L. Devetta, U. Fleck, M. Johannsen, F. Kienzle, T. Netscher, H. Pauling, M. Schneider and S. Wang for stimulating discussions and collaborations over many years. References 1

2

3

4 5

6 7

8 9

10

K. ZOSEL.(a) Angm. Chem. 1978, 90, 784; (b) Angew. Chem. Int. Ed. Engl. 1978, 17, 702. B. SUBRAMANIAN, M. A. MCHUGH, lnd. Eng. Chem. Process Des. Deu. 1986, 25, 1. H. TILSCHER, H. HOFMANN, Chem. Eng. Sci. 1987, 42, 959. B. M. TROST,Angm. Chem. 1995, 107, 285. R. A. SHELDON, CHEMTECH 1994, 38. R. A. SHELDON, J. Mol. Catal. A: Chemical 1996, 107, 75. R. A. SHELDON, Chem. Ind. (Lond) 1992, 903. R. A. SHELDON, J. Chem. Tech. Biotechnol. 1997, G8, 381. M. POLIAKOFF, M. W. GEORGE, S. M. HOWDIE, in Chemistry Under Extreme or Non-Classical Conditions, R. v. ELDICK,C. D. HUBBARD (Eds), I. Wiley & Sons and Spektrum Akademischer Verlag co-publication, New York, 1997. E. DINJUS,R. FORNIKA, M. SCHULZ, in Chemistry Under Extreme or NonClassical Conditions, R. v. ELDICK,C. D. HUBBARD (Eds), J. Wiley & Sons and

Spektrum Akademischer Verlag copublication, New York, 1997. 11 H. SCHMIEDER, N. DAHMEN, J. SCHON,G. WIEGAND, i n Chemistry Under Extreme or Non-Classical Conditions, R. v. ELDICK,C. D. HUBBARD (Eds), 1. Wiley & Sons and Spektrum Akademischer Verlag copublication, New York, 1997. 12 P. G. JESSOP, W. LEITNER(Eds), Chemical Synthesis Using Supercritical Fluids, Wiley-VCH; Weinheim, 1999. 13 P. G. JESSOP, Top. Catalysis 1998, 5, 95. 14 A. BAIKER, Chem. Rev. 1999, 99, 453. 15 P. G. JESSOP, T. IKARIYA, R. NOYORI, Chem. Rev. 1999, 99, 474. 16 J. A. DARR,M. POLIAKOF?, Chem. Rev. 1999, 99, 495. 17 A. WIDMER, (a) in Chemische Rundschau, 2000, No 13, July 1999; (b) @EN, 2000, April 3, 16. 18 T. L. CHESTER, J. D. PINKSTON, D. E. RAYNIE, Anal. Chem. 1998, 70, R301. 19 C. 1. CHANG, K. L. CHIU,Y.-L. CHEN, C.-Y. CHANG,Food. Chem. 2000, 68, 109. 20 J. JANKUM, S. H. SEIAM,R. SWIERZ, Nature 1997, 387, 561.

I

419

420

I

13 Application ofSupercritica1 Fluids in the Fine Chemical Industry 21

22

23

24 25

26

27 28 29 30

31

32

33

34 35

36 37

M. D. A. SALDANA, R. S. MOHAMED, M. G. BAER,P. MAZZAFERA,]. Agric. Food Chem. 1999, 47, 3804. M. JUNGFER, G. BRUNNER, 5” Conference on Supercritical Fluids and their Application, Garda (Italy), June 13-16,1999, K. OHGAXI,I. TSUKAHARA, K. SEMBA, T. KATAYAMA, Int. Chem. Eng. 1989, 29, 303. U. FLECK,EP 1043051 A2. U. FLECK, G. BRUNNER,R. KARGE, 5‘ International Symposium on Supercritical Fluids, Atlanta April 8-12, 2000. (a) J. L. ZEVENBERG, M. RUDRUM, Fat. Sci. Technol. 1993, 95, 456; (b) S. ENDRES,R. DE CATARINA, E. B. SCHMIDT,S. D. KRISTENSEN, Eur. J . Clin. Invest 1995, 25, 629; (c) S. WILLICH,K. WINTHER,Dtsch. Med. Wsch. 1995, 120, 227; (d) S. E. CARLSON, INFORM 1995, 6, 940; (e) M. HAMOSH,].Am. Coll. Nutr. 2994, 13, 546. U. FLECK,C. TIEGS,G. BRUNNER, J. Supercrit. Fluids 1998, 14, 67. V. RIHA, G. BRUNNER, J. Supercrit. Fluids 2000, 17, 55. E. KLESPER, A. H. CORWIN,D. A. TURNER,].Org. Chem. 1962, 27, 1962. M. YOSHIOKA, S. PARVEZ, T. MIYAZAKI, H. PARVEZ (Eds), Supercritical Fluid Chromatography and Micro-HPLC, VSP, Utrecht, The Netherlands, 1989. B. WENCLAWIAK (Ed.), Analysis with Supercritical Fluids: Extraction and Chromatography, Springer, Berlin, 1982. M. SAITO,Y. YAMAUCHI, T. OKUYAMA (Eds), Fractionation by Packed-Column SFC and SFE: Principles and Applications, VCH, New York, 1994. L. LABLER, in Znlmann’s Encyclopedia of Industrial Chemistry, Vol. A27, VCH, Weinheim, 1996, 469f. M. JOHANNSEN, EP 969001 A2. P. E. SAVAGE, S. GOPALN, T. I. MIZAN, C. J. MARTINO,E. E. BROCK,AICHEJ. 1995, 41, 1723. C. Y. TSANG,W. B. STREET,Chem. Eng. Sci. 1981, 36, 993. W. LEITNER, (a) Angav. Chem. 1995, 107, 2391; (b) Angav. Chem Int. Ed. Engl. 1995, 34, 2207.

P. G. JESSOP, Y. HSIAO,T. IKARIYA, R. NOYORI, J. Chem. Soc., Chem. Commun. 1995, 707. 39 P. G. JESSOP, Y. HSIAO,T. IKARIYA,R. NOYORI, I. Organomet. Chern. 1994, 475, 257. 40 J. XIAO,S. C. A. NEFKENS, P. G. JESSOP, T. IKARIYA, R. NOYORI, Tetrahedron Lett. 1996, 37, 2813. 41 R. M’HAMDI, J. F. BOCQUET, K. CHHOR,C. POMMIER,C, J . Supercrit. Fluids 1996, 4, 55. 42 G. DI GIACOMO, P. P. BOATTINI, Chim. Ind. (Milan) 1991, 73, 549. 43 A. BERTUCCO, P. CANU,L. DEVEITA, Ind. Eng. Chem. Res. 1997, 3G, 2626. 44 L. D E V E ~ PhD A , thesis, University Padua, Italy, 1998. 45 M. G. HITZLER, M. POLIAKOFF, Chem. Commun. 1997, 1667. 46 B. MINDER, T. MALLAT,K.-H. PICKEL, K. STEINER, A. BAIKER,Catal. Lett. 1995, 34, 1. 47 (a) M. HARROD,M.-B. MACHER, J. HOGBERG, P. MOLLER,in Proceedings of dth Italian Conference on Supercritical fluids and their Applications, Capri (Italy), 1997, 319; (b) M. HARROD,P. MOLLER, i n Proceedings ~ f 3 ‘ ~ International Symposium on High Pressure Chemical Engineering, Ziirich Switzerland, 1996, 43. 48 0. ISLER, G. BRUBACHER, in Vitamine I , Georg Thieme Verlag, Stuttgart, 1982, 126. 49 K.-U. BALDENIUS, L. V. d. BUSSCHEHBNNEFELD, E. HILGEMANN, P. HOPPE,R. STURMER,UElmann’s Enzyclopedia of Industrial Chemistry, Vol. A 27, VCH, Weinheim, 1996, 478f. 50 (a) S. M. WILLING, EP 178400; (b) L. WEISLER,US 2,640,058; (c) P. LECHTEN, U. HORCHER,B. JESSEL, EP 338429; (d) L. WEISLER,US 2,486,539; (e) J. GREEN,S. 2. MARCINKIEWICZ, US 2,992,235; ( f ) J. G. BAXTER,US 2,592,630; (g) N. S. BALDWIN, EP 159018. 51 M. BREUNINGER, EP 882722 Al. 52 M. G. HITZLER, F. R. SMAIL,S. K. Ross, M. POLIAKOFF, Chem. Commun. 1998, 359. 38

WANG,W. BONRATH, H. PAULING, F. KIENZLE,].Supercrit. Fluids 2000, 17, 135. 54 Chemical Marketing Reporter, November 10,1997, 5. 55 G. BRUNNER, T. MALCHOW. K. S ~ U E R K E N T. , G O ~ S C H A/. USupercrit. , Fluids 1991, 4, 72. 56 K. LOWACK,J. MEYER, M. EGGERSDORFER, P. GRAFEN, EP 603695 Al. 57 (a) D. A. MORGENSTERN, R. M. LELACHEUR, D. K. MORITA,S. L. BORKOWSKY,S . FENG,G. H. BROWN, 1.. LUAN,M. F. GROSS,M. J. BURK, W. TUMAS,in Green Chemistry, ACS Symp. T. C. Ser. 626, P. T. ANASTAS, (Eds), Am. Chem. SOC., WILLIAMSON Washington, DC 1996, 132ff. 58 D. R. PESIRI,D. K. MOBITA,W. GLAZE, W. TUMAS,Chern. Commun. 1998,1015. 53 S.

HAAS.G. R., KOLIS, 1. W., Tetrahedron Lett. 1998, 39, 5923-5926. 60 C . D. FROHLING, C. W. KOHLPAINTNER, in Applied 59

Homogeneous Catalysis with Organometallic Compounds, Vol. 1, B. (Eds), CORNILS, W. A. HERMANN VCH; Weinheim, 1996, 61 (a) 1. W. RATHKE, R. J . KLINGLER, T. R. KRAUSE,Organometalllics 1991, 10, 1350; (b) R. J . KLINGLER, J. W. RATHKE,].Am. Chem. SOC.1994, 116,4772. 62 D. KOCH, W. LEITNER,].Am. Chem. SOC.1998, 120, 13398. 63 D. R. PALO,C. ERKEY, Ind. Eng. Chem. Res. 1999, 38, 2163. 64 S. WANG,M. KARPF, F. KIENZLE, J. Supercrit. Fluids 1999, 15, 157.

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

14

Applications of Supercritical Water Eckhard Dinjus+c and Andrea Kruse 14.1

Introduction

Supercritical water is completely miscible with many organic compounds and gases. This opens new opportunities for chemical reactions and technical processes. Supercritical water (SCW) can be a suitable reaction medium for reactions usually carried out in organic solvents. Examples include reactions with organometallic complexes, which are unexpectedly stable in supercritical water. The differences compared to organic solvents are firstly, that the solubility exists only in the supercritical state, such that separation of non-polar products after the reaction is facile and cheap. Secondly, water does not pyrolyze or oxidize under the reaction conditions; therefore it is suitable as a reaction medium in e.g. single-phase oxidations of organic compounds. From the macroscopic point of view, SCW is a “non-polar” solvent; from a microscopic point of view, it is a molecule with a strong dipole of 1.85 Debye. Water in the supercritical state is able to react with different compounds. Therefore water is simultaneously solvent and reactant in a variety of reactions. Water at high temperature and high pressure is a variable reaction medium [ 13. Important properties like density, ionic product and relative dielectric constant, and the solubility of organic or inorganic compounds depend strongly on temperature and pressure. This means that adjusting temperature and pressure changes the selectivity of reactions in near- and supercritical water. In near-critical water, the ionic product is some orders of magnitudes higher than in ambient water, thus reactions normally needing the addition of acids or bases as catalysts show very high reaction rates without these additions. For SCW at rather low density, the ionic product is low and therefore different reactions occur. The most drastic change in properties is for instance, the solubility of inorganic compounds near the critical point. This chapter focuses on reactions that take advantage of the special properties or the tunability of properties of near- and supercritical water. These are on the one hand, synthesis reactions near the critical temperature of water, and on the other, decomposition reactions at higher temperatures. First of all, the properties of nearand supercritical water and their influence on chemical reactions will be discussed.

74.2 Physico-Chemical Properties of Water at High Temperature and Pressure

I

423

14.2 Physico-Chemical Properties of Water at High Temperature and Pressure and their Relation to Applications

Water is an ecologically safe substance widespread throughout nature. Below the critical point, the vapor pressure curve separates the liquid and gaseous phase. Here a continuous density variation from gas-like to liquid-like densities is not possible [ 2 ] . The vapor pressure ends at the critical point (T, = 373 "C, pc = 22.1 MPa, pc = 320 kg m-3). Beyond the critical point, the density of SCW can be varied continuously from liquid-like to gas-like values without phase transition, over a wide range of conditions. Liquid water under standard conditions (T = 25 "C, p = 0.1 MPa) is poorly miscible with hydrocarbons and gases. In contrast, it is a good solvent for salts because of its high relative dielectric constant of 78.5 at a high density of 997 kg m-3. At nearly critical temperature and pressure the relative dielectric constant is in the range of 10 [3]; this is nearly the relative dielectric constant of methylene chloride under standard conditions, and it further decreases with increasing temperature. Therefore, SCW at low densities becomes a poor solvent for ionic species like inorganic salts. On the other hand, SCW is completely miscible with many organic compounds and gases. As a consequence of the strong dependence of the solvent power on temperature and density of water, changes in temperature and pressure near the critical point can be used to precipitate particles of a special structure or size. The complete miscibility of supercritical water and gases, as well as with organic compounds, makes SCW an excellent solvent for homogeneous reactions of organic compounds with gases, like the oxidation of organic compounds with oxygen and air. The absence of phase boundaries leads to a fast and complete reaction. Pressure: 25 MPa

-.__ .-0

,

~

L c

6 0 :

,

,

7 0 :

Salts are soluble

100

0 s

2s

.a,

(I)

9: z a 0

Organic compounds and

50

--.__ ---__.

gases are --.__. ----.-..__.__

0 -5

soluble

c

h

0

-10

3

g -

u

-15

.O

-25

g

K

0 -

v

Fig. 14.1.

. ..

...Ambient water: -14

-------..._.____-_-

-20

0

100

200

300

400

500

600

700

T/"C Selected properties of water at high temperature and high pressure.

424

I

14 Applications of Supercritical Water

The properties of a solvent also influence the rates of chemical reactions. During a reaction, the transition state may be of higher or lower polarity than the initial state. A high relative dielectric constant lowers the activation energy of the reaction with a transition state of higher polarity than the initial state. By variation of the relative dielectric constant, achieved by adjusting the temperature and pressure, the reaction rates may be controlled. As a consequence, these reactions show a high activation volume. Supercritical water is not only a solvent; it is a solvent at high pressure. Some of the effects observed in reactions carried out in supercritical water are caused by high pressure. Examples are reactions of free radicals. The reaction rates of small free radicals are increased by the enhanced energy equilibration rate due to the high collision frequency [4].This is very important during the oxidation in supercritical water because all reactions involving OH radicals are accelerated (see below). On the other hand, reactions of high molecular mass free radicals, e.g. occurring during pyrolysis, are slowed down by a so-called cage effect caused by solvent molecules at high pressure [S]. In some cases these effects may be reinforced by the special properties of water. A similar effect may be the reason why organometallic complexes are able to exist and even act as catalysts at rather high temperatures in supercritical water. The ionic product of water increases slightly with temperature up to around lop1' in the range 200-300 "C. Above the critical temperature the ionic product decreases drastically with temperature, but increases with pressure. The ionic product in subcritical water and in supercritical water at high pressures can be some orders of magnitude higher than that obtained in ambient water. Under these conditions water may play the role of an acid or base catalyst because of the rather high concentration of H30f and OH- ions. Acid- or base-catalyzed reactions in water at high pressures and high temperatures show a characteristic non-Arrhenius kinetic behavior near the critical point of water [G, 81. Below the critical temperature of water, the reaction rates usually increase with temperature until the critical temperature is reached. At the critical point, the reaction rate decreases drastically. Diffusion rates are high and viscosity is low in a supercritical aqueous mixture. Transport properties and miscibility are important parameters, which influence the rate of chemical reactions. High diffusion rates and low viscosity, together with the complete miscibility with many substances, make supercritical water an excellent medium for homogeneous, fast, and efficient reactions. In addition, SCW is an excellent reaction medium with heterogeneous catalysts, because the high diffusion rate avoids mass transfer limitations and efficient solubility prevents coke formation on, or poisoning of the catalyst. Water is a molecule with a permanent dipole and is therefore reactive in a variety of reactions. For instance in hydrolysis reactions or in the water-gas shift reaction, water reacts with the organic compound or carbon monoxide. In some important reaction steps in the overall oxidation in SCW, it is assumed to participate in the activated complex [9, 101. By forming a complex, the activation energy is lowered,

14.3 Supercritical Water in Chemical Synthesis

I

425

which means that water is a catalyst for these reactions. Examples of applications that benefit from these unusual properties of water are given below.

14.3

Supercritical Water in Chemical Synthesis

In many organic reactions such as hydrolysis or certain rearrangements, water is the solvent and catalyst via self-dissociation, and sometimes also a reactant [11, 121. The advantage of the use of water is that the addition of acids and bases may be avoided. This means that cleaning the effluent is easier and less expensive. The ionic product of water increases with pressure (under supercritical conditions); therefore reaction rates e.g. of acid- or base-catalyzed reactions also increase. On the other hand, the reaction of free radicals, which are undesirable during pyrolysis, decreases with pressure (see Introduction), thus high selectivities can be achieved. In organometallic reactions, water is a thermally very stable solvent: but may also be a reactant. In this case water is a non-polar solvent from a macroscopic point of view, and a polar molecule from a microscopic point of view. This opens new opportunities for unusual reactions. A particular advantage of carrying out those reactions which are usually performed in organic solvents in supercritical water, is that solubility only exists at supercritical conditions. Following the reaction and cooling to ambient conditions, water and organic compounds separate. No distillation or other expensive separation techniques are necessary. Near the critical point, water is a very variable solvent with respect to the solubility of salts. Inorganic compounds are completely soluble below the critical point and precipitate at slightly higher temperatures. The variability of solvent properties opens the opportunity to generate crystals of defined size and morphology. 14.3.1

Organic Reactions

A variety of reactions in aqueous media can be accelerated by the addition of acids or bases. Here examples of reactions are given, which proceed at very high reaction

rates under conditions of high ionic product of water without addition of acids or bases. These reactions usually show the highest reaction rates in near-critical water, at the maximum of the ionic product. Undesirable side reaction such as C-C scissions occur at low densities to a certain extent. Decarboxylation is also reported as a side reaction for organic reactions in supercritical water. 14.3.1.1

Hydrolysis Reactions

During hydrolysis, water acts simultaneously as solvent, reactant and catalyst via self-dissociation. Often the addition of a further catalyst, usually acids or bases, is necessary to avoid undesirable side reactions. Hydrolysis reactions which have been investigated include the hydrolysis of amines [14-16],amides [17], nitriles

426

I

14 Applications ofSupercritical Water

1171, esters [7, 18-20], ethers [8, 21-24] and anhydrides [13]. The successful hydrolysis of aromatic amines like aniline [14] and benzylphenylamine [15, 161 requires the assistance of a further catalyst. The hydrolysis of aniline was carried out in a silver-lined tube reactor at pressures between 40 and 70 MPa, and at temperatures up to 450 "C. Phosphoric acid and its sodium salts were used as catalysts. The activation energy was lower in the subcritical than in the supercritical region. Under supercritical conditions the reaction rate increases with pressure. Both observations are in accordance with the assumption that the dissociation of the catalyst is the rate-determining step of the reaction. The solvent properties of water lead to an improved dissociation of acids and salts at subcritical conditions, and also at higher densities in the supercritical region. The hydrolysis of benzylphenylamine [ 15, 161 was investigated in a batch autoclave at 385 "C in the pressure range of 22 to 100 MPa, and was found to lead to the formation of aniline, benzyl alcohol and toluene as the main products. The selectivity towards hydrolysis is increased at higher pressure and in the presence of NaC1. The hydrolysis of nitriles to amides and further to alcohols has been investigated mainly at subcritical conditions in batch reactors (Scheme 14.1). I

I

I

Scheme 14.1.

I

Hydrolysis of nitriles and amides to alcohols.

The hydrolysis of acetamide, acetonitrile and benzonitrile was also investigated at higher temperatures (350-450 "C, 28-32 MPa) in a tubular reactor [17] without addition of catalysts. The measured activation energy of acetonitrile decreases with pressure, which was assumed to be a consequence of catalysis by H30+ ions. At higher pressure the self-dissociation of water increases, leading to an increased concentration of OH- and H30+ ions. The hydrolysis of esters is of technical interest; therefore many different esters such as acetates [ 181, phthalates [ 191, natural fats [ 201 and others were investigated. A detailed investigation of the hydrolysis of ethylacetate (tubular reactor, 23-30 MPa, 250-450 "C, 4-230 s) [7] without the addition of a catalyst shows a lower activation energy at subcritical conditions than at supercritical conditions, indicating two different reaction mechanisms. Under subcritical conditions nucleophilic attack on a protonated ester is assumed to be the rate-determining step of the hydrolysis process. The formation of a protonated ester is favored in the subcritical region because here the self-dissociation of water and the dissociation of the acid, formed via hydrolysis, increase. At 350 "C, 30 MPa, 170 s reaction time, and without additional acid, the conversion to acid and alcohol was 96 %, which is the equilibrium value. In other cases, mostly with unsaturated esters, the acids formed undergo decarboxylation, which leads to poorer yields [ 121.

14.3 Supercritical Water in Chemical Synthesis

Similar to the hydrolysis of esters, the hydrolysis o f ethers occurs at high pressures without the addition of acid catalysts. As for other hydrolysis reactions, high density and the addition of NaCl improves the reaction rate and selectivity of hydrolysis relative to other degradation reactions. Under optimal conditions, the reaction leads only to the respective alcohols. Examples of ethers investigated are methoxynaphthalenes [21], dibenzylether [7, 221,anisols [23], and from cellulose to glucose, fructose and oligorners [24]. 14.3.1.2 Condensations

The high degree of self-dissociation of water at high densities leads to catalysis of water elimination from alcohols and the formation of double bonds. In the case of tert-butanol [25], complete conversion to isobutene is achieved in 30 s at subcritical temperatures without addition of acids. In other cases, such as the elimination of water from ethanol [ 2 6 ] , propanol [27, 281, glycerol 1291, glycol [30], fructose [31, 321, glucose [31, 321, lactic acid [33], and cyclohexanol [34, 351, the addition of a mineral acid is necessary to prevent the breakage of C-C bonds and to obtain satisfactory yields. Also the formation of carbonyl groups ( pyruvaldehyde formation from glyceraldehyde and dihydroxyacetone [ 361) and ethers (formation of tetrahydrohran [37] and dibenzylether [38, 351) in water is observed in the near- and supercritical region of water without the addition of acids. Aldol condensations are typical base-catalyzed reactions. The reaction of nbutyraldehyde to 2-ethyl-3-hexanal at 275 "C occurs with 100 % conversion of butyraldehyde and 85 % selectivity [ 391 (Scheme 14.2).

Scheme 14.2.

Aldol condensation.

Other base-catalyzed reactions such as the Canizarro reaction of benzaldehyde could not be successfully carried out at these rather low temperatures [38, 391, but the reaction was completed in the supercritical region at 397 "C [40]without the addition of a base. 14.3.1.3 Diels-Alder Reactions

Diels-Alder reactions in water under ambient conditions have been investigated for many years. In contrast to ambient water, most of the imaginable dienophiles and dienes are completely miscible in SCW (Scheme 14.3). Diels-Alder-reactions possess a rather high intrinsic activation volume of around -25 to -50 cm3 mol-' [41]. Thus pressure increases the reaction rate, and SCW acts as solvent and pressure medium. The activation volumes for the different isomeric products, e.g. the endo and the exo product, differ slightly. The endo-

I

427

428

I

14 Applications of Supercriticd Water

I Scheme 14.3.

Diels-Alder reaction.

Diels-Alder reaction shows a more negative activation volume than the corresponding exo reaction (around -2.5 cm3 mol-I) [41]. Therefore, high pressures have an impact on regio- and stereoselectivity, e.g. by increasing the endo: exo ratio [ 42, 431. On the other hand, ambient water at low pressure also influences stereoselectivity. A large increase in stereoselectivity relative to organic solvents is found due to the so-called hydrophobic effect in ambient water [44, 451. The hydrophobic effect is the tendency of non-polar molecules and molecular segments in aqueous solution to avoid contact with water molecules [4G]. Around 15 different combinations of dienophiles and dienes were tested at nearand supercritical conditions [47]. In most cases the reaction rate at near critical conditions is faster than that under ambient, conventional reaction conditions. The reaction is highly sensitive to steric inhibitions. For non-steric inhibited reactants, like butadime and acrylnitrile an dienuphile the yields of the Diels-Alder products are between 49 and 100 %. During the reaction of cyclopentadiene with methylacrylate, higher temperatures in combination with longer reaction times lead to a shift in the endo: exo ratio from 1:1 to 3 :1. This is contradictory to frontier orbital calculations predicting a loss of endo:exo selectivity in supercritical water because of the loss of the hydrophobic effect [481. 14.3.1.4

Rearrangements

The pinacol/pinacolone rearrangement, which is a typical rearrangement catalyzed by acids, occurs in sub- and supercritical water in a certain temperature range without any addition of acids (Scheme 14.4). The reaction was carried out in the temperature range of 20 to 450 "C [49]. Below 300 "C no reaction was observed. Under most reaction conditions investigated, pinacolone was the only reaction product until conversion was complete. Between 375 and 380 "C and 22.5-25 MPa, 1,2,4trimethyl-4-isopropencyclohexeneinstead of pinacolone is the main product. This product is formed by the elimination of two molecules of water from pinacol and the subsequent reaction of the elimination product via a Diels-Alder reaction. In the same temperature and pressure range, the Beckmann rearrangement was also investigated (Scheme 14.5). This commercially important reaction to produce &-caprolactam,which is the precursor of perlon8 (nylon G), is therefore im-

14.3 Supercritical Water in Chemical Synthesis

I

429

YH3

CH,

300-450°C

;/c-f-CH3 CH3

20-35 MPa/

Scheme 14.4.

Pinacol/pinacolone rearrangement

portant for a variety of industrial polyamide products. In the industrial process, the reaction is carried out in the presence of sulfuric acid. In laboratory experiments at conditions close to the critical point of water, conversion to only E-caprolactam without the addition of an acid was observed [49].

Scheme 14.5.

Beckmann rearrangement.

Catalysts for rearrangement reactions were also investigated. For example, SnC12 acts as a Lewis acid to catalyze the formation of methylcyclopentene from cyclohexanol or cyclohexene (up to 62 % yield at 375 "C [34]). 14.3.1.5

Friedel-Crafts Reactions

The Friedel-Crafts alkylation of phenol and p-cresol with tert-butanol, isopropanol and n-propanol was investigated [SO, 511 (Scheme 14.6).At 275 "C, concentrations of H30+ and OH- reached a maximum, this is therefore the temperature applied for this acid-catalyzed reactions. The fastest alkylation of phenol was the reaction of tert-butanol with phenol. After a short reaction time, up to 17 % 2-tert-butylphenol was found. This yield decreases to the equilibrium value of around 10 % (yields relative to initial phenol content). The yield of 4-tertbutylphenol reaches 20 % and the residual phenol content was around 70 % after attaining equilibrium.

430

I

74 Applications of Supercrhical Water

I

OH

Scheme 14.6.

FriedelLCrafts alkylation.

Acylation was also carried out in a similar manner to alkylation without the addition of an acid catalyst, but with less favorable equilibrium yields [52] (Scheme 14.7). Phenol and resorcinol can be acetylated to the corresponding esters and ketones in aqueous acetic acid at high temperature (250-300 "C) to give substantial equilibrium conversions without any added acid catalysts. Usually strong mineral acids or Lewis acids are necessary to catalyze this process, and these are not easy to remove after the reaction. In aqueous acetic acid at 290 "C, phenol was primarily converted to 2-hydroxyacetophenone, 4-hydroxyacetophenone, and phenyl acetate in nearly equal amounts, with a combined equilibrium yield of less than 1 %. Under the same conditions, resorcinol was converted primarily to 2,4dihydroxyacetophenone with a modest equilibrium yield of 4 %. Due to the equilibrium limitation of these reactions, the yields are around 10-fold higher if the reaction is carried out in neat acetic acid. Water is a product of the reaction and therefore excess water decreases the yield [52].

I

0

Scheme 14.7.

0

OH

Friedel-Crafts acylation.

There is another application for electrophilic substitution in SCW. The use of supercritical D 2 0 presents the opportunity of producing deuterated organic compounds, e.g. complete deuterated aniline (78 % yield of completely deuterated product, 400 "C, 12 h reaction time [53]) in the presence of sodium deuteroxide. 14.3.1.6

Partial Oxidations

The oxidation of methane to methanol was studied in an attempt to find a new process for the production of methanol without synthesis of gas as an intermediate step. The idea was that SCW would stabilize methanol so that it can be isolated in higher quantities than those resulting from the usual oxidation processes. In the

74.3 Supercritical Water in Chemical Synthesis

I

431

latter case the formation of methanol is faster than the oxidation and no remarkable yields of methanol are obtained. Experiments show that the oxidation of methane in SCW without a catalyst leads to very low yields of methanol, although a small stabilization effect seems to occur [54]. Using Cr203/A1203and Mn02/Ce0 as heterogeneous catalysts leads to maximum selectivities of 1.7 % methanol, with a 6 % conversion of methane [55], and formic acid as the main products. This is a better yield than in the uncatalyzed reaction, but a very poor result in view of the technical application. In spite of these results, research on the partial oxidation of alkanes with heterogeneous catalysts is continuing [ 561. Much higher yields are found for the oxidation of alkylarenes to aldehydes, ketones and carbon acids by oxygen in the presence of transition metal compounds (MnBr2, CoBr2, CuBr) as catalysts [57]. For example, a yield of 30 % benzaldehyde and only 10 % benzoic acid was found for the oxidation of toluene. Also the dehydrogenation of cyclohexene and cyclohexane to benzene with PtOz as catalyst was investigated at 375 "C [34]. 14.3.1.7

Reduction Reactions

In subcritical water, the reduction of nitroarenes to the corresponding amines and quinolines with zinc was reported [58, 591. Also under subcritical conditions, alkynes and alkenes are reduced to alkanes by NaCO2H as the reducing agent and Pt on carbon support as the catalyst [60]. For example a yield of up to 78 % decane was found for the reduction of 1-decene. Without a catalyst, but with NaC02H, aldehydes were selectively reduced in the presence of ketones [61]. Aldehyde reduction was observed at 250-300 "C and around 8.27 MPa with reaction times of up to 3 h; at 300 "C or higher, cyclic ketones show substantial reduction, and acyclic ketones afford only poor conversion to alcohols at 340 "C even for extended reaction times. The yields for most alcohols formed from aldehydes are between 40 and 74 %. At 340 "C yields of 53 and 51 % were found for cyclohexanone and cyclopentanone, respectively. Under these conditions the yields of the acyclic ketones investigated were 10 % or less. Under supercritical conditions, hydrogen produced via the water-gas shift reaction is used as the reducing agent, for example in the hydrogenation of naphthalenes [62,63]. 14.3.2

Organornetallic Reactions

Organometallic complexes are usually assumed to be rather unstable, for instance with respect to high temperatures. It is an interesting fact that organometalliccatalyzed reactions are successfully carried out in SCW. The advantage of organometallic reactions with respect to their technical application can be the facile separation of products upon cooling. 14.3.2.1

Heck Coupling

The reaction of iodobenzene with alkenes [6466] was carried out in hot compressed water at 260 "C and also in SCW at 400 "C. The pre-catalyst used in most

432

I

14 Applications of Supercritical Water

cases was P ~ ( O A Ctogether )~ with a reduction agent (e.g. N(Et)3),to form the Pd(0) complex in situ, which is the catalyst for the Heck reaction (Scheme 14.8). Other pre-catalysts show no significant difference. The presence of a reducing agent is not necessary at supercritical conditions; perhaps the metal wall of the autoclave takes this role. NH4HC03was found to be most efficient in the reaction with the acid formed. It leads for instance, to a yield of 30 % of both coupling products following the reaction of iodobenzene and styrene. The reaction proceeds in similar way to the reaction in an organic solvent, but is more sensitive to the nature and steric structure of the arene relative to classical reaction conditions. For example, stylbene reacts with iodobenzene, whereas methylstilbene does not. The alkene necessary for coupling can be formed in situ by elimination of acids from halogenated alkanes.

dR+ Qf26-o Pd(0Ach

v

R: -OH, -02C-CH3, -Br

260"C, H 2 0 -RI R: -CI, -0 2C-CH3, -Br

Scheme 14.8.

14.3.2.2

+@Po+..

Heck coupling.

Cyclotrimerization of Alkynes

The cyclotrimerization of alkynes is also a typical reaction which proceeds via organometallic complexes and which is usually carried out in organic solvents (Scheme 14.9). At 400 "C this reaction was successfully carried out in SCW, using CpCo(q4-H2C=CH-CH=CH2) and CpCo(C0)2 [67, 681 as pre-catalysts. The yields and relative amounts of both benzene isomers formed are comparable to results obtained from catalysis in organic solvents [69-731. With phenyl as the remaining R, conversion rates of >95 % are achieved at 380 "C, 25 MPa with a yield of 24 % of the symmetric and 71 % of the second isomer. Analysis of the organic and organometallic products formed in the cyclotrimerization of acetylenes in SCW, leads to a mechanism in accordance with that known from catalysis in organic solvents. The formation of all possible isomers of the organometallic products has been observed and quantified by GC-MS. Two of the three isomers of the bis-substituted CpCo(cyclopentadienone) derivatives were crystallized and characterized by singlecrystal X-ray diffractometry.

74.3 Supercritical Water in Chemical Synthesis

I

433

Scheme 14.9.

14.3.2.3

Cyclotrimerization o f alkynes.

Hydroforrnylation

The hydroformylation of hexene and cyclohexene in the presence of Co-carbonyl complexes in SCW has been examined showing that hydroformylation under these conditions is possible, although no organometallic products could be isolated after the reaction [67, 741. The product distribution of the hydroformylation has been investigated by varying the amount and composition of the synthesis gas and the catalyst. Depending on the conditions, the following features were observed: [Co(CO)3PPhj]Z is less reactive than Co2(CO)xbut more selective towards the linear products, and no formation of ketones is observed. Isomerization always takes place and is much faster than hydroformylation. Hydroformylation of hexene is also observed in the absence of hydrogen, since under the selected reaction conditions, hydrogen is formed by the water-gas shift reaction. In these experiments ca. 10 % hydrogen was found in the gas phase after the reaction. In the absence of added hydrogen, less hydrogenation of the aldehydes is observed, leading to higher selectivities towards aldehydes and less formation of alcohols. In the presence of carbon monoxide and catalysts such as Co2(CO)x, [Co(C0)3PPh3I2 and H I ~ ( C O ) hydroformylation ~~, of alkenes was observed (Scheme 14.10). The hydrogen required for this reaction was produced from carbon monoxide via the water-gas shift reaction. This reaction seems to be catalyzed by the same catalytic species as the hydroformylation reaction. Up to 55 % hydroformylated products are formed from cyclohexene and 1-hexene. Hydrogenation is the main side reaction.

HRC -CH,

I

HRC=CH,

+

CO + H,

CO,(CO)

8

sew

*

+ H,RC-CH,-CHO

CHO HRC - CH, + H,RC-

I

CH,OH Scheme 14.10.

Hydroformylation o f alkenes.

CH,-

CH,OH

434

I

74 Applications of Supercritical Water

14.3.3

Inorganic Reactions

The excellent solubility of SiOz or other inorganic compounds in near-critical water leads to an homogeneous solution. A small increase in temperature decreases the solubility drastically; therefore crystals grow from the solutions with defined size and morphology. The crystal growth can be precisely controlled by temperature and pressure. Quartz crystals produced by hydrothermal growth methods (350-400 "C, 80-200 MPa, e.g. [75]) possess excellent properties and are widely used in the electronics industry. Supercritical water can be an excellent reaction environment for hydrothermal crystallization of e.g. metal oxide particles [76]. The production of defined particles of complex metal oxides, like barium hexaferrite (BaFel2019) used in high-density recording media, metal-doped oxide (A15(Y Tb)3012, YAG/Tb) used in phosphor screens, and lithium cobalt oxide (LiCoOz) for Li ion battery cathodes, may possess remarkable potential for technical applications [761. Metal oxides are also formed by hydrolysis of acetates [77], nitrates [78, 791 or mixtures of both [80] in SCW. For example the decomposition of La(CH3C00)3and Cu(CH3C00)2 at 400-500 "C in SCW leads to La2Cu04with a high surface area and enhanced oxygen mobility. La2Cu04 is an important, catalytically active perovskit, which is produced in SCW in a more facile way than by conventional preparation techniques [771. Rapid expansion to lower pressure leads to a drastic decrease in solubility. The so-called RESS process, well known for particle formation from solutions in supercritical carbon dioxide, is adapted to form fine particles or films by fast expansion of an aqueous solution [81]. Under more drastic reaction conditions (7.7 GPa, 1600 "C) even the formation of diamond was observed in supercritical water-COz mixtures [821.

+

14.4

Supercritical Water in Decomposition Reactions

The interest in supercritical water as a reaction medium firstly focuses on degradation reactions. The reason is obvious: the rather high critical temperature seems to favor reactions which lead to small, thermodynamically stable compounds at high temperature. 14.4.1

Oxidation in Supercritical Water

The complete miscibility of organic compounds and gases with supercritical water makes it possible to carry out an oxidation in an inert solvent. Since there are no phase boundaries, incomplete conversion resulting from incomplete mixing of the fluid, does not occur.

14.4 Supercritical Water in Decomposition Reactions

I

435

The oxidation of harmful organic compounds contained in aqueous waste effluents known as Supercritical Water Oxidation (SCWO) has been investigated since the 1980s (reviews: 83, 891). These studies are based on very fundamental work concerning the properties of supercritical water [ 21 and on pioneering experiments [ 90, 911 concerning conversions in supercritical water. At this time, subcritical Wet Air Oxidation (WAO, review: [ 921) at lower pressure had already been successfully used, but the new process promised significant advantages [83]: Higher oxidation efficiency at higher reaction temperature and density. Much lower residence time caused partly by homogeneous mixing of reactants, which avoids interfacial transport limitations leading to compact reactors. To meet the release standards in one process step for the gaseous and aqueous effluents without additional process steps as needed for WAO. From the engineering point of view, two major difficulties emerged under supercritical operating conditions [83]: Increased corrosion of reactor and heat exchanger construction materials. Drastically decreased solubility of salts resulting in precipitation within the reactor and feed pre-heater causing fouling and even blockage. During the SCWO process, the organic compounds react completely with the oxidant, mostly oxygen, forming COz and HzO. The hetero atoms chlorine, sulfur, or phosphor present in the organic wastes are transformed into the mineral acids HC1, HzS04, or H3P03, respectively. Organic bound nitrogen predominantly forms NZand small amounts of NzO. Undesired by-products known from incineration, such as dioxins and NO,, are usually not formed [89]. To achieve the desired conversion efficiency of about 98 % (for some hazardous wastes even higher), usually temperatures of 500-GOO “C at pressures between 25 and 35 MPa and a reactor residence time of up to 1 min are applied. This means that a commercial SCWO plant has to include a pre-heater for efficient recovery of heat from the reactor effluent. This is necessary for practical applications in order to be able to compete with traditional waste-treatment processes like incineration, landfill storage, etc. Essentially three reactor concepts were developed and studied [93-99] : tubular reactor (e.g. (93-95]), tank reactor with the reaction zone in the upper part and a cooling zone in the lower part of the tank to dissolve the salts (e.g. [9G]),and the “transpiring wall reactor” with an inner porous pipe which is rinsed with water to prevent salt deposits on the wall (e.g. 94, 97-99]. A fourth concept is the hydrothermal burner, which cools the wall by coaxial injection of large amounts of water [ 1001. As oxidants, mainly air, oxygen, and hydrogen peroxide were tested. Mostly Ni-based alloys were used as reactor construction materials. An unusual reactor concept should also be mentioned here. An alternative to the above is the deep-well reactor consisting of concentric tubes of 1.4 to 3.0 m length

436

I

74 Applications of Supercritical Water

inserted in a conventionally drilled and cased well hole in the earth [ l o l l . Such reactors were initially used for wet oxidation. The application of such reactors is not very promising because of blockage problems. For waste streams with low amounts of salts, a simple tube reactor is a useful tool since blockage is not likely. In order to find a suitable material for such SCWO reactors, a series of corrosion studies with different metals, alloys, and ceramics was carried out [ 102-1091. Investigations on the corrosion phenomena of nickelbased alloy 625 (UNSNOGG25),which is regarded as representative of other nickelbased alloys, and stainless steels in strongly oxidizing solutions of different acids (HF, HC1, HBr, HzS04,HNO3, H3P04), salts (NaC1, NaHS04) and bases (NaOH, Na2S04), showed an interesting temperature dependence. In all cases (except in the presence of NaOH and H3P04),the corrosion was strongest in the subcritical region. Above a certain temperature, corrosion effects disappeared, which is correlated with the change in water properties at the critical point. At higher temperature and lower density, the dmociation of acids and the solubility of salts is decreased, which leads to an increased stability of Ni(I1) oxide which forms a protective layer. On the other hand, in the case of NaOH and H3P04, corrosion was most drastic in supercritical water. This means for most reaction systems and in many SCWO plants that corrosion occurs mainly in the pre-heater and less in the reactor [ 1021. For the pre-heater, titanium liners have been shown to be suitable except in the presence of fluoride [ 108,891. In addition to the investigation of numerous model compounds, real wastes from chemical, pharmaceutical and food industry, from municipal sewage treatment plants, and from military and nuclear power facilities were tested in bench and pilot scale plants [110]. For a better understanding of supercritical water oxidation, single components like 2,4-dinitrotoluene, acetic acid, ammonia, aniline, cyanide, dichloromethane, ethanol, formic acid, hexachlorocyclohexane,hydrogen, phenol, PVC, DDT, pyridine, thiophene, toluene, trichloroethylene, and 1,1,1trichloroethane were studied. From these experiments, kinetic data were obtained. The destruction efficiency, which is the ratio between the residual total organic carbon content (TOC) and the initial TOC achieved for these compounds is up to 99.999 % [83]. Also flames in supercritical water, e.g. by oxidation of methane with oxygen, have been studied [ 111,1121. Due to the formation of dioxin during common gas phase oxidation, special attention has been paid to the oxidation of halogenated aromatic compounds like 2,4dichlorobenzene, 2,4,G-trichlorophenol,tetrabromobisphenol A, 3-chlorobiphenyl, 4-chlorophenol, PCB, and pentachlorophenol. For all these compounds a destruction efficiency equal or higher than 99 % was obtained [83]. For dioxins the destruction efficiency was 99.99999 %. Real waste materials investigated include radioactive sludge from the nuclear industry, brewery effluents, electronic scrap, sewage sludge, municipal sludge, navy hazardous wastes, paper mill effluents, percolate, chemical and pharmaceutical industry waste, polymers, and rocky flats. Here destruction efficiencies of higher than 99 % were achieved [83]. As a consequence of the high potential of SCWO, research and development activities all over the world have been initiated (see references [83, 851). The first and up until now,

74.4 Supercritical Water in Decomposition Reactions

the only commercial plant for civil waste treatment has been in operation in Texas since 1994 by the Huntsman Corporation [95]. Further plants for the treatment of different wastes are planned or are presently operated on pilot plant scale (see references [83, 851). To describe the oxidation of simple compounhs in supercritical water, different groups have developed kinetic models. These models consist of elementary reactions [ 113-1151 or combined chemical reactions [ 1161 in the mathematical form of ordinary differential equations. From a chemical point of view these models are very similar to small differences in the values of the activation energy and the preexponential factor. These modeling studies are supported by spectroscopic measurements on intermediates [117]. The basis for the models consisting of elementary reactions [113-1151, are well-investigated gas phase models [118, 1191. The gas phase models are transformed to high-pressure conditions by increasing the reaction rate of the elementary reactions as a consequence of the increased energy transfer at higher pressure [4]. No specific solvent effect of water on free radical reactions is considered. Water is regarded mainly as a pressure medium and not as a solvent with a solvent effect on the chemical reaction. From experiments on the decomposition of H202 [9, 101, it is assumed that water is able to take part in the activated complex of free radical reactions. Measurements of the fugacity coefficient [ 1201 show that water is a solvent with a specific solvent effect. Consideration of these specific influences of water during SCWO may lead to improved elementary reaction models [121]. The very good solubility of organic compounds in supercritical water opens up the possibility of extracting hazardous compounds e.g. from soil. Such an extraction was combined with a SCWO process, in which the oxygen is produced electrochemically [ 122, 1231. Catalysts are now being employed to enhance SCWO operations. In comparison to the catalytic wet air oxidation, the low viscosity and the high diffusion coefficients prevent mass transfer limitations [124, 1251. Another advantage is that the high solubility of organic compounds prevents coke formation, because coke precursors, if they are formed on the surface, are quickly removed [124, 1251. For the catalytic supercritical water oxidation (CSCWO),heterogeneous oxidation catalysts made of transition metal oxides or noble metals dispersed on metal oxide supports were investigated (review: [ 1251). For model compounds like acetic acids, ammonia and phenol, higher yields at lower temperatures relative to SCWO were achieved [125, 1261. The opportunity to achieve complete conversion at lower temperatures makes this process economically attractive [ 1271. The challenge is to find a catalyst, which is stable and active at supercritical water conditions and also in the presence of waste water [ 1251. 14.4.2

Gasification in Supercritical Water

Energy from biomass may provide a significant contribution towards satisfying the growing future energy demand. Energy from biomass avoids the net increase of

I

437

438

1

14 Applications of5upercriticaf Water

carbon dioxide in the atmosphere and would help to fulfill the obligations of the European Union to reduce carbon dioxide emissions. Waste from e.g. agriculture and the food industry is referred to as biomass, which can be gasified to produce fuel gases. These gases can be used subsequently to produce electricity. A large part of the biomass is wet biomass containing up to 95 % water. Such wastes may have high negative costs (disposal costs) which are beneficial for the gasification process. For a water content of more than 40 %, the thermal efficiency of a traditional steam (reforming) gasification plant decreases drastically. It is only about 10 % at a water content of 80 % 11281. A very promising alternative to wet biomass is the less investigated hydrothermal gasification (expected thermal efficiency of 70 % at a water content of 90 % [ 1281). In principle there are three different ways to use biomass for energy production: 1. Liquefaction: formation of liquid fuels near the critical temperature (300400 "C) of water. 2. Gasification to methane: at 200-400 "C in the presence of Ni and alkali salts, methane is produced from wastes. 3. Gasification to hydrogen: at 600-700 "C hydrogen is the main product of biomass conversion in supercritical water. The presence of alkali salts often improves the hydrogen yield.

One benefit of near- and supercritical water applies in all these processes for the conversion of biomass. The good solubility of organic compounds, which could be the precursor of tar, and the high reactivity of biomass in near- and supercritical water, decrease the formation of char and coke, and increase the yields of the desired products. In the temperature range between 250 and 300 "C, hydrolysis reactions are very fast because of the high ionic product of water. Cellulose, which is the major component of biomass, is fractionated very rapidly to glucose and fructose [24, 1291. Glucose and fructose eliminate water to form unsaturated compounds such as furfural derivatives [31]. Without a catalyst the gas yield is rather low. The liquid formed possesses a lower relative content of oxygen because of the elimination of water; it is therefore of higher fuel quality (higher gross calorific value) than biomass. The fuel can be burned or gasified. Liquefaction experiments of model compounds and biomass were carried out under sub- and supercritical conditions. In some cases a catalyst was added to improve the yield of the liquid products 190, 130-1321. One modification of this process is the so-called HTU process carried out at 330-350 "C and 12-18 MPa [133, 1341. Thermodynamic calculations of biomass gasification show [1351 that at low temperatures the exothermic formation of methane and at high temperatures the endothermic formation of hydrogen, is preferred. The calculated yield of carbon monoxide is very low for biomass with a high water content. Experiments show that both the formation of methane as well as the formation of hydrogen is kinetically inhibited, and that the carbon monoxide content may be much higher than expected [ 1421. If methane is the desired product, the temperature has to be in the range of

14.4 Supercritica/ Water in Decomposition Reactions

I

439

200-400 "C. In this temperature range the gasification rate is rather low, but it can be improved by the addition of alkali salts [ 136,137,1421. In addition, the presence

of a hydrogenation catalyst such as Ni is necessary for the formation of methane. For example the conversion of cellulose was studied in autoclaves with residence times of up to 1 h at 200-400 "C and 8-18 MPa [138, 1391. It was found that sodium carbonate as a catalyst suppresses the formation of char and oil and mainly water-soluble products were formed. At 400 "C and with a Ni catalyst, CH4 and COz were found as major products in the gas phase. Batch reactor experiments [140: 1411 used for the formation of a CH4-rich gas from biomass, waste model compounds and real waste waters, were also carried out at 350 "C, 20 MPa, and reaction times of 60-120 min. It was shown that aromatic and aliphatic hydrocarbons as well as oxygenates are converted to a CH4-rich fuel gas in the presence of hydrogenation catalysts. The results were confirmed in continuous-flow reactor experiments with residence times of 10 min and longer for conversions of 90 % or more. In any case, without the high reactivity of biomass in and with near-critical water, methane formation at low temperatures would not be possible. The presence of alkali salts also improves the hydrogen yield at high temperatures by accelerating the water-gas shift reaction [140, 141, 144, 1451. Investigation of the gasification of pyrocatechol as a model compound for lignin, shows an increase in the hydrogen content by a factor of 3 corresponding to a dramatic decrease in CO formation at 500 "C by adding up to 5 wt % KOH (5 wt. % pyrocatechol, 250 bar, 1-h reaction time [1421). Experiments in tubular reactors show that at temperatures of 600 "C, complete gasification of the model compounds (glucose, pyrocatchol) used can be achieved 1142, 1351. Aromatic rings included in biomass as lignin are rather stable under supercritical water conditions [5, 1431. Therefore, their degradation behavior during gasification is of particular interest. Usually real biomass already contains alkali salts, such that the effect of salt addition is less effective [ 135, 144, 1451. A similar effect of improving the gasification efficiency and hydrogen yield by alkali salts was found for the gasification of coal in SCW [146]. An alternative for improving gas and hydrogen yields from biomass is the use of coke as catalyst. The coke-catalyzed gasification of model substances, biomass (aquatic plants, etc.) and wastes at GOO-650 "C, 280 or 34.5 MPa and reaction times of about 30 s, was investigated in a tubular flow reactor [147, 148, 149J. At 600 "C and 34.5 MPa, glucose and other feeds at concentrations of up to about 0.2 mole L-' were gasified completely to a hydrogen-rich gas. Coke or tar formation was not observed. Military wastes were also completely gasified under these conditions. This process is possible because biomass is much more reactive in SCW than coke. In comparison to the traditional gasification processes for hydrothermal gasification, the following advantages for a wet biomass/organic waste feedstock can be expected: much higher thermal efficiency,

- a hydrogen-rich gas with low CO content can be produced in a single process - step, soot and tar formation can be suppressed, and

440

I

14 Applications of Supercritical Water

heteroatomes (S, N, and halogens) leave the process with the aqueous effluent avoiding an expensive gas cleaning process. Further experiments need to be carried out in order to optimize the process parameters (pressure, additives), especially in view of higher feed concentrations (> 10 wt % organic), which are necessary to achieve a thermal efficiency high enough to establish an economic process. For the process development, the next indispensable step is the construction of a pilot plant of a representative scale to optimize the technical components and to demonstrate the interconnected operation'). The following technical hurdles have to be overcome: to find a reliable high pressure feeding system for slurries,

- fouling problems in the heat exchanger, pre-heater and reactor caused by salty precipitates have to be avoided, and to find a construction material resistant to hydrogen embrittlement under SCW conditions. In some applications degradation to lower molecular weight compounds rather than gasification is desired. For example the extraction of oil-contaminated soil, which is not biologically degradable, leads to a mixture that can be treated in a conventional sewage treatment process [ 150, 1511. Coal and oil can be upgraded by treatment with SCW, which leads to higher quality oils [152] and coal [153]. Degradation or depolymerization of polymers to monomers and other low molecular weight compounds is regarded as a suitable process for recycling (e.g. [154, 1551).

14.5

Conclusions

Supercritical water is characterized by special, adjustable properties making it a reaction medium for a variety of chemical reactions. One particular property to be mentioned is the higher solubility of organic substances in SCW; although the solubility is similar to that in an organic solvent, the thermal stability is high. SCW can therefore be used as a substitute for organic solvents at high temperatures. Examples include the cyclotrimerization of alkynes, hydroformylation and the Heck reaction. These reactions proceed via organometallic intermediates, which show an unexpected stability in supercritical water. The reason for this stability might be a pressure or solvent effect caused by the SCW environment. A detailed investigation of the organometallic species shows that in principle the same reaction mechanism occurs as in organic solvents. SCW possesses macroscopic properties similar to an organic solvent, but of course a single molecule remains 1)

Such a pilot plant has now been built at the Forschungszentrum Karlsruhe (Karlsmhe Research Center.)

14.6 Outlook 14.41

unchanged and can react. During hydroformylation in SCW only the addition of carbon monoxide and not of hydrogen is necessary; hydrogen is formed via the water-gas shift reaction. After cooling down, the products of these reactions separate themselves from the aqueous phase. The increased ion product of water at higher densities, in many cases obviates the need to add acids and bases as catalysts in technically important reactions, such as hydrolyses, rearrangements, aldol condensations, Friedel-Crafts alkylation and acylation, as well as condensation reactions. This could help to save such additives and to avoid the associated waste management problems. The excellent solubility of many inorganic compounds in near critical water, and the decrease in solubility with temperature, opens up the possibility of generating specific crystals from solutions. The morphology and size of these crystals can be controlled by varying the solubility, which can be optimized by adjusting the temperature and pressure. Because organic compounds and gases are completely miscible with SCW, hazardous compounds can be completely oxidized with oxygen to carbon dioxide in a single phase. This process has been demonstrated successfully for many single compounds and wastes. Challenges like corrosion and salt deposition are overcome e.g. by special reactor designs. Another process in which the properties of SCW offer special benefits, is the gasification of wet biomass, i.e. waste arising e.g. in the food industry and in agriculture, but also of sewage sludge and coal. Advantages over classical gasification processes are the increased reactivity, i.e. the lower process temperature, the generation of hydrogen and carbon dioxide instead of synthesis gas, the low expense in off-gas cleaning, and the avoidance of costly drying processes.

14.6 Outlook

Studies on organic and organometallic synthesis reactions in SCW have just begun. The results are very promising, but there is still more work to be done in order to understand chemical reactions and their dependence on the properties of this special reaction medium. The task will be to optimize selectivities and reaction rates by adjusting the reaction parameters. The decomposition reactions can be applied in industry in a relative short period of time. The total oxidation of wastes is very well investigated and difficulties such as corrosion and salt deposition seem to be solved e.g. by advanced reactor designs. Now it is up to industry to apply this knowledge if the costs of this process can compete with e.g. incineration. Biomass gasification and the recycling processes mentioned above have not been studied in as much detail as SCWO, but the laboratory scale experiments are very promising. From an engineering point of view, there are still many questions which need to be answered before industrial application is possible. Therefore, the way forward is to rectify any problems at the pilot plant stage before scaling up the procedures for industrial application.

442

I

74 Applications of Supercritical Water References

C. A. ECKERT,D. BURK,J. S. BROWN, C. L. L I O ~ AInd. , Eng. Chem. Res. 2000, 39,4615-4621. 2 E. U. FRANCK, H. WEINGARTNER, IUPAC - Chemical Thermodynamics, T. M. LETCHER,(Ed.), Blackwell Science Ltd., Oxford, 1999, 105. 3 C. A. MEYER,R. B. MCCLINTOCK, G. J. SILVESTRI, R. C. SPENCER JR., Steam Tables - Thermodynamic and Transport Properties of Steam, 6th edn, ASME, New York, 1992. 4 H. HIPPLER, in Proceedings of the 4th ItaEian Conference on Supercritical Fluids and their Applications, E. REVERCHON (Ed.), CUES, Salemo, Italy, 1997, 279-286. 5 H. J. EDERER, A. KRUSE,C. MAS,K. H. EBERT,J.Supercrit. Fluids 1999, 15,191. 6 M. T. KLEIN,L. A. TORRY, B. C. Wu, S. H. TOWNSEND, S. C. PASPEK, J. Supercrit. Fluids 1990, 3, 222. 7 P. KRAMMER, H. VOGEL,]. Supercrit. Fluids 2000, 16, 189. 8 M. T. KLEIN,B. Izzo, C. L. HARRELL, T. ADSCHIRI,i n Proceedings of the 4th International Symposium on Supercritical Fluids, K. ARAI(Ed.), Tohoku University Press, Sendai, Japan, 1997, 543. 9 N. AKIYA,P. E. SAVAGE,].Phys. Chem. A 2000, 104,4433-4440. 10 N. AKIYA,P. E. SAVAGE,].Phys. Chem. A 2000, 104,4441-4448. 11 P. E. SAVAGE, Chem. Rev. 1999, 99, 603-321. 12 J. AN, I. BAGNELL, T. CABLEWSKI, C. R. STRAUSS, R. W. TRAINOR,]. Org. Chem. 1997, 62, 2505. 13 W. C. BELL,K. S. BOOKSH, M. L. MYRICK,Anal. Chem. 1998, 70, 332339. 14 F. PATAT,Monatshefte Chemie 1945, 77, 352-375. 15 S. H. TOWNSEND, M. A. ABRAHAM, G. L. HUPPERT,M. T. KLEIN,S. C. PASPEK,Ind. Eng. Chem. Res. 1988, 27, 143-149. 16 L. A. TORRY,R. KAMINSKY, M. T. KLEIN,M. R. KLOTZ;]. Supercrit. Fluids 1992, 5, 163-168. 1

17 18

19 20 21

22

23

24

25

26

27

28 29

30

31

32

33

A. KRLMER,S. MITTELSTADT, H. VOGEL, Chem. lng. Tech. 1999, 71, 261-267. Y. TANABE, J. TORIYA,I. KAAHAHRA, 0. KURASHIKI, Ger. Pat. Appl. DE 2645030,1977. J. M. PENNINGER, Fuel 1988, 67,490496. P. MBLLER, PCT Intern. Pat. Appl. WO 97/07187,1994. J. M. L. PENNINGER, J. M. M. KOLMSCHATE: Supercritical Fluid Science and Technology, K. P. JOHNSON, J. M. L. PENNINGER, (Eds), ACS Symp. Series 406 1989, 242-258. L. A. TORRY,R. KAMINSKY, M. T. KLEIN,M. R. KLOTZ;J. Supercrit. Fluids 1992, 5, 163-168. M. T. KLEIN,Y. G. MENTHA,L. A. TORRY,Ind. Eng. Chem. Res. 1992, 32, 182-187. M. SASAKI,B. KABYEMELA, R. MALALUAN, S. HIROSE,N. TAKEDA, T. ADSCHIRI,K. ARAI,]. Supercrit. Fluids 1998, 13, 261-268. X. Xu, M. J. ANTALJR., D. G. M. ANDERSON, Ind. Eng. Chem. Res. 1997, 36,23-41. X. Xu, C. DEALMEIDA, M. J. ANTAL JR.,]. Supercrit. Fluids 1990, 3, 228232. M. J. ANTALJR., D. G. M. ANDERSON, X. Xu, Ind. Eng. Chem. Res. 1998, 37, 3820-3829. R. NARAYAN, M. J. ANTALJR.,]. Am. Chem. SOC.1990, 122, 1927-1931. M. J. ANTALJR., W. S. L. MOK,J. C. ROY,A. T. RArssr, J. Anal. Appl. Pyrolysis 1985, 8, 291-303. S. RAMAYYA, A. B R I ~ A I NC., DEALMEIDA, W. Mox, Fuel 1978, 66, 1364-1371. M. J. ANTAL,JR., W. S. MOK,i n Research in 2kermochemistry and Biomass Conversion, A. V. BRIDGEWATER (Ed.), J. L. Publisher, Elsevier, London, UK, 1988, 464-472. M. J. ANTAL,JR., W. S. MOK,G. N. RICHARDS, Carbohyd. Res. 1990, 199, 91-109. C. T. LIRA,P. J. MCCRACKIN, Ind. Eng. Chem. Res. 1993, 32, 2608-2613.

R. C. CRITTENDON, E. J. PARSONS, Organometallics 1994, 13, 2587-2591. 35 B. KUHLMANN, E. M. ARNETT,M. J. SISKIN,Org. Chem. 1994,59, 30983101. 36 B. M. KABYEMEIA, T. ADSCHIRI,R. MAIALUAN, K. ARAI,Ind. Eng. Chem. Res. 1997,36, 2025-2030. 37 P. KRAMMER, S. MITTELSTADT, H.VOGEL, Chem. Eng. Technol. 1999,22, 126-130. 38 A. R. KATRITZKY, M. BALASUBKAMANIAN, M. SISKIN.Energy @ Fuels 1990,4,499-505. 39 R. G L ~ S E R J.,S. BROWN,S. A. NOLEN, C. L. LIOTTA,C. A. ECKERT,Preprints at Symposia-ACS Division of Fuel Chemistry, 44(2), 1999,385-388. 40 Y. IKUSHIMA, K. KATAKEDA, 0. %TO, T. YOKOYAMA, M. ARAI,Angav. Chem. 2001, 113,216-219. 41 F.-G. KLLRNER,M. K. DIEDRICH,A. E. WIGGER,Chemistry under Extreme or Non-classical Conditions, R. VAN ELDIK, C. D. HUBBARD (Eds), Wiley, 1997, 103-162. 42 J. JURCZAK, D. T. GRYKO:Chemistry under Extreme or Non-classical Conditions R. VAN ELDIK,C. D. HUBBARD (Eds), Wiley, 1997,163-188. 43 E. DINJUS,R. FORNIKA,M. SCHOU: Chemistry under Extreme or Nonclassical Conditions R. VAN ELDIK,C. D. HUBBARD (Eds), Wiley, 1997,219-272. 44 R. BRESLOW, U. MAITRA,D. RIDEOUT, Tetrahedron Lett. 1983,24, 1901. 45 D. RIDEOUT,R. BRESLOW, J . Am. Chem. Soc. 1980, 102, 7816-7817. 46 A. KUMAR,Chem. Rev. 2001,101,l-19. 47 M. B. KORZENSKI, J. W. KOLIS, Tetrahedron Lett. 1997,38, 5611-5614. 48 Y. HARANO,H. SATO,F. HIRATA, Chem. Physics 2000, 258, 151-161. 49 Y. IKSUHIMA, K. HATAKEDA. 0. SATO, T. YOKOYAMA, M. ARAI, Angew. Chem. Int. Ed. 1999,38, 2910-2914. 50 K. CHANDLER, F. DENG,A. K. DILLOW, C. L. LIOTTA,C. A. ECKERT,Ind. Eng. Chem. Res. 1997,36, 5175-5179. 51 A. R. KATRIZKY, S . M. ALLIN,M. SISKIN,Acc. Chem. Res. 1996,29, 399406. 52 J. S. BROWN,R. GLLsER, C. L. LIOTTA, C. A. ECKERT,Chem. Comm. 2000, 1295-1296. 34

53

54 55 56 57

T. J U N KW. , J. CATALLO, L. D. CIVILS, J. Labelled Comp. Radiophamceut. 1997,39, 625-630. P. E. SAVAGE, R. LI, J . T. SANTINI,JR.. J . Supercrit. Fluids 1994, 7, 135-144. S. N. V. K. AKI, M. A. ABRAHAM, J. Supercrit. Fluids 1994,7, 259-263. D. BROLL,A. K R ~ M E R H ,. VOGEL, Chem. Ing. Tech. 2000, 72, 377-382. R. L. HOLLIDAY, B. Y. M. JONG,J. W. KOIIS, /. Supercrit. Fluids 1998, 12,

255-260. C. BOIX, M. POLIAKOFF, i n Proceedings of the 6th Meeting Supercritical FluidsChemistry and Materials, 10-13 April, Nottingham, M, POLIAKOFF(ed.) INPL, Vandoeuvre, France 1999,247-249. 59 C. BOIX,J. MARTINEZ DE LA FUENTE, M. POLIAKOFF, N a v J . Chem 1999,23, 641-643. 60 J. M. JENNINGS, T. A. BRYSON,I. M. GIBSON,Green Chem. 2000, 2, 87-88. 61 T. A. BRYSON, J. M. JENNINGS,J. M. GIBSON,Tetrahedron Lett. 2000,41, 3523-3526. 62 T. ADSCHIRI,R. SHIBATA,T. SATO,M. WATANABE, K. ARAI, Ind. Eng. Chem. Res. 1998,37,2634-2638. 63 T. ADSCHIRI,S. OKAZAKI,M. MOCHIDUKI,S. KUROSAWA, K. ARAI; rnt. J. Soc. Mat. Eng. Resour. 1999,7, 273-281. 64 E. J. PARSONS, Chemtech. 1996,30-34. 65 P. READON,S. METTS,C. CRITTENDON, P. DAUGHERITY, E. J. PARSONS, Organometallics 1995, 14, 3810-3816. 66 J. DIMINNIE,S. MEITS, E. J. PARSONS, Organometallics 1995, 14, 40234025. 67 H. BORWIECK, Wissenschajliche Berichte, FZKA-6302, (Mai) 1999, Doctoral thesis, University Heidelberg, 1999. 68 H . BORWIECK, 0. WALTER,E. DINJUS, J. RIBIZANT;J . Organometallic Chem. 1998,570, 121-127. 69 M. D. RAUSCH,R. A. GENETTI,J . Org. Chem. 1970,35, 3888. 70 H. BOENNEMANN, W. BRIJOUX,R. BRINKMANN, W. MEURERS,R. MYNOTT,J . Organomet. Chem. 1984, 272, 231. 71 H . BOENNEMANN, W. BRIJOUX,NavJ. Chem. 1987, 11, 549. 58

444

I

14 Applications of Supercritical Water

H. BOENNEMANN, R. BRINKMANN, H. SCHENKLUHN, Synthesis 1974, 575. 73 H. BOENNEMANN, Angew. Chem. lnt. Ed. Engl. 1985, 24, 248. 74 E. DINJUS,W. RIFFEL,H. BORWIECK, Ger. Patent DE-PS 19853371, 2000. 75 G. R. J O H N S O N , in Proceedings of the Joint 6th l S H R @ 4th ICSTR, Kochi, Japan, 2000,40. 76 T. ADSCHIRI,Y. HAKUTA, K. ARAI, Ind. Eng. Chem. Res. 2000, 39,4901-4907. 77 A. A. GALKIN, B. G. KOSTYUK, V. V. LUNIN,M. POLIAKOFF, Angew. Chern. 2000, 112,2850-2852. 78 T. ADSCHIRI, K. QNAZAWA, K. ARAI,]. Am. Chern. Soc. 1992, 75, 1019. 79 Y. HAKUTA, S. ONAI, S. TERAYAMA, T. ADSCHIRI,K. ARAI,I.Mater. Sci. Lett. 1998, 17, 1211. 80 A. CABANAS, J. A. DARR,E. LESTER,M. POLIAKOFF, Chem. Commun. 2000, 901-902. 81 J. I. BRAND, A. 0. SEZER:i n Proceedings ofthe lnternational Meeting of the GVC-Fachauschu. “Hochdmckvevfahrenstechnik” E. DINJUS,N. DAHMEN (eds), Wissenschaftliche Berichte FZKA 6271, Karlsruhe, 1999, 97-99. 82 M. AKAISHI,M. D. S. KUMAR, H. KANDA,S. YAMAOKA, Diamond Rel. Mat. 2000, 9, 1945-1950. 83 H. SCHMIEDER, J. ABELN,Chem. Eng. Techn. 1999, 11, 903-908. 84 J. W. TESTER, J. A. CLINE,Comosion 1999, 55, 1088. 85 R. W. SHAW,N. DAHMEN:Supercritical Fluids - Fundamentals and Applications, E. KIRAN,P. G. DEBENEDETTI, C. J. PETERS(Eds), Kluwer Academic Publisher, Dortrecht, 2000,425437. 86 E. F. GLOYNA, L. LI, Waste Manag. 1993, 13, 379-394. 87 E. F. GLOYNA, L. LI: Encyclopedia of Chemical Processing and Design, Vol. 65, JOHN J. M E K E ~ (Ed.), A , Marcel Dekker, Inc. Publisher, 1998. 88 E. F. GLOYNA, L. LI, Environ. Prog. 1995, 14, 182-192. 89 P. KRITZER, E. DINJUS,Chem. Eng. J . 2001, 83, 207-214. 90 M. MODELL, R. C. REID,S. I. AMIN, US. Patent 4 1978, 113, 446. 72

M. MODELL,U.S. Patent 4, 1982, 338, 199. 92 S. T. KOLACZKOWSKI, P. PLUCINSKI, F. J. RIVAS,D. B. F. J. BLETRAN, MCLURGH, Chem. Eng. J . 73,1999, 143-160. 93 K. C. SWALLOW, W. R. KILLILEA,K. C. Waste MALINOWSKI, C. N. STASZAK, Manag. 1989, 9, 19-26. 94 J. ABELN, M. KLUTH, G. PETRICH,H. SCHMIEDER, in Proceedings ofthe Joint Gth lSHR@4th ICSTR, K. YANUGISAWA, Q. FENG(eds.) Nishimira Tosha-Do Ltd., Kochi, Japan, 2000. 95 Eco waste Technologies, Supercritical Water oxidation (SCWO) News Letter, Eco waste Technologies Company, Austin, Texas, Vol. 3, No. 1,1998. 96 H. E. BARNER, C. Y. HUANG,T. JOHNSON, G. JACOBS, M. A. MARTCH, W. R. KILLILEA, J. Hazardous Mat. 1992, 31,l-17. 97 T. C. MCGUINNESS, US Patent 5,384,051, 1995. 98 H. H. MUEGGENBURG, D. C. ROUSSAR, M. F. YOUNG, US Patent 5,387,398, 1995. 99 P. J. CROOKER, K. S. AHLUWALIA, 2. FAN,J. PRINCE,Ind. Eng. Chem. Res. 2000, 39,4865-4870. 100 H. L. LAROCHE,M. WEBER, C. TREPP,Chem. Eng. Techn. 1997, 20, 208. 101 D. KODRA, V. BALKOTAIAH, AlChE Journal 1992, 38,988-1002. 102 M. SCHACHT,N. BOUKIS,E. DINJUS,]. Mat. Sci. 2000, 35, 6251-6258. 103 P. KRITZER, N. BOUKIS,E. DINJUS, Corrosion 2000, 56, 1093-1104. 104 Y. S. KIM, D. B. MITTON,R. M. LATANISION, Korean]. Chem. Eng. 2000, 1 7 , 5 8 4 6 . 105 K. W. DOWNEY, R. H. SNOW, D. A. HAZLEBECK, A. J. ROBERTS, in Innovations in Supercritical Fluids, K. W. NUTCHENSON, N. R. FOSTER (Eds.) ACS Symp. Ser. 608, American Chemical Society (ACS), Washington, DC, 1993, 313. 106 R. M. LATANISION, R. W. SHAW,MIT The Energy Laboratory, Workshop Summary, Massachusetts Institute of Technology (MIT), Boston, MA, 1994, 14. 91

References A. RUCK, J. NOVOTNY:J . KONYS, J. HAUSSELT, Proceedings of EUROCORR 98. European Federation of Corrosion, Frankfurt/Main, Germany, 1998, on CD. 108 N. BOUKIS, G. FRANZ,W. HABICHT, E. DINJUS,Corrosion 2001, Paper 01353. NACE Int., Houston, Texas, USA, 107

2001. 109

M. SCHACHT, N. BOUKIS,E. DINJUS, K. EBERT,R. J A N S S E N , F. MESCHKE, N. CLAUSSEN, J. Eur. Cer. Soc. 1998, 18, 2373-2376.

110

E. F. GLOYNA, L. LI, Waste Manag. 1992, 13, 379-94.

111

W. SCHILLING, E. U. FRANCK, Ber. Bunsenges. Phys. Chem. 1988, 92, 631-

112

G. M. POHSNER, E. U. FRANCK, Ber. Bunsenges. Phys. Chem., 1994, 98,

113

P. E. SAVAGE, R. LI, J. T. SANTINI,]. Supercrit. Fluids 1994, 7, 135-144. H. R. HOLGATE, j. W. TESTER, ]. Phys. Chem. 1994, 98, 810-822. P. DAGAUT,B. DANEY D E MARCILIAC, Y. TAN,M. CKTHONNET, J. c . J . Chim. Phys. 1995, 92, BOEITNER,

636.

1082-1090.

114 115

1124-1141. 116 117 118

119 120

L. T. BOOCK, M. T. KLEIN, Ind. Eng. Chem. Res. 1993, 32, 2464-2473. S. R. RICE,E. CROISET,Znd. Eng. Chem. Res. 2001, 40,86-93. R. D. WILK,N. P. CERNANSKY, W. J. Prlz, C. K. WESTBROOK, Combust. Flame 1989, 77, 145. T. S. NORTON,F. L. DRYER, Combust. Sci. Technol. 1989, 63, 107. T. 1. MIZAN,P.E. SAVAGE, R. M. ZIFF,AIChEJournal 1997, 43, 12871299.

121

122

123 124

N. AKIYA,P. E. SAVAGE, in Proceedings ofthe 5th International Symposium of Supercritical Fluids, D. BIJSCH,(Ed.), 2000, CD. A. FIRUS,G. BRUNNER, in High Pressure Chemical Engineering, R. VON ROHR,C. TREPP,(Eds), Elsevier Science, 1996, 179-184. B. MISCH,A. FIRUS,G. BRUNNER, J. Supercrit. Fluids 2000, 17, 227-237. G. MANO,in Natural Microporous Materials in Enuironmental Technology, P. MISAFLIDES, (Ed.), Nato Science

Serics, Series E: Applied Science, 362; Kluwer Acad. Publ., 1999, 283-302. 125 2 . Y. DING,M. A. FRISCH, L. LI, E. F. Znd. Eng. Chem. Res. 1996. GLOYNA. 35, 3257-3279. 126 1. Yu, P.E. SAVAGE, Enuiron. Sci. Technol. 2000. 34, 3191-3198. 127 S. N. V. K. AKI,M. A. ABRAHAM, Enuiron. Prog. 1998, 17, 146-255. 128 J. M. L. PENNINGER, personal communication. 129 S. SAKA,T. UENO,i n Proceedings ofthe 10th European Conference and Technology Exhibition Biornassfor Energy and Industry, H. KOPETZ. T. WEBER, W. PALZ,P. CHARTIER, J. L. FIRRIRO (Eds), CARMER, Strarihing, Germany 1998, 1815-1818. 130 D. BECKMAN, D. G. BOOCOCK, Can. J . Chem. Eng. 1983, GI, 80-86. 131 G. DELRE,G. GIACOMO, V. BRANDINI,in Proceedings of the 4th Italian Conference on Supercritical Fluids and Their Applications, E. RIVERCIION (Eds) CUES, Salemo, Italy 1997, 35. 132 M. MODELL, in Fundamentals of Thermochemical Biomass Conversion, A. V. BRIDGWATER, (Ed.), Elsevier, Amsterdam, 1985, 95. 133 F. GOUDRIAAN, D. G. R. PEFEROEN, Chem. Eng. Sci. 45,1990, 2729-2734. 134 F. GOUDRIAAN: B. VAN DE BELD,F. R. NABER, S. VAND E R WAL,J. A. ZEEVALKINK, in Progress in Thennochemical Biomass Cotwersion, A. V. BRIDGEWATER (Ed.), Blackwell Science, Oxford, 2001, Vol. 2, 1312-1325. 135 A. KRUSE,J. ABELN, E. DINTUS, M. M. SCHACHT, E. KLurH, G. PETRICH, SADRI,H. SCHMIEDER, in Proceedings of AIRAPT-17: International Conference on High Pressure Science and W. J. Technology, M. H. MUNGHNANI, MELLIS,M. F. NICOL(Eds.) Universities Press (India), Limited, Hydriguida, India 1999,954-957. 136 M. SISKIN,A. R. KATRITZKY, Science 1991, 254, 231. 137 L. SEALOCK, D. C. ELLIOIT,E. G. BAKER, R. S. BUTNER,Ind. Eng. Chem. Res. 32, 1993, 1535. 138 T. MINOWA, F. ZHEN,T. OGI,J . Supercrit. Fluids 1998, 13, 253-260.

I445

446

I

14 Applications of Supercritical Water 139

140

141

142

143 144

T. MINOWA,T. OGI, Y. DOTE, S. YOKOYAMA, Renewable Energy 1994, 5, 813-815. D. C. ELLIOTT,L. J. SEALOCK, JR., E. G. C. BAKER,Ind. Eng. Chem. Res. 1994, 33, 558-565. D. C. ELLIOTT,M. R. PHELPS,L. J. SEALOCK,JR., E. G. C. BAKER, Ind. Eng. Chem. Res. 1994, 33, 566574. A. KRUSE,D. MEIER,P. RIMBRECHT, M. SCHACHT,Ind. Eng. Chem. Res. 2000, 39,4842-4848. A. KRUSE,K. H. EBERT,Ber. Bunsenges. Phps. Chern. 1996, 100, 80-83. A. KRUSE,J. ABELN,N. BOUKIS,E. DINJUS, M. KLUTH, G. PETRICH,E. SADRI, M. SCHACHT,H. SCHMIEDER, in Proceedings of the International

Meeting ofthe GVC-FachauschuJ3 “Hochdruckuerfhrenstechnik”, E. DINJN’S, N. DAHMIN (Eds.), Wissenschaftliche Berichte FZKA 6271, Karlsruhe, Germany 1999,111145

146

147 148

149

150 151

152 153

154

114. H. SCHMIEDER,J. ABELN,N. BOUKIS,

E. DINJUS,A. KRUSE,M. KLUTH, G. PETRICH,E. SADRI, M. SCHACHT,J . Supercrit. Fluids 2000, 17, 145.

155

M. NYWLT,S. PETER,Erdol und Kohle Erdgas - Petrochemie uereinigt mit Brennsto$-Chernie, 1992, 45, 349. D. Yu, M. AIHARA, M. J. ANTAL,JR., Energy @ Fuels 1993, 7, 574-577. X . Xu, Y. MATSUMURA,J. STEINBERG, M. J. ANTAL, JR., Ind. Eng. Chern. Res. 1996, 35, 2522-2530. M. J. ANTAL,JR., S. G. ALLEN,D. SCHULMAN, X . Xu; Ind. Eng. Chem. Res. 2000, 39,4040-4053. K. NOWAK,G. BRUNNER;Chem. Ing. Tech. 1992, 64, 188-189. B. S. KOCHER,F. 0. AZZAM,T. J. CUTRIGHT,S. LEE, Energy Sources 1995, 17,213-222. J. R. KERSHAW,J.Supercrit. Fluids 1989, 2, 35-45. T. J. HOUSER,Y. ZHOU, C.-C. TSAO,X. LIU, Supercritical Fluid Engineering Science, ACS Symp. Ser. 1993, 514, 327-337. K. KODAMA,K. MURAYAMA, T. KUMAKI,F. KASUYA, Y. NAGASE,M. NAKATA,S. NISHIDA, Eur. Patent EP 1006143,1999. M. WATANABE,H. HIRAKOSO,S. SAWAMOTO, T. ADSCHIRI,K. ARAI; J. Supercrit. Fluids 2000, 13, 247-252.

High Pressure Chemistry Synthetic, Mechanistic, and Supercritical Applications

Edited by Rudi van Eldik and Frank-Gerrit Kliirner

0WILEY-VCH Verlag GmbH, 2002

I

447

Index a ab initio calculations 132, 138, 157

3-amino sugars 254 apparent activation volume 188f acetophenone 229 aquacobalamin 14 acid reactions 424 aquation 13 acroleine 341 arene-arene interactions (rr-rr, CH-rr, cation-rr) acrylate 348 - resin bound 285 ascorbic acid (vitamine C) acrylonitrile 311, 325, 327, 335, 341, 343 - oxidation of with quinones 86 activation energy 131, 138, 143 association 348 activation enthalpy 144 association constant K, 355 activation of dioxygen 27 associative 12, 97, 134 activation of small molecules 25 associative mechanism 20, 203 activation parameters 152 asymmetric hydroformylation 391f activation volumes 4, 139, 153, 162, 179, 383 asymmetric hydrogenation 387,410 - estimation 42ff asymmetric induction 277 - of organic reactions 240 aza Diels-Alder reactions 270 - of pericyclic cycloadditions 52 aziridines 265 - of various elementary processes 45 azodyes 357 - of competing [4 21 and ;2 + 2; or (4 41 cydodimerizations 49 b acylation 430 base hydrolysis 13 addition reactions 30 base quenching 211 adiabatic 164, 168 base-catalyzed reactions 427 -electron transfer 171, 173 basic principles 3 - outersphere 163 batch reactors 378 adsorption 403 Baylis-Hillman reaction 278, 305, 324, 338, Agmon-Hopfield reaction coordinate (AH) 343ff 99 Beckmann rearrangement 428 AH see Agmon-Hopfield reaction coordinate benzodicyclobutene 58, 243 aldol condensations 427 benzoin condensation 324 alkene oxidation 417 p-benzoquinone 339, 356 1-alkoxycyclohexenes 289 N-benzoylpyrrole alkyl vinyl ethers 254 p-benzylcalix[5larene 354 alkylation reactions 379f benzylamine 265 allenylchloromethylsulfone 62 benzylazide 65 dlylzincation 279 benzylphenylnitrone 65 p-aminoester 337, 342 benzyl vinyl ether 314 p-aminonitriles 342 biactivation 306, 313, 317, 341 p-aminosterase 315 binding of CO 29

+

+

448

I

lndex binuclear complexes 191 biomass 439 biomass gasification 438 biphasic systems 377 biradical intermediates 48, 50, 67, 253 biradicals 253 bismuth salts - catalysis by 319 boron trifluoride etherate 227 bromobenzene 230 Brownian motion 98 Buckminster fullerene C60 61, 351, 354 bulk solvent 12, 141 bullvalene - Cope rearrangement 72 1,3htadiene - dimerization 48,67 - reaction with a-acetoxyacrylonitrile 48 - reaction with ethene 52ff tert-butylamine 325, 338 tert-butylammoniumion 353 C

C-C bond formation 41 C-C coupling reactions 381 C60 see Buckminster fullerene calixarenes 353 (*)-cantharidin 240, 310 carbocyaninecations - Z/E isomerization 114 carbonyl complexes 203 carceplex 352 CAS-SCF 138 catalysis - heterogenous 375 -homogenous 377f catalyst 376 - heterogenous 371 -homogenous 371, 386 catalyst efficiency 388 catalytic hydrogenation 408 catalytic reactions 371, 373, 405 catecholborane (CBH) 236 [3] catenanes 89f CBH see catecholborane change in mechanism 154 changeover in mechanism 17, 20 charge-transfer(CT) band 28 cheletropic reactions 70, 83 chemical activation 306 chemoselectivity 393 5,6-CHIRAGEN 363 chiral catalysts 275 chiral solvents 277

p-chloranil 85 chloroprene - dimerization 48 trans-cinnamaldehyde 262 cinnamic esters 235 Claisen rearrangements 71, 324 - volumes of activation 73 Claisen-Ireland rearrangement 62 classical MD simulations 146 13C-NMR spectroscopy - pressure dependent 354 combinatorial chemistry 284ff compactin 268 complexation-inducedshift 355 complex-formation 13 compressibility 206, 378 compressibilitycoefficient 366 computer simulations 146, 156 concerted mechanisms 155 condensations 427 - microwave assisted 301 conformational changes 216 Connolly’smolecular surface 366 Connolly’smolecular volumes 366 continuous flow reactors 379, 383, 385 continuous reactor 376 Cope rearrangements 71 - volumes of activation 73 corrosion phenomena 436 Coulombic contributions 176 coumarin 248 counter-ion catalysis 166 coupling reactions - palladium-catalyzed 234 critical point 373, 425 cross coupling 231 cross reactions 162, 170 cross relation of Marcus 169 crotonaldehydes - reactions of 314, 321, 341 (E)-3-crotonyl-l,3-oxazolidin-2-one - Diels-Alder reaction 275 crown ethers 348, 353 [2.2.2]cryptand 353 crystal field splitting 144f crystal growth 434 cyanoacetylene 58 - trimerization 58 2-cyanobicyclo[2.2.21octa-2,s-diene 58 cyclic enol ethers 298 cyclic staircase voltammetry 170 cyclic voltammetry 170 cyclic voltammogram 171 cyclization

Index I 4 4 9

endo-trig 87 87 cyclization reactions 380 cycloaddition 30, 45 - Lewis acid catalysis 224 - multicomponent domino 286 [2 + 21 cycloaddition 66 [3 21 cycloaddition - anti-selective 299 - exo, anti-selective 290, 298 [4 21 cydoaddition - endo-selective 290 [4 2]/[3 21 cydoaddition - one-pot three-component 287 - on solid phase 299 [6 41 cycloaddition 68 (8 21 cycloaddition 68 cyclodextrines 336 a-cyclodextrin 357, 361 a-cyclodextrin cavity 357 /I-cyclodextrin 337 cycloheptatriene 246, 311 1,3,5-cycloheptatriene 61 1,3-cyclohexadiene 335 - dimerization 253 1,4-cyclohexadiene - dehydrogenation 85 cyclohex-2-en-1-one 294 cyclohexene - packing coefficient 72 1,5-cydooctadiene 81 cyclooctatetraene - antiaromaticity 57 cyclopentadiene 225, 268, 274 cyclopentenone 265 [2 21 cycloreversion 68 cyclotrimerization 432

decaffeination 373f, 398 decarboxylation 425 (E)-1,3,9-deca~ene 76 decomposition reactions 434 degree of bond-making 134 dehydration 380 denaturant 349 density 373 density functional theory 144 density measurements 82 (E)-1-deuteriochloroprene - dimerization 48 (Z)-/I-deuteriostyrene 67 DFTcalculations 132, 138 dialltoxy aluminium chloride 227 diamond anvil cells 186f, 193 diastereoselectivity - pressure-induced increase 262 1,4-diazabicydo[2.2.2loctane (DABCO) 343 1,4-diazabicydo[2.2.2]octanes, 2,3-substituted (DABCO) 278 diazomethane 65 2,3-dichloro-5,6-dicyano-l,4-quinone (DDQ) 85 dicyanoacetylene (DCA) 56, 58, 243 p-dicyanobenzene 356 1,2-dicyano-l,3-cydobutadiene59f dicyclohexylcarbodiimide(DCC) 241 2,4-dicyclohexyl-2-methylpentane (DCMP) 102 Z,Z-1,4-dideuterio-1,3-butadiene - dimerization 50 1,l-diethoxyethene 69 dielectric constant 162, 168, 210, 373, 422f - pressure dependence 322 dielectric continuum model 326 Diels-Alder reaction 298, 404, 427f d - activation volumes 45 DABCO see 1,4-diazabicylo[2.2.2]octan - aqueous 324, 327 Danishefsky’s diene 319 - asymmetric 269 DBBA see N-[4-(dimethylamino)benzylidene]- intramolecular 76, 227, 320 4-bromoaniline - reaction volumes 45 DBEA see N-[4-(dimethylamino)benzylidene]-- repetitive 62,64 4-ethoxycarbonylaniline - salt effect in 328 DBNA see N-[C(dimethylamino) - volume data 46ff benzylidene]-4-nitroaniline -with inverse electron-demand 287 DCA see dicyanoacetylene diethylaluminium chloride 269 DCC see dicyclohexyl carbodiimide diethyl azodicarboxylate(DEAD) 79 DCMP see 2,4-dicyclohexylmethylpentane 3,4-diethyl-3,4-diphenylhexane DDQ see 2,3-dichloro-5,6-dicyano-1,4- dissociation 85 benzoquinone diethyl ketomalonate 311 DEAD see diethylazodicarboxylate 3,3 ’-diethyloxacarbocyanineiodide (DOCI) 114 Debye-Hiidtel contributions 176 -

- exo-trig

+

+ +

+ +

+

+

3,3’-diethyloxacarbocyanineiodide (DOCI) (cont.) - isoviscous activation energies for the Z/E isomerization 119 - pressure effects on the Z/E isomerization 115 - viscosity dependence of the Z/E isomerization 120 3,3’-diethyloxadicarbocyanineiodide (DODCI) 114 - isobaric activation energies for the Z/E isomerization 119 - isoviscous activation energies for the Z/E isomerization 119 - pressure effects on the Z/E isomerization 116 - viscosity dependence of the Z/E isomerization 120 difference between the volumes of activation 243ff difference in the volume of reaction 243ff differential pulse voltammetry (DPV) 170 diffusion 373 - coefficients 171 -limit 208, 214 -rates 424 1,l-difluoroallene 67 difuranocydooctane 62 9,lO-dihydroanthracene 86 dihydrobarrelene 54 dihydrofuran 257 2,3-dihydrofuran 233, 236, 279, 289 1,2-dihydronaphthalene 86 1,4-dihydronaphthalene 86 9,lO-dihydrophenanthrene 86 dihydropyrans - preparation of 255f 3,4-dihydropyran 289 dihydropyridines 270 dihydrothiopyrans - phosphone-substituted 260 dihydroxyvitamine D3 314 diketopiperazines 286 di-bmenthyl acetoxymethylenemalonate 269 dimethylacetylene dicarboxylate 62, 246, 311 N-[4-(dimethylamino)benzylidene] -4bromoaniline (DBBA) - Z/E isomerization 101 - viscosity dependence of the Z/E isomerization 108 N-[4-(dimethylamino)benzylidene]-4ethoxycarbonylaniline(DBEA) - Z/E isornerization 101

N-[4-( dimethylamino)benzylidene]-4nitroaniline (DBNA) - Z/E isomerization 101 - Arrhenius plots for the Z/E isomerization 108 - viscosity dependence of the Z/E isomerization 108 - Z/E isomerization of in DCMP 105, 107 - Z/E isomerization of in GTA 104, 106 - Z/E isomerization of in MPD 106 4-( dimethylamino)benzylidene]-4’nitroazobenzene (DNAB) - Z/E isomerization 101, 108, 110, 122 - Z/E isomerization in silicone oil 123 4’-(dimethyIamino)-2-methoxy-4nitroazobenzene (DMNAB) - Z/E isomerization 101, 108 2,6-dimethylbenzoquinone 310 2,3-dimethylbutadiene 83, 286 2,3-dimethyl-1,3hutadiene 70 2,6-dimethyl-l-cyclohexenyl methyl ether 289 dimethylfulvene 54 2,s-dimethylfuran 339 (tris)[2,2-dimethyl-G,G,7,7,8,8,8-heptafluoro-3,5octanedionato]) (fod) 313 diolmycin A2 229 dioxygen complex 28 1,3-diphenylallylradicals 74 1,4-diphenylbuta-l,3-diene (DTB) - E/Z isomerization 99 diphenyldiazornethane 83 diphenylketene 66

l,l-diphenyl-2-methylhyI-2-(4-methoxyphenyl)oxirane 277 1,3-dipolarcydoadditions - activation and reaction volumes 64 diradical see biradical (+)-discondermolide 242 dispersion forces 348 disproportionation 85 dissociative 12, 134, 197 - activation 142 - substitution mechanism 216 distortions 167 c~s-1,2-divinylcyc~obutane 81 trans-1,2-divinylcyclobutane 67, 81 DMNAB see 4’-(dimethylamino)-2-methoxy4-nitroazobenzene DNAB see 4-(dimethylamino)-4’-nitroazobenzene DOCI see 3,3’-diethyloxacarbocyanineiodide DODCI see 3,3’-diethyloxadicarbocyanine iodide domino process 234

domino reactions 284 domino-Knoevenagel-hetero Diels-Alder reaction 261 donor strength 17 D P R see 1.4-diphen yl- 1 3-butadiene drug chemistry 319 drug discovery 284 dynamic solvent effects 97 dynamic viscosities 373 ~

e eight-coordinate 132, 145. 151 El'yanov equation 45, 80, 270 electrical double layer 173 electrochemical determination 22 electrochemical methods 10 electrochemical transfer coefficient 171 electrochemistry 161 electrocyclic rearrangements 74 electrode reactions 162 electron beam 33 electron transfer - heterogenous 171 - intramolecular 25 - non-symmetrical 22 - outer-sphere 161 - quenching 204 - reactions 2 1f - single (SET) 85 electronic coupling 164 electrostriction 25. 44,85, 135, 206, 309, 315, 322, 354, 363 elimination reactions 30 emission 186, 203 emission intensities and lifetimes 193 emission lifetimes 212 emission spectra 188, 194 enamine carbaldehyde 259 8-enaminoesters 315 enaminoketones 254,259 - hetero-Diels-Alder reactions 257 enantioselective hydrogenation 386, 388 enantioselective reduction 387 encounter complex 2 10 endo-selective 14 + 21 cycloaddition 298 ene reactions 77, 247 - volume data 80 - ) 6 + 4 j 253 energy transfer reactions 207 enolethers 66, 228 entropy of association 354, 356 entropy of formation 83 enviromental factor 399 epoxides

- aminolysis 319 - ring opening 229 ethene sulfonates - intramolecular Diels-Alder reaction 273 etherification 380 ethyl acrylate 230 ethyl cinnamate 230 ethyl vinyl ether 255. 260. 288, 301, 339 ethylene glycol 323 Eu(f0d)l 225 - catalysis by 313 exchange reactions 143 exciplex 212 excited states 184, 186 excited state deactivation 206 excited state tuning 193 Exner increments 82 expansion volume 44,81 experimental methods 6 external pressure 310 extraction 400f. 403 extraction of caffeine 374f extraction with COz 391

f

s-facial selectivity 274 FeCl, 277 fifty-percent rule 173, 178 fine chemical industry 399 first coordination shell 131 first coordination sphere 1 2 Fisher carbene 31 five-coordinate 217 flash photolysis 28, 196, 217 fluidity of the solvent 174 fluorescence 187 fluorophenylcarbene 83 fod see (tris)[2.2-dimethyl-6,6,7,7,8,8,8heptafluoro-3,5-octanedionato]) force-field calculations 352 formamide 323 formation of diamond 434 formation of dioxin 436 Franck-Condon Principle 184 free-radical cyclizations 87 free-radical reactions 84 frequency-dependent friction 99 Friedel-Crafts -acylation 430 - alkylation 379>412, 430 - reactions 399,429 fumaronitrile 55 furan 310, 327. 340 - Diels-Alder reaction with acrylonitrile 5 4

452

I

Index furanobenzocyclophane 56, 58,60 furfuryl fumarates - intramolecular Diels-Alder reaction 272

g gasification 438 GdmCl see guanidinium chloride a-D-glucose - cyclic oligomers of 357 glycerol hiacetate (GTA) 102 green chemistry 398 GTA see glycerol hiacetate guanidinium chloride (GdmC1) 349

h Haber-Bosch process 400 Hartree-Fock 138 Hartree-Fock level 152 Heck coupling 381, 431f Heck reaction 230ff, 235, 279, 286 - intramolecular 271f helicates 363 helium atoms 351 heme 213 hemerythrin 26 hemicarceplex 352 hemicarcerand 351 hemoproteins 213, 216 hepta-coordinated 132 6-heptenyl 87 hetero-Diels-Alderreactions 61, 225, 254, 308, 313 - intramolecular 249, 24G, 278 - solvent effect on the selectivity 257 heterolp~cbond dissociation 44 hexachlorocyclopentadiene 327 hexa-coordinated 132 1,S-hexadiene - Cope rearrangement 72 hexamethylbis(methylene)cyclopentane 48 hexamethyl-Dewarbenzene 75 (2)-1,3,5-hexatriene - electrocydization 75 1-hexene 311 5-hexenyl 87 high-pressure cell 7 high-pressure electrochemical measurements 170, 175 high-pressure electrochemistry 169, 180 high pressure-promoted domino [4 + 2]/ [4 2]/[3 21 cydoaddition 298 high-pressure pulse-radiolysis 34 high-pressure stopped-flow 9 high-spin to low-spin transition 30

+

+

high spin/low spin 195 high spin/low spin equilibria 194 high-to-lowspin 167 Hildebrands theory 310 historical development 3 I H - N M R spectroscopy - pressure dependent 8, 11, 72, 157, 228, 275, 356, 365 'H-NMR titrations 355, 361 homo Diels-Alder reactions 55, 57, 308, 311 - volume data 57 homofuran 55 homolytic bond cleavage 41 homolytic bond dissociations 84 Homer-Wadsworth-Emmons (HWE) reaction 234, 286 host-guest complexes 351, 354,357 host-guest interactions - reaction volumes of 356 - volume data for 357 HWE see Horner-Wadsworth-Emmons reaction hydrated metal ions 131 hydration energies 141 hydration shells 147 hydroboration - rhodium-catalyzed 236 hydroformylation 378, 390,417,433 - of allenes 433 - of hexene 433 - of 1-octene 391 - of olefins 418 - reactions 389, 393 hydrogen atom transfer 88 hydrogen bonding 335, 348 hydrogen bonding effects 336 hydrogen transfer reactions 85 hydrogenation 378, 385, 389,406,412 - catalysts 382 - of cydohexene 384 - of isophorone 384f -homogenous 406 - reaction 383, 386, 409 hydrogen-bonded network 3225 340 hydrolysis - of acetamide 426 - of benzylphenylamine 426 - of esters 426 - reactions 425, 438 hydrophobic effects 336 hydrophobic hydration shell 323 hydrophobic/solvophobicinteractions 323 hydrostatic pressure 310 hydrothermal burner 435

lndex I 4 5 3

hydrothermal crystallization 434 hydrovinylation 382 9-hydroxymethylanthracene 324 I

inclusion reactions profiles for 360 indole 229 - alkaloids 317 inert 131 inner-sphere 164 interchange 12, 134 - mechanisms 197 - rate constant 14 internal free-energy barrier 179 internal pressure 310 internd reorganization energy 178 intramolecular reorganization 161 intrinsic 5, 135 - molar volume 43 -volume 197 - - calculation 44 --definition 44 iodobenzene 230 ion pairing 166, 348 ionic reactions 88 isochromanes 272 isocyanates 228 isoprene 54, 227, 275, 307, 327, 335f, 338 - dimerization 45, 310 isopropenyl methyl ether 257 isoquinolines - synthesis 234 isothermal compressibilities 164 - volume

j Jablonski diagram 184, 199 jahn-Teller 167 Jahn-Teller distortion 136

k ketenes 66 P-ketoesters 317 Kezdy-Swinbourne procedure 249 KGH see Kramers-Grote-Hynes reaction coordinate Kirkwood-Onsagertheory 326 Kramers’ model 98 Kramers-Grote-Hynes reaction coordinate (KGH) 99

I labile 131 p-lactams 66

- bicyclic 293 - tricyclic 293 lanthanide(111) ions 145 lanthanides 144 - catalysis by 312 Ieuco crystal violet - dehydrogenation 85 Lewis acid catalysis 267ff, 287, 30Gff Lewis bases 212f lifetimes 193 ligand exchange 224 - processes 135 - reactions 232 ligand field 186 -bands 186 - states 186 ligand substitution 203 - mechanisms 13 - reactions 12 LIOAS see optoacoustic spectroscopy liquefaction 438 lithium hexamethyldisilazide 270 lithium perchlorate - properties 311 - solution in diethylether (LPDE) 309ff lithium tehafluoroborate 311 lithium hiflate 311 low-spin Fe(I1) 30 low-spinlhigh-spin 166f low-spin/low-spin 176 LPDE see lithium perchlorate, solution in diethylether luminescence 208

m macrocycles 6lff maleic anhydride 54 maleonitrile 55 Marcus approach 174 Marcus theory 169 MD simulations 148 mechanistic changeover 136 mechanistic discrimination 17 Menshutkin reaction 89 mesityl oxide 3 14 metal-carbon bond 15 - formation 34 metal-carbon n bond 20 methacroleine 341 methacrylonitrile 337f p-methoxybenzyl vinyl ether 288, 290, 292, 295 1-methoxybutadiene 313 - hetero-Diels-Alder reaction 267

454

I

Index

4-nitrobenzaldehyde 279 3-[(E)-2-nitro-l-ethenyl] pyridine 295 nitromethane 315, 325 nitronates 287 nitrones 65 nitroprusside 206 nitrosoacetals 287 - tetracyclic 295 - tricyclic 289 nitrostyrenes - resin-bound 286 B-nitrostyrene 288, 298 NMR line-broadening 168 NO, reaction with 216 non-adiabatic 163, 167, 179 non-aqueous media 181 non-aqueous solvents 167, 180 (E)-1,3,8-nonatriene 76 (Z)-1,3,8-nonatriene 77, 247 non-radiativedeactivation 185, 190ff, 199, 205 non-steady state 186 norbornadiene 62, 311 norcaradiene 247, 311 nucleophilicities 142

2-methoxy-l,3-butadiene 298 methoxycyclohexene 295 trcms-l-methoxy-l,3-diene 225 truns-l-methoxy-3-trimethylsilyloxy-l,3butadiene 319 methyl acrylate 288, 327, 336, 341 methyl tert-butylcyclohexylidenebro-moacetate 265 methyl cinnamate 293 methyl crotonate 293 2-methyldihydrofuran 257f N-methylindole 319 methyl methoxymethacrylate 65 2-methylpentane-2,4-diol(MPD) 102 - pressure dependence of viscosity 102 methyl vinyl ketone 279, 310, 315, 325, 327, 335, 340 methylaluminium bis(2,6-diphenylphenoxide) 287 methylaluminium bis(2,6-di-tert-butyl-4methylphenoxide) 287 methylation 411,414 methyl-azulene-1-carboxylate 69 3-methylenecydohexene 67 metmyoglobin 217 micellar catalysis 324 micelles 323, 336 Michael addition see also Michael reaction 265, 273 Michael reaction 315, 325, 342ff microorganisms 349 molar volume 245 - definition 43ff molecular container 351 molecular dynamics 132, 157 - simulation 145, 149 molecular libraries 284 molecular orbital calculations 136 molecular recognition 348 molecular tweezers 354 monoally1 succinate 241 Monte Carlo simulations 132, 157 MPD see 2-methylpentane-2,4-diol Mukaiyama aldol reaction 252, 324 multiactivation 225 multicomponent domino reactions 284 multinuclear coordination species - self-assembled 363 myoglobin 26

oil-contaminatedsoil 440 oligomers, ribbon-type 64 online FT-IR spectroscopy 249, 261 optoacoustic spectroscopy (LIOAS) 116 organometallic complexes 424 organometdlic reactions 431 osmium-catalyzeddihydroxylation 232 outer-spheremechanism 144 outer-sphereself-exchangereactions 180 oxabicyclo[2.2.1lheptene derivatives 272 l-oxa-1,3butadienes 258 - hetero-Diels-Alderreaction 255 oxanorbornadiene 60 oxidation 28,415,430,434 - in supercritical water 437 - of alkylarenes 431 - of methane 431 - of organic compounds 423 oxidative addition 32 oxygenation 26 oxyhemerythrin 26 oxymyoglobin 26

n

P

Narasaka catalyst 275, 278 neutron diffraction 138 nine-coordinateed 132, 144

- of

0

packing coefficient 52f, 71, 75, 83, 234 cyclohexene 52 - definition 43ff

fndex I 4 5 5

(+)-palasonin 240 palladium catalyzed reactions 230 [2 2lparacyclophane 60 partial molar volume 170, 245, 258. 353 - definition 43ff - estimation 43 -ofCO 30 partial volumes 81 Pauson-Khand reaction 382 PBH see pinacol borane Pd(Oac)2/PPh3 232f, 272 peak separation 171 pericyclic rearrangements 71 per-methylation 412 pharmaceuticals 383, 386 phase boundaries 409 cis phenyl-aminocyclohexene 286 phenylhalogenocarbenes 70 N-phenylmaleic imide 54, 291, 307, 324, 336 1-phenyl-2-nitropropene 293 phenyl nonaflat 233, 279 1-phenylphosphole 65 phenyl vinylsulphone 292,343 phosphines - catalysis by 320 phospholipids 349 phosphorescence 190 photoacoustic calorimetry 193 photoactivity 205 photoaquation 197, 200f, 203 photochemical 184 - cycloadditions - - pressure effect 265ff photochemistry 201,203 photoexcitation 196 photolabilization 200, 213 photolysis experiments 186 photophysical 184 photophysics 203 photoreactions 184, 196 photoredox 205 photosolvolysis 198 photosubstitution 196, 201f, 204f PhPBr2 70 PhPC12 70 physical activation 306 physicochemical activation 306 pill-box cuvette 6 pill-box optical cell 7 pinacolborane (PBH) 236 (-)-5,6-pinene bipyridine 363 polarization of water 145 porphyrin complexes 30, 187

+

potential energy surface 100 precursor complex 14, 207 precursors 163 pressure - coefficients 103 - dependence of the volume of activation 4 - kinetic effects 97 -units of 41 - effect - - chemical equilibria 42 - - conductivity 41 41 --density -- dielectric constant 41 - - melting points 41ff - - rates of reaction 42 - - solubility 41 - - solvent dependence 52 --viscosity 41 - generating system 6 - induced freezing 188 -tuning 161 propionaldehyde 343 protein folding 348 proteins 349 proton transfer 85 pulse-radiolysis 32 -techniques 23 purification 400 pyridiminiumoxides 84 pyrolysis 424 2-pyrones see y-pyrones y-pyrones - cycloaddition 242, 314

4 quadricyclane 62 quantum chemical calculations 356 quantum mechanical calculations 48 quantum yields 189 quenching 207ff, 213 oquinodimethane - dimerization 48 quinone oxidations 85 r radiation-induced 32 radiative 185 -decay 189 radical pairs -caged 85 - solvent-separated 85 rate constants - viscosity dependence 98

reaction volume 4, 42 42ff reactive intermediates 186, 213 reactor concepts 435 redox quenching 21 1 reduction reactions 431 reductive elimination 32 reductive quenching 209 regioselectivity - Diels-Alder reactions 249 relaxation lifetime 196 replacement of a water molecule 131 retro Diels-Alder reactions 54 reveromycin A 241 reversible substitution 14 ribonuclease A 351 ring closing metathesis 382f ring size - pressure effect 81 rotaxanes 89f - estimation

5

salting-in effect 336 salting-out agents 328 saltingout effects 323 scandium triflates 313 SDS see sodium dodecyl sulfate second coordination shell 155 second coordination sphere 141 second hydration shell 150 selectivity - influence of pressure on 239, 245 self-assembly 363 self-assembly/self-organization 348 self-exchange 164, 168 - reactions 22, 162, 167, 169, 173, 179f separation 400 - of caffeine 401 SET see electron transfer, single sharpless epoxidation 417 shear viscosities 103 1,Cshifts - benzhydrylic 84 -benzylic 84 sigmatropic[ 1,s) hydrogen shift 77 1,5-sigmatropicrearrangement 247 3,3-sigmatropicrearrangement 246 sigmatropic [3,3] shifts 71 silver helicate - Connolly's molecular surface of [Ag&]4t and [Ag6L6I6+ 366 -crystal structure of [Ag6klG+ 364 singlet carbene 70 sodium dodecyl sulfate (SDS) 336

solid phase chemistry 284 solid phase synthesis 301 solubility 422 solvation changes 197 solvation sheat 162 solvation shell rearrangement 97 solvational contributions 5 solvent dynamical effects 179f solvent dynamics 174f solvent effects 162, 252 solvent electrostriction 208 - see also electrostriction solvent exchange reaction 131 solvent friction 174, 176 solvent parameter 31 solvent properties 371 solvent viscosity 214 solvolysis 13 solvophobic activation 321 solvophobic interactions 328, 335 Spanish Fly 240 spectator ligand 18 spin equilibrium 173 spin isomers 196 spin multiplicity 185 square-planar complexes 19 steam (reforming) 438 steric effects 136 steric hindrance 186 256, 264 stilbene 99 stopped-flow instruments 7 stopped-flowkinetics - pressure dependent 359 strong coupling 190, 201 styrene 288, 301, 327, 336 subcritical 426 sucrose - inversion of 305 supercritical 426 - carbon dioxide 371 - chromatography 398 - fluids 10, 371, 398,400, 405 - - Chromatography 403 - - extraction 398,400,402 -hydrogenation 410 -phase 371 -water 422, 4245 434, 436, 438 -- oxidation 415,435 superexo 28 supermolecules 348 supramoledar chemistry 61, 348 supramolecular complexes 349, 356 surface tension 377 synthetic receptors 348, 354

lndex I 4 5 7 L

tandem reactions 65 tank reactor 435 TCN E see tetracyanoethene temperature-jump technique 9 tendamistat 349 - folding - - volume profiles 350 terephthalic aldehyde 356 tetracyanoethene (TCNE) 67 - reaction with 1.1-dimethylbutadiene 52 tetradecyltrimethylammonium bromide 336 tetrahydropyndines 270 tetraline - dehydrogenation 85 tetramethylethene 83 1,3,4,6-tetraphenyl-l,5-hexadiene - racemization and diastereomerization 73 I-thia-1,3-butadiene 260 threoninal 277 thymoquinone 85 titanium tetrachloride 269 TOF see turnover frequency toluquinone 327, 338 TON see turnover number (S)-p-tolylsulfinyltrialkoxycarbonyl ethene 274 transformations - chemo-and regioselective 245 - diastereoselective 253 - enantioselective 275 transition 186 - metal ions 142 - metal-catalyzed reactions 223 - state 24, 29, 155 --early 36 --late 36 - - location of 25 --theory (TST) 4,43,48,97, 174, 196, 305 - structures - - bicyclic 247 -- endo 244, 256, 264, 271 - - endo-E-syn 255, 259, 261 - - endo-2-anti 259f - - ex0 244, 256, 264 - - exo-E-anti 256ff -- exo-2-syn 255, 259f - - monocyclic 247 - - pericychc 48 transpiring wall reactor 435 triactivation 343 1.2,4-tricyanobenzene 59, 75 2,3,5-tricyano-Dewar-benzene 59, 75

triflate 313 trifluoromethanesulfonato 313 trimethylsilylacetylene 65 trimethylsilylcyanide 229 tropone 68ff TS'T see transition state theory tube reactor 436 turnover frequency (TOF) 223 turnover number (TON) 223, 233 U

Ugi reaction 286 unimolecular photoreactions 212 urea 328 UVjVIS spectroscopy 6 V

valence-bond isomerization 246 6-valerolactam 320 van der Waals distance 41 van der Waals interactions 357, 363 van der Waals radii 44, 274 van der Waals volumes 43, 71, 79ff vibronic relaxation 184 vinyl acetate 292 1-vinylcyclobutene 70 4-vinylcydohexene 67 vinylsulfonic esters - intramolecular cycloaddition 265 viscosity 98 viscous medium 214 void volume 44,81 voltammetric techniques 24 volume changes 154 volume of activation see activation volume volume of association 356 volume of reaction see reaction volume volume profile 4, 14f, 17, 24ff, 29, 33ff, 140, 150, 198, 201, 218,252 W

Wagner-Meenvein rearrangement 277 Wang resin 285,288, 300 wastes treatment plants 436 water - cohesion energy density 323 - exchange 34, 136ff, 144f, 152f, 156 - - on metal ions 131 - -gas shift 433 - - reaction 424,439 weak coupling 190 Wilkinson' catalyst 236 Wittig reaction 262 Wittig reagent 242

458

I

Index X

1,3-xylyl-18-crown-5 354

Y Yb(fod)3 269 ytterbium triflate - catalysis by 229, 318

z zinc chloride 225, 228, 269 ZnClz see zinc chloride zwitterionic intermediates 31, 50 zwitterions 253

Metal clusters in chemistry

Read more

Metal Clusters in Chemistry

Read more

Advances in Metal and Semiconductor Clusters, Volume 5 (Advances in Metal and Semiconductor Clusters)

Read more

Inorganic Chemistry Metal Carbonyl Chemistry

Read more

Surface Diffusion: Metals, Metal Atoms, and Clusters

Read more

Surface Diffusion: Metals, Metal Atoms, and Clusters

Read more

Organotransition Metal Chemistry

Read more

Organotransition Metal Chemistry

Read more

Metal Amide Chemistry

Read more

Advances in Metal and Semiconductor Clusters (Vol. 4): Cluster Materials

Read more

Spectroscopy and Structure (Advances in Metal and Semiconductor Clusters)

Read more

Landmarks in Organo-Transition Metal Chemistry

Read more

Transition Metal Chemistry: The Valence Shell in d-Block Chemistry

Read more

The Chemistry of Metal Alkoxides

Read more

The Chemistry of Metal CVD

Read more

Chemical Reactions in Clusters (Topics in Physical Chemistry)

Read more

Transition Metal Chemistry: The Valence Shell in d-Block Chemistry

Read more

Metal Oxide Chemistry and Synthesis Jolivet

Read more

Chemistry of Transition Metal Carbides and Nitrides

Read more

The Chemistry of Metal Enolates (Chemistry of Functional Groups)

Read more

Metal-DNA Chemistry (Acs Symposium Series)

Read more

Metal Catalyzed Cascade Reactions (Topics in Organometallic Chemistry, Volume 19)

Read more

Catalysis in C1 Chemistry (Catalysis by Metal Complexes)

Read more

Merging Processes in Galaxy Clusters

Read more

Molecular clusters

Read more

Landmarks in Organo-Transition Metal Chemistry: A Personal View (Profiles in Inorganic Chemistry)

Read more

Metal Ions in Fungi

Read more

Metal

Read more

Metal Nanoparticles in Microbiology

Read more

Complexes, Clusters and Crystal Chemistry (Structure and Bonding, Volume 79)

Read more

Recommend Documents

Metal clusters in chemistry

Matthias Driess, Heinrich No¨th (Eds.) Molecular Clusters of the Main Group Elements Further Titles of Interest: G. S...

Metal Clusters in Chemistry

Metal Clusters in Chemistry Edited by P. Braunstein L. A. Oro P. R. Raithby Further Titles of Interest A. Togni, R....

Advances in Metal and Semiconductor Clusters, Volume 5 (Advances in Metal and Semiconductor Clusters)

Advances in Metal and Semiconductor Clusters Volumes 1–4 published by JAI Press Inc., available from Elsevier Science. ...

Inorganic Chemistry Metal Carbonyl Chemistry

Surface Diffusion: Metals, Metal Atoms, and Clusters

Surface Diffusion Metals, Metal Atoms, and Clusters For the ﬁrst time, this book unites the theory, experimental techni...

Surface Diffusion: Metals, Metal Atoms, and Clusters

Surface Diffusion Metals, Metal Atoms, and Clusters For the ﬁrst time, this book unites the theory, experimental techni...

Organotransition Metal Chemistry

Organotransition Metal Chemistry

Cover images 0Murray Robertsonlvisual elements 1998-99, taken from the 109 Visual Elements Periodic Table, available at...

Metal Amide Chemistry

Metal Amide Chemistry Metal Amide Chemistry Michael Lappert, Andrey Protchenko, Philip Power and Alexandra Seeber © 20...

Advances in Metal and Semiconductor Clusters (Vol. 4): Cluster Materials

ADVANCES IN METAL A N D SEMICONDUCTOR CLUSTERS Volume4 9 1998 CLUSTER MATERIALS This Page Intentionally Left Blank...